Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
/*-
|
2000-01-03 10:26:59 +00:00
|
|
|
|
* Copyright (c) 1998,1999,2000 S<EFBFBD>ren Schmidt
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
* All rights reserved.
|
|
|
|
|
|
*
|
|
|
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
|
|
* modification, are permitted provided that the following conditions
|
|
|
|
|
|
* are met:
|
|
|
|
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
|
|
|
|
* notice, this list of conditions and the following disclaimer,
|
|
|
|
|
|
* without modification, immediately at the beginning of the file.
|
|
|
|
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
|
|
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
|
|
|
|
* documentation and/or other materials provided with the distribution.
|
|
|
|
|
|
* 3. The name of the author may not be used to endorse or promote products
|
|
|
|
|
|
* derived from this software without specific prior written permission.
|
|
|
|
|
|
*
|
|
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
|
|
|
|
|
|
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
|
|
|
|
|
|
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
|
|
|
|
|
|
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
|
|
|
|
|
|
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
|
|
|
|
|
|
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
|
|
|
|
|
|
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
|
|
|
|
|
|
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
|
|
|
|
|
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
|
|
|
|
|
|
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
|
|
|
|
*
|
1999-08-28 01:08:13 +00:00
|
|
|
|
* $FreeBSD$
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
|
|
#include "ata.h"
|
1999-03-03 21:10:29 +00:00
|
|
|
|
#include "isa.h"
|
2000-02-18 20:57:33 +00:00
|
|
|
|
#include "card.h"
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
#include "pci.h"
|
1999-03-03 21:10:29 +00:00
|
|
|
|
#include "atadisk.h"
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#include "atapicd.h"
|
|
|
|
|
|
#include "atapifd.h"
|
|
|
|
|
|
#include "atapist.h"
|
1999-04-18 20:48:15 +00:00
|
|
|
|
#include "opt_global.h"
|
1999-10-09 19:57:13 +00:00
|
|
|
|
#include "opt_ata.h"
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
#include <sys/param.h>
|
|
|
|
|
|
#include <sys/systm.h>
|
|
|
|
|
|
#include <sys/kernel.h>
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#include <sys/disk.h>
|
1999-04-18 20:48:15 +00:00
|
|
|
|
#include <sys/module.h>
|
|
|
|
|
|
#include <sys/bus.h>
|
2000-05-05 09:59:14 +00:00
|
|
|
|
#include <sys/bio.h>
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
#include <sys/malloc.h>
|
|
|
|
|
|
#include <sys/devicestat.h>
|
2000-01-27 19:00:51 +00:00
|
|
|
|
#include <sys/sysctl.h>
|
2000-01-24 20:45:24 +00:00
|
|
|
|
#include <machine/stdarg.h>
|
1999-04-18 20:48:15 +00:00
|
|
|
|
#include <machine/resource.h>
|
|
|
|
|
|
#include <machine/bus.h>
|
|
|
|
|
|
#include <sys/rman.h>
|
1999-05-17 15:58:47 +00:00
|
|
|
|
#if NPCI > 0
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
#include <pci/pcivar.h>
|
|
|
|
|
|
#include <pci/pcireg.h>
|
1999-05-17 15:58:47 +00:00
|
|
|
|
#endif
|
1999-04-18 20:48:15 +00:00
|
|
|
|
#include <isa/isavar.h>
|
1999-04-16 21:22:55 +00:00
|
|
|
|
#include <isa/isareg.h>
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#include <machine/clock.h>
|
2000-02-18 20:57:33 +00:00
|
|
|
|
#ifdef __alpha__
|
|
|
|
|
|
#include <machine/md_var.h>
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#endif
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
#include <dev/ata/ata-all.h>
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#include <dev/ata/ata-disk.h>
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
#include <dev/ata/atapi-all.h>
|
|
|
|
|
|
|
|
|
|
|
|
/* misc defines */
|
2000-02-18 20:57:33 +00:00
|
|
|
|
#define IOMASK 0xfffffffc
|
|
|
|
|
|
#define ATA_IOADDR_RID 0
|
2000-02-25 09:48:23 +00:00
|
|
|
|
#define ATA_ALTADDR_RID 1
|
2000-02-18 20:57:33 +00:00
|
|
|
|
#define ATA_BMADDR_RID 2
|
2000-10-04 08:28:37 +00:00
|
|
|
|
#define ATA_MASTERDEV(dev) ((pci_get_progif(dev) & 0x80) && \
|
|
|
|
|
|
(pci_get_progif(dev) & 0x05) != 0x05)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
|
|
|
|
|
/* prototypes */
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static int ata_probe(device_t);
|
|
|
|
|
|
static int ata_attach(device_t);
|
|
|
|
|
|
static int ata_detach(device_t);
|
|
|
|
|
|
static int ata_resume(device_t);
|
|
|
|
|
|
static void ata_boot_attach(void);
|
|
|
|
|
|
static void ata_intr(void *);
|
2000-09-19 11:08:39 +00:00
|
|
|
|
static int ata_getparam(struct ata_softc *, int, u_int8_t);
|
|
|
|
|
|
static int ata_service(struct ata_softc *);
|
|
|
|
|
|
static char *active2str(int);
|
|
|
|
|
|
static void bswap(int8_t *, int);
|
|
|
|
|
|
static void btrim(int8_t *, int);
|
|
|
|
|
|
static void bpack(int8_t *, int8_t *, int);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
1999-10-09 19:57:13 +00:00
|
|
|
|
/* local vars */
|
1999-04-18 20:48:15 +00:00
|
|
|
|
static devclass_t ata_devclass;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static devclass_t ata_pci_devclass;
|
|
|
|
|
|
static struct intr_config_hook *ata_delayed_attach = NULL;
|
|
|
|
|
|
static char ata_conf[256];
|
1999-10-09 19:57:13 +00:00
|
|
|
|
MALLOC_DEFINE(M_ATA, "ATA generic", "ATA driver generic layer");
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
1999-04-18 20:48:15 +00:00
|
|
|
|
#if NISA > 0
|
1999-09-01 20:53:43 +00:00
|
|
|
|
static struct isa_pnp_id ata_ids[] = {
|
1999-10-09 19:57:13 +00:00
|
|
|
|
{0x0006d041, "Generic ESDI/IDE/ATA controller"}, /* PNP0600 */
|
|
|
|
|
|
{0x0106d041, "Plus Hardcard II"}, /* PNP0601 */
|
|
|
|
|
|
{0x0206d041, "Plus Hardcard IIXL/EZ"}, /* PNP0602 */
|
|
|
|
|
|
{0x0306d041, "Generic ATA"}, /* PNP0603 */
|
1999-09-01 20:53:43 +00:00
|
|
|
|
{0}
|
|
|
|
|
|
};
|
1999-04-18 20:48:15 +00:00
|
|
|
|
|
|
|
|
|
|
static int
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_isa_probe(device_t dev)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
2000-02-18 20:57:33 +00:00
|
|
|
|
struct ata_softc *scp = device_get_softc(dev);
|
1999-04-18 20:48:15 +00:00
|
|
|
|
struct resource *port;
|
|
|
|
|
|
int rid;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
u_long tmp;
|
1999-04-18 20:48:15 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* check isapnp ids */
|
1999-09-01 20:53:43 +00:00
|
|
|
|
if (ISA_PNP_PROBE(device_get_parent(dev), dev, ata_ids) == ENXIO)
|
2000-01-24 20:45:24 +00:00
|
|
|
|
return ENXIO;
|
1999-09-01 20:53:43 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* allocate the port range */
|
|
|
|
|
|
rid = ATA_IOADDR_RID;
|
|
|
|
|
|
port = bus_alloc_resource(dev, SYS_RES_IOPORT, &rid, 0, ~0,
|
|
|
|
|
|
ATA_IOSIZE, RF_ACTIVE);
|
1999-04-18 20:48:15 +00:00
|
|
|
|
if (!port)
|
2000-01-24 20:45:24 +00:00
|
|
|
|
return ENOMEM;
|
1999-04-18 20:48:15 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* alloctate the altport range */
|
|
|
|
|
|
if (bus_get_resource(dev, SYS_RES_IOPORT, 1, &tmp, &tmp)) {
|
|
|
|
|
|
bus_set_resource(dev, SYS_RES_IOPORT, 1,
|
2000-09-24 18:19:43 +00:00
|
|
|
|
rman_get_start(port) + ATA_ALTOFFSET,
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ATA_ALTIOSIZE);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
bus_release_resource(dev, SYS_RES_IOPORT, 0, port);
|
2000-10-13 11:21:27 +00:00
|
|
|
|
scp->channel = 0;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
scp->flags |= ATA_USE_16BIT;
|
|
|
|
|
|
return ata_probe(dev);
|
|
|
|
|
|
}
|
1999-04-18 20:48:15 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static device_method_t ata_isa_methods[] = {
|
|
|
|
|
|
/* device interface */
|
|
|
|
|
|
DEVMETHOD(device_probe, ata_isa_probe),
|
|
|
|
|
|
DEVMETHOD(device_attach, ata_attach),
|
|
|
|
|
|
DEVMETHOD(device_resume, ata_resume),
|
|
|
|
|
|
{ 0, 0 }
|
|
|
|
|
|
};
|
1999-04-18 20:48:15 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static driver_t ata_isa_driver = {
|
|
|
|
|
|
"ata",
|
|
|
|
|
|
ata_isa_methods,
|
|
|
|
|
|
sizeof(struct ata_softc),
|
|
|
|
|
|
};
|
1999-04-18 20:48:15 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
DRIVER_MODULE(ata, isa, ata_isa_driver, ata_devclass, 0, 0);
|
|
|
|
|
|
#endif
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
#if NCARD > 0
|
1999-04-18 20:48:15 +00:00
|
|
|
|
static int
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_pccard_probe(device_t dev)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
2000-02-18 20:57:33 +00:00
|
|
|
|
struct ata_softc *scp = device_get_softc(dev);
|
1999-04-18 20:48:15 +00:00
|
|
|
|
struct resource *port;
|
2000-02-25 09:48:23 +00:00
|
|
|
|
int rid, len;
|
1999-04-18 20:48:15 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* allocate the port range */
|
|
|
|
|
|
rid = ATA_IOADDR_RID;
|
2000-02-25 09:48:23 +00:00
|
|
|
|
len = bus_get_resource_count(dev, SYS_RES_IOPORT, rid);
|
|
|
|
|
|
port = bus_alloc_resource(dev, SYS_RES_IOPORT, &rid, 0, ~0, len, RF_ACTIVE);
|
1999-04-18 20:48:15 +00:00
|
|
|
|
if (!port)
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return ENOMEM;
|
1999-04-18 20:48:15 +00:00
|
|
|
|
|
2000-02-25 09:48:23 +00:00
|
|
|
|
/*
|
|
|
|
|
|
* if we got more than the default ATA_IOSIZE ports, this is likely
|
|
|
|
|
|
* a pccard system where the altio ports are located just after the
|
|
|
|
|
|
* normal io ports, so no need to allocate them.
|
|
|
|
|
|
*/
|
|
|
|
|
|
if (len <= ATA_IOSIZE) {
|
|
|
|
|
|
bus_set_resource(dev, SYS_RES_IOPORT, ATA_ALTADDR_RID,
|
2000-09-24 18:19:43 +00:00
|
|
|
|
rman_get_start(port) + ATA_ALTOFFSET, ATA_ALTIOSIZE);
|
1999-04-18 20:48:15 +00:00
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
bus_release_resource(dev, SYS_RES_IOPORT, 0, port);
|
2000-10-13 11:21:27 +00:00
|
|
|
|
scp->channel = 0;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
scp->flags |= ATA_USE_16BIT;
|
|
|
|
|
|
return ata_probe(dev);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static device_method_t ata_pccard_methods[] = {
|
|
|
|
|
|
/* device interface */
|
|
|
|
|
|
DEVMETHOD(device_probe, ata_pccard_probe),
|
|
|
|
|
|
DEVMETHOD(device_attach, ata_attach),
|
|
|
|
|
|
DEVMETHOD(device_detach, ata_detach),
|
1999-04-18 20:48:15 +00:00
|
|
|
|
{ 0, 0 }
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
};
|
|
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static driver_t ata_pccard_driver = {
|
1999-05-20 09:12:06 +00:00
|
|
|
|
"ata",
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_pccard_methods,
|
|
|
|
|
|
sizeof(struct ata_softc),
|
1999-04-18 20:48:15 +00:00
|
|
|
|
};
|
|
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
DRIVER_MODULE(ata, pccard, ata_pccard_driver, ata_devclass, 0, 0);
|
1999-04-18 20:48:15 +00:00
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
|
|
#if NPCI > 0
|
2000-02-18 20:57:33 +00:00
|
|
|
|
struct ata_pci_softc {
|
|
|
|
|
|
struct resource *bmio;
|
|
|
|
|
|
struct resource bmio_1;
|
|
|
|
|
|
struct resource bmio_2;
|
|
|
|
|
|
struct resource *irq;
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int irqcnt;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
};
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int
|
|
|
|
|
|
ata_find_dev(device_t dev, u_int32_t type, u_int32_t revid)
|
2000-02-18 20:57:33 +00:00
|
|
|
|
{
|
|
|
|
|
|
device_t *children, child;
|
|
|
|
|
|
int nchildren, i;
|
|
|
|
|
|
|
|
|
|
|
|
if (device_get_children(device_get_parent(dev), &children, &nchildren))
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < nchildren; i++) {
|
|
|
|
|
|
child = children[i];
|
|
|
|
|
|
|
|
|
|
|
|
/* check that it's on the same silicon and the device we want */
|
|
|
|
|
|
if (pci_get_slot(dev) == pci_get_slot(child) &&
|
|
|
|
|
|
pci_get_vendor(child) == (type & 0xffff) &&
|
2000-04-09 15:08:27 +00:00
|
|
|
|
pci_get_device(child) == ((type & 0xffff0000) >> 16) &&
|
|
|
|
|
|
pci_get_revid(child) >= revid) {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
free(children, M_TEMP);
|
|
|
|
|
|
return 1;
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
free(children, M_TEMP);
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
static const char *
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_pci_match(device_t dev)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
1999-10-16 09:00:49 +00:00
|
|
|
|
if (pci_get_class(dev) != PCIC_STORAGE)
|
1999-10-13 18:56:49 +00:00
|
|
|
|
return NULL;
|
|
|
|
|
|
|
1999-10-07 02:17:21 +00:00
|
|
|
|
switch (pci_get_devid(dev)) {
|
1999-10-09 19:57:13 +00:00
|
|
|
|
/* supported chipsets */
|
1999-10-07 02:17:21 +00:00
|
|
|
|
case 0x12308086:
|
1999-12-07 22:07:18 +00:00
|
|
|
|
return "Intel PIIX ATA controller";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
1999-10-07 02:17:21 +00:00
|
|
|
|
case 0x70108086:
|
1999-12-07 22:07:18 +00:00
|
|
|
|
return "Intel PIIX3 ATA controller";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
1999-10-07 02:17:21 +00:00
|
|
|
|
case 0x71118086:
|
1999-12-21 20:18:56 +00:00
|
|
|
|
case 0x71998086:
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return "Intel PIIX4 ATA33 controller";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
2000-08-06 19:51:58 +00:00
|
|
|
|
case 0x24218086:
|
|
|
|
|
|
return "Intel ICH0 ATA33 controller";
|
|
|
|
|
|
|
2000-01-17 02:04:19 +00:00
|
|
|
|
case 0x24118086:
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return "Intel ICH ATA66 controller";
|
2000-01-17 02:04:19 +00:00
|
|
|
|
|
2000-08-06 19:51:58 +00:00
|
|
|
|
case 0x244b8086:
|
|
|
|
|
|
return "Intel ICH2 ATA100 controller";
|
2000-01-17 02:04:19 +00:00
|
|
|
|
|
1999-10-09 19:57:13 +00:00
|
|
|
|
case 0x522910b9:
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return "AcerLabs Aladdin ATA33 controller";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
case 0x05711106:
|
2000-04-09 15:08:27 +00:00
|
|
|
|
if (ata_find_dev(dev, 0x05861106, 0))
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return "VIA 82C586 ATA33 controller";
|
2000-04-09 15:08:27 +00:00
|
|
|
|
if (ata_find_dev(dev, 0x05961106, 0x12))
|
2000-08-16 07:09:04 +00:00
|
|
|
|
return "VIA 82C596 ATA66 controller";
|
2000-04-09 15:08:27 +00:00
|
|
|
|
if (ata_find_dev(dev, 0x05961106, 0))
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return "VIA 82C596 ATA33 controller";
|
2000-04-09 15:08:27 +00:00
|
|
|
|
if (ata_find_dev(dev, 0x06861106, 0))
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return "VIA 82C686 ATA66 controller";
|
1999-12-07 22:07:18 +00:00
|
|
|
|
return "VIA Apollo ATA controller";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
1999-12-13 10:19:39 +00:00
|
|
|
|
case 0x55131039:
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return "SiS 5591 ATA33 controller";
|
|
|
|
|
|
|
2000-10-03 13:12:36 +00:00
|
|
|
|
case 0x06491095:
|
|
|
|
|
|
return "CMD 649 ATA100 controller";
|
|
|
|
|
|
|
|
|
|
|
|
case 0x06481095:
|
|
|
|
|
|
return "CMD 648 ATA66 controller";
|
|
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
case 0x06461095:
|
|
|
|
|
|
return "CMD 646 ATA controller";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
2000-02-25 09:48:23 +00:00
|
|
|
|
case 0xc6931080:
|
|
|
|
|
|
if (pci_get_subclass(dev) == PCIS_STORAGE_IDE)
|
|
|
|
|
|
return "Cypress 82C693 ATA controller";
|
|
|
|
|
|
break;
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
case 0x01021078:
|
|
|
|
|
|
return "Cyrix 5530 ATA33 controller";
|
|
|
|
|
|
|
1999-12-21 20:18:56 +00:00
|
|
|
|
case 0x74091022:
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return "AMD 756 ATA66 controller";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
2000-10-03 13:12:36 +00:00
|
|
|
|
case 0x02111166:
|
|
|
|
|
|
return "ServerWorks ROSB4 ATA33 controller";
|
|
|
|
|
|
|
1999-12-13 10:19:39 +00:00
|
|
|
|
case 0x4d33105a:
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return "Promise ATA33 controller";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
1999-12-13 10:19:39 +00:00
|
|
|
|
case 0x4d38105a:
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return "Promise ATA66 controller";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
case 0x0d30105a:
|
2000-08-06 19:51:58 +00:00
|
|
|
|
case 0x4d30105a:
|
|
|
|
|
|
return "Promise ATA100 controller";
|
|
|
|
|
|
|
1999-12-13 10:19:39 +00:00
|
|
|
|
case 0x00041103:
|
2000-08-06 19:51:58 +00:00
|
|
|
|
switch (pci_get_revid(dev)) {
|
|
|
|
|
|
case 0x00:
|
|
|
|
|
|
case 0x01:
|
|
|
|
|
|
return "HighPoint HPT366 ATA66 controller";
|
|
|
|
|
|
case 0x02:
|
|
|
|
|
|
return "HighPoint HPT368 ATA66 controller";
|
|
|
|
|
|
case 0x03:
|
|
|
|
|
|
case 0x04:
|
|
|
|
|
|
return "HighPoint HPT370 ATA100 controller";
|
|
|
|
|
|
default:
|
|
|
|
|
|
return "Unknown revision HighPoint ATA controller";
|
|
|
|
|
|
}
|
1999-10-09 19:57:13 +00:00
|
|
|
|
|
|
|
|
|
|
/* unsupported but known chipsets, generic DMA only */
|
2000-02-18 20:57:33 +00:00
|
|
|
|
case 0x10001042:
|
|
|
|
|
|
case 0x10011042:
|
|
|
|
|
|
return "RZ 100? ATA controller !WARNING! buggy chip data loss possible";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
case 0x06401095:
|
|
|
|
|
|
return "CMD 640 ATA controller !WARNING! buggy chip data loss possible";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* unknown chipsets, try generic DMA if it seems possible */
|
1999-10-07 02:17:21 +00:00
|
|
|
|
default:
|
|
|
|
|
|
if (pci_get_class(dev) == PCIC_STORAGE &&
|
1999-10-09 19:57:13 +00:00
|
|
|
|
(pci_get_subclass(dev) == PCIS_STORAGE_IDE))
|
2000-05-28 07:51:23 +00:00
|
|
|
|
return "Generic PCI ATA controller";
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
return NULL;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
1999-04-18 20:48:15 +00:00
|
|
|
|
static int
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_pci_probe(device_t dev)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
2000-02-18 20:57:33 +00:00
|
|
|
|
const char *desc = ata_pci_match(dev);
|
1999-12-07 22:07:18 +00:00
|
|
|
|
|
1999-04-18 20:48:15 +00:00
|
|
|
|
if (desc) {
|
|
|
|
|
|
device_set_desc(dev, desc);
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
}
|
|
|
|
|
|
else
|
|
|
|
|
|
return ENXIO;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_pci_add_child(device_t dev, int unit)
|
1999-04-18 20:48:15 +00:00
|
|
|
|
{
|
2000-02-18 20:57:33 +00:00
|
|
|
|
device_t child;
|
|
|
|
|
|
|
|
|
|
|
|
/* check if this is located at one of the std addresses */
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev)) {
|
2000-02-25 09:48:23 +00:00
|
|
|
|
if (!(child = device_add_child(dev, "ata", unit)))
|
|
|
|
|
|
return ENOMEM;
|
|
|
|
|
|
}
|
|
|
|
|
|
else {
|
|
|
|
|
|
if (!(child = device_add_child(dev, "ata", 2)))
|
|
|
|
|
|
return ENOMEM;
|
|
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return 0;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
|
ata_pci_attach(device_t dev)
|
|
|
|
|
|
{
|
|
|
|
|
|
struct ata_pci_softc *sc = device_get_softc(dev);
|
1999-04-18 20:48:15 +00:00
|
|
|
|
u_int8_t class, subclass;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
u_int32_t type, cmd;
|
|
|
|
|
|
int rid;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
|
|
|
|
|
/* set up vendor-specific stuff */
|
1999-04-18 20:48:15 +00:00
|
|
|
|
type = pci_get_devid(dev);
|
|
|
|
|
|
class = pci_get_class(dev);
|
|
|
|
|
|
subclass = pci_get_subclass(dev);
|
|
|
|
|
|
cmd = pci_read_config(dev, PCIR_COMMAND, 4);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-10-03 13:12:36 +00:00
|
|
|
|
/* is busmastering supported ? */
|
|
|
|
|
|
if ((cmd & (PCIM_CMD_PORTEN | PCIM_CMD_BUSMASTEREN)) ==
|
|
|
|
|
|
(PCIM_CMD_PORTEN | PCIM_CMD_BUSMASTEREN)) {
|
|
|
|
|
|
|
|
|
|
|
|
/* is there a valid port range to connect to ? */
|
|
|
|
|
|
rid = 0x20;
|
|
|
|
|
|
sc->bmio = bus_alloc_resource(dev, SYS_RES_IOPORT, &rid,
|
|
|
|
|
|
0, ~0, 1, RF_ACTIVE);
|
2000-10-13 11:21:27 +00:00
|
|
|
|
if (!sc->bmio)
|
|
|
|
|
|
device_printf(dev, "Busmastering DMA not supported\n");
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
1999-11-08 21:36:00 +00:00
|
|
|
|
|
1999-11-26 19:24:26 +00:00
|
|
|
|
/* do extra chipset specific setups */
|
|
|
|
|
|
switch (type) {
|
2000-01-03 10:26:59 +00:00
|
|
|
|
case 0x522910b9: /* Aladdin need to activate the ATAPI FIFO */
|
1999-11-08 21:36:00 +00:00
|
|
|
|
pci_write_config(dev, 0x53,
|
|
|
|
|
|
(pci_read_config(dev, 0x53, 1) & ~0x01) | 0x02, 1);
|
1999-11-26 19:24:26 +00:00
|
|
|
|
break;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
case 0x4d38105a: /* Promise 66 & 100 need their clock changed */
|
|
|
|
|
|
case 0x4d30105a:
|
|
|
|
|
|
case 0x0d30105a:
|
2000-02-18 20:57:33 +00:00
|
|
|
|
outb(rman_get_start(sc->bmio) + 0x11,
|
|
|
|
|
|
inb(rman_get_start(sc->bmio) + 0x11) | 0x0a);
|
2000-02-04 10:20:22 +00:00
|
|
|
|
/* FALLTHROUGH */
|
1999-11-08 21:36:00 +00:00
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
case 0x4d33105a: /* Promise (all) need burst mode to be turned on */
|
2000-02-18 20:57:33 +00:00
|
|
|
|
outb(rman_get_start(sc->bmio) + 0x1f,
|
|
|
|
|
|
inb(rman_get_start(sc->bmio) + 0x1f) | 0x01);
|
1999-11-26 19:24:26 +00:00
|
|
|
|
break;
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
case 0x00041103: /* HighPoint */
|
2000-08-06 19:51:58 +00:00
|
|
|
|
switch (pci_get_revid(dev)) {
|
|
|
|
|
|
case 0x00:
|
|
|
|
|
|
case 0x01:
|
2000-09-19 11:08:39 +00:00
|
|
|
|
/* turn off interrupt prediction */
|
2000-08-06 19:51:58 +00:00
|
|
|
|
pci_write_config(dev, 0x51,
|
|
|
|
|
|
(pci_read_config(dev, 0x51, 1) & ~0x80), 1);
|
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
|
|
case 0x02:
|
|
|
|
|
|
case 0x03:
|
|
|
|
|
|
case 0x04:
|
2000-09-19 11:08:39 +00:00
|
|
|
|
/* turn off interrupt prediction */
|
2000-08-06 19:51:58 +00:00
|
|
|
|
pci_write_config(dev, 0x51,
|
|
|
|
|
|
(pci_read_config(dev, 0x51, 1) & ~0x02), 1);
|
|
|
|
|
|
pci_write_config(dev, 0x55,
|
|
|
|
|
|
(pci_read_config(dev, 0x55, 1) & ~0x02), 1);
|
2000-09-19 11:08:39 +00:00
|
|
|
|
/* turn on interrupts */
|
2000-08-06 19:51:58 +00:00
|
|
|
|
pci_write_config(dev, 0x5a,
|
|
|
|
|
|
(pci_read_config(dev, 0x5a, 1) & ~0x10), 1);
|
|
|
|
|
|
|
|
|
|
|
|
}
|
2000-02-04 10:20:22 +00:00
|
|
|
|
break;
|
|
|
|
|
|
|
1999-11-26 19:24:26 +00:00
|
|
|
|
case 0x05711106:
|
2000-01-03 10:26:59 +00:00
|
|
|
|
case 0x74091022: /* VIA 82C586, 82C596, 82C686 & AMD 756 default setup */
|
1999-11-26 19:24:26 +00:00
|
|
|
|
/* set prefetch, postwrite */
|
|
|
|
|
|
pci_write_config(dev, 0x41, pci_read_config(dev, 0x41, 1) | 0xf0, 1);
|
|
|
|
|
|
|
|
|
|
|
|
/* set fifo configuration half'n'half */
|
|
|
|
|
|
pci_write_config(dev, 0x43,
|
|
|
|
|
|
(pci_read_config(dev, 0x43, 1) & 0x90) | 0x2a, 1);
|
|
|
|
|
|
|
|
|
|
|
|
/* set status register read retry */
|
|
|
|
|
|
pci_write_config(dev, 0x44, pci_read_config(dev, 0x44, 1) | 0x08, 1);
|
|
|
|
|
|
|
|
|
|
|
|
/* set DMA read & end-of-sector fifo flush */
|
|
|
|
|
|
pci_write_config(dev, 0x46,
|
|
|
|
|
|
(pci_read_config(dev, 0x46, 1) & 0x0c) | 0xf0, 1);
|
|
|
|
|
|
|
|
|
|
|
|
/* set sector size */
|
|
|
|
|
|
pci_write_config(dev, 0x60, DEV_BSIZE, 2);
|
|
|
|
|
|
pci_write_config(dev, 0x68, DEV_BSIZE, 2);
|
2000-01-17 02:04:19 +00:00
|
|
|
|
|
2000-04-09 15:08:27 +00:00
|
|
|
|
/* prepare for ATA-66 on the 82C686 and rev 0x12 and newer 82C596's */
|
|
|
|
|
|
if (ata_find_dev(dev, 0x06861106, 0) ||
|
|
|
|
|
|
ata_find_dev(dev, 0x05961106, 0x12)) {
|
2000-01-17 02:04:19 +00:00
|
|
|
|
pci_write_config(dev, 0x50,
|
|
|
|
|
|
pci_read_config(dev, 0x50, 4) | 0x070f070f, 4);
|
2000-02-04 10:20:22 +00:00
|
|
|
|
}
|
1999-11-26 19:24:26 +00:00
|
|
|
|
break;
|
|
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
2000-02-25 09:48:23 +00:00
|
|
|
|
/*
|
|
|
|
|
|
* the Cypress chip is a mess, it contains two ATA functions, but
|
|
|
|
|
|
* both channels are visible on the first one.
|
|
|
|
|
|
* simply ignore the second function for now, as the right
|
|
|
|
|
|
* solution (ignoring the second channel on the first function)
|
|
|
|
|
|
* doesn't work with the crappy ATA interrupt setup on the alpha.
|
|
|
|
|
|
*/
|
|
|
|
|
|
if (pci_get_devid(dev) == 0xc6931080 && pci_get_function(dev) > 1)
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_pci_add_child(dev, 0);
|
|
|
|
|
|
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev) || pci_read_config(dev, 0x18, 4) & IOMASK)
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_pci_add_child(dev, 1);
|
|
|
|
|
|
|
|
|
|
|
|
return bus_generic_attach(dev);
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
|
ata_pci_print_child(device_t dev, device_t child)
|
|
|
|
|
|
{
|
|
|
|
|
|
struct ata_softc *scp = device_get_softc(child);
|
|
|
|
|
|
int retval = 0;
|
|
|
|
|
|
|
|
|
|
|
|
retval += bus_print_child_header(dev, child);
|
|
|
|
|
|
retval += printf(": at 0x%x", scp->ioaddr);
|
|
|
|
|
|
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev))
|
2000-10-13 11:21:27 +00:00
|
|
|
|
retval += printf(" irq %d", 14 + scp->channel);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
|
|
|
|
|
retval += bus_print_child_footer(dev, child);
|
|
|
|
|
|
|
|
|
|
|
|
return retval;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static struct resource *
|
|
|
|
|
|
ata_pci_alloc_resource(device_t dev, device_t child, int type, int *rid,
|
|
|
|
|
|
u_long start, u_long end, u_long count, u_int flags)
|
|
|
|
|
|
{
|
|
|
|
|
|
struct ata_pci_softc *sc = device_get_softc(dev);
|
2000-10-13 11:21:27 +00:00
|
|
|
|
int channel = ((struct ata_softc *)device_get_softc(child))->channel;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
int myrid;
|
|
|
|
|
|
|
|
|
|
|
|
if (type == SYS_RES_IOPORT) {
|
|
|
|
|
|
switch (*rid) {
|
|
|
|
|
|
case ATA_IOADDR_RID:
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev)) {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
myrid = 0;
|
2000-10-13 11:21:27 +00:00
|
|
|
|
start = (channel == 0 ? IO_WD1 : IO_WD2);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
end = start + ATA_IOSIZE - 1;
|
|
|
|
|
|
count = ATA_IOSIZE;
|
|
|
|
|
|
}
|
|
|
|
|
|
else
|
2000-10-13 11:21:27 +00:00
|
|
|
|
myrid = 0x10 + 8 * channel;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
break;
|
|
|
|
|
|
|
2000-02-25 09:48:23 +00:00
|
|
|
|
case ATA_ALTADDR_RID:
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev)) {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
myrid = 0;
|
2000-10-13 11:21:27 +00:00
|
|
|
|
start = (channel == 0 ? IO_WD1 : IO_WD2) + ATA_ALTOFFSET;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
end = start + ATA_ALTIOSIZE - 1;
|
|
|
|
|
|
count = ATA_ALTIOSIZE;
|
|
|
|
|
|
}
|
|
|
|
|
|
else
|
2000-10-13 11:21:27 +00:00
|
|
|
|
myrid = 0x14 + 8 * channel;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
|
|
case ATA_BMADDR_RID:
|
|
|
|
|
|
/* the busmaster resource is shared between the two channels */
|
|
|
|
|
|
if (sc->bmio) {
|
2000-10-13 11:21:27 +00:00
|
|
|
|
if (channel == 0) {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
sc->bmio_1 = *sc->bmio;
|
|
|
|
|
|
sc->bmio_1.r_end = sc->bmio->r_start + ATA_BM_OFFSET1;
|
|
|
|
|
|
return &sc->bmio_1;
|
|
|
|
|
|
} else {
|
|
|
|
|
|
sc->bmio_2 = *sc->bmio;
|
|
|
|
|
|
sc->bmio_2.r_start = sc->bmio->r_start + ATA_BM_OFFSET1;
|
|
|
|
|
|
sc->bmio_2.r_end = sc->bmio_2.r_start + ATA_BM_OFFSET1;
|
|
|
|
|
|
return &sc->bmio_2;
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
|
return 0;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev))
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* make the parent just pass through the allocation. */
|
|
|
|
|
|
return BUS_ALLOC_RESOURCE(device_get_parent(dev), child,
|
|
|
|
|
|
SYS_RES_IOPORT, &myrid,
|
|
|
|
|
|
start, end, count, flags);
|
|
|
|
|
|
else
|
|
|
|
|
|
/* we are using the parent resource directly. */
|
|
|
|
|
|
return BUS_ALLOC_RESOURCE(device_get_parent(dev), dev,
|
|
|
|
|
|
SYS_RES_IOPORT, &myrid,
|
|
|
|
|
|
start, end, count, flags);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
|
|
|
|
|
if (type == SYS_RES_IRQ) {
|
|
|
|
|
|
if (*rid != 0)
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev)) {
|
2000-09-29 16:50:31 +00:00
|
|
|
|
#ifdef __alpha__
|
2000-10-13 11:21:27 +00:00
|
|
|
|
return alpha_platform_alloc_ide_intr(channel);
|
2000-09-29 16:50:31 +00:00
|
|
|
|
#else
|
2000-10-13 11:21:27 +00:00
|
|
|
|
int irq = (channel == 0 ? 14 : 15);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
|
|
|
|
|
return BUS_ALLOC_RESOURCE(device_get_parent(dev), child,
|
|
|
|
|
|
SYS_RES_IRQ, rid,
|
|
|
|
|
|
irq, irq, 1, flags & ~RF_SHAREABLE);
|
1999-04-16 21:22:55 +00:00
|
|
|
|
#endif
|
2000-01-25 20:14:51 +00:00
|
|
|
|
} else {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* primary and secondary channels share the same interrupt */
|
|
|
|
|
|
sc->irqcnt++;
|
|
|
|
|
|
if (!sc->irq)
|
|
|
|
|
|
sc->irq = BUS_ALLOC_RESOURCE(device_get_parent(dev), dev,
|
|
|
|
|
|
SYS_RES_IRQ, rid, 0, ~0, 1, flags);
|
|
|
|
|
|
return sc->irq;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
}
|
1999-04-18 20:48:15 +00:00
|
|
|
|
return 0;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static int
|
|
|
|
|
|
ata_pci_release_resource(device_t dev, device_t child, int type, int rid,
|
|
|
|
|
|
struct resource *r)
|
1999-12-18 20:06:30 +00:00
|
|
|
|
{
|
2000-02-18 20:57:33 +00:00
|
|
|
|
struct ata_pci_softc *sc = device_get_softc(dev);
|
2000-10-13 11:21:27 +00:00
|
|
|
|
int channel = ((struct ata_softc *)device_get_softc(child))->channel;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
int myrid = 0;
|
|
|
|
|
|
|
|
|
|
|
|
if (type == SYS_RES_IOPORT) {
|
|
|
|
|
|
switch (rid) {
|
|
|
|
|
|
case ATA_IOADDR_RID:
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev))
|
2000-02-18 20:57:33 +00:00
|
|
|
|
myrid = 0;
|
|
|
|
|
|
else
|
2000-10-13 11:21:27 +00:00
|
|
|
|
myrid = 0x10 + 8 * channel;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
break;
|
1999-12-18 20:06:30 +00:00
|
|
|
|
|
2000-02-25 09:48:23 +00:00
|
|
|
|
case ATA_ALTADDR_RID:
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev))
|
2000-02-18 20:57:33 +00:00
|
|
|
|
myrid = 0;
|
|
|
|
|
|
else
|
2000-10-13 11:21:27 +00:00
|
|
|
|
myrid = 0x14 + 8 * channel;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
break;
|
1999-12-18 20:06:30 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
case ATA_BMADDR_RID:
|
|
|
|
|
|
return 0;
|
1999-12-18 20:06:30 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
default:
|
|
|
|
|
|
return ENOENT;
|
1999-12-18 20:06:30 +00:00
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev))
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* make the parent just pass through the allocation. */
|
|
|
|
|
|
return BUS_RELEASE_RESOURCE(device_get_parent(dev), child,
|
|
|
|
|
|
SYS_RES_IOPORT, myrid, r);
|
|
|
|
|
|
else
|
|
|
|
|
|
/* we are using the parent resource directly. */
|
|
|
|
|
|
return BUS_RELEASE_RESOURCE(device_get_parent(dev), dev,
|
|
|
|
|
|
SYS_RES_IOPORT, myrid, r);
|
1999-12-18 20:06:30 +00:00
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (type == SYS_RES_IRQ) {
|
|
|
|
|
|
if (rid != 0)
|
|
|
|
|
|
return ENOENT;
|
|
|
|
|
|
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev)) {
|
2000-09-29 16:50:31 +00:00
|
|
|
|
#ifdef __alpha__
|
2000-10-13 11:21:27 +00:00
|
|
|
|
return alpha_platform_release_ide_intr(channel, r);
|
2000-09-29 16:50:31 +00:00
|
|
|
|
#else
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return BUS_RELEASE_RESOURCE(device_get_parent(dev),
|
|
|
|
|
|
child, SYS_RES_IRQ, rid, r);
|
|
|
|
|
|
#endif
|
|
|
|
|
|
}
|
|
|
|
|
|
else {
|
|
|
|
|
|
if (--sc->irqcnt)
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
|
|
return BUS_RELEASE_RESOURCE(device_get_parent(dev),
|
|
|
|
|
|
dev, SYS_RES_IRQ, rid, r);
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
return EINVAL;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
|
ata_pci_setup_intr(device_t dev, device_t child, struct resource *irq,
|
|
|
|
|
|
int flags, driver_intr_t *intr, void *arg,
|
|
|
|
|
|
void **cookiep)
|
|
|
|
|
|
{
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev)) {
|
2000-09-29 16:50:31 +00:00
|
|
|
|
#ifdef __alpha__
|
2000-10-05 23:09:57 +00:00
|
|
|
|
return alpha_platform_setup_ide_intr(child, irq, intr, arg, cookiep);
|
2000-09-29 16:50:31 +00:00
|
|
|
|
#else
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return BUS_SETUP_INTR(device_get_parent(dev), child, irq,
|
|
|
|
|
|
flags, intr, arg, cookiep);
|
|
|
|
|
|
#endif
|
|
|
|
|
|
}
|
|
|
|
|
|
else
|
|
|
|
|
|
return BUS_SETUP_INTR(device_get_parent(dev), dev, irq,
|
|
|
|
|
|
flags, intr, arg, cookiep);
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
|
ata_pci_teardown_intr(device_t dev, device_t child, struct resource *irq,
|
|
|
|
|
|
void *cookie)
|
|
|
|
|
|
{
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(dev)) {
|
2000-09-29 16:50:31 +00:00
|
|
|
|
#ifdef __alpha__
|
2000-10-05 23:09:57 +00:00
|
|
|
|
return alpha_platform_teardown_ide_intr(child, irq, cookie);
|
2000-09-29 16:50:31 +00:00
|
|
|
|
#else
|
|
|
|
|
|
return BUS_TEARDOWN_INTR(device_get_parent(dev), child, irq, cookie);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
#endif
|
|
|
|
|
|
}
|
|
|
|
|
|
else
|
|
|
|
|
|
return BUS_TEARDOWN_INTR(device_get_parent(dev), dev, irq, cookie);
|
1999-12-18 20:06:30 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
1999-04-18 20:48:15 +00:00
|
|
|
|
static device_method_t ata_pci_methods[] = {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* device interface */
|
|
|
|
|
|
DEVMETHOD(device_probe, ata_pci_probe),
|
|
|
|
|
|
DEVMETHOD(device_attach, ata_pci_attach),
|
|
|
|
|
|
DEVMETHOD(device_shutdown, bus_generic_shutdown),
|
|
|
|
|
|
DEVMETHOD(device_suspend, bus_generic_suspend),
|
|
|
|
|
|
DEVMETHOD(device_resume, bus_generic_resume),
|
|
|
|
|
|
|
|
|
|
|
|
/* bus methods */
|
|
|
|
|
|
DEVMETHOD(bus_print_child, ata_pci_print_child),
|
|
|
|
|
|
DEVMETHOD(bus_alloc_resource, ata_pci_alloc_resource),
|
|
|
|
|
|
DEVMETHOD(bus_release_resource, ata_pci_release_resource),
|
|
|
|
|
|
DEVMETHOD(bus_activate_resource, bus_generic_activate_resource),
|
|
|
|
|
|
DEVMETHOD(bus_deactivate_resource, bus_generic_deactivate_resource),
|
|
|
|
|
|
DEVMETHOD(bus_setup_intr, ata_pci_setup_intr),
|
|
|
|
|
|
DEVMETHOD(bus_teardown_intr, ata_pci_teardown_intr),
|
1999-04-18 20:48:15 +00:00
|
|
|
|
{ 0, 0 }
|
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
static driver_t ata_pci_driver = {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
"atapci",
|
1999-04-18 20:48:15 +00:00
|
|
|
|
ata_pci_methods,
|
2000-02-18 20:57:33 +00:00
|
|
|
|
sizeof(struct ata_pci_softc),
|
1999-04-18 20:48:15 +00:00
|
|
|
|
};
|
|
|
|
|
|
|
2000-02-27 12:41:07 +00:00
|
|
|
|
DRIVER_MODULE(atapci, pci, ata_pci_driver, ata_pci_devclass, 0, 0);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static int
|
|
|
|
|
|
ata_pcisub_probe(device_t dev)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
2000-02-18 20:57:33 +00:00
|
|
|
|
struct ata_softc *scp = device_get_softc(dev);
|
2000-10-13 11:21:27 +00:00
|
|
|
|
device_t *list;
|
|
|
|
|
|
int count, i;
|
|
|
|
|
|
|
|
|
|
|
|
/* find channel number on this controller */
|
|
|
|
|
|
device_get_children(device_get_parent(dev), &list, &count);
|
|
|
|
|
|
for (i = 0; i < count; i++) {
|
|
|
|
|
|
if (list[i] == dev)
|
|
|
|
|
|
scp->channel = i;
|
|
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
2000-08-16 07:09:04 +00:00
|
|
|
|
scp->chiptype = pci_get_devid(device_get_parent(dev));
|
2000-10-13 11:21:27 +00:00
|
|
|
|
|
|
|
|
|
|
/* is this an ATA RAID setup ? */
|
|
|
|
|
|
if (((pci_get_subclass(device_get_parent(dev)) == PCIS_STORAGE_RAID) &&
|
|
|
|
|
|
(scp->chiptype == 0x4d33105a || scp->chiptype == 0x4d38105a ||
|
|
|
|
|
|
scp->chiptype == 0x4d30105a || scp->chiptype == 0x0d30105a)) ||
|
|
|
|
|
|
scp->chiptype == 0x00041103) {
|
|
|
|
|
|
scp->flags |= ATA_RAID;
|
|
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return ata_probe(dev);
|
|
|
|
|
|
}
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static device_method_t ata_pcisub_methods[] = {
|
|
|
|
|
|
/* device interface */
|
|
|
|
|
|
DEVMETHOD(device_probe, ata_pcisub_probe),
|
|
|
|
|
|
DEVMETHOD(device_attach, ata_attach),
|
|
|
|
|
|
DEVMETHOD(device_detach, ata_detach),
|
|
|
|
|
|
DEVMETHOD(device_resume, ata_resume),
|
|
|
|
|
|
{ 0, 0 }
|
|
|
|
|
|
};
|
1999-10-09 19:57:13 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static driver_t ata_pcisub_driver = {
|
|
|
|
|
|
"ata",
|
|
|
|
|
|
ata_pcisub_methods,
|
|
|
|
|
|
sizeof(struct ata_softc),
|
|
|
|
|
|
};
|
1999-10-09 19:57:13 +00:00
|
|
|
|
|
2000-02-27 12:41:07 +00:00
|
|
|
|
DRIVER_MODULE(ata, atapci, ata_pcisub_driver, ata_devclass, 0, 0);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
|
ata_probe(device_t dev)
|
|
|
|
|
|
{
|
|
|
|
|
|
struct ata_softc *scp = device_get_softc(dev);
|
|
|
|
|
|
struct resource *io = 0;
|
|
|
|
|
|
struct resource *altio = 0;
|
|
|
|
|
|
struct resource *bmio = 0;
|
|
|
|
|
|
int rid;
|
2000-09-19 11:08:39 +00:00
|
|
|
|
u_int32_t ioaddr, altioaddr, bmaddr;
|
|
|
|
|
|
int mask = 0;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
u_int8_t status0, status1;
|
|
|
|
|
|
|
|
|
|
|
|
if (!scp || scp->flags & ATA_ATTACHED)
|
|
|
|
|
|
return ENXIO;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* initialize the softc basics */
|
|
|
|
|
|
scp->active = ATA_IDLE;
|
|
|
|
|
|
scp->dev = dev;
|
|
|
|
|
|
scp->devices = 0;
|
|
|
|
|
|
|
|
|
|
|
|
rid = ATA_IOADDR_RID;
|
2000-02-25 09:48:23 +00:00
|
|
|
|
io = bus_alloc_resource(dev, SYS_RES_IOPORT, &rid, 0, ~0,
|
|
|
|
|
|
ATA_IOSIZE, RF_ACTIVE);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (!io)
|
|
|
|
|
|
goto failure;
|
|
|
|
|
|
ioaddr = rman_get_start(io);
|
|
|
|
|
|
|
2000-02-25 09:48:23 +00:00
|
|
|
|
rid = ATA_ALTADDR_RID;
|
|
|
|
|
|
altio = bus_alloc_resource(dev, SYS_RES_IOPORT, &rid, 0, ~0,
|
|
|
|
|
|
ATA_ALTIOSIZE, RF_ACTIVE);
|
2000-09-24 18:19:43 +00:00
|
|
|
|
if (altio) {
|
2000-10-03 13:12:36 +00:00
|
|
|
|
if (ATA_MASTERDEV(device_get_parent(dev)))
|
2000-09-24 18:19:43 +00:00
|
|
|
|
altioaddr = rman_get_start(altio);
|
|
|
|
|
|
else
|
|
|
|
|
|
altioaddr = rman_get_start(altio) + 0x02;
|
|
|
|
|
|
}
|
2000-02-25 09:48:23 +00:00
|
|
|
|
else
|
2000-09-24 18:19:43 +00:00
|
|
|
|
altioaddr = ioaddr + ATA_IOSIZE;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
|
|
|
|
|
rid = ATA_BMADDR_RID;
|
|
|
|
|
|
bmio = bus_alloc_resource(dev, SYS_RES_IOPORT, &rid, 0, ~0, 1, RF_ACTIVE);
|
|
|
|
|
|
bmaddr = bmio ? rman_get_start(bmio) : 0;
|
|
|
|
|
|
|
|
|
|
|
|
/* store the IO resources for eventual later release */
|
|
|
|
|
|
scp->r_io = io;
|
|
|
|
|
|
scp->r_altio = altio;
|
|
|
|
|
|
scp->r_bmio = bmio;
|
|
|
|
|
|
|
|
|
|
|
|
/* store the physical IO addresse for easy access */
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
scp->ioaddr = ioaddr;
|
|
|
|
|
|
scp->altioaddr = altioaddr;
|
1999-12-21 20:18:56 +00:00
|
|
|
|
scp->bmaddr = bmaddr;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
if (bootverbose)
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, -1, "iobase=0x%04x altiobase=0x%04x bmaddr=0x%04x\n",
|
|
|
|
|
|
scp->ioaddr, scp->altioaddr, scp->bmaddr);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
|
|
|
|
|
/* do we have any signs of ATA/ATAPI HW being present ? */
|
|
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, ATA_D_IBM | ATA_MASTER);
|
|
|
|
|
|
DELAY(1);
|
|
|
|
|
|
status0 = inb(scp->ioaddr + ATA_STATUS);
|
|
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, ATA_D_IBM | ATA_SLAVE);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
DELAY(1);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
status1 = inb(scp->ioaddr + ATA_STATUS);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if ((status0 & 0xf8) != 0xf8 && status0 != 0xa5)
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
mask |= 0x01;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if ((status1 & 0xf8) != 0xf8 && status1 != 0xa5)
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
mask |= 0x02;
|
|
|
|
|
|
if (bootverbose)
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, -1, "mask=%02x status0=%02x status1=%02x\n",
|
|
|
|
|
|
mask, status0, status1);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (!mask)
|
|
|
|
|
|
goto failure;
|
|
|
|
|
|
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
ata_reset(scp, &mask);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
if (bootverbose)
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, -1, "devices = 0x%x\n", scp->devices);
|
2000-08-16 07:09:04 +00:00
|
|
|
|
|
|
|
|
|
|
if (!mask)
|
2000-02-18 20:57:33 +00:00
|
|
|
|
goto failure;
|
2000-08-16 07:09:04 +00:00
|
|
|
|
|
Nine'th update to the new ATA/ATAPI driver:
The atapi subsystem has gotten better error handeling and timeouts,
it also tries a REQUEST SENSE command when devices returns errors,
to give a little more info as to what went wrong. It might be a
little verbose for now, but I'm interested in as much feedback on
errors as possible, especially timeouts, as I'm a bit in doubt if
I've chosen resonable default values everywhere.
The disk driver has been changed a bit to prepare for tagged queing,
which is next on my list.
The disk driver has grown a dump routine, I got one implementation
from Darrell Anderson <anderson@cs.duke.edu> which also did
partial dumps (usefull on big memory machines) I left out the
partial stuff for now, and changed the rest alot to fit into the new
ad_request framework.
Some minor cleanups and rearrangements as well.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, again you have been warned :)
Notebook owners should be carefull that their machines dont suspend
as this might cause trouble...
But please tell me how it works for you!
Enjoy!
-Søren
1999-06-25 09:03:07 +00:00
|
|
|
|
TAILQ_INIT(&scp->ata_queue);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
TAILQ_INIT(&scp->atapi_queue);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
|
|
failure:
|
|
|
|
|
|
if (io)
|
|
|
|
|
|
bus_release_resource(dev, SYS_RES_IOPORT, ATA_IOADDR_RID, io);
|
|
|
|
|
|
if (altio)
|
2000-02-25 09:48:23 +00:00
|
|
|
|
bus_release_resource(dev, SYS_RES_IOPORT, ATA_ALTADDR_RID, altio);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (bmio)
|
|
|
|
|
|
bus_release_resource(dev, SYS_RES_IOPORT, ATA_BMADDR_RID, bmio);
|
|
|
|
|
|
if (bootverbose)
|
|
|
|
|
|
ata_printf(scp, -1, "probe allocation failed\n");
|
|
|
|
|
|
return ENXIO;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static int
|
|
|
|
|
|
ata_attach(device_t dev)
|
2000-01-18 21:02:59 +00:00
|
|
|
|
{
|
2000-02-18 20:57:33 +00:00
|
|
|
|
struct ata_softc *scp = device_get_softc(dev);
|
2000-02-25 09:48:23 +00:00
|
|
|
|
int error, rid = 0;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
|
|
|
|
|
if (!scp || scp->flags & ATA_ATTACHED)
|
|
|
|
|
|
return ENXIO;
|
|
|
|
|
|
|
|
|
|
|
|
scp->r_irq = bus_alloc_resource(dev, SYS_RES_IRQ, &rid, 0, ~0, 1,
|
|
|
|
|
|
RF_SHAREABLE | RF_ACTIVE);
|
|
|
|
|
|
if (!scp->r_irq) {
|
|
|
|
|
|
ata_printf(scp, -1, "unable to allocate interrupt\n");
|
|
|
|
|
|
return ENXIO;
|
|
|
|
|
|
}
|
2000-02-25 09:48:23 +00:00
|
|
|
|
if ((error = bus_setup_intr(dev, scp->r_irq, INTR_TYPE_BIO, ata_intr,
|
|
|
|
|
|
scp, &scp->ih)))
|
|
|
|
|
|
return error;
|
2000-01-18 21:02:59 +00:00
|
|
|
|
|
|
|
|
|
|
/*
|
2000-02-18 20:57:33 +00:00
|
|
|
|
* do not attach devices if we are in early boot, this is done later
|
|
|
|
|
|
* when interrupts are enabled by a hook into the boot process.
|
|
|
|
|
|
* otherwise attach what the probe has found in scp->devices.
|
2000-01-18 21:02:59 +00:00
|
|
|
|
*/
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (!ata_delayed_attach) {
|
|
|
|
|
|
if (scp->devices & ATA_ATA_SLAVE)
|
|
|
|
|
|
if (ata_getparam(scp, ATA_SLAVE, ATA_C_ATA_IDENTIFY))
|
|
|
|
|
|
scp->devices &= ~ATA_ATA_SLAVE;
|
|
|
|
|
|
if (scp->devices & ATA_ATAPI_SLAVE)
|
|
|
|
|
|
if (ata_getparam(scp, ATA_SLAVE, ATA_C_ATAPI_IDENTIFY))
|
|
|
|
|
|
scp->devices &= ~ATA_ATAPI_SLAVE;
|
|
|
|
|
|
if (scp->devices & ATA_ATA_MASTER)
|
|
|
|
|
|
if (ata_getparam(scp, ATA_MASTER, ATA_C_ATA_IDENTIFY))
|
|
|
|
|
|
scp->devices &= ~ATA_ATA_MASTER;
|
|
|
|
|
|
if (scp->devices & ATA_ATAPI_MASTER)
|
|
|
|
|
|
if (ata_getparam(scp, ATA_MASTER,ATA_C_ATAPI_IDENTIFY))
|
|
|
|
|
|
scp->devices &= ~ATA_ATAPI_MASTER;
|
|
|
|
|
|
#if NATADISK > 0
|
|
|
|
|
|
if (scp->devices & ATA_ATA_MASTER)
|
|
|
|
|
|
ad_attach(scp, ATA_MASTER);
|
|
|
|
|
|
if (scp->devices & ATA_ATA_SLAVE)
|
|
|
|
|
|
ad_attach(scp, ATA_SLAVE);
|
|
|
|
|
|
#endif
|
|
|
|
|
|
#if NATAPICD > 0 || NATAPIFD > 0 || NATAPIST > 0
|
|
|
|
|
|
if (scp->devices & ATA_ATAPI_MASTER)
|
|
|
|
|
|
atapi_attach(scp, ATA_MASTER);
|
|
|
|
|
|
if (scp->devices & ATA_ATAPI_SLAVE)
|
|
|
|
|
|
atapi_attach(scp, ATA_SLAVE);
|
|
|
|
|
|
#endif
|
2000-01-18 21:02:59 +00:00
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
scp->flags |= ATA_ATTACHED;
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
|
ata_detach(device_t dev)
|
|
|
|
|
|
{
|
|
|
|
|
|
struct ata_softc *scp = device_get_softc(dev);
|
|
|
|
|
|
|
|
|
|
|
|
if (!scp || !(scp->flags & ATA_ATTACHED))
|
|
|
|
|
|
return ENXIO;
|
2000-01-18 21:02:59 +00:00
|
|
|
|
|
2000-01-19 07:25:46 +00:00
|
|
|
|
#if NATADISK > 0
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (scp->devices & ATA_ATA_MASTER)
|
2000-03-13 12:27:11 +00:00
|
|
|
|
ad_detach(scp->dev_softc[0]);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (scp->devices & ATA_ATA_SLAVE)
|
2000-03-13 12:27:11 +00:00
|
|
|
|
ad_detach(scp->dev_softc[1]);
|
2000-01-19 07:25:46 +00:00
|
|
|
|
#endif
|
|
|
|
|
|
#if NATAPICD > 0 || NATAPIFD > 0 || NATAPIST > 0
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (scp->devices & ATA_ATAPI_MASTER)
|
2000-03-13 12:27:11 +00:00
|
|
|
|
atapi_detach(scp->dev_softc[0]);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (scp->devices & ATA_ATAPI_SLAVE)
|
2000-03-13 12:27:11 +00:00
|
|
|
|
atapi_detach(scp->dev_softc[1]);
|
2000-01-19 07:25:46 +00:00
|
|
|
|
#endif
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (scp->dev_param[ATA_DEV(ATA_MASTER)]) {
|
|
|
|
|
|
free(scp->dev_param[ATA_DEV(ATA_MASTER)], M_ATA);
|
|
|
|
|
|
scp->dev_param[ATA_DEV(ATA_MASTER)] = NULL;
|
2000-01-18 21:02:59 +00:00
|
|
|
|
}
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (scp->dev_param[ATA_DEV(ATA_SLAVE)]) {
|
|
|
|
|
|
free(scp->dev_param[ATA_DEV(ATA_SLAVE)], M_ATA);
|
|
|
|
|
|
scp->dev_param[ATA_DEV(ATA_SLAVE)] = NULL;
|
|
|
|
|
|
}
|
2000-03-13 12:04:58 +00:00
|
|
|
|
scp->dev_softc[ATA_DEV(ATA_MASTER)] = NULL;
|
|
|
|
|
|
scp->dev_softc[ATA_DEV(ATA_SLAVE)] = NULL;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
scp->mode[ATA_DEV(ATA_MASTER)] = ATA_PIO;
|
|
|
|
|
|
scp->mode[ATA_DEV(ATA_SLAVE)] = ATA_PIO;
|
2000-02-25 09:48:23 +00:00
|
|
|
|
bus_teardown_intr(dev, scp->r_irq, scp->ih);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
bus_release_resource(dev, SYS_RES_IRQ, 0, scp->r_irq);
|
2000-02-25 09:48:23 +00:00
|
|
|
|
if (scp->r_bmio)
|
|
|
|
|
|
bus_release_resource(dev, SYS_RES_IOPORT, ATA_BMADDR_RID, scp->r_bmio);
|
|
|
|
|
|
if (scp->r_altio)
|
|
|
|
|
|
bus_release_resource(dev, SYS_RES_IOPORT, ATA_ALTADDR_RID,scp->r_altio);
|
2000-02-18 20:57:33 +00:00
|
|
|
|
bus_release_resource(dev, SYS_RES_IOPORT, ATA_IOADDR_RID, scp->r_io);
|
|
|
|
|
|
scp->flags &= ~ATA_ATTACHED;
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
|
ata_resume(device_t dev)
|
|
|
|
|
|
{
|
|
|
|
|
|
struct ata_softc *scp = device_get_softc(dev);
|
|
|
|
|
|
|
|
|
|
|
|
ata_reinit(scp);
|
|
|
|
|
|
return 0;
|
2000-01-18 21:02:59 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
static int
|
|
|
|
|
|
ata_getparam(struct ata_softc *scp, int device, u_int8_t command)
|
2000-01-18 21:02:59 +00:00
|
|
|
|
{
|
|
|
|
|
|
struct ata_params *ata_parm;
|
|
|
|
|
|
int8_t buffer[DEV_BSIZE];
|
2000-01-24 20:45:24 +00:00
|
|
|
|
int retry = 0;
|
2000-01-18 21:02:59 +00:00
|
|
|
|
|
|
|
|
|
|
/* select drive */
|
|
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, ATA_D_IBM | device);
|
|
|
|
|
|
DELAY(1);
|
2000-01-24 20:45:24 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
/* enable interrupt */
|
2000-09-24 18:19:43 +00:00
|
|
|
|
outb(scp->altioaddr, ATA_A_4BIT);
|
2000-01-25 20:14:51 +00:00
|
|
|
|
DELAY(1);
|
|
|
|
|
|
|
2000-01-24 20:45:24 +00:00
|
|
|
|
/* apparently some devices needs this repeated */
|
|
|
|
|
|
do {
|
|
|
|
|
|
if (ata_command(scp, device, command, 0, 0, 0, 0, 0, ATA_WAIT_INTR)) {
|
|
|
|
|
|
ata_printf(scp, device, "identify failed\n");
|
|
|
|
|
|
return -1;
|
|
|
|
|
|
}
|
2000-01-25 20:14:51 +00:00
|
|
|
|
if (retry++ > 4) {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_printf(scp, device, "identify retries exceeded\n");
|
2000-01-24 20:45:24 +00:00
|
|
|
|
return -1;
|
|
|
|
|
|
}
|
2000-01-25 20:14:51 +00:00
|
|
|
|
} while (ata_wait(scp, device,
|
|
|
|
|
|
((command == ATA_C_ATAPI_IDENTIFY) ?
|
|
|
|
|
|
ATA_S_DRQ : (ATA_S_READY | ATA_S_DSC | ATA_S_DRQ))));
|
2000-01-24 20:45:24 +00:00
|
|
|
|
|
2000-01-18 21:02:59 +00:00
|
|
|
|
insw(scp->ioaddr + ATA_DATA, buffer, sizeof(buffer)/sizeof(int16_t));
|
|
|
|
|
|
ata_parm = malloc(sizeof(struct ata_params), M_ATA, M_NOWAIT);
|
|
|
|
|
|
if (!ata_parm) {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_printf(scp, device, "malloc for identify data failed\n");
|
2000-01-18 21:02:59 +00:00
|
|
|
|
return -1;
|
|
|
|
|
|
}
|
|
|
|
|
|
bcopy(buffer, ata_parm, sizeof(struct ata_params));
|
|
|
|
|
|
if (command == ATA_C_ATA_IDENTIFY ||
|
|
|
|
|
|
!((ata_parm->model[0] == 'N' && ata_parm->model[1] == 'E') ||
|
|
|
|
|
|
(ata_parm->model[0] == 'F' && ata_parm->model[1] == 'X')))
|
|
|
|
|
|
bswap(ata_parm->model, sizeof(ata_parm->model));
|
|
|
|
|
|
btrim(ata_parm->model, sizeof(ata_parm->model));
|
|
|
|
|
|
bpack(ata_parm->model, ata_parm->model, sizeof(ata_parm->model));
|
|
|
|
|
|
bswap(ata_parm->revision, sizeof(ata_parm->revision));
|
|
|
|
|
|
btrim(ata_parm->revision, sizeof(ata_parm->revision));
|
|
|
|
|
|
bpack(ata_parm->revision, ata_parm->revision, sizeof(ata_parm->revision));
|
2000-01-24 20:45:24 +00:00
|
|
|
|
scp->dev_param[ATA_DEV(device)] = ata_parm;
|
2000-01-18 21:02:59 +00:00
|
|
|
|
return 0;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
static void
|
|
|
|
|
|
ata_boot_attach(void)
|
|
|
|
|
|
{
|
|
|
|
|
|
struct ata_softc *scp;
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int ctlr;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
|
* run through all ata devices and look for real ATA & ATAPI devices
|
|
|
|
|
|
* using the hints we found in the early probe, this avoids some of
|
|
|
|
|
|
* the delays probing of non-exsistent devices can cause.
|
|
|
|
|
|
*/
|
|
|
|
|
|
for (ctlr=0; ctlr<devclass_get_maxunit(ata_devclass); ctlr++) {
|
|
|
|
|
|
if (!(scp = devclass_get_softc(ata_devclass, ctlr)))
|
|
|
|
|
|
continue;
|
|
|
|
|
|
if (scp->devices & ATA_ATA_SLAVE)
|
|
|
|
|
|
if (ata_getparam(scp, ATA_SLAVE, ATA_C_ATA_IDENTIFY))
|
|
|
|
|
|
scp->devices &= ~ATA_ATA_SLAVE;
|
|
|
|
|
|
if (scp->devices & ATA_ATAPI_SLAVE)
|
|
|
|
|
|
if (ata_getparam(scp, ATA_SLAVE, ATA_C_ATAPI_IDENTIFY))
|
|
|
|
|
|
scp->devices &= ~ATA_ATAPI_SLAVE;
|
|
|
|
|
|
if (scp->devices & ATA_ATA_MASTER)
|
|
|
|
|
|
if (ata_getparam(scp, ATA_MASTER, ATA_C_ATA_IDENTIFY))
|
|
|
|
|
|
scp->devices &= ~ATA_ATA_MASTER;
|
|
|
|
|
|
if (scp->devices & ATA_ATAPI_MASTER)
|
|
|
|
|
|
if (ata_getparam(scp, ATA_MASTER,ATA_C_ATAPI_IDENTIFY))
|
|
|
|
|
|
scp->devices &= ~ATA_ATAPI_MASTER;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
#if NATADISK > 0
|
|
|
|
|
|
/* now we know whats there, do the real attach, first the ATA disks */
|
|
|
|
|
|
for (ctlr=0; ctlr<devclass_get_maxunit(ata_devclass); ctlr++) {
|
|
|
|
|
|
if (!(scp = devclass_get_softc(ata_devclass, ctlr)))
|
|
|
|
|
|
continue;
|
|
|
|
|
|
if (scp->devices & ATA_ATA_MASTER)
|
|
|
|
|
|
ad_attach(scp, ATA_MASTER);
|
|
|
|
|
|
if (scp->devices & ATA_ATA_SLAVE)
|
|
|
|
|
|
ad_attach(scp, ATA_SLAVE);
|
|
|
|
|
|
}
|
|
|
|
|
|
#endif
|
|
|
|
|
|
#if NATAPICD > 0 || NATAPIFD > 0 || NATAPIST > 0
|
|
|
|
|
|
/* then the atapi devices */
|
|
|
|
|
|
for (ctlr=0; ctlr<devclass_get_maxunit(ata_devclass); ctlr++) {
|
|
|
|
|
|
if (!(scp = devclass_get_softc(ata_devclass, ctlr)))
|
|
|
|
|
|
continue;
|
|
|
|
|
|
if (scp->devices & ATA_ATAPI_MASTER)
|
|
|
|
|
|
atapi_attach(scp, ATA_MASTER);
|
|
|
|
|
|
if (scp->devices & ATA_ATAPI_SLAVE)
|
|
|
|
|
|
atapi_attach(scp, ATA_SLAVE);
|
|
|
|
|
|
}
|
|
|
|
|
|
#endif
|
|
|
|
|
|
if (ata_delayed_attach) {
|
|
|
|
|
|
config_intrhook_disestablish(ata_delayed_attach);
|
|
|
|
|
|
free(ata_delayed_attach, M_ATA);
|
|
|
|
|
|
ata_delayed_attach = NULL;
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
1999-03-03 21:10:29 +00:00
|
|
|
|
static void
|
2000-02-18 20:57:33 +00:00
|
|
|
|
ata_intr(void *data)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
2000-01-03 10:26:59 +00:00
|
|
|
|
struct ata_softc *scp = (struct ata_softc *)data;
|
2000-10-05 08:28:06 +00:00
|
|
|
|
struct ata_pci_softc *sc = device_get_softc(device_get_parent(scp->dev));
|
2000-09-19 11:08:39 +00:00
|
|
|
|
u_int8_t dmastat = 0;
|
2000-02-04 10:20:22 +00:00
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
|
* since we might share the IRQ with another device, and in some
|
2000-02-18 20:57:33 +00:00
|
|
|
|
* cases with our twin channel, we only want to process interrupts
|
2000-02-04 10:20:22 +00:00
|
|
|
|
* that we know this channel generated.
|
|
|
|
|
|
*/
|
|
|
|
|
|
switch (scp->chiptype) {
|
|
|
|
|
|
#if NPCI > 0
|
2000-08-06 19:51:58 +00:00
|
|
|
|
case 0x00041103: /* HighPoint HPT366/368/370 */
|
2000-09-19 11:08:39 +00:00
|
|
|
|
if (((dmastat = ata_dmastatus(scp)) &
|
|
|
|
|
|
(ATA_BMSTAT_ACTIVE | ATA_BMSTAT_INTERRUPT)) != ATA_BMSTAT_INTERRUPT)
|
2000-02-04 10:20:22 +00:00
|
|
|
|
return;
|
|
|
|
|
|
outb(scp->bmaddr + ATA_BMSTAT_PORT, dmastat | ATA_BMSTAT_INTERRUPT);
|
|
|
|
|
|
break;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-10-03 13:12:36 +00:00
|
|
|
|
case 0x06481095: /* CMD 648 */
|
|
|
|
|
|
case 0x06491095: /* CMD 649 */
|
|
|
|
|
|
if (!(pci_read_config(device_get_parent(scp->dev), 0x71, 1) &
|
2000-10-13 11:21:27 +00:00
|
|
|
|
(scp->channel ? 0x08 : 0x04)))
|
2000-10-03 13:12:36 +00:00
|
|
|
|
return;
|
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
case 0x4d33105a: /* Promise Ultra/Fasttrak 33 */
|
|
|
|
|
|
case 0x4d38105a: /* Promise Ultra/Fasttrak 66 */
|
|
|
|
|
|
case 0x4d30105a: /* Promise Ultra/Fasttrak 100 */
|
|
|
|
|
|
case 0x0d30105a: /* Promise OEM ATA100 */
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (!(inl(rman_get_start(sc->bmio) + 0x1c) &
|
2000-10-13 11:21:27 +00:00
|
|
|
|
(scp->channel ? 0x00004000 : 0x00000400)))
|
2000-01-03 10:26:59 +00:00
|
|
|
|
return;
|
2000-10-05 08:28:06 +00:00
|
|
|
|
/* FALLTHROUGH */
|
2000-10-03 13:12:36 +00:00
|
|
|
|
out:
|
2000-02-04 10:20:22 +00:00
|
|
|
|
#endif
|
|
|
|
|
|
default:
|
|
|
|
|
|
if (scp->flags & ATA_DMA_ACTIVE) {
|
2000-09-19 11:08:39 +00:00
|
|
|
|
if (((dmastat = ata_dmastatus(scp)) &
|
|
|
|
|
|
(ATA_BMSTAT_ACTIVE|ATA_BMSTAT_INTERRUPT))!=ATA_BMSTAT_INTERRUPT)
|
2000-02-04 10:20:22 +00:00
|
|
|
|
return;
|
2000-09-19 11:08:39 +00:00
|
|
|
|
outb(scp->bmaddr + ATA_BMSTAT_PORT, dmastat | ATA_BMSTAT_INTERRUPT);
|
2000-02-04 10:20:22 +00:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
DELAY(1);
|
2000-01-03 10:26:59 +00:00
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
/* if drive is busy it didn't interrupt */
|
2000-09-24 18:19:43 +00:00
|
|
|
|
if (inb(scp->altioaddr) & ATA_S_BUSY)
|
1999-10-09 19:57:13 +00:00
|
|
|
|
return;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
/* clear interrupt and get status */
|
|
|
|
|
|
scp->status = inb(scp->ioaddr + ATA_STATUS);
|
|
|
|
|
|
|
|
|
|
|
|
if (scp->status & ATA_S_ERROR)
|
|
|
|
|
|
scp->error = inb(scp->ioaddr + ATA_ERROR);
|
|
|
|
|
|
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
/* find & call the responsible driver to process this interrupt */
|
|
|
|
|
|
switch (scp->active) {
|
1999-03-03 21:10:29 +00:00
|
|
|
|
#if NATADISK > 0
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
case ATA_ACTIVE_ATA:
|
2000-09-19 11:08:39 +00:00
|
|
|
|
if (!scp->running || ad_interrupt(scp->running) == ATA_OP_CONTINUES)
|
1999-11-06 16:50:21 +00:00
|
|
|
|
return;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
break;
|
1999-03-03 21:10:29 +00:00
|
|
|
|
#endif
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#if NATAPICD > 0 || NATAPIFD > 0 || NATAPIST > 0
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
case ATA_ACTIVE_ATAPI:
|
2000-09-19 11:08:39 +00:00
|
|
|
|
if (!scp->running || atapi_interrupt(scp->running) == ATA_OP_CONTINUES)
|
1999-11-06 16:50:21 +00:00
|
|
|
|
return;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
break;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#endif
|
1999-03-05 09:43:30 +00:00
|
|
|
|
case ATA_WAIT_INTR:
|
|
|
|
|
|
wakeup((caddr_t)scp);
|
|
|
|
|
|
break;
|
|
|
|
|
|
|
1999-11-06 16:50:21 +00:00
|
|
|
|
case ATA_WAIT_READY:
|
|
|
|
|
|
break;
|
|
|
|
|
|
|
1999-10-09 19:57:13 +00:00
|
|
|
|
case ATA_REINITING:
|
|
|
|
|
|
return;
|
|
|
|
|
|
|
1999-03-03 21:10:29 +00:00
|
|
|
|
case ATA_IDLE:
|
2000-09-19 11:08:39 +00:00
|
|
|
|
if (scp->flags & ATA_QUEUED) {
|
|
|
|
|
|
scp->active = ATA_ACTIVE;
|
|
|
|
|
|
if (ata_service(scp) == ATA_OP_CONTINUES)
|
|
|
|
|
|
return;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
}
|
2000-09-19 11:08:39 +00:00
|
|
|
|
/* FALLTHROUGH */
|
|
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
|
#ifdef ATA_DEBUG
|
|
|
|
|
|
{
|
|
|
|
|
|
static int intr_count = 0;
|
|
|
|
|
|
|
|
|
|
|
|
if (intr_count++ < 10)
|
|
|
|
|
|
ata_printf(scp, -1, "unwanted interrupt %d status = %02x\n",
|
|
|
|
|
|
intr_count, scp->status);
|
|
|
|
|
|
}
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#endif
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
1999-03-07 21:49:14 +00:00
|
|
|
|
scp->active = ATA_IDLE;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
scp->running = NULL;
|
1999-03-07 21:49:14 +00:00
|
|
|
|
ata_start(scp);
|
2000-09-19 11:08:39 +00:00
|
|
|
|
return;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
|
ata_start(struct ata_softc *scp)
|
|
|
|
|
|
{
|
2000-02-04 10:20:22 +00:00
|
|
|
|
#if NATADISK > 0
|
Nine'th update to the new ATA/ATAPI driver:
The atapi subsystem has gotten better error handeling and timeouts,
it also tries a REQUEST SENSE command when devices returns errors,
to give a little more info as to what went wrong. It might be a
little verbose for now, but I'm interested in as much feedback on
errors as possible, especially timeouts, as I'm a bit in doubt if
I've chosen resonable default values everywhere.
The disk driver has been changed a bit to prepare for tagged queing,
which is next on my list.
The disk driver has grown a dump routine, I got one implementation
from Darrell Anderson <anderson@cs.duke.edu> which also did
partial dumps (usefull on big memory machines) I left out the
partial stuff for now, and changed the rest alot to fit into the new
ad_request framework.
Some minor cleanups and rearrangements as well.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, again you have been warned :)
Notebook owners should be carefull that their machines dont suspend
as this might cause trouble...
But please tell me how it works for you!
Enjoy!
-Søren
1999-06-25 09:03:07 +00:00
|
|
|
|
struct ad_request *ad_request;
|
2000-02-04 10:20:22 +00:00
|
|
|
|
#endif
|
|
|
|
|
|
#if NATAPICD > 0 || NATAPIFD > 0 || NATAPIST > 0
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
struct atapi_request *atapi_request;
|
2000-02-04 10:20:22 +00:00
|
|
|
|
#endif
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
if (scp->active != ATA_IDLE)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
return;
|
2000-09-19 11:08:39 +00:00
|
|
|
|
|
2000-01-27 19:00:51 +00:00
|
|
|
|
scp->active = ATA_ACTIVE;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
Fourth update to the new ATA/ATAPI driver:
Well, better late than newer, but things has been hectic
around here, sorry for the long delay.
DMA support has been added to the ATA disk driver.
This only works on Intel PIIX3/4, Acer Aladdin and Promise controllers.
The promise support works without the BIOS on the board,
and timing modes are set to support up to UDMA speed. This
solves the problems with having more than one promise controller
in the same system.
There is support for "generic" DMA, that might work on other
controllers, but now you have been warned :)
More chipset specific code will come soon, I have to find testers
with the approbiate HW, more on that when I have it ready.
The system now uses its own major numbers, please run MAKEDEV
with the devices you need (ad?, acd?, afd?, ast?).
For now the disk driver will also attach to the old wd major
so one can at least boot without this step, but be warned, this
will eventually go away. The bootblocks will have to be changed
before one can boot directly from an "ad" device though.
Fixed problems:
All known hang problems should be solved
The probe code has been sligthly changed, this should solve
the reports I have lying around (I hope).
Hangs when accessing ata & atapi device on the same channel simultaniously.
A real braino in ata_start caused this, fixed.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, agaiin you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-03-28 18:57:20 +00:00
|
|
|
|
#if NATADISK > 0
|
1999-10-09 19:57:13 +00:00
|
|
|
|
/* find & call the responsible driver if anything on the ATA queue */
|
2000-03-13 12:04:58 +00:00
|
|
|
|
if (TAILQ_EMPTY(&scp->ata_queue)) {
|
|
|
|
|
|
if (scp->devices & (ATA_ATA_MASTER) && scp->dev_softc[0])
|
|
|
|
|
|
ad_start((struct ad_softc *)scp->dev_softc[0]);
|
|
|
|
|
|
if (scp->devices & (ATA_ATA_SLAVE) && scp->dev_softc[1])
|
|
|
|
|
|
ad_start((struct ad_softc *)scp->dev_softc[1]);
|
|
|
|
|
|
}
|
Nine'th update to the new ATA/ATAPI driver:
The atapi subsystem has gotten better error handeling and timeouts,
it also tries a REQUEST SENSE command when devices returns errors,
to give a little more info as to what went wrong. It might be a
little verbose for now, but I'm interested in as much feedback on
errors as possible, especially timeouts, as I'm a bit in doubt if
I've chosen resonable default values everywhere.
The disk driver has been changed a bit to prepare for tagged queing,
which is next on my list.
The disk driver has grown a dump routine, I got one implementation
from Darrell Anderson <anderson@cs.duke.edu> which also did
partial dumps (usefull on big memory machines) I left out the
partial stuff for now, and changed the rest alot to fit into the new
ad_request framework.
Some minor cleanups and rearrangements as well.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, again you have been warned :)
Notebook owners should be carefull that their machines dont suspend
as this might cause trouble...
But please tell me how it works for you!
Enjoy!
-Søren
1999-06-25 09:03:07 +00:00
|
|
|
|
if ((ad_request = TAILQ_FIRST(&scp->ata_queue))) {
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
TAILQ_REMOVE(&scp->ata_queue, ad_request, chain);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
scp->active = ATA_ACTIVE_ATA;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
scp->running = ad_request;
|
2000-09-19 11:08:39 +00:00
|
|
|
|
if (ad_transfer(ad_request) == ATA_OP_CONTINUES)
|
|
|
|
|
|
return;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
2000-09-19 11:08:39 +00:00
|
|
|
|
|
Fourth update to the new ATA/ATAPI driver:
Well, better late than newer, but things has been hectic
around here, sorry for the long delay.
DMA support has been added to the ATA disk driver.
This only works on Intel PIIX3/4, Acer Aladdin and Promise controllers.
The promise support works without the BIOS on the board,
and timing modes are set to support up to UDMA speed. This
solves the problems with having more than one promise controller
in the same system.
There is support for "generic" DMA, that might work on other
controllers, but now you have been warned :)
More chipset specific code will come soon, I have to find testers
with the approbiate HW, more on that when I have it ready.
The system now uses its own major numbers, please run MAKEDEV
with the devices you need (ad?, acd?, afd?, ast?).
For now the disk driver will also attach to the old wd major
so one can at least boot without this step, but be warned, this
will eventually go away. The bootblocks will have to be changed
before one can boot directly from an "ad" device though.
Fixed problems:
All known hang problems should be solved
The probe code has been sligthly changed, this should solve
the reports I have lying around (I hope).
Hangs when accessing ata & atapi device on the same channel simultaniously.
A real braino in ata_start caused this, fixed.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, agaiin you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-03-28 18:57:20 +00:00
|
|
|
|
#endif
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#if NATAPICD > 0 || NATAPIFD > 0 || NATAPIST > 0
|
2000-03-18 22:42:45 +00:00
|
|
|
|
/* find & call the responsible driver if anything on the ATAPI queue */
|
2000-03-13 12:04:58 +00:00
|
|
|
|
if (TAILQ_EMPTY(&scp->atapi_queue)) {
|
|
|
|
|
|
if (scp->devices & (ATA_ATAPI_MASTER) && scp->dev_softc[0])
|
|
|
|
|
|
atapi_start((struct atapi_softc *)scp->dev_softc[0]);
|
|
|
|
|
|
if (scp->devices & (ATA_ATAPI_SLAVE) && scp->dev_softc[1])
|
|
|
|
|
|
atapi_start((struct atapi_softc *)scp->dev_softc[1]);
|
|
|
|
|
|
}
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
if ((atapi_request = TAILQ_FIRST(&scp->atapi_queue))) {
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
TAILQ_REMOVE(&scp->atapi_queue, atapi_request, chain);
|
|
|
|
|
|
scp->active = ATA_ACTIVE_ATAPI;
|
|
|
|
|
|
scp->running = atapi_request;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
atapi_transfer(atapi_request);
|
Fourth update to the new ATA/ATAPI driver:
Well, better late than newer, but things has been hectic
around here, sorry for the long delay.
DMA support has been added to the ATA disk driver.
This only works on Intel PIIX3/4, Acer Aladdin and Promise controllers.
The promise support works without the BIOS on the board,
and timing modes are set to support up to UDMA speed. This
solves the problems with having more than one promise controller
in the same system.
There is support for "generic" DMA, that might work on other
controllers, but now you have been warned :)
More chipset specific code will come soon, I have to find testers
with the approbiate HW, more on that when I have it ready.
The system now uses its own major numbers, please run MAKEDEV
with the devices you need (ad?, acd?, afd?, ast?).
For now the disk driver will also attach to the old wd major
so one can at least boot without this step, but be warned, this
will eventually go away. The bootblocks will have to be changed
before one can boot directly from an "ad" device though.
Fixed problems:
All known hang problems should be solved
The probe code has been sligthly changed, this should solve
the reports I have lying around (I hope).
Hangs when accessing ata & atapi device on the same channel simultaniously.
A real braino in ata_start caused this, fixed.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, agaiin you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-03-28 18:57:20 +00:00
|
|
|
|
return;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#endif
|
2000-01-27 19:00:51 +00:00
|
|
|
|
scp->active = ATA_IDLE;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void
|
2000-09-19 11:08:39 +00:00
|
|
|
|
ata_reset(struct ata_softc *scp, int *mask)
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
{
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int timeout;
|
2000-08-16 07:09:04 +00:00
|
|
|
|
u_int8_t status0 = ATA_S_BUSY, status1 = ATA_S_BUSY;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
|
|
|
|
|
|
/* reset channel */
|
|
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, ATA_D_IBM | ATA_MASTER);
|
|
|
|
|
|
DELAY(1);
|
|
|
|
|
|
inb(scp->ioaddr + ATA_STATUS);
|
2000-09-24 18:19:43 +00:00
|
|
|
|
outb(scp->altioaddr, ATA_A_IDS | ATA_A_RESET);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
DELAY(10000);
|
2000-09-24 18:19:43 +00:00
|
|
|
|
outb(scp->altioaddr, ATA_A_IDS);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
DELAY(10000);
|
|
|
|
|
|
inb(scp->ioaddr + ATA_ERROR);
|
1999-10-09 19:57:13 +00:00
|
|
|
|
DELAY(3000);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
|
|
|
|
|
|
/* wait for BUSY to go inactive */
|
1999-10-09 19:57:13 +00:00
|
|
|
|
for (timeout = 0; timeout < 310000; timeout++) {
|
2000-08-16 07:09:04 +00:00
|
|
|
|
if (status0 & ATA_S_BUSY) {
|
|
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, ATA_D_IBM | ATA_MASTER);
|
|
|
|
|
|
DELAY(1);
|
|
|
|
|
|
status0 = inb(scp->ioaddr + ATA_STATUS);
|
|
|
|
|
|
if (!(status0 & ATA_S_BUSY)) {
|
|
|
|
|
|
/* check for ATAPI signature while its still there */
|
|
|
|
|
|
if (inb(scp->ioaddr + ATA_CYL_LSB) == ATAPI_MAGIC_LSB &&
|
|
|
|
|
|
inb(scp->ioaddr + ATA_CYL_MSB) == ATAPI_MAGIC_MSB)
|
|
|
|
|
|
scp->devices |= ATA_ATAPI_MASTER;
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
if (status1 & ATA_S_BUSY) {
|
|
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, ATA_D_IBM | ATA_SLAVE);
|
|
|
|
|
|
DELAY(1);
|
|
|
|
|
|
status1 = inb(scp->ioaddr + ATA_STATUS);
|
|
|
|
|
|
if (!(status1 & ATA_S_BUSY)) {
|
|
|
|
|
|
/* check for ATAPI signature while its still there */
|
|
|
|
|
|
if (inb(scp->ioaddr + ATA_CYL_LSB) == ATAPI_MAGIC_LSB &&
|
|
|
|
|
|
inb(scp->ioaddr + ATA_CYL_MSB) == ATAPI_MAGIC_MSB)
|
|
|
|
|
|
scp->devices |= ATA_ATAPI_SLAVE;
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
if (*mask == 0x01) /* wait for master only */
|
1999-10-09 19:57:13 +00:00
|
|
|
|
if (!(status0 & ATA_S_BUSY))
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
break;
|
|
|
|
|
|
if (*mask == 0x02) /* wait for slave only */
|
1999-10-09 19:57:13 +00:00
|
|
|
|
if (!(status1 & ATA_S_BUSY))
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
break;
|
|
|
|
|
|
if (*mask == 0x03) /* wait for both master & slave */
|
1999-10-09 19:57:13 +00:00
|
|
|
|
if (!(status0 & ATA_S_BUSY) && !(status1 & ATA_S_BUSY))
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
break;
|
2000-08-11 10:34:22 +00:00
|
|
|
|
DELAY(100);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
}
|
1999-10-09 19:57:13 +00:00
|
|
|
|
DELAY(1);
|
2000-09-24 18:19:43 +00:00
|
|
|
|
outb(scp->altioaddr, ATA_A_4BIT);
|
1999-10-09 19:57:13 +00:00
|
|
|
|
if (status0 & ATA_S_BUSY)
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
*mask &= ~0x01;
|
1999-10-09 19:57:13 +00:00
|
|
|
|
if (status1 & ATA_S_BUSY)
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
*mask &= ~0x02;
|
|
|
|
|
|
if (bootverbose)
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, -1, "mask=%02x status0=%02x status1=%02x\n",
|
|
|
|
|
|
*mask, status0, status1);
|
2000-08-16 07:09:04 +00:00
|
|
|
|
if (!mask) {
|
|
|
|
|
|
scp->devices = 0;
|
|
|
|
|
|
return;
|
|
|
|
|
|
}
|
|
|
|
|
|
/*
|
|
|
|
|
|
* OK, we have at least one device on the chain, checks for ATAPI
|
|
|
|
|
|
* already done, if none check if its a good old ATA device.
|
|
|
|
|
|
*/
|
|
|
|
|
|
if (status0 != 0x00 && !(scp->devices & ATA_ATAPI_MASTER)) {
|
|
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, (ATA_D_IBM | ATA_MASTER));
|
|
|
|
|
|
DELAY(1);
|
|
|
|
|
|
outb(scp->ioaddr + ATA_ERROR, 0x58);
|
|
|
|
|
|
outb(scp->ioaddr + ATA_CYL_LSB, 0xa5);
|
|
|
|
|
|
if (inb(scp->ioaddr + ATA_ERROR) != 0x58 &&
|
|
|
|
|
|
inb(scp->ioaddr + ATA_CYL_LSB) == 0xa5) {
|
|
|
|
|
|
scp->devices |= ATA_ATA_MASTER;
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
if (status1 != 0x00 && !(scp->devices & ATA_ATAPI_SLAVE)) {
|
|
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, (ATA_D_IBM | ATA_SLAVE));
|
|
|
|
|
|
DELAY(1);
|
|
|
|
|
|
outb(scp->ioaddr + ATA_ERROR, 0x58);
|
|
|
|
|
|
outb(scp->ioaddr + ATA_CYL_LSB, 0xa5);
|
|
|
|
|
|
if (inb(scp->ioaddr + ATA_ERROR) != 0x58 &&
|
|
|
|
|
|
inb(scp->ioaddr + ATA_CYL_LSB) == 0xa5) {
|
|
|
|
|
|
scp->devices |= ATA_ATA_SLAVE;
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
ata_reinit(struct ata_softc *scp)
|
|
|
|
|
|
{
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int mask = 0, omask;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
|
1999-10-09 19:57:13 +00:00
|
|
|
|
scp->active = ATA_REINITING;
|
|
|
|
|
|
scp->running = NULL;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
if (scp->devices & (ATA_ATA_MASTER | ATA_ATAPI_MASTER))
|
|
|
|
|
|
mask |= 0x01;
|
|
|
|
|
|
if (scp->devices & (ATA_ATA_SLAVE | ATA_ATAPI_SLAVE))
|
|
|
|
|
|
mask |= 0x02;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (mask) {
|
|
|
|
|
|
omask = mask;
|
|
|
|
|
|
ata_printf(scp, -1, "resetting devices .. ");
|
|
|
|
|
|
ata_reset(scp, &mask);
|
|
|
|
|
|
if (omask != mask)
|
|
|
|
|
|
printf(" device dissapeared! %d ", omask & ~mask);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
|
|
|
|
|
|
#if NATADISK > 0
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (scp->devices & (ATA_ATA_MASTER) && scp->dev_softc[0])
|
|
|
|
|
|
ad_reinit((struct ad_softc *)scp->dev_softc[0]);
|
|
|
|
|
|
if (scp->devices & (ATA_ATA_SLAVE) && scp->dev_softc[1])
|
|
|
|
|
|
ad_reinit((struct ad_softc *)scp->dev_softc[1]);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#endif
|
|
|
|
|
|
#if NATAPICD > 0 || NATAPIFD > 0 || NATAPIST > 0
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (scp->devices & (ATA_ATAPI_MASTER) && scp->dev_softc[0])
|
|
|
|
|
|
atapi_reinit((struct atapi_softc *)scp->dev_softc[0]);
|
|
|
|
|
|
if (scp->devices & (ATA_ATAPI_SLAVE) && scp->dev_softc[1])
|
|
|
|
|
|
atapi_reinit((struct atapi_softc *)scp->dev_softc[1]);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
#endif
|
2000-02-18 20:57:33 +00:00
|
|
|
|
printf("done\n");
|
|
|
|
|
|
}
|
1999-10-09 19:57:13 +00:00
|
|
|
|
scp->active = ATA_IDLE;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
ata_start(scp);
|
|
|
|
|
|
return 0;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
static int
|
|
|
|
|
|
ata_service(struct ata_softc *scp)
|
|
|
|
|
|
{
|
|
|
|
|
|
/* do we have a SERVICE request from the drive ? */
|
|
|
|
|
|
if ((scp->status & (ATA_S_SERVICE|ATA_S_ERROR|ATA_S_DRQ)) == ATA_S_SERVICE){
|
|
|
|
|
|
outb(scp->bmaddr + ATA_BMSTAT_PORT, ata_dmastatus(scp) | ATA_BMSTAT_INTERRUPT);
|
|
|
|
|
|
#if NATADISK > 0
|
|
|
|
|
|
if ((inb(scp->ioaddr + ATA_DRIVE) & ATA_SLAVE) == ATA_MASTER) {
|
|
|
|
|
|
if ((scp->devices & ATA_ATA_MASTER) && scp->dev_softc[0])
|
|
|
|
|
|
return ad_service((struct ad_softc *)scp->dev_softc[0], 0);
|
|
|
|
|
|
}
|
|
|
|
|
|
else {
|
|
|
|
|
|
if ((scp->devices & ATA_ATA_SLAVE) && scp->dev_softc[1])
|
|
|
|
|
|
return ad_service((struct ad_softc *)scp->dev_softc[1], 0);
|
|
|
|
|
|
}
|
|
|
|
|
|
#endif
|
|
|
|
|
|
}
|
|
|
|
|
|
return ATA_OP_FINISHED;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
int
|
|
|
|
|
|
ata_wait(struct ata_softc *scp, int device, u_int8_t mask)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int timeout = 0;
|
2000-09-24 18:19:43 +00:00
|
|
|
|
int statio = scp->ioaddr + ATA_STATUS;
|
2000-01-17 02:04:19 +00:00
|
|
|
|
|
1999-12-13 10:19:39 +00:00
|
|
|
|
DELAY(1);
|
2000-01-17 02:04:19 +00:00
|
|
|
|
while (timeout < 5000000) { /* timeout 5 secs */
|
2000-09-24 18:19:43 +00:00
|
|
|
|
scp->status = inb(statio);
|
Fourth update to the new ATA/ATAPI driver:
Well, better late than newer, but things has been hectic
around here, sorry for the long delay.
DMA support has been added to the ATA disk driver.
This only works on Intel PIIX3/4, Acer Aladdin and Promise controllers.
The promise support works without the BIOS on the board,
and timing modes are set to support up to UDMA speed. This
solves the problems with having more than one promise controller
in the same system.
There is support for "generic" DMA, that might work on other
controllers, but now you have been warned :)
More chipset specific code will come soon, I have to find testers
with the approbiate HW, more on that when I have it ready.
The system now uses its own major numbers, please run MAKEDEV
with the devices you need (ad?, acd?, afd?, ast?).
For now the disk driver will also attach to the old wd major
so one can at least boot without this step, but be warned, this
will eventually go away. The bootblocks will have to be changed
before one can boot directly from an "ad" device though.
Fixed problems:
All known hang problems should be solved
The probe code has been sligthly changed, this should solve
the reports I have lying around (I hope).
Hangs when accessing ata & atapi device on the same channel simultaniously.
A real braino in ata_start caused this, fixed.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, agaiin you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-03-28 18:57:20 +00:00
|
|
|
|
|
|
|
|
|
|
/* if drive fails status, reselect the drive just to be sure */
|
2000-01-17 02:04:19 +00:00
|
|
|
|
if (scp->status == 0xff) {
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, device, "no status, reselecting device\n");
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, ATA_D_IBM | device);
|
1999-03-07 21:49:14 +00:00
|
|
|
|
DELAY(1);
|
2000-09-24 18:19:43 +00:00
|
|
|
|
scp->status = inb(statio);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
}
|
2000-01-17 02:04:19 +00:00
|
|
|
|
|
|
|
|
|
|
/* are we done ? */
|
|
|
|
|
|
if (!(scp->status & ATA_S_BUSY))
|
|
|
|
|
|
break;
|
|
|
|
|
|
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
if (timeout > 1000) {
|
|
|
|
|
|
timeout += 1000;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
DELAY(1000);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
}
|
|
|
|
|
|
else {
|
|
|
|
|
|
timeout += 10;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
DELAY(10);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
}
|
2000-01-17 02:04:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
if (scp->status & ATA_S_ERROR)
|
|
|
|
|
|
scp->error = inb(scp->ioaddr + ATA_ERROR);
|
|
|
|
|
|
if (timeout >= 5000000)
|
|
|
|
|
|
return -1;
|
|
|
|
|
|
if (!mask)
|
|
|
|
|
|
return (scp->status & ATA_S_ERROR);
|
|
|
|
|
|
|
|
|
|
|
|
/* Wait 50 msec for bits wanted. */
|
|
|
|
|
|
timeout = 5000;
|
|
|
|
|
|
while (timeout--) {
|
2000-09-24 18:19:43 +00:00
|
|
|
|
scp->status = inb(statio);
|
2000-01-17 02:04:19 +00:00
|
|
|
|
if ((scp->status & mask) == mask) {
|
|
|
|
|
|
if (scp->status & ATA_S_ERROR)
|
|
|
|
|
|
scp->error = inb(scp->ioaddr + ATA_ERROR);
|
|
|
|
|
|
return (scp->status & ATA_S_ERROR);
|
|
|
|
|
|
}
|
|
|
|
|
|
DELAY (10);
|
|
|
|
|
|
}
|
|
|
|
|
|
return -1;
|
|
|
|
|
|
}
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int
|
|
|
|
|
|
ata_command(struct ata_softc *scp, int device, u_int8_t command,
|
|
|
|
|
|
u_int16_t cylinder, u_int8_t head, u_int8_t sector,
|
|
|
|
|
|
u_int8_t count, u_int8_t feature, int flags)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int error = 0;
|
Fourth update to the new ATA/ATAPI driver:
Well, better late than newer, but things has been hectic
around here, sorry for the long delay.
DMA support has been added to the ATA disk driver.
This only works on Intel PIIX3/4, Acer Aladdin and Promise controllers.
The promise support works without the BIOS on the board,
and timing modes are set to support up to UDMA speed. This
solves the problems with having more than one promise controller
in the same system.
There is support for "generic" DMA, that might work on other
controllers, but now you have been warned :)
More chipset specific code will come soon, I have to find testers
with the approbiate HW, more on that when I have it ready.
The system now uses its own major numbers, please run MAKEDEV
with the devices you need (ad?, acd?, afd?, ast?).
For now the disk driver will also attach to the old wd major
so one can at least boot without this step, but be warned, this
will eventually go away. The bootblocks will have to be changed
before one can boot directly from an "ad" device though.
Fixed problems:
All known hang problems should be solved
The probe code has been sligthly changed, this should solve
the reports I have lying around (I hope).
Hangs when accessing ata & atapi device on the same channel simultaniously.
A real braino in ata_start caused this, fixed.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, agaiin you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-03-28 18:57:20 +00:00
|
|
|
|
#ifdef ATA_DEBUG
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, device, "ata_command: addr=%04x, cmd=%02x, "
|
2000-09-19 11:08:39 +00:00
|
|
|
|
"c=%d, h=%d, s=%d, count=%d, feature=%d, flags=%02x\n",
|
|
|
|
|
|
scp->ioaddr, command, cylinder, head, sector,
|
|
|
|
|
|
count, feature, flags);
|
Fourth update to the new ATA/ATAPI driver:
Well, better late than newer, but things has been hectic
around here, sorry for the long delay.
DMA support has been added to the ATA disk driver.
This only works on Intel PIIX3/4, Acer Aladdin and Promise controllers.
The promise support works without the BIOS on the board,
and timing modes are set to support up to UDMA speed. This
solves the problems with having more than one promise controller
in the same system.
There is support for "generic" DMA, that might work on other
controllers, but now you have been warned :)
More chipset specific code will come soon, I have to find testers
with the approbiate HW, more on that when I have it ready.
The system now uses its own major numbers, please run MAKEDEV
with the devices you need (ad?, acd?, afd?, ast?).
For now the disk driver will also attach to the old wd major
so one can at least boot without this step, but be warned, this
will eventually go away. The bootblocks will have to be changed
before one can boot directly from an "ad" device though.
Fixed problems:
All known hang problems should be solved
The probe code has been sligthly changed, this should solve
the reports I have lying around (I hope).
Hangs when accessing ata & atapi device on the same channel simultaniously.
A real braino in ata_start caused this, fixed.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, agaiin you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-03-28 18:57:20 +00:00
|
|
|
|
#endif
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
/* disable interrupt from device */
|
|
|
|
|
|
if (scp->flags & ATA_QUEUED)
|
2000-09-24 18:19:43 +00:00
|
|
|
|
outb(scp->altioaddr, ATA_A_IDS | ATA_A_4BIT);
|
2000-09-19 11:08:39 +00:00
|
|
|
|
|
|
|
|
|
|
/* select device */
|
|
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, ATA_D_IBM | device);
|
|
|
|
|
|
|
1999-03-07 21:49:14 +00:00
|
|
|
|
/* ready to issue command ? */
|
Fourth update to the new ATA/ATAPI driver:
Well, better late than newer, but things has been hectic
around here, sorry for the long delay.
DMA support has been added to the ATA disk driver.
This only works on Intel PIIX3/4, Acer Aladdin and Promise controllers.
The promise support works without the BIOS on the board,
and timing modes are set to support up to UDMA speed. This
solves the problems with having more than one promise controller
in the same system.
There is support for "generic" DMA, that might work on other
controllers, but now you have been warned :)
More chipset specific code will come soon, I have to find testers
with the approbiate HW, more on that when I have it ready.
The system now uses its own major numbers, please run MAKEDEV
with the devices you need (ad?, acd?, afd?, ast?).
For now the disk driver will also attach to the old wd major
so one can at least boot without this step, but be warned, this
will eventually go away. The bootblocks will have to be changed
before one can boot directly from an "ad" device though.
Fixed problems:
All known hang problems should be solved
The probe code has been sligthly changed, this should solve
the reports I have lying around (I hope).
Hangs when accessing ata & atapi device on the same channel simultaniously.
A real braino in ata_start caused this, fixed.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, agaiin you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-03-28 18:57:20 +00:00
|
|
|
|
if (ata_wait(scp, device, 0) < 0) {
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, device,
|
|
|
|
|
|
"timeout waiting to give command=%02x s=%02x e=%02x\n",
|
|
|
|
|
|
command, scp->status, scp->error);
|
1999-10-09 19:57:13 +00:00
|
|
|
|
return -1;
|
1999-03-05 09:43:30 +00:00
|
|
|
|
}
|
2000-09-19 11:08:39 +00:00
|
|
|
|
|
Fourth update to the new ATA/ATAPI driver:
Well, better late than newer, but things has been hectic
around here, sorry for the long delay.
DMA support has been added to the ATA disk driver.
This only works on Intel PIIX3/4, Acer Aladdin and Promise controllers.
The promise support works without the BIOS on the board,
and timing modes are set to support up to UDMA speed. This
solves the problems with having more than one promise controller
in the same system.
There is support for "generic" DMA, that might work on other
controllers, but now you have been warned :)
More chipset specific code will come soon, I have to find testers
with the approbiate HW, more on that when I have it ready.
The system now uses its own major numbers, please run MAKEDEV
with the devices you need (ad?, acd?, afd?, ast?).
For now the disk driver will also attach to the old wd major
so one can at least boot without this step, but be warned, this
will eventually go away. The bootblocks will have to be changed
before one can boot directly from an "ad" device though.
Fixed problems:
All known hang problems should be solved
The probe code has been sligthly changed, this should solve
the reports I have lying around (I hope).
Hangs when accessing ata & atapi device on the same channel simultaniously.
A real braino in ata_start caused this, fixed.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, agaiin you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-03-28 18:57:20 +00:00
|
|
|
|
outb(scp->ioaddr + ATA_FEATURE, feature);
|
2000-09-19 11:08:39 +00:00
|
|
|
|
outb(scp->ioaddr + ATA_COUNT, count);
|
|
|
|
|
|
outb(scp->ioaddr + ATA_SECTOR, sector);
|
1999-03-05 09:43:30 +00:00
|
|
|
|
outb(scp->ioaddr + ATA_CYL_MSB, cylinder >> 8);
|
2000-09-19 11:08:39 +00:00
|
|
|
|
outb(scp->ioaddr + ATA_CYL_LSB, cylinder);
|
Fourth update to the new ATA/ATAPI driver:
Well, better late than newer, but things has been hectic
around here, sorry for the long delay.
DMA support has been added to the ATA disk driver.
This only works on Intel PIIX3/4, Acer Aladdin and Promise controllers.
The promise support works without the BIOS on the board,
and timing modes are set to support up to UDMA speed. This
solves the problems with having more than one promise controller
in the same system.
There is support for "generic" DMA, that might work on other
controllers, but now you have been warned :)
More chipset specific code will come soon, I have to find testers
with the approbiate HW, more on that when I have it ready.
The system now uses its own major numbers, please run MAKEDEV
with the devices you need (ad?, acd?, afd?, ast?).
For now the disk driver will also attach to the old wd major
so one can at least boot without this step, but be warned, this
will eventually go away. The bootblocks will have to be changed
before one can boot directly from an "ad" device though.
Fixed problems:
All known hang problems should be solved
The probe code has been sligthly changed, this should solve
the reports I have lying around (I hope).
Hangs when accessing ata & atapi device on the same channel simultaniously.
A real braino in ata_start caused this, fixed.
As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, agaiin you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-03-28 18:57:20 +00:00
|
|
|
|
outb(scp->ioaddr + ATA_DRIVE, ATA_D_IBM | device | head);
|
1999-03-05 09:43:30 +00:00
|
|
|
|
|
|
|
|
|
|
switch (flags) {
|
|
|
|
|
|
case ATA_WAIT_INTR:
|
1999-10-09 19:57:13 +00:00
|
|
|
|
if (scp->active != ATA_IDLE)
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, device, "WARNING: WAIT_INTR active=%s\n",
|
|
|
|
|
|
active2str(scp->active));
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
scp->active = ATA_WAIT_INTR;
|
2000-01-25 20:14:51 +00:00
|
|
|
|
asleep((caddr_t)scp, PRIBIO, "atacmd", 10 * hz);
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
outb(scp->ioaddr + ATA_CMD, command);
|
2000-09-19 11:08:39 +00:00
|
|
|
|
|
|
|
|
|
|
/* enable interrupt */
|
|
|
|
|
|
if (scp->flags & ATA_QUEUED)
|
2000-09-24 18:19:43 +00:00
|
|
|
|
outb(scp->altioaddr, ATA_A_4BIT);
|
2000-09-19 11:08:39 +00:00
|
|
|
|
|
2000-01-25 20:14:51 +00:00
|
|
|
|
if (await(PRIBIO, 10 * hz)) {
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, device, "ata_command: timeout waiting for intr\n");
|
1999-03-07 21:49:14 +00:00
|
|
|
|
scp->active = ATA_IDLE;
|
2000-09-19 11:08:39 +00:00
|
|
|
|
error = -1;
|
1999-03-07 21:49:14 +00:00
|
|
|
|
}
|
1999-03-05 09:43:30 +00:00
|
|
|
|
break;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
1999-11-06 16:50:21 +00:00
|
|
|
|
case ATA_WAIT_READY:
|
1999-10-09 19:57:13 +00:00
|
|
|
|
if (scp->active != ATA_IDLE && scp->active != ATA_REINITING)
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, device, "WARNING: WAIT_READY active=%s\n",
|
|
|
|
|
|
active2str(scp->active));
|
|
|
|
|
|
if (scp->active != ATA_REINITING)
|
|
|
|
|
|
scp->active = ATA_WAIT_READY;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
outb(scp->ioaddr + ATA_CMD, command);
|
1999-11-06 16:50:21 +00:00
|
|
|
|
if (ata_wait(scp, device, ATA_S_READY) < 0) {
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, device,
|
|
|
|
|
|
"timeout waiting for command=%02x s=%02x e=%02x\n",
|
|
|
|
|
|
command, scp->status, scp->error);
|
2000-09-19 11:08:39 +00:00
|
|
|
|
error = -1;
|
1999-11-06 16:50:21 +00:00
|
|
|
|
}
|
2000-01-24 20:45:24 +00:00
|
|
|
|
if (scp->active != ATA_REINITING)
|
|
|
|
|
|
scp->active = ATA_IDLE;
|
1999-03-05 09:43:30 +00:00
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
|
|
case ATA_IMMEDIATE:
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
outb(scp->ioaddr + ATA_CMD, command);
|
1999-03-05 09:43:30 +00:00
|
|
|
|
break;
|
1999-10-09 19:57:13 +00:00
|
|
|
|
|
|
|
|
|
|
default:
|
2000-01-24 20:45:24 +00:00
|
|
|
|
ata_printf(scp, device, "DANGER: illegal interrupt flag=%s\n",
|
|
|
|
|
|
active2str(flags));
|
1999-03-05 09:43:30 +00:00
|
|
|
|
}
|
2000-09-19 11:08:39 +00:00
|
|
|
|
/* enable interrupt */
|
|
|
|
|
|
if (scp->flags & ATA_QUEUED)
|
2000-09-24 18:19:43 +00:00
|
|
|
|
outb(scp->altioaddr, ATA_A_4BIT);
|
2000-09-19 11:08:39 +00:00
|
|
|
|
return error;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
int
|
2000-09-19 11:08:39 +00:00
|
|
|
|
ata_get_lun(u_int32_t *map)
|
|
|
|
|
|
{
|
|
|
|
|
|
int lun = ffs(~*map) - 1;
|
|
|
|
|
|
|
|
|
|
|
|
*map |= (1 << lun);
|
|
|
|
|
|
return lun;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
|
ata_free_lun(u_int32_t *map, int lun)
|
|
|
|
|
|
{
|
|
|
|
|
|
*map &= ~(1 << lun);
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
int
|
|
|
|
|
|
ata_printf(struct ata_softc *scp, int device, const char * fmt, ...)
|
2000-02-18 20:57:33 +00:00
|
|
|
|
{
|
|
|
|
|
|
va_list ap;
|
|
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
|
|
|
|
if (device == -1)
|
|
|
|
|
|
ret = printf("ata%d: ", device_get_unit(scp->dev));
|
|
|
|
|
|
else
|
|
|
|
|
|
ret = printf("ata%d-%s: ", device_get_unit(scp->dev),
|
|
|
|
|
|
(device == ATA_MASTER) ? "master" : "slave");
|
|
|
|
|
|
va_start(ap, fmt);
|
|
|
|
|
|
ret += vprintf(fmt, ap);
|
|
|
|
|
|
va_end(ap);
|
|
|
|
|
|
return ret;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
char *
|
|
|
|
|
|
ata_mode2str(int mode)
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
{
|
|
|
|
|
|
switch (mode) {
|
2000-01-27 19:00:51 +00:00
|
|
|
|
case ATA_PIO: return "BIOSPIO";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
case ATA_PIO0: return "PIO0";
|
|
|
|
|
|
case ATA_PIO1: return "PIO1";
|
|
|
|
|
|
case ATA_PIO2: return "PIO2";
|
|
|
|
|
|
case ATA_PIO3: return "PIO3";
|
|
|
|
|
|
case ATA_PIO4: return "PIO4";
|
|
|
|
|
|
case ATA_WDMA2: return "WDMA2";
|
|
|
|
|
|
case ATA_UDMA2: return "UDMA33";
|
|
|
|
|
|
case ATA_UDMA4: return "UDMA66";
|
2000-08-06 19:51:58 +00:00
|
|
|
|
case ATA_UDMA5: return "UDMA100";
|
2000-01-27 19:00:51 +00:00
|
|
|
|
case ATA_DMA: return "BIOSDMA";
|
2000-01-03 10:26:59 +00:00
|
|
|
|
default: return "???";
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int
|
|
|
|
|
|
ata_pio2mode(int pio)
|
2000-01-03 10:26:59 +00:00
|
|
|
|
{
|
|
|
|
|
|
switch (pio) {
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
default:
|
2000-01-03 10:26:59 +00:00
|
|
|
|
case 0: return ATA_PIO0;
|
|
|
|
|
|
case 1: return ATA_PIO1;
|
|
|
|
|
|
case 2: return ATA_PIO2;
|
|
|
|
|
|
case 3: return ATA_PIO3;
|
|
|
|
|
|
case 4: return ATA_PIO4;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
int
|
|
|
|
|
|
ata_pmode(struct ata_params *ap)
|
|
|
|
|
|
{
|
|
|
|
|
|
if (ap->atavalid & ATA_FLAG_64_70) {
|
|
|
|
|
|
if (ap->apiomodes & 2)
|
|
|
|
|
|
return 4;
|
|
|
|
|
|
if (ap->apiomodes & 1)
|
|
|
|
|
|
return 3;
|
|
|
|
|
|
}
|
|
|
|
|
|
if (ap->opiomode == 2)
|
|
|
|
|
|
return 2;
|
|
|
|
|
|
if (ap->opiomode == 1)
|
|
|
|
|
|
return 1;
|
|
|
|
|
|
if (ap->opiomode == 0)
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
return -1;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
int
|
|
|
|
|
|
ata_wmode(struct ata_params *ap)
|
|
|
|
|
|
{
|
|
|
|
|
|
if (ap->wdmamodes & 4)
|
|
|
|
|
|
return 2;
|
|
|
|
|
|
if (ap->wdmamodes & 2)
|
|
|
|
|
|
return 1;
|
|
|
|
|
|
if (ap->wdmamodes & 1)
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
return -1;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
int
|
|
|
|
|
|
ata_umode(struct ata_params *ap)
|
|
|
|
|
|
{
|
|
|
|
|
|
if (ap->atavalid & ATA_FLAG_88) {
|
2000-08-06 19:51:58 +00:00
|
|
|
|
if (ap->udmamodes & 0x20)
|
|
|
|
|
|
return 5;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (ap->udmamodes & 0x10)
|
2000-05-23 19:05:56 +00:00
|
|
|
|
return 4;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (ap->udmamodes & 0x08)
|
2000-05-23 19:05:56 +00:00
|
|
|
|
return 3;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (ap->udmamodes & 0x04)
|
|
|
|
|
|
return 2;
|
|
|
|
|
|
if (ap->udmamodes & 0x02)
|
|
|
|
|
|
return 1;
|
|
|
|
|
|
if (ap->udmamodes & 0x01)
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
}
|
|
|
|
|
|
return -1;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2000-09-19 11:08:39 +00:00
|
|
|
|
static char *
|
|
|
|
|
|
active2str(int active)
|
1999-10-09 19:57:13 +00:00
|
|
|
|
{
|
2000-01-03 10:26:59 +00:00
|
|
|
|
static char buf[8];
|
|
|
|
|
|
|
1999-10-09 19:57:13 +00:00
|
|
|
|
switch (active) {
|
|
|
|
|
|
case ATA_IDLE:
|
|
|
|
|
|
return("ATA_IDLE");
|
2000-02-04 10:20:22 +00:00
|
|
|
|
case ATA_IMMEDIATE:
|
|
|
|
|
|
return("ATA_IMMEDIATE");
|
1999-10-09 19:57:13 +00:00
|
|
|
|
case ATA_WAIT_INTR:
|
|
|
|
|
|
return("ATA_WAIT_INTR");
|
2000-02-04 10:20:22 +00:00
|
|
|
|
case ATA_WAIT_READY:
|
|
|
|
|
|
return("ATA_WAIT_READY");
|
|
|
|
|
|
case ATA_ACTIVE:
|
|
|
|
|
|
return("ATA_ACTIVE");
|
1999-10-09 19:57:13 +00:00
|
|
|
|
case ATA_ACTIVE_ATA:
|
|
|
|
|
|
return("ATA_ACTIVE_ATA");
|
|
|
|
|
|
case ATA_ACTIVE_ATAPI:
|
|
|
|
|
|
return("ATA_ACTIVE_ATAPI");
|
|
|
|
|
|
case ATA_REINITING:
|
|
|
|
|
|
return("ATA_REINITING");
|
|
|
|
|
|
default:
|
2000-01-03 10:26:59 +00:00
|
|
|
|
sprintf(buf, "0x%02x", active);
|
|
|
|
|
|
return buf;
|
1999-10-09 19:57:13 +00:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2000-01-18 21:02:59 +00:00
|
|
|
|
static void
|
2000-09-19 11:08:39 +00:00
|
|
|
|
bswap(int8_t *buf, int len)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
2000-02-18 20:57:33 +00:00
|
|
|
|
u_int16_t *ptr = (u_int16_t*)(buf + len);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
while (--ptr >= (u_int16_t*)buf)
|
|
|
|
|
|
*ptr = ntohs(*ptr);
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
2000-01-18 21:02:59 +00:00
|
|
|
|
static void
|
2000-09-19 11:08:39 +00:00
|
|
|
|
btrim(int8_t *buf, int len)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
2000-02-18 20:57:33 +00:00
|
|
|
|
int8_t *ptr;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-02-18 20:57:33 +00:00
|
|
|
|
for (ptr = buf; ptr < buf+len; ++ptr)
|
|
|
|
|
|
if (!*ptr)
|
|
|
|
|
|
*ptr = ' ';
|
|
|
|
|
|
for (ptr = buf + len - 1; ptr >= buf && *ptr == ' '; --ptr)
|
|
|
|
|
|
*ptr = 0;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
2000-01-18 21:02:59 +00:00
|
|
|
|
static void
|
2000-09-19 11:08:39 +00:00
|
|
|
|
bpack(int8_t *src, int8_t *dst, int len)
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
{
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int i, j, blank;
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
|
2000-01-24 20:45:24 +00:00
|
|
|
|
for (i = j = blank = 0 ; i < len; i++) {
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
if (blank && src[i] == ' ') continue;
|
|
|
|
|
|
if (blank && src[i] != ' ') {
|
|
|
|
|
|
dst[j++] = src[i];
|
|
|
|
|
|
blank = 0;
|
|
|
|
|
|
continue;
|
|
|
|
|
|
}
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
if (src[i] == ' ') {
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
blank = 1;
|
Ten'th update to the new ATA/ATAPI driver:
It been awhile since the last major update, as a benefit there
are some cool things in this one (and new bugs probably :) )...
The ATA driver has grown "real" timeout support for all devices.
This means that it should be possible to get in contact with
(especially) lost ATAPI devices. It also means that the ATA
driver is now usable on notebooks as it will DTRT on resume.
An experimental hack at utilizing the Promise66's at UDMA66 is
in there, but I cant test it. If someone feels like sending
me one, give me a ping.
The ATAPI DMA enableling scheme has been changed, also better DMA
support for the Aladdin chipset has been implemented for ATAPI
devices. Note that the Aladdin apparently only can do DMA reads
on ATAPI devices, and the Promise cant do ATAPI DMA at all.
I have seen problems on some ATAPI devices that should be able
to run in DMA mode, so if you encounter problems with hanging
atapi devices during the probe, or during access, disable DMA
in atapi-all.c, and let me know. It might be nessesary to do this
via a "white list" for known good devices...
The ATAPI CDROM driver can now use eject/close without hanging and
the bug that caused reading beyond the end of a CD has been fixed.
Media change is also handled proberly. DVD drives are identified
and are usable as CDROM devices at least, I dont have the HW to
test this further, see above :).
The ATAPI tape driver has gotten some support for using the DSC
method for not blocking the IDE channel during read/write when
the device has full buffers. It knows about the OnStream DI-30
device, support is not completed yet, but it can function as a
primitive backup medium, without filemarks, and without bad media
handeling. This is because the OnStream device doesn't handle this
(like everybody else) in HW. It also now supports getting/setting
the record position on devices that supports it.
Some rather major cleanups and rearrangements as well (cvs -b diff
is your freind). I'm closing in on declaring this for beta code,
most of the infrastruture is in place by now.
As usual USE AT YOUR OWN RISK!!, this is still alpha level code.
This driver can hose your disk real bad if anything goes wrong, but
now you have been warned :)
But please tell me how it works for you!
Enjoy!
-Søren
1999-09-21 19:50:40 +00:00
|
|
|
|
if (i == 0)
|
|
|
|
|
|
continue;
|
|
|
|
|
|
}
|
Finally!!
The much roumored replacement for our current IDE/ATA/ATAPI is
materialising in the CVS repositories around the globe.
So what does this bring us:
A new reengineered ATA/ATAPI subsystem, that tries to overcome
most of the deficiencies with the current drivers.
It supports PCI as well as ISA devices without all the hackery
in ide_pci.c to make PCI devices look like ISA counterparts.
It doesn't have the excessive wait problem on probe, in fact you
shouldn't notice any delay when your devices are getting probed.
Probing and attaching of devices are postponed until interrupts
are enabled (well almost, not finished yet for disks), making
things alot cleaner.
Improved performance, although DMA support is still WIP and not
in this pre alpha release, worldstone is faster with the new
driver compared to the old even with DMA.
So what does it take away:
There is NO support for old MFM/RLL/ESDI disks.
There is NO support for bad144, if your disk is bad, ditch it, it has
already outgrown its internal spare sectors, and is dying.
For you to try this out, you will have to modify your kernel config
file to use the "ata" controller instead of all wdc? entries.
example:
# for a PCI only system (most modern machines)
controller ata0
device atadisk0 # ATA disks
device atapicd0 # ATAPI CDROM's
device atapist0 # ATAPI tapes
#You should add the following on ISA systems:
controller ata1 at isa? port "IO_WD1" bio irq 14
controller ata2 at isa? port "IO_WD2" bio irq 15
You can leave it all in there, the system knows how to manage.
For now this driver reuses the device entries from the old system
(that will probably change later), but remember that disks are
now numbered in the sequence they are found (like the SCSI system)
not as absolute positions as the old system.
Although I have tested this on all the systems I can get my hands on,
there might very well be gremlins in there, so use AT YOU OWN RISK!!
This is still WIP, so there are lots of rough edges and unfinished
things in there, and what I have in my lab might look very different
from whats in CVS at any given time. So please have all eventual
changes go through me, or chances are they just dissapears...
I would very much like to hear from you, both good and bad news
are very welcome.
Enjoy!!
-Søren
1999-03-01 21:19:19 +00:00
|
|
|
|
dst[j++] = src[i];
|
|
|
|
|
|
}
|
2000-01-24 20:45:24 +00:00
|
|
|
|
if (j < len)
|
|
|
|
|
|
dst[j] = 0x00;
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2000-01-27 19:00:51 +00:00
|
|
|
|
static void
|
2000-09-19 11:08:39 +00:00
|
|
|
|
ata_change_mode(struct ata_softc *scp, int device, int mode)
|
2000-01-27 19:00:51 +00:00
|
|
|
|
{
|
2000-09-19 11:08:39 +00:00
|
|
|
|
int s = splbio();
|
2000-01-27 19:00:51 +00:00
|
|
|
|
|
|
|
|
|
|
while (scp->active != ATA_IDLE)
|
|
|
|
|
|
tsleep((caddr_t)&s, PRIBIO, "atachm", hz/4);
|
|
|
|
|
|
scp->active = ATA_REINITING;
|
2000-09-19 11:08:39 +00:00
|
|
|
|
ata_dmainit(scp, device, ata_pmode(ATA_PARAM(scp, device)),
|
2000-01-27 19:00:51 +00:00
|
|
|
|
mode < ATA_DMA ? -1 : ata_wmode(ATA_PARAM(scp, device)),
|
|
|
|
|
|
mode < ATA_DMA ? -1 : ata_umode(ATA_PARAM(scp, device)));
|
|
|
|
|
|
scp->active = ATA_IDLE;
|
|
|
|
|
|
ata_start(scp);
|
|
|
|
|
|
splx(s);
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
2000-07-04 11:25:35 +00:00
|
|
|
|
sysctl_hw_ata(SYSCTL_HANDLER_ARGS)
|
2000-01-27 19:00:51 +00:00
|
|
|
|
{
|
2000-02-18 20:57:33 +00:00
|
|
|
|
struct ata_softc *scp;
|
|
|
|
|
|
int ctlr, error, i;
|
2000-01-27 19:00:51 +00:00
|
|
|
|
|
|
|
|
|
|
/* readout internal state */
|
|
|
|
|
|
bzero(ata_conf, sizeof(ata_conf));
|
2000-02-18 20:57:33 +00:00
|
|
|
|
for (ctlr=0; ctlr<devclass_get_maxunit(ata_devclass); ctlr++) {
|
|
|
|
|
|
if (!(scp = devclass_get_softc(ata_devclass, ctlr)))
|
|
|
|
|
|
continue;
|
|
|
|
|
|
for (i = 0; i < 2; i++) {
|
|
|
|
|
|
if (!scp->dev_softc[i])
|
|
|
|
|
|
strcat(ata_conf, "---,");
|
|
|
|
|
|
else if (scp->mode[i] >= ATA_DMA)
|
|
|
|
|
|
strcat(ata_conf, "dma,");
|
|
|
|
|
|
else
|
|
|
|
|
|
strcat(ata_conf, "pio,");
|
|
|
|
|
|
}
|
2000-01-27 19:00:51 +00:00
|
|
|
|
}
|
|
|
|
|
|
error = sysctl_handle_string(oidp, ata_conf, sizeof(ata_conf), req);
|
|
|
|
|
|
if (error == 0 && req->newptr != NULL) {
|
|
|
|
|
|
char *ptr = ata_conf;
|
|
|
|
|
|
|
|
|
|
|
|
/* update internal state */
|
|
|
|
|
|
i = 0;
|
|
|
|
|
|
while (*ptr) {
|
|
|
|
|
|
if (!strncmp(ptr, "pio", 3) || !strncmp(ptr, "PIO", 3)) {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if ((scp = devclass_get_softc(ata_devclass, i >> 1)) &&
|
|
|
|
|
|
scp->dev_softc[i & 1] && scp->mode[i & 1] >= ATA_DMA)
|
|
|
|
|
|
ata_change_mode(scp, (i & 1)?ATA_SLAVE:ATA_MASTER, ATA_PIO);
|
2000-01-27 19:00:51 +00:00
|
|
|
|
}
|
|
|
|
|
|
else if (!strncmp(ptr, "dma", 3) || !strncmp(ptr, "DMA", 3)) {
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if ((scp = devclass_get_softc(ata_devclass, i >> 1)) &&
|
|
|
|
|
|
scp->dev_softc[i & 1] && scp->mode[i & 1] < ATA_DMA)
|
|
|
|
|
|
ata_change_mode(scp, (i & 1)?ATA_SLAVE:ATA_MASTER, ATA_DMA);
|
2000-01-27 19:00:51 +00:00
|
|
|
|
}
|
|
|
|
|
|
else if (strncmp(ptr, "---", 3))
|
|
|
|
|
|
break;
|
|
|
|
|
|
ptr+=3;
|
2000-02-18 20:57:33 +00:00
|
|
|
|
if (*ptr++ != ',' ||
|
|
|
|
|
|
++i > (devclass_get_maxunit(ata_devclass) << 1))
|
2000-01-27 19:00:51 +00:00
|
|
|
|
break;
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
return error;
|
|
|
|
|
|
}
|
|
|
|
|
|
SYSCTL_PROC(_hw, OID_AUTO, atamodes, CTLTYPE_STRING | CTLFLAG_RW,
|
|
|
|
|
|
0, sizeof(ata_conf), sysctl_hw_ata, "A", "");
|
2000-02-18 20:57:33 +00:00
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
|
ata_init(void)
|
|
|
|
|
|
{
|
|
|
|
|
|
/* register boot attach to be run when interrupts are enabled */
|
|
|
|
|
|
if (!(ata_delayed_attach = (struct intr_config_hook *)
|
|
|
|
|
|
malloc(sizeof(struct intr_config_hook),
|
|
|
|
|
|
M_TEMP, M_NOWAIT))) {
|
|
|
|
|
|
printf("ata: malloc of delayed attach hook failed\n");
|
|
|
|
|
|
return;
|
|
|
|
|
|
}
|
|
|
|
|
|
bzero(ata_delayed_attach, sizeof(struct intr_config_hook));
|
|
|
|
|
|
|
|
|
|
|
|
ata_delayed_attach->ich_func = (void*)ata_boot_attach;
|
|
|
|
|
|
if (config_intrhook_establish(ata_delayed_attach) != 0) {
|
|
|
|
|
|
printf("ata: config_intrhook_establish failed\n");
|
|
|
|
|
|
free(ata_delayed_attach, M_TEMP);
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
SYSINIT(atadev, SI_SUB_DRIVERS, SI_ORDER_SECOND, ata_init, NULL)
|