sometime around 1.51, the check for minphys dissappeared out of
transfers for disks..
we weren't hecking that the adapter could handle a transfer of
the size we were requesting..
Peter!?
:)
this explains the rash of failures I've seen reported recently
with "too many DMA segments" on raw devices
(added one for st as well)
o don't use polled mode after the device probe phase
o don't biodone() a NULL buffer pointer
o increase the timeout for scsi_read_capacity(); WORMs are slooow
o make WORMMs known to scsiconf at all
This brings the driver in a state where it at least doesn't
immediately panic, nor hangs the controller any more. Unfortunately,
at least the YAMAHA CDR100 i've been testing with answered my write
attempts with an "Incorrect command sequence" response. Perhaps other
CD burners might work however.
Reviewed by: dufault
This change forces the controller drivers to allocate a scsibus_data struct
via a call to scsi_alloc_bus(), fill in the adapter_link field, and optionally
modify any other fields of the struct. Scsi_alloc_bus() initializes all fields
to the default, so the changes in most drivers are very minimal. For drivers
that support Wide controllers, the maxtarg field will have to be updated to
allow probing of all targets (for an example, look at the aic7xxx driver).
Scsi_attachdevs() now takes a scsibus_data* as its argument instead of an
sc_link*. This allows us to expand the role of the scsibus_data struct for
other bus level configuration setings (max number of transactions, current
transaction opennings, etc for better tagged queuing support).
Reviewed by: Rodney Grimes <rgrimes>, Peter Dufault <dufault>, Julian Elischer <julian>
for the kernel, but gcc provides an inline version of it if the
kernel is compiled with -O.
The inline memcmp() is OK for small compares and is better than
the dumb kernel bcmp() in all cases, but it has been hiding the
library memcmp() which is 4 times faster for large compares.
This is performed by using a line similar to:
controller scbus0 at ahc0 bus 1
to wire scbus0 to the second bus on an adaptec 2742T controller.
Reviewed by: Peter Dufault(dufault@hda.com), Rod Grimes(rgrimes@FreeBSD.org)
probes). Apart from there being no reason to set SCSI_NOSLEEP on every
tape command, this prevents controller drivers from sleeping when resources
are fully utilized causing unecessary "Oops not queued" errors. This is
only noticed for controllers that can run out of resources like the
27/2842 adaptec controllers. Before this fix, it is almost impossible to
perform extended tape operations if more than one scsi disk is on the
bus with the tape drive with these controllers. This does not address a
similar problem that could occur if devices are probed while other targets
are active since SCSI_NOSLEEP will still be set in that case.
>Synopsis: Booting w/scsi tape in drive causes first use to fail
Booting with a tape in a SCSI tape drive will cause the first
use of the tape to fail with the following message:
st0: bad request, must be between 0 and 0.
Submitted by: mpp@legarto.minn.net (Mike Pritchard)
loading kernel
worm.o: Undefined symbol `_STUNIT' referenced from text segment
I copied the STUNIT definition from the old scsiconf.c into this file to
work around this problem.
Added CONTROL device that only does user-ioctl and nothing else;
Added protection so user-ioctl requires write access;
Clean up scsiconf.h a little. It needs more work.
user-level part has already been commited.)
Note that i've lost the "official" code for this; it went into the
system after 1.1.5.1. The commited code is my own version, but it has
proven to work for me for more than a year now.
flags parameter to all xxstart routines so that the correct information can
be passed down into the device specific routines. This is needed to ensure
that ccb/scb allocation routines don't hang.
Submitted by: John Dyson
function calls. Broken tracking of open partitions gets fixed as a side
effect. Races in sdopen() and sdioctl() become more numerous than before.
Close the device in sdsize(). Media removal could not have worked right
when unused swap devices were held open. There may now be a problem
with used swap devices not being held open.
Fix ancient operator precedence bug in sddump(). It may have been
possible to attempt dumping to unloaded media.
Remove NetBSD ifdefs. We're too different now, and the NetBSD driver was
already too different for them to be relevant.
Remove support for soft write protection. There are no ioctls to set it,
and this driver never set it either. It was used in the wd driver to
disable writing to unlabelled disk, but we want to support writing to
foreign partitions on unlabeled disks.
Clean up some formatting and restore some old-style function headers.
sd.c: Julian's removal of subdriver requests
st.c, scsi_tape.h: Julian's suport of compressed tape drives
Note: compressed tape drives are still not working fully.
scsiconf.h, scsi_base.c, scsi_driver.c: address problems in probes
and error console logs
Fix it so that calls to extend_get won't dereference 0 when "swap
on sd.." calls sdsize() on systems with no SCSI devices.
This probably really fixes Poul's "boot with no SCSI" problem.
tape drive. I traced it to sys/scsi/scsi_base.c where some code were added
to print the return values of the sense command. My tape drive returned a
extended flags value of 0x20, so the key that is the lower 4 bits, is 0.
The code uses "key-1" to index into a table and then the kernel go to never-
never land.
Here is a fix for this. Will somebody please apply it?
Submitted by: John Hay
Reenabled "SCIOCOMAND" ioctl.
Restructured so low level drivers can easily request retries.
Added preliminary fixed SCSI devices (should be revisited before 2.1)
Added "ssc" device that can have its' (HBA, ID, LUN) set via ioctl.
Change the sc_link structure from 8 targets with 8 lun's to
16 targets with 8 lun's. Wide scsi and the ncr 825 does support 16 targets.
scsiconf.c:
move the addition of the trailing NUL's out of a then clause, they are also
needed in the else clause for an unknown device.
scsi adapter. Let it work on every adapter. Someone want to rewrite
this to be coupled with the adapter specification in the config file,
so that you can say which adapters need it ?
device announcement; assume a sector size of 512 instead (likely to be
right at all).
This case happens when booting with a removable disk device attached
(e.g. an MOD), but no medium inserted.
and all SCSI devices (except that it's not done quite the way I want). New
information added includes:
- A text description of the device
- A ``state''---unknown, unconfigured, idle, or busy
- A generic parent device (with support in the m.i. code)
- An interrupt mask type field (which will hopefully go away) so that
. ``doconfig'' can be written
This requires a new version of the `lsdev' program as well (next commit).
drives that return sector counts of 0 and cause division by 0 traps during
the probe).
Reviewed by: Dave <root@hclb.demon.co.uk>
Obtained from: FreeBSD 1.1.5.1
Changes are only effective, if NEW_SCSICONF is defined:
1) Added u_int16 quirks to struct scsi_link for device quirks.
2) Added void *devmodes to struct scsi_link for tape density modes.
3) Modified selectdev() to return first full match (with wildcard).
With this patch all device dependent configuration is done in one place
(scsiconf.c), while there were some specifications in st.c and some
hardware dependent drivers (ie. ncr.c) before.
Calculation of 'cd->cmdscount' now as in 'sd.c' (required for ncr scsi).
CDOUTSTANDING defined as '1' (as the comment says), was '2' in fact ...
Debugger takes an (char*) argument, changed macro definition accordingly.
with BOUNCE_BUFFERS. This is more intuitive, and is better for future
multiplatform support. Added BOUNCE_BUFFERS option to the GENERIC and
LINT kernel config files.
in your kernel config now).
2) Added ps ddb function from 1.1.5. Cleaned it up a bit and moved into its
own file.
3) Added \r handing in db_printf.
4) Added missing memory usage stats to statclock().
5) Added dummy function to pseudo_set so it will be emitted if there
are no other pseudo declarations.
use of timeout_t -> timeout_func_t in aha1542 and aha1742 drivers.
2) fix a bug in the portalfs that was uncovered by better prototyping -
specifically, the time must be converted from timeval to timespec
before storing in va_atime.
3) fixed/added some miscellaneous prototypes
``changes'' are actually not changes at all, but CVS sometimes has trouble
telling the difference.
This also includes support for second-directory compiles. This is not
quite complete yet, as `config' doesn't yet do the right thing. You can
still make it work trivially, however, by doing the following:
rm /sys/compile
mkdir /usr/obj/sys/compile
ln -s M-. /sys/compile
cd /sys/i386/conf
config MYKERNEL
cd ../../compile/MYKERNEL
ln -s /sys @
rm machine
ln -s @/i386/include machine
make depend
make
1) check va before clearing the page clean flag. Not doing so was
causing the vnode pager error 5 messages when paging from
NFS. (pmap.c)
2) put back interrupt protection in idle_loop. Bruce didn't think
it was necessary, John insists that it is (and I agree). (swtch.s)
3) various improvements to the clustering code (vm_machdep.c). It's
now enabled/used by default.
4) bad disk blocks are now handled properly when doing clustered IOs.
(wd.c, vm_machdep.c)
5) bogus bad block handling fixed in wd.c.
6) algorithm improvements to the pageout/pagescan daemons. It's amazing
how well 4MB machines work now.
1) fixed some bugs related to the bounce buffer code
2) vnode pager now supports clustered pageouts
3) experimental code for clustering all I/O via a new "cldisksort"
4) added >16MB check to Bustek driver
5) made some experimental algorithmic changes to the pageout daemon
6) fixed bugs in truncating mapped files (esp when mapped via NFS)
7) reorganized vnode pager I/O code
a zero cylinder and/or head count, set them and the sector count
to some fictious values. The same already happens if the request
sense is unsuccessfull.
>From: Adam David <adam@veda.is>
Subject: SCSI timeout in sdopen()
Date: Sun, 6 Feb 1994 21:40:48 GMT
Index: sys FreeBSD-current
Reference: FreeBSD-current/sys/9
Description:
sdopen() times out on slow SCSI devices. This code was moved into
a different file, which might explain why it got missed again.
Repeat-By:
using slow SCSI disks
cd.c:
Initialize channel info in CDIOCSETVOL ioctl.
Correct CDIOCSTOP and CDIOCEJCET ioctls to use scsi_stop_unit
instead of scsi_start_unit.
Add CDIOCALLOW and CDIOCPREVENT ioctls.
ch.h:
Return EBUSY instead of ENXIO if the device is already in use.
scsi_base.c:
Add scsi_stop_unit routine.
sd.c:
Add mising indirection through sc_link to sd_get_parms routine
when checking for media loaded.
st.c:
Return EBUSY instead of ENXIO if the device is already in use.
Clear the SDEV_WAITING flag in ststart if we do the wakeup call.
a binary link-kit. Make all non-optional options (pagers, procfs) standard,
and update LINT to reflect new symtab requirements.
NB: -Wtraditional will henceforth be forgotten. This editing pass was
primarily intended to detect any constructions where the old code might
have been relying on traditional C semantics or syntax. These were all
fixed, and the result of fixing some of them means that -Wall is now a
realistic possibility within a few weeks.
Here is the fix for the 'hanging' bug.
This bug happenned whenever two operations were already underway on the
disk and a third (non-IO) command was requested..
in this case the process submitting the NON-IO command was requested to wait,
and a flag set so that on completion of the IO commands, the Non-io command
was given priority over any pending IO commands. (the queue is not
allowed to drain while there are pending "special" ops).
The flag that indicated this was not being reset, so further IO commands
were prohibited from that moment on.
Date: Wed, 20 Oct 1993 18:28:44 +0300
Editors Note:
Line ended with comma (``,'') instead of semicolon (``;''),
this should not have effected the code.
Submitter made no comments other than it was ``simple''.
Subject: Re: Error reading from DAT (fwd)
Date: Mon, 11 Oct 1993 03:02:39 -40962758 (WST)
Some tapes won't do anything until you tell them to load.
Other tapes think that being asked to load is a great excuse to
make the next command return a "unit attention" (media changed) error.
Since we need to do the load, throw in a dummy instruction to cop the
error.
specifically for DEC DAT drives.
>Date: Sat, 11 Sep 93 12:59:39 +0800
1/ fix bug where cd0a was unusable because it was compared against
the number of 2k blocks not the number of 512byte blocks.. gave
an error for all reads past 1/4 way through the disk (cd0d was ok).
2/ fix open code so that it now notices if a disk has been changed
and updates the disklabel.
3/ use the new error handling code from st.c
4/ start adding stuff to get ready for using cd_scsi_cmd from
cd_start (like is done in st and sd)
5/ more general cleanups including making dmesg report
sector size of disk.. (believe it or not SUN drives use 512 byte)
in your config file from:
device cd0
device cd1
...
device cdn
to either
device cd0
or
psuedo-device cd0
The driver now dynamically allocates all per unit dependent structs, so it
will never run out of units (okay, so now we need 32 bit device major/minor
numbers!). It will allocate unit numbers as it needs them.
data. cd_data is now malloced on a per unit basis.
Protected all if ({scsi,cd}_debug).... with #ifdef CDDEBUG to reduce the
the size of the kernel when not debugging.
