1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-24 11:29:10 +00:00
Commit Graph

11797 Commits

Author SHA1 Message Date
Justin T. Gibbs
ebeec632cd Correct a bug in the timeout handler. When queuing a BDR response to
a timeout, we must remove the pending SCB from the disconnected list
or risk list corruption when our BDR request using the same SCB is placed
on the waiting list.

Eradicate some silly uses of u_int8_t that just serve to slow the code down.
1998-09-20 05:06:10 +00:00
Justin T. Gibbs
c279502fc6 Correct compilation with option DIAGNOSTIC set.
Upgrade to the latest firmware.
1998-09-20 05:04:05 +00:00
Justin T. Gibbs
ecdf111388 Only deregister out configuration hook manually if there are no SCSI
busses to configure.
1998-09-20 05:03:34 +00:00
Bruce Evans
a6796db6ce Ensure that the i8254 timecounter doesn't go backards. It sometimes
went backwards when interrupts were masked for more than one i8254
interrupt period.  It sometimes went backwards when the i8254 counter
was reprogrammed.  Neither of these should happen in normal operation.

Update the i8254 timecounter support variables atomically.  Calling
timecounter functions from fast interrupt handlers may actually work
in all cases now.
1998-09-20 03:47:54 +00:00
John Polstra
cc59ec0b9e Null-terminate the message for a warning symbol in the __warn_references()
macro for ELF.  Ian@cygnus.com says that is the proper way to use
the ".gnu.warning.SYMBOL" sections.  However, he has accepted Doug
Rabson's BFD patch anyway.
1998-09-20 03:24:55 +00:00
Kenneth D. Merry
bcc6a3da92 Change the devstat generation number from an int to a long. The int-sized
generation was causing unaligned access faults on the Alpha.

I have incremented the devstat version number, since this is an interface
change.  You'll need to recompile libdevstat, systat, iostat, vmstat and
rpc.rstatd along with your kernel.

Partially Submitted by:	Andrew Gallatin <gallatin@cs.duke.edu>
1998-09-20 00:11:23 +00:00
John Polstra
dac949d859 Fix breakage of ELF kernel builds caused by the addition of an
assembly language reference to a C symbol without the addition of
the corresponding entry in asnames.h.
1998-09-19 17:16:48 +00:00
KATO Takenori
e185bae6e1 Sync with sys/i386/conf/SMP-GENERIC revision up to 1.15. 1998-09-19 13:36:13 +00:00
KATO Takenori
770e1ffc56 Sync with sys/i386/conf/GENERIC revision 1.118. 1998-09-19 13:35:18 +00:00
Jordan K. Hubbard
534484343b Note that we're in BETA status. 1998-09-19 11:45:37 +00:00
Doug Rabson
4e6823805e Add splsoftvm().
Submitted by: Andrew Gallatin <gallatin@cs.duke.edu>
1998-09-19 09:29:40 +00:00
Justin T. Gibbs
9eec0f7d5a Don't leave the device queue in a frozen state if the Synchronize Cache
command on close fails.
1998-09-19 04:59:35 +00:00
Mike Smith
c9758db4ad Make libstand movable.
This completes the basic work required to make the loader work with BTX. We
now have a third-stage bootstrap shared by the i386 and Alpha.
1998-09-19 01:35:53 +00:00
Mike Smith
b6dc8cf656 - Improve debugging code.
- Make the "what do we do with a drunken disklabel" if-then-else-regardless
   tangle easier to read.
 - Don't count on the v86 structure being preserved between loop iterations,
   as it may be trampled eg. by the DEBUG call.
1998-09-19 01:33:29 +00:00
Mike Smith
acd1d918fe Add a generic hexdump tool for debugging purposes. 1998-09-19 01:31:28 +00:00
Kenneth D. Merry
66411419a6 Fix error recovery in scsi_interpret_sense(). It turns out that ERESTART
wasn't getting sent back for most errors, even if there were retries left
on the command.  I'm not sure how I ever let this slip by before...

In any case, we now send back ERESTART if there are retries left for the
command, and send back the default error code when there are no retries
left.

Reviewed by:	gibbs
1998-09-19 01:23:04 +00:00
Justin T. Gibbs
0532933e60 Set QUIRK_NOMSG for all devices as was the case in the non-CAM version
of the driver.  Why this is necessary (and how identify message handling
is performed when this is set) I don't comprehend...
1998-09-18 22:41:12 +00:00
Kenneth D. Merry
37b9efd37b Fix the CAM code so that people can compile kernels with the CD driver but
without the DA driver.

The problem was that the CD driver depended on scsi_read_write() and
scsi_start_stop(), which were defined in scsi_da.c.

I moved both functions, and their associated data structures and defines
from scsi_da.* to scsi_all.*.  This is technically the "wrong" thing to do
since those commands are really only for direct-access type devices, not
for all SCSI devices.  I think, though, that the advantage (allowing people
to compile kernels without the disk driver) outweighs any architectural
purity arguments.

PR:		kern/7969
Reviewed by:	gibbs
1998-09-18 22:33:59 +00:00
Luigi Rizzo
65d5a8999d fix an mbuf leak when using ipfw to filger bridged packets
(from -stable, since this code is not yet active in -current)
1998-09-18 20:55:50 +00:00
Eivind Eklund
7c0daaa801 Make LINT compile and link again after the CAM merge. The little
annoying #!CAM# indicators are used to be clear, in the expectation
that the places they show will be either fixed or diked out reasonably
quickly.

Reviewed by:	ken
1998-09-18 20:17:41 +00:00
Kenneth D. Merry
6747168dc9 Change the Atlas II quirk entries so they work with differential Atlas
II's.  Also, add a quirk entry for the 2 gig Atlas II.

Partially Submitted by:	Ted Buswell <tbuswell@mediaone.net>
1998-09-18 19:55:34 +00:00
Matt Jacob
090040e092 Bump request queue size up to full amount (now that we have
a contiguous CCB memory allocation).
1998-09-18 18:58:07 +00:00
Doug Rabson
4ada0d2731 Change sd to da and comment out the non-cam TurboChannel scsi drivers.
Fix a typo which prevented VGA consoles from working.
1998-09-18 18:43:52 +00:00
Doug Rabson
c6f34b8987 Change version number. 1998-09-18 18:41:49 +00:00
Doug Rabson
2a8d2696cd Cam changes sd to da. 1998-09-18 18:40:59 +00:00
Justin T. Gibbs
a90c881ed7 Don't interpret SCB index 254 as a target mode completion when not in
target mode.  This hack of using the normal command complete mechanism
for target cdb completion will go away soon.

Submitted by:	Tor.Egge@fast.no
1998-09-18 03:42:16 +00:00
Mike Smith
6bc32d7d5e Enable the biosdisk driver, duplicate -lstand as it both calls and is called
by the i386 platform library.
1998-09-18 02:03:30 +00:00
Mike Smith
819ab203c3 Synch with development version. Compiles and opens but doesn't work yet. 1998-09-18 02:02:33 +00:00
Mike Smith
c676c40b0d Remove 'panic' command (it works), don't ask for a keypress in panic, as
exit() does.  Perhaps it shouldn't?
1998-09-18 02:01:38 +00:00
Mike Smith
b521106fc5 We lost all the files in crt/, so define the BIOS sector size here instead. 1998-09-18 01:12:46 +00:00
Mike Smith
a3e06d5cc1 Use a.out.h to get all the correct bits in one place. 1998-09-18 01:12:23 +00:00
Matt Jacob
3dd37e4387 (requested by gibbs) Remove the SCSI_CAM option (and rework the isp driver
that had depended on it for compilation within or without CAM to use
__FreeBSD_version instead).
1998-09-18 00:46:42 +00:00
Mike Smith
6d10d056f3 Oops, missed these. Machine-independant ISA PnP enumerator. 1998-09-18 00:24:25 +00:00
Kenneth D. Merry
5623443753 Move SCSI_DELAY and SCSI_CAM from the undocumented options section to the
CAM options section.

Document that SCSI_DELAY is in milliseconds, not seconds.

Tell users that SCSI_CAM is only needed if you've got the QLogic driver in
your kernel.

Reviewed by:	gibbs
1998-09-18 00:01:39 +00:00
Kenneth D. Merry
a822eb2a8b Fix a formatting error.
Fix a problem reported by bde:  setting SCSI_DELAY to 0 doesn't work.  Now,
when the user sets SCSI_DELAY to 0, we re-set it to the minimum allowable
bus settle delay (100ms).

Fix a potential panic in xptfinishconfigfunc() if the CCB passed in is
NULL.  Reported by, I think, Nicolas Souchu.  Fix a memory leak in the same
function (we created a path, but didn't free it) by allocating the getdev
CCB and path on the stack.

Reviewed by:	gibbs
1998-09-17 23:58:53 +00:00
Mike Smith
948486abe3 Initial integration of the i386 bootloader and BTX.
- Discard large amounts of BIOS-related code in favour of the more compact
   BTX vm86 interface.
 - Build the loader module as ELF, although the resulting object is a.out,
   make gensetdefs 32/64-bit sensitive and use a single copy of it.
 - Throw away installboot, as it's no longer required.
 - Use direct bcopy operations in the i386_copy module, as BTX
   maps the first 16M of memory.  Check operations against the
   detected size of actual memory.
1998-09-17 23:52:16 +00:00
Matt Jacob
a70ac9fd83 per bde (who is right about this) that an inlined fucntion with const
char * strings being returned defined in a header file included several
places but only used in one module, is, uh, silly.
1998-09-17 23:20:29 +00:00
Matt Jacob
7ac6b5a314 Cleanliness. Don't leave defined a const char array that's only used
if target mode is defined (which it isn't, yet).
1998-09-17 22:53:35 +00:00
Justin T. Gibbs
761fe626ea Really correct ncr_freeze_devq now. We scan backwards from the current
insertion point into the start queue looking for entries to remove and
mark them with the 'skip' address, recording the entry furthest from the
insertion point that needs to be removed.  We then go through a second
loop starting at the furthest entry to be removed and compress the start
queue.  The old algorithm started at (old insert point + 1) and wrapped
through the whole queue which would end up moving the start position in
the queue out from under the nose of the scrip processor.
1998-09-17 22:29:02 +00:00
Mike Smith
4e138a28a0 Add support for glibc consumers using the new ld-linux.so.2 linker.
I can't say this was the most inspired fix, but it matches the design OK.
1998-09-17 22:08:34 +00:00
Nicolas Souchu
f7d09fba58 vpo comments updated for cam, nlpt suggested instead of lpt 1998-09-17 21:54:17 +00:00
Matt Jacob
4873663cf8 A major amount of cleaning up:
+ Change some messages about CCB memory allocation
	+ Turn a failure to DMA map all of a transaction due to lack of
	  ISP queue entries into a requeue operation (instead of the
	  case where it had been treated the same as a DMA too big
	  operation).
	+ put back splsoftvm around bus_dmamap_load calls.
	+ cleanup (and fix a glaring bug) in the and of the dma setup
	  routine. Also, the dma setup routines either return CMD_QUEUED
	  (for success) or CMD_COMPLETE (for failure) or CMD_EAGAIN
	  (for requeuing for resource shortage reasons).
1998-09-17 21:11:21 +00:00
Matt Jacob
0c1b701510 Clean up a comment. Make the default debugging level contingent upon CAMDEBUG. 1998-09-17 21:06:38 +00:00
Matt Jacob
c6904df981 Remove commented out and no longer pertinent EAGAIN message. Redo CMD_COMPLETE
logic for returns from inner layer (just make sure that a CAM_REQ_INPROG
doesn't get xpt_done'd).
1998-09-17 21:05:40 +00:00
Matt Jacob
42b7427595 ISP_DMASETUP now returns a value to be possibly punted to outer layers.
Turn request queue overflow messages into debug messages. Ensure on
isp_restarts that we nullify the xflist array.
1998-09-17 21:03:45 +00:00
Matt Jacob
4d65e73e67 Roll revision, clean up a comment. 1998-09-17 21:02:18 +00:00
Søren Schmidt
d74e86d9e3 Fix the sreensavers so the work again with the new syscons & friends.
Submitted by: Kazutaka YOKOTA <yokota@zodiac.mech.utsunomiya-u.ac.jp>
1998-09-17 19:40:30 +00:00
Robert Nordier
131bc817a0 Correct description of get_diskinfo result. 1998-09-17 18:43:18 +00:00
Bill Fenner
52b65dbe85 Fix the bind security fix introduced in rev 1.38 to work with multicast:
- Don't bother checking for conflicting sockets if we're binding to a
  multicast address.
- Don't return an error if we're binding to INADDR_ANY, the conflicting
  socket is bound to INADDR_ANY, and the conflicting socket has SO_REUSEPORT
  set.

PR:		kern/7713
1998-09-17 18:42:16 +00:00
KATO Takenori
7ef16b04e6 Sync with sys/i386/i386/userconfig.c revision 1.109. 1998-09-17 13:14:14 +00:00