This patch series is against scsi-misc-2.5.

These patches modify the scsi error handler to process cmds needing
recovery off a list. The error handler policy has been altered to do the
following:
	1.) Check for legacy behavior of requesting sense.
	2.) Abort commands marked needing to be canceled.
	3.) Ready devices through tur and eh handlers.
	4.) disposition each command on the list to be retried or
	finished.

00_serror-cmd-list-1.diff:
	- Add per host eh_cmd_list list head.
	- Add per cmd eh_list list head.

01_serror-scmd-add-1.diff:
	- Add scsi_eh_scmd_add function.

02_serror-hndlr-1.diff:
	- Change to using eh_cmd_list.
	- Change scsi_unjam_host to get sense, abort cmds, ready
	  devices, and disposition cmds for retry or finish.
	- Moved retries outside of eh.

03_serror-dev-offline-1.diff:
	- Add scsi_set_device_offline interface.

I have tested both the timeout and offline cases with the scsi_debug. I
have ran some other testing, but my machine stability is causing it to
oops in other areas.

-andmike
--
Michael Anderson
andmike@us.ibm.com

 hosts.c |   16 +---------------
 hosts.h |    2 +-
 scsi.h  |    9 +++++----
 3 files changed, 7 insertions(+), 20 deletions(-)

Add include of linux/interrupts from Rolf Eike Beer <eike-kernel@sf-tec.de>

Improve the interrupt probing so it works for CPQ4411

  http://linux.bkbits.net:8080/linux-2.5/diffs/drivers/scsi/aic7xxx/aic79xx_osm.c@1.16?nav=index.html|src/|src/drivers|src/drivers/scsi|src/drivers/scsi/aic7xxx|hist/drivers/scsi/aic7xxx/aic79xx_osm.c

    This change(latest aic79xx driver and scsi_cmnd changes merging)
    is dropped `hscb->lun = cmd->device->lun;' (in aic79xx_osm.c line
    4272). This change cause lun detect problem. I believe it is still
    needed.

Last changelog entry is from 1996 and it had no changes ecept global
search and replace since start of the bitkeeper history.

The alternate driver eata is maintained and the Kconfig file contains
a stong advice on using it.

I think this patch takes care of scsi/imm.c build errors
in 2.5.60.

Please review and apply if correct.

For kernel bugzilla #330.

Both of these patches are originally by Grant Grundler.  They've been
tested in the PA-RISC tree for a few months, and I'm forwarding them along.

First patch:

pci_set_dma_mask returns 0 if the card can perform DMA as specified,
so the sense of the test is wrong.

Try to keep all queuing-related stuff in one file and sanitize global
symbols.  BTW, any chance you could bk rename scsi_lib.c to
scsi_queue.c?

This is userland testing code that leaked into the kernel tree somehow.

James, your sim710 changes look nice, but I think there's some more
stuff to do in this driver.  I've started with the following (all
untested due to lack of the hardware):

  - move to scsi_add_host/scsi_remove_host interface
  - remove a bunch of unneeded ifdefs
  - add missing statics
  - unregister driver templates on unload

and I think there's more that should be done, but this would change
user-visiable attributes:

  - remove the ugly single module option code and the command line
    parsing in favour of Rusty's new module_param stuff
  - split the driver into two drivers: 53c700_eisa and 53c700_mca.
    There's no logic shared between those two busses, just a little
    bit of helper in the setup/remove code

And after looking a this driver I have some rants about the new
mac/eisa code:

  - mca_register_driver/mca_unregister_driver should check for
    the precense of an MCA bus by themselves instead of leaving
    it to the caller
  - eisa_driver_register should really return 0 for sucess
  - mca_register_driver/mca_unregister_driver should be named
    mca_driver_register/mca_driver_unregister to be more similar
    to the other *driver_(un)registers.

Fix cmd->lun to cmd->device->lun

Most of the instances of <module>.o in Kconfig files have had the ".o"
extension trimmed.  This change came from GertJan Spoelman through Rusty
"Trivial" Russell.

However, there are a few files that didn't get trimmed.  This brings
them line with the rest of the tree.

so that there isn't any window for running before the signal
handler has been invoced.

from certain states.

This simplifies "default_wake_function()", and makes it possible for
signal handling to wake up only the processes it _should_ wake up
without races.

From:  GertJan Spoelman <kl@gjs.cc>

  OK, here is a new patch, I edited the old patch and took out the .ko's
  so now the extension is trimmed instead.

[ Acked by AKPM --RR ]
From:  Kazuto MIYOSHI <miyoshi@hpc.bs1.fc.nec.co.jp>

  On 64-bit platforms, issuing write(2) with buffer larger than
  2GB will return -1 and broken errno (such as 2147483640)
  Requested data itself is written correctly.

  That is because generic_file_write() and other relating functions
  store 'ssize_t written' into 'int err'. Written byte is trimmed to
  int and then sign-extended to a negative ssize_t value, which
  wrongly indicates an error.

  (On 64bit platform, current glibc defines SSIZE_MAX as 'LONG_MAX')

From:  Steven Cole <elenstev@mesatop.com>

  Here are some help texts from 2.4.21-pre3 Configure.help which are
  needed in 2.5.59 drivers/media/video/Kconfig.

From:  Pablo Menichini <pablo@menichini.com.ar>

From:  Steven Cole <elenstev@mesatop.com>

  Here are some help texts from 2.4.21-pre3 Configure.help which are
  needed in 2.5.59 drivers/char/watchdog/Kconfig.

From:  Steven Cole <elenstev@mesatop.com>

  Here are some help texts from 2.4.21-pre3 Configure.help which are
  needed in 2.5.59 drivers/ide/Kconfig.

From:  Pavel Machek <pavel@ucw.cz>

  This patch makes kernel ignore bad signature on suspend device when
  "noresume" is given, and cleans things up a little bit. Please apply,

[ Verified that no text changed with tr and cmp --RR ]
From:  ookhoi@humilis.net

  With this patch I tried to make Documentation/arm/SA1100/CERF more
  readible by fixing the linewrap.

From:  Steven Cole <elenstev@mesatop.com>

  This fixes the only instance of "kernal" in 2.5.59.

From:  David Gibson <david@gibson.dropbear.id.au>

  is never used, so this patch removes it.

From:  William Lee Irwin III <wli@holomorphy.com>

  This trivially corrects an unused variable warning in nfs/write.c:

From:  Adrian Bunk <bunk@fs.tum.de>

  drivers/net/tulip/de4x5.h in 2.5.54 contains a definition of
  LinuxVersionCode. LinuxVersionCode isn't used and it's anyway obsoleted
  by KERNEL_VERSION in version.h.

From:  Pavel Machek <pavel@ucw.cz>

  Second part of this patch never got in (and I was told it was not bug
  in ASUS but in linux), so it is useless junk... Please apply,

From:  Pablo Menichini <pablo@menichini.com.ar>

  	This patch returns correct error codes if init_modules fail.
  Because of this, we can take the printks indicating the error as these
  corrected error codes return miningfull information.

From:  Pablo Menichini <pablo@menichini.com.ar>

(Included in 2.4)
From:  Paul Gortmaker <p_gortmaker@yahoo.com>

  Summary: Wildcards in RTC alarm settings failed to work

  Description:
   The RTC has provision for wildcards when setting the alarm; to
   use them you have to write a value higher than 0xc0 to the
   appropriate hr/min/sec entry.  The driver used 0xff, which is
   fine, but it mistakenly fed the 0xff through BIN_TO_BCD before
   writing them (which is < 0xc0) and so wildcards didn't work.
   (Thanks to Gerhard Kurz for reporting the bug.)

From:  Pablo Menichini <pablo@menichini.com.ar>

From:  Muli Ben-Yehuda <mulix@mulix.org>

  The i2c driver_lock is needlessly exported.
  This makes it static.

From:  Bob Miller <rem@osdl.org>

  Include string.h to remove a compiler warning.

From:  Steven Cole <elenstev@mesatop.com>

  Here are some help texts from 2.4.21-pre3 Configure.help which are
  needed in 2.5.59 drivers/i2c/Kconfig.

From:  Pablo Menichini <pablo@menichini.com.ar>

From:  Frank Davis <fdavis@si.rr.com>

     The ver_linux script is still using rmmod to determine
  module-init-tools version. The following patch uses depmod,
  which produces the appropriate result.