pandora-kernel.git
18 years ago[IA64] Update tiger_defconfig
Tony Luck [Wed, 29 Jun 2005 22:45:41 +0000 (15:45 -0700)]
[IA64] Update tiger_defconfig

Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years agoAuto merge with /home/aegl/GIT/linus
Tony Luck [Wed, 29 Jun 2005 22:21:41 +0000 (15:21 -0700)]
Auto merge with /home/aegl/GIT/linus

18 years ago[PATCH] Fix get_request nastiness
Hugh Dickins [Wed, 29 Jun 2005 14:15:40 +0000 (15:15 +0100)]
[PATCH] Fix get_request nastiness

get_request is now expected to be holding on to queue_lock, with interrupts
disabled, when it returns NULL; but one path forgot that, causing all kinds
of nastiness under swap load - badness backtraces, strange failures, BUGs.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoRemove duplicate mention of "edd" in Documentation/kernel-parameters.txt
Linus Torvalds [Wed, 29 Jun 2005 15:43:49 +0000 (08:43 -0700)]
Remove duplicate mention of "edd" in Documentation/kernel-parameters.txt

18 years agoLinux v2.6.13-rc1 v2.6.13-rc1
Linus Torvalds [Wed, 29 Jun 2005 05:57:29 +0000 (22:57 -0700)]
Linux v2.6.13-rc1

Ok, a lot of things were pending after the 2.6.12 release, let's try to
start calming things down again.

18 years ago[PATCH] ieee80211.h build fix
Jeff Garzik [Wed, 29 Jun 2005 05:17:55 +0000 (01:17 -0400)]
[PATCH] ieee80211.h build fix

This crept in with the resync-to-mainline.  Nothing uses 802.11-crypt in
mainline, so we can safely comment it out for now.

Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge rsync://rsync.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 29 Jun 2005 04:24:32 +0000 (21:24 -0700)]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

18 years ago[PATCH] sis 760 support.
Dave Jones [Wed, 29 Jun 2005 00:08:29 +0000 (20:08 -0400)]
[PATCH] sis 760 support.

This patch adds the SiS 760 ID to the amd64-agp driver, so that agpgart can be
used on Athlon64 boards based on this chip.

Signed-off-by: Daniel Drake <dsd@gentoo.org>
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] sparc32: Kconfig fixups
William Lee Irwin III [Wed, 29 Jun 2005 03:45:31 +0000 (20:45 -0700)]
[PATCH] sparc32: Kconfig fixups

Something reverted most of the arch/sparc/Kconfig changes, leaving
arch/sparc/ unconfigurable.  This patch re-removes the parts made redundant
by drivers/Kconfig in addition to a mysterious, spurious second instance of
source "mm/Kconfig".  cvs strikes again?

Signed-off-by: William Irwin <wli@holomorphy.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Doc/Submitting: corrections, additions
Randy Dunlap [Wed, 29 Jun 2005 03:45:30 +0000 (20:45 -0700)]
[PATCH] Doc/Submitting: corrections, additions

Corrections to Documentation/Submitting{Drivers,Patches}
- update LANANA info.
- fix some typos
- update 2.2 kernel maintainer info.
- update 'dontdiff' info.
- update URLs for patch scripts
- add Trivial Patch Monkey URL
- add more references for submitting patches

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] V4L: documentation changes - mostly new cards included
Mauro Carvalho Chehab [Wed, 29 Jun 2005 03:45:28 +0000 (20:45 -0700)]
[PATCH] V4L: documentation changes - mostly new cards included

New cards included.
V4L1 api renamed. Message included informing it is obsoleted by V4L2 API.
V4L2 api included.
Mark all 7135 cards as 7133.

Signed-off-by: Luc Saillard <luc@saillard.org>.
Signed-off-by: Nickolay V Shmyrev <nshmyrev@yandex.ru>
Signed-off-by: Hermann Pitton <hermann.pitton@onlinehome.de>
Signed-off-by: Michael Krufky <mkrufky@m1k.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@brturbo.com.br>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] V4L: API new webcam formats included
Mauro Carvalho Chehab [Wed, 29 Jun 2005 03:45:27 +0000 (20:45 -0700)]
[PATCH] V4L: API new webcam formats included

Add Philips Webcam format.

Signed-off-by: Mauro Carvalho Chehab <mchehab@brturbo.com.br>
Signed-off-by: Luc Saillard <luc@saillard.org>.
Signed-off-by: Nickolay V Shmyrev <nshmyrev@yandex.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] v4l: bttv new insmod parameters
Mauro Carvalho Chehab [Wed, 29 Jun 2005 03:45:25 +0000 (20:45 -0700)]
[PATCH] v4l: bttv new insmod parameters

* bttv-driver.c, bttvp.h:

- New bttv module params:

- uv_ratio : allow a ratio of saturation between u and v. If you
        have a ratio of 40 and a saturation of 100, usat will be 80 and
        vstat 120. Useful to correct a bad color balance.
- full_luma_range : provide a better contrast in using the full
        range 0-253 of values instead of 16-253.
- coring : to have a better black level.
- radio range is now defined on tuner-core.c. Cleaning up.

* bttvp.h:

- Fix gcc 4.0 compilation

Signed-off-by: Jorik Jonker <jorik@dnd.utwente.nl>
Signed-off-by: Sylvain Meyer <sylvain.meyer@worldonline.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@brturbo.com.br>
Signed-off-by: Nickolay V Shmyrev <nshmyrev@yandex.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] v4l: tuner improvements
Mauro Carvalho Chehab [Wed, 29 Jun 2005 03:45:21 +0000 (20:45 -0700)]
[PATCH] v4l: tuner improvements

*tuner-core.c:
- some tuner_info msgs will be generated only if insmod opt
        tuner_debug enabled.
- Implemented tuner-core support for VIDIO_S_TUNER to allow
        changing mono/stereo mode
- Remove unneeded config options.
- I2C_CLIENT_MULTI option removed.
- support for Philips FMD12ME hybrid tuner
- allow to initialize with another tuner
- Move PHILIPS_FMD initialization code to set_type function,

* tda8290:

- Fix dumb error in tda8290 tunning.
- Radio tuner uses high-precision step instead of 62.5 KHz.

*tea5767.c:
- tuner_info msgs will be generated only if insmod tuner option
        tuner_debug enabled.
- some cleanups for better reading.
- Radio tuner uses high-precision step instead of 62.5 KHz.
- Changing radio mode stereo/mono for tea5767 working.

*tuner-simple.c:
- TNF9533-D/IF UHF fixup.
- Radio tuners now uses high-precision step instead of 62.5 KHz.

*mt20xx.c:
        - Radio tuner uses high-precision step instead of 62.5 KHz.

*tda9887.c:
        - tab and blank spaces corrections.

Signed-off-by: Mauro Carvalho Chehab <mchehab@brturbo.com.br>
Signed-off-by: Gerd Knorr <kraxel@bytesex.org>
Signed-off-by: Nickolay V Shmyrev <nshmyrev@yandex.ru>
Signed-off-by: Hartmut Hackmann <hartmut.hackmann@t-online.de>
Signed-off-by: Michael Krufky <mkrufky@m1k.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] V4L maintainer patch
Mauro Carvalho Chehab [Wed, 29 Jun 2005 03:45:20 +0000 (20:45 -0700)]
[PATCH] V4L maintainer patch

This patch updates maintainer info for BTTV and V4L.

Signed-off-by: Mauro Carvalho Chehab <mchehab@brturbo.com.br>
Acked-by: Gerd Knorr <kraxel@bytesex.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] irqpoll
Alan Cox [Wed, 29 Jun 2005 03:45:18 +0000 (20:45 -0700)]
[PATCH] irqpoll

Anyone reporting a stuck IRQ should try these options.  Its effectiveness
varies we've found in the Fedora case.  Quite a few systems with misdescribed
IRQ routing just work when you use irqpoll.  It also fixes up the VIA systems
although thats now fixed with the VIA quirk (which we could just make default
as its what Redmond OS does but Linus didn't like it historically).

A small number of systems have jammed IRQ sources or misdescribes that cause
an IRQ that we have no handler registered anywhere for.  In those cases it
doesn't help.

Signed-off-by: Alan Cox <number6@the-village.bc.nu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ext3: reduce allocate-with-reservation lock latencies
Mingming Cao [Wed, 29 Jun 2005 03:45:16 +0000 (20:45 -0700)]
[PATCH] ext3: reduce allocate-with-reservation lock latencies

Currently in ext3 block reservation code, the global filesystem reservation
tree lock (rsv_block) is hold during the process of searching for a space
to make a new reservation window, including while scaning the block bitmap
to verify if the avalible window has a free block.  Holding the lock during
bitmap scan is unnecessary and could possibly cause scalability issue and
latency issues.

This patch tries to address this by dropping the lock before scan the
bitmap.  Before that we need to reserve the open window in case someone
else is targetting at the same window.  Question was should we reserve the
whole free reservable space or just the window size we need.  Reserve the
whole free reservable space will possibly force other threads which
intended to do block allocation nearby move to another block group(cause
bad layout).  In this patch, we just reserve the desired size before drop
the lock and scan the block bitmap.  This patch fixed a ext3 reservation
latency issue seen on a cvs check out test.  Patch is tested with many fsx,
tiobench, dbench and untar a kernel test.

Signed-Off-By: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] blk: light iocontext ops
Nick Piggin [Wed, 29 Jun 2005 03:45:15 +0000 (20:45 -0700)]
[PATCH] blk: light iocontext ops

get_io_context needlessly turned off interrupts and checked for racing io
context creations.  Both of which aren't needed, because the io context can
only be created while in process context of the current process.

Also, split the function in 2.  A light version, current_io_context does not
elevate the reference count specifically, but can be used when in process
context, because the process holds a reference itself.

Signed-off-by: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] blk: reduce locking
Nick Piggin [Wed, 29 Jun 2005 03:45:14 +0000 (20:45 -0700)]
[PATCH] blk: reduce locking

Change around locking a bit for a result of 1-2 less spin lock unlock pairs in
request submission paths.

Signed-off-by: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] blk: __make_request efficiency
Nick Piggin [Wed, 29 Jun 2005 03:45:13 +0000 (20:45 -0700)]
[PATCH] blk: __make_request efficiency

In the case where the request is not able to be merged by the elevator, don't
retake the lock and retry the merge mechanism after allocating a new request.

Instead assume that the chance of a merge remains slim, and now that we've
done most of the work allocating a request we may as well just go with it.

Also be rid of the GFP_ATOMIC allocation: we've got working mempools for the
block layer now, so let's save atomic memory for things like networking.

Lastly, in get_request_wait, do an initial get_request call before going into
the waitqueue.  This is reported to help efficiency.

Signed-off-by: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] coverity: tty_ldisc_ref return null check
KAMBAROV, ZAUR [Wed, 29 Jun 2005 03:45:12 +0000 (20:45 -0700)]
[PATCH] coverity: tty_ldisc_ref return null check

We add a check of the return value of tty_ldisc_ref(), which
is checked 7 out of 8 times, e.g.:

149   ld = tty_ldisc_ref(tty);
150   if (ld != NULL) {
151   if (ld->set_termios)
152   (ld->set_termios)(tty, &old_termios);
153   tty_ldisc_deref(ld);
154   }

This defect was found automatically by Coverity Prevent, a static analysis
tool.

(akpm: presumably `ld' is never NULL.  Oh well)

Signed-off-by: Zaur Kambarov <zkambarov@coverity.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] coverity: fs/ext3/super.c: match_int return check
KAMBAROV, ZAUR [Wed, 29 Jun 2005 03:45:11 +0000 (20:45 -0700)]
[PATCH] coverity: fs/ext3/super.c: match_int return check

The return value of  "match_int" is  checked  27 out of 28 times

In lib/parser.c
142   /**
143    * match_int: - scan a decimal representation of an integer from a substring_t
144    * @s: substring_t to be scanned
145    * @result: resulting integer on success
146    *
147    * Description: Attempts to parse the &substring_t @s as a decimal integer. On
148    * success, sets @result to the integer represented by the string and returns 0.
149    * Returns either -ENOMEM or -EINVAL on failure.
150    */
151   int match_int(substring_t *s, int *result)
152   {
153   return match_number(s, result, 0);
154   }

Signed-off-by: Zaur Kambarov <zkambarov@coverity.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] coverity: fs/udf/namei.c null check
KAMBAROV, ZAUR [Wed, 29 Jun 2005 03:45:10 +0000 (20:45 -0700)]
[PATCH] coverity: fs/udf/namei.c null check

"dir" was dereferenced before null check

Signed-off-by: Zaur Kambarov <zkambarov@coverity.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] coverity: ipmi_msghandler() channels array overrun fix
KAMBAROV, ZAUR [Wed, 29 Jun 2005 03:45:08 +0000 (20:45 -0700)]
[PATCH] coverity: ipmi_msghandler() channels array overrun fix

We fix the check in 1084, which was

1084  if (addr->channel > IPMI_NUM_CHANNELS) {
1085  spin_lock_irqsave(&intf->counter_lock, flags);
1086  intf->sent_invalid_commands++;
1087  spin_unlock_irqrestore(&intf->counter_lock, flags);
1088  rv = -EINVAL;
1089  goto out_err;
1090  }

addr->channel is used in

1092  if (intf->channels[addr->channel].medium

Definitions involved:

221   struct ipmi_channel channels[IPMI_MAX_CHANNELS];

134   #define IPMI_MAX_CHANNELS       8

In /linux-2.6.12-rc6/include/linux/ipmi.h
148   #define IPMI_NUM_CHANNELS 0x10

Signed-off-by: Zaur Kambarov <zkambarov@coverity.com>
Cc: Corey Minyard <minyard@acm.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] coverity: i386: scsi_lib buffer overrun fix
KAMBAROV, ZAUR [Wed, 29 Jun 2005 03:45:06 +0000 (20:45 -0700)]
[PATCH] coverity: i386: scsi_lib buffer overrun fix

The check in

627   BUG_ON(index > SG_MEMPOOL_NR);

with SG_MEMPOOL_NR defined in

32    #define SG_MEMPOOL_NR (sizeof(scsi_sg_pools)/sizeof(struct scsi_host_sg_pool))

was not sufficient.

sgp, set in

629   sgp = scsi_sg_pools + index;

is dereferenced in

630   mempool_free(sgl, sgp->pool);

Signed-off-by: Zaur Kambarov <zkambarov@coverity.com>
Cc: <linux-scsi@vger.kernel.org>
Cc: James Bottomley <James.Bottomley@steeleye.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] coverity: i386: build.c: negative return to unsigned fix
KAMBAROV, ZAUR [Wed, 29 Jun 2005 03:45:06 +0000 (20:45 -0700)]
[PATCH] coverity: i386: build.c: negative return to unsigned fix

Variable "c" was declared as an unsigned int, but used in:

125   for (i=0 ; (c=read(fd, buf, sizeof(buf)))>0 ; i+=c )
126   if (write(1, buf, c) != c)
127   die("Write call failed");

(akpm: read() can return -1.  If it does, we fill the disk up with garbage).

Signed-off-by: Zaur Kambarov <zkambarov@coverity.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] headers: include linux/types.h for usb_ch9.h
GOTO Masanori [Wed, 29 Jun 2005 03:45:05 +0000 (20:45 -0700)]
[PATCH] headers: include linux/types.h for usb_ch9.h

This patch for usb_ch9.h includes linux/types.h instead of asm/types.h so that
__le16 and so on is explicitly defined.  It also cleans up non standard //
comment.

Signed-off-by: GOTO Masanori <gotom@debian.or.jp>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] headers: include linux/compiler.h for __user
GOTO Masanori [Wed, 29 Jun 2005 03:45:04 +0000 (20:45 -0700)]
[PATCH] headers: include linux/compiler.h for __user

This patch lets i2c-dev.h include linux/compiler.h so that __user is defined.

Signed-off-by: GOTO Masanori <gotom@debian.or.jp>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] headers: enable ppc64 ___arch__swab16 and ___arch__swab32
GOTO Masanori [Wed, 29 Jun 2005 03:45:03 +0000 (20:45 -0700)]
[PATCH] headers: enable ppc64 ___arch__swab16 and ___arch__swab32

This patch cleans up asm-ppc64/byteorder.h to enable ___arch__swab16 and
___arch__swab32 which are marked TODO currently.  It removes ___arch__swab64
because ppc64 does not have short instruction combinations for swab64, the
recent gcc generates enough smart code that is equivalent to hand assembled
code under my tests.

Signed-off-by: GOTO Masanori <gotom@debian.or.jp>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] aio-retry-fix: fix aio retry work queueing
Sébastien Dugu [Wed, 29 Jun 2005 03:44:59 +0000 (20:44 -0700)]
[PATCH] aio-retry-fix: fix aio retry work queueing

In the case of buffered AIO, in the aio retry path (aio_run_iocb), when the
retry method returns EIOCBRETRY the kicked iocb is added to the context run
list but is never queued onto the work queue.  The request therefore is
never completed.

This patch fixes that by adding the appropriate call to aio_queue_work in
aio_run_aiocb so that subsequent retries will be handled by the aio worker
thread.

Signed-off-by: Sébastien Dugué <sebastien.dugue@bull.net>
Acked-by: Benjamin LaHaise <benjamin.c.lahaise@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] really remove xattr_acl.h
Christoph Hellwig [Wed, 29 Jun 2005 03:44:58 +0000 (20:44 -0700)]
[PATCH] really remove xattr_acl.h

Looks like it sneaked back with the NFS ACL merge..

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Cc: Neil Brown <neilb@cse.unsw.edu.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] drivers/isdn/: make some code static
Adrian Bunk [Wed, 29 Jun 2005 03:44:56 +0000 (20:44 -0700)]
[PATCH] drivers/isdn/: make some code static

This patch makes some needlessly global code static.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Armin Schindler <armin@melware.de>
Signed-off-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] rename wakeup_bdflush to wakeup_pdflush
Pekka J Enberg [Wed, 29 Jun 2005 03:44:55 +0000 (20:44 -0700)]
[PATCH] rename wakeup_bdflush to wakeup_pdflush

Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] swabb.h warning fixes
Andrew Morton [Wed, 29 Jun 2005 03:44:54 +0000 (20:44 -0700)]
[PATCH] swabb.h warning fixes

In file included from drivers/media/dvb/ttpci/av7110_hw.c:38:
include/linux/byteorder/swabb.h:96: warning: type qualifiers ignored on function return type
include/linux/byteorder/swabb.h:110: warning: type qualifiers ignored on function return type
In file included from drivers/media/dvb/ttpci/av7110_v4l.c:36:
include/linux/byteorder/swabb.h:96: warning: type qualifiers ignored on function return type
include/linux/byteorder/swabb.h:110: warning: type qualifiers ignored on function return type
In file included from drivers/media/dvb/ttpci/av7110_av.c:37:
include/linux/byteorder/swabb.h:96: warning: type qualifiers ignored on function return type
include/linux/byteorder/swabb.h:110: warning: type qualifiers ignored on function return type
drivers/isdn/icn/icn.c:719:4: warning: #warning TODO test headroom or use skb->nb to flag ACK
In file included from drivers/media/dvb/ttpci/av7110_ca.c:39:
include/linux/byteorder/swabb.h:96: warning: type qualifiers ignored on function return type
include/linux/byteorder/swabb.h:110: warning: type qualifiers ignored on function return type
In file included from drivers/media/dvb/ttpci/av7110.c:41:
include/linux/byteorder/swabb.h:96: warning: type qualifiers ignored on function return type
include/linux/byteorder/swabb.h:110: warning: type qualifiers ignored on function return type

Does declaring a function to return a const value actually mean something to
gcc?

Dunno.  Kill it and replace sone `__inline__'s with `inline' too.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] hisax warning fixes
Andrew Morton [Wed, 29 Jun 2005 03:44:53 +0000 (20:44 -0700)]
[PATCH] hisax warning fixes

drivers/isdn/hisax/hfc4s8s_l1.c:317: warning: type qualifiers ignored on function return type
drivers/isdn/hisax/hfc4s8s_l1.c:329: warning: type qualifiers ignored on function return type

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] request_firmware(): avoid race conditions
Stanislaw W. Gruszka [Wed, 29 Jun 2005 03:44:51 +0000 (20:44 -0700)]
[PATCH] request_firmware(): avoid race conditions

Avoid race occurs when some process have open file descriptor for class
device attributes and already firmware allocated memory are freed.  Don't
allow negative loading timeout.

Signed-off-by: Stanislaw W. Gruszka <stf_xl@wp.pl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Adapt drivers/char/vt_ioctl.c to non-x86
Emmanuel Colbus [Wed, 29 Jun 2005 03:44:49 +0000 (20:44 -0700)]
[PATCH] Adapt drivers/char/vt_ioctl.c to non-x86

This code uses the x86 (non-AMD-ELAN) value of CLOCK_TICK_RATE instead of
CLOCK_TICK_RATE itself, which is wrong for other archs.

Signed-off-by: Emmanuel Colbus <emmanuel.colbus@ensimag.imag.fr>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ITIMER_REAL: fix possible deadlock and race
Oleg Nesterov [Wed, 29 Jun 2005 03:44:47 +0000 (20:44 -0700)]
[PATCH] ITIMER_REAL: fix possible deadlock and race

As Steven Rostedt pointed out, there are 2 problems with ITIMER_REAL
timers.

1. do_setitimer() does not call del_timer_sync() in case
   when the timer is not pending (it_real_value() returns 0).
   This is wrong, the timer may still be running, and it can
   rearm itself.

2. It calls del_timer_sync() with tsk->sighand->siglock held.
   This is deadlockable, because timer's handler needs this
   lock too.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Acked-by: Steven Rostedt <rostedt@goodmis.org>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ixp4xx/ixp2000 watchdog driver typo
Lennert Buytenhek [Wed, 29 Jun 2005 03:44:46 +0000 (20:44 -0700)]
[PATCH] ixp4xx/ixp2000 watchdog driver typo

Fix the same typo in the ixp4xx and ixp2000 watchdog drivers.

Signed-off-by: Lennert Buytenhek <buytenh+lkml@wantstofly.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] drivers/char/tipar.c: off by one array access
Luc Van Oostenryck [Wed, 29 Jun 2005 03:44:44 +0000 (20:44 -0700)]
[PATCH] drivers/char/tipar.c: off by one array access

In the setup function, the delay variable is initialized with ints[2],
but ints is declared as:
int ints[2];

Since the module parameter should correspond to:
tipar=timeout,delay

I suppose that the following patch fix the problem.

Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@looxix.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Using msleep() instead of HZ
Luca Falavigna [Wed, 29 Jun 2005 03:44:42 +0000 (20:44 -0700)]
[PATCH] Using msleep() instead of HZ

Use msleep() in a few places.

Signed-off-by: Luca Falavigna <dktrkranz@gmail.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fix semaphore handling in __unregister_chrdev_region
Wen-chien Jesse Sung [Wed, 29 Jun 2005 03:44:41 +0000 (20:44 -0700)]
[PATCH] fix semaphore handling in __unregister_chrdev_region

This up() should be down() instead.

Signed-off-by: Wen-chien Jesse Sung <jesse@cola.voip.idv.tw>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[NET]: drivers/net/slip.c needs linux/delay.h
David S. Miller [Tue, 28 Jun 2005 23:27:32 +0000 (16:27 -0700)]
[NET]: drivers/net/slip.c needs linux/delay.h

For msleep_interruptible().

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETFILTER]: Fix connection tracking bug in 2.6.12
Patrick McHardy [Tue, 28 Jun 2005 23:04:44 +0000 (16:04 -0700)]
[NETFILTER]: Fix connection tracking bug in 2.6.12

In 2.6.12 we started dropping the conntrack reference when a packet
leaves the IP layer. This broke connection tracking on a bridge,
because bridge-netfilter defers calling some NF_IP_* hooks to the bridge
layer for locally generated packets going out a bridge, where the
conntrack reference is no longer available. This patch keeps the
reference in this case as a temporary solution, long term we will
remove the defered hook calling. No attempt is made to drop the
reference in the bridge-code when it is no longer needed, tc actions
could already have sent the packet anywhere.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: Add missing include to linux/netdevice.h
Arnd Bergmann [Tue, 28 Jun 2005 22:58:50 +0000 (15:58 -0700)]
[NET]: Add missing include to linux/netdevice.h

linux/etherdevice.h can't be included standalone at the moment, which
is required in order to sort the header files in the recommended
alphabetic order. This patch fixes that and is needed to build spider_net.

Signed-off-by: Arnd Bergmann <arndb@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: Micro optimization in eth_header()
Denis Vlasenko [Tue, 28 Jun 2005 22:49:06 +0000 (15:49 -0700)]
[NET]: Micro optimization in eth_header()

Signed-off-by: Denis Vlasenko <vda@ilport.com.ua>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPV6]: remove more unused IPV6_AUTHHDR things.
YOSHIFUJI Hideaki [Tue, 28 Jun 2005 22:46:24 +0000 (15:46 -0700)]
[IPV6]: remove more unused IPV6_AUTHHDR things.

Remove two more unused IPV6_AUTHHDR option things,
which I failed to remove them last time,
plus, mark IPV6_AUTHHDR obsolete.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPVS]: Close race conditions on ip_vs_conn_tab list modification
Neil Horman [Tue, 28 Jun 2005 22:40:02 +0000 (15:40 -0700)]
[IPVS]: Close race conditions on ip_vs_conn_tab list modification

In an smp system, it is possible for an connection timer to expire, calling
ip_vs_conn_expire while the connection table is being flushed, before
ct_write_lock_bh is acquired.

Since the list iterator loop in ip_vs_con_flush releases and re-acquires the
spinlock (even though it doesn't re-enable softirqs), it is possible for the
expiration function to modify the connection list, while it is being traversed
in ip_vs_conn_flush.

The result is that the next pointer gets set to NULL, and subsequently
dereferenced, resulting in an oops.

Signed-off-by: Neil Horman <nhorman@redhat.com>
Acked-by: JulianAnastasov
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: Remove gratuitous use of skb->tail in network drivers.
David S. Miller [Tue, 28 Jun 2005 22:25:31 +0000 (15:25 -0700)]
[NET]: Remove gratuitous use of skb->tail in network drivers.

Many drivers use skb->tail unnecessarily.

In these situations, the code roughly looks like:

dev = dev_alloc_skb(...);

[optional] skb_reserve(skb, ...);

... skb->tail ...

But even if the skb_reserve() happens, skb->data equals
skb->tail.  So it doesn't make any sense to use anything
other than skb->data in these cases.

Another case was the s2io.c driver directly mucking with
the skb->data and skb->tail pointers.  It really just wanted
to do an skb_reserve(), so that's what the code was changed
to do instead.

Another reason I'm making this change as it allows some SKB
cleanups I have planned simpler to merge.  In those cleanups,
skb->head, skb->tail, and skb->end pointers are removed, and
replaced with skb->head_room and skb->tail_room integers.

Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[IPV4]: Broken memory allocation in fib_trie
Robert Olsson [Tue, 28 Jun 2005 22:00:39 +0000 (15:00 -0700)]
[IPV4]: Broken memory allocation in fib_trie

This should help up the insertion... but the resize is more crucial.
and complex and needs some thinking.

Signed-off-by: Robert Olsson <robert.olsson@its.uu.se>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Tue, 28 Jun 2005 22:00:08 +0000 (15:00 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/pci-2.6
Linus Torvalds [Tue, 28 Jun 2005 21:59:07 +0000 (14:59 -0700)]
Merge /pub/scm/linux/kernel/git/gregkh/pci-2.6

18 years ago[PATCH] Tweak idle thread setup semantics
Ingo Molnar [Tue, 28 Jun 2005 14:40:42 +0000 (16:40 +0200)]
[PATCH] Tweak idle thread setup semantics

This patch tweaks idle thread setup semantics a bit: instead of setting
NEED_RESCHED in init_idle(), we do an explicit schedule() before calling
into cpu_idle().

This patch, while having no negative side-effects, enables wider use of
cond_resched()s.  (which might happen in the stock kernel too, but it's
particulary important for voluntary-preempt)

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ll_rw_blk: prevent huge request allocations
Jens Axboe [Tue, 28 Jun 2005 14:35:11 +0000 (16:35 +0200)]
[PATCH] ll_rw_blk: prevent huge request allocations

Currently we cap request allocations at q->nr_requests, but we allow a
batching io context to allocate up to 32 more (default setting).  This
can flood the queue with request allocations, with only a few batching
processes.  The real fix would be to limit the number of batchers, but
as that isn't currently tracked, I suggest we just cap the maximum
number of allocated requests to eg 50% over the limit.

This was observed in real life, users typically see this as vmstat bo
numbers going off the wall with seconds of no queueing afterwards.
Behaviour this bursty is not beneficial.

Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] md: bio leak fix
Neil Brown [Tue, 28 Jun 2005 05:29:34 +0000 (22:29 -0700)]
[PATCH] md: bio leak fix

insert a missing bio_put when writting the md superblock.

Without this we have a steady growth in the "bio" slab.

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] kexec: fix sparse warnings
Alexey Dobriyan [Tue, 28 Jun 2005 05:29:33 +0000 (22:29 -0700)]
[PATCH] kexec: fix sparse warnings

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] cciss_ioctl() warning fix
Andrew Morton [Tue, 28 Jun 2005 05:29:31 +0000 (22:29 -0700)]
[PATCH] cciss_ioctl() warning fix

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ARM: 2763/1: S3C24XX - mark IRQ_LCD as valid
Ben Dooks [Tue, 28 Jun 2005 21:42:06 +0000 (22:42 +0100)]
[PATCH] ARM: 2763/1: S3C24XX - mark IRQ_LCD as valid

Patch from Ben Dooks

Fix the IRQ_LCD so that it is marked as valid
since we no longer de-mux this in the main IRQ
handler.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[SCTP] Make init & delayed sack timeouts configurable by user.
Vlad Yasevich [Tue, 28 Jun 2005 20:24:23 +0000 (13:24 -0700)]
[SCTP] Make init & delayed sack timeouts configurable by user.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: Sridhar Samudrala <sri@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPV4]: ipconfig.c: fix dhcp timeout behaviour
Maxime Bizon [Tue, 28 Jun 2005 20:21:12 +0000 (13:21 -0700)]
[IPV4]: ipconfig.c: fix dhcp timeout behaviour

I think there is a small bug in ipconfig.c in case IPCONFIG_DHCP is set
and dhcp is used.

When a DHCPOFFER is received, ip address is kept until we get DHCPACK.
If no ack is received, ic_dynamic() returns negatively, but leaves the
offered ip address in ic_myaddr.

This makes the main loop in ip_auto_config() break and uses the maybe
incomplete configuration.

Not sure if it's the best way to do, but the following trivial patch
correct this.

Signed-off-by: Maxime Bizon <mbizon@freebox.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPV4]: Snmpv2 Mib IP counter ipInAddrErrors support
Dietmar Eggemann [Tue, 28 Jun 2005 20:06:23 +0000 (13:06 -0700)]
[IPV4]: Snmpv2 Mib IP counter ipInAddrErrors support

I followed Thomas' proposal to see every martian destination as a case
where the ipInAddrErrors counter has to be incremented. There are
two advantages by doing so: (1) The relation between the ipInReceive
counter and all the other ipInXXX counters is more accurate in the
case the RTN_UNICAST code check fails and (2) it makes the code in
ip_route_input_slow easier.

Signed-off-by: Dietmar Eggemann <dietmar.eggemann@gmx.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[PATCH] ARM: 2760/1: Add a warning on dynamic tick timekeeping on some platforms
Tony Lindgren [Tue, 28 Jun 2005 20:01:16 +0000 (21:01 +0100)]
[PATCH] ARM: 2760/1: Add a warning on dynamic tick timekeeping on some platforms

Patch from Tony Lindgren

This patch was suggested by RMK, and adds a warning on the accuracy
of timekeeping when using dynamic tick on some platforms. Depending
on the timer implementation, dynamic tick may affect the accuracy of
timekeeping.
Currently at least OMAP is known to have accurate timekeeping with
dynamic tick.

Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[PATCH] ARM: 2762/1: S3C24XX Audio platform data
Ben Dooks [Tue, 28 Jun 2005 20:01:03 +0000 (21:01 +0100)]
[PATCH] ARM: 2762/1: S3C24XX Audio platform data

Patch from Ben Dooks

This provides the s3c24xx audio platform data which can be
supplied from any of the board specific drivers.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[PATCH] ARM: 2761/1: OProfile: Add call graphing support for arm
Richard Purdie [Tue, 28 Jun 2005 20:01:03 +0000 (21:01 +0100)]
[PATCH] ARM: 2761/1: OProfile: Add call graphing support for arm

Patch from Richard Purdie

Add functions to generate backtraces of both kernel and user processes
which allows oprofile's call graphing functionality to be used on arm.
This requires unstripped binaries/libs which use a frame pointer.

Signed-off-by: Richard Purdie
Signed-off-by: Zwane Mwaikambo <zwane@arm.linux.org.uk>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[IPV6]: Don't dump temporary addresses twice
YOSHIFUJI Hideaki [Tue, 28 Jun 2005 20:00:30 +0000 (13:00 -0700)]
[IPV6]: Don't dump temporary addresses twice

Each IPv6 Temporary Address (w/ CONFIG_IPV6_PRIVACY) is dumped twice
to netlink.

Because temporary addresses are listed in idev->addr_list,
there's no need to dump idev->tempaddr separately.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETLINK]: Missing padding fields in dumped structures
Patrick McHardy [Tue, 28 Jun 2005 19:56:45 +0000 (12:56 -0700)]
[NETLINK]: Missing padding fields in dumped structures

Plug holes with padding fields and initialized them to zero.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETLINK]: Missing initializations in dumped data
Patrick McHardy [Tue, 28 Jun 2005 19:55:30 +0000 (12:55 -0700)]
[NETLINK]: Missing initializations in dumped data

Mostly missing initialization of padding fields of 1 or 2 bytes length,
two instances of uninitialized nlmsgerr->msg of 16 bytes length.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETLINK]: Clear padding in netlink messages
Patrick McHardy [Tue, 28 Jun 2005 19:54:43 +0000 (12:54 -0700)]
[NETLINK]: Clear padding in netlink messages

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETFILTER]: ipt_CLUSTERIP: fix ARP mangling
Harald Welte [Tue, 28 Jun 2005 19:49:30 +0000 (12:49 -0700)]
[NETFILTER]: ipt_CLUSTERIP: fix ARP mangling

This patch adds mangling of ARP requests (in addition to replies),
since ARP caches are made from snooping both requests and replies.

Signed-off-by: Harald Welte <laforge@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[EBTABLES]: Fix thinkos in ebt_log.c
David S. Miller [Tue, 28 Jun 2005 19:39:40 +0000 (12:39 -0700)]
[EBTABLES]: Fix thinkos in ebt_log.c

When converting over the skb_header_pointer(), I converted parts of
this module incorrectly.  Kill the 'u' union in ebt_log() and all the
bogus references to it.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years agoMerge head 'upstream-20050628-1' of rsync://rsync.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 28 Jun 2005 17:20:11 +0000 (10:20 -0700)]
Merge head 'upstream-20050628-1' of /linux/kernel/git/jgarzik/libata-dev

18 years ago[IA64] Fix another IA64 preemption problem
Peter Chubb [Fri, 24 Jun 2005 04:14:00 +0000 (21:14 -0700)]
[IA64] Fix another IA64 preemption problem

There's another problem shown up by Ingo's recent patch to make
smp_processor_id() complain if it's called with preemption enabled.
local_finish_flush_tlb_mm() calls activate_context() in a situation
where it could be rescheduled to another processor.  This patch
disables preemption around the call.

Signed-off-by: Peter Chubb <peterc@gelato.unsw.edu.au>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64] Replace stale KDB-code with useful MAGIC_SYSRQ code in simserial.c
David Mosberger-Tang [Fri, 10 Jun 2005 05:40:00 +0000 (22:40 -0700)]
[IA64] Replace stale KDB-code with useful MAGIC_SYSRQ code in simserial.c

Patch makes it possible to use the "F4" function key to do
magic sysrq in the HP Ski simulator.

Signed-off-by: David Mosberger-Tang <davidm@hpl.hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] Altix patch to tiocx, add subsys_initcall
Bruce Losure [Tue, 24 May 2005 15:30:00 +0000 (08:30 -0700)]
[IA64-SGI] Altix patch to tiocx, add subsys_initcall

This patch fixes an ordering issue between the init code for the
tiocx bus driver and tiocx-related device drivers.   Also adds
a new brick to the list of known FPGA bricks.

Signed-off-by: Bruce Losure <blosure@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64] sparse cleanup of TIOCA files
Prarit Bhargava [Tue, 10 May 2005 19:40:00 +0000 (12:40 -0700)]
[IA64] sparse cleanup of TIOCA files

This patch is a sparse compile cleanup of tioca_provider.c, sn_hwperf.h, and
tioca_provider.h.  Each of these files had sparse warnings when
compiled.

Signed-off-by: Prarit Bhargava <prarit@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] Fix TIO IOSPACE MMR Addres
Colin Ngam [Tue, 10 May 2005 16:12:00 +0000 (09:12 -0700)]
[IA64-SGI] Fix TIO IOSPACE MMR Addres

This patches provides support on Shub2 for the separate TIO IOSPACE MMR.  This
patch is SN specific.

Signed-off-by: Colin Ngam <cngam@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] - new macros for SGI SN simulator
Jack Steiner [Tue, 10 May 2005 15:01:00 +0000 (08:01 -0700)]
[IA64-SGI] - new macros for SGI SN simulator

This patch changes some macros that are used when running kernel on the
SGI simulator.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64] sparse cleanup of shub_mmr.h
Prarit Bhargava [Mon, 9 May 2005 12:07:00 +0000 (05:07 -0700)]
[IA64] sparse cleanup of shub_mmr.h

This patch is a sparse compile cleanup of shub_mmr.h using both the defconfig
and the sn2_defconfig config files.

The issue with this file was the missing usage of __IA64_UL_CONST wrapper.
This wrapper is defined in include/asm-ia64/types.h and wraps a long
constant definition with UL or with nothing depending on its usage in the
kernel.  The missing wrapper caused many sparse compile errors like

        warning: constant 0x0x0000000010000380 so big it is long

Signed-off-by: Prarit Bhargava <prarit@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64] Speed up lfetch.fault [NULL]
David Mosberger-Tang [Wed, 4 May 2005 20:25:00 +0000 (13:25 -0700)]
[IA64] Speed up lfetch.fault [NULL]

This patch greatly speeds up the handling of lfetch.fault instructions
which result in NaT consumption. Due to the NaT-page mapped at address
0, this is guaranteed to happen when lfetch.fault'ing a NULL pointer.
With this patch in place, we can even define prefetch()/prefetchw() as
lfetch.fault without significant performance degradation.  More
importantly, it allows compilers to be more aggressive with using
lfetch.fault on pointers that might be NULL.

Signed-off-by: David Mosberger-Tang <davidm@hpl.hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64] Recognize HP sx2000 chipset
Bjorn Helgaas [Tue, 3 May 2005 19:07:00 +0000 (12:07 -0700)]
[IA64] Recognize HP sx2000 chipset

No functional change, just identify the device nicely:

  -IOC: Unknown (103c:12ec) 0.1 HPA 0xf8020002000 IOVA space 1024Mb at 0x40000000
  +IOC: sx2000 0.1 HPA 0xf8020002000 IOVA space 1024Mb at 0x40000000

We used to create fake PCI devices for these chips, but we no longer do that.
So I don't think there's any reason to touch pci_ids.h now.

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] altix: enable vgacon support
Mark Maule [Tue, 26 Apr 2005 15:01:00 +0000 (08:01 -0700)]
[IA64-SGI] altix: enable vgacon support

Altix patch to enable use of vgacon driver on that platform.  Depends on the
PCDP generalization patch discussed at:

http://marc.theaimsgroup.com/?l=linux-ia64&m=111446235101939&w=2

Signed-off-by: Mark Maule <maule@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] pcdp: add PCDP pci interface support
Mark Maule [Mon, 25 Apr 2005 20:51:00 +0000 (13:51 -0700)]
[IA64-SGI] pcdp: add PCDP pci interface support

Resend 2 with changes per Bjorn Helgaas comments.  Changes from original:

+ Change globals to vga_console_iobase/vga_console_membase and make them
  unconditional.
+ Address style-related comments.

Patch to extend the PCDP vga setup code to support PCI io/mem translations
for the legacy vga ioport and ram spaces on architectures (e.g. altix) which
need them.

Summary of the changes:

drivers/firmware/pcdp.c
drivers/firmware/pcdp.h
-----------------------
+ add declaration for the spec-defined PCI interface struct (pcdp_if_pci)
  as well as support macros.

+ extend setup_vga_console() to know about pcdp_if_pci and add a couple of
  globals to hold the io and mem translation offsets if present.

arch/ia64/kernel/setup.c
------------------------
+ tweek early_console_setup() to allow multiple early console setup routines
  to be called.

include/asm-ia64/vga.h
----------------------
+ make VGA_MAP_MEM vga_console_membase aware

Signed-off-by: Mark Maule <maule@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years agoAuto merge with /home/aegl/GIT/ia64-test
Tony Luck [Tue, 28 Jun 2005 15:24:49 +0000 (08:24 -0700)]
Auto merge with /home/aegl/GIT/ia64-test

18 years agoMerge rsync://rsync.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6
Greg KH [Tue, 28 Jun 2005 05:07:56 +0000 (22:07 -0700)]
Merge /pub/scm/linux/kernel/git/torvalds/linux-2.6

18 years ago[PATCH] PCI: use the MCFG table to properly access pci devices (x86-64)
Greg Kroah-Hartman [Fri, 24 Jun 2005 00:35:56 +0000 (17:35 -0700)]
[PATCH] PCI: use the MCFG table to properly access pci devices (x86-64)

Now that we have access to the whole MCFG table, let's properly use it
for all pci device accesses (as that's what it is there for, some boxes
don't put all the busses into one entry.)

If, for some reason, the table is incorrect, we fallback to the "old
style" of mmconfig accesses, namely, we just assume the first entry in
the table is the one for us, and blindly use it.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] PCI: use the MCFG table to properly access pci devices (i386)
Greg Kroah-Hartman [Fri, 24 Jun 2005 00:35:56 +0000 (17:35 -0700)]
[PATCH] PCI: use the MCFG table to properly access pci devices (i386)

Now that we have access to the whole MCFG table, let's properly use it
for all pci device accesses (as that's what it is there for, some boxes
don't put all the busses into one entry.)

If, for some reason, the table is incorrect, we fallback to the "old
style" of mmconfig accesses, namely, we just assume the first entry in
the table is the one for us, and blindly use it.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] PCI: add proper MCFG table parsing to ACPI core.
Greg Kroah-Hartman [Fri, 24 Jun 2005 00:35:56 +0000 (17:35 -0700)]
[PATCH] PCI: add proper MCFG table parsing to ACPI core.

This patch is the first step in properly handling the MCFG PCI table.
It defines the structures properly, and saves off the table so that the
pci mmconfig code can access it.  It moves the parsing of the table a
little later in the boot process, but still before the information is
needed.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] PCI: make drivers use the pci shutdown callback instead of the driver core...
Greg Kroah-Hartman [Fri, 24 Jun 2005 00:35:56 +0000 (17:35 -0700)]
[PATCH] PCI: make drivers use the pci shutdown callback instead of the driver core callback.

Now we can change the pci core to always set this pointer, as pci drivers
should use it, not the driver core callback.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] cpqphp: fix oops during unload without probe
Keith Moore [Thu, 2 Jun 2005 10:42:37 +0000 (12:42 +0200)]
[PATCH] cpqphp: fix oops during unload without probe

drivers/pci/hotplug/cpqphp_core.c calls cpqphp_event_start_thread()
in one_time_init(), which is called whenever the hardware is probed.
Unfortunately, cpqphp_event_stop_thread() is *always* called when
the module is unloaded. If the hardware is never probed, then
cpqphp_event_stop_thread() tries to manipulate a couple of
uninitialized mutexes.

Signed-off-by: Keith Moore <keithmo@exmsft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] PCI: clean up the MSI code a bit.
Greg Kroah-Hartman [Tue, 7 Jun 2005 06:07:46 +0000 (23:07 -0700)]
[PATCH] PCI: clean up the MSI code a bit.

Mostly just cleans up the irq handling logic to be smaller and a bit more
descriptive as to what it really does.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] PCI: fix up errors after dma bursting patch and CONFIG_PCI=n
Andrew Morton [Tue, 7 Jun 2005 06:07:46 +0000 (23:07 -0700)]
[PATCH] PCI: fix up errors after dma bursting patch and CONFIG_PCI=n

With CONFIG_PCI=n:

In file included from include/linux/pci.h:917,
                 from lib/iomap.c:6:
include/asm/pci.h:104: warning: `enum pci_dma_burst_strategy' declared inside parameter list
include/asm/pci.h:104: warning: its scope is only this definition or declaration, which is probably not what you want.
include/asm/pci.h: In function `pci_dma_burst_advice':
include/asm/pci.h:106: dereferencing pointer to incomplete type
include/asm/pci.h:106: `PCI_DMA_BURST_INFINITY' undeclared (first use in this function)
include/asm/pci.h:106: (Each undeclared identifier is reported only once
include/asm/pci.h:106: for each function it appears in.)
make[1]: *** [lib/iomap.o] Error 1

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] PCI: DMA bursting advice
David S. Miller [Thu, 2 Jun 2005 19:55:50 +0000 (12:55 -0700)]
[PATCH] PCI: DMA bursting advice

After seeing, at best, "guesses" as to the following kind
of information in several drivers, I decided that we really
need a way for platforms to specifically give advice in this
area for what works best with their PCI controller implementation.

Basically, this new interface gives DMA bursting advice on
PCI.  There are three forms of the advice:

1) Burst as much as possible, it is not necessary to end bursts
   on some particular boundary for best performance.

2) Burst on some byte count multiple.  A DMA burst to some multiple of
   number of bytes may be done, but it is important to end the burst
   on an exact multiple for best performance.

   The best example of this I am aware of are the PPC64 PCI
   controllers, where if you end a burst mid-cacheline then
   chip has to refetch the data and the IOMMU translations
   which hurts performance a lot.

3) Burst on a single byte count multiple.  Bursts shall end
   exactly on the next multiple boundary for best performance.

   Sparc64 and Alpha's PCI controllers operate this way.  They
   disconnect any device which tries to burst across a cacheline
   boundary.

   Actually, newer sparc64 PCI controllers do not have this behavior.
   That is why the "pdev" is passed into the interface, so I can
   add code later to check which PCI controller the system is using
   and give advice accordingly.

Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] PCI: fix-pci-mmap-on-ppc-and-ppc64.patch
Michael Ellerman [Fri, 13 May 2005 07:44:10 +0000 (17:44 +1000)]
[PATCH] PCI: fix-pci-mmap-on-ppc-and-ppc64.patch

This is an updated version of Ben's fix-pci-mmap-on-ppc-and-ppc64.patch
which is in 2.6.12-rc4-mm1.

It fixes the patch to work on PPC iSeries, removes some debug printks
at Ben's request, and incorporates your
fix-pci-mmap-on-ppc-and-ppc64-fix.patch also.

Originally from Benjamin Herrenschmidt <benh@kernel.crashing.org>

This patch was discussed at length on linux-pci and so far, the last
iteration of it didn't raise any comment.  It's effect is a nop on
architecture that don't define the new pci_resource_to_user() callback
anyway.  It allows architecture like ppc who put weird things inside of
PCI resource structures to convert to some different value for user
visible ones.  It also fixes mmap'ing of IO space on those archs.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] ACPI based I/O APIC hot-plug: acpiphp support
Kenji Kaneshige [Thu, 28 Apr 2005 07:25:59 +0000 (00:25 -0700)]
[PATCH] ACPI based I/O APIC hot-plug: acpiphp support

This patch adds PCI based I/O xAPIC hot-add support to ACPIPHP
driver. When PCI root bridge is hot-added, all PCI based I/O xAPICs
under the root bridge are hot-added by this patch. Hot-remove support
is TBD.

Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] ACPI based I/O APIC hot-plug: ia64 support
Kenji Kaneshige [Thu, 28 Apr 2005 07:25:58 +0000 (00:25 -0700)]
[PATCH] ACPI based I/O APIC hot-plug: ia64 support

This is an ia64 implementation of acpi_register_ioapic() and
acpi_unregister_ioapic() interfaces.

Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] ACPI based I/O APIC hot-plug: add interfaces
Kenji Kaneshige [Thu, 28 Apr 2005 07:25:58 +0000 (00:25 -0700)]
[PATCH] ACPI based I/O APIC hot-plug: add interfaces

This patch adds the following new interfaces for I/O xAPIC
hotplug. The implementation of these interfaces depends on each
architecture.

    o int acpi_register_ioapic(acpi_handle handle, u64 phys_addr,
       u32 gsi_base);

        This new interface is to add a new I/O xAPIC specified by
        phys_addr and gsi_base pair. phys_addr is the physical address
        to which the I/O xAPIC is mapped and gsi_base is global system
        interrupt base of the I/O xAPIC. acpi_register_ioapic returns
        0 on success, or negative value on error.

    o int acpi_unregister_ioapic(acpi_handle handle, u32 gsi_base);

        This new interface is to remove a I/O xAPIC specified by
        gsi_base. acpi_unregister_ioapic returns 0 on success, or
        negative value on error.

Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] acpi hotplug: decouple slot power state changes from physical hotplug
Rajesh Shah [Thu, 28 Apr 2005 07:25:57 +0000 (00:25 -0700)]
[PATCH] acpi hotplug: decouple slot power state changes from physical hotplug

Current acpiphp code does not distinguish between the physical presence and
power state of a device/slot.  That is, if a device has to be disabled, it
also tries to physically ejects the device.  This patch decouples power state
from physical presence.  You can now echo to the corresponding sysfs power
control file to repeatedly enable and disable a device without having to
physically re-insert it.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] acpi hotplug: aCPI based root bridge hot-add
Rajesh Shah [Thu, 28 Apr 2005 07:25:56 +0000 (00:25 -0700)]
[PATCH] acpi hotplug: aCPI based root bridge hot-add

acpiphp changes to support acpi based root bridge hot-add.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] acpi hotplug: fix slot power-down problem with acpiphp
Rajesh Shah [Thu, 28 Apr 2005 07:25:55 +0000 (00:25 -0700)]
[PATCH] acpi hotplug: fix slot power-down problem with acpiphp

Earlier I reported that Matthew's acpiphp rewrite had problem in powering down
slot on my i386 system.  The following patch is needed to get the acpiphp
rewrite properly powering down the slot.

Signed-off-by: Dely Sy <dely.l.sy@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] acpi hotplug: clean up notify handlers on acpiphp unload
Rajesh Shah [Thu, 28 Apr 2005 07:25:54 +0000 (00:25 -0700)]
[PATCH] acpi hotplug: clean up notify handlers on acpiphp unload

A root bridge may not have directly attached hotpluggable slots under it.
Instead, it may have p2p bridges with slots under it.  In this case, we need
to clean up the p2p bridges and slots properly too.  Patch below applies on
top of the original patch, and fixes this problem.  Without this, acpiphp
leaves behind notify handlers on module unload, and subsequent module load
attempts don't work properly too.  Patch was tested on an ia64 Tiger4 box.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>