* (trivial) endianness annotations
* don't bother with del_timer() from the inside of timer handler itself
* disable_ast() really ought to do del_timer_sync(), not del_timer()
* clean the timer handling in general.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

* descriptors inside the rx and tx rings are l-e
* don't cpu_to_le32() the argument of outl()
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

* 'for-linus' of git://oss.sgi.com:8090/xfs/xfs-2.6:
  [XFS] fix unaligned access in readdir

* git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6.24:
  sh: Force __access_ok() to obey address space limit.
  sh: Fix argument page dcache flushing regression.

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] Replace 40c7869b kludge
  [MIPS] Lasat: Fix built in separate object directory.
  [MIPS] Malta: Fix software reset on big endian
  [MIPS] pnx8xxx: move to clocksource
  [MIPS] Wrong CONFIG option prevents setup of DMA zone.

Since the introduction of the acquire_console_sem calls in
0333d835, kexecing can cause the
kernel to deadlock:

 ps3fb_shutdown()
  -> unregister_framebuffer()
  -> fb_notifier_call_chain(FB_EVENT_FB_UNBIND)
  -> fbcon_fb_unbind()
  -> unbind_con_driver()
  -> bind_con_driver()
	[ acquires console_sem ]
  -> fbcon_deinit()
  -> fbops->fb_release(newinfo, 0)
  -> ps3fb_release()
  -> ps3fb_sync()
	[ acquires console_sem ]

This change avoids the deadlock by moving the acquire_console_sem()
out of ps3fb_sync(), and puts it into the two other callsites, leaving
ps3fb_release() to call ps3fb_sync() without the console semaphore.

[Geert]
  - Corrected call sequence above
  - ps3fb_release() may be called with and without console_sem held. This is an
    inconsistency that should be fixed at the fb level, but for now, try to
    acquire console_sem in ps3fb_release().

    I think it's safer to let ps3fb_release() try to acquire console_sem and
    not refresh the screen if it fails, than to call ps3fb_sync() without
    holding console_sem, as ps3fb_par may be modified at the same time, causing
    crashes or lockups.

    Besides, ps3fb_release() only calls ps3fb_sync() to refresh the screen
    when display flipping is disabled, which is an uncommon case (except during
    shutdown/kexec).
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

In ps3fb_shutdown, freeing the framebuffer will cause fb_info (in
dev->core.driver_data) to be free()ed, which we potentially access
from the ps3fbd kthread.

This change frees the framebuffer after stopping the ps3fbd kthread.
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  blktrace: kill the unneeded initcall
  block: fix blktrace timestamps
  loop: fix bad bio_alloc() nr_iovec request
  Don't blatt first element of prv in sg_chain()

The bug causes corruptions of data read from flash.

The original code performs cache invalidation from "adr" to "adr + len"
in do_write_buffer().  Since len and adr could be updated in the code
before invalidation - it causes improper setting of cache invalidation
regions.
Signed-off-by: Massimo Cirillo <maxcir@gmail.com>
Signed-off-by: Giuseppe D'Eliseo <giuseppedeliseo@gmail.com>
Acked-by: Nicolas Pitre <nico@cam.org>
Acked-by: Jörn Engel <joern@logfs.org>
Signed-off-by: David Woohouse <dwmw2@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Fix oops reported by Trond.
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

Signed-off-by: WANG Cong <xiyou.wangcong@gmail.com>

[Ralf: The LDSCRIPT script needed fixing, too]
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

I noticed that the commit f1974653
(MIPS Tech: Get rid of volatile in core code) broke the software
reset functionality for MIPS Malta boards in big-endian mode.

According to the MIPS Malta board user's manual, writing the magic
32-bit GORESET value into the SOFTRES register initiates board soft
reset. My experimentation has shown that the endianness of the GORESET
integer should thereby be the same as the endianness, which has been
set for the CPU itself. The writew() function used to write the magic
value in the code introduced by the commit mentioned above, however,
swaps bytes for big-endian kernels and transfers 16 bits instead of 32.

The patch below replaces the writew() function by the __raw_writel()
routine, which leaves the byte order intact and transfers the whole
MIPS machine word. Trivial code cleanup (replacing spaces by a tab
and cutting oversized lines to make checkpatch.pl happy) is also
included.

The patch was tested using a Malta evaluation board running in both
BE and LE modes. For both modes, software reset was fully functional
after the change.

P.S. I suspect that the same commit broke the "standby" functionality
for MIPS Atlas boards. However, I did not touch the Atlas code as I
don't have such board at my disposal and also because the linux-mips.org
Web site claims that Atlas support is scheduled for removal.
Signed-off-by: Dmitri Vorobiev <dmitri.vorobiev@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

This patch converts PNX8XXX system timer to clocksource restoring PNX8550
support back to live.
Signed-off-by: Vitaly Wool <vitalywool@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

It just inits the mutex, we can do that with DEFINE_MUTEX() instead.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

David Dillow reported broken blktrace timestamps. The reason
is cpu_clock() which is not a global time source.

Fix bkltrace timestamps by using ktime_get() like the networking
code does for packet timestamps. This also removes a whole lot
of complexity from bkltrace.c and shrinks the code by 500 bytes:

   text    data     bss     dec     hex filename
   2888     124      44    3056     bf0 blktrace.o.before
   2390     116      44    2550     9f6 blktrace.o.after
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

Don't allocate room for an iovec when it is not needed.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

I realize that sg chaining is a ploy to make the rest of the kernel
devs feel the pain of the SCSI subsystem.  But this was a little
unsubtle.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Acked-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

This patch should fix the issue seen on Alpha with unaligned accesses in
the new readdir code. By aligning each dirent to sizeof(u64) we'll avoid
unaligned accesses. To make doubly sure we're not hitting problems also
rearrange struct hack_dirent to avoid holes.

SGI-PV: 975411
SGI-Modid: xfs-linux-melb:xfs-kern:30302a
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>

When the thread_info->addr_limit changes were introduced, __access_ok()
was missed in the conversion, allowing user processes to perform P1/P2
accesses under certain conditions.

This has already been corrected with the nommu refactoring in later
kernels.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  [CRYPTO] padlock: Fix alignment fault in aes_crypt_copy

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  libata and starting/stopping ATAPI floppy devices
  sata_sil24: prevent hba lockup when pass-through ATA commands are used
  Update kernel parameter document for libata DMA mode setting knobs.
  libata: don't normalize UNKNOWN to NONE after reset
  libata-pmp: propagate timeout to host link
  libata-pmp: 4726 hates SRST
  pata_ixp4xx_cf: fix compilation introduced by ata_port_desc() conversion
  pata_pdc202xx_old: Further fixups
  libata-sff: PCI IRQ handling fix
  sata_qstor: use hardreset instead of softreset

* git://git.kernel.org/pub/scm/linux/kernel/git/bart/ide-2.6:
  trm290: do hook dma_host_{on,off} methods (take 2)
  ide: fix cable detection for SATA bridges
  ide: workaround suspend bug for ACPI IDE

Prevent libata from starting/stopping non-ATA devices (like ATAPI floppy
drives) as they don't seem to like it:

sd 1:0:1:0: [sdb] Starting disk
ata2.01: configured for PIO2
sd 1:0:1:0: [sdb] Result: hostbyte=0x00 driverbyte=0x08
sd 1:0:1:0: [sdb] Sense Key : 0x2 [current]
sd 1:0:1:0: [sdb] ASC=0x3a ASCQ=0x0
Signed-off-by: Ondrej Zary <linux@rainbow-software.org>
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Using default methods caused the chip's DMA PRD count registers, inadvertently
starting DMA!  While fixing it, also do:

- get rid of the 'ide_' prefixes in several functions for which the prefix in
  the method's name has been 'ide_' ectomized already;

- align the code hooking the IDE DMA methods in init_hwif_trm290()...
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Signed-off-by: George Kibardin <george-kibardin@yandex.ru>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

http://bugzilla.kernel.org/show_bug.cgi?id=9673

ACPI _PS3 cause S4 breaks in the second attempt. The system has a _PS3
method for IDE, which will call into SMM mode. Currently we haven't clue
why just the second attempt fails, as it's totally in BIOS code, so
blacklist the system so far for 2.6.24.

A possible suspect is ACPI NVS isn't save/restore, we will revisit the
bug after linux does ACPI NVS save/restore.

Bart:
- fix scripts/checkpatch.pl complaints
- const-ify ide_acpi_dmi_table[]
Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Reported-by: Mikko Vinni <mmvinni@yahoo.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Fix commands timeout with Sil3124/3132 based HBA when pass-through ATA
commands [where ATA_QCFLAG_RESULT_TF is set] are used while other
commands are active on other devices connected to the same port with a
Port Multiplier.  Due to a hardware bug, these commands must be sent
alone, like ATAPI commands.
Signed-off-by: Gwendal Grignou <gwendal@google.com>
Acked-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

Signed-off-by: Jeff Garzik <jeff@garzik.org>

After non-classifying reset, ehc->classes[] could contain
ATA_DEV_UNKNOWN which used to be normalized to ATA_DEV_NONE for
consistency.  However, this causes unfortunate side effect for drivers
which have non-classifying hardresets (e.g. sata_nv) by making
hardreset report ATA_DEV_NONE for non-classifying resets and thus
makes EH believe that the port is unoccupied and recovery can be
skipped.  The end result is that after a device is swapped with
another one, the new device isn't attached after the old one is
detached.

This patch makes ata_eh_reset() not normalize UNKNOWN to NONE after
non-classifying resets.  This fixes the above problem.  As UNKNOWN and
NONE are handled differently by only EH hotplug logic, this doesn't
cause other behavior changes.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Cc: Robert Hancock <hancockr@shaw.ca>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

Timeout on downstream command may indicate transmission problem on
host link.  Propagate timeouts to host link.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

4726 hates SRST even on non-config ports.  Don't use it.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

Fixes a compilation error caused by ata_port_desc() conversion
(cbcdd875).
Signed-off-by: Rod Whitby <rod@whitby.id.au>
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

Turns out distros always enabled burst mode and it is pretty essential so
do the same. Also sort out the post DMA mode restore properly.

My 20263 card now seems happy but needs some four drive tests done yet
(when I've persuaded the kernel not to hang in the edd boot code if I
plug them in ..)
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

It is legitimate (although annoying and silly) for a PCI IDE controller
not to be assigned an interrupt and to be polled. The libata-sff code
should therefore not try and request IRQ 0 in this case.
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

During conversion to new EH, sata_qstor was accidentaly changed to use
softreset, which is buggy on this chip, instead of hardreset.  This
patch updates sata_qstor such that it uses hardreset again.

This fixes bugzilla bug 9631.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Cc: Mark Lord <mlord@pobox.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

Sharing the open sequence queue causes a deadlock when we try to take
both a lock sequence id and and open sequence id.

This fixes the regression reported by Dimitri Puzin and Jeff Garzik: See

	http://bugzilla.kernel.org/show_bug.cgi?id=9712

for details.
Reported-and-tested-by: Dimitri Puzin <bugs@psycast.de>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Tested-by: Jeff Garzik <jgarzik@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

The previous patch fixed spurious read faults from occuring by copying
the data if we happen to have a single block at the end of a page.  It
appears that gcc cannot guarantee 16-byte alignment in the kernel with
__attribute__.  The following report from Torben Viets shows a buffer
that's only 8-byte aligned:

> eneral protection fault: 0000 [#1]
> Modules linked in: xt_TCPMSS xt_tcpmss iptable_mangle ipt_MASQUERADE
> xt_tcpudp xt_mark xt_state iptable_nat nf_nat nf_conntrack_ipv4
> iptable_filter ip_tables x_tables pppoe pppox af_packet ppp_generic slhc
> aes_i586
> CPU:    0
> EIP:    0060:[<c035b828>]    Not tainted VLI
> EFLAGS: 00010292   (2.6.23.12 #7)
> EIP is at aes_crypt_copy+0x28/0x40
> eax: f7639ff0   ebx: f6c24050   ecx: 00000001   edx: f6c24030
> esi: f7e89dc8   edi: f7639ff0   ebp: 00010000   esp: f7e89dc8

Since the hardware must have 16-byte alignment, the following patch fixes
this by open coding the alignment adjustment.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

There's an error remaining in the 32 bit descriptor code after the
conversion to dma accessors:  req_cnt is left uninitialised.

qla1280_32bit_start_scsi gives the following warnings:

drivers/scsi/qla1280.c: In function 'qla1280_32bit_start_scsi':
drivers/scsi/qla1280.c:3044: warning: unused variable 'dma_handle'
drivers/scsi/qla1280.c: In function 'qla1280_queuecommand':
drivers/scsi/qla1280.c:3060: warning: 'req_cnt' is used uninitialized in this function
drivers/scsi/qla1280.c:3042: note: 'req_cnt' was declared here
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>