linux-2.6
18 years ago[PATCH] libata: PIO 0
Alan Cox [Mon, 22 May 2006 15:52:06 +0000 (16:52 +0100)] 
[PATCH] libata: PIO 0

Ensure the pio_mode is always setup. Don't do any setup on the controller b
just ensure the mode reporting is valid to avoid tons of special cases
in PATA driver code when mode switching on the fly.

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: minor fix for irq-pio merge
Albert Lee [Tue, 23 May 2006 10:12:30 +0000 (18:12 +0800)] 
[PATCH] libata: minor fix for irq-pio merge

Minor fix to put the ATA_FLAG_NO_ATAPI flag back.

Signed-off-by: Albert Lee <albertcc@tw.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years agoMerge branch 'upstream-fixes' into upstream
Jeff Garzik [Wed, 24 May 2006 05:49:12 +0000 (01:49 -0400)] 
Merge branch 'upstream-fixes' into upstream

Conflicts:

drivers/scsi/libata-core.c

18 years ago[PATCH] libata: add pio flush for via atapi (was: Re: TR: ASUS A8V Deluxe, x86_64)
Albert Lee [Mon, 22 May 2006 03:43:46 +0000 (11:43 +0800)] 
[PATCH] libata: add pio flush for via atapi (was: Re: TR: ASUS A8V Deluxe, x86_64)

Backport the "pio flush" from the libata major update to 2.6.17 for via atapi.

Signed-off-by: Albert Lee <albertcc@tw.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[libata sata_promise] Add PATA cable detection.
Jeff Garzik [Wed, 24 May 2006 05:43:25 +0000 (01:43 -0400)] 
[libata sata_promise] Add PATA cable detection.

Original patch from Phillip Jordan <phillip.m.jordan@gmail.com>
Cleanups and fixes by me.

18 years ago[PATCH] sata_nv: Add MCP61 support
Andrew Chew [Thu, 20 Apr 2006 22:54:26 +0000 (15:54 -0700)] 
[PATCH] sata_nv: Add MCP61 support

Added MCP61 SATA support to sata_nv.

Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years agoMerge branch 'master' into upstream
Jeff Garzik [Wed, 24 May 2006 05:32:42 +0000 (01:32 -0400)] 
Merge branch 'master' into upstream

18 years ago[BRIDGE]: need to ref count the LLC sap
Stephen Hemminger [Tue, 23 May 2006 22:20:25 +0000 (15:20 -0700)] 
[BRIDGE]: need to ref count the LLC sap

Bridge will OOPS on removal if other application has the SAP open.
The bridge SAP might be shared with other usages, so need
to do reference counting on module removal rather than explicit
close/delete.

Since packet might arrive after or during removal, need to clear
the receive function handle, so LLC only hands it to user (if any).

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETFILTER]: SNMP NAT: fix memleak in snmp_object_decode
Chris Wright [Tue, 23 May 2006 22:08:13 +0000 (15:08 -0700)] 
[NETFILTER]: SNMP NAT: fix memleak in snmp_object_decode

If kmalloc fails, error path leaks data allocated from asn1_oid_decode().

Signed-off-by: Chris Wright <chrisw@sous-sol.org>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETFILTER]: H.323 helper: fix sequence extension parsing
Patrick McHardy [Tue, 23 May 2006 22:07:46 +0000 (15:07 -0700)] 
[NETFILTER]: H.323 helper: fix sequence extension parsing

When parsing unknown sequence extensions the "son"-pointer points behind
the last known extension for this type, don't try to interpret it.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETFILTER]: H.323 helper: fix parser error propagation
Patrick McHardy [Tue, 23 May 2006 22:07:07 +0000 (15:07 -0700)] 
[NETFILTER]: H.323 helper: fix parser error propagation

The condition "> H323_ERROR_STOP" can never be true since H323_ERROR_STOP
is positive and is the highest possible return code, while real errors are
negative, fix the checks. Also only abort on real errors in some spots
that were just interpreting any return value != 0 as error.

Fixes crashes caused by use of stale data after a parsing error occured:

BUG: unable to handle kernel paging request at virtual address bfffffff
 printing eip:
c01aa0f8
*pde = 1a801067
*pte = 00000000
Oops: 0000 [#1]
PREEMPT
Modules linked in: ip_nat_h323 ip_conntrack_h323 nfsd exportfs sch_sfq sch_red cls_fw sch_hfsc  xt_length ipt_owner xt_MARK iptable_mangle nfs lockd sunrpc pppoe pppoxx
CPU:    0
EIP:    0060:[<c01aa0f8>]    Not tainted VLI
EFLAGS: 00210646   (2.6.17-rc4 #8)
EIP is at memmove+0x19/0x22
eax: d77264e9   ebx: d77264e9   ecx: e88d9b17   edx: d77264e9
esi: bfffffff   edi: bfffffff   ebp: de6a7680   esp: c0349db8
ds: 007b   es: 007b   ss: 0068
Process asterisk (pid: 3765, threadinfo=c0349000 task=da068540)
Stack: <0>00000006 c0349e5e d77264e3 e09a2b4e e09a38a0 d7726052 d7726124 00000491
       00000006 00000006 00000006 00000491 de6a7680 d772601e d7726032 c0349f74
       e09a2dc2 00000006 c0349e5e 00000006 00000000 d76dda28 00000491 c0349f74
Call Trace:
 [<e09a2b4e>] mangle_contents+0x62/0xfe [ip_nat]
 [<e09a2dc2>] ip_nat_mangle_tcp_packet+0xa1/0x191 [ip_nat]
 [<e0a2712d>] set_addr+0x74/0x14c [ip_nat_h323]
 [<e0ad531e>] process_setup+0x11b/0x29e [ip_conntrack_h323]
 [<e0ad534f>] process_setup+0x14c/0x29e [ip_conntrack_h323]
 [<e0ad57bd>] process_q931+0x3c/0x142 [ip_conntrack_h323]
 [<e0ad5dff>] q931_help+0xe0/0x144 [ip_conntrack_h323]
...

Found by the PROTOS c07-h2250v4 testsuite.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years agoV4L/DVB (4041): Fix compilation on PPC 64
Mauro Carvalho Chehab [Tue, 23 May 2006 19:02:03 +0000 (16:02 -0300)] 
V4L/DVB (4041): Fix compilation on PPC 64

Those functions don't exist on PPC64 architecture.

Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (4040a): Fix the following section warnings:
Jean Delvare [Tue, 23 May 2006 18:56:50 +0000 (15:56 -0300)] 
V4L/DVB (4040a): Fix the following section warnings:

reference to .init.text: from .text between 'dvb_bt8xx_probe'
(at offset 0x122c) and 'dvb_bt8xx_remove'
reference to .init.text: from .text between 'dvb_bt8xx_probe'
(at offset 0x1267) and 'dvb_bt8xx_remove'

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (4037): Make the bridge devices that depend on I2C dependant on I2C
Manu Abraham [Sat, 20 May 2006 16:17:00 +0000 (13:17 -0300)] 
V4L/DVB (4037): Make the bridge devices that depend on I2C dependant on I2C

Ref: Bugzilla 6179, 6589

Signed-off-by: Manu Abraham <manu@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (3927): Fix VIDEO_DEV=m, VIDEO_V4L1_COMPAT=y
Adrian Bunk [Tue, 23 May 2006 18:56:20 +0000 (15:56 -0300)] 
V4L/DVB (3927): Fix VIDEO_DEV=m, VIDEO_V4L1_COMPAT=y

If CONFIG_VIDEO_DEV=m and CONFIG_VIDEO_V4L1_COMPAT=y, v4l1-compat should
be built as a module (currently, it isn't built at all leading to
problems with modules using it).

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Tue, 23 May 2006 17:40:43 +0000 (10:40 -0700)] 
Merge /pub/scm/linux/kernel/git/davem/sparc-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SPARC64]: Respect gfp_t argument to dma_alloc_coherent().

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Tue, 23 May 2006 17:40:19 +0000 (10:40 -0700)] 
Merge /pub/scm/linux/kernel/git/davem/net-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [NETFILTER]: SNMP NAT: fix memory corruption
  [IRDA]: fixup type of ->lsap_state
  [IRDA]: fix 16/32 bit confusion
  [NET]: Fix "ntohl(ntohs" bugs
  [BNX2]: Use kmalloc instead of array
  [BNX2]: Fix bug in bnx2_nvram_write()
  [TG3]: Add some missing rx error counters

18 years ago[PATCH] blk: fix gendisk->in_flight accounting during barrier sequence
Jens Axboe [Tue, 23 May 2006 09:23:49 +0000 (11:23 +0200)] 
[PATCH] blk: fix gendisk->in_flight accounting during barrier sequence

While executing barrrier sequence, the bar_rq which carries actual
write was accounted as normal IO on completion, while it wasn't on
queueing.  This caused gendisk->in_flight to be decremented by 1 after
each barrier thus messed up statistics.

This patch makes bar_rq not accounted as normal IO.  As the containing
barrier request as a whole is accounted, part of it shouldn't be.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] powerpc: wire up sys_[gs]et_robust_list
David Woodhouse [Tue, 23 May 2006 14:46:40 +0000 (07:46 -0700)] 
[PATCH] powerpc: wire up sys_[gs]et_robust_list

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Paul Mackerras <paulus@samba.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] powerpc: check Cell SPU syscall number range _before_ using it
David Woodhouse [Tue, 23 May 2006 14:46:39 +0000 (07:46 -0700)] 
[PATCH] powerpc: check Cell SPU syscall number range _before_ using it

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Paul Mackerras <paulus@samba.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] powerpc: fill hole in Cell SPU syscall table
David Woodhouse [Tue, 23 May 2006 14:46:38 +0000 (07:46 -0700)] 
[PATCH] powerpc: fill hole in Cell SPU syscall table

Syscall number 224 was absent from the table, which I believe means that
the SPU can cause an oops by attempting to use it.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Paul Mackerras <paulus@samba.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] orinoco: possible null pointer dereference in orinoco_rx_monitor()
Florin Malita [Tue, 23 May 2006 05:35:30 +0000 (22:35 -0700)] 
[PATCH] orinoco: possible null pointer dereference in orinoco_rx_monitor()

If the skb allocation fails, the current error path calls
dev_kfree_skb_irq() with a NULL argument.  Also, 'err' is not being used.

Coverity CID: 275.

Signed-off-by: Florin Malita <fmalita@gmail.com>
Cc: "John W. Linville" <linville@tuxdriver.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] swsusp: fix typo in cr0 handling
Pavel Machek [Tue, 23 May 2006 05:35:29 +0000 (22:35 -0700)] 
[PATCH] swsusp: fix typo in cr0 handling

Writing cr0 to cr2 register can't be right.  This fixes the typo.  I wonder
how it could survive so long.

Signed-off-by: Pavel Machek <pavel@suse.cz>
Cc: Zachary Amsden <zach@vmware.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] tty_insert_flip_string_flags() license fix
Tobias Powalowski [Tue, 23 May 2006 05:35:28 +0000 (22:35 -0700)] 
[PATCH] tty_insert_flip_string_flags() license fix

We still don't have the tty layer licensing compatibility quite right.

tty_insert_flip_char() used to be inlined in include/linux/tty_flip.h.  It
is now out-of-lined and hence needs EXPORT_SYMBOL() to be back-compatible.

One known offender is the Intel Modem driver.

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] md: Make sure bi_max_vecs is set properly in bio_split
NeilBrown [Tue, 23 May 2006 05:35:27 +0000 (22:35 -0700)] 
[PATCH] md: Make sure bi_max_vecs is set properly in bio_split

Else a subsequent bio_clone might make a mess.

Signed-off-by: Neil Brown <neilb@suse.de>
Cc: "Don Dupuis" <dondster@gmail.com>
Acked-by: 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] md: fix possible oops when starting a raid0 array
NeilBrown [Tue, 23 May 2006 05:35:26 +0000 (22:35 -0700)] 
[PATCH] md: fix possible oops when starting a raid0 array

This loop that sets up the hash_table has problems.

Careful examination will show that the last time through, everything but
the first line is pointless.  This is because all it does is change 'cur'
and 'size' and neither of these are used after the loop.  This should ring
warning bells...  That last time through the loop,

        size += conf->strip_zone[cur].size

can index off the end of the strip_zone array.  Depending on what it finds
there, it might exit the loop cleanly, or it might spin going further and
further beyond the array until it hits an unmapped address.

This patch rearranges the code so that the last, pointless, iteration of
the loop never happens.  i.e.  the one statement of the last loop that is
needed is moved the the end of the previous loop - or to before the loop
starts - and the loop counter starts from 1 instead of 0.

Cc: "Don Dupuis" <dondster@gmail.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Fix two problems that can cause rmmod nfsd to die
NeilBrown [Tue, 23 May 2006 05:35:25 +0000 (22:35 -0700)] 
[PATCH] knfsd: Fix two problems that can cause rmmod nfsd to die

Both cause the 'entries' count in the export cache to be non-zero at module
removal time, so unregistering that cache fails and results in an oops.

1/ exp_pseudoroot (used for NFSv4 only) leaks a reference to an export
   entry.
2/ sunrpc_cache_update doesn't increment the entries count when it adds
   an entry.

Thanks to "david m.  richter" <richterd@citi.umich.edu> for triggering the
problem and finding one of the bugs.

Cc: "david m. richter" <richterd@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] sys_sync_file_range(): move exported flags outside __KERNEL__
Andrew Morton [Tue, 23 May 2006 05:35:24 +0000 (22:35 -0700)] 
[PATCH] sys_sync_file_range(): move exported flags outside __KERNEL__

These flags are needed by userspace - move them outside __KERNEL__

(Pointed out by dwmw2)

Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[SPARC64]: Respect gfp_t argument to dma_alloc_coherent().
David S. Miller [Tue, 23 May 2006 09:07:22 +0000 (02:07 -0700)] 
[SPARC64]: Respect gfp_t argument to dma_alloc_coherent().

Using asm-generic/dma-mapping.h does not work because pushing
the call down to pci_alloc_coherent() causes the gfp_t argument
of dma_alloc_coherent() to be ignored.

Fix this by implementing things directly, and adding a gfp_t
argument we can use in the internal call down to the PCI DMA
implementation of pci_alloc_coherent().

This fixes massive memory corruption when using the sound driver
layer, which passes things like __GFP_COMP down into these
routines and (correctly) expects that to work.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETFILTER]: SNMP NAT: fix memory corruption
Patrick McHardy [Mon, 22 May 2006 23:55:14 +0000 (16:55 -0700)] 
[NETFILTER]: SNMP NAT: fix memory corruption

Fix memory corruption caused by snmp_trap_decode:

- When snmp_trap_decode fails before the id and address are allocated,
  the pointers contain random memory, but are freed by the caller
  (snmp_parse_mangle).

- When snmp_trap_decode fails after allocating just the ID, it tries
  to free both address and ID, but the address pointer still contains
  random memory. The caller frees both ID and random memory again.

- When snmp_trap_decode fails after allocating both, it frees both,
  and the callers frees both again.

The corruption can be triggered remotely when the ip_nat_snmp_basic
module is loaded and traffic on port 161 or 162 is NATed.

Found by multiple testcases of the trap-app and trap-enc groups of the
PROTOS c06-snmpv1 testsuite.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IRDA]: fixup type of ->lsap_state
Alexey Dobriyan [Mon, 22 May 2006 23:54:30 +0000 (16:54 -0700)] 
[IRDA]: fixup type of ->lsap_state

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IRDA]: fix 16/32 bit confusion
Alexey Dobriyan [Mon, 22 May 2006 23:54:08 +0000 (16:54 -0700)] 
[IRDA]: fix 16/32 bit confusion

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: Fix "ntohl(ntohs" bugs
Alexey Dobriyan [Mon, 22 May 2006 23:53:22 +0000 (16:53 -0700)] 
[NET]: Fix "ntohl(ntohs" bugs

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[BNX2]: Use kmalloc instead of array
Michael Chan [Mon, 22 May 2006 23:39:20 +0000 (16:39 -0700)] 
[BNX2]: Use kmalloc instead of array

Use kmalloc() instead of a local array in bnx2_nvram_write().

Update version to 1.4.40.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[BNX2]: Fix bug in bnx2_nvram_write()
Michael Chan [Mon, 22 May 2006 23:38:38 +0000 (16:38 -0700)] 
[BNX2]: Fix bug in bnx2_nvram_write()

Fix a bug in bnx2_nvram_write() caused by a counter variable not
correctly incremented by 4.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[TG3]: Add some missing rx error counters
Michael Chan [Mon, 22 May 2006 23:36:27 +0000 (16:36 -0700)] 
[TG3]: Add some missing rx error counters

Add some missing rx error counters for 5705 and newer chips.

Update version to 3.58.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC]: Add robust futex syscall entries.
David S. Miller [Mon, 22 May 2006 04:22:53 +0000 (21:22 -0700)] 
[SPARC]: Add robust futex syscall entries.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SUNSU]: Fix license.
Andrew Morton [Mon, 15 May 2006 21:10:11 +0000 (14:10 -0700)] 
[SUNSU]: Fix license.

FATAL: modpost: GPL-incompatible module sunsu uses the GPL-only symbol tty_insert_flip_string_flags

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years agoRevert "[PATCH] sched: fix interactive task starvation"
Linus Torvalds [Mon, 22 May 2006 01:54:09 +0000 (18:54 -0700)] 
Revert "[PATCH] sched: fix interactive task starvation"

This reverts commit 5ce74abe788a26698876e66b9c9ce7e7acc25413 (and its
dependent commit 8a5bc075b8d8cf7a87b3f08fad2fba0f5d13295e), because of
audio underruns.

Reported by Rene Herman <rene.herman@keyaccess.nl>, who also pinpointed
the exact cause of the underruns:

  "Audio underruns galore, with only ogg123 and firefox (browsing the
   GIT tree online is also a nice trigger by the way).

   If I back it out, everything is fine for me again."

Cc: Rene Herman <rene.herman@keyaccess.nl>
Cc: Mike Galbraith <efault@gmx.de>
Acked-by: Con Kolivas <kernel@kolivas.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/mchehab/v4l-dvb
Linus Torvalds [Mon, 22 May 2006 01:31:53 +0000 (18:31 -0700)] 
Merge /pub/scm/linux/kernel/git/mchehab/v4l-dvb

* master.kernel.org:/pub/scm/linux/kernel/git/mchehab/v4l-dvb: (33 commits)
  V4L/DVB (3965): Fix CONFIG_VIDEO_VIVI=y build bug
  V4L/DVB (3964): Bt8xx/bttv-cards.c: fix off-by-one errors
  V4L/DVB (3914): Vivi build fix
  V4L/DVB (3912): Sparc32 vivi fix
  V4L/DVB (3832): Get_dvb_firmware: download nxt2002 firmware from new driver location
  V4L/DVB (3829): Fix frequency values in the ranges structures of the LG TDVS H06xF tuners
  V4L/DVB (3826): Saa7134: Missing 'break' in Terratec Cinergy 400 TV initialization
  V4L/DVB (3825): Remove broken 'fast firmware load' from cx25840.
  V4L/DVB (3819): Cxusb-bluebird: bug-fix: power down corrupts frontend
  V4L/DVB (3813): Add support for TCL M2523_5N_E tuner.
  V4L/DVB (3804): Tweak bandselect setup fox cx24123
  V4L/DVB (3803): Various correctness fixes to tuning.
  V4L/DVB (3797): Always wait for diseqc queue to become ready before transmitting a diseqc message
  V4L/DVB (3796): Add several debug messages to cx24123 code
  V4L/DVB (3795): Fix for CX24123 & low symbol rates
  V4L/DVB (3792): Kbuild: DVB_BT8XX must select DVB_ZL10353
  V4L/DVB (3790): Use after free in drivers/media/video/em28xx/em28xx-video.c
  V4L/DVB (3788): Fix compilation with V4L1_COMPAT
  V4L/DVB (3782): Removed uneeded stuff from pwc Makefile
  V4L/DVB (3775): Add VIVI Kconfig stuff
  ...

18 years ago[PATCH] ad1848 section fix
Andrew Morton [Sat, 20 May 2006 22:00:36 +0000 (15:00 -0700)] 
[PATCH] ad1848 section fix

WARNING: sound/oss/ad1848.o - Section mismatch: reference to .init.data:ad1848_isapnp_list from .text between 'ad1848_init_generic' (at offset 0x46f0) and 'kmalloc'
WARNING: sound/oss/ad1848.o - Section mismatch: reference to .init.data:ad1848_isapnp_list from .text between 'ad1848_init_generic' (at offset 0x46f8) and 'kmalloc'
WARNING: sound/oss/ad1848.o - Section mismatch: reference to .init.data:ad1848_isapnp_list from .text between 'ad1848_init_generic' (at offset 0x4818) and 'kmalloc'

Also,

sound/oss/ad1848.c: In function `ad1848_init':
sound/oss/ad1848.c:2029: warning: cast to pointer from integer of different size
sound/oss/ad1848.c: In function `ad1848_unload':
sound/oss/ad1848.c:2178: warning: cast to pointer from integer of different size
sound/oss/ad1848.c: In function `adintr':
sound/oss/ad1848.c:2207: warning: cast from pointer to integer of different size

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] nm256_audio section fix
Andrew Morton [Sat, 20 May 2006 22:00:35 +0000 (15:00 -0700)] 
[PATCH] nm256_audio section fix

WARNING: sound/oss/nm256_audio.o - Section mismatch: reference to .init.text:nm256_peek_for_sig from .text between 'nm256_install' (at offset 0x3ba4) and 'nm256_probe'                                                                         WARNING: sound/oss/nm256_audio.o - Section mismatch: reference to .init.text:nm256_peek_for_sig from .text between 'nm256_install' (at offset 0x3bac) and 'nm256_probe'                                                                         WARNING: sound/oss/nm256_audio.o - Section mismatch: reference to .init.text: from .text between 'nm256_install' (at offset 0x3dcc) and 'nm256_probe'           WARNING: sound/oss/nm256_audio.o - Section mismatch: reference to .init.text: from .text between 'nm256_install' (at offset 0x3dd0) and 'nm256_probe'

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] es18xx build fix
Andrew Morton [Sat, 20 May 2006 22:00:35 +0000 (15:00 -0700)] 
[PATCH] es18xx build fix

sound/isa/es18xx.c: In function `snd_es18xx_identify':
sound/isa/es18xx.c:1606: warning: implicit declaration of function `udelay'

Cc: Jaroslav Kysela <perex@suse.cz>
Cc: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] mpu401 section fix
Andrew Morton [Sat, 20 May 2006 22:00:34 +0000 (15:00 -0700)] 
[PATCH] mpu401 section fix

WARNING: sound/drivers/mpu401/snd-mpu401.o - Section mismatch: reference to .init.text: from .text between 'snd_mpu401_pnp_probe' (at offset 0x1f7) and 'snd_mpu401_pnp_remove'

Cc: Jaroslav Kysela <perex@suse.cz>
Cc: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i810 section fix
Andrew Morton [Sat, 20 May 2006 22:00:33 +0000 (15:00 -0700)] 
[PATCH] i810 section fix

WARNING: drivers/video/i810/i810fb.o - Section mismatch: reference to .init.data: from .text between 'i810_fix_offsets' (at offset 0x1b88) and 'i810_alloc_agp_mem'
WARNING: drivers/video/i810/i810fb.o - Section mismatch: reference to .init.data: from .text between 'i810_fix_offsets' (at offset 0x1b8f) and 'i810_alloc_agp_mem'
WARNING: drivers/video/i810/i810fb.o - Section mismatch: reference to .init.data: from .text between 'i810_fix_offsets' (at offset 0x1ba3) and 'i810_alloc_agp_mem'
WARNING: drivers/video/i810/i810fb.o - Section mismatch: reference to .init.data: from .text between 'i810_fix_offsets' (at offset 0x1bb5) and 'i810_alloc_agp_mem'
WARNING: drivers/video/i810/i810fb.o - Section mismatch: reference to .init.data: from .text between 'i810_fix_offsets' (at offset 0x1bc6) and 'i810_alloc_agp_mem'
WARNING: drivers/video/i810/i810fb.o - Section mismatch: reference to .init.data: from .text between 'i810_init_defaults' (at offset 0x1dd8) and 'i810_init_device'
WARNING: drivers/video/i810/i810fb.o - Section mismatch: reference to .init.data: from .text between 'i810_init_defaults' (at offset 0x1dfb) and 'i810_init_device'

Cc: "Antonino A. Daplas" <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] pd6729 section fix
Andrew Morton [Sat, 20 May 2006 22:00:32 +0000 (15:00 -0700)] 
[PATCH] pd6729 section fix

WARNING: drivers/pcmcia/pd6729.o - Section mismatch: reference to .init.text: from .text between 'pd6729_pci_probe' (at offset 0x9a8) and 'pd6729_pci_remove'

Cc: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Align the node_mem_map endpoints to a MAX_ORDER boundary
Bob Picco [Sat, 20 May 2006 22:00:31 +0000 (15:00 -0700)] 
[PATCH] Align the node_mem_map endpoints to a MAX_ORDER boundary

Andy added code to buddy allocator which does not require the zone's
endpoints to be aligned to MAX_ORDER.  An issue is that the buddy allocator
requires the node_mem_map's endpoints to be MAX_ORDER aligned.  Otherwise
__page_find_buddy could compute a buddy not in node_mem_map for partial
MAX_ORDER regions at zone's endpoints.  page_is_buddy will detect that
these pages at endpoints are not PG_buddy (they were zeroed out by bootmem
allocator and not part of zone).  Of course the negative here is we could
waste a little memory but the positive is eliminating all the old checks
for zone boundary conditions.

SPARSEMEM won't encounter this issue because of MAX_ORDER size constraint
when SPARSEMEM is configured.  ia64 VIRTUAL_MEM_MAP doesn't need the logic
either because the holes and endpoints are handled differently.  This
leaves checking alloc_remap and other arches which privately allocate for
node_mem_map.

Signed-off-by: Bob Picco <bob.picco@hp.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] kobject: quiet errors in kobject_add
Greg Kroah-Hartman [Sat, 20 May 2006 22:00:30 +0000 (15:00 -0700)] 
[PATCH] kobject: quiet errors in kobject_add

People don't like released kernels yelling at them, no matter how real the
error might be.  So only report it if CONFIG_KOBJECT_DEBUG is enabled.

Sent on request of Andrew Morton.

(akpm: should bring this back post-2.6.17)

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] rtc subsystem: use ENOIOCTLCMD and ENOTTY where appropriate
Alessandro Zummo [Sat, 20 May 2006 22:00:29 +0000 (15:00 -0700)] 
[PATCH] rtc subsystem: use ENOIOCTLCMD and ENOTTY where appropriate

Appropriately use -ENOIOCTLCMD and -ENOTTY when the ioctl is not
implemented by a driver.

(akpm: we're not allowed to return -ENOIOCTLCMD to userspace.  This patch does
the right thing).

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] kbuild: fix modpost segfault for 64bit mipsel kernel
Atsushi Nemoto [Sat, 20 May 2006 22:00:28 +0000 (15:00 -0700)] 
[PATCH] kbuild: fix modpost segfault for 64bit mipsel kernel

Here is an updated r_info layout fix.  Please apply "check SHT_REL
sections" patch before this.

64bit mips has different r_info layout.  This patch fixes modpost
segfault for 64bit little endian mips kernel.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] kbuild: check SHT_REL sections
Atsushi Nemoto [Sat, 20 May 2006 22:00:26 +0000 (15:00 -0700)] 
[PATCH] kbuild: check SHT_REL sections

I found that modpost can not detect section mismatch on mips and i386.  On
mips64, the modpost (with r_info layout fix) can detect it.  The current
modpst only checks SHT_RELA section but I suppose SHT_REL section should be
checked also.  This patch does not contain r_info layout fix.  I'll post an
updated r_info layout fix on next mail.

Check SHT_REL sections as like as SHT_RELA sections to detect section
mismatch.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] s390: next_timer_interrupt overflow in stop_hz_timer
Martin Schwidefsky [Sat, 20 May 2006 22:00:25 +0000 (15:00 -0700)] 
[PATCH] s390: next_timer_interrupt overflow in stop_hz_timer

The 32 bit unsigned substraction (next - jiffies) in stop_hz_timer can
overflow if jiffies gets advanced between next_timer_interrupt and the read
under the xtime lock.  The cast to a u64 then results in a large value
which causes the cpu to wait too long.  Fix this by casting next and
jiffies independently to u64 before subtracting them.

(Spotted by Zachary Amsden <zach@vmware.com>)

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Fix a NO_IDLE_HZ timer bug
Zachary Amsden [Sat, 20 May 2006 22:00:24 +0000 (15:00 -0700)] 
[PATCH] Fix a NO_IDLE_HZ timer bug

Under certain timing conditions, a race during boot occurs where timer
ticks are being processed on remote CPUs.  The remote timer ticks can
increment jiffies, and if this happens during a window when a timeout is
very close to expiring but a local tick has not yet been delivered, you can
end up with

1) No softirq pending
2) A local timer wheel which is not synced to jiffies
3) No high resolution timer active
4) A local timer which is supposed to fire before the current jiffies value.

In this circumstance, the comparison in next_timer_interrupt overflows,
because the base of the comparison for high resolution timers is jiffies,
but for the softirq timer wheel, it is relative the the current base of the
wheel (jiffies_base).

Signed-off-by: Zachary Amsden <zach@vmware.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] missing newline in scsi/st.c
Rene Herman [Sat, 20 May 2006 22:00:22 +0000 (15:00 -0700)] 
[PATCH] missing newline in scsi/st.c

st: Version 20050830, fixed bufsize 32768, s/g segs 256
st 0:0:6:0: Attached scsi tape st0<4>st0: try direct i/o: yes (alignment 512 B)

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] kprobes: bad manipulation of 2 byte opcode on x86_64
Satoshi Oshima [Sat, 20 May 2006 22:00:21 +0000 (15:00 -0700)] 
[PATCH] kprobes: bad manipulation of 2 byte opcode on x86_64

Problem:

If we put a probe onto a callq instruction and the probe is executed,
kernel panic of Bad RIP value occurs.

Root cause:

If resume_execution() found 0xff at first byte of p->ainsn.insn, it must
check the _second_ byte.  But current resume_execution check _first_ byte
again.

I changed it checks second byte of p->ainsn.insn.

Kprobes on i386 don't have this problem, because the implementation is a
little bit different from x86_64.

Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Satoshi Oshima <soshima@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i386 kdump boot cpu physical apicid fix
Vivek Goyal [Sat, 20 May 2006 22:00:21 +0000 (15:00 -0700)] 
[PATCH] i386 kdump boot cpu physical apicid fix

o Kdump second kernel boot fails after a system crash if second kernel
  is UP and acpi=off and if crash occurred on a non-boot cpu.

o Issue here is that MP tables report boot cpu lapic id as 0 but second
  kernel is booting on a different processor and MP table data is stale
  in this context. Hence apic_id_registered() check fails in setup_local_APIC()
  when called from APIC_init_uniprocessor().

o Problem is not seen if ACPI is enabled as in that case
  boot_cpu_physical_apicid is read from the LAPIC.

o Problem is not seen with SMP kernels as well because in this case also
  boot_cpu_physical_apicid is read from LAPIC. (smp_boot_cpus()).

o The problem is fixed by reading boot_cpu_physical_apicid from LAPIC
  if it is a UP kernel and CRASH_DUMP is enabled.

Signed-off-by: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] pxa2xx-spi update
Stephen Street [Sat, 20 May 2006 22:00:19 +0000 (15:00 -0700)] 
[PATCH] pxa2xx-spi update

Fix some outstanding issues with the pxa2xx_spi driver when running on a
PXA270:

- Wrong timeout calculation in the setup function due to different
  peripheral clock rates in the PXAxxx family.

- Bad handling of SSSR_TFS interrupts in interrupt_transfer function.

- Added locking to interface between the pump_messages workqueue and the
  pump_transfers tasklet.

Much thanks to Juergen Beisert for the extensive testing on the PXA270.

Signed-off-by: Stephen Street <stephen@streetfiresound.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] S3C24XX: hardware SPI driver
Ben Dooks [Sat, 20 May 2006 22:00:18 +0000 (15:00 -0700)] 
[PATCH] S3C24XX: hardware SPI driver

Hardware based SPI driver for Samsung S3C24XX SoC systems

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Cc: David Brownell <david-b@pacbell.net>
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] S3C24XX: GPIO based SPI driver
Ben Dooks [Sat, 20 May 2006 22:00:17 +0000 (15:00 -0700)] 
[PATCH] S3C24XX: GPIO based SPI driver

SPI driver for SPI by GPIO on the Samsung S3C24XX series of SoC processors.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Cc: Greg KH <greg@kroah.com>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] drivers/base/firmware_class.c: cleanups
Adrian Bunk [Sat, 20 May 2006 22:00:16 +0000 (15:00 -0700)] 
[PATCH] drivers/base/firmware_class.c: cleanups

- remove the following global function that is both unused and
  unimplemented:
  - register_firmware()

- make the following needlessly global function static:
  - firmware_class_uevent()

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] spi: add spi master driver for Freescale MPC83xx SPI controller
Kumar Gala [Sat, 20 May 2006 22:00:15 +0000 (15:00 -0700)] 
[PATCH] spi: add spi master driver for Freescale MPC83xx SPI controller

This driver supports the SPI controller on the MPC83xx SoC devices from
Freescale.  Note, this driver supports only the simple shift register SPI
controller and not the descriptor based CPM or QUICCEngine SPI controller.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
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] minor SPI doc fix
dmitry pervushin [Sat, 20 May 2006 22:00:14 +0000 (15:00 -0700)] 
[PATCH] minor SPI doc fix

Because several developers asked me about referenced but missing
spi_add_master(), I think that this patch should be applied ...  it
corrects comments so they refer to spi_register_master() instead.

Signed-off-by: dmitry pervushin <dpervushin@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Update ext2/ext3/jbd MAINTAINERS entries
Theodore Tso [Sat, 20 May 2006 22:00:13 +0000 (15:00 -0700)] 
[PATCH] Update ext2/ext3/jbd MAINTAINERS entries

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Clarify maintainers and include linux-security info
Alan Cox [Sat, 20 May 2006 22:00:12 +0000 (15:00 -0700)] 
[PATCH] Clarify maintainers and include linux-security info

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Overrun in isdn_tty.c
Eric Sesterhenn [Sat, 20 May 2006 22:00:12 +0000 (15:00 -0700)] 
[PATCH] Overrun in isdn_tty.c

This fixes coverity bug id #1237.  After the while loop, it is possible for
i == ISDN_LMSNLEN.  If this happens the terminating '\0' is written after
the end of the array.

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Cc: 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] cpuset: might_sleep_if check in cpuset_zones_allowed
Paul Jackson [Sat, 20 May 2006 22:00:11 +0000 (15:00 -0700)] 
[PATCH] cpuset: might_sleep_if check in cpuset_zones_allowed

It's too easy to incorrectly call cpuset_zone_allowed() in an atomic
context without __GFP_HARDWALL set, and when done, it is not noticed until
a tight memory situation forces allocations to be tried outside the current
cpuset.

Add a 'might_sleep_if()' check, to catch this earlier on, instead of
waiting for a similar check in the mutex_lock() code, which is only rarely
invoked.

Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] cpuset: update cpuset_zones_allowed comment
Paul Jackson [Sat, 20 May 2006 22:00:10 +0000 (15:00 -0700)] 
[PATCH] cpuset: update cpuset_zones_allowed comment

Update the kernel/cpuset.c:cpuset_zone_allowed() comment.

The rule for when mm/page_alloc.c should call cpuset_zone_allowed()
was intended to be:

  Don't call cpuset_zone_allowed() if you can't sleep, unless you
  pass in the __GFP_HARDWALL flag set in gfp_flag, which disables
  the code that might scan up ancestor cpusets and sleep.

The explanation of this rule in the comment above cpuset_zone_allowed() was
stale, as a result of a restructuring of some __alloc_pages() code in
November 2005.

Rewrite that comment ...

Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Cpuset: might sleep checking zones allowed fix
Paul Jackson [Sat, 20 May 2006 22:00:09 +0000 (15:00 -0700)] 
[PATCH] Cpuset: might sleep checking zones allowed fix

Fix a couple of infrequently encountered 'sleeping function called from
invalid context' in the cpuset hooks in __alloc_pages.  Could sleep while
interrupts disabled.

The routine cpuset_zone_allowed() is called by code in mm/page_alloc.c
__alloc_pages() to determine if a zone is allowed in the current tasks
cpuset.  This routine can sleep, for certain GFP_KERNEL allocations, if the
zone is on a memory node not allowed in the current cpuset, but might be
allowed in a parent cpuset.

But we can't sleep in __alloc_pages() if in interrupt, nor if called for a
GFP_ATOMIC request (__GFP_WAIT not set in gfp_flags).

The rule was intended to be:
  Don't call cpuset_zone_allowed() if you can't sleep, unless you
  pass in the __GFP_HARDWALL flag set in gfp_flag, which disables
  the code that might scan up ancestor cpusets and sleep.

This rule was being violated in a couple of places, due to a bogus change
made (by myself, pj) to __alloc_pages() as part of the November 2005 effort
to cleanup its logic, and also due to a later fix to constrain which swap
daemons were awoken.

The bogus change can be seen at:
  http://linux.derkeiler.com/Mailing-Lists/Kernel/2005-11/4691.html
  [PATCH 01/05] mm fix __alloc_pages cpuset ALLOC_* flags

This was first noticed on a tight memory system, in code that was disabling
interrupts and doing allocation requests with __GFP_WAIT not set, which
resulted in __might_sleep() writing complaints to the log "Debug: sleeping
function called ...", when the code in cpuset_zone_allowed() tried to take
the callback_sem cpuset semaphore.

We haven't seen a system hang on this 'might_sleep' yet, but we are at
decent risk of seeing it fairly soon, especially since the additional
cpuset_zone_allowed() check was added, conditioning wakeup_kswapd(), in
March 2006.

Special thanks to Dave Chinner, for figuring this out, and a tip of the hat
to Nick Piggin who warned me of this back in Nov 2005, before I was ready
to listen.

Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] pci: correctly allocate return buffers for osc calls
Kristen Accardi [Sat, 20 May 2006 22:00:08 +0000 (15:00 -0700)] 
[PATCH] pci: correctly allocate return buffers for osc calls

The OSC set and query functions do not allocate enough space for return
values, and set the output buffer length to a false, too large value.  This
causes the acpi-ca code to assume that the output buffer is larger than it
actually is, and overwrite memory when copying acpi return buffers into
this caller provided buffer.  In some cases this can cause kernel oops if
the memory that is overwritten is a pointer.  This patch will change these
calls to use a dynamically allocated output buffer, thus allowing the
acpi-ca code to decide how much space is needed.

Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Cc: "Brown, Len" <len.brown@intel.com>
Cc: "Yu, Luming" <luming.yu@intel.com>
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] fix NULL dereference in inotify_ignore
Amy Griffis [Sat, 20 May 2006 22:00:07 +0000 (15:00 -0700)] 
[PATCH] fix NULL dereference in inotify_ignore

Don't reassign to watch.  If idr_find() returns NULL, then
put_inotify_watch() will choke.

Signed-off-by: Amy Griffis <amy.griffis@hp.com>
Cc: John McCutchan <john@johnmccutchan.com>
Cc: Robert Love <rlove@rlove.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fix race in inotify_release
Amy Griffis [Sat, 20 May 2006 22:00:06 +0000 (15:00 -0700)] 
[PATCH] fix race in inotify_release

While doing some inotify stress testing, I hit the following race.  In
inotify_release(), it's possible for a watch to be removed from the lists
in between dropping dev->mutex and taking inode->inotify_mutex.  The
reference we hold prevents the watch from being freed, but not from being
removed.

Checking the dev's idr mapping will prevent a double list_del of the
same watch.

Signed-off-by: Amy Griffis <amy.griffis@hp.com>
Acked-by: John McCutchan <john@johnmccutchan.com>
Cc: Robert Love <rml@novell.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] SPARSEMEM incorrectly calculates section number
Mike Kravetz [Sat, 20 May 2006 22:00:05 +0000 (15:00 -0700)] 
[PATCH] SPARSEMEM incorrectly calculates section number

A bad calculation/loop in __section_nr() could result in incorrect section
information being put into sysfs memory entries.  This primarily impacts
memory add operations as the sysfs information is used while onlining new
memory.

Fix suggested by Dave Hansen.

Note that the bug may not be obvious from the patch.  It actually occurs in
the function's return statement:

return (root_nr * SECTIONS_PER_ROOT) + (ms - root);

In the existing code, root_nr has already been multiplied by
SECTIONS_PER_ROOT.

Signed-off-by: Mike Kravetz <kravetz@us.ibm.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] build fix: CONFIG_MEMORY_HOTPLUG=y on i386
KAMEZAWA Hiroyuki [Sat, 20 May 2006 22:00:03 +0000 (15:00 -0700)] 
[PATCH] build fix: CONFIG_MEMORY_HOTPLUG=y on i386

typo in #ifdefs.  Fixes http://bugme.osdl.org/show_bug.cgi?id=6538

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] selinux: endian fix
Alexey Dobriyan [Sat, 20 May 2006 22:00:02 +0000 (15:00 -0700)] 
[PATCH] selinux: endian fix

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Stephen Smalley <sds@tycho.nsa.gov>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] binfmt_flat: don't check for EMFILE
Andrew Morton [Sat, 20 May 2006 22:00:01 +0000 (15:00 -0700)] 
[PATCH] binfmt_flat: don't check for EMFILE

Bernd Schmidt points out that binfmt_flat is now leaving the exec file open
while the application runs.  This offsets all the application's fd numbers.
We should have closed the file within exec(), not at exit()-time.

But there doesn't seem to be a lot of point in doing all this just to avoid
going over RLIMIT_NOFILE by one fd for a few microseconds.  So take the EMFILE
checking out again.  This will cause binfmt_flat to again fail LTP's
exec-should-return-EMFILE-when-fdtable-is-full test.  That test appears to be
wrong anyway - Open Group specs say nothing about exec() returning EMFILE.

Cc: Bernd Schmidt <bernd.schmidt@analog.com>
Cc: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] HID read busywait fix
Micon, David [Sat, 20 May 2006 21:59:59 +0000 (14:59 -0700)] 
[PATCH] HID read busywait fix

Make a read of a HID device block until data is available.  Without it, the
read goes into a busy-wait loop until data is available.

Cc: Greg KH <greg@kroah.com>
Acked-by: Vojtech Pavlik <vojtech@suse.cz>
Cc: Dmitry Torokhov <dtor_core@ameritech.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] nfsd: sign conversion obscuring errors in nfsd_set_posix_acl()
Florin Malita [Sat, 20 May 2006 21:59:58 +0000 (14:59 -0700)] 
[PATCH] nfsd: sign conversion obscuring errors in nfsd_set_posix_acl()

Assigning the result of posix_acl_to_xattr() to an unsigned data type
(size/size_t) obscures possible errors.

Coverity CID: 1206.

Signed-off-by: Florin Malita <fmalita@gmail.com>
Acked-by: NeilBrown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] md: Fix inverted test for 'repair' directive.
NeilBrown [Sat, 20 May 2006 21:59:57 +0000 (14:59 -0700)] 
[PATCH] md: Fix inverted test for 'repair' directive.

We should be able to write 'repair' to /sys/block/mdX/md/sync_action,
however due to and inverted test, that always given EINVAL.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] NFS server subtree_check returns dubious value
Peter Staubach [Sat, 20 May 2006 21:59:56 +0000 (14:59 -0700)] 
[PATCH] NFS server subtree_check returns dubious value

Address a problem found when a Linux NFS server uses the "subtree_check"
export option.

The "subtree_check" NFS export option was designed to prohibit a client
from using a file handle for which it should not have permission.  The
algorithm used is to ensure that the entire path to the file being
referenced is accessible to the user attempting to use the file handle.  If
some part of the path is not accessible, then the operation is aborted and
the appropriate version of ESTALE is returned to the NFS client.

The error, ESTALE, is unfortunate in that it causes NFS clients to make
certain assumptions about the continued existence of the file.  They assume
that the file no longer exists and refuse to attempt to access it again.
In this case, the file really does exist, but access was denied by the
server for a particular user.

A better error to return would be an EACCES sort of error.  This would
inform the client that the particular operation that it was attempting was
not allowed, without the nasty side effects of the ESTALE error.

Signed-off-by: Peter Staubach <staubach@redhat.com>
Acked-By: NeilBrown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Kdump maintainer info update
Vivek Goyal [Sat, 20 May 2006 21:59:55 +0000 (14:59 -0700)] 
[PATCH] Kdump maintainer info update

Update MAINTAINERS file for info regarding kdump maintainership.

Signed-off-by: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] powerpc: Fix ide-pmac sysfs entry
Benjamin Herrenschmidt [Sat, 20 May 2006 21:59:53 +0000 (14:59 -0700)] 
[PATCH] powerpc: Fix ide-pmac sysfs entry

It looks like the generic ide code now wants ide_init_hwif_ports() to set
the parent struct device into the ide_hw structure (new field ?).  Without
this, the mac ide code can cause the ide probing code to explode in flames
in sysfs registration due to what looks like a stale pointer in there
(happens when removing/re-inserting one of the hotswap media bays on some
laptops).

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Bartlomiej Zolnierkiewicz <B.Zolnierkiewicz@elka.pw.edu.pl>
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] i386: remove junk from stack dump
Chuck Ebbert [Sat, 20 May 2006 21:59:52 +0000 (14:59 -0700)] 
[PATCH] i386: remove junk from stack dump

i386 stack dump has a "<0>" in the middle of the line and an extra space
between columns in multicolumn mode.  Remove those and also remove an extra
blank line of source code.

Signed-off-by: Chuck Ebbert <76306.1226@compuserve.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] matroxfb: fix DVI setup to be more compatible
Paul A. Clarke [Sat, 20 May 2006 21:59:51 +0000 (14:59 -0700)] 
[PATCH] matroxfb: fix DVI setup to be more compatible

There has been a longstanding problem with the Matrox G450 and perhaps
other similar cards, with modes "above" 1280x1024-60 on ppc/ppc64 boxes
running Linux.  Higher resolutions and/or higher refresh rates resulted in
a very noticably "jittery" display, and sometimes no display, depending on
the physical monitor.  This patch fixes that problem on the systems I have
easy access to...

I've tested with SLES9SP3 (2.6.5+ kernel) and 2.6.16-rc6 custom kernels on
an IBM eServer p5 520 w/G450 (a.k.a GXT135P on IBM's ppc64 systems), and a
colleague of mine (Ian Romanick) tested it successfully on an Apple ppc32
box (w/GXT135P).  I also tested it on IA32 box I have with a GXT135P to
verify that it didn't obviously break anything.  In my testing, I covered
single-card, single and dual-head setups using both HD15 and DVI-D signals,
on both the IA32 and ppc64 boxes.  While everything appeared fine on both
boxes, I did encounter one problem: I can't get any signal on the DVI-D
output on the ppc64 box.  However, this is also the case without my patch.

I just noticed that screen-blanking only occurs on the primary display as
well.

Signed-off-by: Paul A. Clarke <pc@us.ibm.com>
Signed-off-by: Ian Romanick <idr@us.ibm.com>
Signed-off-by: Petr Vandrovec <petr@vandrovec.name>
Cc: "Antonino A. Daplas" <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] NFS: fix error handling on access_ok in compat_sys_nfsservctl
Lin Feng Shen [Sat, 20 May 2006 21:59:49 +0000 (14:59 -0700)] 
[PATCH] NFS: fix error handling on access_ok in compat_sys_nfsservctl

Functions compat_nfs_svc_trans, compat_nfs_clnt_trans,
compat_nfs_exp_trans, compat_nfs_getfd_trans and compat_nfs_getfs_trans,
which are called by compat_sys_nfsservctl(fs/compat.c), don't handle the
return value of access_ok properly.  access_ok return 1 when the addr is
valid, and 0 when it's not, but these functions have the reversed
understanding.  When the address is valid, they always return -EFAULT to
compat_sys_nfsservctl.

An example is to run /usr/sbin/rpc.nfsd(32bit program on Power5).  It
doesn't function as expected.  strace showes that nfsservctl returns
-EFAULT.

The patch fixes this by correcting the error handling on the return value
of access_ok in the five functions.

Signed-off-by: Lin Feng Shen <shenlinf@cn.ibm.com>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Acked-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] forcedeth: fix multi irq issues
Ayaz Abdulla [Sat, 20 May 2006 21:59:48 +0000 (14:59 -0700)] 
[PATCH] forcedeth: fix multi irq issues

With Manfred Spraul <manfred@colorfullife.com> and
     Andrew Morton <akpm@osdl.org>

Bring back this recently-reverted patch, only fixed.

Original changelog:

    From: Ayaz Abdulla <aabdulla@nvidia.com>

    This patch fixes the issues with multiple irqs.

    I am resending based on feedback. I decoupled the dma mask for
    consistent memory and fixed leak with multiple irq in error path.

    Thanks to Manfred for catching the spin lock problem.

Fix it:

    From: Manfred Spraul <manfred@colorfullife.com>

    Fix bug introduced by ebf34c9b6fcd22338ef764b039b3ac55ed0e297b, covered in
    http://bugzilla.kernel.org/show_bug.cgi?id=6568.

    Remove second instance of the request_irq() calls: they were moved
    from nv_open into nv_request_irq.

    Thanks to Alistair Strachan <alistair@devzero.co.uk> for reporting and
    persisting.

Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: Manfred Spraul <manfred@colorfullife.com>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge branch 'upstream-fixes' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Sat, 20 May 2006 17:35:41 +0000 (10:35 -0700)] 
Merge branch 'upstream-fixes' of /linux/kernel/git/jgarzik/libata-dev

* 'upstream-fixes' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  [PATCH] libata-core: fix current kernel-doc warnings
  [PATCH] sata_mv: version bump
  [PATCH] sata_mv: endian fix
  [PATCH] sata_mv: remove local copy of queue indexes
  [PATCH] sata_mv: spurious interrupt workaround
  [PATCH] sata_mv: chip initialization fixes
  [PATCH] sata_mv: deal with interrupt coalescing interrupts
  [PATCH] sata_mv: prevent unnecessary double-resets

18 years ago[PATCH] Add PCI ID for the Intel IDE Controller which is in the Intel Mac Minis shipp...
Thomas Glanzmann [Fri, 12 May 2006 08:00:41 +0000 (10:00 +0200)] 
[PATCH] Add PCI ID for the Intel IDE Controller which is in the Intel Mac Minis shipped in first quarter 2006

Signed-off-by: Thomas Glanzmann <sithglan@stud.uni-erlangen.de>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] SCSI: make scsi_implement_eh() generic API for SCSI transports
Tejun Heo [Fri, 19 May 2006 12:07:05 +0000 (21:07 +0900)] 
[PATCH] SCSI: make scsi_implement_eh() generic API for SCSI transports

libata implemented a feature to schedule EH without an associated EH
by manipulating shost->host_eh_scheduled in ata_scsi_schedule_eh()
directly.  Move this function to scsi_error.c and rename it to
scsi_schedule_eh().  It is now an exported API for SCSI transports and
exported via new header file drivers/scsi/scsi_transport_api.h

This patch also de-export scsi_eh_wakeup() which was exported
specifically for ata_scsi_schedule_eh().

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: enforce default EH actions
Tejun Heo [Tue, 16 May 2006 03:58:24 +0000 (12:58 +0900)] 
[PATCH] libata: enforce default EH actions

LLDDs rely on libata that certain EH actions are automatically taken
on some errors.  If the port is frozen or one or more qc's have failed
with HSM violation or timeout, softreset is enforced (LLDD can ask for
storonger EH action at will).  If any other error condition exists,
libata EH always revalidates.

This behavior existed in earlier revisions of new EH but lost during
development process.  This patch restores it.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: use qc->result_tf for temp taskfile storage
Albert Lee [Thu, 18 May 2006 09:51:10 +0000 (17:51 +0800)] 
[PATCH] libata: use qc->result_tf for temp taskfile storage

Use qc->result_tf for temp taskfile storage.

Signed-off-by: Albert Lee <albertcc@tw.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: Fix the HSM error_mask mapping (was: Re: libata-tj and SMART)
Albert Lee [Fri, 19 May 2006 03:43:04 +0000 (11:43 +0800)] 
[PATCH] libata: Fix the HSM error_mask mapping (was: Re: libata-tj and SMART)

Fix the HSM error_mask mapping.

Changes:
- Better mapping in ac_err_mask()
- In HSM_ST_FIRST ans HSM_ST state, check ATA_ERR|ATA_DF and map it to AC_ERR_DEV instead of AC_ERR_HSM.
- In HSM_ST_FIRST and HSM_ST state, map DRQ=1 ERR=1 to AC_ERR_HSM.
- For PIO data in and DRQ=1 ERR=1, add check after the junk data block is read.

Signed-off-by: Albert Lee <albertcc@tw.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years agoMerge branch 'upstream-fixes' into upstream
Jeff Garzik [Sat, 20 May 2006 04:36:08 +0000 (00:36 -0400)] 
Merge branch 'upstream-fixes' into upstream

Conflicts:

drivers/scsi/libata-core.c

18 years ago[PATCH] libata-core: fix current kernel-doc warnings
Randy Dunlap [Thu, 18 May 2006 17:50:18 +0000 (10:50 -0700)] 
[PATCH] libata-core: fix current kernel-doc warnings

Fix all current kernel-doc warnings.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_mv: version bump
Mark Lord [Fri, 19 May 2006 20:41:27 +0000 (16:41 -0400)] 
[PATCH] sata_mv: version bump

Increment the version number inside sata_mv.c.

Signed-off-by: Mark Lord <liml@rtr.ca>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_mv: endian fix
Mark Lord [Fri, 19 May 2006 20:40:15 +0000 (16:40 -0400)] 
[PATCH] sata_mv: endian fix

This fixes a byte-swap issue on PPC, found by Zang Roy-r61911
on the powerpc platform.  His original patch also had some other
platform-specific changes in #ifdef's, but I'm not sure yet how to
incorporate them.  Look for another patch for those (soon).

Signed-off-by: Mark Lord <liml@rtr.ca>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_mv: remove local copy of queue indexes
Mark Lord [Fri, 19 May 2006 20:36:36 +0000 (16:36 -0400)] 
[PATCH] sata_mv: remove local copy of queue indexes

The driver currently keeps local copies of the hardware request/response queue indexes.
But it expends significant effort ensuring consistency between the two views,
and still gets it wrong after an error or reset occurs.

This patch removes the local copies, in favour of just accessing the hardware
whenever we need them.  Eventually this may need to be tweaked again for NCQ,
but for now this works and solves problems some users were seeing.

Signed-off-by: Mark Lord <liml@rtr.ca>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_mv: spurious interrupt workaround
Mark Lord [Fri, 19 May 2006 20:33:03 +0000 (16:33 -0400)] 
[PATCH] sata_mv: spurious interrupt workaround

The 60xx chips, and possibly others, incorrectly assert DEV_IRQ interrupts
on a regular basis.  The cause of this is under investigation (by me and
in theory by Marvell also), but regardless we do need to deal with these events.

This patch tidies up some interrupt handler code, and ensures that we ignore
DEV_IRQ interrupts when the drive still  has ATA_BUSY asserted.

Signed-off-by: Mark Lord <liml@rtr.ca>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_mv: chip initialization fixes
Mark Lord [Fri, 19 May 2006 20:29:21 +0000 (16:29 -0400)] 
[PATCH] sata_mv: chip initialization fixes

The interface control register of the 60xx (and later) Marvell chip
requires certain bits to always be set when writing to it.  These bits
incorrectly read-back as zeros, so the pattern must be ORed in
with each write of the register.  Also, bit 12 should NOT be set
(note that Marvell's own driver also had bit-12 wrong here).

While we're at it, we also now do pci_set_master() in the init code.

Signed-off-by: Mark Lord <liml@rtr.ca>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_mv: deal with interrupt coalescing interrupts
Mark Lord [Fri, 19 May 2006 20:24:56 +0000 (16:24 -0400)] 
[PATCH] sata_mv: deal with interrupt coalescing interrupts

In some systems, it is possible that the BIOS may have enabled interrupt coalescing
for the Marvell controllers which support it.  This patch adds code to detect/ack
interrupts from the chip's coalescing (combing) logic.

Signed-off-by: Mark Lord <liml@rtr.ca>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_mv: prevent unnecessary double-resets
Mark Lord [Fri, 19 May 2006 20:21:03 +0000 (16:21 -0400)] 
[PATCH] sata_mv: prevent unnecessary double-resets

The mv_err_intr() function is invoked from the driver's interrupt handler,
as well as from the timeout function.  This patch prevents it from triggering
a one-after-the-other double reset of the controller when invoked
from the timeout function.

This also adds a check for a timeout race condition that has been observed
to occur with this driver in earlier kernels.  This should not be needed,
in theory, but in practice it has caught bugs.  Maybe nuke it at a later date.

Signed-off-by: Mark Lord <liml@rtr.ca>
Signed-off-by: Jeff Garzik <jeff@garzik.org>