GitHub/MotorolaMobilityLLC/kernel-slsi.git
17 years agofix console change race exposed by CFS
Jan Lübbe [Sat, 29 Sep 2007 16:47:51 +0000 (18:47 +0200)]
fix console change race exposed by CFS

The new behaviour of CFS exposes a race which occurs if a switch is
requested when vt_mode.mode is VT_PROCESS.

The process with vc->vt_pid is signaled before vc->vt_newvt is set.
This causes the switch to fail when triggered by the monitoing process
because the target is still -1.

[ If the signal sending fails, the subsequent "reset_vc(vc)" will then
  reset vt_newvt to -1, so this works for that case too.   - Linus ]

Signed-off-by: Jan Lübbe <jluebbe@lasnet.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Sat, 29 Sep 2007 16:25:55 +0000 (09:25 -0700)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/netdev-2.6

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6:
  mv643xx_eth: Check ETH_INT_CAUSE_STATE bit

17 years agoi386: remove bogus comment about memory barrier
Nick Piggin [Sat, 29 Sep 2007 13:28:48 +0000 (15:28 +0200)]
i386: remove bogus comment about memory barrier

The comment being removed by this patch is incorrect and misleading.

In the following situation:

1. load  ...
2. store 1 -> X
3. wmb
4. rmb
5. load  a <- Y
6. store ...

4 will only ensure ordering of 1 with 5.
3 will only ensure ordering of 2 with 6.

Further, a CPU with strictly in-order stores will still only provide that
2 and 6 are ordered (effectively, it is the same as a weakly ordered CPU
with wmb after every store).

In all cases, 5 may still be executed before 2 is visible to other CPUs!

The additional piece of the puzzle that mb() provides is the store/load
ordering, which fundamentally cannot be achieved with any combination of
rmb()s and wmb()s.

This can be an unexpected result if one expected any sort of global ordering
guarantee to barriers (eg. that the barriers themselves are sequentially
consistent with other types of barriers).  However sfence or lfence barriers
need only provide an ordering partial ordering of memory operations -- Consider
that wmb may be implemented as nothing more than inserting a special barrier
entry in the store queue, or, in the case of x86, it can be a noop as the store
queue is in order. And an rmb may be implemented as a directive to prevent
subsequent loads only so long as their are no previous outstanding loads (while
there could be stores still in store queues).

I can actually see the occasional load/store being reordered around lfence on
my core2. That doesn't prove my above assertions, but it does show the comment
is wrong (unless my program is -- can send it out by request).

So:
   mb() and smp_mb() always have and always will require a full mfence
   or lock prefixed instruction on x86.  And we should remove this comment.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Cc: Paul McKenney <paulmck@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agomv643xx_eth: Check ETH_INT_CAUSE_STATE bit
Dale Farnsworth [Fri, 28 Sep 2007 13:30:43 +0000 (06:30 -0700)]
mv643xx_eth: Check ETH_INT_CAUSE_STATE bit

Commit 468d09f8946d40228c56de26fe4874b2f98067ed masked the "state"
interrupt (bit 20 of the cause register). This results in Radstone's
PPC7D repeatedly re-entering the interrupt routine, locking up the
board. The following patch returns the required handling for this
interrupt.

Signed-off-by: Martyn Welch <martyn.welch@radstone.co.uk>
Signed-off-by: Dale Farnsworth <dale@farnsworth.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoFix bogus PCI quirk for e100
Benjamin Herrenschmidt [Fri, 28 Sep 2007 23:06:21 +0000 (09:06 +1000)]
Fix bogus PCI quirk for e100

Linas reported me that some machines were crashing at boot in
quirk_e100_interrupt. It appears that this quirk is doing an ioremap
directly on a PCI BAR value, which isn't legal and will cause all sorts
of bad things to happen on architectures where PCI BARs don't directly
match processor bus addresses.

This fixes it by using the proper PCI resources instead which is possible
since the quirk has been moved by a previous commit to happen late enough
for that.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Linas Vepstas <linas@austin.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Fri, 28 Sep 2007 22:44:44 +0000 (15:44 -0700)]
Merge branch 'master' of /linux/kernel/git/davem/net-2.6

* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [TCP]: Fix MD5 signature handling on big-endian.
  [NET]: Zero length write() on socket should not simply return 0.

17 years agoInput: xpad - fix dependancy on LEDS class
Dmitry Torokhov [Thu, 27 Sep 2007 04:09:29 +0000 (00:09 -0400)]
Input: xpad - fix dependancy on LEDS class

Input: xpad - fix dependancy on LEDS class

The driver can not be built-in when LEDS class is a module.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoNFS: Fix an Oops in encode_lookup()
Trond Myklebust [Fri, 28 Sep 2007 16:27:41 +0000 (12:27 -0400)]
NFS: Fix an Oops in encode_lookup()

It doesn't look as if the NFS file name limit is being initialised correctly
in the struct nfs_server. Make sure that we limit whatever is being set in
nfs_probe_fsinfo() and nfs_init_server().

Also ensure that readdirplus and nfs4_path_walk respect our file name
limits.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge branch 'for-2.6.23' of master.kernel.org:/pub/scm/linux/kernel/git/galak/powerpc
Linus Torvalds [Fri, 28 Sep 2007 22:33:03 +0000 (15:33 -0700)]
Merge branch 'for-2.6.23' of /linux/kernel/git/galak/powerpc

* 'for-2.6.23' of master.kernel.org:/pub/scm/linux/kernel/git/galak/powerpc:
  [POWERPC] mpc8349emitx.dts: Setup USB-DR for peripheral mode.
  [POWERPC] Fix mpc834x USB-MPH configuration.
  [POWERPC] Fix cpm_uart driver for cpm1 machines
  [PPC] Fix cpm_dpram_addr returning phys mem instead of virt mem
  [POWERPC] Fix copy'n'paste typo in commproc.c

17 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Fri, 28 Sep 2007 22:32:29 +0000 (15:32 -0700)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/netdev-2.6

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6:
  e1000: Add device IDs of blade version of the 82571 quad port
  sky2: fix transmit state on resume
  sky2: FE+ vlan workaround
  sky2: sky2 FE+ receive status workaround

17 years ago[TCP]: Fix MD5 signature handling on big-endian.
David S. Miller [Fri, 28 Sep 2007 22:18:35 +0000 (15:18 -0700)]
[TCP]: Fix MD5 signature handling on big-endian.

Based upon a report and initial patch by Peter Lieven.

tcp4_md5sig_key and tcp6_md5sig_key need to start with
the exact same members as tcp_md5sig_key.  Because they
are both cast to that type by tcp_v{4,6}_md5_do_lookup().

Unfortunately tcp{4,6}_md5sig_key use a u16 for the key
length instead of a u8, which is what tcp_md5sig_key
uses.  This just so happens to work by accident on
little-endian, but on big-endian it doesn't.

Instead of casting, just place tcp_md5sig_key as the first member of
the address-family specific structures, adjust the access sites, and
kill off the ugly casts.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Fri, 28 Sep 2007 19:02:25 +0000 (12:02 -0700)]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] Fix fallocate on o32 binary compat ABI
  [MIPS] Fix CONFIG_BUILD_ELF64 kernels with symbols in CKSEG0.
  [MIPS] IP32: Fix initialization of UART base addresses.

17 years ago[MIPS] Fix fallocate on o32 binary compat ABI
Kyle McMartin [Fri, 28 Sep 2007 17:15:20 +0000 (13:15 -0400)]
[MIPS] Fix fallocate on o32 binary compat ABI

MIPS was mistakenly forgetting to use the fallocate compat wrapper, which
I noticed while cleaning up all the duplicate fallocate wrappers.

Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] Fix CONFIG_BUILD_ELF64 kernels with symbols in CKSEG0.
Ralf Baechle [Tue, 11 Sep 2007 07:50:40 +0000 (08:50 +0100)]
[MIPS] Fix CONFIG_BUILD_ELF64 kernels with symbols in CKSEG0.

The __pa() for those did assume that all symbols have XKPHYS values and
the math fails for any other address range.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] IP32: Fix initialization of UART base addresses.
Ralf Baechle [Wed, 26 Sep 2007 12:01:37 +0000 (13:01 +0100)]
[MIPS] IP32: Fix initialization of UART base addresses.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux...
Linus Torvalds [Fri, 28 Sep 2007 18:24:20 +0000 (11:24 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/hpa/linux-2.6-x86setup

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-x86setup:
  [x86 setup] Correct the SMAP check for INT 0x15, AX=0xe820

17 years ago[x86 setup] Correct the SMAP check for INT 0x15, AX=0xe820
H. Peter Anvin [Fri, 28 Sep 2007 00:17:12 +0000 (17:17 -0700)]
[x86 setup] Correct the SMAP check for INT 0x15, AX=0xe820

The e820 probe code was checking %edx, not %eax, for the SMAP
signature on return.  This worked on *almost* all systems, since %edx
still contained SMAP from the call on entry, but on a handful of
systems it failed -- plus, we would have missed real mismatches.

The error output is "=d" to make sure gcc knows %edx is clobbered
here.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
17 years ago[POWERPC] mpc8349emitx.dts: Setup USB-DR for peripheral mode.
jacmet@sunsite.dk [Fri, 28 Sep 2007 14:21:15 +0000 (16:21 +0200)]
[POWERPC] mpc8349emitx.dts: Setup USB-DR for peripheral mode.

Setup dr_mode for USB-DR to peripheral as the default (host mode) doesn't make
much sense for the mini-AB connector on the ITX board.

Peripheral mode is preferable to OTG as the fsl_usb2_udc.c driver doesn't yet
properly support it.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
17 years ago[POWERPC] Fix mpc834x USB-MPH configuration.
jacmet@sunsite.dk [Fri, 28 Sep 2007 14:21:14 +0000 (16:21 +0200)]
[POWERPC] Fix mpc834x USB-MPH configuration.

mpc834x USB-MPH configuration got broken by commit
6f442560021aecf08658e26ed9a37e6928ef0fa1. The selection bits in SICRL
should be cleared rather than set to configure the USB MUXes for the MPH.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
17 years ago[POWERPC] Fix cpm_uart driver for cpm1 machines
Jochen Friedrich [Mon, 24 Sep 2007 17:15:43 +0000 (19:15 +0200)]
[POWERPC] Fix cpm_uart driver for cpm1 machines

in cpm_uart_cpm1.h, DPRAM_BASE is assigned an address derived from cpmp.
On ARC=ppc, this is a physical address with 1:1 DMA mapping which can't
be used for arithmetric compare operations with virtual addresses
returned by cpm_dpram_addr. This patch changes the assignment to use
cpm_dpram_addr as well, like in cpm_uart_cpm2.h.

Signed-off-by: Jochen Friedrich <jochen@scram.de>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
17 years ago[PPC] Fix cpm_dpram_addr returning phys mem instead of virt mem
Jochen Friedrich [Mon, 24 Sep 2007 17:14:57 +0000 (19:14 +0200)]
[PPC] Fix cpm_dpram_addr returning phys mem instead of virt mem

cpm_dpram_addr returns physical memory of the DP RAM instead of
iomapped virtual memory. As there usually is a 1:1 MMU map of
the IMMR area, this is often not noticed. However, cpm_dpram_phys
assumes this iomapped virtual memory and returns garbage on the
1:1 mapped memory causing CPM1 uart console to fail.

This patch fixes the problem (copied from the powerpc tree).

Signed-off-by: Jochen Friedrich <jochen@scram.de>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
17 years ago[POWERPC] Fix copy'n'paste typo in commproc.c
Jochen Friedrich [Mon, 24 Sep 2007 17:13:46 +0000 (19:13 +0200)]
[POWERPC] Fix copy'n'paste typo in commproc.c

The powerpc version of commproc.c exports cpm_dpram_addr twice
and cpm_dpram_phys not at all due to a typo. This patch fixes this
problem.

CC      arch/powerpc/sysdev/commproc.o
arch/powerpc/sysdev/commproc.c:398: error: redefinition of '__kcrctab_cpm_dpram_addr'
arch/powerpc/sysdev/commproc.c:392: error: previous definition of '__kcrctab_cpm_dpram_addr' was here
arch/powerpc/sysdev/commproc.c:398: error: redefinition of '__kstrtab_cpm_dpram_addr'
arch/powerpc/sysdev/commproc.c:392: error: previous definition of '__kstrtab_cpm_dpram_addr' was here
arch/powerpc/sysdev/commproc.c:398: error: redefinition of '__ksymtab_cpm_dpram_addr'
arch/powerpc/sysdev/commproc.c:392: error: previous definition of '__ksymtab_cpm_dpram_addr' was here
make[1]: *** [arch/powerpc/sysdev/commproc.o] Error 1
make: *** [arch/powerpc/sysdev] Error 2

Signed-off-by: Jochen Friedrich <jochen@scram.de>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
17 years agoe1000: Add device IDs of blade version of the 82571 quad port
Auke Kok [Thu, 30 Aug 2007 18:23:58 +0000 (11:23 -0700)]
e1000: Add device IDs of blade version of the 82571 quad port

This blade-specific board form factor is identical to the 82571EB
board.

Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agosky2: fix transmit state on resume
Stephen Hemminger [Thu, 27 Sep 2007 19:38:12 +0000 (12:38 -0700)]
sky2: fix transmit state on resume

This should fix http://bugzilla.kernel.org/show_bug.cgi?id=8667

After resume, driver has reset the chip so the current state
of transmit checksum offload state machine and DMA state machine
will be undefined.

The fix is to set the state so that first Tx will set MSS and offset
values.

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agosky2: FE+ vlan workaround
Stephen Hemminger [Thu, 27 Sep 2007 19:32:44 +0000 (12:32 -0700)]
sky2: FE+ vlan workaround

The FE+ workaround means the driver can no longer trust the status register
to indicate VLAN tagged frames.  The fix for this is to just disable VLAN
acceleration for that chip version. Tested and works fine.

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agosky2: sky2 FE+ receive status workaround
Stephen Hemminger [Thu, 27 Sep 2007 00:58:47 +0000 (17:58 -0700)]
sky2: sky2 FE+ receive status workaround

The Yukon FE+ chip appears to have a hardware glitch that causes bogus
receive status values to be posted. The data in the packet is good, but
the status value is random garbage.  As a temporary workaround until the
problem is better understood, implement the workaround the vendor driver
used of ignoring the status value on this chip.

Since this means trusting dodgy hardware values; add additional checking
of the receive packet length.

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Fri, 28 Sep 2007 02:47:53 +0000 (19:47 -0700)]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] Fix CONFIG_BUILD_ELF64 kernels with symbols in CKSEG0.
  [MIPS] IP32: Fix initialization of UART base addresses.

17 years agoi915: make vbl interrupts work properly on i965g/gm hw.
Dave Airlie [Fri, 28 Sep 2007 01:46:28 +0000 (11:46 +1000)]
i915: make vbl interrupts work properly on i965g/gm hw.

This code is ported from the DRM git tree and allows the vblank interrupts
to function on the i965 hw. It also requires a change in Mesa's 965 driver
to actually use them.

[ Without this patch, my 965GM drops vblank interrupts  - Jesse ]

Signed-off-by: Dave Airlie <airlied@linux.ie>
Acked-by: Jesse Barnes <jesse.barnes@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[MIPS] Fix CONFIG_BUILD_ELF64 kernels with symbols in CKSEG0.
Ralf Baechle [Tue, 11 Sep 2007 07:50:40 +0000 (08:50 +0100)]
[MIPS] Fix CONFIG_BUILD_ELF64 kernels with symbols in CKSEG0.

The __pa() for those did assume that all symbols have XKPHYS values and
the math fails for any other address range.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] IP32: Fix initialization of UART base addresses.
Ralf Baechle [Wed, 26 Sep 2007 12:01:37 +0000 (13:01 +0100)]
[MIPS] IP32: Fix initialization of UART base addresses.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[NET]: Zero length write() on socket should not simply return 0.
David S. Miller [Thu, 27 Sep 2007 20:52:00 +0000 (13:52 -0700)]
[NET]: Zero length write() on socket should not simply return 0.

This fixes kernel bugzilla #5731

It should generate an empty packet for datagram protocols when the
socket is connected, for one.

The check is doubly-wrong because all that a write() can be is a
sendmsg() call with a NULL msg_control and a single entry iovec.  No
special semantics should be assigned to it, therefore the zero length
check should be removed entirely.

This matches the behavior of BSD and several other systems.

Alan Cox notes that SuSv3 says the behavior of a zero length write on
non-files is "unspecified", but that's kind of useless since BSD has
defined this behavior for a quarter century and BSD is essentially
what application folks code to.

Based upon a patch from Stephen Hemminger.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoRevert "[PATCH] x86-64: fix x86_64-mm-sched-clock-share"
Linus Torvalds [Wed, 26 Sep 2007 22:52:17 +0000 (15:52 -0700)]
Revert "[PATCH] x86-64: fix x86_64-mm-sched-clock-share"

This reverts commit 184c44d2049c4db7ef6ec65794546954da2c6a0e.

As noted by Dave Jones:
   "Linus, please revert the above cset.  It doesn't seem to be
    necessary (it was added to fix a miscompile in 'make allnoconfig'
    which doesn't seem to be repeatable with it reverted) and actively
   breaks the ARM SA1100 framebuffer driver."

Requested-by: Dave Jones <davej@redhat.com>
Cc: Russell King <rmk+lkml@arm.linux.org.uk>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux...
Linus Torvalds [Wed, 26 Sep 2007 22:43:55 +0000 (15:43 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/hpa/linux-2.6-x86setup

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-x86setup:
  [x86 setup] Handle case of improperly terminated E820 chain

17 years agoRevert "x86-64: Disable local APIC timer use on AMD systems with C1E"
Linus Torvalds [Wed, 26 Sep 2007 22:21:33 +0000 (15:21 -0700)]
Revert "x86-64: Disable local APIC timer use on AMD systems with C1E"

This reverts commit e66485d747505e9d960b864fc6c37f8b2afafaf0, since
Rafael Wysocki noticed that the change only works for his in -mm, not in
mainline (and that both "noapictimer" _and_ "apicmaintimer" are broken
on his hardware, but that's apparently not a regression, just a symptom
of the same issue that causes the automatic apic timer disable to not
work).

It turns out that it really doesn't work correctly on x86-64, since
x86-64 doesn't use the generic clock events for timers yet.

Thanks to Rafal for testing, and here's the ugly details on x86-64 as
per Thomas:

  "I just looked into the code and the logic vs.  noapictimer on SMP is
   completely broken.

   On i386 the noapictimer option not only disables the local APIC
   timer, it also registers the CPUs for broadcasting via IPI on SMP
   systems.

   The x86-64 code uses the broadcast only when the local apic timer is
   active, i.e.  "noapictimer" is not on the command line.  This defeats
   the whole purpose of "noapictimer".  It should be there to make boxen
   work, where the local APIC timer actually has a hardware problem,
   e.g.  the nx6325.

   The current implementation of x86_64 only fixes the ACPI c-states
   related problem where the APIC timer stops in C3(2), nothing else.

   On nx6325 and other AMD X2 equipped systems which have the C1E
   enabled we run into the following:

   PIT keeps jiffies (and the system) running, but the local APIC timer
   interrupts can get out of sync due to this C1E effect.

   I don't think this is a critical problem, but it is wrong
   nevertheless.

   I think it's safe to revert the C1E patch and postpone the fix to the
   clock events conversion."

On further reflection, Thomas noted:

   "It's even worse than I thought on the first check:

    "noapictimer" on the command line of an SMP box prevents _ONLY_ the
    boot CPU apic timer from being used.  But the secondary CPU is still
    unconditionally setting up the APIC timer and uses the non
    calibrated variable calibration_result, which is of course 0, to
    setup the APIC timer.  Wreckage guaranteed."

so we'll just have to wait for the x86 merge to hopefully fix this up
for x86-64.

Tested-and-requested-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[x86 setup] Handle case of improperly terminated E820 chain
H. Peter Anvin [Wed, 26 Sep 2007 21:11:43 +0000 (14:11 -0700)]
[x86 setup] Handle case of improperly terminated E820 chain

At least one system (a Geode system with a Digital Logic BIOS) has
been found which suddenly stops reporting the SMAP signature when
reading the E820 memory chain.  We can't know what, exactly, broke in
the BIOS, so if we detect this situation, declare the E820 data
unusable and fall back to E801.

Also, revert to original behavior of always probing all memory
methods; that way all the memory information is available to the
kernel.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Cc: Jordan Crouse <jordan.crouse@amd.com>
Cc: Joerg Pommnitz <pommnitz@yahoo.com>
17 years agoxen: execve's error paths don't pin the mm before unpinning
Jeremy Fitzhardinge [Tue, 25 Sep 2007 18:50:00 +0000 (11:50 -0700)]
xen: execve's error paths don't pin the mm before unpinning

execve's error paths don't activate (and therefore pin) the mm before
calling exit_mmap to free it up, so don't try to unpin unless it is
actually pinned.  This prevents a BUG_ON from triggering.

Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Cc: Christian Ostheimer <osth@freesurf.ch>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agox86-64: Disable local APIC timer use on AMD systems with C1E
Thomas Gleixner [Tue, 25 Sep 2007 19:37:01 +0000 (21:37 +0200)]
x86-64: Disable local APIC timer use on AMD systems with C1E

commit 3556ddfa9284a86a59a9b78fe5894430f6ab4eef titled

 [PATCH] x86-64: Disable local APIC timer use on AMD systems with C1E

solves a problem with AMD dual core laptops e.g. HP nx6325 (Turion 64
X2) with C1E enabled:

When both cores go into idle at the same time, then the system switches
into C1E state, which is basically the same as C3. This stops the local
apic timer.

This was debugged right after the dyntick merge on i386 and despite the
patch title it fixes only the 32 bit path.

x86_64 is still missing this fix. It seems that mainline is not really
affected by this issue, as the PIT is running and keeps jiffies
incrementing, but that's just waiting for trouble.

-mm suffers from this problem due to the x86_64 high resolution timer
patches.

This is a quick and dirty port of the i386 code to x86_64.

I spent quite a time with Rafael to debug the -mm / hrt wreckage until
someone pointed us to this. I really had forgotten that we debugged this
half a year ago already.

Sigh, is it just me or is there something yelling arch/x86 into my ear?

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoSilent drivers/char/hpet.c build warnings on i386
S.Çağlar Onur [Wed, 26 Sep 2007 09:15:33 +0000 (12:15 +0300)]
Silent drivers/char/hpet.c build warnings on i386

Following patch silents;

...
drivers/char/hpet.c:72: warning: 'clocksource_hpet' defined but not used
drivers/char/hpet.c:81: warning: 'hpet_clocksource' defined but not used
...

build warnings on i386, they appeared after commit 3b2b64fd311c92f2137eb7cee7025794cd854057

Signed-off-by: S.Çağlar Onur <caglar@pardus.org.tr>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
--
 drivers/char/hpet.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

17 years agoNLM: Fix a circular lock dependency in lockd
Trond Myklebust [Tue, 25 Sep 2007 19:55:03 +0000 (15:55 -0400)]
NLM: Fix a circular lock dependency in lockd

The problem is that the garbage collector for the 'host' structures
nlm_gc_hosts(), holds nlm_host_mutex while calling down to
nlmsvc_mark_resources, which, eventually takes the file->f_mutex.

We cannot therefore call nlmsvc_lookup_host() from within
nlmsvc_create_block, since the caller will already hold file->f_mutex, so
the attempt to grab nlm_host_mutex may deadlock.

Fix the problem by calling nlmsvc_lookup_host() outside the file->f_mutex.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agolguest example launcher truncates block device file to 0 length on problems
Chris Malley [Wed, 26 Sep 2007 04:19:18 +0000 (14:19 +1000)]
lguest example launcher truncates block device file to 0 length on problems

The function should also use ftruncate64() rather than ftruncate() to prevent
files over 4GB (not uncommon for a root filesystem) being zeroed.

Signed-off-by: Chris Malley <mail@chrismalley.co.uk>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agofix sctp_del_bind_addr() last argument type
Al Viro [Wed, 26 Sep 2007 00:54:32 +0000 (01:54 +0100)]
fix sctp_del_bind_addr() last argument type

It gets pointer to fastcall function, expects a pointer to normal
one and calls the sucker.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agohibernation doesn't even build on frv - tons of helpers are missing
Al Viro [Wed, 26 Sep 2007 00:54:12 +0000 (01:54 +0100)]
hibernation doesn't even build on frv - tons of helpers are missing

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Acked-By: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agomv643xx_eth: duplicate methods in initializer
Al Viro [Wed, 26 Sep 2007 00:53:52 +0000 (01:53 +0100)]
mv643xx_eth: duplicate methods in initializer

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agousbvision-video: buggered method tables
Al Viro [Wed, 26 Sep 2007 00:53:42 +0000 (01:53 +0100)]
usbvision-video: buggered method tables

duplicated .mmap in one, .vidioc_s_audio misspelled as .vidioc_g_audio
in other

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Acked-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoAdd explicit zeroing to "envp" array in device 'show' method
Linus Torvalds [Wed, 26 Sep 2007 16:16:21 +0000 (09:16 -0700)]
Add explicit zeroing to "envp" array in device 'show' method

As Stephen Hemminger says, this is a "belt and suspenders" patch that
zeroes the envp array at allocation time, even though all the users
should NULL-terminate it anyway (and we've hopefully fixed everybody
that doesn't do that).

And we'll apparently clean the whole envp thing up for 2.6.24 anyway.

But let's just be robust, and do both this *and* make sure that all
users are doing the right thing.

Acked-by: Stephen Hemminger <shemminger@linux-foundation.org>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6
Linus Torvalds [Wed, 26 Sep 2007 16:08:04 +0000 (09:08 -0700)]
Merge /linux/kernel/git/jejb/scsi-rc-fixes-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6:
  [SCSI] esp: fix instance numbering.

17 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Linus Torvalds [Wed, 26 Sep 2007 16:00:04 +0000 (09:00 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/paulus/powerpc

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
  [POWERPC] spufs: fix mismerge, making context signal{1,2} files readable again

17 years agoMerge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 26 Sep 2007 15:59:41 +0000 (08:59 -0700)]
Merge branch 'master' of /linux/kernel/git/davem/net-2.6

* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [PPP_MPPE]: Don't put InterimKey on the stack
  SCTP : Add paramters validity check for ASCONF chunk
  SCTP: Discard OOTB packetes with bundled INIT early.
  SCTP: Clean up OOTB handling and fix infinite loop processing
  SCTP: Explicitely discard OOTB chunks
  SCTP: Send ABORT chunk with correct tag in response to INIT ACK
  SCTP: Validate buffer room when processing sequential chunks
  [PATCH] mac80211: fix initialisation when built-in
  [PATCH] net/mac80211/wme.c: fix sparse warning
  [PATCH] cfg80211: fix initialisation if built-in
  [PATCH] net/wireless/sysfs.c: Shut up build warning

17 years agoMerge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Wed, 26 Sep 2007 15:58:45 +0000 (08:58 -0700)]
Merge branch 'master' of /linux/kernel/git/davem/sparc-2.6

* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SUNSAB]: Fix broken SYSRQ.

17 years agoMerge branch 'trivial' of git://git.kernel.dk/data/git/linux-2.6-block
Linus Torvalds [Wed, 26 Sep 2007 15:58:13 +0000 (08:58 -0700)]
Merge branch 'trivial' of git://git.kernel.dk/data/git/linux-2.6-block

* 'trivial' of git://git.kernel.dk/data/git/linux-2.6-block:
  cdrom_open() forgets to unlock on -EROFS failure exits

17 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Wed, 26 Sep 2007 15:57:27 +0000 (08:57 -0700)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  pata_sis: add missing UDMA5 timing value in sis_66_set_dmamode()
  sata_sil24: fix IRQ clearing race when PCIX_IRQ_WOC is used

17 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux...
Linus Torvalds [Wed, 26 Sep 2007 15:56:28 +0000 (08:56 -0700)]
Merge branch 'release' of git://git./linux/kernel/git/lenb/linux-acpi-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6:
  ACPI: hpet: ACPI Error (utglobal-0126): Unknown exception code: 0xFFFFFFF0
  ACPI: CONFIG_ACPI_SLEEP=n power off regression in 2.6.23-rc8 (NOT in rc7)
  ACPI: suspend: build-fix for CONFIG_SUSPEND=n and CONFIG_HIBERNATION=y

17 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Wed, 26 Sep 2007 15:55:54 +0000 (08:55 -0700)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/netdev-2.6

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6:
  [PATCH] WE : Add missing auth compat-ioctl
  [PATCH] softmac: Fix inability to associate with WEP networks

17 years ago[POWERPC] spufs: fix mismerge, making context signal{1,2} files readable again
Jeremy Kerr [Wed, 26 Sep 2007 00:53:45 +0000 (10:53 +1000)]
[POWERPC] spufs: fix mismerge, making context signal{1,2} files readable again

The commit 8b6f50ef1d5cc86b278eb42bc91630fad455fb10 seems to have
been affected by a mismerge of a duplicate patch
(d054b36ffd302ec65aabec16a0c60ddd9e6b5a62) - both the
spufs_dir_contents and spufs_dir_nosched_contents have been given
write-only signal notification files.

This change reverts the spufs_dir_contents array to use the
readable signal notification file implementation.

Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Acked-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
17 years agocdrom_open() forgets to unlock on -EROFS failure exits
Al Viro [Wed, 26 Sep 2007 06:20:44 +0000 (08:20 +0200)]
cdrom_open() forgets to unlock on -EROFS failure exits

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years ago[PPP_MPPE]: Don't put InterimKey on the stack
Michal Schmidt [Wed, 26 Sep 2007 06:05:39 +0000 (23:05 -0700)]
[PPP_MPPE]: Don't put InterimKey on the stack

Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoSCTP : Add paramters validity check for ASCONF chunk
Wei Yongjun [Wed, 19 Sep 2007 09:19:52 +0000 (17:19 +0800)]
SCTP : Add paramters validity check for ASCONF chunk

If ADDIP is enabled, when an ASCONF chunk is received with ASCONF
paramter length set to zero, this will cause infinite loop.
By the way, if an malformed ASCONF chunk is received, will cause
processing to access memory without verifying.

This is because of not check the validity of parameters in ASCONF chunk.
This patch fixed this.

Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
17 years agoSCTP: Discard OOTB packetes with bundled INIT early.
Vlad Yasevich [Mon, 17 Sep 2007 19:14:28 +0000 (15:14 -0400)]
SCTP: Discard OOTB packetes with bundled INIT early.

RFC 4460 and future RFC 4960 (2960-bis) specify that packets
with bundled INIT chunks need to be dropped.  We currenlty do
that only after processing any leading chunks.  For OOTB chunks,
since we already walk the entire packet, we should discard packets
with bundled INITs.

There are other chunks chunks that MUST NOT be bundled, but the spec
is silent on theire treatment.  Thus, we'll leave their teatment
alone for the moment.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Acked-by: Wei Yongjun <yjwei@cn.fujitsu.com>
17 years agoSCTP: Clean up OOTB handling and fix infinite loop processing
Vlad Yasevich [Fri, 7 Sep 2007 20:30:54 +0000 (16:30 -0400)]
SCTP: Clean up OOTB handling and fix infinite loop processing

While processing OOTB chunks as well as chunks with an invalid
length of 0, it was possible to SCTP to get wedged inside an
infinite loop because we didn't catch the condition correctly,
or didn't mark the packet for discard correctly.
This work is based on original findings and work by
Wei Yongjun <yjwei@cn.fujitsu.com>

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
17 years agoSCTP: Explicitely discard OOTB chunks
Vlad Yasevich [Fri, 7 Sep 2007 15:47:45 +0000 (11:47 -0400)]
SCTP: Explicitely discard OOTB chunks

Explicitely discard OOTB chunks, whether the result is a
SHUTDOWN COMPLETE or an ABORT.  We need to discard the OOTB
SHUTDOWN ACK to prevent bombing attackes since responsed
MUST NOT be bundled.  We also explicietely discard in the
ABORT case since that function is widely used internally.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
17 years agoSCTP: Send ABORT chunk with correct tag in response to INIT ACK
Wei Yongjun [Fri, 31 Aug 2007 02:03:58 +0000 (10:03 +0800)]
SCTP: Send ABORT chunk with correct tag in response to INIT ACK

When SCTP client received an INIT ACK chunk with missing mandatory
parameter such as "cookie parameter", it will send back a ABORT
with T-bit not set and verification tag is set to 0.
This is because before we accept this INIT ACK chunk, we do not know
the peer's tag.  This patch change to reflect vtag when responding to
INIT ACK with missing mandatory parameter.

Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
17 years agoSCTP: Validate buffer room when processing sequential chunks
Vlad Yasevich [Wed, 5 Sep 2007 19:53:58 +0000 (15:53 -0400)]
SCTP: Validate buffer room when processing sequential chunks

When we process bundled chunks, we need to make sure that
the skb has the buffer for each header since we assume it's
always there.  Some malicious node can send us something like
DATA + 2 bytes and we'll try to walk off the end refrencing
potentially uninitialized memory.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
17 years ago[PATCH] mac80211: fix initialisation when built-in
Johannes Berg [Tue, 11 Sep 2007 10:50:32 +0000 (12:50 +0200)]
[PATCH] mac80211: fix initialisation when built-in

When mac80211 is built into the kernel it needs to init earlier
so that device registrations are run after it has initialised.
The same applies to rate control algorithms.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
17 years ago[PATCH] net/mac80211/wme.c: fix sparse warning
Johannes Berg [Mon, 10 Sep 2007 11:55:08 +0000 (13:55 +0200)]
[PATCH] net/mac80211/wme.c: fix sparse warning

wme.c triggers a sparse warning; it wasn't noticed before because until
recently ARRAY_SIZE triggered a sparse error.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
17 years ago[PATCH] cfg80211: fix initialisation if built-in
Johannes Berg [Mon, 10 Sep 2007 11:44:45 +0000 (13:44 +0200)]
[PATCH] cfg80211: fix initialisation if built-in

When cfg80211 is built into the kernel it needs to init earlier
so that device registrations are run after it has initialised.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
17 years ago[PATCH] net/wireless/sysfs.c: Shut up build warning
Satyam Sharma [Sun, 2 Sep 2007 20:11:34 +0000 (01:41 +0530)]
[PATCH] net/wireless/sysfs.c: Shut up build warning

net/wireless/sysfs.c:108: warning: ‘wiphy_uevent’ defined but not used

when CONFIG_HOTPLUG=n is because the only usage site of this function
is #ifdef'ed as such, so let's #ifdef the definition also.

Signed-off-by: Satyam Sharma <satyam@infradead.org>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
17 years ago[SUNSAB]: Fix broken SYSRQ.
Thomas Bogendoerfer [Wed, 26 Sep 2007 05:49:37 +0000 (22:49 -0700)]
[SUNSAB]: Fix broken SYSRQ.

Include SYSRQ support for SUNSAB if SUNSAB_CONSOLE is selected

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[SCSI] esp: fix instance numbering.
David Miller [Sat, 25 Aug 2007 05:25:58 +0000 (22:25 -0700)]
[SCSI] esp: fix instance numbering.

Because the ->unique_id is set too late, the ESP scsi host
instance numbers in the kernel log during probing are
wrong.

Bug reported by Meelis Roos.

Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years agopata_sis: add missing UDMA5 timing value in sis_66_set_dmamode()
Tejun Heo [Fri, 21 Sep 2007 07:29:05 +0000 (16:29 +0900)]
pata_sis: add missing UDMA5 timing value in sis_66_set_dmamode()

sis_66_set_dmamode() also handles early UDMA100 (SIS630 ET) but is
missing udma timing value for UDMA100.  According to sis5513, this
should be 0x8000.  This caused UDMA100 device to fail on pata_sis till
it downgrades to UDMA66 while it works fine on sis5513 at UDMA100.

Reported by Adam Blech.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Cc: Adam Blech <desaster.area@addcom.de>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agosata_sil24: fix IRQ clearing race when PCIX_IRQ_WOC is used
Tejun Heo [Sun, 23 Sep 2007 03:37:05 +0000 (12:37 +0900)]
sata_sil24: fix IRQ clearing race when PCIX_IRQ_WOC is used

When PCIX_IRQ_WOC is used, sil24 has an inherent race condition
between clearing IRQ pending and reading IRQ status.  If IRQ pending
is cleared after reading IRQ status, there's possibility of lost IRQ.
If IRQ pending is cleared before reading IRQ status, spurious IRQs
will occur.

sata_sil24 till now cleared IRQ pending after reading IRQ status thus
losing IRQs on machines where PCIX_IRQ_WOC was used.  Reverse the
order and ignore spurious IRQs if PCIX_IRQ_WOC.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoACPI: hpet: ACPI Error (utglobal-0126): Unknown exception code: 0xFFFFFFF0
Zhao Yakui [Fri, 21 Sep 2007 01:23:13 +0000 (21:23 -0400)]
ACPI: hpet: ACPI Error (utglobal-0126): Unknown exception code: 0xFFFFFFF0

If hpet has been initialized before registering hpet driver, the callback
function of hpet_resources will return the status code of -EBUSY, which is
not defined in the ACPI exception table.  So when ACPI checks the status
code of callback function, it will report the unknown exception code.

So the status code in ACPI is used instead of the generic error code in the
ACPI callback function of hpet_resources.
For example: -EBUSY is replaced by AE_ALREADY_EXISTS
     -EINVAL is replaced by AE_NO_MEMORY

http://bugzilla.kernel.org/show_bug.cgi?id=8630

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoACPI: CONFIG_ACPI_SLEEP=n power off regression in 2.6.23-rc8 (NOT in rc7)
Alexey Starikovskiy [Tue, 25 Sep 2007 14:45:15 +0000 (18:45 +0400)]
ACPI: CONFIG_ACPI_SLEEP=n power off regression in 2.6.23-rc8 (NOT in rc7)

Signed-off-by: Alexey Starikovskiy <astarikovskiy@suse.de>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoACPI: suspend: build-fix for CONFIG_SUSPEND=n and CONFIG_HIBERNATION=y
Alexey Starikovskiy [Mon, 24 Sep 2007 12:33:21 +0000 (14:33 +0200)]
ACPI: suspend: build-fix for CONFIG_SUSPEND=n and CONFIG_HIBERNATION=y

This fixes compilation with CONFIG_SUSPEND unset and CONFIG_HIBERNATION set
(raf. http://marc.info/?l=linux-acpi&m=119055289723895&w=4).

Signed-off-by: Alexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoMerge branch 'fixes-jgarzik' of git://git.kernel.org/pub/scm/linux/kernel/git/linvill...
Jeff Garzik [Tue, 25 Sep 2007 19:47:12 +0000 (15:47 -0400)]
Merge branch 'fixes-jgarzik' of git://git./linux/kernel/git/linville/wireless-2.6 into upstream-fixes

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6
Linus Torvalds [Tue, 25 Sep 2007 18:19:03 +0000 (11:19 -0700)]
Merge /linux/kernel/git/jejb/scsi-rc-fixes-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6:
  [SCSI] scsi_transport_spi: fix domain validation failure from incorrect width setting

17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Tue, 25 Sep 2007 18:18:10 +0000 (11:18 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: appletouch - fix idle reset logic

17 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Linus Torvalds [Tue, 25 Sep 2007 15:54:35 +0000 (08:54 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/paulus/powerpc

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
  [POWERPC] Ensure FULL_REGS on exec

17 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Tue, 25 Sep 2007 15:52:26 +0000 (08:52 -0700)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/netdev-2.6

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6:
  Revert "drivers/net/pcmcia/3c589_cs: fix port configuration switcheroo"
  sky2: be more selective about FIFO watchdog
  sky2: FE+ Phy initialization
  r8169: workaround against ignored TxPoll writes (8168)
  r8169: correct phy parameters for the 8110SC

17 years agofix modules oopsing in lguest guests
Rusty Russell [Tue, 25 Sep 2007 04:24:44 +0000 (21:24 -0700)]
fix modules oopsing in lguest guests

The assembly templates for lguest guest patching are in the .init.text
section.  This means that modules get patched with "cc cc cc cc" or similar
junk.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agotypo fix Kernel config option
Danny ter Haar [Tue, 25 Sep 2007 04:24:43 +0000 (21:24 -0700)]
typo fix Kernel config option

Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agomissing null termination in one wire uevent
Stephen Hemminger [Tue, 25 Sep 2007 04:24:42 +0000 (21:24 -0700)]
missing null termination in one wire uevent

Need to null terminate environment.  Found by inspection while looking for
similar problems to platform uevent bug

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@suse.de>
Acked-by: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agofix "mspec: handle shrinking virtual memory areas"
Cliff Wickman [Tue, 25 Sep 2007 04:24:41 +0000 (21:24 -0700)]
fix "mspec: handle shrinking virtual memory areas"

The vma_data structure may be shared by vma's from multiple tasks, with no
way of knowing which areas are shared or not shared, so release/clear pages
only when the refcount (of vma's) goes to zero.

Signed-off-by: Cliff Wickman <cpw@sgi.com>
Cc: Jes Sorensen <jes@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoufs: fix sun state
Evgeniy Dushistov [Tue, 25 Sep 2007 04:24:39 +0000 (21:24 -0700)]
ufs: fix sun state

Different types of ufs hold state in different places, to hide complexity
of this, there is ufs_get_fs_state, it returns state according to
"UFS_SB(sb)->s_flags", but during mount ufs_get_fs_state is called, before
setting s_flags, this cause message for ufs types like sun ufs: "fs need
fsck", and remount in readonly state.

Signed-off-by: Evgeniy Dushistov <dushistov@mail.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[POWERPC] Ensure FULL_REGS on exec
Roland McGrath [Mon, 24 Sep 2007 23:52:44 +0000 (16:52 -0700)]
[POWERPC] Ensure FULL_REGS on exec

When PTRACE_O_TRACEEXEC is used, a ptrace call to fetch the registers at
the PTRACE_EVENT_EXEC stop (PTRACE_PEEKUSR) will oops in CHECK_FULL_REGS.
With recent versions, "gdb --args /bin/sh -c 'exec /bin/true'" and "run" at
the (gdb) prompt is sufficient to produce this.  I also have written an
isolated test case, see https://bugzilla.redhat.com/show_bug.cgi?id=301791#c15.

This change fixes the problem by clearing the low bit of pt_regs.trap in
start_thread so that FULL_REGS is true again.  This is correct since all of
the GPRs that "full" refers to are cleared in start_thread.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
17 years agoMerge branch 'r8169-for-jeff-20070919' of git://electric-eye.fr.zoreil.com/home/romie...
Jeff Garzik [Tue, 25 Sep 2007 04:14:03 +0000 (00:14 -0400)]
Merge branch 'r8169-for-jeff-20070919' of git://electric-eye.fr.zoreil.com/home/romieu/linux-2.6 into tmp

17 years agoRevert "drivers/net/pcmcia/3c589_cs: fix port configuration switcheroo"
Jeff Garzik [Tue, 25 Sep 2007 04:11:34 +0000 (00:11 -0400)]
Revert "drivers/net/pcmcia/3c589_cs: fix port configuration switcheroo"

This reverts commit fadacb1b80e35e0b36a90d43e21ef91eec4b889b.

The change being reverted made the driver consistent with
include/linux/netdevice.h, but then inconsistent with the other PCMCIA
ethernet drivers.

Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoInput: appletouch - fix idle reset logic
Thomas Rohwer [Tue, 25 Sep 2007 04:06:25 +0000 (00:06 -0400)]
Input: appletouch - fix idle reset logic

Idle count should only be incremented when touchpad button
is not pressed, otherwise reset may happen at a wrong time
and touchpad will never report button release event.

Signed-off-by: Thomas Rohwer <trohwer@tng.de>
Acked-by: Matthew Garrett <mjg59@srcf.ucam.org>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agosky2: be more selective about FIFO watchdog
Stephen Hemminger [Thu, 20 Sep 2007 20:03:49 +0000 (13:03 -0700)]
sky2: be more selective about FIFO watchdog

Be more selective about when to enable the ram buffer watchdog code.
It is unnecessary on XL A3 or later revs, and with Yukon FE
the buffer is so small (4K) that the watchdog detects false positives.

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agosky2: FE+ Phy initialization
Stephen Hemminger [Tue, 25 Sep 2007 02:34:51 +0000 (19:34 -0700)]
sky2: FE+ Phy initialization

One more snippet of PHY initialization required for FE+ chips.
Discovered in latest sk98lin 10.21.1.3 driver.

Please apply to 2.6.23.

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoLinux 2.6.23-rc8
Linus Torvalds [Tue, 25 Sep 2007 00:33:10 +0000 (17:33 -0700)]
Linux 2.6.23-rc8

Getting there...

17 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Tue, 25 Sep 2007 00:32:16 +0000 (17:32 -0700)]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] SMTC: Make ack_bad_irq() safe with no IM backstop.

17 years agoMerge branch 'async-tx-fixes-for-linus' of git://lost.foo-projects.org/~dwillia2...
Linus Torvalds [Tue, 25 Sep 2007 00:25:30 +0000 (17:25 -0700)]
Merge branch 'async-tx-fixes-for-linus' of git://lost.foo-projects.org/~dwillia2/git/iop

* 'async-tx-fixes-for-linus' of git://lost.foo-projects.org/~dwillia2/git/iop:
  raid5: fix 2 bugs in ops_complete_biofill
  async_tx: fix dma_wait_for_async_tx
  async_tx: usage documentation and developer notes (v2)

17 years agoraid5: fix 2 bugs in ops_complete_biofill
Dan Williams [Mon, 24 Sep 2007 17:06:13 +0000 (10:06 -0700)]
raid5: fix 2 bugs in ops_complete_biofill

1/ ops_complete_biofill tried to avoid calling handle_stripe since all the
state necessary to return read completions is available.  However the
process of determining whether more read requests are pending requires
locking the stripe (to block add_stripe_bio from updating dev->toead).
ops_complete_biofill can run in tasklet context, so rather than upgrading
all the stripe locks from spin_lock to spin_lock_bh this patch just
unconditionally reschedules handle_stripe after completing the read
request.

2/ ops_complete_biofill needlessly qualified processing R5_Wantfill with
dev->toread.  The result being that the 'biofill' pending bit is cleared
before handling the pending read-completions on dev->read.  R5_Wantfill can
be unconditionally handled because the 'biofill' pending bit prevents new
R5_Wantfill requests from being seen by ops_run_biofill and
ops_complete_biofill.

Found-by: Yuri Tikhonov <yur@emcraft.com>
[neilb@suse.de: simpler fix for bug 1 than moving code]
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
17 years agoasync_tx: fix dma_wait_for_async_tx
Dan Williams [Fri, 21 Sep 2007 20:27:04 +0000 (13:27 -0700)]
async_tx: fix dma_wait_for_async_tx

Fix dma_wait_for_async_tx to not loop forever in the case where a
dependency chain is longer than two entries.  This condition will not
happen with current in-kernel drivers, but fix it for future drivers.

Found-by: Saeed Bishara <saeed.bishara@gmail.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
17 years agoasync_tx: usage documentation and developer notes (v2)
Dan Williams [Thu, 20 Sep 2007 22:49:08 +0000 (15:49 -0700)]
async_tx: usage documentation and developer notes (v2)

Changes in v2:
* cleanups from Randy and Shannon

Reviewed-by: Randy Dunlap <randy.dunlap@oracle.com>
Reviewed-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
17 years ago[MIPS] SMTC: Make ack_bad_irq() safe with no IM backstop.
Ralf Baechle [Fri, 21 Sep 2007 16:13:55 +0000 (17:13 +0100)]
[MIPS] SMTC: Make ack_bad_irq() safe with no IM backstop.

Issue reported and original patch by Kevin Kissel, cleaner (imho)
implementation by me.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years agoConvert snd-page-alloc proc file to use seq_file
Takashi Iwai [Mon, 17 Sep 2007 19:55:10 +0000 (21:55 +0200)]
Convert snd-page-alloc proc file to use seq_file

Use seq_file for the proc file read/write of snd-page-alloc module.
This automatically fixes bugs in the old proc code.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland...
Linus Torvalds [Mon, 24 Sep 2007 05:39:57 +0000 (22:39 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/roland/infiniband

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB/mlx4: Fix data corruption triggered by wrong headroom marking order

17 years agoMerge git://git.infradead.org/battery-2.6
Linus Torvalds [Sun, 23 Sep 2007 20:04:01 +0000 (13:04 -0700)]
Merge git://git.infradead.org/battery-2.6

* git://git.infradead.org/battery-2.6:
  [PATCH] missing null termination in power supply uevent

17 years agoIB/mlx4: Fix data corruption triggered by wrong headroom marking order
Jack Morgenstein [Wed, 19 Sep 2007 16:52:25 +0000 (09:52 -0700)]
IB/mlx4: Fix data corruption triggered by wrong headroom marking order

This is an addendum to commit 0e6e7416 ("IB/mlx4: Handle new FW
requirement for send request prefetching").  We also need to handle
prefetch marking properly for S/G segments, or else the HCA may end up
processing S/G segments that are not fully written and end up sending
the wrong data.  This can actually cause data corruption in practice,
especially on systems with relatively slow CPUs (where the HCA is more
likely to prefetch while the CPU is in the middle of writing a work
request into memory).

We write S/G segments in reverse order into the WQE, in order to
guarantee that the first dword of all cachelines containing S/G
segments is written last (overwriting the headroom invalidation
pattern).  The entire cacheline will thus contain valid data when the
invalidation pattern is overwritten.

Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>