linux-2.6
18 years ago[PATCH] md: Fix issues with referencing rdev in md/raid1
NeilBrown [Fri, 1 Sep 2006 04:27:36 +0000 (21:27 -0700)] 
[PATCH] md: Fix issues with referencing rdev in md/raid1

We need to be careful when referencing mirrors[i].rdev.  It can disappear
under us at various times.

So:
  fix a couple of problem places.
  comment a couple of non-problem places
  move an 'atomic_add' which deferences rdev down a little
    way to some where where it is sure to not be NULL.

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] ZVC: Scale thresholds depending on the size of the system
Christoph Lameter [Fri, 1 Sep 2006 04:27:35 +0000 (21:27 -0700)] 
[PATCH] ZVC: Scale thresholds depending on the size of the system

The ZVC counter update threshold is currently set to a fixed value of 32.
This patch sets up the threshold depending on the number of processors and
the sizes of the zones in the system.

With the current threshold of 32, I was able to observe slight contention
when more than 130-140 processors concurrently updated the counters.  The
contention vanished when I either increased the threshold to 64 or used
Andrew's idea of overstepping the interval (see ZVC overstep patch).

However, we saw contention again at 220-230 processors.  So we need higher
values for larger systems.

But the current default is already a bit of an overkill for smaller
systems.  Some systems have tiny zones where precision matters.  For
example i386 and x86_64 have 16M DMA zones and either 900M ZONE_NORMAL or
ZONE_DMA32.  These are even present on SMP and NUMA systems.

The patch here sets up a threshold based on the number of processors in the
system and the size of the zone that these counters are used for.  The
threshold should grow logarithmically, so we use fls() as an easy
approximation.

Results of tests on a system with 1024 processors (4TB RAM)

The following output is from a test allocating 1GB of memory concurrently
on each processor (Forking the process.  So contention on mmap_sem and the
pte locks is not a factor):

                       X                   MIN
TYPE:               CPUS       WALL       WALL        SYS     USER     TOTCPU
fork                   1      0.552      0.552      0.540    0.012      0.552
fork                   4      0.552      0.548      2.164    0.036      2.200
fork                  16      0.564      0.548      8.812    0.164      8.976
fork                 128      0.580      0.572     72.204    1.208     73.412
fork                 256      1.300      0.660    310.400    2.160    312.560
fork                 512      3.512      0.696   1526.836    4.816   1531.652
fork                1020     20.024      0.700  17243.176    6.688  17249.863

So a threshold of 32 is fine up to 128 processors. At 256 processors contention
becomes a factor.

Overstepping the counter (earlier patch) improves the numbers a bit:

fork                   4      0.552      0.548      2.164    0.040      2.204
fork                  16      0.552      0.548      8.640    0.148      8.788
fork                 128      0.556      0.548     69.676    0.956     70.632
fork                 256      0.876      0.636    212.468    2.108    214.576
fork                 512      2.276      0.672    997.324    4.260   1001.584
fork                1020     13.564      0.680  11586.436    6.088  11592.523

Still contention at 512 and 1020. Contention at 1020 is down by a third.
256 still has a slight bit of contention.

After this patch the counter threshold will be set to 125 which reduces
contention significantly:

fork                 128      0.560      0.548     69.776    0.932     70.708
fork                 256      0.636      0.556    143.460    2.036    145.496
fork                 512      0.640      0.548    284.244    4.236    288.480
fork                1020      1.500      0.588   1326.152    8.892   1335.044

[akpm@osdl.org: !SMP build fix]
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ZVC: Overstep counters
Christoph Lameter [Fri, 1 Sep 2006 04:27:34 +0000 (21:27 -0700)] 
[PATCH] ZVC: Overstep counters

Increments and decrements are usually grouped rather than mixed.  We can
optimize the inc and dec functions for that case.

Increment and decrement the counters by 50% more than the threshold in
those cases and set the differential accordingly.  This decreases the need
to update the atomic counters.

The idea came originally from Andrew Morton.  The overstepping alone was
sufficient to address the contention issue found when updating the global
and the per zone counters from 160 processors.

Also remove some code in dec_zone_page_state.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland...
Linus Torvalds [Fri, 1 Sep 2006 04:27:29 +0000 (21:27 -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/mthca: Use IRQ safe locks to protect allocation bitmaps

18 years agoIB/mthca: Use IRQ safe locks to protect allocation bitmaps
Roland Dreier [Thu, 31 Aug 2006 23:43:06 +0000 (16:43 -0700)] 
IB/mthca: Use IRQ safe locks to protect allocation bitmaps

It is supposed to be OK to call mthca_create_ah() and mthca_destroy_ah()
from any context.  However, for mem-full HCAs, these functions use the
mthca_alloc() and mthca_free() bitmap helpers, and those helpers use
non-IRQ-safe spin_lock() internally.  Lockdep correctly warns that
this could lead to a deadlock.  Fix this by changing mthca_alloc() and
mthca_free() to use spin_lock_irqsave().

Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years ago[IPV4]: Fix SNMPv2 "ipFragFails" counter error
Wei Dong [Thu, 31 Aug 2006 22:24:48 +0000 (15:24 -0700)] 
[IPV4]: Fix SNMPv2 "ipFragFails" counter error

  When I tested Linux kernel 2.6.17.7 about statistics
"ipFragFails",found that this counter couldn't increase correctly. The
criteria is RFC2011:
RFC2011
  ipFragFails OBJECT-TYPE
    SYNTAX      Counter32
    MAX-ACCESS  read-only
    STATUS      current
    DESCRIPTION
            "The number of IP datagrams that have been discarded because
            they needed to be fragmented at this entity but could not
            be, e.g., because their Don't Fragment flag was set."
    ::= { ip 18 }

When I send big IP packet to a router with DF bit set to 1 which need to
be fragmented, and router just sends an ICMP error message
ICMP_FRAG_NEEDED but no increments for this counter(in the function
ip_fragment).

Signed-off-by: Wei Dong <weid@nanjing-fnst.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: Rate limiting for socket allocation failure messages.
Akinobu Mita [Thu, 31 Aug 2006 22:21:50 +0000 (15:21 -0700)] 
[NET]: Rate limiting for socket allocation failure messages.

This patch limits the warning messages when socket allocation failures
happen. It happens under memory pressure.

Signed-off-by: Akinobu Mita <mita@miraclelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPV6]: Fix kernel OOPs when setting sticky socket options.
YOSHIFUJI Hideaki [Thu, 31 Aug 2006 21:52:17 +0000 (14:52 -0700)] 
[IPV6]: Fix kernel OOPs when setting sticky socket options.

Bug noticed by Remi Denis-Courmont <rdenis@simphalempin.com>.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years agoMerge branch 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6
Linus Torvalds [Thu, 31 Aug 2006 21:42:07 +0000 (14:42 -0700)] 
Merge branch 'for-linus' of git://git390.osdl.marist.edu/linux-2.6

* 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6:
  [S390] cio: unsolicited interrupts during sense pgid.
  [S390] cio: no path after machine check.
  [S390] cio: kernel stack overflow.
  [S390] dasd: fix device shutdown process.
  [S390] broken copy_in_user function.

18 years ago[PATCH] i386: rwlock.h fix smp alternatives fix
Chris Wright [Thu, 31 Aug 2006 07:53:22 +0000 (00:53 -0700)] 
[PATCH] i386: rwlock.h fix smp alternatives fix

Commit 8c74932779fc6f61b4c30145863a17125c1a296c ("i386: Remove
alternative_smp") did not actually compile on x86 with CONFIG_SMP.

This fixes the __build_read/write_lock helpers.  I've boot tested on
SMP.

[ Andi: "Oops, I think that was a quilt unrefreshed patch.  Sorry.  I
  fixed those before testing, but then still send out the old patch." ]

Signed-off-by: Chris Wright <chrisw@sous-sol.org>
Cc: Gerd Hoffmann <kraxel@suse.de>
Acked-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Linus Torvalds [Thu, 31 Aug 2006 04:44:06 +0000 (21:44 -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] Fix return value from memcpy
  [POWERPC] iseries: Define insw et al. so libata/ide will compile
  [POWERPC] Fix irq enable/disable in smp_generic_take_timebase
  [POWERPC] Fix problem with time not advancing on 32-bit platforms
  [POWERPC] Restore copyright notice in arch/powerpc/kernel/fpu.S
  [POWERPC] Fix up ibm_architecture_vec definition
  [POWERPC] Make OF irq map code detect more error cases
  [POWERPC] Support for "weird" MPICs and fixup mpc7448_hpc2
  [POWERPC] Fix MPIC sense codes in documentation
  [POWERPC] Fix performance regression in IRQ radix tree locking
  [POWERPC] Add mpc7448hpc2 device tree source file
  [POWERPC] Add MPC8349E MDS device tree source file to arch/powerpc/boot/dts
  [POWERPC] modify mpc83xx platforms to use new IRQ layer
  [POWERPC] Adapt ipic driver to new host_ops interface, add set_irq_type to set IRQ sense
  [POWERPC] back up old school ipic.[hc] to arch/ppc
  [POWERPC] Use mpc8641hpcn PIC base address from dev tree.
  [POWERPC] Allow MPC8641 HPCN to build with CONFIG_PCI disabled too.
  [POWERPC] Fix powerpc 44x_mmu build
  [POWERPC] Remove flush_dcache_all export

18 years ago[PATCH] ppc32: fix last_jiffy time comparison
Paul Mackerras [Wed, 30 Aug 2006 06:13:16 +0000 (16:13 +1000)] 
[PATCH] ppc32: fix last_jiffy time comparison

This fixes a hang on ppc32.

The problem was that I was comparing a 32-bit quantity with a 64-bit
quantity, and consequently time wasn't advancing.  This makes us use a
64-bit quantity on all platforms, which ends up simplifying the code
since we can now get rid of the tb_last_stamp variable (which actually
fixes another bug that Ben H and I noticed while going carefully through
the code).

This works fine on my G4 tibook.  Let me know how it goes on your
machines.

Acked-by: Olaf Hering <olaf@aepfle.de>
Acked-by: Mikael Pettersson <mikpe@it.uu.se>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fbdev: Fix crashes in various fbdev's blank routines
Benjamin Herrenschmidt [Thu, 31 Aug 2006 04:04:34 +0000 (14:04 +1000)] 
[PATCH] fbdev: Fix crashes in various fbdev's blank routines

The backlight changes that went in had a bug where they could cause the
kernel to access an unitialized pointer when blanking if there is no
backlight control on a machine.

The bug affects atyfb, aty128fb, nvidiafb and rivafb.  radeonfb seems to
be ok.  This fixes it.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[POWERPC] Fix return value from memcpy
Paul Mackerras [Thu, 31 Aug 2006 03:22:58 +0000 (13:22 +1000)] 
[POWERPC] Fix return value from memcpy

As pointed out by Herbert Xu <herbert@gondor.apana.org.au>, our
memcpy implementation didn't return the destination pointer as its
return value, and there is code in the kernel that expects that.
This fixes it.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Thu, 31 Aug 2006 00:12:11 +0000 (17:12 -0700)] 
Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] Increase default nodes shift to 10, nr_cpus to 1024
  [IA64] remove redundant local_irq_save() calls from sn_sal.h
  [IA64] panic if topology_init kzalloc fails
  [IA64-SGI] Silent data corruption caused by XPC V2.

18 years ago[PATCH] drm: radeon flush TCL VAP for vertex program enable/disable
Roland Scheidegger [Wed, 30 Aug 2006 22:17:55 +0000 (23:17 +0100)] 
[PATCH] drm: radeon flush TCL VAP for vertex program enable/disable

The radeon requires a VAP state flush when enabling/disabling
vertex programs on the r200 cards.

Signed-off-by: Dave Airlie <airlied@linux.ie>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Missing PCI id update for VIA IDE
Alan Cox [Wed, 30 Aug 2006 18:14:25 +0000 (19:14 +0100)] 
[PATCH] Missing PCI id update for VIA IDE

The following change from -mm is important to 2.6.18 (actually to 2.6.17
but its too late for that). This was contributed over three months ago
by VIA to Bartlomiej and nothing happened. As a result the new chipset
is now out and Linux won't run on it. By the time 2.6.18 is finalised
this will be the defacto standard VIA chipset so support would be a good
plan.

Tested in -mm for a while, its essentially a PCI ident update but for
the bridge chip because VIA do things in weird ways.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86_64: Don't write out segments from vsyscall32 DSO if it is not mapped
Suleiman Souhlal [Wed, 30 Aug 2006 17:37:20 +0000 (19:37 +0200)] 
[PATCH] x86_64: Don't write out segments from vsyscall32 DSO if it is not mapped

It's possible to get an invalid page fault in kernel mode when we try to
write out segments from vsyscall32 when dumping core for a 32bit process if
the vsyscall32 DSO is not mapped in its address space (which can happen if,
for example, ulimit -v 100 is run).

Signed-off-by: Suleiman Souhlal <suleiman@google.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86_64: Save original IST values for checking stack addresses
Keith Owens [Wed, 30 Aug 2006 17:37:19 +0000 (19:37 +0200)] 
[PATCH] x86_64: Save original IST values for checking stack addresses

The values in init_tss.ist[] can change when an IST event occurs.  Save
the original IST values for checking stack addresses when debugging or
doing stack traces.

Signed-off-by: Keith Owens <kaos@ocs.com.au>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i386: Remove __KERNEL__ ifdef around _syscall*()
Andi Kleen [Wed, 30 Aug 2006 17:37:18 +0000 (19:37 +0200)] 
[PATCH] i386: Remove __KERNEL__ ifdef around _syscall*()

After all their only point is having them in user space.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i386: Fix stack switching in do_IRQ
Andi Kleen [Wed, 30 Aug 2006 17:37:17 +0000 (19:37 +0200)] 
[PATCH] i386: Fix stack switching in do_IRQ

There was a bogus hunk from the genirq merge that essentially
broke stack switching for hard interrupts. Remove it since it isn't
needed.

Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86_64: Remove __KERNEL__ ifdef around _syscall*()
Andi Kleen [Wed, 30 Aug 2006 17:37:16 +0000 (19:37 +0200)] 
[PATCH] x86_64: Remove __KERNEL__ ifdef around _syscall*()

After all their only point is having them in user space. On x86-64
they don't even work in kernel space.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86: Disable MMCONFIG on Intel SDV using DMI blacklist
Andi Kleen [Wed, 30 Aug 2006 17:37:15 +0000 (19:37 +0200)] 
[PATCH] x86: Disable MMCONFIG on Intel SDV using DMI blacklist

As a replacement for the earlier removal of the e820 MCFG check
we blacklist the Intel SDV with the original BIOS bug that
motivated that check. On those machines don't use MMCONFIG.

This also adds a new pci=mmconf parameter to override the blacklist.

Cc: Greg KH <gregkh@suse.de>
Cc: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i386: Remove alternative_smp
Andi Kleen [Wed, 30 Aug 2006 17:37:14 +0000 (19:37 +0200)] 
[PATCH] i386: Remove alternative_smp

The .fill causes miscompilations with some binutils version.

Instead just patch the lock prefix in the lock constructs. That is the
majority of the cost and should be good enough.

Cc: Gerd Hoffmann <kraxel@suse.de>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86_64: Remove alternative_smp
Andi Kleen [Wed, 30 Aug 2006 17:37:13 +0000 (19:37 +0200)] 
[PATCH] x86_64: Remove alternative_smp

The .fill causes miscompilations with some binutils version.

Instead just patch the lock prefix in the lock constructs. That is the
majority of the cost and should be good enough.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86_64: Recover 1MB of kernel memory
Andi Kleen [Wed, 30 Aug 2006 17:37:12 +0000 (19:37 +0200)] 
[PATCH] x86_64: Recover 1MB of kernel memory

Noticed by Jan Beulich.

When the kernel was moved from 1MB to 2MB in 2.6.17 the kernel reservation
code wasn't adjusted and it still reserved starting with 1MB. This means 1MB always
were lost.

This patch fixes this by reserving only starting with _text.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86: Make backtracer fallback logic more bullet-proof
Jan Beulich [Wed, 30 Aug 2006 17:37:11 +0000 (19:37 +0200)] 
[PATCH] x86: Make backtracer fallback logic more bullet-proof

The unwinder fallback logic still had potential for falling through to
the legacy stack trace code without printing an indication (at once
serving as a separator) of this.

Further, the stack pointer retrieval for the fallback should be as
restrictive as possible (in order to avoid having the legacy stack
tracer try to access invalid memory). The patch tightens that, but
this could certainly be further improved.

Also making the call_trace command line option now conditional upon
CONFIG_STACK_UNWIND (as it's meaningless otherwise).

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86: fix x86 cpuid keys used in alternative_smp()
Jan Beulich [Wed, 30 Aug 2006 17:37:10 +0000 (19:37 +0200)] 
[PATCH] x86: fix x86 cpuid keys used in alternative_smp()

By hard-coding the cpuid keys for alternative_smp() rather than using
the symbolic constant it turned out that incorrect values were used on
both i386 (0x68 instead of 0x69) and x86-64 (0x66 instead of 0x68).

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i386: Add kernel thread stack frame termination for properly stopping stack...
Andi Kleen [Wed, 30 Aug 2006 17:37:09 +0000 (19:37 +0200)] 
[PATCH] i386: Add kernel thread stack frame termination for properly stopping stack unwinds.

One open question: Should this added push perhaps be made conditional
upon CONFIG_STACK_UNWIND or CONFIG_UNWIND_INFO?
[AK: not needed, these are all very slow paths]

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86_64: Add kernel thread stack frame termination for properly stopping stack...
Andi Kleen [Wed, 30 Aug 2006 17:37:08 +0000 (19:37 +0200)] 
[PATCH] x86_64: Add kernel thread stack frame termination for properly stopping stack unwinds.

One open question: Should these added pushes perhaps be made
conditional upon CONFIG_STACK_UNWIND or CONFIG_UNWIND_INFO?
[AK: Not needed -- these are all very slow paths]

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86: Revert e820 MCFG heuristics
Andi Kleen [Wed, 30 Aug 2006 17:37:06 +0000 (19:37 +0200)] 
[PATCH] x86: Revert e820 MCFG heuristics

The check for the MCFG table being reserved in the e820 map was originally
added to detect a broken BIOS in a preproduction Intel SDV. However it also
breaks the Apple x86 Macs, which can't supply this properly, but need
a working MCFG. With this patch they wouldn't use the MCFG and not work.

After some discussion I think it's best to remove the heuristic again.
It also failed on some other boxes (although it didn't cause much
problems there because old style port access for PCI config space
still works as fallback), but the preproduction SDVs can just use
pci=nommcfg. Supporting production machines properly is more
important.

Edgar Hucek did all the debugging work.

Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Edgar Hucek <hostmaster@ed-soft.at>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86_64: Update defconfig
Andi Kleen [Wed, 30 Aug 2006 17:37:05 +0000 (19:37 +0200)] 
[PATCH] x86_64: Update defconfig

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Wed, 30 Aug 2006 22:54:55 +0000 (15:54 -0700)] 
Merge /pub/scm/linux/kernel/git/davem/sparc-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SUNLANCE]: Fix probing problem.
  [SPARC64]: Fix X server hangs due to large pages.

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 30 Aug 2006 22:54:35 +0000 (15:54 -0700)] 
Merge /pub/scm/linux/kernel/git/davem/net-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [STRIP]: Fix neighbour table refcount leak.
  [IPV6]: ipv6_add_addr should install dstentry earlier
  [NETLINK]: Call panic if nl_table allocation fails
  [TCP]: Two RFC3465 Appropriate Byte Count fixes.
  [IPV6]: SNMPv2 "ipv6IfStatsInAddrErrors" counter error
  [E100]: Add module option to ignore bad EEPROM checksums.
  [SCTP]: Fix sctp_primitive_ABORT() call in sctp_close().

18 years ago[S390] cio: unsolicited interrupts during sense pgid.
Stefan Bader [Wed, 30 Aug 2006 12:33:39 +0000 (14:33 +0200)] 
[S390] cio: unsolicited interrupts during sense pgid.

Calls to set a device online with path grouping may get stuck in
some cases because certain device conditions where discarded after
unsolicited interrupts.
Check subchannel activity after unsolicited interrupts and retry
the operation if the subchannel is idle.

Signed-off-by: Stefan Bader <shbader@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
18 years ago[S390] cio: no path after machine check.
Peter Oberparleiter [Wed, 30 Aug 2006 12:33:37 +0000 (14:33 +0200)] 
[S390] cio: no path after machine check.

Devices enter no-path state after disabling a channel path
via the SE even though another path has been reenabled at the SE.
The devices are set into no-path state before triggering path
verification even though other paths may have become available.
To fix this trigger path verification before setting a device into
no-path state.

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
18 years ago[S390] cio: kernel stack overflow.
Heiko Carstens [Wed, 30 Aug 2006 12:33:35 +0000 (14:33 +0200)] 
[S390] cio: kernel stack overflow.

Use different kind of assignment to make sure gcc doesn't create code
that creates temp variables on the stack, assigns values to it and
copies the content of the whole temp variable to the destination.
This reduces stack usage of e.g. ccwgroup_driver_register from 976
to 48 bytes instead.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
18 years ago[S390] dasd: fix device shutdown process.
Horst Hummel [Wed, 30 Aug 2006 12:33:33 +0000 (14:33 +0200)] 
[S390] dasd: fix device shutdown process.

Fix clear_IO handling (need to wait for interrupt) and
introduced error-handling in shutdown processing.

Signed-off-by: Horst Hummel <horst.hummel@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
18 years ago[S390] broken copy_in_user function.
Martin Schwidefsky [Wed, 30 Aug 2006 12:33:30 +0000 (14:33 +0200)] 
[S390] broken copy_in_user function.

The copy_in_user primitive does not work as advertised. If the source
and target area are available copy_in_user copies one byte too much.
If one of the memory areas is not available it does not copy as much
data as it can, but up to 257 bytes less.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
18 years ago[POWERPC] iseries: Define insw et al. so libata/ide will compile
Stephen Rothwell [Wed, 30 Aug 2006 07:11:34 +0000 (17:11 +1000)] 
[POWERPC] iseries: Define insw et al. so libata/ide will compile

These are build fixes that enable (for example) libata and the ide
code to actually build on iSeries.  The associated hardware will never
be supported on legacy iSeries, so the code paths don't actually need
to work, but it is useful (especially for a combined kernel) if the
code can build.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] Fix irq enable/disable in smp_generic_take_timebase
Paul Mackerras [Wed, 30 Aug 2006 06:10:47 +0000 (16:10 +1000)] 
[POWERPC] Fix irq enable/disable in smp_generic_take_timebase

Eran Ben-Avi <eranpublic@yahoo.com> pointed out that the arch/ppc version
of smp_generic_take_timebase disables interrupts on entry but exits without
restoring them.  However, both it and the arch/powerpc version have another
problem, which is that they use local_irq_disable/enable rather than
local_irq_save/restore, and they are called with interrupts disabled.

This fixes both problems; it changes a return to a break in the arch/ppc
version, and changes both versions to use local_irq_save/restore.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] Fix problem with time not advancing on 32-bit platforms
Paul Mackerras [Wed, 30 Aug 2006 05:55:32 +0000 (15:55 +1000)] 
[POWERPC] Fix problem with time not advancing on 32-bit platforms

This fixes a problem introduced in 5db9fa9593e2ff69f2b95f9d59229dc4faaa564d.
The last_jiffy per-cpu variable is only 32 bits on 32-bit machines, but it
was being compared with a 64-bit quantity (tb_next_jiffy), which resulted in
time not advancing.

This fixes it by changing last_jiffy to be 64 bits on all platforms.  With
this, we no longer need tb_last_stamp as a 32-bit version of tb_last_jiffy,
so this gets rid of tb_last_stamp and we just use tb_last_jiffy instead.
This also fixes a bug when the boot cpu is not online, because using
tb_last_stamp could have caused the wrong timebase origin value to be used
when calculating the time of day.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] Restore copyright notice in arch/powerpc/kernel/fpu.S
Paul Mackerras [Wed, 30 Aug 2006 04:45:35 +0000 (14:45 +1000)] 
[POWERPC] Restore copyright notice in arch/powerpc/kernel/fpu.S

This code got moved from head.S but the copyright notice on head.S didn't
get transferred with it.  Noticed by Cort Dougan <cort@fsmlabs.com>.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] Fix up ibm_architecture_vec definition
Will Schmidt [Fri, 25 Aug 2006 20:46:59 +0000 (15:46 -0500)] 
[POWERPC] Fix up ibm_architecture_vec definition

This problem was noticed by one of the Phyp firmware folks.
Our ibm,client-architecture-support call was failing.
This corrects the vector length parameters being passed in.

Signed-off-by: Will Schmidt <will_schmidt@vnet.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] Make OF irq map code detect more error cases
Benjamin Herrenschmidt [Fri, 25 Aug 2006 04:46:23 +0000 (14:46 +1000)] 
[POWERPC] Make OF irq map code detect more error cases

Device-tree bugs on js20 with some versions of SLOF were causing the
interrupt for IDE to not be parsed correctly and fail to boot. This
patch adds a bit more sanity checking to the parser to detect some of
those errors and fail instead of returning bogus information.  The
powerpc PCI code can then trigger a fallback that works on those
machines.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] Support for "weird" MPICs and fixup mpc7448_hpc2
Zang Roy-r61911 [Fri, 25 Aug 2006 04:16:30 +0000 (14:16 +1000)] 
[POWERPC] Support for "weird" MPICs and fixup mpc7448_hpc2

This adds a new hardware information table for mpic. This enables
the mpic code to deal with mpic controllers with different register
layouts and hardware behaviours.

This introduces CONFIG_MPIC_WEIRD.  For boards with non standard mpic
controllers, select CONFIG_MPIC_WEIRD and add its hardware information
in the mpic_infos[] array.

TSI108/109 PIC takes the first index of weird hardware information
table.  :)  The table can be extended. The Tsi108/109 PIC looks like
standard OpenPIC but, in fact, is different in register mapping and
behavior.

The patch does not affect the behavior of standard mpic.  If
CONFIG_MPIC_WEIRD is not defined, the code is essentially identical to
the current code.

[benh@kernel.crashing.org:
This patch is a slightly cleaned up version of Zang Roy's support for
the TSI108 MPIC variant. It also fixes up MPC7448_hpc2 to use the new
version of the type macros and changes the way MPIC is selected in
Kconfig to better match what is done for other system devices.
]

Signed-off-by: Roy Zang <tie-fei.zang@freescale.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[SUNLANCE]: Fix probing problem.
Krzysztof Helt [Tue, 29 Aug 2006 06:28:16 +0000 (23:28 -0700)] 
[SUNLANCE]: Fix probing problem.

The current probe table causes ledma and lebuffer
"le" devices to get probed twice which is not what
we want.

Match just "le" and look directly at the parent to get the correct
top-level node information.

Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Fix X server hangs due to large pages.
David S. Miller [Mon, 28 Aug 2006 07:33:03 +0000 (00:33 -0700)] 
[SPARC64]: Fix X server hangs due to large pages.

This problem was introduced by changeset
14778d9072e53d2171f66ffd9657daff41acfaed

Unlike the hugetlb code paths, the normal fault code is not setup to
propagate PTE changes for large page sizes correctly like the ones we
make for I/O mappings in io_remap_pfn_range().

It is absolutely necessary to update all sub-ptes of a largepage
mapping on a fault.  Adding special handling for this would add
considerably complexity to tlb_batch_add().  So let's just side-step
the issue and forcefully dirty any writable PTEs created by
io_remap_pfn_range().

The only other real option would be to disable to large PTE code of
io_remap_pfn_range() and we really don't want to do that.

Much thanks to Mikael Pettersson for tracking down this problem and
testing debug patches.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[STRIP]: Fix neighbour table refcount leak.
Stephen Hemminger [Wed, 30 Aug 2006 04:18:45 +0000 (21:18 -0700)] 
[STRIP]: Fix neighbour table refcount leak.

Found by inspection. The STRIP driver does neigh_lookup() but never
releases.  This driver shouldn't being doing gratuitous arp anyway.

Untested, obviously, because of lack of hardware.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPV6]: ipv6_add_addr should install dstentry earlier
Keir Fraser [Tue, 29 Aug 2006 09:43:49 +0000 (02:43 -0700)] 
[IPV6]: ipv6_add_addr should install dstentry earlier

ipv6_add_addr allocates a struct inet6_ifaddr and a dstentry, but it
doesn't install the dstentry in ifa->rt until after it releases the
addrconf_hash_lock. This means other CPUs will be able to see the new
address while it hasn't been initialized completely yet.

One possible fix would be to grab the ifp->lock spinlock when
creating the address struct; a simpler fix is to just move the
assignment.

Acked-by: jbeulich@novell.com
Acked-by: okir@suse.de
Acked-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETLINK]: Call panic if nl_table allocation fails
Akinobu Mita [Tue, 29 Aug 2006 09:15:24 +0000 (02:15 -0700)] 
[NETLINK]: Call panic if nl_table allocation fails

This patch makes crash happen if initialization of nl_table fails
in initcalls. It is better than getting use after free crash later.

Signed-off-by: Akinobu Mita <mita@miraclelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[TCP]: Two RFC3465 Appropriate Byte Count fixes.
Daikichi Osuga [Tue, 29 Aug 2006 09:01:44 +0000 (02:01 -0700)] 
[TCP]: Two RFC3465 Appropriate Byte Count fixes.

1) fix slow start after retransmit timeout
2) fix case of L=2*SMSS acked bytes comparison

Signed-off-by: Daikichi Osuga <osugad@s1.nttdocomo.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPV6]: SNMPv2 "ipv6IfStatsInAddrErrors" counter error
Lv Liangying [Tue, 29 Aug 2006 07:00:47 +0000 (00:00 -0700)] 
[IPV6]: SNMPv2 "ipv6IfStatsInAddrErrors" counter error

   When I tested Linux kernel 2.6.17.7 about statistics
"ipv6IfStatsInAddrErrors", found that this counter couldn't increase
correctly. The criteria is RFC2465:
 ipv6IfStatsInAddrErrors OBJECT-TYPE
      SYNTAX      Counter32
      MAX-ACCESS  read-only
      STATUS      current
      DESCRIPTION
         "The number of input datagrams discarded because
         the IPv6 address in their IPv6 header's destination
         field was not a valid address to be received at
         this entity.  This count includes invalid
         addresses (e.g., ::0) and unsupported addresses
         (e.g., addresses with unallocated prefixes).  For
         entities which are not IPv6 routers and therefore
         do not forward datagrams, this counter includes
         datagrams discarded because the destination address
         was not a local address."
      ::= { ipv6IfStatsEntry 5 }

When I send packet to host with destination that is ether invalid
address(::0) or unsupported addresses(1::1), the Linux kernel just
discard the packet, and the counter doesn't increase(in the function
ip6_pkt_discard).

Signed-off-by: Lv Liangying <lvly@nanjing-fnst.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[E100]: Add module option to ignore bad EEPROM checksums.
David S. Miller [Tue, 29 Aug 2006 05:12:54 +0000 (22:12 -0700)] 
[E100]: Add module option to ignore bad EEPROM checksums.

Several people run into the situation where the E100
EEPROM contents are fine, but the checksum hasn't been
set properly.  This renders the device useless for
them even though it would function correctly.

The default is off, which retains the current behavior.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SCTP]: Fix sctp_primitive_ABORT() call in sctp_close().
Sridhar Samudrala [Mon, 28 Aug 2006 20:53:01 +0000 (13:53 -0700)] 
[SCTP]: Fix sctp_primitive_ABORT() call in sctp_close().

With the recent fix, the callers of sctp_primitive_ABORT()
need to create an ABORT chunk and pass it as an argument rather
than msghdr that was passed earlier.

Signed-off-by: Sridhar Samudrala <sri@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[XFS] Fix char size overflow in bmap_alloc call for unwritten extent
Adrian Bunk [Wed, 30 Aug 2006 03:41:58 +0000 (13:41 +1000)] 
[XFS] Fix char size overflow in bmap_alloc call for unwritten extent
conversion.

Since bma.conv is a char and XFS_BMAPI_CONVERT is 0x1000, bma.conv was
always assigned zero. Spotted by the GNU C compiler (SVN version).

SGI-PV: 947312
SGI-Modid: xfs-linux-melb:xfs-kern:26887a

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Nathan Scott <nathans@sgi.com>
18 years ago[XFS] Update the MAINTAINERS file entry for XFS.
Nathan Scott [Wed, 30 Aug 2006 03:41:23 +0000 (13:41 +1000)] 
[XFS] Update the MAINTAINERS file entry for XFS.

Signed-off-by: Nathan Scott <nathans@sgi.com>
18 years ago[POWERPC] Fix MPIC sense codes in documentation
Benjamin Herrenschmidt [Tue, 29 Aug 2006 22:58:00 +0000 (08:58 +1000)] 
[POWERPC] Fix MPIC sense codes in documentation

The booting-without-of.txt had incorrect definition for the sense codes
for an OpenPIC controller

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] Fix performance regression in IRQ radix tree locking
Benjamin Herrenschmidt [Mon, 28 Aug 2006 01:17:37 +0000 (11:17 +1000)] 
[POWERPC] Fix performance regression in IRQ radix tree locking

When reworking the powerpc irq code, I figured out that we were using
the radix tree in a racy way. As a temporary fix, I put a spinlock in
there. However, this can have a significant impact on performances. This
patch reworks that to use a smarter technique based on the fact that
what we need is in fact a rwlock with extremely rare writers (thus
optimized for the read path).

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] Add mpc7448hpc2 device tree source file
Zang Roy-r61911 [Fri, 25 Aug 2006 08:43:25 +0000 (16:43 +0800)] 
[POWERPC] Add mpc7448hpc2 device tree source file

This patch adds the mpc7448hpc2 device tree source file.

Signed-off-by: Roy Zang <tie-fei.zang@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] Add MPC8349E MDS device tree source file to arch/powerpc/boot/dts
Kim Phillips [Tue, 29 Aug 2006 23:13:31 +0000 (18:13 -0500)] 
[POWERPC] Add MPC8349E MDS device tree source file to arch/powerpc/boot/dts

Add MPC8349E MDS device tree source file to arch/powerpc/boot/dts

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] modify mpc83xx platforms to use new IRQ layer
Kim Phillips [Fri, 25 Aug 2006 16:59:22 +0000 (11:59 -0500)] 
[POWERPC] modify mpc83xx platforms to use new IRQ layer

This fixes MPC834x MDS (formerly SYS) and ITX platform code to get IRQ data (including PCI) from the device tree, and to use the new IPIC code.

renamed defconfig (sys -> mds), left one redundant NULL assignment in mpc83xx_pcibios_fixup to keep the compiler happy.

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Li Yang <leoli@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] Adapt ipic driver to new host_ops interface, add set_irq_type to set IRQ...
Kim Phillips [Fri, 25 Aug 2006 16:59:07 +0000 (11:59 -0500)] 
[POWERPC] Adapt ipic driver to new host_ops interface, add set_irq_type to set IRQ sense

This converts ipic code to Benh's IRQ mods.  For the IPIC, IRQ sense values in the device tree equal those in include/linux/irq.h; that's 8 for low assertion (most internal IRQs on mpc83xx), and 2 for high-to-low change.

spinlocks added to [un]mask, ack operations; default handler and type now set in host_map; and redundant condition check eliminated.

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Li Yang <leoli@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[POWERPC] back up old school ipic.[hc] to arch/ppc
Kim Phillips [Fri, 25 Aug 2006 16:58:53 +0000 (11:58 -0500)] 
[POWERPC] back up old school ipic.[hc] to arch/ppc

Keep from breaking 83xx arch/ppc build.  Back up old school arch/powerpc/sysdev/ipic.[hc] to arch/ppc/syslib.

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6
Linus Torvalds [Tue, 29 Aug 2006 22:54:07 +0000 (15:54 -0700)] 
Merge git://git./linux/kernel/git/sfrench/cifs-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
  [CIFS] Do not send Query All EAs SMB when mount option nouser_xattr
  [CIFS] endian errors in lanman protocol support
  [CIFS] Fix oops in cifs_close due to unitialized lock sem and list in
  [CIFS] Fix oops when negotiating lanman and no password specified
  [CIFS]
  [CIFS] Allow cifsd to suspend if connection is lost
  [CIFS] Make midState usage more consistent
  [CIFS] spinlock protect read of last srv response time in timeout path
  [CIFS] Do not time out posix brl requests when using new posix setfileinfo

18 years ago[IA64] Increase default nodes shift to 10, nr_cpus to 1024
Christoph Lameter [Wed, 23 Aug 2006 02:43:27 +0000 (19:43 -0700)] 
[IA64] Increase default nodes shift to 10, nr_cpus to 1024

Change both the NODES_SHIFT and the NR_CPUS so that even big machines
can boot all nodes and processors with a generic kernel.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64] remove redundant local_irq_save() calls from sn_sal.h
Russ Anderson [Thu, 24 Aug 2006 16:08:52 +0000 (11:08 -0500)] 
[IA64] remove redundant local_irq_save() calls from sn_sal.h

sn_change_memprotect() does a local_irq_save() then calls
ia64_sal_oemcall_nolock() which calls SAL_CALL_NOLOCK()
which also does a local_irq_save().

This patch removes the redundant local_irq_save() and local_irq_restore()
calls in sn_change_memprotect() and sn_inject_error().

Signed-off-by: Russ Anderson <rja@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years agoMerge master.kernel.org:/home/rmk/linux-2.6-serial
Linus Torvalds [Tue, 29 Aug 2006 03:19:55 +0000 (20:19 -0700)] 
Merge master.kernel.org:/home/rmk/linux-2.6-serial

* master.kernel.org:/home/rmk/linux-2.6-serial:
  [SERIAL] Support for Intashield 2 port PCI serial card

18 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Tue, 29 Aug 2006 03:19:16 +0000 (20:19 -0700)] 
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm:
  [ARM] 3761/1: fix armv4t breakage after adding thumb interworking to userspace helpers
  [ARM] Add Integrator support for glibc outb() and friends
  [ARM] Move prototype for register_isa_ports to asm/io.h
  [ARM] Arrange for isa.c to use named initialisers
  [ARM] 3741/1: remove sa1111.c build warning on non-sa1100 systems
  [ARM] 3760/1: This patch adds timeouts while working with SSP registers. Such timeouts were en
  [ARM] 3758/1: Preserve signalling NaNs in conversion
  [ARM] 3749/3: Correct VFP single/double conversion emulation
  [ARM] 3748/3: Correct error check in vfp_raise_exceptions

18 years ago[ARM] 3761/1: fix armv4t breakage after adding thumb interworking to userspace helpers
Lennert Buytenhek [Mon, 28 Aug 2006 11:51:20 +0000 (12:51 +0100)] 
[ARM] 3761/1: fix armv4t breakage after adding thumb interworking to userspace helpers

Patch from Lennert Buytenhek

On armv4t systems, we have always compiled the kernel with -march=armv4
instead of -march=armv4t, which means that any use of bx will bomb out.

Commit ba9b5d76372dc290b6ca04dad93927a22c2ac49a introduced the use of
bx in the kernel, which means we need to compile with -march=armv4t on
armv4t systems now.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] Add Integrator support for glibc outb() and friends
Russell King [Mon, 28 Aug 2006 11:47:05 +0000 (12:47 +0100)] 
[ARM] Add Integrator support for glibc outb() and friends

Add the necessary call to register_isa_ports() so that glibc knows
where these are found on Integrator platforms.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] Move prototype for register_isa_ports to asm/io.h
Russell King [Mon, 28 Aug 2006 11:45:16 +0000 (12:45 +0100)] 
[ARM] Move prototype for register_isa_ports to asm/io.h

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] Arrange for isa.c to use named initialisers
Russell King [Mon, 28 Aug 2006 11:44:03 +0000 (12:44 +0100)] 
[ARM] Arrange for isa.c to use named initialisers

Convert isa.c (the glibc interface for emulating ISA IO) to use
named initialisers.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years agoLinux v2.6.18-rc5 v2.6.18-rc5
Linus Torvalds [Mon, 28 Aug 2006 03:41:48 +0000 (20:41 -0700)] 
Linux v2.6.18-rc5

18 years ago[PATCH] Fix tty layer DoS and comment relevant code
Alan Cox [Sun, 27 Aug 2006 08:24:02 +0000 (01:24 -0700)] 
[PATCH] Fix tty layer DoS and comment relevant code

Unlike the other tty comment patch this one has code changes.  Specifically
it limits the queue size for a tty to 64K characters (128Kbytes) worst case
even if the tty is ignoring tty->throttle.  This is because certain drivers
don't honour the throttle value correctly, although it is a useful
safeguard anyway.

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] tty layer comment the locking assumptions and functions somewhat
Alan Cox [Sun, 27 Aug 2006 08:24:01 +0000 (01:24 -0700)] 
[PATCH] tty layer comment the locking assumptions and functions somewhat

Doesn't fix them but does show up some interesting areas that need review
and fixing.

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] cdrom/gdsc: fix printk format warning
Randy Dunlap [Sun, 27 Aug 2006 08:24:00 +0000 (01:24 -0700)] 
[PATCH] cdrom/gdsc: fix printk format warning

Fix printk format warning:
drivers/cdrom/gscd.c:269: warning: format â\80\98%luâ\80\99 expects type â\80\98long unsigned intâ\80\99, but argument 2 has type â\80\98unsigned intâ\80\99

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Cc: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86: NUMAQ Kconfig fix
KAMEZAWA Hiroyuki [Sun, 27 Aug 2006 08:24:00 +0000 (01:24 -0700)] 
[PATCH] x86: NUMAQ Kconfig fix

When we select NUMA with i386, the system is only X86_NUMAQ or using ACPI.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] /proc/meminfo: don't put spaces in names
Andrew Morton [Sun, 27 Aug 2006 08:23:58 +0000 (01:23 -0700)] 
[PATCH] /proc/meminfo: don't put spaces in names

None of the other /proc/meminfo lines have a space in the identifier.  This
post-2.6.17 addition has the potential to break existing parsers, so use an
underscore instead (like Committed_AS).

Cc: Christoph Lameter <clameter@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fix up lockdep trace in fs/exec.c
Dave Jones [Sun, 27 Aug 2006 08:23:57 +0000 (01:23 -0700)] 
[PATCH] fix up lockdep trace in fs/exec.c

This fixes the locking error noticed by lockdep:

  =============================================
  [ INFO: possible recursive locking detected ]
  ---------------------------------------------
  init/1 is trying to acquire lock:
   (&sighand->siglock){....}, at: [<c047a78a>] flush_old_exec+0x3ae/0x859

  but task is already holding lock:
   (&sighand->siglock){....}, at: [<c047a77a>] flush_old_exec+0x39e/0x859

  other info that might help us debug this:
  2 locks held by init/1:
   #0:  (tasklist_lock){..--}, at: [<c047a76a>] flush_old_exec+0x38e/0x859
   #1:  (&sighand->siglock){....}, at: [<c047a77a>] flush_old_exec+0x39e/0x859

  stack backtrace:
   [<c04051e1>] show_trace_log_lvl+0x54/0xfd
   [<c040579d>] show_trace+0xd/0x10
   [<c04058b6>] dump_stack+0x19/0x1b
   [<c043b33a>] __lock_acquire+0x773/0x997
   [<c043bacf>] lock_acquire+0x4b/0x6c
   [<c060630b>] _spin_lock+0x19/0x28
   [<c047a78a>] flush_old_exec+0x3ae/0x859
   [<c0498053>] load_elf_binary+0x4aa/0x1628
   [<c0479cab>] search_binary_handler+0xa7/0x24e
   [<c047b577>] do_execve+0x15b/0x1f9
   [<c04022b4>] sys_execve+0x29/0x4d
   [<c0403faf>] syscall_call+0x7/0xb

Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] lockdep: annotate reiserfs
Ingo Molnar [Sun, 27 Aug 2006 08:23:56 +0000 (01:23 -0700)] 
[PATCH] lockdep: annotate reiserfs

reiserfs seems to have another locking level layer for the i_mutex due to the
xattrs-are-a-directory thing.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] lockdep: annotate idescsi_pc_intr()
Ingo Molnar [Sun, 27 Aug 2006 08:23:56 +0000 (01:23 -0700)] 
[PATCH] lockdep: annotate idescsi_pc_intr()

idescsi_pc_intr() uses local_irq_enable() in IRQ context: annotate it.

(this has no effect on kernels with lockdep disabled.  On kernels with lockdep
enabled this means that we wont actually disable interrupts, and the warning
message will go away as well.)

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] CONFIG_ACPI_SRAT NUMA build fix
KAMEZAWA Hiroyuki [Sun, 27 Aug 2006 08:23:55 +0000 (01:23 -0700)] 
[PATCH] CONFIG_ACPI_SRAT NUMA build fix

  In file included from include/asm/mmzone.h:18,
                   from include/linux/mmzone.h:439,
  <snip>
  include/asm/srat.h:31:2: error: #error CONFIG_ACPI_SRAT not defined, and srat.h header has been included
  make[1]: *** [arch/i386/kernel/asm-offsets.s] Error 1

This can happen with CONFIG_NUMA && !CONFIG_ACPI && !CONFIG_X86_NUMAQ

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] cpuset: oom panic fix
Nick Piggin [Sun, 27 Aug 2006 08:23:54 +0000 (01:23 -0700)] 
[PATCH] cpuset: oom panic fix

cpuset_excl_nodes_overlap always returns 0 if current is exiting.  This caused
customer's systems to panic in the OOM killer when processes were having
trouble getting memory for the final put_user in mm_release.  Even though
there were lots of processes to kill.

Change to returning 1 in this case.  This achieves parity with !CONFIG_CPUSETS
case, and was observed to fix the problem.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Acked-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] register_one_node() compile fix
KAMEZAWA Hiroyuki [Sun, 27 Aug 2006 08:23:52 +0000 (01:23 -0700)] 
[PATCH] register_one_node() compile fix

register_one_node()'s should be defined under CONFIG_NUMA=n.
fixes following bug.

  CC   init/version.o
  LD   init/built-in.o
  LD   .tmp_vmlinux1
  mm/built-in.o: In function `add_memory': undefined reference to `register_one_node'

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Manage jbd allocations from its own slabs
Badari Pulavarty [Sun, 27 Aug 2006 08:23:52 +0000 (01:23 -0700)] 
[PATCH] Manage jbd allocations from its own slabs

JBD currently allocates commit and frozen buffers from slabs.  With
CONFIG_SLAB_DEBUG, its possible for an allocation to cross the page
boundary causing IO problems.

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=200127

So, instead of allocating these from regular slabs - manage allocation from
its own slabs and disable slab debug for these slabs.

[akpm@osdl.org: cleanups]
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] cpuset: top_cpuset tracks hotplug changes to cpu_online_map
Paul Jackson [Sun, 27 Aug 2006 08:23:51 +0000 (01:23 -0700)] 
[PATCH] cpuset: top_cpuset tracks hotplug changes to cpu_online_map

Change the list of cpus allowed to tasks in the top (root) cpuset to
dynamically track what cpus are online, using a CPU hotplug notifier.  Make
this top cpus file read-only.

On systems that have cpusets configured in their kernel, but that aren't
actively using cpusets (for some distros, this covers the majority of
systems) all tasks end up in the top cpuset.

If that system does support CPU hotplug, then these tasks cannot make use
of CPUs that are added after system boot, because the CPUs are not allowed
in the top cpuset.  This is a surprising regression over earlier kernels
that didn't have cpusets enabled.

In order to keep the behaviour of cpusets consistent between systems
actively making use of them and systems not using them, this patch changes
the behaviour of the 'cpus' file in the top (root) cpuset, making it read
only, and making it automatically track the value of cpu_online_map.  Thus
tasks in the top cpuset will have automatic use of hot plugged CPUs allowed
by their cpuset.

Thanks to Anton Blanchard and Nathan Lynch for reporting this problem,
driving the fix, and earlier versions of this patch.

Signed-off-by: Paul Jackson <pj@sgi.com>
Cc: Nathan Lynch <ntl@pobox.com>
Cc: Anton Blanchard <anton@samba.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] md: fix recent breakage of md/raid1 array checking
NeilBrown [Sun, 27 Aug 2006 08:23:50 +0000 (01:23 -0700)] 
[PATCH] md: fix recent breakage of md/raid1 array checking

A recent patch broke the ability to do a user-request check of a raid1.
This patch fixes the breakage and also moves a comment that was dislocated
by the same patch.

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] md: avoid backward event updates in md superblock when degraded.
NeilBrown [Sun, 27 Aug 2006 08:23:49 +0000 (01:23 -0700)] 
[PATCH] md: avoid backward event updates in md superblock when degraded.

If we
  - shut down a clean array,
  - restart with one (or more) drive(s) missing
  - make some changes
  - pause, so that they array gets marked 'clean',
the event count on the superblock of included drives
will be the same as that of the removed drives.
So adding the removed drive back in will cause it
to be included with no resync.

To avoid this, we only update the eventcount backwards when the array
is not degraded.  In this case there can (should) be no non-connected
drives that we can get confused with, and this is the particular case
where updating-backwards is valuable.

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] eventpoll.c compile fix
Masoud Asgharifard Sharbiani [Sun, 27 Aug 2006 08:23:48 +0000 (01:23 -0700)] 
[PATCH] eventpoll.c compile fix

Fix two compile failures in eventpoll.c code which would happen if
DEBUG_EPOLL is bigger than zero.

Signed-off-by: Masoud Sharbiani <masouds@google.com>
Cc: Davide Libenzi <davidel@xmailserver.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Documentation update for relay interface
Tom Zanussi [Sun, 27 Aug 2006 08:23:47 +0000 (01:23 -0700)] 
[PATCH] Documentation update for relay interface

Here's updated documentation for the relay interface, rewritten to match
the relayfs->relay changes.  It also moves relayfs.txt to relay.txt in the
process.

It includes the changes to relayfs.txt previously posted by Randy Dunlap,
thanks for those.

The relay-apps examples have also been updated to match, and can be found
on the sourceforge relayfs website.

Signed-off-by: Tom Zanussi <zanussi@us.ibm.com>
Cc: "Randy.Dunlap" <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Remove redundant up() in stop_machine()
Yingchao Zhou [Sun, 27 Aug 2006 08:23:46 +0000 (01:23 -0700)] 
[PATCH] Remove redundant up() in stop_machine()

An up() is called in kernel/stop_machine.c on failure, and also in the
caller (unconditionally).

Signed-off-by: Zhou Yingchao <yingchao.zhou@gmail.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ufs: truncate correction
Evgeniy Dushistov [Sun, 27 Aug 2006 08:23:46 +0000 (01:23 -0700)] 
[PATCH] ufs: truncate correction

1) When we allocated last fragment in ufs_truncate, we read page, check
   if block mapped to address, and if not trying to allocate it.  This is
   wrong behaviour, fragment may be NOT allocated, but mapped, this
   happened because of "block map" function not checked allocated fragment
   or not, it just take address of the first fragment in the block, add
   offset of fragment and return result, this is correct behaviour in
   almost all situation except call from ufs_truncate.

2) Almost all implementation of UFS, which I can investigate have such
   "defect": if you have full disk, and try truncate file, for example 3GB
   to 2MB, and have hole in this region, truncate return -ENOSPC.  I tried
   evade from this problem, but "block allocation" algorithm is tied to
   right value of i_lastfrag, and fix of this corner case may slow down of
   ordinaries scenarios, so this patch makes behavior of "truncate"
   operations similar to what other UFS implementations do.

Signed-off-by: Evgeniy Dushistov <dushistov@mail.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ufs: write to hole in big file
Evgeniy Dushistov [Sun, 27 Aug 2006 08:23:45 +0000 (01:23 -0700)] 
[PATCH] ufs: write to hole in big file

On UFS, this scenario:
open(O_TRUNC)
lseek(1024 * 1024 * 80)
write("A")
lseek(1024 * 2)
write("A")

may cause access to invalid address.

This happened because of "goal" is calculated in wrong way in block
allocation path, as I see this problem exists also in 2.4.

We use construction like this i_data[lastfrag], i_data array of pointers to
direct blocks, indirect and so on, it has ceratain size ~20 elements, and
lastfrag may have value for example 40000.

Also this patch fixes related to handling such scenario issues, wrong
zeroing metadata, in case of block(not fragment) allocation, and wrong goal
calculation, when we allocate block

Signed-off-by: Evgeniy Dushistov <dushistov@mail.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ext3 filesystem bogus ENOSPC with reservation fix
Mingming Cao [Sun, 27 Aug 2006 08:23:44 +0000 (01:23 -0700)] 
[PATCH] ext3 filesystem bogus ENOSPC with reservation fix

To handle the earlier bogus ENOSPC error caused by filesystem full of block
reservation, current code falls back to non block reservation, starts to
allocate block(s) from the goal allocation block group as if there is no
block reservation.

Current code needs to re-load the corresponding block group descriptor for
the initial goal block group in this case.  The patch fixes this.

Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ext2: prevent div-by-zero on corrupted fs
Andries Brouwer [Sun, 27 Aug 2006 08:23:43 +0000 (01:23 -0700)] 
[PATCH] ext2: prevent div-by-zero on corrupted fs

Mounting an ext2 filesystem with zero s_inodes_per_group will cause a
divide error.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Fix for minix crash
Andries Brouwer [Sun, 27 Aug 2006 08:23:42 +0000 (01:23 -0700)] 
[PATCH] Fix for minix crash

Mounting a (corrupt) minix filesystem with zero s_zmap_blocks
gives a spectacular crash on my 2.6.17.8 system, no doubt
because minix/inode.c does an unconditional
minix_set_bit(0,sbi->s_zmap[0]->b_data);

[akpm@osdl.org: make labels conistent while we're there]

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] MTD NAND: Fix ams-delta after core conversion
Jonathan McDowell [Sun, 27 Aug 2006 08:23:41 +0000 (01:23 -0700)] 
[PATCH] MTD NAND: Fix ams-delta after core conversion

The recent hwctrl core conversion for MTD NAND devices broke the Amstrad
Delta driver.  This fixes it up and uses the existing control line defines
rather than unclear magic numbers.

Signed-off-by: Jonathan McDowell <noodles@earth.li>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] futex_find_get_task(): remove an obscure EXIT_ZOMBIE check
Oleg Nesterov [Sun, 27 Aug 2006 08:23:40 +0000 (01:23 -0700)] 
[PATCH] futex_find_get_task(): remove an obscure EXIT_ZOMBIE check

futex_find_get_task:

if (p->state == EXIT_ZOMBIE || p->exit_state == EXIT_ZOMBIE)
return NULL;

I can't understand this.  First, p->state can't be EXIT_ZOMBIE.  The
->exit_state check looks strange too.  Sub-threads or tasks whose ->parent
ignores SIGCHLD go directly to EXIT_DEAD state (I am ignoring a ptrace
case).  Why EXIT_DEAD tasks should be ok?  Yes, EXIT_ZOMBIE is more
important (a task may stay zombie for a long time), but this doesn't mean
we should explicitely ignore other EXIT_XXX states.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Acked-by: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] vcsa attribute bits -> ioctl(VT_GETHIFONTMASK)
Samuel Thibault [Sun, 27 Aug 2006 08:23:39 +0000 (01:23 -0700)] 
[PATCH] vcsa attribute bits -> ioctl(VT_GETHIFONTMASK)

When reading /dev/vcsa while a font with more than 256 characters is
loaded, one of the attribute bits records the 9th bit of the character.
But depending on the console driver (vgacon or fbcon for instance), that's
bit 3 or bit 0.  And there is no way for userland to know that, thus no way
for userland to safely grab the screen content.  So here is a (tested)
patch:

Add a VT_GETHIFONTMASK ioctl for knowing which bit is the 9th bit for VC
text (vc_hi_font_mask field of the vc_data structure).

Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>