GitHub/moto-9609/android_kernel_motorola_exynos9610.git
18 years ago[MMC] sdhci: reset sdhci controller early
Pierre Ossman [Fri, 30 Jun 2006 09:22:30 +0000 (02:22 -0700)]
[MMC] sdhci: reset sdhci controller early

The specification states that the capabilities register might need a reset to
get correct values after boot up.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: check controller version
Pierre Ossman [Fri, 30 Jun 2006 09:22:29 +0000 (02:22 -0700)]
[MMC] sdhci: check controller version

Check the interface version of the controller and bail out if it's an unknown
version.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: check only relevant inhibit bits
Pierre Ossman [Fri, 30 Jun 2006 09:22:28 +0000 (02:22 -0700)]
[MMC] sdhci: check only relevant inhibit bits

Conform to the sdhci specification as to which inhibit bits should be checked
at different times.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: Test for invalid block size
Pierre Ossman [Sun, 2 Jul 2006 15:52:10 +0000 (16:52 +0100)]
[MMC] sdhci: Test for invalid block size

The controller has an upper limit on the block size. Make sure we do not
cross it.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: Avoid sdhci DMA boundaries
Pierre Ossman [Sun, 2 Jul 2006 15:51:35 +0000 (16:51 +0100)]
[MMC] sdhci: Avoid sdhci DMA boundaries

The sdhci controllers will issue an interrupt when a configurable number of
bytes have been transfered using DMA. The purpose is to handle multiple,
scattered memory pages.

Unfortunately, it requires that all transfers are completely aligned to
memory pages, which we cannot guarantee. So we just disable the function.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] Fix sdhci PIO routines
Pierre Ossman [Sun, 2 Jul 2006 15:50:59 +0000 (16:50 +0100)]
[MMC] Fix sdhci PIO routines

The sdhci controllers operate with blocks, not bytes. The PIO routines must
therefore make sure that the minimum unit transfered is a complete block.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: fix interrupt handling
Pierre Ossman [Fri, 30 Jun 2006 09:22:26 +0000 (02:22 -0700)]
[MMC] sdhci: fix interrupt handling

The specification says that interrupts should be cleared before the source is
removed.  We should also not set unknown bits.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: correct register order
Pierre Ossman [Fri, 30 Jun 2006 09:22:25 +0000 (02:22 -0700)]
[MMC] sdhci: correct register order

The sdhci specification states that some registers must be written to in a
specific order.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: proper timeout handling
Pierre Ossman [Fri, 30 Jun 2006 09:22:25 +0000 (02:22 -0700)]
[MMC] sdhci: proper timeout handling

Use the give timeout clock and calculate a proper timeout instead of using the
maximum at all times.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: fix sdhci reset timeout
Pierre Ossman [Fri, 30 Jun 2006 09:22:24 +0000 (02:22 -0700)]
[MMC] sdhci: fix sdhci reset timeout

The reset register is automatically cleared when the reset has completed.
Hence, we should busy wait and not have a fixed delay.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: fix timeout loops in sdhci
Pierre Ossman [Fri, 30 Jun 2006 09:22:23 +0000 (02:22 -0700)]
[MMC] sdhci: fix timeout loops in sdhci

The current timeout loop assume that jiffies are updated.  This might not be
the case depending on locks and if the kernel is compiled without preemption.
Change the system to use a counter and fixed delays.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: support for multiple voltages
Pierre Ossman [Fri, 30 Jun 2006 09:22:23 +0000 (02:22 -0700)]
[MMC] sdhci: support for multiple voltages

The sdhci controllers can support up to three voltage levels.  Detect which
and report back to the MMC layer.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: print device id
Pierre Ossman [Fri, 30 Jun 2006 09:22:22 +0000 (02:22 -0700)]
[MMC] sdhci: print device id

As sdhci is a generic driver, it is helpful to see some more specific
identification of the actual hardware in dmesg.  PCI vendor, device and
revision is sufficient in most cases.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[MMC] sdhci: check SDHCI base clock
Pierre Ossman [Fri, 30 Jun 2006 09:22:21 +0000 (02:22 -0700)]
[MMC] sdhci: check SDHCI base clock

A base clock value of 0 means that the driver must get the base clock through
some other means.  As we have no other way of getting it, we must abort.

Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[PATCH] powermac: Fix some 64b resource damage
Benjamin Herrenschmidt [Sun, 2 Jul 2006 01:55:03 +0000 (11:55 +1000)]
[PATCH] powermac: Fix some 64b resource damage

The 64 bits resource patches did a bit of damage on PowerMac causing a
buffer overflow in macio_asic and a warning in a sound driver. The
former is fixed by reverting the sprintf of the bus_id to %08x as it was
before. The bus_id used for macio devices is always a 32 bits value
(macio always sits in 32 bits space) and since it's exposed to userland,
the format of the string shouldn't be changed like that anyway. The
second by using the proper type for printk.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge branch 'audit.b22' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/audit...
Linus Torvalds [Sat, 1 Jul 2006 16:59:08 +0000 (09:59 -0700)]
Merge branch 'audit.b22' of git://git./linux/kernel/git/viro/audit-current

* 'audit.b22' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/audit-current:
  [PATCH] audit syscall classes
  [PATCH] audit: support for object context filters
  [PATCH] audit: rename AUDIT_SE_* constants
  [PATCH] add rule filterkey

18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild
Linus Torvalds [Sat, 1 Jul 2006 16:58:43 +0000 (09:58 -0700)]
Merge git://git./linux/kernel/git/sam/kbuild

* git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild:
  kbuild: documentation change on allowing checkers besides sparse
  kbuild: warn when a moduled uses a symbol marked UNUSED
  kbuild: fix segv in modpost
  kconfig: enhancing accessibility of lxdialog
  kbuild: fix ia64 breakage after introducing make -rR

18 years ago[PATCH] pnp: suppress request_irq() warning
Andrew Morton [Sat, 1 Jul 2006 11:36:37 +0000 (04:36 -0700)]
[PATCH] pnp: suppress request_irq() warning

Suppress the "setup_irq: irq handler mismatch" coming out of pnp_check_irq():
failures are expected here.

Cc: <stable@kernel.org>
Cc: Santiago Garcia Mantinan <manty@manty.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] drivers/block/nbd.c compile fix
Ingo van Lil [Sat, 1 Jul 2006 11:36:36 +0000 (04:36 -0700)]
[PATCH] drivers/block/nbd.c compile fix

The Network Block Device driver doesn't compile if NDEBUG is defined.

Signed-off-by: Ingo van Lil <inguin@gmx.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] kernel-doc: make man/text mode function output same
Randy Dunlap [Sat, 1 Jul 2006 11:36:36 +0000 (04:36 -0700)]
[PATCH] kernel-doc: make man/text mode function output same

Make output of function descriptions in text mode match contents of 'man'
mode by adding Name: plus function-short-description ("purpose") and
changing Function: to Synopsis:.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Fix problem with ATAPI DMA on IT8212 in Linux
Alan Cox [Sat, 1 Jul 2006 11:36:34 +0000 (04:36 -0700)]
[PATCH] Fix problem with ATAPI DMA on IT8212 in Linux

Missing variable initialisation would mean it would sometimes not put ATAPI
devices into DMA by default.

Signed-off-by: Alan Cox <alan@redhat.com>
Cc: <Jack.Lee@ite.com.tw>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] kernel-doc: consistent text/man mode output
Randy.Dunlap [Sat, 1 Jul 2006 11:36:34 +0000 (04:36 -0700)]
[PATCH] kernel-doc: consistent text/man mode output

Add a space between data type and struct field name in man-mode
bitfield struct output so that they don't run together.

For text-mode struct output, print the struct 'purpose' or
short description (as done in man-mode output).

For text-mode enum output, print the enum 'purpose' or
short description (as done in man-mode output).

For text-mode typedef output, print the typedef 'purpose' or
short description (as done in man-mode output).

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] reiserfs: update ctime and mtime on expanding truncate
Vladimir Saveliev [Sat, 1 Jul 2006 11:36:32 +0000 (04:36 -0700)]
[PATCH] reiserfs: update ctime and mtime on expanding truncate

Reiserfs does not update ctime and mtime on expanding truncate via
truncate().  This patch fixes it.

Signed-off-by: Vladimir Saveliev <vs@namesys.com>
Cc: Hans Reiser <reiser@namesys.com>
Cc: Michael Kerrisk <mtk-manpages@gmx.net>
Cc: Chris Mason <mason@suse.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] s390: put sys_call_table into .rodata section and write protect it
Heiko Carstens [Sat, 1 Jul 2006 11:36:31 +0000 (04:36 -0700)]
[PATCH] s390: put sys_call_table into .rodata section and write protect it

Put s390's syscall tables into .rodata section and write protect this
section to prevent misuse of it.  Suggested by Arjan van de Ven
<arjan@infradead.org>.

Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] add __[start|end]_rodata sections to asm-generic/sections.h
Heiko Carstens [Sat, 1 Jul 2006 11:36:30 +0000 (04:36 -0700)]
[PATCH] add __[start|end]_rodata sections to asm-generic/sections.h

Add __start_rodata and __end_rodata to sections.h to avoid extern
declarations.  Needed by s390 code (see following patch).

[akpm@osdl.org: update architectures]
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Andi Kleen <ak@muc.de>
Acked-by: Kyle McMartin <kyle@mcmartin.ca>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] hotcpu_notifier-fixes
Andrew Morton [Sat, 1 Jul 2006 11:36:30 +0000 (04:36 -0700)]
[PATCH] hotcpu_notifier-fixes

Always use do {} while (0).  Failing to do so can cause subtle compile
failures or bugs.

Cc: Chandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC: class driver for Samsung S3C series SoC
Ben Dooks [Sat, 1 Jul 2006 11:36:26 +0000 (04:36 -0700)]
[PATCH] RTC: class driver for Samsung S3C series SoC

This is a renamed and tested version of the previous S3C24XX RTC class
driver.

The driver has been renamed from the original s3c2410-rtc, which is now too
narrow for the range of devices supported.

The rtc-s3c has been chosen as there is the distinct possibility of this
driver being carried forward into newer Samsung SoC silicon.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Cc: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Fix IS_ERR Threshold Value
Ralf Baechle [Sat, 1 Jul 2006 11:36:25 +0000 (04:36 -0700)]
[PATCH] Fix IS_ERR Threshold Value

 o Raise the maximum error number in IS_ERR_VALUE to 4095.
 o Make that number available as a new constant MAX_ERRNO.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ufs: truncate should allocate block for last byte
Evgeniy Dushistov [Sat, 1 Jul 2006 11:36:24 +0000 (04:36 -0700)]
[PATCH] ufs: truncate should allocate block for last byte

This patch fixes buggy behaviour of UFS
in such kind of scenario:
open(, O_TRUNC...)
ftruncate(, 1024)
ftruncate(, 0)

Such a scenario causes ufs_panic and remount read-only.  This happen
because of according to specification UFS should always allocate block for
last byte, and many parts of our implementation rely on this, but
`ufs_truncate' doesn't care about this.

To make possible return error code and to know about old size, this patch
removes `truncate' from ufs inode_operations and uses `setattr' method to
call ufs_truncate.

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] uml: rename and improve actually_do_remove()
Paolo 'Blaisorblade' Giarrusso [Sat, 1 Jul 2006 11:36:24 +0000 (04:36 -0700)]
[PATCH] uml: rename and improve actually_do_remove()

Rename actually_do_remove() to remove_files_and_dir(), make it call
closedir(), make it ignore ENOENT (I see it frequently enough).

ENOENT is probably due to multiple threads calling the exitcall functions
together*, but fixing that is non-trivial; and ignoring it is perfectly ok
in any case.

* it can surely happen: last_ditch_exit() is installed as SIGTERM handler
  at boot, and it's not removed on thread creation.  So killall vmlinux
  (which I do) surely causes that.  I've seen also a crash which seems to
  do the same.

Installing the handler on only the main thread would make UML do no cleanup
when another thread exits, and we're not sure we want that.  And mutual
exclusion in that context is tricky - we can't use spinlock in code not on
a kernel stack (spinlock debugging uses "current" a lot).

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Cc: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: fix not_dead_yet when directory is in bad state
Paolo 'Blaisorblade' Giarrusso [Sat, 1 Jul 2006 11:36:23 +0000 (04:36 -0700)]
[PATCH] uml: fix not_dead_yet when directory is in bad state

The bug occurred to me when a UML left an empty ~/.uml/Sarge-norm folder -
when trying to reuse not_dead_yet() failed one of its check.  The comment
says that's ok and means that we can take the directory, but while normally
not_dead_yet() removes it and returns 0 (i.e.  go on, use this), on failure
it returns 0 but forgets to remove it.  The fix is to remove it anytime
we're going to return 0.

But since "not_dead_yet" didn't make the interface so clear, causing this
bug, and I couldn't find a convenient name for the mix of things it did, I
split it into two parts:

is_umdir_used()      - returns a boolean, contains all checks of not_dead_yet()
umdir_take_if_dead   - tries to remove the dir unless it's used - returns
whether it removed it, that is we now own it.

With this changes the control flow is IMHO a bit clearer and needs less
comment for control flow.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Cc: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: make copy_*_user atomic
Paolo 'Blaisorblade' Giarrusso [Sat, 1 Jul 2006 11:36:19 +0000 (04:36 -0700)]
[PATCH] uml: make copy_*_user atomic

Make __copy_*_user_inatomic really atomic to avoid "Sleeping function called in
atomic context" warnings, especially from futex code.

This is made by adding another kmap_atomic slot and making copy_*_user_skas
use kmap_atomic; also copy_*_user() becomes atomic, but that's true and is not
a problem for i386 (and we can always add might_sleep there as done
elsewhere).  For TT mode kmap is not used, so there's no need for this.

I've had to use another slot since both KM_USER0 and KM_USER1 are used
elsewhere and could cause conflicts.  Till now we reused the kmap_atomic slot
list from the subarch, but that's not needed as that list must contain the
common ones (used by generic code) + the ones used in architecture specific
code (and Uml till now used none); so I've taken the i386 one after comparing
it with ones from other archs, and added KM_UML_USERCOPY.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Cc: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] SMP alternatives: skip with UP kernels
Gerd Hoffmann [Sat, 1 Jul 2006 11:36:18 +0000 (04:36 -0700)]
[PATCH] SMP alternatives: skip with UP kernels

Hide the magic in alternative.h and provide some dummy inline functions
for the UP case (gcc should manage to optimize away these calls).  No
changes in module.c.

Cc: Dave Jones <davej@codemonkey.org.uk>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: namespace cleanup: replace ips with ipath
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:17 +0000 (04:36 -0700)]
[PATCH] IB/ipath: namespace cleanup: replace ips with ipath

Remove ips namespace from infinipath drivers.  This renames ips_common.h to
ipath_common.h.  Definitions, data structures, etc.  that were not used by
kernel modules have moved to user-only headers.  All names including ips have
been renamed to ipath.  Some names have had an ipath prefix added.

Signed-off-by: Christian Bell <christian.bell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: ignore receive queue size if SRQ is specified
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:16 +0000 (04:36 -0700)]
[PATCH] IB/ipath: ignore receive queue size if SRQ is specified

The receive work queue size should be ignored if the QP is created to use a
shared receive queue according to the IB spec.

Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: remove some #if 0 code related to lockable memory
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:15 +0000 (04:36 -0700)]
[PATCH] IB/ipath: remove some #if 0 code related to lockable memory

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: fix a bug that results in addresses near 0 being written via DMA
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:15 +0000 (04:36 -0700)]
[PATCH] IB/ipath: fix a bug that results in addresses near 0 being written via DMA

We can't tell for sure if any packets are in the infinipath receive buffer
when we shut down a chip port.  Normally this is taken care of by orderly
shutdown, but when processes are terminated, or sending process has a bug, we
can continue to receive packets.  So rather than writing zero to the address
registers for the closing port, we point it at a dummy memory.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: read/write correct sizes through diag interface
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:14 +0000 (04:36 -0700)]
[PATCH] IB/ipath: read/write correct sizes through diag interface

We must increment uaddr by size we are reading or writing, since it's passed
as a char *, not a pointer to the appropriate size.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: support more models of InfiniPath hardware
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:13 +0000 (04:36 -0700)]
[PATCH] IB/ipath: support more models of InfiniPath hardware

We do a few more explicit checks for specific models, and now also support the
old PathScale serial number style, or new QLogic style.

This is backwards compatible with previous versions of software and hardware.
That is, older software will see a plausible serial number and correct GUID
when used with a new board, while newer software will correctly handle an
older board.

Signed-off-by: Mike Albaugh <mike.albaugh@qlogic.com>
Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: drop the "stats" sysfs attribute group
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:12 +0000 (04:36 -0700)]
[PATCH] IB/ipath: drop the "stats" sysfs attribute group

This attribute group made it into the original driver, but should not have.

Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: purge sps_lid and sps_mlid arrays
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:11 +0000 (04:36 -0700)]
[PATCH] IB/ipath: purge sps_lid and sps_mlid arrays

The two arrays only had space for 4 units.

Also changed from ipath_set_sps_lid() to ipath_set_lid(); the sps was
leftover.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: rC receive interrupt performance changes
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:10 +0000 (04:36 -0700)]
[PATCH] IB/ipath: rC receive interrupt performance changes

This patch separates QP state used for sending and receiving RC packets so the
processing in the receive interrupt handler can be done mostly without locks
being held.  ACK packets are now sent without requiring synchronization with
the send tasklet.

Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: fixes to performance get counters for IB compliance
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:09 +0000 (04:36 -0700)]
[PATCH] IB/ipath: fixes to performance get counters for IB compliance

This patch fixes some problems uncovered during IB compliance testing to
return the right values for error counters returned by the Performance Get
Counters packet.

Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: check for valid LID and multicast LIDs
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:08 +0000 (04:36 -0700)]
[PATCH] IB/ipath: check for valid LID and multicast LIDs

Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: removed redundant statements
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:08 +0000 (04:36 -0700)]
[PATCH] IB/ipath: removed redundant statements

The tail register read became redundant as the result of earlier receive
interrupt bug fixes.

Drop another unneeded register read.

And another line that got duplicated.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: don't confuse the max message size with the MTU
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:07 +0000 (04:36 -0700)]
[PATCH] IB/ipath: don't confuse the max message size with the MTU

Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: disallow send of invalid packet sizes over UD
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:06 +0000 (04:36 -0700)]
[PATCH] IB/ipath: disallow send of invalid packet sizes over UD

Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: fix lost interrupts on HT-400
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:05 +0000 (04:36 -0700)]
[PATCH] IB/ipath: fix lost interrupts on HT-400

Do an extra check to see if in-memory tail changed while processing packets,
and if so, going back through the loop again (but only once per call to
ipath_kreceive()).  In practice, this seems to be enough to guarantee that if
we crossed the clearing of an interrupt at start of ipath_intr with a
scheduled tail register update, that we'll process the "extra" packet that
lost the interrupt because we cleared it just as it was about to arrive.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: fixed bug 9776 for real
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:05 +0000 (04:36 -0700)]
[PATCH] IB/ipath: fixed bug 9776 for real

The problem was that I was updating the head register multiple times in the
rcvhdrq processing loop, and setting the counter on each update.  Since that
meant that the tail register was ahead of head for all but the last update, we
would get extra interrupts.  The fix was to not write the counter value except
on the last update.

I also changed to update rcvhdrhead and rcvegrindexhead at most every 16
packets, if there were lots of packets in the queue (and of course, on the
last packet, regardless).

I also made some small cleanups while debugging this.

With these changes, xeon/monty typically sees two openib packets per interrupt
on sdp and ipoib, opteron/monty is about 1.25 pkts/intr.

I'm seeing about 3800 Mbit/s monty/xeon, and 5000-5100 opteron/monty with
netperf sdp.  Netpipe doesn't show as good as that, peaking at about 4400 on
opteron/monty sdp.  Plain ipoib xeon is about 2100+ netperf, opteron 2900+, at
128KB

Signed-off-by: olson@eng-12.pathscale.com
Signed-off-by: Bryan O'Sullivan <bos@pathscale.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: reduce overhead on receive interrupts
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:04 +0000 (04:36 -0700)]
[PATCH] IB/ipath: reduce overhead on receive interrupts

Also count the number of interrupts where that works (fastrcvint).  On any
interrupt where the port0 head and tail registers are not equal, just call the
ipath_kreceive code without reading the interrupt status, thus saving the
approximately 0.25usec processor stall waiting for the read to return.  If any
other interrupt bits are set, or head==tail, take the normal path, but that
has been reordered to handle read ahead of pioavail.  Also no longer call
ipath_kreceive() from ipath_qcheck(), because that just seems to make things
worse, and isn't really buying us anything, these days.

Also no longer loop in ipath_kreceive(); better to not hold things off too
long (I saw many cases where we would loop 4-8 times, and handle thousands (up
to 3500) in a single call).

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: memory management cleanups
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:03 +0000 (04:36 -0700)]
[PATCH] IB/ipath: memory management cleanups

Made in-memory rcvhdrq tail update be in dma_alloc'ed memory, not random user
or special kernel (needed for ppc, also "just the right thing to do").

Some cleanups to make unexpected link transitions less likely to produce
complaints about packet errors, and also to not leave SMA packets stuck and
unable to go out.

A few other random debug and comment cleanups.

Always init rcvhdrq head/tail registers to 0, to avoid race conditions (should
have been that way some time ago).

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: use vmalloc to allocate struct ipath_devdata
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:02 +0000 (04:36 -0700)]
[PATCH] IB/ipath: use vmalloc to allocate struct ipath_devdata

This is not a DMA target, so no need to use dma_alloc_coherent on it.

Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: use more appropriate gfp flags
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:01 +0000 (04:36 -0700)]
[PATCH] IB/ipath: use more appropriate gfp flags

This helps us to survive better when memory is fragmented.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: enable freeze mode when shutting down device
Bryan O'Sullivan [Sat, 1 Jul 2006 11:36:00 +0000 (04:36 -0700)]
[PATCH] IB/ipath: enable freeze mode when shutting down device

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: print better debug info when handling 32/64-bit DMA mask problems
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:59 +0000 (04:35 -0700)]
[PATCH] IB/ipath: print better debug info when handling 32/64-bit DMA mask problems

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: removed unused field ipath_kregvirt from struct ipath_devdata
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:59 +0000 (04:35 -0700)]
[PATCH] IB/ipath: removed unused field ipath_kregvirt from struct ipath_devdata

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: enforce device resource limits
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:58 +0000 (04:35 -0700)]
[PATCH] IB/ipath: enforce device resource limits

These limits are somewhat artificial in that we don't actually have any
device limits.  However, the verbs layer expects that such limits exist
and are enforced, so we make up arbitrary (but sensible) limits.

Signed-off-by: Robert Walsh <robert.walsh@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: report correct device identification information in /sys
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:57 +0000 (04:35 -0700)]
[PATCH] IB/ipath: report correct device identification information in /sys

Signed-off-by: Robert Walsh <robert.walsh@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: return an error for unknown multicast GID
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:56 +0000 (04:35 -0700)]
[PATCH] IB/ipath: return an error for unknown multicast GID

Signed-off-by: Robert Walsh <robert.walsh@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: fix some memory leaks on failure paths
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:55 +0000 (04:35 -0700)]
[PATCH] IB/ipath: fix some memory leaks on failure paths

Signed-off-by: Robert Walsh <robert.walsh@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: don't allow resources to be created with illegal values
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:55 +0000 (04:35 -0700)]
[PATCH] IB/ipath: don't allow resources to be created with illegal values

Signed-off-by: Robert Walsh <robert.walsh@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: remove some duplicate code
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:54 +0000 (04:35 -0700)]
[PATCH] IB/ipath: remove some duplicate code

Signed-off-by: Robert Walsh <robert.walsh@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: update some comments and fix typos
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:53 +0000 (04:35 -0700)]
[PATCH] IB/ipath: update some comments and fix typos

Signed-off-by: Robert Walsh <robert.walsh@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: allow diags on any unit
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:52 +0000 (04:35 -0700)]
[PATCH] IB/ipath: allow diags on any unit

There is no longer a /dev/ipath_diag file; instead, there's
/dev/ipath_diag0, 1, etc.

It's still not possible to have diags run on more than one unit at a time,
but that's easy to fix at some point.

Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: fix shared receive queues for RC
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:51 +0000 (04:35 -0700)]
[PATCH] IB/ipath: fix shared receive queues for RC

Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: fix an indenting problem
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:51 +0000 (04:35 -0700)]
[PATCH] IB/ipath: fix an indenting problem

Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: share more common code between RC and UC protocols
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:50 +0000 (04:35 -0700)]
[PATCH] IB/ipath: share more common code between RC and UC protocols

Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: update copyrights and other strings to reflect new company name
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:49 +0000 (04:35 -0700)]
[PATCH] IB/ipath: update copyrights and other strings to reflect new company name

Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IB/ipath: name zero counter offsets so it's clear they aren't counters
Bryan O'Sullivan [Sat, 1 Jul 2006 11:35:48 +0000 (04:35 -0700)]
[PATCH] IB/ipath: name zero counter offsets so it's clear they aren't counters

Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Cc: "Michael S. Tsirkin" <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ACPI: fix not power-manageable device message
Jae-hyeon Park [Sat, 1 Jul 2006 11:35:47 +0000 (04:35 -0700)]
[PATCH] ACPI: fix not power-manageable device message

Fix typo in message.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] EDAC: fix module names quoted in sysfs
Doug Thompson [Sat, 1 Jul 2006 11:35:45 +0000 (04:35 -0700)]
[PATCH] EDAC: fix module names quoted in sysfs

Fix the quoted module name in the sysfs for EDAC modules and reported by several
people.

Instead of  ../_edac_e752x_/   now the following will be presented, like other
modules:   ../edac_e752x/

Signed-off-by: Doug Thompson <norsk5@xmission.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IRQ: warning message cleanup
Bjorn Helgaas [Sat, 1 Jul 2006 11:35:45 +0000 (04:35 -0700)]
[PATCH] IRQ: warning message cleanup

Make warnings more consistent.

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-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] IRQ: Use SA_PERCPU_IRQ, not IRQ_PER_CPU, for irqaction.flags
Bjorn Helgaas [Sat, 1 Jul 2006 11:35:44 +0000 (04:35 -0700)]
[PATCH] IRQ: Use SA_PERCPU_IRQ, not IRQ_PER_CPU, for irqaction.flags

IRQ_PER_CPU is a bit in the struct irq_desc "status" field, not in the
struct irqaction "flags", so the previous code checked the wrong bit.

SA_PERCPU_IRQ is only used by drivers/char/mmtimer.c for SGI ia64 boxes.

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Acked-by: 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] pi-futex: futex_wake() lockup fix
Ingo Molnar [Sat, 1 Jul 2006 11:35:46 +0000 (04:35 -0700)]
[PATCH] pi-futex: futex_wake() lockup fix

Fix futex_wake() exit condition bug when handling the robust-list with PI
futexes on them.

(reported by Ulrich Drepper, debugged by the lock validator.)

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Ulrich Drepper <drepper@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] pi-futex: fix mm_struct memory leak
Vernon Mauery [Sat, 1 Jul 2006 11:35:42 +0000 (04:35 -0700)]
[PATCH] pi-futex: fix mm_struct memory leak

lock_queue was getting called essentially twice in a row and was
continually incrementing the mm_count ref count, thus causing a memory
leak.

Dinakar Guniguntala provided a proper fix for the problem that simply grabs
the spinlock for the hash bucket queue rather than calling lock_queue.

The second time we do a queue_lock in futex_lock_pi, we really only need to
take the hash bucket lock.

Signed-off-by: Dinakar Guniguntala <dino@in.ibm.com>
Signed-off-by: Vernon Mauery <vernux@us.ibm.com>
Acked-by: Paul E. McKenney <paulmck@us.ibm.com>
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 agokbuild: documentation change on allowing checkers besides sparse
Dustin Kirkland [Tue, 23 May 2006 20:57:23 +0000 (15:57 -0500)]
kbuild: documentation change on allowing checkers besides sparse

Minor documentation change on allowing checkers besides sparse

This patch cleans up a couple of mentions of sparse in the inline
toplevel Makefile documentation such that it's clear that other checkers
besides sparse can override CHECK and CHECKFLAGS.

Signed-off-by: Dustin Kirkland <dustin.kirkland@us.ibm.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
18 years ago[PATCH] audit syscall classes
Al Viro [Sat, 1 Jul 2006 07:56:16 +0000 (03:56 -0400)]
[PATCH] audit syscall classes

Allow to tie upper bits of syscall bitmap in audit rules to kernel-defined
sets of syscalls.  Infrastructure, a couple of classes (with 32bit counterparts
for biarch targets) and actual tie-in on i386, amd64 and ia64.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
18 years agokbuild: warn when a moduled uses a symbol marked UNUSED
Sam Ravnborg [Sat, 1 Jul 2006 09:44:23 +0000 (11:44 +0200)]
kbuild: warn when a moduled uses a symbol marked UNUSED

We now have infrastructure in place to mark an EXPORTed symbol
as unused. So the natural next step is to warn during buildtime when
a module uses a symbol marked UNUSED.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
18 years ago[PATCH] audit: support for object context filters
Darrel Goeddel [Thu, 29 Jun 2006 21:57:08 +0000 (16:57 -0500)]
[PATCH] audit: support for object context filters

This patch introduces object audit filters based on the elements
of the SELinux context.

Signed-off-by: Darrel Goeddel <dgoeddel@trustedcs.com>
Acked-by: Stephen Smalley <sds@tycho.nsa.gov>
 kernel/auditfilter.c           |   25 +++++++++++++++++++++++++
 kernel/auditsc.c               |   40 ++++++++++++++++++++++++++++++++++++++++
 security/selinux/ss/services.c |   18 +++++++++++++++++-
 3 files changed, 82 insertions(+), 1 deletion(-)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
18 years ago[PATCH] audit: rename AUDIT_SE_* constants
Darrel Goeddel [Thu, 29 Jun 2006 21:56:39 +0000 (16:56 -0500)]
[PATCH] audit: rename AUDIT_SE_* constants

This patch renames some audit constant definitions and adds
additional definitions used by the following patch.  The renaming
avoids ambiguity with respect to the new definitions.

Signed-off-by: Darrel Goeddel <dgoeddel@trustedcs.com>
 include/linux/audit.h          |   15 ++++++++----
 kernel/auditfilter.c           |   50 ++++++++++++++++++++---------------------
 kernel/auditsc.c               |   10 ++++----
 security/selinux/ss/services.c |   32 +++++++++++++-------------
 4 files changed, 56 insertions(+), 51 deletions(-)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
18 years ago[PATCH] add rule filterkey
Amy Griffis [Wed, 14 Jun 2006 22:45:21 +0000 (18:45 -0400)]
[PATCH] add rule filterkey

Add support for a rule key, which can be used to tie audit records to audit
rules.  This is useful when a watched file is accessed through a link or
symlink, as well as for general audit log analysis.

Because this patch uses a string key instead of an integer key, there is a bit
of extra overhead to do the kstrdup() when a rule fires.  However, we're also
allocating memory for the audit record buffer, so it's probably not that
significant.  I went ahead with a string key because it seems more
user-friendly.

Note that the user must ensure that filterkeys are unique.  The kernel only
checks for duplicate rules.

Signed-off-by: Amy Griffis <amy.griffis@hpd.com>
18 years agokbuild: fix segv in modpost
Sam Ravnborg [Sat, 1 Jul 2006 08:10:19 +0000 (10:10 +0200)]
kbuild: fix segv in modpost

Parsing an old Modules.symvers file casued modpost to SEGV.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
18 years agokconfig: enhancing accessibility of lxdialog
Samuel Thibault [Fri, 30 Jun 2006 09:18:50 +0000 (02:18 -0700)]
kconfig: enhancing accessibility of lxdialog

Some fix that I forgot for good accessibility of lxdialog (the cursor
should always be left at the focus location):

Have the checklist display the currently highlighted entry last, for having
the cursor left on it (rather than on the last line of the list).

Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
18 years agokbuild: fix ia64 breakage after introducing make -rR
Sam Ravnborg [Sat, 1 Jul 2006 07:58:02 +0000 (09:58 +0200)]
kbuild: fix ia64 breakage after introducing make -rR

kbuild used $ยค(*F to get filename of target without extension.
This was used in several places all over kbuild, but introducing
make -rR broke his for all cases where we specified full path to
target/prerequsite. It is assumed that make -rR disables old style
suffix-rules which is why is suddenly failed.

ia64 was impacted by this change because several div* routines in
arch/ia64/lib are build using explicit paths and then kbuild failed.

Thanks to David Mosberger-Tang <David.Mosberger@acm.org> for an explanation
what was the root-cause and for testing on ia64.

This patch also fixes two uses of $(*F) in arch/um

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Fri, 30 Jun 2006 22:40:35 +0000 (15:40 -0700)]
Merge /pub/scm/linux/kernel/git/davem/sparc-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SPARC64]: Kill sun4v virtual device layer.
  [SERIAL] sunhv: Convert to of_driver layer.
  [SPARC64]: Mask out top 8-bits in physical address when building resources.
  [SERIAL] sunsu: Missing return statement in su_probe().

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Fri, 30 Jun 2006 22:40:17 +0000 (15:40 -0700)]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [IPV6]: Added GSO support for TCPv6
  [NET]: Generalise TSO-specific bits from skb_setup_caps
  [IPV6]: Added GSO support for TCPv6
  [IPV6]: Remove redundant length check on input
  [NETFILTER]: SCTP conntrack: fix crash triggered by packet without chunks
  [TG3]: Update version and reldate
  [TG3]: Add TSO workaround using GSO
  [TG3]: Turn on hw fix for ASF problems
  [TG3]: Add rx BD workaround
  [TG3]: Add tg3_netif_stop() in vlan functions
  [TCP]: Reset gso_segs if packet is dodgy

18 years agoMerge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband
Linus Torvalds [Fri, 30 Jun 2006 22:39:49 +0000 (15:39 -0700)]
Merge branch 'for-linus' of /linux/kernel/git/roland/infiniband

* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:
  IB/core: Set alternate port number when initializing QP attributes
  IB/uverbs: Set correct user handle for user SRQs

18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/bunk/trivial
Linus Torvalds [Fri, 30 Jun 2006 22:39:30 +0000 (15:39 -0700)]
Merge git://git./linux/kernel/git/bunk/trivial

* git://git.kernel.org/pub/scm/linux/kernel/git/bunk/trivial:
  Remove obsolete #include <linux/config.h>
  remove obsolete swsusp_encrypt
  arch/arm26/Kconfig typos
  Documentation/IPMI typos
  Kconfig: Typos in net/sched/Kconfig
  v9fs: do not include linux/version.h
  Documentation/DocBook/mtdnand.tmpl: typo fixes
  typo fixes: specfic -> specific
  typo fixes in Documentation/networking/pktgen.txt
  typo fixes: occuring -> occurring
  typo fixes: infomation -> information
  typo fixes: disadvantadge -> disadvantage
  typo fixes: aquire -> acquire
  typo fixes: mecanism -> mechanism
  typo fixes: bandwith -> bandwidth
  fix a typo in the RTC_CLASS help text
  smb is no longer maintained

Manually merged trivial conflict in arch/um/kernel/vmlinux.lds.S

18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6/
Linus Torvalds [Fri, 30 Jun 2006 22:36:52 +0000 (15:36 -0700)]
Merge git://git./linux/kernel/git/brodo/pcmcia-2.6/

* git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6/:
  [PATCH] pcmcia: fix deadlock in pcmcia_parse_events
  [PATCH] com20020_cs: more device support
  [PATCH] au1xxx: pcmcia: fix __init called from non-init
  [PATCH] kill open-coded offsetof in cm4000_cs.c ZERO_DEV()
  [PATCH] pcmcia: convert pcmcia_cs to kthread
  [PATCH] pcmcia: fix kernel-doc function name
  [PATCH] pcmcia: hostap_cs.c - 0xc00f,0x0000 conflicts with pcnet_cs
  [PATCH] pcmcia: at91_cf suspend/resume/wakeup
  [PATCH] pcmcia: Make ide_cs work with the memory space of CF-Cards if IO space is not available
  [PATCH] pcmcia: TI PCIxx12 CardBus controller support
  [PATCH] pcmcia: warn if driver requests exclusive, but gets a shared IRQ
  [PATCH] pcmcia: expose tool in pcmcia/Documentation/pcmcia/
  [PATCH] pcmcia: another ID for serial_cs.c
  [PATCH] yenta: fix hidden PCI bus numbers
  [PATCH] yenta: do power-up only after socket is configured

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/mchehab/v4l-dvb
Linus Torvalds [Fri, 30 Jun 2006 22:36:19 +0000 (15:36 -0700)]
Merge /pub/scm/linux/kernel/git/mchehab/v4l-dvb

* master.kernel.org:/pub/scm/linux/kernel/git/mchehab/v4l-dvb:
  V4L/DVB (4290): Add support for the TCL M2523_3DB_E tuner.
  V4L/DVB (4289): Missing statement in drivers/media/dvb/frontends/cx22700.c
  V4L/DVB (4288): Clean out a zillion sparse warnings in pvrusb2
  V4L/DVB (4287): Pvrusb2/: possible cleanups
  V4L/DVB (4285): Cx88: add support for Geniatech Digistar / Digiwave 103g
  V4L/DVB (4284): Cx24123: fix set_voltage function according to the specs
  V4L/DVB (4282): Fix: use swzigzag for swalgo
  V4L/DVB (4281): TDA9887_SET_CONFIG should only be handled by the tda9887.
  V4L/DVB (4277): Fix CI interface on PRO KNC1 cards
  V4L/DVB (4276): Fix CI on old KNC1 DVBC cards
  V4L/DVB (4275): The FE_SET_FRONTEND_TUNE_MODE ioctl always returns EOPNOTSUPP
  V4L/DVB (4274): Eliminate use of tda9887 from pvrusb2 driver
  V4L/DVB (4273): Always log pvrusb2 device register / unregister events
  V4L/DVB (4272): Fix tveeprom supported standards
  V4L/DVB (4270): Add tda9887-specific tuner configuration
  V4L/DVB (4269): Subject: videocodec: make 1-bit fields unsigned
  V4L/DVB (4267): Remove all instances of request_module("tda9887")
  V4L/DVB (4264): Cx88-blackbird: implement VIDIOC_QUERYCTRL and VIDIOC_QUERYMENU

18 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux...
Linus Torvalds [Fri, 30 Jun 2006 22:34:15 +0000 (15:34 -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: (25 commits)
  ACPI: Kconfig: ACPI_SRAT depends on ACPI
  ACPI: drivers/acpi/scan.c: make acpi_bus_type static
  ACPI: fixup memhotplug debug message
  ACPI: ACPICA 20060623
  ACPI: C-States: only demote on current bus mastering activity
  ACPI: C-States: bm_activity improvements
  ACPI: C-States: accounting of sleep states
  ACPI: additional blacklist entry for ThinkPad R40e
  ACPI: restore comment justifying 'extra' P_LVLx access
  ACPI: fix battery on HP NX6125
  ACPIPHP: prevent duplicate slot numbers when no _SUN
  ACPI: static-ize handle_hotplug_event_func()
  ACPIPHP: use ACPI dock driver
  ACPI: dock driver
  KEVENT: add new uevent for dock
  ACPI: asus_acpi_init: propagate correct return value
  [ACPI] Print error message if remove/install notify handler fails
  ACPI: delete tracing macros from drivers/acpi/*.c
  ACPI: HW P-state coordination support
  ACPI: un-export ACPI_ERROR() -- use printk(KERN_ERR...)
  ...

18 years ago[SPARC64]: Kill sun4v virtual device layer.
David S. Miller [Fri, 30 Jun 2006 07:07:40 +0000 (00:07 -0700)]
[SPARC64]: Kill sun4v virtual device layer.

Replace with a simple IRQ translater in the PROM
device tree builder.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SERIAL] sunhv: Convert to of_driver layer.
David S. Miller [Fri, 30 Jun 2006 07:00:22 +0000 (00:00 -0700)]
[SERIAL] sunhv: Convert to of_driver layer.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Mask out top 8-bits in physical address when building resources.
David S. Miller [Fri, 30 Jun 2006 02:58:28 +0000 (19:58 -0700)]
[SPARC64]: Mask out top 8-bits in physical address when building resources.

These top 8-bits are supposed to be ignored in the ranges and
top-level reg properties on this platform.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SERIAL] sunsu: Missing return statement in su_probe().
David S. Miller [Fri, 30 Jun 2006 02:40:19 +0000 (19:40 -0700)]
[SERIAL] sunsu: Missing return statement in su_probe().

If we have a keyboard/mouse port, don't drop through to
calling sunsu_autoconfig().

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPV6]: Added GSO support for TCPv6
Herbert Xu [Fri, 30 Jun 2006 20:37:03 +0000 (13:37 -0700)]
[IPV6]: Added GSO support for TCPv6

This patch adds GSO support for IPv6 and TCPv6.  This is based on a patch
by Ananda Raju <Ananda.Raju@neterion.com>.  His original description is:

This patch enables TSO over IPv6. Currently Linux network stacks
restricts TSO over IPv6 by clearing of the NETIF_F_TSO bit from
"dev->features". This patch will remove this restriction.

This patch will introduce a new flag NETIF_F_TSO6 which will be used
to check whether device supports TSO over IPv6. If device support TSO
over IPv6 then we don't clear of NETIF_F_TSO and which will make the
TCP layer to create TSO packets. Any device supporting TSO over IPv6
will set NETIF_F_TSO6 flag in "dev->features" along with NETIF_F_TSO.

In case when user disables TSO using ethtool, NETIF_F_TSO will get
cleared from "dev->features". So even if we have NETIF_F_TSO6 we don't
get TSO packets created by TCP layer.

SKB_GSO_TCPV4 renamed to SKB_GSO_TCP to make it generic GSO packet.
SKB_GSO_UDPV4 renamed to SKB_GSO_UDP as UFO is not a IPv4 feature.
UFO is supported over IPv6 also

The following table shows there is significant improvement in
throughput with normal frames and CPU usage for both normal and jumbo.

--------------------------------------------------
|          |     1500        |      9600         |
|          ------------------|-------------------|
|          | thru     CPU    |  thru     CPU     |
--------------------------------------------------
| TSO OFF  | 2.00   5.5% id  |  5.66   20.0% id  |
--------------------------------------------------
| TSO ON   | 2.63   78.0 id  |  5.67   39.0% id  |
--------------------------------------------------

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: Generalise TSO-specific bits from skb_setup_caps
Herbert Xu [Fri, 30 Jun 2006 20:36:35 +0000 (13:36 -0700)]
[NET]: Generalise TSO-specific bits from skb_setup_caps

This patch generalises the TSO-specific bits from sk_setup_caps by adding
the sk_gso_type member to struct sock.  This makes sk_setup_caps generic
so that it can be used by TCPv6 or UFO.

The only catch is that whoever uses this must provide a GSO implementation
for their protocol which I think is a fair deal :) For now UFO continues to
live without a GSO implementation which is OK since it doesn't use the sock
caps field at the moment.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPV6]: Added GSO support for TCPv6
Herbert Xu [Fri, 30 Jun 2006 20:36:15 +0000 (13:36 -0700)]
[IPV6]: Added GSO support for TCPv6

This patch adds GSO support for IPv6 and TCPv6.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPV6]: Remove redundant length check on input
Herbert Xu [Fri, 30 Jun 2006 20:35:46 +0000 (13:35 -0700)]
[IPV6]: Remove redundant length check on input

We don't need to check skb->len when we're just about to call
pskb_may_pull since that checks it for us.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETFILTER]: SCTP conntrack: fix crash triggered by packet without chunks
Patrick McHardy [Fri, 30 Jun 2006 04:40:23 +0000 (21:40 -0700)]
[NETFILTER]: SCTP conntrack: fix crash triggered by packet without chunks

When a packet without any chunks is received, the newconntrack variable
in sctp_packet contains an out of bounds value that is used to look up an
pointer from the array of timeouts, which is then dereferenced, resulting
in a crash. Make sure at least a single chunk is present.

Problem noticed by George A. Theall <theall@tenablesecurity.com>

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>