linux-2.6
18 years ago[PATCH] remove unused tmp_buf_sem's
Adrian Bunk [Sat, 14 Jan 2006 02:08:16 +0000 (03:08 +0100)] 
[PATCH] remove unused tmp_buf_sem's

tmp_buf_sem sems to be a common name for something completely unused...

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de> ("usb portion")
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] docs: update some updated code docs
Xose Vazquez Perez [Sat, 14 Jan 2006 11:57:41 +0000 (12:57 +0100)] 
[PATCH] docs: update some updated code docs

Based on comments from Randy Dunlap on my previous commit
5b0ed2c64d8fdafb5fcfb3baabdd288628b1ff9b

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ufs cleanup
Evgeniy [Sat, 14 Jan 2006 08:42:06 +0000 (11:42 +0300)] 
[PATCH] ufs cleanup

Here is update of ufs cleanup patch, brought on by the recently fixed
ubh_get_usb_second() bug that made some ugly code rather painfully
obvious.  It also includes

 - fix compilation warnings which appears if debug mode turn on
 - remove unnecessary duplication of code to support UFS2

I tested it on ufs1 and ufs2 file-systems.

Signed-off-by: Evgeniy Dushistov <dushistov@mail.ru>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Sat, 14 Jan 2006 17:55:28 +0000 (09:55 -0800)] 
Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Sat, 14 Jan 2006 17:49:16 +0000 (09:49 -0800)] 
Merge /pub/scm/linux/kernel/git/dtor/input

18 years agoInput: HID - add support for fn key on Apple PowerBooks
Michael Hanselmann [Sat, 14 Jan 2006 15:08:06 +0000 (10:08 -0500)] 
Input: HID - add support for fn key on Apple PowerBooks

This patch implements support for the fn key on Apple PowerBooks using
USB based keyboards and makes them behave like their ADB counterparts.

Signed-off-by: Michael Hanselmann <linux-kernel@hansmi.ch>
Acked-by: Rene Nussbaumer <linux-kernel@killerfox.forkbomb.ch>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
18 years agoInput: wacom - fix compile on PowerPC
Dmitry Torokhov [Sat, 14 Jan 2006 05:28:04 +0000 (00:28 -0500)] 
Input: wacom - fix compile on PowerPC

Rename G4 (new Graphire4) to WACOM_G4 to avoid clashes on PowerPC

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
18 years agoInput: HID - add more simulation usages
Dmitry Torokhov [Sat, 14 Jan 2006 05:27:51 +0000 (00:27 -0500)] 
Input: HID - add more simulation usages

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
18 years agoInput: psmouse - attempt to re-synchronize mouse every 5 seconds
Dmitry Torokhov [Sat, 14 Jan 2006 05:27:37 +0000 (00:27 -0500)] 
Input: psmouse - attempt to re-synchronize mouse every 5 seconds

This should help driver to deal vith KVMs that reset mice when
switching between boxes.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
18 years agoInput: HID - fix an oops in PID initialization code
Dmitry Torokhov [Sat, 14 Jan 2006 05:26:15 +0000 (00:26 -0500)] 
Input: HID - fix an oops in PID initialization code

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
18 years agoInput: HID - add support for Cherry Cymotion keyboard
Vojtech Pavlik [Sat, 14 Jan 2006 05:25:39 +0000 (00:25 -0500)] 
Input: HID - add support for Cherry Cymotion keyboard

The Cherry Cymotion is a special Linux keyboard made by Cherry, with
only one little problem: it doesn't work with Linux. This patch
(originally by hexten.net, cleaned up by me) makes it work including
all the special keys.

Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc-merge
Linus Torvalds [Sat, 14 Jan 2006 05:24:55 +0000 (21:24 -0800)] 
Merge git://git./linux/kernel/git/paulus/powerpc-merge

18 years agoInput: i8042 - add Sony Vaio FSC-115b to MUX blacklist
Vojtech Pavlik [Sat, 14 Jan 2006 05:24:06 +0000 (00:24 -0500)] 
Input: i8042 - add Sony Vaio FSC-115b to MUX blacklist

Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
18 years ago[PATCH] powerpc: Recognize /chaos bridge on old pmacs as PCI
Paul Mackerras [Sat, 14 Jan 2006 04:08:50 +0000 (15:08 +1100)] 
[PATCH] powerpc: Recognize /chaos bridge on old pmacs as PCI

The first generation of PCI powermacs had a host bridge called /chaos
which was for all intents and purposes a PCI host bridge, but has a
device_type of "vci" in the device tree (presumably it's not really
PCI at the hardware level or something).

The OF parsing stuff in arch/powerpc/kernel/prom_parse.c currently
doesn't recognize it as a PCI bridge, which means that controlfb.c
can't get its device addresses.

This makes prom_parse.c recognize a device_type of "vci" as indicating
a PCI host bridge.  With this, controlfb works again.

Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agopowerpc: Avoid unaligned loads and stores in boot memcpy code
Paul Mackerras [Sat, 14 Jan 2006 04:06:51 +0000 (15:06 +1100)] 
powerpc: Avoid unaligned loads and stores in boot memcpy code

The 601 processor will generate an alignment exception for accesses
which cross a page boundary.  In the boot wrapper code, OF is still
handling all exceptions, and it doesn't have an alignment exception
handler that emulates the instruction and continues.

This changes the memcpy and memmove routines in the boot wrapper to
avoid doing unaligned accesses.  If the source and destination are
misaligned with respect to each other, we just copy one byte at a
time.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agopowerpc: Make COFF zImages for old 32-bit powermacs
Paul Mackerras [Sat, 14 Jan 2006 04:04:06 +0000 (15:04 +1100)] 
powerpc: Make COFF zImages for old 32-bit powermacs

This adds code to build zImage.coff and/or zImage.initrd.coff when
CONFIG_PPC32 and CONFIG_PPC_PMAC are defined.  It also restructures
the OF client code and adds some workarounds for OF quirks on the
older machines.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: macio-adb build fix
David Woodhouse [Sat, 14 Jan 2006 00:15:19 +0000 (00:15 +0000)] 
[PATCH] powerpc: macio-adb build fix

This makes macio-adb.c build again. Entirely untested.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Make CHRP build again
David Woodhouse [Sat, 14 Jan 2006 00:13:49 +0000 (00:13 +0000)] 
[PATCH] powerpc: Make CHRP build again

This makes CHRP build again, although it's untested because my Pegasos
is currently in pieces.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Sat, 14 Jan 2006 01:13:44 +0000 (17:13 -0800)] 
Merge /pub/scm/linux/kernel/git/davem/net-2.6

18 years ago[PATCH] powerpc: special-case ibm,suspend-me RTAS call
Dave C Boutcher [Sat, 14 Jan 2006 00:39:24 +0000 (18:39 -0600)] 
[PATCH] powerpc: special-case ibm,suspend-me RTAS call

Handle the ibm,suspend-me RTAS call specially.  It needs
to be wrapped in a set of synchronization hypervisor calls
(H_Join).  When the H_Join calls are made on all CPUs, the
intent is that only one will return with H_Continue, meaning
that he is the "last man standing".  That CPU then issues the
ibm,suspend-me call.  What is interesting, of course, is that
the CPU running when the rtas syscall is made, may NOT be the
CPU that ultimately executes the ibm,suspend-me rtas call.

Signed-off-by: Dave Boutcher <sleddog@us.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoppc: Remove duplicate export of get_wchan
Paul Mackerras [Sat, 14 Jan 2006 00:17:32 +0000 (11:17 +1100)] 
ppc: Remove duplicate export of get_wchan

The arch/powerpc version of process.c exports get_wchan itself.  When
I moved ARCH=ppc over to using arch/powerpc/kernel/process.c the
get_wchan export in arch/ppc/kernel/ppc_ksyms.c became redundant, so
remove it.

Signed-off-by: Paul Mackerras <paulus@samba.org>
(cherry picked from 9871166ad692121d6b944159ef3f053570158ea8 commit)

18 years ago[NETFILTER] x-tables: Missing linux/ipv6.h includes.
David S. Miller [Sat, 14 Jan 2006 00:19:44 +0000 (16:19 -0800)] 
[NETFILTER] x-tables: Missing linux/ipv6.h includes.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[PATCH] powerpc/8xx: Use 8MB D-TLB's for kernel static mapping faults
Marcelo Tosatti [Fri, 13 Jan 2006 16:16:12 +0000 (14:16 -0200)] 
[PATCH] powerpc/8xx: Use 8MB D-TLB's for kernel static mapping faults

The following implements support for instantiation of 8MB D-TLB
entries for the kernel direct virtual mapping on 8xx, thus reducing TLB
space consumed for the kernel.

Test used: writing 40MB from /dev/zero to file in ext2fs over
RAMDISK.

$ time dd if=/dev/zero of=file bs=4k count=10000

VANILLA 8MB kernel data pages

real    0m11.485s real    0m11.267s
user    0m0.218s        user    0m0.250s
sys     0m8.939s sys     0m9.108s

real    0m11.518s real    0m10.978s
user    0m0.203s  user    0m0.222s
sys     0m9.585s sys     0m9.138s

real    0m11.554s real    0m10.967s
user    0m0.228s     user    0m0.222s
sys     0m9.497s sys     0m9.127s

real    0m11.633s real 0m11.286s
user    0m0.214s user    0m0.196s
sys     0m9.529s sys     0m9.134s

and averages for both:

real 11.54750 real 11.12450

Which is a 3.6% improvement in execution time. More improvement is
expected for loads with larger kernel data footprint (real workloads).

Signed-off-by: Marcelo Tosatti <marcelo.tosatti@cyclades.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Add MPC834x SYS board to arch/powerpc
Kumar Gala [Fri, 13 Jan 2006 17:19:58 +0000 (11:19 -0600)] 
[PATCH] powerpc: Add MPC834x SYS board to arch/powerpc

Add the first MPC83xx board that uses a flat device tree to arch/powerpc.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Add FSL SOC library and setup code
Kumar Gala [Fri, 13 Jan 2006 17:19:13 +0000 (11:19 -0600)] 
[PATCH] powerpc: Add FSL SOC library and setup code

Parse the flat device tree for devices on Freescale SOC's that we know
about (gianfar, gianfar_mdio, i2c, mpc83xx_wdt).  We need to setup
platform devices and platform data for these devices to match arch/ppc
usage.

Also add a helper function (get_immrbase) that reports the base
address of the MMIO registers on the SOC.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Allow for ppc_md restart, power_off, and halt to be NULL
Kumar Gala [Fri, 13 Jan 2006 16:15:17 +0000 (10:15 -0600)] 
[PATCH] powerpc: Allow for ppc_md restart, power_off, and halt to be NULL

On a number of embedded reference boards there isn't always a
way to reset, power_off, or halt the board.  Rather than having
each board implement a spin loop just let the generic code do
it.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: oprofile cpu type names clash with other code
Andy Whitcroft [Fri, 13 Jan 2006 12:35:49 +0000 (12:35 +0000)] 
[PATCH] powerpc: oprofile cpu type names clash with other code

In 2.6.15-git6 a change was commited in the oprofile support in
the powerpc architecture.  It introduced the powerpc_oprofile_type
which contains the define G4.  This causes a name clash with the
existing wacom usb tablet driver.

      CC [M]  drivers/usb/input/wacom.o
    drivers/usb/input/wacom.c:98: error: conflicting types for `G4'
    include/asm/cputable.h:37: error: previous declaration of `G4'
      CC [M]  drivers/usb/mon/mon_text.o
    make[3]: *** [drivers/usb/input/wacom.o] Error 1
    make[2]: *** [drivers/usb/input] Error 2

The elements of an enum declared in global scope are effectivly
global identifiers themselves.  As such we need to ensure the names
are unique.  This patch updates the later oprofile support to use
unique names.

Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoMerge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband
Linus Torvalds [Fri, 13 Jan 2006 23:29:07 +0000 (15:29 -0800)] 
Merge branch 'for-linus' of /linux/kernel/git/roland/infiniband

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Fri, 13 Jan 2006 23:28:10 +0000 (15:28 -0800)] 
Merge /pub/scm/linux/kernel/git/davem/net-2.6

18 years agopowerpc: Provide a suitable AT_PLATFORM value
Paul Mackerras [Fri, 13 Jan 2006 23:11:39 +0000 (10:11 +1100)] 
powerpc: Provide a suitable AT_PLATFORM value

The glibc folks want to use AT_PLATFORM to select between possible
alternative versions of shared libraries.  This commit makes the kernel
supply an AT_PLATFORM string that indicates what class of processor
we are running on.  Processors with the same set of user-level
instructions and roughly the same instruction scheduling characteristics
are given the same AT_PLATFORM value; for example, 821, 823 and 860
are all reported as "ppc823", and 7447, 7447A, 7448, 7450, 7451, 7455
are all called "ppc7450".

The intention is that the AT_PLATFORM values match the values that
gcc accepts for the -mcpu= option.  For values which are numeric
(e.g. -mcpu=750), "ppc" has been prepended.

This also adds a PPC_FEATURE_BOOKE bit to the AT_HWCAP value and sets
it for the 440 family and the Freescale 85xx family.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[IA64] build broken for ia64 simserial.c
Andreas Schwab [Fri, 13 Jan 2006 22:46:38 +0000 (23:46 +0100)] 
[IA64] build broken for ia64 simserial.c

TTY layer buffering revamp broke ia64 in commit
 33f0f88f1c51ae5c2d593d26960c760ea154c2e2

  CC      arch/ia64/hp/sim/simserial.o
arch/ia64/hp/sim/simserial.c: In function `receive_chars':
arch/ia64/hp/sim/simserial.c:170: error: structure has no member named `flip'
 ... and so on ...
make[1]: *** [arch/ia64/hp/sim/simserial.o] Error 1

Patch from Andreas Schwab.

Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years agoIB: convert from semaphores to mutexes
Ingo Molnar [Fri, 13 Jan 2006 22:51:39 +0000 (14:51 -0800)] 
IB: convert from semaphores to mutexes

semaphore to mutex conversion by Ingo and Arjan's script.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
[ Sanity-checked on real IB hardware ]
Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years ago[IA64] prevent accidental modification of args in jprobe handler
Zhang Yanmin [Fri, 13 Jan 2006 22:45:21 +0000 (14:45 -0800)] 
[IA64] prevent accidental modification of args in jprobe handler

When jprobe is hit, the function parameters of the original function
should be saved before jprobe handler is executed, and restored it after
jprobe handler is executed, because jprobe handler might change the
register values due to tail call optimization by the gcc.

Signed-off-by: Zhang Yanmin <yanmin.zhang@intel.com>
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[PKT_SCHED]: Change default clock source to gettimeofday
Patrick McHardy [Fri, 13 Jan 2006 22:36:55 +0000 (14:36 -0800)] 
[PKT_SCHED]: Change default clock source to gettimeofday

The default of using jiffies is very bad and results in
underutilization except with very low bandwidth.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[XFRM]: IPsec tunnel wildcard address support
Patrick McHardy [Fri, 13 Jan 2006 22:34:36 +0000 (14:34 -0800)] 
[XFRM]: IPsec tunnel wildcard address support

When the source address of a tunnel is given as 0.0.0.0 do a routing lookup
to get the real source address for the destination and fill that into the
acquire message. This allows to specify policies like this:

spdadd 172.16.128.13/32 172.16.0.0/20 any -P out ipsec
        esp/tunnel/0.0.0.0-x.x.x.x/require;
spdadd 172.16.0.0/20 172.16.128.13/32 any -P in ipsec
        esp/tunnel/x.x.x.x-0.0.0.0/require;

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: Clean up comments for sk_chk_filter()
Kris Katterjohn [Fri, 13 Jan 2006 22:33:06 +0000 (14:33 -0800)] 
[NET]: Clean up comments for sk_chk_filter()

This removes redundant comments, and moves one comment to a better
location.

Signed-off-by: Kris Katterjohn <kjak@users.sourceforge.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: Use NIP6_FMT in kernel.h
Joe Perches [Fri, 13 Jan 2006 22:29:07 +0000 (14:29 -0800)] 
[NET]: Use NIP6_FMT in kernel.h

There are errors and inconsistency in the display of NIP6 strings.
ie: net/ipv6/ip6_flowlabel.c

There are errors and inconsistency in the display of NIPQUAD strings too.
ie: net/netfilter/nf_conntrack_ftp.c

This patch:
adds NIP6_FMT to kernel.h
changes all code to use NIP6_FMT
fixes net/ipv6/ip6_flowlabel.c
adds NIPQUAD_FMT to kernel.h
fixes net/netfilter/nf_conntrack_ftp.c
changes a few uses of "%u.%u.%u.%u" to NIPQUAD_FMT for symmetry to NIP6_FMT

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IA64] Add hotplug cpu to salinfo.c, replace semaphore with mutex
Keith Owens [Thu, 5 Jan 2006 23:36:06 +0000 (10:36 +1100)] 
[IA64] Add hotplug cpu to salinfo.c, replace semaphore with mutex

Add hotplug cpu support to salinfo.c.

The cpu_event field is a cpumask so use the cpu_* macros consistently,
replacing the existing mixture of cpu_* and *_bit macros.

Instead of counting the number of outstanding events in a semaphore and
trying to track that count over user space context, interrupt context,
non-maskable interrupt context and cpu hotplug, replace the semaphore
with a test for "any bits set" combined with a mutex.

Modify the locking to make the test for "work to do" an atomic
operation.

Signed-off-by: Keith Owens <kaos@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64] Handle debug traps in fsys mode
Jason Uhlenkott [Fri, 30 Dec 2005 10:27:01 +0000 (02:27 -0800)] 
[IA64] Handle debug traps in fsys mode

We need to handle debug traps in fsys mode non-fatally.  They can
happen now that we have fsyscalls which contain probe instructions.

Signed-off-by: Jason Uhlenkott <jasonuhl@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years agoMerge ../linux-2.6
Paul Mackerras [Fri, 13 Jan 2006 22:15:28 +0000 (09:15 +1100)] 
Merge ../linux-2.6

18 years ago[IA64-SGI] Fix sn_flush_device_kernel & spinlock initialization
Prarit Bhargava [Fri, 23 Dec 2005 18:33:25 +0000 (13:33 -0500)] 
[IA64-SGI] Fix sn_flush_device_kernel & spinlock initialization

This patch separates the sn_flush_device_list struct into kernel and
common (both kernel and PROM accessible) structures.  As it was, if the
size of a spinlock_t changed (due to additional CONFIG options, etc.) the
sal call which populated the sn_flush_device_list structs would erroneously
write data (and cause memory corruption and/or a panic).

This patch does the following:

1.  Removes sn_flush_device_list and adds sn_flush_device_common and
sn_flush_device_kernel.

2.  Adds a new SAL call to populate a sn_flush_device_common struct per
device, not per widget as previously done.

3.  Correctly initializes each device's sn_flush_device_kernel spinlock_t
struct (before it was only doing each widget's first device).

Signed-off-by: Prarit Bhargava <prarit@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64] Hole in IA64 TLB flushing from system threads
Jack Steiner [Thu, 22 Dec 2005 19:45:41 +0000 (13:45 -0600)] 
[IA64] Hole in IA64 TLB flushing from system threads

I originally thought this was an bug only in the SN code, but I think I
also see a hole in the generic IA64 tlb code. (Separate patch was sent
for the SN problem).

It looks like there is a bug in the TLB flushing code. During context switch,
kernel threads (kswapd, for example) inherit the mm of the task that was
previously running on the cpu. Normally, this is ok because the previous context
is still loaded into the RR registers. However, if the owner of the mm
migrates to another cpu, changes it's context number, and references a
page before kswapd issues a tlb_purge for that same page, the purge will be
done with a stale context number (& RR registers).

Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] Altix BTE error handling fixes
Russ Anderson [Fri, 16 Dec 2005 23:19:01 +0000 (17:19 -0600)] 
[IA64-SGI] Altix BTE error handling fixes

Altix (shub2) pushes the BTE clean-up into SAL.
This patch correctly interfaces with the now implemented SAL call.
It also fixes a bug when delaying clean-up to allow busy BTEs to
complete (or error out).

Signed-off-by: Russ Anderson <rja@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64] Fix conversion of pal_min_state physical address
Francois Wellenrieter [Fri, 13 Jan 2006 22:01:01 +0000 (14:01 -0800)] 
[IA64] Fix conversion of pal_min_state physical address

On return from INIT handler we must convert the address of the
minstate area from a kernel virtual uncached address (0xC...)
to physical uncached (0x8...).  A typo (or thinko?) in the code
converted to physical cached.

Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[PATCH] genetlink: don't touch module ref count
Per Liden [Fri, 13 Jan 2006 21:06:40 +0000 (13:06 -0800)] 
[PATCH] genetlink: don't touch module ref count

Increasing the module ref count at registration will block the module from
ever being unloaded. In fact, genetlink should not care about the owner at
all. This patch removes the owner field from the struct registered with
genetlink.

Signed-off-by: Per Liden <per.liden@ericsson.com>
Signed-off-by: Jamal Hadi Salim <hadi@cyberus.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IA64-SGI] move xpc.h to include/asm-ia64/sn (cleanup)
Dean Nelson [Tue, 10 Jan 2006 17:12:32 +0000 (11:12 -0600)] 
[IA64-SGI] move xpc.h to include/asm-ia64/sn (cleanup)

Cleanup a few items after moving xpc.h from arch/ia64/sn/kernel to
include/asm-ia64/sn.

Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] move xpc.h to include/asm-ia64/sn
Dean Nelson [Tue, 10 Jan 2006 17:09:48 +0000 (11:09 -0600)] 
[IA64-SGI] move xpc.h to include/asm-ia64/sn

Move xpc.h from arch/ia64/sn/kernel to include/asm-ia64/sn without change.

Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] move xpc_system_reboot()
Dean Nelson [Tue, 10 Jan 2006 17:08:55 +0000 (11:08 -0600)] 
[IA64-SGI] move xpc_system_reboot()

Move xpc_system_reboot() to be closer to the file it calls for readability
reasons (which are indeed subjective).

Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] ignoring loss of heartbeat while XPC is in kdebug
Dean Nelson [Tue, 10 Jan 2006 17:08:00 +0000 (11:08 -0600)] 
[IA64-SGI] ignoring loss of heartbeat while XPC is in kdebug

Allow for the loss of heartbeat while in kdebug to be ignored by remote
partitions.

Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] XPC and unregistering from notifier lists
Dean Nelson [Tue, 10 Jan 2006 17:07:19 +0000 (11:07 -0600)] 
[IA64-SGI] XPC and unregistering from notifier lists

Only unregister from notifier lists if XPC is unloading.

Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] cleanup XPC disengage related messages
Dean Nelson [Fri, 6 Jan 2006 15:48:21 +0000 (09:48 -0600)] 
[IA64-SGI] cleanup XPC disengage related messages

Cleanup the XPC disengage related messages that are printed to the log.

Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64-SGI] ensure XPC disengage request is processed
Dean Nelson [Thu, 22 Dec 2005 20:32:56 +0000 (14:32 -0600)] 
[IA64-SGI] ensure XPC disengage request is processed

This patch fixes a problem in XPC disengage processing whereby it was not
seeing the request to disengage from a remote partition, so the disengage
wasn't happening. The disengagement is suppose to transpire during the time
a XPC channel is disconnecting, and should be completed before the channel
is declared to be disconnected.

Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[IA64] Add stub entry to fsys.S for sys_migrate_pages
Tony Luck [Fri, 13 Jan 2006 18:03:58 +0000 (10:03 -0800)] 
[IA64] Add stub entry to fsys.S for sys_migrate_pages

When this new syscall was added to ia64 in commit

  39743889aaf76725152f16aa90ca3c45f6d52da3

fsys.S was forgotten.  Add a ".data8 0" there to keep
it in step.  [Reported by Stephane Eranian]

Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years ago[PATCH] Increase AT_VECTOR_SIZE
Paul Mackerras [Fri, 13 Jan 2006 03:23:25 +0000 (14:23 +1100)] 
[PATCH] Increase AT_VECTOR_SIZE

On PowerPC, we want to be able to provide an AT_PLATFORM aux table
entry to userspace, so that glibc can choose optimized libraries for
the processor we're running on.  Unfortunately that would be the 21st
aux table entry on powerpc, meaning that the aux table including the
terminating null entry would overflow the mm->saved_auxv[] array,
leading to userland programs segfaulting.

This increases the size of the mm->saved_auxv array to be large enough
to accommodate an AT_PLATFORM entry on powerpc.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Fix oops in ufs_fill_super at mount time
Evgeniy [Fri, 13 Jan 2006 15:12:15 +0000 (18:12 +0300)] 
[PATCH] Fix oops in ufs_fill_super at mount time

There's a lack of parenthesis in fs/ufs/utils.h, so instead of the 512th
byte of buffer, the usb2 pointer will point to the nth structure of type
ufs_super_block_second.

This can cause a mount-time oops if you're unlucky (especially with
DEBUG_PAGEALLOC, which is how Alexey Dobriyan saw this problem)

Signed-off-by: Evgeniy Dushistov <dushistov@mail.ru>
Acked-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] powerpc: reformat atomic_add_unless
Anton Blanchard [Fri, 13 Jan 2006 05:51:52 +0000 (16:51 +1100)] 
[PATCH] powerpc: reformat atomic_add_unless

It makes my eyes hurt.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: use lwsync in atomics, bitops, lock functions
Anton Blanchard [Fri, 13 Jan 2006 04:37:17 +0000 (15:37 +1100)] 
[PATCH] powerpc: use lwsync in atomics, bitops, lock functions

eieio is only a store - store ordering. When used to order an unlock
operation loads may leak out of the critical region. This is potentially
buggy, one example is if a user wants to atomically read a couple of
values.

We can solve this with an lwsync which orders everything except store - load.

I removed the (now unused) EIEIO_ON_SMP macros and the c versions
isync_on_smp and eieio_on_smp now we dont use them. I also removed some
old comments that were used to identify inline spinlocks in assembly,
they dont make sense now our locks are out of line.

Another interesting thing was that read_unlock was using an eieio even
though the rest of the spinlock code had already been converted to
use lwsync.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Remove lppaca structure from the PACA
David Gibson [Thu, 12 Jan 2006 23:26:42 +0000 (10:26 +1100)] 
[PATCH] powerpc: Remove lppaca structure from the PACA

At present the lppaca - the structure shared with the iSeries
hypervisor and phyp - is contained within the PACA, our own low-level
per-cpu structure.  This doesn't have to be so, the patch below
removes it, making a separate array of lppaca structures.

This saves approximately 500*NR_CPUS bytes of image size and kernel
memory, because we don't need aligning gap between the Linux and
hypervisor portions of every PACA.  On the other hand it means an
extra level of dereference in many accesses to the lppaca.

The patch also gets rid of several places where we assign the paca
address to a local variable for no particular reason.

Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Cleanup LOADADDR etc. asm macros
David Gibson [Fri, 13 Jan 2006 03:56:25 +0000 (14:56 +1100)] 
[PATCH] powerpc: Cleanup LOADADDR etc. asm macros

This patch consolidates the variety of macros used for loading 32 or
64-bit constants in assembler (LOADADDR, LOADBASE, SET_REG_TO_*).  The
idea is to make the set of macros consistent across 32 and 64 bit and
to make it more obvious which is the appropriate one to use in a given
situation.  The new macros and their semantics are described in the
comments in ppc_asm.h.

In the process, we change several places that were unnecessarily using
immediate loads on ppc64 to use the GOT/TOC.  Likewise we cleanup a
couple of places where we were clumsily subtracting PAGE_OFFSET with
asm instructions to use assemble-time arithmetic or the toreal() macro
instead.

Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Updated platforms that use gianfar to match driver
Kumar Gala [Fri, 13 Jan 2006 03:04:23 +0000 (21:04 -0600)] 
[PATCH] powerpc: Updated platforms that use gianfar to match driver

The gianfar driver changed how it required MDIO bus and phy id's
to be passed to it.  Also, it no longer passes the physical address
of the MDIO bus.  Instead we have a proper platform device.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc/boot: Better use of defaultimage-
Tom Rini [Thu, 12 Jan 2006 23:55:58 +0000 (16:55 -0700)] 
[PATCH] powerpc/boot: Better use of defaultimage-

The following reworks how defaultimage- is used.  We default to zImage
here and then override it on platforms that need something more (uImage
in the future) or less (vmlinux on iSeries).

Signed-off-by: Tom Rini <trini@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Add support for changing properties from userspace
Dave C Boutcher [Thu, 12 Jan 2006 22:10:22 +0000 (16:10 -0600)] 
[PATCH] powerpc: Add support for changing properties from userspace

Add support to reconfigure the device tree through the existing
proc filesystem interface.  Add "add_property", "remove_property",
and "update_property" commands to the existing interface.

Signed-off-by: Dave Boutcher <sleddog@us.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Add of_find_property function
Dave C Boutcher [Thu, 12 Jan 2006 22:09:29 +0000 (16:09 -0600)] 
[PATCH] powerpc: Add of_find_property function

Add an of_find_property function that returns a struct property
given a property name.  Then change the get_property function to
use that routine internally.

Signed-off-by: Dave Boutcher <sleddog@us.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Add/remove/update properties in firmware device tree
Dave C Boutcher [Thu, 12 Jan 2006 22:08:27 +0000 (16:08 -0600)] 
[PATCH] powerpc: Add/remove/update properties in firmware device tree

Add support for updating and removing device tree
properties.  Since we hand out pointers to properties with gay
abandon, we can't just free the property storage.  Instead we
move deleted, or the old copy of an updated property, to a
"dead properties" list.

Also note, its not feasable to kref device tree properties.
we call get_property() all over the kernel in a wild variety
of contexts.

One consequence of this change is that we now take a
read_lock(&devtree_lock) when doing get_property().

Signed-off-by: Dave Boutcher <sleddog@us.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Add/remove/update properties in /proc/device-tree
Dave C Boutcher [Thu, 12 Jan 2006 22:07:17 +0000 (16:07 -0600)] 
[PATCH] powerpc: Add/remove/update properties in /proc/device-tree

Add support to the proc_device_tree file for removing
and updating properties.  Remove just removes the
proc file, update changes the data pointer within
the proc file.  The remainder of the device-tree
changes occur elsewhere.

Signed-off-by: Dave Boutcher <sleddog@us.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Add some more pSeries hypervisor call constants
Dave C Boutcher [Thu, 12 Jan 2006 22:05:35 +0000 (16:05 -0600)] 
[PATCH] powerpc: Add some more pSeries hypervisor call constants

Adds a few more hypervisor call constants.

Signed-off-by: Dave Boutcher <sleddog@us.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] Remove bzImage target
Tom Rini [Thu, 12 Jan 2006 21:25:50 +0000 (14:25 -0700)] 
[PATCH] Remove bzImage target

Since the documentation no longer tells people to 'make bzImage', and
with the previous patch nothing more than 'make' is required to get the
right bootable images (just like on i386 now), this removes the bzImage
-> zImage target redirect on ARCH=powerpc

Signed-off-by: Tom Rini <trini@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoIB/mthca: Initialize grh_present before using it
Michael S. Tsirkin [Thu, 12 Jan 2006 23:55:41 +0000 (15:55 -0800)] 
IB/mthca: Initialize grh_present before using it

build_mlx_header() was using sqp->ud_header.grh_present before it was
initialized by mthca_read_ah().  Furthermore, header->grh_present is
set by ib_ud_header_init, so there's no need to set it again in
mthca_read_ah().

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years agoIB/mthca: Cosmetic: use the ALIGN macro
Michael S. Tsirkin [Thu, 12 Jan 2006 23:43:58 +0000 (15:43 -0800)] 
IB/mthca: Cosmetic: use the ALIGN macro

Use the ALIGN macro to simplify some rounding code.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Thu, 12 Jan 2006 23:43:39 +0000 (15:43 -0800)] 
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/netdev-2.6

18 years agoIB/mthca: Fix memory leaks in error handling
Jack Morgenstein [Thu, 12 Jan 2006 23:35:15 +0000 (15:35 -0800)] 
IB/mthca: Fix memory leaks in error handling

Fix memory leaks in mthca_create_qp() and mthca_create_srq()
error handling.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years agoIB/mthca: Fix memory leak of multicast group structures
Ishai Rabinovitz [Thu, 12 Jan 2006 23:24:51 +0000 (15:24 -0800)] 
IB/mthca: Fix memory leak of multicast group structures

Convert "/ (1 << lg)" to ">> lg" for a slight code size reduction.

add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-24 (-24)
function                                     old     new   delta
mthca_map_cmd                                613     589     -24

Signed-off-by: Ishai Rabinovitz <ishai@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years agoIPoIB: Fix memory leak of multicast group structures
Eli Cohen [Thu, 12 Jan 2006 22:32:20 +0000 (14:32 -0800)] 
IPoIB: Fix memory leak of multicast group structures

The current handling of multicast groups in IPoIB ends up never
freeing send-only multicast groups.  It turns out the logic was much
more complicated than it needed to be; we can fix this bug and
completely kill ipoib_mcast_dev_down() at the same time.

Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years ago[NETFILTER] x_tables: Abstraction layer for {ip,ip6,arp}_tables
Harald Welte [Thu, 12 Jan 2006 21:30:04 +0000 (13:30 -0800)] 
[NETFILTER] x_tables: Abstraction layer for {ip,ip6,arp}_tables

This monster-patch tries to do the best job for unifying the data
structures and backend interfaces for the three evil clones ip_tables,
ip6_tables and arp_tables.  In an ideal world we would never have
allowed this kind of copy+paste programming... but well, our world
isn't (yet?) ideal.

o introduce a new x_tables module
o {ip,arp,ip6}_tables depend on this x_tables module
o registration functions for tables, matches and targets are only
  wrappers around x_tables provided functions
o all matches/targets that are used from ip_tables and ip6_tables
  are now implemented as xt_FOOBAR.c files and provide module aliases
  to ipt_FOOBAR and ip6t_FOOBAR
o header files for xt_matches are in include/linux/netfilter/,
  include/linux/netfilter_{ipv4,ipv6} contains compatibility wrappers
  around the xt_FOOBAR.h headers

Based on this patchset we're going to further unify the code,
gradually getting rid of all the layer 3 specific assumptions.

Signed-off-by: Harald Welte <laforge@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[TIPC]: Fix 64-bit build warnings.
David S. Miller [Thu, 12 Jan 2006 21:22:32 +0000 (13:22 -0800)] 
[TIPC]: Fix 64-bit build warnings.

When storing u32 values in a pointer, need to do
some long casts to keep GCC happy.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[TIPC] Added entry for TIPC in MAINTAINERS
Per Liden [Thu, 12 Jan 2006 15:45:18 +0000 (16:45 +0100)] 
[TIPC] Added entry for TIPC in MAINTAINERS

Signed-off-by: Per Liden <per.liden@nospam.ericsson.com>
18 years ago[TIPC] More updates of file headers
Per Liden [Wed, 11 Jan 2006 18:14:19 +0000 (19:14 +0100)] 
[TIPC] More updates of file headers

Updated copyright notice to include the year the file was
actually created. Information about file creation dates
was extracted from the files in the old CVS repository
at tipc.sourceforge.net.

Signed-off-by: Per Liden <per.liden@nospam.ericsson.com>
18 years ago[TIPC] Update of file headers
Per Liden [Wed, 11 Jan 2006 17:40:41 +0000 (18:40 +0100)] 
[TIPC] Update of file headers

The copyright statements from different parts of Ericsson
have been merged into one.

Signed-off-by: Per Liden <per.liden@nospam.ericsson.com>
18 years ago[TIPC] Cleaned up info/warn/err macros
Per Liden [Wed, 11 Jan 2006 12:52:51 +0000 (13:52 +0100)] 
[TIPC] Cleaned up info/warn/err macros

Signed-off-by: Per Liden <per.liden@nospam.ericsson.com>
18 years ago[TIPC] License header update
Per Liden [Wed, 11 Jan 2006 12:30:43 +0000 (13:30 +0100)] 
[TIPC] License header update

The license header in each file now more clearly state that this
code is licensed under a dual BSD/GPL. Before this was only
evident if you looked at the MODULE_LICENSE line in core.c.

Signed-off-by: Per Liden <per.liden@nospam.ericsson.com>
18 years ago[TIPC] Moved configuration interface into tipc_config.h
Per Liden [Wed, 11 Jan 2006 11:28:47 +0000 (12:28 +0100)] 
[TIPC] Moved configuration interface into tipc_config.h

Restored the old tipc_config.h to get a cleaner division between the
interfaces used by normal TIPC users and TIPC administration utilities.

Signed-off-by: Per Liden <per.liden@nospam.ericsson.com>
18 years ago[TIPC} Fixed bug in disc_timeout()
Jon Maloy [Tue, 10 Jan 2006 18:54:24 +0000 (18:54 +0000)] 
[TIPC} Fixed bug in disc_timeout()

Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
18 years ago[TIPC] Use dynamically allocated family id with NETLINK_GENERIC
Per Liden [Thu, 5 Jan 2006 15:34:00 +0000 (16:34 +0100)] 
[TIPC] Use dynamically allocated family id with NETLINK_GENERIC

Signed-off-by: Per Liden <per.liden@nospam.ericsson.com>
18 years ago[TIPC] Initial merge
Per Liden [Mon, 2 Jan 2006 18:04:38 +0000 (19:04 +0100)] 
[TIPC] Initial merge

TIPC (Transparent Inter Process Communication) is a protocol designed for
intra cluster communication. For more information see
http://tipc.sourceforge.net

Signed-off-by: Per Liden <per.liden@nospam.ericsson.com>
18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-fixes-2.6
Linus Torvalds [Thu, 12 Jan 2006 22:00:47 +0000 (14:00 -0800)] 
Merge git://git./linux/kernel/git/brodo/pcmcia-fixes-2.6

18 years agoMerge branch 'drm-forlinus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Thu, 12 Jan 2006 21:53:40 +0000 (13:53 -0800)] 
Merge branch 'drm-forlinus' of git://git./linux/kernel/git/airlied/drm-2.6

18 years ago[PATCH] via-velocity: use NETIF_F_IP_CSUM (hardware only support IPv4)
John W. Linville [Fri, 9 Dec 2005 15:36:09 +0000 (10:36 -0500)] 
[PATCH] via-velocity: use NETIF_F_IP_CSUM (hardware only support IPv4)

At least some versions of the via-velocity hardware only support
checksumming IPv4 frames in hardware.  However, the driver is currently
setting the NETIF_F_HW_CSUM flag, which indicates support for more than
just IPv4.  This results in errors when trying to use IPv6 over
via-velocity hardware.

Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[PATCH] fix wrong comments in ieee80211.h
Johannes Berg [Tue, 27 Dec 2005 01:14:13 +0000 (02:14 +0100)] 
[PATCH] fix wrong comments in ieee80211.h

The comments in ieee80211.h claim that one doesn't need to set the len
parameter of the stats struct. But if one doesn't, the management frames
are read far over the memory they actually occupy causing badness.

Signed-Off-By: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[PATCH] bonding: UPDATED hash-table corruption in bond_alb.c
Jay Vosburgh [Mon, 9 Jan 2006 20:14:00 +0000 (12:14 -0800)] 
[PATCH] bonding: UPDATED hash-table corruption in bond_alb.c

I believe I see the race Michael refers to (tlb_choose_channel
may set head, which tlb_init_slave clears), although I was not able to
reproduce it.  I have updated his patch for the current netdev-2.6.git
tree and added a version update.  His original comment follows:

Our systems have been crashing during testing of PCI HotPlug
support in the various networking components.  We've faulted in
the bonding driver due to a bug in bond_alb.c:tlb_clear_slave()

In that routine, the last modification to the TLB hash table is
made without protection of the lock, allowing a race that can lead
tlb_choose_channel() to select an invalid table element.

-J

Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[PATCH] wireless/atmel: add IWENCODEEXT, IWAUTH, and association event support
Dan Williams [Tue, 10 Jan 2006 05:56:11 +0000 (00:56 -0500)] 
[PATCH] wireless/atmel: add IWENCODEEXT, IWAUTH, and association event support

This patch allows the Atmel driver to work correctly with wpa_supplicant
and other programs that require some conformance with WEXT-18.  It
should not affect current behavior of the driver.  The patch does four
things:

1) Implements SIOCSIWENCODEEXT, SIOCGIWENCODEEXT, SIOCSIWAUTH, and
SIOCGIWAUTH calls for unencrypted and WEP operation

2) Accepts zero-filled addresses for SIOCSIWAP, which are legal and
should turn off any previous forced WAP address

3) Sends association and de-association events to userspace at most of
the appropriate times

4) Fixes erroneous order of CIPHER_SUITE_WEP_* arguments in one location
which are actually unused anyway

Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[PATCH] replace MODULE_PARM in tulip/uli526x.c
Eric Sesterhenn / snakebyte [Tue, 10 Jan 2006 12:16:03 +0000 (13:16 +0100)] 
[PATCH] replace MODULE_PARM in tulip/uli526x.c

Replace the MODULE_PARM usage in uli526x.c with module_param.

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[PATCH] gianfar: Use new PHY_ID_FMT macro
Kumar Gala [Wed, 11 Jan 2006 19:27:33 +0000 (11:27 -0800)] 
[PATCH] gianfar: Use new PHY_ID_FMT macro

Make the driver produce the string used by phy_connect and have board specific
code pass the integer mii bus id and phy device id for the specific controller
instance.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[PATCH] phy: Added a macro to represent the string format used to match a phy device
Kumar Gala [Wed, 11 Jan 2006 19:27:33 +0000 (11:27 -0800)] 
[PATCH] phy: Added a macro to represent the string format used to match a phy device

Add the PHY_ID_FMT macro to ensure that the format of the id string used by a
driver to match to its specific phy is consistent between the mdio_bus and the
driver.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[PATCH] gianfar mii: Use proper resource for MII memory region
Kumar Gala [Wed, 11 Jan 2006 19:27:32 +0000 (11:27 -0800)] 
[PATCH] gianfar mii: Use proper resource for MII memory region

We can now have the gianfar mii platform device have a proper resource for the
IO memory region for its registers.  Previously we passed this information
that the platform_data structure because we couldn't handle overlapping memory
regions for platform devices.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[PATCH] gfar: fix compile error
Kumar Gala [Wed, 11 Jan 2006 19:26:25 +0000 (11:26 -0800)] 
[PATCH] gfar: fix compile error

Missing include of <linux/in.h> to get definition of IPPROTO_UDP.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[PATCH] CONFIG_AIRO needs CONFIG_CRYPTO
dann frazier [Wed, 11 Jan 2006 19:26:24 +0000 (11:26 -0800)] 
[PATCH] CONFIG_AIRO needs CONFIG_CRYPTO

airo.c currently has MICSUPPORT enabled, which requires CONFIG_CRYPTO.  A
user reported a build failure which is due to the lack of a Kconfig
dependency.  See http://bugs.debian.org/344205.

This patch makes Kconfig enforce this dependency.

Signed-off-by: dann frazier <dannf@debian.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years ago[PATCH] corruption during e100 MDI register access
ODonnell, Michael [Wed, 11 Jan 2006 19:26:22 +0000 (11:26 -0800)] 
[PATCH] corruption during e100 MDI register access

We have identified two related bugs in the e100 driver.

Both bugs are related to manipulation of the MDI control register.

The first problem is that the Ready bit is being ignored when writing to
the Control register; we noticed this because the Linux bonding driver
would occasionally come to the spurious conclusion that the link was down
when querying Link State.  It turned out that by failing to wait for a
previous command to complete it was selecting what was essentially a random
register in the MDI register set.  When we added code that waits for the
Ready bit (as shown in the patch file below) all such problems ceased.

The second problem is that, although access to the MDI registers involves
multiple steps which must not be intermixed, nothing was defending against
two or more threads attempting simultaneous access.  The most obvious
situation where such interference could occur involves the watchdog versus
ioctl paths, but there are probably others, so we recommend the locking
shown in our patch file.

Signed-off-by: Michael O'Donnell <Michael.ODonnell@stratus.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jeff Garzik <jgarzik@pobox.com>
Cc: John Ronciak <john.ronciak@intel.com>
Cc: Ganesh Venkatesan <ganesh.venkatesan@intel.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
18 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Thu, 12 Jan 2006 20:23:49 +0000 (12:23 -0800)] 
Merge master.kernel.org:/home/rmk/linux-2.6-arm

18 years agoMerge master.kernel.org:/home/rmk/linux-2.6-serial
Linus Torvalds [Thu, 12 Jan 2006 20:23:03 +0000 (12:23 -0800)] 
Merge master.kernel.org:/home/rmk/linux-2.6-serial

18 years agoMerge master.kernel.org:/home/rmk/linux-2.6-mmc
Linus Torvalds [Thu, 12 Jan 2006 20:22:17 +0000 (12:22 -0800)] 
Merge master.kernel.org:/home/rmk/linux-2.6-mmc