pandora-kernel.git
13 years agoia64: iosapic: Cleanup irq_desc access
Thomas Gleixner [Fri, 25 Mar 2011 19:16:05 +0000 (20:16 +0100)]
ia64: iosapic: Cleanup irq_desc access

Use irq_to_desc() and use accessors for setting chip and handler.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoia64: sn: Drop unused function
Thomas Gleixner [Fri, 25 Mar 2011 18:40:39 +0000 (19:40 +0100)]
ia64: sn: Drop unused function

Leftover from the irq chip conversion.

Reported-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoia64: Convert iosapic to new irq_chip functions
Thomas Gleixner [Fri, 4 Feb 2011 19:31:21 +0000 (20:31 +0100)]
ia64: Convert iosapic to new irq_chip functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoia64: Convert lsapic to new irq_chip functions
Thomas Gleixner [Fri, 4 Feb 2011 19:29:37 +0000 (20:29 +0100)]
ia64: Convert lsapic to new irq_chip functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoia64: Convert msi to new irq_chip functions
Thomas Gleixner [Fri, 4 Feb 2011 19:18:43 +0000 (20:18 +0100)]
ia64: Convert msi to new irq_chip functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoia64: Convert sn to new irq_chip functions
Thomas Gleixner [Fri, 4 Feb 2011 19:08:36 +0000 (20:08 +0100)]
ia64: Convert sn to new irq_chip functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoia64: Convert msi_sn to new irq_chip functions
Thomas Gleixner [Fri, 4 Feb 2011 19:02:09 +0000 (20:02 +0100)]
ia64: Convert msi_sn to new irq_chip functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoia64: hp-sim: Cleanup direct access to irq_desc
Thomas Gleixner [Sat, 5 Feb 2011 19:23:29 +0000 (20:23 +0100)]
ia64: hp-sim: Cleanup direct access to irq_desc

Use accessor functions instead of open coded access.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoia64: Convert hp-sim to new irq_chip functions
Thomas Gleixner [Fri, 4 Feb 2011 19:12:06 +0000 (20:12 +0100)]
ia64: Convert hp-sim to new irq_chip functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoia64: Remove stale irq_chip.end
Thomas Gleixner [Thu, 3 Feb 2011 00:46:48 +0000 (00:46 +0000)]
ia64: Remove stale irq_chip.end

irq_chip.end got obsolete with the removal of __do_IRQ().

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tony Luck <tony.luck@intel.com>
LKML-Reference: <20110203004210.143127544@linutronix.de>

13 years agoia64: Cleanup migrate_irqs()
Thomas Gleixner [Fri, 25 Mar 2011 19:12:33 +0000 (20:12 +0100)]
ia64: Cleanup migrate_irqs()

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoia64: Convert migrate_platform_irqs() to new irq chip functions
Thomas Gleixner [Fri, 25 Mar 2011 18:00:33 +0000 (19:00 +0100)]
ia64: Convert migrate_platform_irqs() to new irq chip functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agobfin: Use proper accessors in trace
Thomas Gleixner [Mon, 28 Mar 2011 11:50:50 +0000 (13:50 +0200)]
bfin: Use proper accessors in trace

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agobfin: Convert irq namespace
Thomas Gleixner [Thu, 24 Mar 2011 16:22:30 +0000 (17:22 +0100)]
bfin: Convert irq namespace

Convert to the new function names. Scripted with coccinelle.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: uclinux-dist-devel@blackfin.uclinux.org
13 years agobfin: Final irq cleanup
Thomas Gleixner [Thu, 24 Mar 2011 16:21:01 +0000 (17:21 +0100)]
bfin: Final irq cleanup

Use the trigger type in irq_data and check level type instead of
looking at desc->handle_irq.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: uclinux-dist-devel@blackfin.uclinux.org
13 years agoalpha: Use generic show_interrupts()
Thomas Gleixner [Fri, 25 Mar 2011 21:20:51 +0000 (22:20 +0100)]
alpha: Use generic show_interrupts()

The only subtle difference is that alpha uses ACTUAL_NR_IRQS and
prints the IRQF_DISABLED flag.

Change the generic implementation to deal with ACTUAL_NR_IRQS if
defined.

The IRQF_DISABLED printing is pointless, as we nowadays run all
interrupts with irqs disabled.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoalpha: Convert to new irq function names
Thomas Gleixner [Fri, 25 Mar 2011 21:17:31 +0000 (22:17 +0100)]
alpha: Convert to new irq function names

Scripted with coccinelle.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: Fold irq_set_chip/irq_set_handler
Thomas Gleixner [Thu, 24 Mar 2011 12:35:09 +0000 (13:35 +0100)]
arm: Fold irq_set_chip/irq_set_handler

Use irq_set_chip_and_handler() instead. Converted with coccinelle.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: Reorder irq_set_ function calls
Thomas Gleixner [Thu, 24 Mar 2011 12:29:39 +0000 (13:29 +0100)]
arm: Reorder irq_set_ function calls

Reorder
irq_set_chip()
irq_set_chip_data()
irq_set_handler()

to

irq_set_chip()
irq_set_handler()
irq_set_chip_data()

so the next patch can combine irq_set_chip() and irq_set_handler() to
irq_set_chip_and_handler().

Automated conversion with coccinelle.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: Cleanup the irq namespace
Thomas Gleixner [Thu, 24 Mar 2011 12:25:22 +0000 (13:25 +0100)]
arm: Cleanup the irq namespace

Convert to the new function names. Automated with coccinelle.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: Use generic show_interrupts()
Thomas Gleixner [Thu, 24 Mar 2011 11:02:11 +0000 (12:02 +0100)]
arm: Use generic show_interrupts()

Use the generic version and just keep the arch specific output.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: dove: Use proper irq accessor functions
Thomas Gleixner [Thu, 24 Mar 2011 11:33:40 +0000 (12:33 +0100)]
arm: dove: Use proper irq accessor functions

Remove the obsolete desc_handle_irq() helper.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: gic: Use proper accessor functions
Thomas Gleixner [Thu, 24 Mar 2011 11:48:54 +0000 (12:48 +0100)]
arm: gic: Use proper accessor functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: at91: Cleanup irq chip
Thomas Gleixner [Thu, 24 Mar 2011 11:48:18 +0000 (12:48 +0100)]
arm: at91: Cleanup irq chip

Avoid the whole lazy disable dance in the demux handler by providing a
irq_disable() callback.

Use the proper accessor functions and tidy up gpio_irq_handler()

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: davinci: Cleanup irq chip code
Thomas Gleixner [Thu, 24 Mar 2011 11:47:04 +0000 (12:47 +0100)]
arm: davinci: Cleanup irq chip code

Make use of the new functionality which ensures that irq_set_type is
called with the chip masked. Unmask is only done when the interrupt is
not disabled.

Retrieve the trigger type from irq_data in unmask

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: ep93xx: Use proper irq accessor functions
Thomas Gleixner [Thu, 24 Mar 2011 11:45:56 +0000 (12:45 +0100)]
arm: ep93xx: Use proper irq accessor functions

No need to write the flow type. Core code does already.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: gemini: Use proper irq accessor functions
Thomas Gleixner [Thu, 24 Mar 2011 11:44:54 +0000 (12:44 +0100)]
arm: gemini: Use proper irq accessor functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: mxc: Use generic_handle_irq()
Thomas Gleixner [Thu, 24 Mar 2011 11:43:25 +0000 (12:43 +0100)]
arm: mxc: Use generic_handle_irq()

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: vt8500: Use proper irq accessors
Thomas Gleixner [Thu, 24 Mar 2011 11:42:50 +0000 (12:42 +0100)]
arm: vt8500: Use proper irq accessors

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: msm: Use proper irq accessor functions
Thomas Gleixner [Thu, 24 Mar 2011 11:41:27 +0000 (12:41 +0100)]
arm: msm: Use proper irq accessor functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: plat-omap: Cleanup irq_desc access
Thomas Gleixner [Thu, 24 Mar 2011 11:40:15 +0000 (12:40 +0100)]
arm: plat-omap: Cleanup irq_desc access

1) Core code stores the flow type already
2) Flow type is accessible in irq_data

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: nomadik: Use local irq state
Thomas Gleixner [Thu, 24 Mar 2011 11:38:50 +0000 (12:38 +0100)]
arm: nomadik: Use local irq state

Store the enabled mask in the local state, so there is no need to
fiddle in the irq descriptor.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: stmp3xxx: Use generic_handle_irq()
Thomas Gleixner [Thu, 24 Mar 2011 11:38:17 +0000 (12:38 +0100)]
arm: stmp3xxx: Use generic_handle_irq()

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: plat-samsung: Use proper irq accessor functions
Thomas Gleixner [Thu, 24 Mar 2011 11:36:46 +0000 (12:36 +0100)]
arm: plat-samsung: Use proper irq accessor functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: Cleanup irq_desc access
Thomas Gleixner [Thu, 24 Mar 2011 11:35:19 +0000 (12:35 +0100)]
arm: Cleanup irq_desc access

Use the proper wrappers and use the flow type in irq_data.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: msm: Convert to new irq chip functions
Thomas Gleixner [Thu, 24 Mar 2011 10:58:31 +0000 (11:58 +0100)]
arm: msm: Convert to new irq chip functions

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: Use irq flag setter function
Thomas Gleixner [Mon, 7 Feb 2011 21:30:49 +0000 (22:30 +0100)]
arm: Use irq flag setter function

Use the proper accessor function instead of fiddling in the status
bits directly.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arm-kernel@lists.infradead.org
13 years agoarm: Use genirq lockdep helper to set lock class
Thomas Gleixner [Tue, 22 Mar 2011 16:11:09 +0000 (17:11 +0100)]
arm: Use genirq lockdep helper to set lock class

Remove the open coded access to irq_desc which will fail on sparse irq
and use the proper wrappers.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: tegra: Remove unused bogus irq enable/disable magic
Thomas Gleixner [Wed, 9 Feb 2011 15:46:14 +0000 (16:46 +0100)]
arm: tegra: Remove unused bogus irq enable/disable magic

The core code handles thees already.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Colin Cross <ccross@android.com>
Cc: linux-tegra@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
13 years agoarm: h720x: Fix irq conversion fallout
Thomas Gleixner [Mon, 28 Mar 2011 08:54:40 +0000 (10:54 +0200)]
arm: h720x: Fix irq conversion fallout

The conversion missed, that one of the irq functions is called from
the init code. Split it out, so the irq number based call works.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: nomadik: Remove non existing cpu id check
Thomas Gleixner [Mon, 28 Mar 2011 11:12:20 +0000 (13:12 +0200)]
arm: nomadik: Remove non existing cpu id check

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: ns9xxx: Remove non exisiting machine checks
Thomas Gleixner [Mon, 28 Mar 2011 10:11:46 +0000 (12:11 +0200)]
arm: ns9xxx: Remove non exisiting machine checks

The machine id cleanup missed to remove the checks for now removed
ids.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: dove: Use correct GPIO_BASE and remove orion_gpio_init() leftover
Thomas Gleixner [Mon, 28 Mar 2011 09:26:09 +0000 (11:26 +0200)]
arm: dove: Use correct GPIO_BASE and remove orion_gpio_init() leftover

commit 9eac6d0 (ARM: Remove dependency of plat-orion GPIO code on mach
directory includes) missed to convert one instance of
DOVE_GPIO_VIRT_BASE and left the orion_gpio_init() in mpp.c

Fix it up.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: at91: at572d940hf: Fix SDRAMC define
Thomas Gleixner [Mon, 28 Mar 2011 10:32:39 +0000 (12:32 +0200)]
arm: at91: at572d940hf: Fix SDRAMC define

That wants to be AT91_SDRAMC0

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: footbridge: Make cksrc_dc21285_disable() void
Thomas Gleixner [Mon, 28 Mar 2011 09:25:40 +0000 (11:25 +0200)]
arm: footbridge: Make cksrc_dc21285_disable() void

This clocksource function needs to be void.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: bios32: Remove non exisiting machine code
Thomas Gleixner [Sun, 27 Mar 2011 17:45:00 +0000 (19:45 +0200)]
arm: bios32: Remove non exisiting machine code

The id removal left this machine check in which breaks the build on
some platforms. Remove it.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agowatchdog: softdog.c: enhancement to optionally invoke panic instead of reboot on...
Anithra P Janakiraman [Mon, 28 Mar 2011 21:29:19 +0000 (14:29 -0700)]
watchdog: softdog.c: enhancement to optionally invoke panic instead of reboot on timer expiry

This is needed for determining the reason for failure when a softdog
timeout occurs.

We use softdog to watch for critical application failures and at the
minimum a snapshot of the system would help to determine the cause.  In
such a scenario the application could fail but there isn't a softlockup as
such, hence the detect softlockup feature does not help.

The patch adds a module parameter soft_panic which when set to 1 causes
softdog to invoke panic instead of reboot when the softdog timer expires.
By invoking panic we execute kdump if it is configured and the vmcore
generated by kdump should provide atleast a minimal idea of the reason for
failure.

Based on an original patch by Ken Sugawara <sugaken.r3@gmail.com>
Signed-off-by: Anithra P J <anithra@linux.vnet.ibm.com>
Reviewed-by: WANG Cong <xiyou.wangcong@gmail.com>
Acked-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
13 years agowatchdog: fix nv_tco section mismatch
Randy Dunlap [Thu, 24 Mar 2011 20:32:44 +0000 (13:32 -0700)]
watchdog: fix nv_tco section mismatch

Fix section mismatch warning:
Mark the called nv_tco_getdevice() as __devinit, just like its caller.

WARNING: drivers/watchdog/nv_tco.o(.devinit.text+0x16): Section mismatch in reference from the function nv_tco_init() to the function .init.text:nv_tco_getdevice()
The function __devinit nv_tco_init() references
a function __init nv_tco_getdevice().
If nv_tco_getdevice is only used by nv_tco_init then
annotate nv_tco_getdevice with a matching annotation.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
13 years agowatchdog: sp5100_tco.c: Check if firmware has set correct value in tcobase.
Yinghai Lu [Thu, 17 Mar 2011 03:01:07 +0000 (20:01 -0700)]
watchdog: sp5100_tco.c: Check if firmware has set correct value in tcobase.

Stefano found SP5100 TCO watchdog driver using wrong address.

[    9.148536] SP5100 TCO timer: SP5100 TCO WatchDog Timer Driver v0.01
[    9.148628] DEBUG __ioremap_caller WARNING address=b8fe00 size=8 valid=1 reserved=1

and e820 said that range is RAM.

We should check if we can use that reading out. BIOS could just program wrong address there.

Reported-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by:Yinghai Lu <yinghai@kernel.org>
Acked-by: Mike Waychison <mikew@google.com>
Tested-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Cc: stable <stable@kernel.org>
13 years agowatchdog: Convert release_resource to release_region/release_mem_region
Julia Lawall [Sat, 26 Feb 2011 16:34:39 +0000 (17:34 +0100)]
watchdog: Convert release_resource to release_region/release_mem_region

Request_mem_region should be used with release_mem_region, not
release_resource.

In pnx4008_wdt.c, a missing clk_put is added as well.

The semantic match that finds the first problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression x,E;
@@
*x = request_mem_region(...)
... when != release_mem_region(x)
    when != x = E
* release_resource(x);
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Cc: stable <stable@kernel.org>
13 years agowatchdog: s3c2410_wdt.c: Convert release_resource to release_region/release_mem_region
Julia Lawall [Sat, 26 Feb 2011 16:34:38 +0000 (17:34 +0100)]
watchdog: s3c2410_wdt.c: Convert release_resource to release_region/release_mem_region

Request_mem_region should be used with release_mem_region, not
release_resource.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression x,E;
@@
*x = request_mem_region(...)
... when != release_mem_region(x)
    when != x = E
* release_resource(x);
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Cc: stable <stable@kernel.org>
13 years agogenirq: Fix harmless typo
Thomas Gleixner [Tue, 29 Mar 2011 09:36:05 +0000 (11:36 +0200)]
genirq: Fix harmless typo

The late night fixup missed to convert the data type from irq_desc to
irq_data, which results in a harmless but annoying warning.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoxfrm: Restrict extended sequence numbers to esp
Steffen Klassert [Mon, 28 Mar 2011 19:48:09 +0000 (19:48 +0000)]
xfrm: Restrict extended sequence numbers to esp

The IPsec extended sequence numbers are fully implemented just for
esp. So restrict the usage to esp until other protocols have
support too.

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoxfrm: Check for esn buffer len in xfrm_new_ae
Steffen Klassert [Mon, 28 Mar 2011 19:47:30 +0000 (19:47 +0000)]
xfrm: Check for esn buffer len in xfrm_new_ae

In xfrm_new_ae() we may overwrite the allocated esn replay state
buffer with a wrong size. So check that the new size matches the
original allocated size and return an error if this is not the case.

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoxfrm: Assign esn pointers when cloning a state
Steffen Klassert [Mon, 28 Mar 2011 19:46:39 +0000 (19:46 +0000)]
xfrm: Assign esn pointers when cloning a state

When we clone a xfrm state we have to assign the replay_esn
and the preplay_esn pointers to the state if we use the
new replay detection method. To this end, we add a
xfrm_replay_clone() function that allocates memory for
the replay detection and takes over the necessary values
from the original state.

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoxfrm: Move the test on replay window size into the replay check functions
Steffen Klassert [Mon, 28 Mar 2011 19:45:52 +0000 (19:45 +0000)]
xfrm: Move the test on replay window size into the replay check functions

As it is, the replay check is just performed if the replay window of the
legacy implementation is nonzero. So we move the test on a nonzero replay
window inside the replay check functions to be sure we are testing for the
right implementation.

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonetdev: bfin_mac: document TE setting in RMII modes
Mike Frysinger [Sun, 27 Mar 2011 22:33:13 +0000 (22:33 +0000)]
netdev: bfin_mac: document TE setting in RMII modes

The current code sometimes generates build warnings due to how it checks
the silicon revision, so clean it up and properly document things.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agodrivers net: Fix declaration ordering in inline functions.
Balaji G [Sun, 27 Mar 2011 17:15:19 +0000 (17:15 +0000)]
drivers net: Fix declaration ordering in inline functions.

The correct usage should be "static inline void" instead of "static void inline"

Signed-off-by: G.Balaji <balajig81@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agocxgb3: Apply interrupt coalescing settings to all queues
Anton Blanchard [Sun, 27 Mar 2011 16:50:41 +0000 (16:50 +0000)]
cxgb3: Apply interrupt coalescing settings to all queues

While testing the performance of different receive interrupt
coalescing settings on a single stream TCP benchmark, I noticed two
very different results. With rx-usecs=50, most of the time a
connection would hit 8280 Mbps but once in a while it would hit
9330 Mbps.

It turns out we are only applying the interrupt coalescing settings
to the first queue and whenever the rx hash would direct us onto
that queue we ran faster.

With this patch applied and rx-usecs=50, I get 9330 Mbps
consistently.

Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: Divy Le Ray <divy@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet: Always allocate at least 16 skb frags regardless of page size
Anton Blanchard [Sun, 27 Mar 2011 14:57:26 +0000 (14:57 +0000)]
net: Always allocate at least 16 skb frags regardless of page size

When analysing performance of the cxgb3 on a ppc64 box I noticed that
we weren't doing much GRO merging. It turns out we are limited by the
number of SKB frags:

#define MAX_SKB_FRAGS (65536/PAGE_SIZE + 2)

With a 4kB page size we have 18 frags, but with a 64kB page size we
only have 3 frags.

I ran a single stream TCP bandwidth test to compare the performance of

Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomd: Fix integrity registration error when no devices are capable
Martin K. Petersen [Tue, 29 Mar 2011 00:09:12 +0000 (20:09 -0400)]
md: Fix integrity registration error when no devices are capable

We incorrectly returned -EINVAL when none of the devices in the array
had an integrity profile.  This in turn prevented mdadm from starting
the metadevice.  Fix this so we only return errors on mismatched
profiles and memory allocation failures.

Reported-by: Giacomo Catenazzi <cate@cateee.net>
Reported-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris...
Linus Torvalds [Tue, 29 Mar 2011 00:51:27 +0000 (17:51 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jmorris/security-testing-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6:
  selinux: Fix regression for Xorg

13 years agoARM: mx51_efika: fix build error due to new mfd changes
Andres Salomon [Tue, 29 Mar 2011 00:22:30 +0000 (17:22 -0700)]
ARM: mx51_efika: fix build error due to new mfd changes

MFD changes in 4ec1b54c ('mfd: mfd_cell is now implicitly available to
mc13xxx drivers') changed the mc13xxx_platform_data struct layout.

At the time all users were changed, but this driver was introduced in
another tree at the same time.  This updates the mc13xxx_platform_data
user, fixing a build error.

Signed-off-by: Andres Salomon <dilinger@queued.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge branches 'irq-cleanup-for-linus' and 'irq-fixes-for-linus' of git://git.kernel...
Linus Torvalds [Tue, 29 Mar 2011 00:39:54 +0000 (17:39 -0700)]
Merge branches 'irq-cleanup-for-linus' and 'irq-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'irq-cleanup-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  vlynq: Convert irq functions

* 'irq-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  genirq; Fix cleanup fallout
  genirq: Fix typo and remove unused variable
  genirq: Fix new kernel-doc warnings
  genirq: Add setter for AFFINITY_SET in irq_data state
  genirq: Provide setter inline for IRQD_IRQ_INPROGRESS
  genirq: Remove handle_IRQ_event
  arm: Ns9xxx: Remove private irq flow handler
  powerpc: cell: Use the core flow handler
  genirq: Provide edge_eoi flow handler
  genirq: Move INPROGRESS, MASKED and DISABLED state flags to irq_data
  genirq: Split irq_set_affinity() so it can be called with lock held.
  genirq: Add chip flag for restricting cpu_on/offline calls
  genirq: Add chip hooks for taking CPUs on/off line.
  genirq: Add irq disabled flag to irq_data state
  genirq: Reserve the irq when calling irq_set_chip()

13 years agoipv4: Don't ip_rt_put() an error pointer in RAW sockets.
David S. Miller [Mon, 28 Mar 2011 23:51:15 +0000 (16:51 -0700)]
ipv4: Don't ip_rt_put() an error pointer in RAW sockets.

Reported-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agogenirq; Fix cleanup fallout
Thomas Gleixner [Mon, 28 Mar 2011 19:59:37 +0000 (21:59 +0200)]
genirq; Fix cleanup fallout

I missed the CONFIG_GENERIC_PENDING_IRQ dependency in the affinity
related functions and the IRQ_LEVEL propagation into irq_data
state. Did not pop up on my main test platforms. :(

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: David Daney <ddaney@caviumnetworks.com>
13 years agoselinux: Fix regression for Xorg
Stephen Smalley [Fri, 25 Mar 2011 14:13:43 +0000 (10:13 -0400)]
selinux: Fix regression for Xorg

Commit 6f5317e730505d5cbc851c435a2dfe3d5a21d343 introduced a bug in the
handling of userspace object classes that is causing breakage for Xorg
when XSELinux is enabled.  Fix the bug by changing map_class() to return
SECCLASS_NULL when the class cannot be mapped to a kernel object class.

Reported-by: "Justin P. Mattock" <justinmattock@gmail.com>
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@namei.org>
13 years agomemstick: make enable_dma less generic in r592
Stephen Rothwell [Mon, 28 Mar 2011 23:24:29 +0000 (16:24 -0700)]
memstick: make enable_dma less generic in r592

Fixes this build error:

  drivers/memstick/host/r592.c:26: error: 'enable_dma' redeclared as different kind of symbol
  arch/powerpc/include/asm/dma.h:189: note: previous definition of 'enable_dma' was here

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Alex Dubov <oakad@yahoo.com>
Acked-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Mon, 28 Mar 2011 22:51:02 +0000 (15:51 -0700)]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: stop using the page cache to back the buffer cache
  xfs: register the inode cache shrinker before quotachecks
  xfs: xfs_trans_read_buf() should return an error on failure
  xfs: introduce inode cluster buffer trylocks for xfs_iflush
  vmap: flush vmap aliases when mapping fails
  xfs: preallocation transactions do not need to be synchronous

Fix up trivial conflicts in fs/xfs/linux-2.6/xfs_buf.c due to plug removal.

13 years agoRelax si_code check in rt_sigqueueinfo and rt_tgsigqueueinfo
Roland Dreier [Mon, 28 Mar 2011 21:13:35 +0000 (14:13 -0700)]
Relax si_code check in rt_sigqueueinfo and rt_tgsigqueueinfo

Commit da48524eb206 ("Prevent rt_sigqueueinfo and rt_tgsigqueueinfo
from spoofing the signal code") made the check on si_code too strict.
There are several legitimate places where glibc wants to queue a
negative si_code different from SI_QUEUE:

 - This was first noticed with glibc's aio implementation, which wants
   to queue a signal with si_code SI_ASYNCIO; the current kernel
   causes glibc's tst-aio4 test to fail because rt_sigqueueinfo()
   fails with EPERM.

 - Further examination of the glibc source shows that getaddrinfo_a()
   wants to use SI_ASYNCNL (which the kernel does not even define).
   The timer_create() fallback code wants to queue signals with SI_TIMER.

As suggested by Oleg Nesterov <oleg@redhat.com>, loosen the check to
forbid only the problematic SI_TKILL case.

Reported-by: Klaus Dittrich <kladit@arcor.de>
Acked-by: Julien Tinnes <jln@google.com>
Cc: <stable@kernel.org>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge branch 'for-linus' of git://www.jni.nu/cris
Linus Torvalds [Mon, 28 Mar 2011 22:44:06 +0000 (15:44 -0700)]
Merge branch 'for-linus' of git://www.jni.nu/cris

* 'for-linus' of git://www.jni.nu/cris:
  Correct auto-restart of syscalls via restartblock
  CRISv10: Fix return before mutex_unlock in pcf8563
  Drop the CRISv32 version of pcf8563

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ecryptfs...
Linus Torvalds [Mon, 28 Mar 2011 22:43:25 +0000 (15:43 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/ecryptfs/ecryptfs-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ecryptfs/ecryptfs-2.6:
  eCryptfs: write lock requested keys
  eCryptfs: move ecryptfs_find_auth_tok_for_sig() call before mutex_lock
  eCryptfs: verify authentication tokens before their use
  eCryptfs: modified size of keysig in the ecryptfs_key_sig structure
  eCryptfs: removed num_global_auth_toks from ecryptfs_mount_crypt_stat
  eCryptfs: ecryptfs_keyring_auth_tok_for_sig() bug fix
  eCryptfs: Unlock page in write_begin error path
  ecryptfs: modify write path to encrypt page in writepage
  eCryptfs: Remove ECRYPTFS_NEW_FILE crypt stat flag
  eCryptfs: Remove unnecessary grow_file() function

13 years agoMerge branch 'for-linus-unmerged' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Mon, 28 Mar 2011 22:31:05 +0000 (15:31 -0700)]
Merge branch 'for-linus-unmerged' of git://git./linux/kernel/git/mason/btrfs-unstable

* 'for-linus-unmerged' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable: (45 commits)
  Btrfs: fix __btrfs_map_block on 32 bit machines
  btrfs: fix possible deadlock by clearing __GFP_FS flag
  btrfs: check link counter overflow in link(2)
  btrfs: don't mess with i_nlink of unlocked inode in rename()
  Btrfs: check return value of btrfs_alloc_path()
  Btrfs: fix OOPS of empty filesystem after balance
  Btrfs: fix memory leak of empty filesystem after balance
  Btrfs: fix return value of setflags ioctl
  Btrfs: fix uncheck memory allocations
  btrfs: make inode ref log recovery faster
  Btrfs: add btrfs_trim_fs() to handle FITRIM
  Btrfs: adjust btrfs_discard_extent() return errors and trimmed bytes
  Btrfs: make btrfs_map_block() return entire free extent for each device of RAID0/1/10/DUP
  Btrfs: make update_reserved_bytes() public
  btrfs: return EXDEV when linking from different subvolumes
  Btrfs: Per file/directory controls for COW and compression
  Btrfs: add datacow flag in inode flag
  btrfs: use GFP_NOFS instead of GFP_KERNEL
  Btrfs: check return value of read_tree_block()
  btrfs: properly access unaligned checksum buffer
  ...

Fix up trivial conflicts in fs/btrfs/volumes.c due to plug removal in
the block layer.

13 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mjg59/platf...
Linus Torvalds [Mon, 28 Mar 2011 22:16:10 +0000 (15:16 -0700)]
Merge branch 'for_linus' of git://git./linux/kernel/git/mjg59/platform-drivers-x86

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mjg59/platform-drivers-x86: (81 commits)
  xo15-ebook: Remove device.wakeup_count
  ips: use interruptible waits in ips-monitor
  acer-wmi: does not poll device status when WMI event is available
  acer-wmi: does not set persistence state by rfkill_init_sw_state
  platform-drivers: x86: fix common misspellings
  acer-wmi: use pr_<level> for messages
  asus-wmi: potential NULL dereference in show_call()
  asus-wmi: signedness bug in read_brightness()
  platform-driver-x86: samsung-laptop: make dmi_check_cb to return 1 instead of 0
  platform-driver-x86: fix wrong merge for compal-laptop.c
  msi-laptop: use pr_<level> for messages
  Platform: add Samsung Laptop platform driver
  acer-wmi: Fix WMI ID
  acer-wmi: deactive mail led when power off
  msi-laptop: send out touchpad on/off key
  acer-wmi: set the touchpad toggle key code to KEY_TOUCHPAD_TOGGLE
  platform-driver-x86: intel_mid_thermal: fix unterminated platform_device_id table
  sony-laptop: potential null dereference
  sony-laptop: handle allocation failures
  sony-laptop: return negative on failure in sony_nc_add()
  ...

13 years agoMerge branch 'for-torvalds' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw...
Linus Torvalds [Mon, 28 Mar 2011 22:14:45 +0000 (15:14 -0700)]
Merge branch 'for-torvalds' of git://git./linux/kernel/git/linusw/linux-stericsson

* 'for-torvalds' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-stericsson:
  mach-ux500: configure board for the TPS61052 regulator v2
  mach-ux500: provide ab8500 init vector
  mach-ux500: board support for AB8500 GPIO driver
  gpio: driver for 42 AB8500 GPIO pins

13 years agoMerge branch 's5p-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Mon, 28 Mar 2011 22:08:48 +0000 (15:08 -0700)]
Merge branch 's5p-fixes-for-linus' of git://git./linux/kernel/git/kgene/linux-samsung

* 's5p-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  ARM: Suspend: Fix dependency of ARCH_SUSPEND_POSSIBLE
  ARM: SAMSUNG: Fix CPU idmask
  ARM: EXYNOS4: Fix addruart macro
  ARM: EXYNOS4: Fix smsc9215 irq polarity on SMDKC210
  ARM: EXYNOS4: Fix smsc9215 irq polarity on SMDKV310
  ARM: EXYNOS4: Fix build warning on regarding SATA_AHCI_PLATFORM
  ARM: S5PV210: Remove duplicated inclusion
  ARM: S5PV210: Fix security engine interrupt names
  ARM: S5P64X0: Fix iodesc array size for S5P6450

13 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Mon, 28 Mar 2011 20:07:49 +0000 (13:07 -0700)]
Merge git://git./linux/kernel/git/herbert/crypto-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: aesni-intel - fixed problem with packets that are not multiple of 64bytes

13 years agoMerge branch 'for-linus' of git://gitserver.sunplusct.com/linux-2.6-score
Linus Torvalds [Mon, 28 Mar 2011 20:06:32 +0000 (13:06 -0700)]
Merge branch 'for-linus' of git://gitserver.sunplusct.com/linux-2.6-score

* 'for-linus' of git://gitserver.sunplusct.com/linux-2.6-score:
  score: Use generic show_interrupts()
  score: Convert to new irq function names
  score: lost a semicolon in asm/irqflags.h
  score: Select GENERIC_HARDIRQS_NO_DEPRECATED
  score: Convert irq_chip to new functions

13 years agoacpi: Remove sysfs_create_link from video driver
Matthew Garrett [Mon, 28 Mar 2011 12:46:02 +0000 (12:46 +0000)]
acpi: Remove sysfs_create_link from video driver

The acpi video driver attempts to explicitly create a sysfs link between
the acpi device and the associated PCI device. However, we're now also
doing this from the backlight core, which means that we get a backtrace
caused by a duplicate file. Remove the code and leave it up to the
backlight core.

Reported-by: Jeff Chua <jeff.chua.linux@gmail.com>
Signed-off-by: Matthew Garrett <mjg@redhat.com>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Tested-by: Alessandro Suardi <alessandro.suardi@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec...
Linus Torvalds [Mon, 28 Mar 2011 20:03:31 +0000 (13:03 -0700)]
Merge branch 'upstream-linus' of git://git./linux/kernel/git/jlbec/ocfs2

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2: (39 commits)
  Treat writes as new when holes span across page boundaries
  fs,ocfs2: Move o2net_get_func_run_time under CONFIG_OCFS2_FS_STATS.
  ocfs2/dlm: Move kmalloc() outside the spinlock
  ocfs2: Make the left masklogs compat.
  ocfs2: Remove masklog ML_AIO.
  ocfs2: Remove masklog ML_UPTODATE.
  ocfs2: Remove masklog ML_BH_IO.
  ocfs2: Remove masklog ML_JOURNAL.
  ocfs2: Remove masklog ML_EXPORT.
  ocfs2: Remove masklog ML_DCACHE.
  ocfs2: Remove masklog ML_NAMEI.
  ocfs2: Remove mlog(0) from fs/ocfs2/dir.c
  ocfs2: remove NAMEI from symlink.c
  ocfs2: Remove masklog ML_QUOTA.
  ocfs2: Remove mlog(0) from quota_local.c.
  ocfs2: Remove masklog ML_RESERVATIONS.
  ocfs2: Remove masklog ML_XATTR.
  ocfs2: Remove masklog ML_SUPER.
  ocfs2: Remove mlog(0) from fs/ocfs2/heartbeat.c
  ocfs2: Remove mlog(0) from fs/ocfs2/slot_map.c
  ...

Fix up trivial conflict in fs/ocfs2/super.c

13 years agogenirq: Fix typo and remove unused variable
Thomas Gleixner [Mon, 28 Mar 2011 18:28:56 +0000 (20:28 +0200)]
genirq: Fix typo and remove unused variable

Sigh, I'm overworked.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agogenirq: Fix new kernel-doc warnings
Randy Dunlap [Fri, 18 Mar 2011 16:33:56 +0000 (09:33 -0700)]
genirq: Fix new kernel-doc warnings

Fix new irq-related kernel-doc warnings in 2.6.38:

Warning(kernel/irq/manage.c:149): No description found for parameter 'mask'
Warning(kernel/irq/manage.c:149): Excess function parameter 'cpumask' description in 'irq_set_affinity'
Warning(include/linux/irq.h:161): No description found for parameter 'state_use_accessors'
Warning(include/linux/irq.h:161): Excess struct/union/enum/typedef member 'state_use_accessor' description in 'irq_data'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
LKML-Reference: <20110318093356.b939558d.randy.dunlap@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agovlynq: Convert irq functions
Thomas Gleixner [Fri, 25 Mar 2011 20:56:37 +0000 (21:56 +0100)]
vlynq: Convert irq functions

Convert to the new irq_chip functions and the new namespace.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Florian Fainelli <florian@openwrt.org>
LKML-Reference: <alpine.LFD.2.00.1103252150180.31464@localhost6.localdomain6>

13 years agoTreat writes as new when holes span across page boundaries
Goldwyn Rodrigues [Thu, 17 Feb 2011 15:44:40 +0000 (09:44 -0600)]
Treat writes as new when holes span across page boundaries

When a hole spans across page boundaries, the next write forces
a read of the block. This could end up reading existing garbage
data from the disk in ocfs2_map_page_blocks. This leads to
non-zero holes. In order to avoid this, mark the writes as new
when the holes span across page boundaries.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.de>
Signed-off-by: jlbec <jlbec@evilplan.org>
13 years agoMerge branch 'mlog_replace_for_39' of git://repo.or.cz/taoma-kernel into ocfs2-merge...
Joel Becker [Mon, 28 Mar 2011 16:44:26 +0000 (09:44 -0700)]
Merge branch 'mlog_replace_for_39' of git://repo.or.cz/taoma-kernel into ocfs2-merge-window-fix

13 years agofs,ocfs2: Move o2net_get_func_run_time under CONFIG_OCFS2_FS_STATS.
Rakib Mullick [Fri, 18 Mar 2011 11:01:08 +0000 (17:01 +0600)]
fs,ocfs2: Move o2net_get_func_run_time under CONFIG_OCFS2_FS_STATS.

When CONFIG_DEBUG_FS=y and CONFIG_OCFS2_FS_STATS=n, we get the
following warning:

fs/ocfs2/cluster/tcp.c:213:16: warning: ‘o2net_get_func_run_time’
defined but not used

Since o2net_get_func_run_time is only called from
o2net_update_recv_stats, so move it under CONFIG_OCFS2_FS_STATS.

Signed-off-by: Rakib Mullick <rakib.mullick@gmail.com>
Signed-off-by: jlbec <jlbec@evilplan.org>
13 years agogenirq: Add setter for AFFINITY_SET in irq_data state
Thomas Gleixner [Mon, 28 Mar 2011 15:11:13 +0000 (17:11 +0200)]
genirq: Add setter for AFFINITY_SET in irq_data state

Some archs want to prevent the default affinity being set on their
chips in the reqeust_irq() path.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agogenirq: Provide setter inline for IRQD_IRQ_INPROGRESS
Thomas Gleixner [Mon, 28 Mar 2011 14:41:14 +0000 (16:41 +0200)]
genirq: Provide setter inline for IRQD_IRQ_INPROGRESS

Special function for demultiplexing handlers which can be disabled via
disable_irq().

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agogenirq: Remove handle_IRQ_event
Thomas Gleixner [Mon, 28 Mar 2011 14:27:31 +0000 (16:27 +0200)]
genirq: Remove handle_IRQ_event

Last user gone.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoarm: Ns9xxx: Remove private irq flow handler
Thomas Gleixner [Wed, 2 Feb 2011 21:41:27 +0000 (21:41 +0000)]
arm: Ns9xxx: Remove private irq flow handler

handle_prio_irq is almost identical with handle_fasteoi_irq. The
subtle differences are

1) The handler checks for IRQ_DISABLED after the device handler has
   been called. In case it's set it masks the interrupt.

2) When the handler sees IRQ_DISABLED on entry it masks the interupt
   in the same way as handle_fastoei_irq, but does not set the
   IRQ_PENDING flag.

3) Instead of gracefully handling a recursive interrupt it crashes the
   kernel.

#1 is just relevant when a device handler calls disable_irq_nosync()
   and it does not matter whether we mask the interrupt right away or
   not. We handle lazy masking for disable_irq anyway, so there is no
   real reason to have this extra mask in place.

#2 will prevent the resend of a pending interrupt, which can result in
   lost interrupts for edge type interrupts. For level type interrupts
   the resend is a noop in the generic code. According to the
   datasheet all interrupts are level type, so marking them as such
   will result in the exact same behaviour as the private
   handle_prio_irq implementation.

#3 is just stupid. Crashing the kernel instead of handling a problem
   gracefully is just wrong. With the current semantics- all handlers
   run with interrupts disabled - this is even more wrong.

Rename ack to eoi, remove the unused mask_ack, switch to
handle_fasteoi_irq and remove the private function.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Acked-by: Uwe Kleine-Koenig <u.kleine-koenig@pengutronix.de>
Cc: linux-arm-kernel@lists.infradead.org
LKML-Reference: <20110202212552.299898447@linutronix.de>

13 years agopowerpc: cell: Use the core flow handler
Thomas Gleixner [Mon, 28 Mar 2011 14:23:12 +0000 (16:23 +0200)]
powerpc: cell: Use the core flow handler

The core handler is a full equivalent replacement.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agogenirq: Provide edge_eoi flow handler
Thomas Gleixner [Mon, 28 Mar 2011 14:13:24 +0000 (16:13 +0200)]
genirq: Provide edge_eoi flow handler

This is a replacment for the cell flow handler which is in the way of
cleanups. Must be selected to avoid general bloat.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agogenirq: Move INPROGRESS, MASKED and DISABLED state flags to irq_data
Thomas Gleixner [Mon, 28 Mar 2011 12:10:52 +0000 (14:10 +0200)]
genirq: Move INPROGRESS, MASKED and DISABLED state flags to irq_data

We really need these flags for some of the interrupt chips. Move it
from internal state to irq_data and provide proper accessors.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: David Daney <ddaney@caviumnetworks.com>
13 years agoMerge branch 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6
Linus Torvalds [Mon, 28 Mar 2011 14:52:58 +0000 (07:52 -0700)]
Merge branch 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6

* 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6:
  NFS: Ensure that rpc_release_resources_task() can be called twice.
  NFS: Don't leak RPC clients in NFSv4 secinfo negotiation
  NFS: Fix a hang in the writeback path

13 years agoMFD: allow cs5535-mfd to build on X86 only
Andres Salomon [Mon, 28 Mar 2011 04:26:03 +0000 (21:26 -0700)]
MFD: allow cs5535-mfd to build on X86 only

Stephen ran into the following build error:

  drivers/mfd/cs5535-mfd.c:30:22: error: asm/olpc.h: No such file or directory

olpc.h exists only on x86 (and in the future, ARM).  Rather than
wrapping the include in an #ifdef, just change cs5535-mfd to only build
on x86.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Andres Salomon <dilinger@queued.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agodrbd: fix up merge error
Linus Torvalds [Mon, 28 Mar 2011 14:42:58 +0000 (07:42 -0700)]
drbd: fix up merge error

In commit 95a0f10cddbf ("drbd: store in-core bitmap little endian,
regardless of architecture") drbd had made the sane choice to use
little-endian bitmap functions everywhere.  However, it used the
horrible old functions names from <asm-generic/bitops/le.h>, that were
never really meant to be exported.

In the meantime, things got cleaned up, and in commit c4945b9ed472
("asm-generic: rename generic little-endian bitops functions") we
renamed the LE bitops to something sane, exactly so that they could be
used in random code without people gouging their eyes out when seeing
the crazy jumble of letters that were the old internal names.

As a result the drbd thing merged cleanly (commit 8d49a77568d1: "Merge
branch 'for-2.6.39/drivers' of git://git.kernel.dk/linux-2.6-block"),
since there was no data conflict - but the end result obviously doesn't
actually compile.

Reported-and-tested-by: Ingo Molnar <mingo@elte.hu>
Cc: Jens Axboe <jaxboe@fusionio.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agommc: SDHI should depend on SUPERH || ARCH_SHMOBILE
Guennadi Liakhovetski [Mon, 28 Mar 2011 07:16:33 +0000 (09:16 +0200)]
mmc: SDHI should depend on SUPERH || ARCH_SHMOBILE

Fix build breakage on platforms, not providing readsw and writesw
functions, e.g., on x86(_64).

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Chris Ball <cjb@laptop.org>
13 years agoNOMMU: support SMP dynamic percpu_alloc
Graf Yang [Mon, 28 Mar 2011 11:53:29 +0000 (12:53 +0100)]
NOMMU: support SMP dynamic percpu_alloc

The percpu code requires more functions to be implemented in the mm core
which nommu currently does not provide.  So add inline implementations
since these are largely meaningless on nommu systems.

Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Greg Ungerer <gerg@uclinux.org>
13 years agoNOMMU: percpu should use is_vmalloc_addr().
David Howells [Mon, 28 Mar 2011 11:53:29 +0000 (12:53 +0100)]
NOMMU: percpu should use is_vmalloc_addr().

per_cpu_ptr_to_phys() uses VMALLOC_START and VMALLOC_END to determine if an
address is in the vmalloc() region or not.  This is incorrect on NOMMU as
there is no real vmalloc() capability (vmalloc() is emulated by kmalloc()).

The correct way to do this is to use is_vmalloc_addr().  This encapsulates the
vmalloc() region test in MMU mode and just returns 0 in NOMMU mode.

On FRV in NOMMU mode, the percpu compilation fails without this patch:

mm/percpu.c: In function 'per_cpu_ptr_to_phys':
mm/percpu.c:1011: error: 'VMALLOC_START' undeclared (first use in this function)
mm/percpu.c:1011: error: (Each undeclared identifier is reported only once
mm/percpu.c:1011: error: for each function it appears in.)
mm/percpu.c:1012: error: 'VMALLOC_END' undeclared (first use in this function)
mm/percpu.c:1018: warning: control reaches end of non-void function

Signed-off-by: David Howells <dhowells@redhat.com>
13 years agoMerge branch 'x86-platform-next' into x86-platform
Matthew Garrett [Mon, 28 Mar 2011 10:47:38 +0000 (06:47 -0400)]
Merge branch 'x86-platform-next' into x86-platform