linux-toradex.git/arch/arm, branch v3.10.32

bpf: do not use reciprocal divide

2014-02-06T19:08:16+00:00

[ Upstream commit aee636c4809fa54848ff07a899b326eb1f9987a2 ]

At first Jakub Zawadzki noticed that some divisions by reciprocal_divide
were not correct. (off by one in some cases)
http://www.wireshark.org/~darkjames/reciprocal-buggy.c

He could also show this with BPF:
http://www.wireshark.org/~darkjames/set-and-dump-filter-k-bug.c

The reciprocal divide in linux kernel is not generic enough,
lets remove its use in BPF, as it is not worth the pain with
current cpus.

Signed-off-by: Eric Dumazet 
Reported-by: Jakub Zawadzki 
Cc: Mircea Gherzan 
Cc: Daniel Borkmann 
Cc: Hannes Frederic Sowa 
Cc: Matt Evans 
Cc: Martin Schwidefsky 
Cc: Heiko Carstens 
Cc: David S. Miller 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

ARM: at91: smc: bug fix in sam9_smc_cs_read()

2014-02-06T19:08:14+00:00

commit 1588c51cf6d782e63a8719681d905ef0ac22ee62 upstream.

There was a copy/paste error when reading the nwe_pulse value.

Signed-off-by: Jean-Jacques Hiblot 
Acked-by: Boris BREZILLON 
Signed-off-by: Nicolas Ferre 
Signed-off-by: Kevin Hilman 
Signed-off-by: Greg Kroah-Hartman

ARM: 7938/1: OMAP4/highbank: Flush L2 cache before disabling

2014-01-25T16:27:12+00:00

commit b25f3e1c358434bf850220e04f28eebfc45eb634 upstream.

Kexec disables outer cache before jumping to reboot code, but it doesn't
flush it explicitly. Flush is done implicitly inside of l2x0_disable().
But some SoC's override default .disable handler and don't flush cache.
This may lead to a corrupted memory during Kexec reboot on these
platforms.

This patch adds cache flush inside of OMAP4 and Highbank outer_cache.disable()
handlers to make it consistent with default l2x0_disable().

Acked-by: Rob Herring 
Acked-by: Santosh Shilimkar 
Acked-by: Tony Lindgren 
Signed-off-by: Taras Kondratiuk 
Signed-off-by: Russell King 
Cc: Wang Nan 
Signed-off-by: Greg Kroah-Hartman

ARM: 7815/1: kexec: offline non panic CPUs on Kdump panic

2014-01-25T16:27:11+00:00

commit 4f9b4fb7a2091eec339413a460b1665758401828 upstream.

In case of normal kexec kernel load, all cpu's are offlined
before calling machine_kexec().But in case crash panic cpus
are relaxed in machine_crash_nonpanic_core() SMP function
but not offlined.

When crash kernel is loaded with kexec and on panic trigger
machine_kexec() checks for number of cpus online.
If more than one cpu is online machine_kexec() fails to load
with below error

kexec: error: multiple CPUs still online

In machine_crash_nonpanic_core() SMP function, offline CPU
before cpu_relax

Signed-off-by: Vijaya Kumar K 
Acked-by: Stephen Warren 
Acked-by: Will Deacon 
Signed-off-by: Russell King 
Cc: l00221744 
Signed-off-by: Greg Kroah-Hartman

ARM: shmobile: mackerel: Fix coherent DMA mask

2014-01-15T23:28:51+00:00

commit b6328a6b7ba57fc84c38248f6f0e387e1170f1a8 upstream.

Commit 4dcfa60071b3d23f0181f27d8519f12e37cefbb9 ("ARM: DMA-API: better
handing of DMA masks for coherent allocations") added an additional
check to the coherent DMA mask that results in an error when the mask is
larger than what dma_addr_t can address.

Set the LCDC coherent DMA mask to DMA_BIT_MASK(32) instead of ~0 to fix
the problem.

Signed-off-by: Laurent Pinchart 
Signed-off-by: Simon Horman 
Signed-off-by: Greg Kroah-Hartman

ARM: shmobile: armadillo: Fix coherent DMA mask

2014-01-15T23:28:51+00:00

commit dcd740b645003b866d7eb30d13d34d0729cce9db upstream.

Commit 4dcfa60071b3d23f0181f27d8519f12e37cefbb9 ("ARM: DMA-API: better
handing of DMA masks for coherent allocations") added an additional
check to the coherent DMA mask that results in an error when the mask is
larger than what dma_addr_t can address.

Set the LCDC coherent DMA mask to DMA_BIT_MASK(32) instead of ~0 to fix
the problem.

Signed-off-by: Laurent Pinchart 
Signed-off-by: Simon Horman 
Signed-off-by: Greg Kroah-Hartman

ARM: shmobile: kzm9g: Fix coherent DMA mask

2014-01-15T23:28:51+00:00

commit 4f387323853c495ac589210832fad4503f75a0e7 upstream.

Commit 4dcfa60071b3d23f0181f27d8519f12e37cefbb9 ("ARM: DMA-API: better
handing of DMA masks for coherent allocations") added an additional
check to the coherent DMA mask that results in an error when the mask is
larger than what dma_addr_t can address.

Set the LCDC coherent DMA mask to DMA_BIT_MASK(32) instead of ~0 to fix
the problem.

Signed-off-by: Laurent Pinchart 
Signed-off-by: Simon Horman 
Signed-off-by: Greg Kroah-Hartman

ARM: dts: exynos5250: Fix MDMA0 clock number

2014-01-15T23:28:51+00:00

commit 8777539479abd7b3efeb691685415dc2b057d0e0 upstream.

Due to incorrect clock specified in MDMA0 node, using MDMA0 controller
could cause system failures, due to wrong clock being controlled. This
patch fixes this by specifying correct clock.

Signed-off-by: Abhilash Kesavan 
Acked-by: Mike Turquette 
[t.figa: Corrected commit message and description.]
Signed-off-by: Tomasz Figa 
Signed-off-by: Greg Kroah-Hartman

ARM: fix "bad mode in ... handler" message for undefined instructions

2014-01-15T23:28:51+00:00

commit 29c350bf28da333e41e30497b649fe335712a2ab upstream.

The array was missing the final entry for the undefined instruction
exception handler; this commit adds it.

Signed-off-by: Russell King 
Signed-off-by: Greg Kroah-Hartman

ARM: fix footbridge clockevent device

2014-01-15T23:28:50+00:00

commit 4ff859fe1dc0da0f87bbdfff78f527898878fa4a upstream.

The clockevents code was being told that the footbridge clock event
device ticks at 16x the rate which it actually does.  This leads to
timekeeping problems since it allows the clocksource to wrap before
the kernel notices.  Fix this by using the correct clock.

Fixes: 4e8d76373c9fd ("ARM: footbridge: convert to clockevents/clocksource")
Signed-off-by: Russell King 
Signed-off-by: Greg Kroah-Hartman