Having missed the ENOMEM return via i915_gem_fault(), there are probably
other paths that I also missed. By not enabling NORETRY by default these
paths can run the shrinker and take memory from the system (but not from
our own inactive lists because our shrinker can not run whilst we hold
the struct mutex) and this may allow the system to survive a little longer
whilst our drivers consume all available memory.

References:
  OOM killer unexpectedly called with kernel 2.6.32
  http://bugzilla.kernel.org/show_bug.cgi?id=14933

v2: Pass gfp into page mapping.
v3: Use new read_cache_page_gfp() instead of open-coding.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Eric Anholt <eric@anholt.net>
Cc: stable@kernel.org
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

In a case where the number of the input data is bigger than the
modulus of the key, the coprocessor adapters will report an 8/72
error. This case is not caught yet, thus the adapter will be taken
offline. To prevent this, we return an -EINVAL instead.
Signed-off-by: NFelix Beck <felix.beck@de.ibm.com>
Signed-off-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>

When opening a SCLP VT220 terminal, the terminal window size is not
initialized (defaults to zero).
Since the SCLP VT220 terminal supports only 80x24, explicitly set
the window size to prevent (n)curses applications from guessing
the default setting.
Signed-off-by: NHendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>

Fix possible NULL pointer in DASD messages and correct discipline
checking.
Signed-off-by: NStefan Haberland <stefan.haberland@de.ibm.com>
Signed-off-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>

I have seen RX stalls on a machine that experienced a suspected
OOM.  After the stall, the RX buffer is empty on the guest side
and there are exactly 16 entries available on the host side.  As
the number of entries is less than that required by a maximal
skb, the host cannot proceed.

The guest did not have a refill job scheduled.

My diagnosis is that an OOM had occured, with the delayed refill
job scheduled.  The job was able to allocate at least one skb, but
not enough to overcome the minimum required by the host to proceed.

As the refill job would only reschedule itself if it failed completely
to allocate any skbs, this would lead to an RX stall.

The following patch removes this stall possibility by always
rescheduling the refill job until the ring is totally refilled.

Testing has shown that the RX stall no longer occurs whereas
previously it would occur within a day.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Acked-by: NRusty Russell <rusty@rustcorp.com.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The low-level MCDI code always uses 32-bit MMIO operations, and
callers must pad input and output buffers to multiples of 4 bytes.
The MCDI NVRAM functions are not doing this.  Also, their buffers are
declared as variable-length arrays with no explicit maximum length.

Switch to a fixed buffer size based on the chunk size used by the
MTD driver (which is a multiple of 4).
Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Due to a hardware bug in the SFC9000 family, the firmware must
transfer raw GMAC statistics to host memory before aggregating them
into the cooked (speed-independent) MAC statistics.  Extend the stats
buffer to support this.

The length of the buffer is explicit in the MAC_STATS command, so this
change is backward-compatible on both sides.
Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently qlge tries to release regions even if they were not allocated.
This causes messages like the following in the kernel log

Trying to free nonexistent resource <00000000006af400-00000000006af4ff>
Trying to free nonexistent resource <00003c04ff9f4000-00003c04ff9f7fff>
Trying to free nonexistent resource <00003c04ffc00000-00003c04ffcfffff>

This patch fixes the goto logic in order to not release the resources
if they were not allocated.
Signed-off-by: NBreno Leitao <leitao@linux.vnet.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The mdelay function was used between I/O access commands, that causes peak
in CPU usage. Fix it by substitution mdelay to msleep.

Expand usage on fitPC2 compatible boards according to DMI identification.
Signed-off-by: NDenis Turischev <denis@compulab.co.il>
Signed-off-by: NWim Van Sebroeck <wim@iguana.be>

When /dev/watchdog gets opened a second time we return -EBUSY, but
we already have got a kref then, so we end up leaking our data struct.
Signed-off-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: stable@kernel.org

MBIF (motherboard identification) is only used to print the name of
the board, it's not essential for the driver; do not fail if it's
missing. Based on Juan's patch.
Signed-off-by: NLuca Tettamanti <kronos.it@gmail.com>
Acked-by: NJuan RP <xtraeme@gmail.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

The mutex gets unlocked after we goto EXIT.
Signed-off-by: NDan Carpenter <error27@gmail.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

smsc47m1_restore is called from sm_smsc47m1_exit, which is an __exit
function, so it can't be __init.
Signed-off-by: NJeff Mahoney <jeffm@suse.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

Obviously, this register had some other impact that is causing
the regression.  Either it is masking some other access or needs
to be reset in some path.

Either, way it is best to just revert the change for 2.6.33

This reverts commit 166a0fd4.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This just adds a little more info to the warning for old -ati/mesa
userspaces.
Signed-off-by: NDave Airlie <airlied@redhat.com>

These printks aren't required anymore.
Signed-off-by: NDave Airlie <airlied@redhat.com>

commit 731b5a15
Author: James Simmons <jsimmons@infradead.org>
Date:   Thu Oct 29 20:39:07 2009 +0000

    drm/kms: properly handle fbdev blanking

uses DRM_MODE_DPMS_ON for FB_BLANK_NORMAL, but DRM_MODE_DPMS_ON
is actually for turning output on instead of blank.

This makes fb blank broken on my T61, it put LVDS on but leave
pipe disabled which made screen totally white or caused some
'burning' effect.

[airlied: James objects to this but at this point in 2.6.33,
I can't see a patch that will fix this properly like he wants coming
in time and otherwise this is a regression - proper fix for 2.6.34
hopefully.]

Cc: James Simmons <jsimmons@infradead.org>
Signed-off-by: NZhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

This is the least invasive fix without migrating the radeon driver
to pm_ops from what I can see. We just always migrate VRAM objects
on IGPs for now and we can fix it up later to migrate depending
on STR vs STD.
Signed-off-by: NDave Airlie <airlied@redhat.com>

Use VRAM whenever there is free space for DMA buffers,
but use system GMR memory if using VRAM would cause an eviction.

This significantly reduces the guest system memory usage for
VMs with a large amount of VRAM allocated.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

This is needed to fix a vmwgfx memory usage bug.
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Resending this with Thomas Hellstrom's signoff for merging into 2.6.33

ttm_bo_delayed_delete has a race condition, because after we do:
kref_put(&nentry->list_kref, ttm_bo_release_list);

we are not holding the list lock and not holding any reference to
objects, and thus every bo in the list can be removed and freed at
this point.

However, we then use the next pointer we stored, which is not guaranteed
to be valid.

This was apparently the cause of some Nouveau oopses I experienced.

This patch rewrites the function so that it keeps the reference to nentry
until nentry itself is freed and we already got a reference to nentry->next.

v2 updated by me according to Thomas Hellstrom's feedback.
v3 proposed by Thomas Hellstrom. Commit comment updated by me.

Both updates fixed minor efficiency/style issues only and all three versions
should be correct.
Signed-off-by: NLuca Barbieri <luca@luca-barbieri.com>
Signed-off-by: NThomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Another hack because of us exposing each encoder block's function as
an encoder rather than exposing a single encoder that deals with them
all.

A proper fix will come, it's just rather invasive so this hack will
do until then.
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

We need to add the buffer to the list even if we fail, otherwise the
validate_fini() call won't unreserve + unreference the GEM object,
making TTM very unhappy.
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Also adds a module option to ignore the status reported via ACPI, in case
we hit systems with broken ACPI.
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

accordingly adapt order of release_mem_region and release_mem_region on
remove.
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Cc: Ben Dooks <ben-linux@fluff.org>
Cc: Wolfram Sang <w.sang@pengutronix.de>
Cc: Richard Zhao <linuxzsc@gmail.com>
Cc: Darius Augulis <augulis.darius@gmail.com>
Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: linux-i2c@vger.kernel.org
Acked-by: NWolfram Sang <w.sang@pengutronix.de>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

After generating the stop bit by changing MSTA from 1 to 0,
the i2c_imx->stopped was immediatly set to 1. The second test
on i2c_imx->stopped then is correct and the controller never
waits if the bus is busy. This patch corrects this.

On mx31moboard, stop bit was not generated on single write transfers.
This was kept unnoticed as other transfers are made afterwards that
help the write recipient to resynchronize.

Thanks to Philippe and Michael for the debugging.
Signed-off-by: NValentin Longchamp <valentin.longchamp@epfl.ch>
Signed-off by: Philippe Rétornaz <philippe.retornaz@epfl.ch>
Reported-by: NMichael Bonani <michael.bonani@epfl.ch>
Acked-by; Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: NBen Dooks <ben-linux@fluff.org>

Fix a bad shift in the post div.

Should fix fdo bug 26145
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Only reset the reg block on the initial execute
table call; nested calls require the reg block not be
reset on each call.  Also reset the fb window and
io mode.  This matches the upstream parser behavior.
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

- split pll adjust into a separate function
- use a union for SetPixelClock params
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

- add a new flag for fixed post div
- pull the pll flags into the struct
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

This makes crtc_prepare and crtc_commit match.
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Y2_HW_WOL_ON/Y2_HW_WOL_OFF should be set and cleared per chip,
not per port.  On dual port cards, Y2_HW_WOL_ON should be
enabled if either sky2 port has WOL enabled.

Found while reviewing code for a WOL regression, though this is
probably not the cause of the regression.
Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Found this problem when testing IPv6 from a KVM guest to a remote
host via e1000e device on the host.
The following patch fixes the check for IPv6 GSO packet in Intel
ethernet drivers to use skb_is_gso_v6(). SKB_GSO_DODGY is also set
when packets are forwarded from a guest.
Signed-off-by: NSridhar Samudrala <sri@us.ibm.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

After removing the skb_dma_map/unmap calls the exception handling in
igb_tx_map_adv is not correct.  The issue is that the count value was not
being correctly handled so as a result we were not rewinding the ring as
back as we should have been.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When testing the "e1000: enhance frame fragment detection" (and e1000e)
patches we found some bugs with reducing the MTU size.  The 1024 byte
descriptor used with the 1000 mtu test also (re) introduced the
(originally) reported bug, and causes us to need the e1000_clean_tx_irq
"enhance frame fragment detection" fix.

So what has occured here is that 2.6.32 is only vulnerable for mtu <
1500 due to the jumbo specific routines in both e1000 and e1000e.
So, 2.6.32 needs the 2kB buffer len fix for those smaller MTUs, but
is not vulnerable to the original issue reported.  It has been pointed
out that this vulnerability needs to be patched in older kernels that
don't have the e1000 jumbo routine.  Without the jumbo routines, we
need the "enhance frame fragment detection" fix the e1000, old
e1000e is only vulnerable for < 1500 mtu, and needs a similar
fix.  We split the patches up to provide easy backport paths.

There is only a slight bit of extra code when this fix and the
original "enhance frame fragment detection" fixes are applied, so
please apply both, even though it is a bit of overkill.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

fmvj18x_cs, serial_cs:
 add new id
 Panasonic lan & modem card (model name:AL-VML101)
Signed-off-by: NKen Kawasaki <ken_kawasaki@spring.nifty.jp>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Only the first two fields of mcc wrb - embedded, payload_len
need to be cpu_to_le32() swapped while issuing a cmd to the hw.
The fields tag0, tag1 are opaque and returned back to cpu as is...
Signed-off-by: NSathya Perla <sathyap@serverengines.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Please add support for Microsoft MN-120 PCMCIA network card. It's an
old card, I know, but adding support is very easy. You just need to
get tulip_core.c to recognise its vendor/device ID.

Patch for kernel 2.6.32.4 (and many previous) attached.

 .....Ron Murray
Signed-off-by: NRon Murray <rjmx@rjmx.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Pages are posted to the rxq in such a way that more than one frag
can share the page. The last frag that uses the page unmaps the
page.  In the case when a page is not fully used (due to lack of space in rxq)
the last frag that uses the page is not being set as a "last_page_user";
instead, the next frag in the rxq is incorrectly being set.

The fix has also been tested on ppc64 with 64k pages...
Signed-off-by: NSathya Perla <sathyap@serverengines.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>