This seems to be the root-cause that breaks resume on my i855gm when I
apply the "drm/i915: fixup the plane->pipe fixup code" patch. And that
code doesn't even run on my machine, so it's pure timing changes
causing the regression.

Furthermore resume has been constantly switching between working and
broken on this machine ever since kms support has been merged,
seemingly with no related change as a root cause. And always with the
same symptoms of the backlight lighting up, but the lvds panel only
displaying black.

Also, of both i855gm variants only one is in the table. And in the
past we've only ever removed entries from this quirk table because it
breaks things.

So let's just remove it - in case there's indeed a bios out there
relying on a running pipe A, we can add back in a more precise quirk
entry, like all the others (save for i830/i845).

Tested-by: Chris Wilson <chris@chris-wilson.co.uk> #855gm
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

It doesn't work since the gtt pte range sits in the middle of the mmio
bar. We didn't notice that since both my and Chris' gen2 machines
don't support PAT and hence all wc io mapping request will
automatically be demoted to uc.

This regression has been introduce in

commit edef7e68
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Fri Sep 14 11:57:47 2012 +0100

    agp/intel: Use a write-combining map for updating PTEs
Reported-by: NEgbert Eich <eich@pdx.freedesktop.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=55834Acked-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Note that just because we have n == MAX elements left, does not imply
that there are only MAX elements left in the scatterlist and so we may
not be on the last chain, and the nth element may in fact be a chain ptr.

This is exercised by the improved hangman tests and the gem_exec_big
test in i-g-t.

This regression has been introduced in

commit 9da3da66
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Fri Jun 1 15:20:22 2012 +0100

   drm/i915: Replace the array of pages with a scatterlist

v2: KISS, replace the direct lookup with a for_each_sg() [danvet]
v3: Try to be clever again.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The intention was to allow the caller to avoid a failure to queue a
request having already written commands to the ring. However, this is a
moot point as the i915_add_request() can fail for other reasons than a
mere allocation failure and those failure cases are more likely than
ENOMEM. So the overlay code already had to handle i915_add_request()
failures, and due to

commit 3bb73aba
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Fri Jul 20 12:40:59 2012 +0100

    drm/i915: Allow late allocation of request for i915_add_request()

the error handling code in intel_overlay.c was subject to causing
double-frees, as found by coverity.

Rather than further complicate i915_add_request() and callers, realise
the battle is lost and adapt intel_overlay.c to take advantage of the
late allocation of requests.

v2: Handle callers passing in a NULL seqno.
v3: Ditto. This time for sure.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

A previous patch, namely:

commit bf97b276
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Wed Apr 11 20:42:41 2012 +0200

    drm/i915: implement w/a for incorrect guarband clipping

accidentally set bit 5 in 3D_CHICKEN, which has nothing to do with
clipping.  This patch changes it to be set in 3D_CHICKEN3, where it
belongs.

The game "Dante" demonstrates random clipping issues when guardband
clipping is enabled and bit 5 of 3D_CHICKEN3 isn't set.  So the
workaround is actually necessary.

Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Oliver McFadden <oliver.mcfadden@linux.intel.com>
Acked-by: NPaul Menzel <paulepanter@users.sourceforge.net>
Signed-off-by: NKenneth Graunke <kenneth@whitecape.org>
Reviewed-by: NMika Kuoppala <mika.kuoppala@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

I've discovered this on my ivb machine while stress-testing the new
flip_tests. Only harmful effect observed is that the timestamp is a
bit bogus.

Note that this is empirical duct-tape: I've noticed that we seem to
only ever miss the very first vblank irq right after enabling the
pipe. And with this hack applied I couldn't reproduce the failure case
anywhere else any more.
Tested-by: NImre Deak <imre.deak@intel.com>
Reviewed-by: NImre Deak <imre.deak@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

starting an old X server causes a kernel BUG since commit 1b50247a:

------------[ cut here ]------------
kernel BUG at drivers/gpu/drm/i915/i915_gem.c:3661!
invalid opcode: 0000 [#1] SMP
Modules linked in: snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss uvcvideo
+videobuf2_core videodev videobuf2_vmalloc videobuf2_memops uhci_hcd ath9k mac80211 snd_hda_codec_realtek ath9k_common microcode
+ath9k_hw psmouse serio_raw sg ath cfg80211 atl1c lpc_ich mfd_core ehci_hcd snd_hda_intel snd_hda_codec snd_hwdep snd_pcm rtc_cmos
+snd_timer snd evdev eeepc_laptop snd_page_alloc sparse_keymap

Pid: 2866, comm: X Not tainted 3.5.6-rc1-eeepc #1 ASUSTeK Computer INC. 1005HA/1005HA
EIP: 0060:[<c12dc291>] EFLAGS: 00013297 CPU: 0
EIP is at i915_gem_entervt_ioctl+0xf1/0x110
EAX: f5941df4 EBX: f5940000 ECX: 00000000 EDX: 00020000
ESI: f5835400 EDI: 00000000 EBP: f51d7e38 ESP: f51d7e20
 DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
CR0: 8005003b CR2: b760e0a0 CR3: 351b6000 CR4: 000007d0
DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
DR6: ffff0ff0 DR7: 00000400
Process X (pid: 2866, ti=f51d6000 task=f61af8d0 task.ti=f51d6000)
Stack:
 00000001 00000000 f5835414 f51d7e84 f5835400 f54f85c0 f51d7f10 c12b530b
 00000001 c151b139 c14751b6 c152e030 00000b32 00006459 00000059 0000e200
 00000001 00000000 00006459 c159ddd0 c12dc1a0 ffffffea 00000000 00000000
Call Trace:
 [<c12b530b>] drm_ioctl+0x2eb/0x440
 [<c12dc1a0>] ? i915_gem_init+0xe0/0xe0
 [<c1052b2b>] ? enqueue_hrtimer+0x1b/0x50
 [<c1053321>] ? __hrtimer_start_range_ns+0x161/0x330
 [<c10530b3>] ? lock_hrtimer_base+0x23/0x50
 [<c1053163>] ? hrtimer_try_to_cancel+0x33/0x70
 [<c12b5020>] ? drm_version+0x90/0x90
 [<c10ca171>] vfs_ioctl+0x31/0x50
 [<c10ca2e4>] do_vfs_ioctl+0x64/0x510
 [<c10535de>] ? hrtimer_nanosleep+0x8e/0x100
 [<c1052c20>] ? update_rmtp+0x80/0x80
 [<c10ca7c9>] sys_ioctl+0x39/0x60
 [<c1433949>] syscall_call+0x7/0xb
Code: 83 c4 0c 5b 5e 5f 5d c3 c7 44 24 04 2c 05 53 c1 c7 04 24 6f ef 47 c1 e8 6e e0 fd ff c7 83 38 1e 00 00 00 00 00 00 e9 3f ff ff
+ff <0f> 0b eb fe 0f 0b eb fe 8d b4 26 00 00 00 00 0f 0b eb fe 8d b6
EIP: [<c12dc291>] i915_gem_entervt_ioctl+0xf1/0x110 SS:ESP 0068:f51d7e20
---[ end trace dd332ec083cbd513 ]---

The crash happens here in i915_gem_entervt_ioctl() :

    3659          BUG_ON(!list_empty(&dev_priv->mm.active_list));
    3660          BUG_ON(!list_empty(&dev_priv->mm.flushing_list));
 -> 3661          BUG_ON(!list_empty(&dev_priv->mm.inactive_list));
    3662          mutex_unlock(&dev->struct_mutex);

Quoting Chris :
  "That BUG_ON there is silly and can simply be removed. The check is to
   verify that no batches were submitted to the kernel whilst the UMS/GEM
   client was suspended - to which the BUG_ONs are a crude approximation.
   Furthermore, the checks are too late, since it means we attempted to
   program the hardware whilst it was in an invalid state, the BUG_ONs are
   the least of your concerns at that point."

Note that this regression has been introduced in

commit 1b50247a
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Apr 24 15:47:30 2012 +0100

    drm/i915: Remove the list of pinned inactive objects

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NWilly Tarreau <w@1wt.eu>
[danvet: Added note about the regressing commit and cc: stable.]
Cc: stable@vger.kernel.org
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

I can't even find how I figured this might be needed anymore. But sure
enough, the value I'm reading back on platforms doesn't match what the
docs recommends.

It seemed to fix Chris' GT1 in limited testing as well.
Tested-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@vger.kernel.org
Signed-off-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Now that we correctly generate it, this hack is no longer required (and
might actually paper over a serious bug).

pageflip timestamps are sanity check in the latest version of the flip-test
in intel-gpu-tools.

v2: Also remove the gettimeofday(&now) which is no longer used.
Noticed by Mario Kleiner.

Reviewed-by: mario.kleiner@tuebingen.mpg.de
Tested-by: NImre Deak <imre.deak@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

... since finish_page_flip needs the vblank timestamp generated
in drm_handle_vblank. Somehow all the gmch platforms get it right,
but all the pch platform irq handlers get is wrong. Hooray for copy&
pasting!

Currently this gets papered over by a gross hack in finish_page_flip.
A second patch will remove that.

Note that without this, the new timestamp sanity checks in flip_test
occasionally get tripped up, hence the cc: stable tag.

Cc: stable@vger.kernel.org
Reviewed-by: mario.kleiner@tuebingen.mpg.de
Tested-by: NImre Deak <imre.deak@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Falling into default case in vmi915_gem_fault is a bug. Be more
verbose about it.
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Subsequent threads returning EBUSY from vm_insert_pfn() was not handled
correctly. As a result concurrent access from new threads to
mmapped data caused SIGBUS.

Note that this fixes i-g-t/tests/gem_threaded_tiled_access.
Tested-by: NMika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: NDmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

In commit cdb0e95b
Author: Keith Packard <keithp@keithp.com>
Date:   Tue Nov 1 20:00:06 2011 -0700

    drm/i915: Try harder during dp pattern 1 link training

extra passes were made to retry the same voltage and then retry a full
clock reset. However, as coverity pointed out, we never tried the full
clock reset as we broke out of the loop early.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We don't have bit 6 swizzling on VLV, so this function is easy.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NAdam Jackson <ajax@redhat.com>
Reviewed-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Note with a comment anything newer than DP 1.1a.

Obviously this needs some work still...
Signed-off-by: NAdam Jackson <ajax@redhat.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

If there's no downstream device, DPCD success is good enough.  If
there's a hotplug-capable downstream device, count the number of
connected sinks in DP_SINK_STATUS and return success if it's non-zero.
Otherwise, probe DDC and report appropriately.

v2: Check DP_SINK_STATUS instead of something unrelated to sink status.
Tested-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NAdam Jackson <ajax@redhat.com>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

v2: Fix parenthesis mismatch, spotted by Jani Nikula
Tested-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NAdam Jackson <ajax@redhat.com>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
[danvet: Fixup merge conflict and MAX_DOWNSTREAM #define as spotted by
Jani.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Sources: DP, eDP, and DP interop specs, and a VESA slideshow about DP
1.2 for the MST bits.
Tested-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NAdam Jackson <ajax@redhat.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Reviewed-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Tested-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NAdam Jackson <ajax@redhat.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This was meant to be the purpose of the
intel_crtc_wait_for_pending_flips() function which is called whilst
preparing the CRTC for a modeset or before disabling. However, as Ville
Syrjala pointed out, we set the pending flip notification on the old
framebuffer that is no longer attached to the CRTC by the time we come
to flush the pending operations. Instead, we can simply wait on the
pending unpin work to be finished on this CRTC, knowning that the
hardware has therefore finished modifying the registers, before proceeding
with our direct access.

Fixes i-g-t/flip_test on non-pch platforms. pch platforms simply
schedule the flip immediately when the pipe is disabled, leading
to other funny issues.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@vger.kernel.org
[danvet: Added i-g-t note and cc: stable]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

A side-effect of commit 7d54a904
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Fri Aug 10 10:18:10 2012 +0100

    drm/i915: Apply post-sync write for pipe control invalidates

was that only a request to emit invalidate flush would result in the
TLB being invalidated (since it requires synchronisation and so incurs a
performance penalty). However, the stated w/a for hardware contexts is
that the TLBs must be invalidated prior to a MI_SET_CONTEXT, yet the w/a
itself did not request the TLBs to be invalidated...

Note this w/a does not prevent the hard system hang I experience when
using hw contexts (with rc6 enabled) on SNB GT1.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

On the EINVAL case we don't release struct_mutex. It should be safe to
grab the lock after checking the parameters, which also resolves the
issues.
Signed-off-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The dithering introduced in

commit 3b5c78a3
Author: Adam Jackson <ajax@redhat.com>
Date:   Tue Dec 13 15:41:00 2011 -0800

    drm/i915/dp: Dither down to 6bpc if it makes the mode fit

stores the INTEL_MODE_DP_FORCE_6BPC flag in the private_flags of the
adjusted mode, while i9xx_crtc_mode_set() and ironlake_crtc_mode_set() use
the original mode, without the flag, so it would never have any
effect. However, the BPC was clamped by VBT settings, making things work by
coincidence, until that part was removed in

commit 4344b813
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Fri Aug 10 11:10:20 2012 +0200

Use adjusted_mode instead of mode when checking for
INTEL_MODE_DP_FORCE_6BPC to make the flag have effect.

v2: Don't forget to fix this in i9xx_crtc_mode_set() also, pointed out by
Daniel both before and after sending the first patch.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=47621
CC: Adam Jackson <ajax@redhat.com>
CC: stable@vger.kernel.org
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Reviewed-by: NAdam Jackson <ajax@redhat.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Finishes commit 02d71956
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Thu Aug 9 16:44:54 2012 +0200

    drm/i915: properly guard ilk ips state

The core functions were annotated with their locking requirements, but
we overlooked that they were exported, without any control over the
locking, to debugfs. So in order to enable debugfs to read the registers
without triggering sanity checks, we change the exported entry points to
properly take the required locks before calling the core routines.
Reported-by: Nyangguang <guang.a.yang@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=55304Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

... even if the actual infoframe is smaller than the maximum possible
size.

If we don't write all the 32 DIP data bytes the InfoFrame ECC may not
be correctly calculated in some cases (e.g., when changing the port),
and this will lead to black screens on HDMI monitors. The ECC value is
generated by the hardware.

I don't see how this should break anything since we're writing 0 and
that should be the correct value, so this patch should be safe.

Notice that on IVB and older we actually have 64 bytes available for
VIDEO_DIP_DATA, but only bytes 0-31 actually store infoframe data: the
others are either read-only ECC values or marked as "reserved". On HSW
we only have 32 bytes, and the ECC value is stored on its own separate
read-only register. See BSpec.

This patch fixes bug #46761, which is marked as a regression
introduced by commit 4e89ee17:
    drm/i915: set the DIP port on ibx_write_infoframe

Before commit 4e89 we were just failing to send AVI infoframes when we
needed to change the port, which can lead to black screens in some
cases. After commit 4e89 we started sending infoframes, but with a
possibly wrong ECC value. After this patch I hope we start sending
correct infoframes.

Version 2:
  - Improve commit message
  - Try to make the code more clear

Cc: stable@vger.kernel.org
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=46761Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Manual backmerge of -rc7 to resolve a silent conflict leading to
compile failure in drivers/gpu/drm/i915/intel_hdmi.c.

This is due to the bugfix in -rc7:

commit b98b6016
Author: Wang Xingchao <xingchao.wang@intel.com>
Date:   Thu Sep 13 07:43:22 2012 +0800

    drm/i915: HDMI - Clear Audio Enable bit for Hot Plug

Since this code moved around a lot in -next git put that snippet at
the wrong spot. I've tried to fix this by making the conflict explicit
by merging a version for next with:

commit 3cce574f
Author: Wang Xingchao <xingchao.wang@intel.com>
Date:   Thu Sep 13 11:19:00 2012 +0800

    drm/i915: HDMI - Clear Audio Enable bit for Hot Plug unconditionally

But that failed to solve the entire problem. To avoid pushing out
further -nightly branch to our QA where this is broken, do the
backmerge and manually add the stuff git adds to -next from the patch
in -fixes.

Note that this doesn't show up in git's merge diff (and hence is also
not handled by git rerere), which adds to the reasons why I'd like to
fix this with a verbose backmerge. The git merge diff only shows a
bunch of trivial conflicts of the "code changed in lines next to each
another" kind.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This should never happen, but the silent "return" makes me wonder
every time I try to debug InfoFrame bugs, so promote this to BUG() to
make sure people will complain if we ever break this.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Pull kbuild fixes from Michal Marek:
 "There are two more kbuild fixes for 3.6.

  One fixes a race between x86's archscripts target and the rule
  (re)building scripts/basic/fixdep.  The second is a fix for the
  previous attempt at fixing make firmware_install with make 3.82.
  This new solution should work with any version of GNU make"

* 'rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
  x86/kbuild: archscripts depends on scripts_basic
  firmware: fix directory creation rule matching with make 3.80

Pull hwmon subsystem fixes from Jean Delvare.

* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
  hwmon: (fam15h_power) Tweak runavg_range on resume
  hwmon: (coretemp) Use get_online_cpus to avoid races involving CPU hotplug
  hwmon: (via-cputemp) Use get_online_cpus to avoid races involving CPU hotplug

Pull SCSI fixes from James Bottomley:
 "This is a set of four essential fixes: two oops related (bnx2i,
  virtio-scsi), one data corruption related (hpsa) and one failure to
  boot due to interrupt routing issues (mpt2ss).

  Signed-off-by: James Bottomley <JBottomley@Parallels.com>"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  [SCSI] hpsa: fix handling of protocol error
  [SCSI] mpt2sas: Fix for issue - Unable to boot from the drive connected to HBA
  [SCSI] bnx2i: Fixed NULL ptr deference for 1G bnx2 Linux iSCSI offload
  [SCSI] scsi: virtio-scsi: Fix address translation failure of HighMem pages used by sg list

Fix potential NULL pointer dereference in edac_unregister_sysfs() on
system boot introduced in 3.6-rc1.

Since commit 7a623c03 ("edac: rewrite the sysfs code to use struct
device") edac_mc_alloc() no longer initializes embedded kobjects in
struct mem_ctl_info.  Therefore edac_mc_free() can no longer simply
decrement a kobject reference count to free the allocated memory unless
the memory controller driver module had also called edac_mc_add_mc().

Now edac_mc_free() will check if the newly embedded struct device has
been registered with sysfs before using either the standard device
release functions or freeing the data structures itself with logic
pulled out of the error path of edac_mc_alloc().

The BUG this patch resolves for me:

  BUG: unable to handle kernel NULL pointer dereference at   (null)
  EIP is at __wake_up_common+0x1a/0x6a
  Process modprobe (pid: 933, ti=f3dc6000 task=f3db9520 task.ti=f3dc6000)
  Call Trace:
    complete_all+0x3f/0x50
    device_pm_remove+0x23/0xa2
    device_del+0x34/0x142
    edac_unregister_sysfs+0x3b/0x5c [edac_core]
    edac_mc_free+0x29/0x2f [edac_core]
    e7xxx_probe1+0x268/0x311 [e7xxx_edac]
    e7xxx_init_one+0x56/0x61 [e7xxx_edac]
    local_pci_probe+0x13/0x15
  ...

Cc: Mauro Carvalho Chehab <mchehab@redhat.com>
Cc: Shaohui Xie <Shaohui.Xie@freescale.com>
Signed-off-by: NShaun Ruffell <sruffell@digium.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

coccinelle warns about:

+ drivers/edac/edac_mc.c:429:9-23: ERROR: reference preceded by free on line 429

   421         if (mci->csrows) {
 > 422                 for (chn = 0; chn < tot_channels; chn++) {
   423                         csr = mci->csrows[chn];
   424                         if (csr) {
 > 425                                 for (chn = 0; chn < tot_channels; chn++)
   426                                          kfree(csr->channels[chn]);
   427                                  kfree(csr);
   428                          }
 > 429                          kfree(mci->csrows[i]);
   430                  }
   431                  kfree(mci->csrows);
   432          }

and that code block seem to mess things up in several ways (double free, memory
leak, out-of-bound reads etc.):

L422: The iterator "chn" and bound "tot_channels" are totally wrong. Should be
      "row" and "tot_csrows" respectively. Which means either memory leak, or
      out-of-bound reads (which if does not trigger an immediate page fault
      error, will further lead to kfree() on random addresses).

L425: The inner loop is reusing the same iterator "chn" as the outer loop,
      which could lead to premature end of the outer loop, and hence memory leak.

L429: The array index 'i' in mci->csrows[i] is a temporary value used in
      previous loops, and won't change at all in the current loop. Which
      means either out-of-bound read and possibly kfree(random number), or the
      same mci->csrows[i] get freed once and again, and possibly double free
      for the kfree(csr) in L427.

L426/L427: a kfree(csr->channels) is needed in between to avoid leaking the memory.

The buggy code was introduced by commit de3910eb ("edac: change the mem
allocation scheme to make Documentation/kobject.txt happy") in the 3.6-rc1
merge window. Fix it by freeing up resources in this order:

  free csrows[i]->channels[j]
  free csrows[i]->channels
  free csrows[i]
  free csrows

CC: Mauro Carvalho Chehab <mchehab@redhat.com>
CC: Shaun Ruffell <sruffell@digium.com>
Signed-off-by: NFengguang Wu <fengguang.wu@intel.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The quirk introduced with commit
00250ec9 (hwmon: fam15h_power: fix
bogus values with current BIOSes) is not only required during driver
load but also when system resumes from suspend. The BIOS might set the
previously recommended (but unsuitable) initilization value for the
running average range register during resume.
Signed-off-by: NAndreas Herrmann <andreas.herrmann3@amd.com>
Tested-by: NAndreas Hartmann <andihartmann@01019freenet.de>
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: stable@vger.kernel.org # 3.0+

coretemp_init loops with for_each_online_cpu, adding platform_devices
and sysfs interfaces, then calls register_hotcpu_notifier.  There is a
race if a CPU is offlined or onlined after the loop, but before
register_hotcpu_notifier.  The race might result in the absence of a
platform_device+sysfs interface for an online CPU, or the presence of
a platform_device+sysfs interface for an offline CPU.  A similar race
occurs during coretemp_exit, after the module calls
unregister_hotcpu_notifier, but before it unregisters all devices, a
CPU might offline and a device for an offline CPU will exist for a
short while.

This fix surrounds for_each_online_cpu and register_hotcpu_notifier
with get_online_cpus+put_online_cpus; and surrounds
unregister_hotcpu_notifier and device unregistering with
get_online_cpus+put_online_cpus.

Build tested.
Signed-off-by: NSilas Boyd-Wickizer <sbw@mit.edu>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

via_cputemp_init loops with for_each_online_cpu, adding
platform_devices, then calls register_hotcpu_notifier.  If a CPU is
offlined between the loop and register_hotcpu_notifier, then later
onlined, via_cputemp_device_add will attempt to add platform devices
with the same ID.  A similar race occurs during via_cputemp_exit,
after the module calls unregister_hotcpu_notifier, a CPU might offline
and a device will exist for a CPU that is offline.

This fix surrounds for_each_online_cpu and register_hotcpu_notifier
with get_online_cpus+put_online_cpus; and surrounds
unregister_hotcpu_notifier and device unregistering with
get_online_cpus+put_online_cpus.

Build tested.
Signed-off-by: NSilas Boyd-Wickizer <sbw@mit.edu>
Acked-by: NHarald Welte <laforge@gnumonks.org>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

Pull MIPS fixes from Ralf Baechle:
 "Random fixes across arch/mips, essentially.

  One fix for an issue in get_user_pages_fast() which previously was
  discovered on x86, a miscalculation in the support for the MIPS MT
  hardware multithreading support, the RTC support for the Malta and a
  fix for a spurious interrupt issue that seems to bite only very
  special Malta configurations."

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
  MIPS: Malta: Don't crash on spurious interrupt.
  MIPS: Malta: Remove RTC Data Mode bootstrap breakage
  MIPS: mm: Add compound tail page _mapcount when mapped
  MIPS: CMP/SMTC: Fix tc_id calculation

Pull ARM and clkdev fixes from Russell King:
 "Two patches for clkdev which resolve the long standing issue that the
  devm_* versions were dependent on clkdev, which they shouldn't have
  been.  Instead, they're dependent on HAVE_CLK instead, which implies
  that you're providing clk_get() and clk_put().

  A small fix to the ARM decompressor to ensure that the page tables are
  properly interpreted by the CPU, and reserve syscall 378 for kcmp (the
  checksyscalls.sh script is unfortunately currently broken so arch
  maintainers aren't getting notified of new syscalls...)

  Lastly, a larger fix for an issue between the common clk subsystem and
  smp_twd which causes warnings to be spat out."

* 'fixes' of git://git.linaro.org/people/rmk/linux-arm:
  ARM: reserve syscall 378 for kcmp
  ARM: 7535/1: Reprogram smp_twd based on new common clk framework notifiers
  ARM: 7537/1: clk: Fix release in devm_clk_put()
  ARM: 7532/1: decompressor: reset SCTLR.TRE for VMSA ARMv7 cores
  ARM: 7534/1: clk: Make the managed clk functions generically available