Chris Wilson [Wed, 29 Sep 2010 11:26:37 +0000 (12:26 +0100)]
drm/i915: Avoid blocking the kworker thread on a stuck mutex
Just reschedule the retire requests again if the device is currently
busy. The request list will be pruned along other paths so will never
grow unbounded and so we can afford to miss the occasional pruning.
Chris Wilson [Fri, 24 Sep 2010 11:52:03 +0000 (12:52 +0100)]
drm/i915/sdvo: Fix GMBUSification
Besides a couple of bugs when writing more than a single byte along the
GMBUS, SDVO was completely failing whilst trying to use GMBUS, so use
bit banging instead.
Chris Wilson [Tue, 28 Sep 2010 09:07:56 +0000 (10:07 +0100)]
Revert "drm/i915: Drop ring->lazy_request"
With multiple rings generating requests independently, the outstanding
requests must also be track independently.
Reported-by: Wang Jinjin <[email protected]>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=30380 Signed-off-by: Chris Wilson <[email protected]>
Alex Deucher [Mon, 27 Sep 2010 14:57:10 +0000 (10:57 -0400)]
drm/radeon/kms: fix potential segfault in r600_ioctl_wait_idle
radeon_gem_wait_idle_ioctl can apparently get called prior to
the vram page being set up or even if accel if false, so make
sure it's valid before using it.
Should fix:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=597636
https://bugs.freedesktop.org/show_bug.cgi?id=29834
Chris Wilson [Mon, 27 Sep 2010 20:28:30 +0000 (21:28 +0100)]
drm: Prune GEM vma entries
Hook the GEM vm open/close ops into the generic drm vm open/close so
that the private vma entries are created and destroy appropriately.
Fixes the leak of the drm_vma_entries during the lifetime of the filp.
The IPS driver needs to know the current power consumption of the GMCH
in order to make decisions about when to increase or decrease the CPU
and/or GPU power envelope. So fix up the divisions to save the results
so the numbers are actually correct (contrary to some earlier comments
and code, these functions do not modify the first argument and use it
for the result).
Chris Wilson [Sun, 26 Sep 2010 10:03:27 +0000 (11:03 +0100)]
drm/i915: Convert the file mutex into a spinlock
Daniel Vetter pointed out that in this case is would be clearer and
cleaner to use a spinlock instead of a mutex to protect the per-file
request list manipulation. Make it so.
Al Viro [Sat, 25 Sep 2010 20:07:14 +0000 (21:07 +0100)]
alpha: fix hae_cache race in RESTORE_ALL
We want interrupts disabled on all paths leading to RESTORE_ALL;
otherwise, we are risking an IRQ coming between the updates of
alpha_mv->hae_cache and *alpha_mv->hae_register and set_hae()
within the IRQ getting badly confused.
RESTORE_ALL used to play with disabling IRQ itself, but that got
removed back in 2002, without making sure we had them disabled
on all paths. It's cheaper to make sure we have them disabled than
to revert to original variant...
Remove the detritus left from that commit back in 2002; we used to
need a reload of $0 and $1 since swpipl would change those, but
doing that had become pointless when we stopped doing swpipl in
there...
Larry Woodman [Fri, 24 Sep 2010 16:04:48 +0000 (12:04 -0400)]
Avoid pgoff overflow in remap_file_pages
Thomas Pollet noticed that the remap_file_pages() system call in
fremap.c has a potential overflow in the first part of the if statement
below, which could cause it to process bogus input parameters.
Specifically the pgoff + size parameters could be wrap thereby
preventing the system call from failing when it should.
The SNDRV_HDSP_IOCTL_GET_CONFIG_INFO and
SNDRV_HDSP_IOCTL_GET_CONFIG_INFO ioctls in hdspm.c and hdsp.c allow
unprivileged users to read uninitialized kernel stack memory, because
several fields of the hdsp{m}_config_info structs declared on the stack
are not altered or zeroed before being copied back to the user. This
patch takes care of it.
Daniel Vetter [Fri, 24 Sep 2010 19:14:22 +0000 (21:14 +0200)]
drm/i915: kill per-ring macros
Two macros that use a base address for HWS_PGA were missing, add them.
Also switch the remaining users of *_ACTHD to the ring-base one.
Kill the other ring-specific macros because they're now unused.
Signed-off-by: Daniel Vetter <[email protected]>
[ickle: And silence checkpatch whilst in the vicinity] Signed-off-by: Chris Wilson <[email protected]>
Chris Wilson [Sat, 25 Sep 2010 09:19:17 +0000 (10:19 +0100)]
drm/i915: Adjust hangcheck EIO semantics
Owain Ainsworth reported an issue between the interaction of the
hangcheck and userspace immediately (and permanently) falling back to
s/w rasterisation. In order to break the mutex and begin resetting the
GPU, we must abort the current operation (usually within the wait) and
climb sufficiently far back up the call chain to drop the mutex. In his
implementation, Owain has a loop within the ioctl handler to detect the
hang and then sleep until the error handler has run. I've chosen to
return to userspace and report an EAGAIN which should trigger the
userspace ioctl handler to repeat the call (simply because it felt less
invasive...). Before hitting a wedged GPU, we then wait upon completion
of the error handler.
Adrian Hunter [Sat, 25 Sep 2010 10:42:55 +0000 (12:42 +0200)]
block: prevent merges of discard and write requests
Add logic to prevent two I/O requests being merged if
only one of them is a discard. Ditto secure discard.
Without this fix, it is possible for write requests
to transform into discard requests. For example:
Submit bio 1 to discard 8 sectors from sector n
Submit bio 2 to write 8 sectors from sector n + 16
Submit bio 3 to write 8 sectors from sector n + 8
Bio 1 becomes request 1. Bio 2 becomes request 2.
Bio 3 is merged with request 2, and then subsequently
request 2 is merged with request 1 resulting in just
one I/O request which discards all 24 sectors.
Signed-off-by: Adrian Hunter <[email protected]>
(Moved the checks above the position checks /Jens)
Merge branch 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/staging
* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/staging:
x86/hwmon: pkgtemp has no dependency on PCI
MAINTAINERS: Update hwmon entry
x86/hwmon: register alternate sibling upon CPU removal
x86/hwmon: fix initialization of pkgtemp
x86/hwmon: fix initialization of coretemp
x86/hwmon: don't leak device attribute file from pkgtemp_probe() and pkgtemp_remove()
x86/hwmon: avoid deadlock on CPU removal in pkgtemp
x86/hwmon: fix module init for hotplug-but-no-device-found case
hwmon: (lis3) Fix Oops with NULL platform data
Chris Metcalf [Fri, 24 Sep 2010 21:19:20 +0000 (17:19 -0400)]
arch/tile: remove dead code from intvec_32.S
This "bpt_code" instruction was killed off in our development line a while
ago (the actual definition of bpt_code that is used is in kernel/traps.c)
but I didn't push it for 2.6.36 because it seemed harmless and I didn't
want to try to push more than absolutely necessary.
However, we recently fixed a bug in our gcc that had been causing
"-gdwarf2" not to be passed to the assembler, and passing this flag causes
an erroneous assembler failure in the presence of code in a data section,
sometimes. While we'd like to track down the bug in the assembler,
we'd also like to make sure 2.6.36 builds with the current toolchain,
so I'm removing this dead code as well.
Thomas Pollet points out that the 'end' variable is broken. It was
computed based on start/size before they were page-aligned, and as such
doesn't actually match any of the other actions we take. The overflow
test on end was also redundant, since we had already tested it with the
properly aligned version.
So just get rid of it entirely. The one remaining use for that broken
variable can just use 'start+size' like all the other cases already did.
Merge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2
* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2:
o2dlm: force free mles during dlm exit
ocfs2: Sync inode flags with ext2.
ocfs2: Move 'wanted' into parens of ocfs2_resmap_resv_bits.
ocfs2: Use cpu_to_le16 for e_leaf_clusters in ocfs2_bg_discontig_add_extent.
ocfs2: update ctime when changing the file's permission by setfacl
ocfs2/net: fix uninitialized ret in o2net_send_message_vec()
Ocfs2: Handle empty list in lockres_seq_start() for dlmdebug.c
Ocfs2: Re-access the journal after ocfs2_insert_extent() in dxdir codes.
ocfs2: Fix lockdep warning in reflink.
ocfs2/lockdep: Move ip_xattr_sem out of ocfs2_xattr_get_nolock.
* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6:
USB: update Kconfig help text for CONFIG_USB_SUSPEND
usb: musb: gadget: restart request on clearing endpoint halt
usb: musb: host: Issue a memory barrier before starting DMA
usb: musb: gadget: fix dma length in txstate
usb: musb: gadget: complete request only if data is transfered over
usb: musb: gadget: fix DMA length for OUT transfer
usb: musb: gadget: enable autoclear for OUT transfer in both DMA 0 and DMA 1
usb: musb: gadget: fix bulk IN infinit hangs in double buffer case
usb: musb: gadget: fix kernel panic if using out ep with FIFO_TXRX style
USB: fix bug in initialization of interface minor numbers
Merge branch 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6
* 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
vgaarb: trivial fix
drm: radeon cleanup fixes...
drm: fix trivial coding errors
drm: ttm sparse fixes.
drm/nouveau: fix panels using straps-based mode detection
drm/ttm: Clear the ghost cpu_writers flag on ttm_buffer_object_transfer.
drm/radeon: don't allow device to be opened if powered down
Chris Wilson [Fri, 24 Sep 2010 15:02:42 +0000 (16:02 +0100)]
drm/i915: Only hold a process-local lock whilst throttling.
Avoid cause latencies in other clients by not taking the global struct
mutex and moving the per-client request manipulation a local per-client
mutex. For example, this allows a compositor to schedule a page-flip
(through X) whilst an OpenGL application is monopolising the GPU.
Jan Beulich [Mon, 13 Sep 2010 10:23:05 +0000 (10:23 +0000)]
x86/hwmon: register alternate sibling upon CPU removal
Just like pkgtemp registers another core of the same package when one
gets removed, coretemp should register another hyperthread (if
available) in that situation.
As pointed out in the patch fixing the respective code in pkgtemp, the
list protectng mutex must be dropped before calling
coretemp_device_add(), and due to the restructured loop (including an
explicit return) the "safe" variant of the list iterator isn't needed
anymore.
Jan Beulich [Fri, 24 Sep 2010 05:21:34 +0000 (22:21 -0700)]
x86/hwmon: fix initialization of coretemp
Using cpuid_eax() to determine feature availability on other than
the current CPU is invalid. And feature availability should also be
checked in the hotplug code path.
Jan Beulich [Mon, 13 Sep 2010 10:14:43 +0000 (10:14 +0000)]
x86/hwmon: don't leak device attribute file from pkgtemp_probe() and pkgtemp_remove()
While apparently inherited from coretemp source, this particular error handling
cleanup and exit path wasn't copied properly (or perhaps got discarded
intermediately and not re-added properly later).
Jan Beulich [Mon, 13 Sep 2010 10:11:05 +0000 (10:11 +0000)]
x86/hwmon: avoid deadlock on CPU removal in pkgtemp
pkgtemp_device_remove(), holding the list protecting mutex, calls
pkgtemp_device_add(), which itself wants to acquire the same mutex.
Holding the mutex over the entire loop body in pkgtemp_device_remove()
isn't really necessary, as long as the loop gets exited after
processing the matched CPU.
Once exiting the loop after removing an eventual match, there's no
need for using the "safe" list iterator anymore.
Jan Beulich [Mon, 13 Sep 2010 10:05:51 +0000 (10:05 +0000)]
x86/hwmon: fix module init for hotplug-but-no-device-found case
In commit 0dca94baeab4a1a514841b0a4c8e3a51dfb4d5ae the call to
platform_driver_unregister() was made conditional upon !HOTPLUG_CPU,
but the return value from coretemp_init() was left to indicate an
error. This isn't correct, as the negative return value indicates to
the module loader that initialization failed, which isn't intended
here and results in dangling pointers.
Alan Stern [Fri, 24 Sep 2010 15:41:46 +0000 (11:41 -0400)]
USB: update Kconfig help text for CONFIG_USB_SUSPEND
This patch (as1429) updates the Kconfig help text for
CONFIG_USB_SUSPEND. The power/level file is now deprecated; we should
tell people to use power/control instead.
Sergei Shtylyov [Sat, 11 Sep 2010 18:23:12 +0000 (13:23 -0500)]
usb: musb: gadget: restart request on clearing endpoint halt
Commit 46034dca515bc4ddca0399ae58106d1f5f0d809f (USB: musb_gadget_ep0: stop
abusing musb_gadget_set_halt()) forgot to restart a queued request after
clearing the endpoint halt feature. This results in a couple of USB resets
while enumerating the file-backed storage gadget due to CSW packet not being
sent for the MODE SENSE(10) command.
usb: musb: host: Issue a memory barrier before starting DMA
This patch fixes the issue which was observed while transfering
a large file ( > 20MB) over USB (OMAP MUSB controller acts as USB host)
to an attached USB thumb drive.
It was found that CDB field of CBW packet was set to 0x0. This was
due to missing a barrier before DMA engine starts transfer.
This buffer is allocated using dma_alloc_coherent which gives
non-cacheble but bufferable memory and hence needed a write
memory barrier to flush the write buffer.
More info on this thread is here:
http://www.spinics.net/lists/linux-omap/msg33987.html
Ming Lei [Mon, 20 Sep 2010 07:32:03 +0000 (10:32 +0300)]
usb: musb: gadget: enable autoclear for OUT transfer in both DMA 0 and DMA 1
This patch fixes one bugs of OUT transfer in double buffer case:
-the current code only enable autoclear for dma mode 1, and not
for dma mode 0
Without this patch, test #5 of usbtest can't be passed if we
configure musb as g_zero and use fifo mode 3 to enable double
buffer mode.
With this patch and the following patch(fix dma length),
on my beagle B5, test#5(queued bulk out) may go beyond
18Mbyte/s(seems dma mode 0 is quicker in double buffer case)
if musb is configured as g_zero and fifo mode 3 is taken, follows
the test command:
Ming Lei [Mon, 20 Sep 2010 07:32:02 +0000 (10:32 +0300)]
usb: musb: gadget: fix bulk IN infinit hangs in double buffer case
This patch fixes one infinite hang of bulk IN transfer in double buffer
case, the hang can be observed easily by test #6 of usbtest if musb is
configured as g_zero and fifo mode 3 is taken to enable double fifo.
In fact, the patch only removes the check for non-empty fifo before
loading data from new request into fifo since the check is not correct:
-in double buffer case, fifo may accommodate more than one packet,
even though it has contained one packet already and is non-empty
-since last DMA is completed before calling musb_g_tx, it is sure
that fifo may accommodate at least one packet
Without applying the patch, new requst enqueued from .complte may not
have a chance to be loaded into fifo, then will never be completed and
cause infinite hangs.
With the patch, on my beagle B5, test#6(queued bulk in) can be passed and
test result may go beyond 33Mbyte/s if musb is configured as g_zero and
fifo mode 3 is taken, follows the test command:
Ming Lei [Mon, 20 Sep 2010 07:32:01 +0000 (10:32 +0300)]
usb: musb: gadget: fix kernel panic if using out ep with FIFO_TXRX style
For shared fifo hw endpoint(with FIFO_TXRX style), only ep_in
field of musb_hw_ep is intialized in musb_g_init_endpoints, and
ep_out is not initialized, but musb_g_rx and rxstate may access
ep_out field of musb_hw_ep by the method below:
musb_ep = &musb->endpoints[epnum].ep_out
which can cause the kernel panic[1] below, this patch fixes the issue
by getting 'musb_ep' from '&musb->endpoints[epnum].ep_in' for shared fifo
endpoint.
Alan Stern [Tue, 21 Sep 2010 19:01:53 +0000 (15:01 -0400)]
USB: fix bug in initialization of interface minor numbers
Recent changes in the usbhid layer exposed a bug in usbcore. If
CONFIG_USB_DYNAMIC_MINORS is enabled then an interface may be assigned
a minor number of 0. However interfaces that aren't registered as USB
class devices also have their minor number set to 0, during
initialization. As a result usb_find_interface() may return the
wrong interface, leading to a crash.
This patch (as1418) fixes the problem by initializing every
interface's minor number to -1. It also cleans up the
usb_register_dev() function, which besides being somewhat awkwardly
written, does not unwind completely on all its error paths.
Pavan Savoy [Thu, 19 Aug 2010 18:08:51 +0000 (14:08 -0400)]
Staging: ti-st: remove st_get_plat_device
In order to support multiple ST platform devices, a new symbol
'st_get_plat_device' earlier needed to be exported by the arch/XX/brd-XX.c
file which intends to add the ST platform device.
On removing this dependency, now inside ST driver maintain the array of
ST platform devices that would be registered.
As of now let id=0, as and when we end up having such platforms
where mutliple ST devices can exist, id would come from
protocol drivers (BT, FM and GPS) as to on which platform device
they want to register to.
Hette Visser [Fri, 24 Sep 2010 09:51:30 +0000 (10:51 +0100)]
drm/i915/dp: Wait for PP_CONTROL to take effect.
This patch fixes the black screen bug on Dell e6510, by
adding two delays to give the eDP panel time to turn on before we
continue with the next write.
300ms is rather arbitray and a rather long sleep, we need to find a way
of refining this value.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=29278 Signed-off-by: Chris Wilson <[email protected]> Acked-by: Jesse Barnes <[email protected]>
Jan Beulich [Fri, 24 Sep 2010 12:25:30 +0000 (13:25 +0100)]
some clean up to intel-gtt.c
In commit e517a5e97080bbe52857bd0d7df9b66602d53c4d the call to
map_page_into_agp() got removed from intel_i830_setup_flush(), but the
counterpart call from intel_i830_fini_flush() to unmap_page_from_agp()
was left in place.
Additionally, the page allocated here never gets its physical address
used for sending to hardware, so there's no need to allocate it with
GFP_DMA32. Nor is __GFP_ZERO really necessary, as the page is used
only to store data to force flushing of some internal processor state.
ALSA: hda - Fix auto-parse of SPDIF input of Realtek codecs
The SPDIF in audio widget must be searched through the list as the widget
that contains the given pin as the connection source. The current code
was implemented in a reverse way.
Ben Skeggs [Thu, 23 Sep 2010 06:37:39 +0000 (16:37 +1000)]
drm/nouveau: fix panels using straps-based mode detection
nouveau_bios_fp_mode() zeroes the mode struct before filling in relevant
entries. This nukes the mode id initialised by drm_mode_create(), and
causes warnings from idr when we try to remove the mode.
Francisco Jerez [Mon, 20 Sep 2010 14:18:28 +0000 (16:18 +0200)]
drm/nouveau: Parse old style perf tables.
Used on nv17-nv28, they contain memory clocks and timings, only one of
the table entries can actually be used, depending on the RAMCFG
straps, and it's usually higher than the frequency programmed on boot
by the BIOS.
The memory timings listed in table version 0x1x are used to init the
0x12xx range but they aren't required for reclocking to work.
projects / linux.git / log