Commit Graph

425935 Commits

Author SHA1 Message Date
Alex Deucher
a1d6f97c8c drm/radeon/cik: enable/disable vce cg when encoding v2
Some of the vce clocks are automatic, others need to
be manually enabled.  For ease, just disable cg when
vce is active.

v2: rebased

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:46 +01:00
Alex Deucher
b9fa188376 drm/radeon: add support for vce 2.0 clock gating
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:44 +01:00
Alex Deucher
44493ba959 drm/radeon/dpm: properly enable/disable vce when vce pg is enabled
The adds the appropriate function calls to properly re-init
vce before it's used after it has been power gated.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:43 +01:00
Alex Deucher
03afe6f648 drm/radeon/dpm: enable dynamic vce state switching v2
enable vce states when vce is active.  When vce is active,
it adjusts the currently selected state (performance, battery,
uvd, etc.)

v2: add code comments

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
2014-02-18 16:11:41 +01:00
Alex Deucher
4233290519 drm/radeon: add vce dpm support for KV/KB
TODO: plug in cik_vce_suspend()/resume() so we can enable
vce powergating. See XXX in code.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:40 +01:00
Alex Deucher
ee35b0024a drm/radeon: enable vce dpm on CI
VCE dpm dynamically adjusts the uvd clocks on
demand.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:38 +01:00
Alex Deucher
8cd366823e drm/radeon: add vce dpm support for CI
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:36 +01:00
Alex Deucher
5ad6bf91ef drm/radeon: fill in set_vce_clocks for CIK asics
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:35 +01:00
Alex Deucher
58bd2a88fa drm/radeon/dpm: fetch vce states from the vbios
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:33 +01:00
Alex Deucher
b62d628bd6 drm/radeon/dpm: fill in some initial vce infrastructure
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:32 +01:00
Alex Deucher
82f79cc54b drm/radeon/dpm: move platform caps fetching to a separate function
It's needed by by both the asic specific functions and the
extended table parser.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:30 +01:00
Alex Deucher
b59b733397 drm/radeon: add callback for setting vce clocks
Similar to uvd clock setting.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-18 16:11:28 +01:00
Christian König
98ccc291ff drm/radeon: add VCE version parsing and checking
Also make the result available to userspace.

Signed-off-by: Christian König <christian.koenig@amd.com>
2014-02-18 16:11:26 +01:00
Christian König
f7ba8b04b2 drm/radeon: add VCE ring query
Signed-off-by: Christian König <christian.koenig@amd.com>
2014-02-18 16:11:24 +01:00
Christian König
d93f79376f drm/radeon: initial VCE support v4
Only VCE 2.0 support so far.

v2: squashing multiple patches into this one
v3: add IRQ support for CIK, major cleanups,
    basic code documentation
v4: remove HAINAN from chipset list

Signed-off-by: Christian König <christian.koenig@amd.com>
2014-02-18 16:11:22 +01:00
Christian König
1c61eae469 drm/radeon: fix CP semaphores on CIK
The CP semaphore queue on CIK has a bug that triggers if uncompleted
waits use the same address while a signal is still pending. Work around
this by using different addresses for each sync.

Signed-off-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
2014-02-18 11:30:59 +01:00
Alex Deucher
d02f8575f1 drm/radeon: add missing include in btc_dpm.c
Fixes a compile error with debugfs disabled.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-06 14:45:35 -05:00
Dave Jones
7c7e867cf0 drm/radeon/dpm: fix uninitialized read from stack in kv_dpm_late_enable
If we take the false branch of the if quoted in the diff below, we
end up doing a return ret, without ever having initialized it.

Picked up by coverity.

Signed-off-by: Dave Jones <davej@fedoraproject.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-06 12:22:49 -05:00
Alex Deucher
4dbffb8f21 drm/radeon: remove useless return
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-06 12:22:48 -05:00
Alex Deucher
299302294e drm/radeon/dpm: use stored max_vddc rather than looking it up
When we parse the power tables use the stored mac_vddc value
rather than lookig it up manually each time.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-06 12:22:47 -05:00
Alex Deucher
9f3f63f24c drm/radeon/dpm: use the driver state for dpm debugfs
For btc and newer, we may modify the power state depending
on the circumstances.  Use the modified state rather than
the base state.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-06 12:22:46 -05:00
Alex Deucher
858a41c853 drm/radeon: fix UVD IRQ support on 7xx
Otherwise decoding isn't really useable.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2014-02-06 12:22:45 -05:00
Christian König
b927e1c204 drm/radeon: fix UVD IRQ support on SI
Otherwise decoding isn't really useable.

bug:
https://bugs.freedesktop.org/show_bug.cgi?id=71448

Signed-off-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-02-06 12:22:44 -05:00
Dave Airlie
7c4c62a04a drm/radeon: allow geom rings to be setup on r600/r700 (v2)
the evergreen CS parser has allowed this for a while, just port
the code to the r600 one.

This is required before geom shaders can be made work.

v2: agd5f: minor cleanup and add additional 7xx reg.

Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
2014-02-06 12:13:52 +10:00
Dave Airlie
9ca5d4b4e1 Merge tag 'vmwgfx-fixes-3.14-2014-02-05' of git://people.freedesktop.org/~thomash/linux into drm-next
A couple of vmwgfx fixes together with missing bits of legacy device
emulation to facilitate old user-space drivers on new devices.

The shader emulation bits are a bit large, but since they mostly touch the
new device code, regressions are unlikely. I figure the gain of having
this from the start clearly outweighs the risc of adding these bits at
this point.

Pull request of 2014-02-05

* tag 'vmwgfx-fixes-3.14-2014-02-05' of git://people.freedesktop.org/~thomash/linux:
  vmwgfx: Fix unitialized stack read in vmw_setup_otable_base
  drm/vmwgfx: Reemit context bindings when necessary v2
  drm/vmwgfx: Detect old user-space drivers and set up legacy emulation v2
  drm/vmwgfx: Emulate legacy shaders on guest-backed devices v2
  drm/vmwgfx: Fix legacy surface reference size copyback
  drm/vmwgfx: Fix SET_SHADER_CONST emulation on guest-backed devices
  drm/vmwgfx: Fix regression caused by "drm/ttm: make ttm reservation calls behave like reservation calls"
  drm/vmwgfx: Don't commit staged bindings if execbuf fails
2014-02-06 12:04:31 +10:00
Dave Airlie
9df5a9b02f Merge tag 'ttm-fixes-3.14-2014-02-05' of git://people.freedesktop.org/~thomash/linux into drm-next
Two ttm regression fixes.

Pull request of 2014-02-05

* tag 'ttm-fixes-3.14-2014-02-05' of git://people.freedesktop.org/~thomash/linux:
  drm/ttm: Don't clear page metadata of imported sg pages
  drm/ttm: Fix TTM object open regression
2014-02-06 11:50:48 +10:00
Dave Airlie
8b7ad1bb3d drm/mgag200,ast,cirrus: fix regression with drm_can_sleep conversion
I totally sign inverted my way out of this one.

Cc: stable@vger.kernel.org
Reported-by: "Sabrina Dubroca" <sd@queasysnail.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
2014-02-06 11:39:03 +10:00
Thomas Hellstrom
1b76af5ce8 drm/ttm: Don't clear page metadata of imported sg pages
These page pointers shouldn't be visible to TTM in the first place, but
until we fix that up, don't clear the page metadata because that
will upset the exporter.

Reported-and-tested-by: Cristoph Haag <haagch.christoph@googleemail.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
2014-02-05 16:03:29 +01:00
Thomas Hellstrom
c66f854338 drm/ttm: Fix TTM object open regression
Commit drm/ttm: ttm object security fixes for render nodes introduced a
regression where, if a TTM object was opened multiple times from the same
open file, the caller would spin uninterruptibly in the kernel.

Fix this.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
2014-02-05 09:29:25 +01:00
Dave Jones
cd9a21a831 vmwgfx: Fix unitialized stack read in vmw_setup_otable_base
One of the error paths in vmw_setup_otable_base causes us to return with
'ret' having never been set to anything causing us to return whatever was
on the stack.

Found with Coverity

Signed-off-by: Dave Jones <davej@fedoraproject.org>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
2014-02-05 08:52:34 +01:00
Thomas Hellstrom
30f82d816d drm/vmwgfx: Reemit context bindings when necessary v2
When a context is first referenced in the command stream, make sure that all
scrubbed (as a result of eviction) bindings are re-emitted. Also make sure that
all bound resources are put on the resource validate list.

This is needed for legacy emulation, since legacy user-space drivers will
typically not re-emit shader bindings. It also removes the requirement for
user-space drivers to re-emit render-target- and texture bindings.

Makes suspend and hibernate now also work with legacy user-space drivers on
guest-backed devices.

v2: Don't rebind on legacy devices.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
2014-02-05 08:41:58 +01:00
Thomas Hellstrom
a6fc955ff9 drm/vmwgfx: Detect old user-space drivers and set up legacy emulation v2
GB aware mesa userspace drivers are detected by the fact that they are
calling the vmw getparam ioctl querying DRM_VMW_PARAM_HW_CAPS to detect
whether the device is Guest-backed object capable. For other drivers,
lie about hardware version and send the 3D capabilities in a format they
expect.

v2:
Use DRM_VMW_PARAM_MAX_MOB_MEMORY to detect gb awareness,
Make sure we don't ovwerwrite bounce buffer or write past user-space buffer
indicated size.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
2014-02-05 08:41:44 +01:00
Thomas Hellstrom
d5bde95663 drm/vmwgfx: Emulate legacy shaders on guest-backed devices v2
Command stream legacy shader creation and destruction is replaced by
NOPs in the command stream, and instead guest-backed shaders are created
and destroyed as part of the command validation process.

v2: Removed some stray debug messages.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
2014-02-05 08:39:19 +01:00
Thomas Hellstrom
c1a21373d2 drm/vmwgfx: Fix legacy surface reference size copyback
Surfaces created using the guest-backed surface interface only keeps the
base mip size, so only copy that if the legacy surface reference
ioctl requests the size information.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
2014-02-05 08:38:03 +01:00
Thomas Hellstrom
0ccbbae43c drm/vmwgfx: Fix SET_SHADER_CONST emulation on guest-backed devices
Emulate the SET_SHADER_CONST legacy command on guest-backed devices by
issuing a SET_GB_SHADERCONSTS_INLINE command.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
2014-02-05 08:37:05 +01:00
Thomas Hellstrom
cf5e341333 drm/vmwgfx: Fix regression caused by "drm/ttm: make ttm reservation calls behave like reservation calls"
The call to ttm_eu_backoff_reservation() as part of an error path would cause
a lock imbalance if the reservation ticket was not initialized. This error is
easily triggered from user-space by submitting a bogus command stream.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
Cc: stable@vger.kernel.org
Cc: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: Dave Airlie <airlied@redhat.com>
2014-02-05 08:36:11 +01:00
Thomas Hellstrom
76c7d18bcd drm/vmwgfx: Don't commit staged bindings if execbuf fails
If execbuf fails and binding commands are never sent to the device,
don't commit the staged context bindings to the tracker.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
2014-02-05 08:35:09 +01:00
Dave Airlie
ec22b4aa99 drm/mgag200: fix typo causing bw limits to be ignored on some chips
mode->mdev otherwise the bw limits never kick in.

Reported in RHEL testing.

Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
2014-02-05 14:31:32 +10:00
Linus Torvalds
38dbfb59d1 Linus 3.14-rc1 2014-02-02 16:42:13 -08:00
Linus Torvalds
69048e0188 Merge branch 'parisc-3.14' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux
Pull parisc updates from Helge Deller:
 "The three major changes in this patchset is a implementation for
  flexible userspace memory maps, cache-flushing fixes (again), and a
  long-discussed ABI change to make EWOULDBLOCK the same value as
  EAGAIN.

  parisc has been the only platform where we had EWOULDBLOCK != EAGAIN
  to keep HP-UX compatibility.  Since we will probably never implement
  full HP-UX support, we prefer to drop this compatibility to make it
  easier for us with Linux userspace programs which mostly never checked
  for both values.  We don't expect major fall-outs because of this
  change, and if we face some, we will simply rebuild the necessary
  applications in the debian archives"

* 'parisc-3.14' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: add flexible mmap memory layout support
  parisc: Make EWOULDBLOCK be equal to EAGAIN on parisc
  parisc: convert uapi/asm/stat.h to use native types only
  parisc: wire up sched_setattr and sched_getattr
  parisc: fix cache-flushing
  parisc/sti_console: prefer Linux fonts over built-in ROM fonts
2014-02-02 16:32:53 -08:00
Mikulas Patocka
1c0b8a7a62 hpfs: optimize quad buffer loading
HPFS needs to load 4 consecutive 512-byte sectors when accessing the
directory nodes or bitmaps.  We can't switch to 2048-byte block size
because files are allocated in the units of 512-byte sectors.

Previously, the driver would allocate a 2048-byte area using kmalloc,
copy the data from four buffers to this area and eventually copy them
back if they were modified.

In the current implementation of the buffer cache, buffers are allocated
in the pagecache.  That means that 4 consecutive 512-byte buffers are
stored in consecutive areas in the kernel address space.  So, we don't
need to allocate extra memory and copy the content of the buffers there.

This patch optimizes the code to avoid copying the buffers.  It checks
if the four buffers are stored in contiguous memory - if they are not,
it falls back to allocating a 2048-byte area and copying data there.

Signed-off-by: Mikulas Patocka <mikulas@artax.karlin.mff.cuni.cz>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-02-02 16:24:07 -08:00
Mikulas Patocka
2cbe5c76fc hpfs: remember free space
Previously, hpfs scanned all bitmaps each time the user asked for free
space using statfs.  This patch changes it so that hpfs scans the
bitmaps only once, remembes the free space and on next invocation of
statfs it returns the value instantly.

New versions of wine are hammering on the statfs syscall very heavily,
making some games unplayable when they're stored on hpfs, with load
times in minutes.

This should be backported to the stable kernels because it fixes
user-visible problem (excessive level load times in wine).

Signed-off-by: Mikulas Patocka <mikulas@artax.karlin.mff.cuni.cz>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-02-02 16:24:07 -08:00
Helge Deller
9dabf60dc4 parisc: add flexible mmap memory layout support
Add support for the flexible mmap memory layout (as described in
http://lwn.net/Articles/91829). This is especially very interesting on
parisc since we currently only support 32bit userspace (even with a
64bit Linux kernel).

Signed-off-by: Helge Deller <deller@gmx.de>
2014-02-02 21:00:13 +01:00
Guy Martin
f5a408d53e parisc: Make EWOULDBLOCK be equal to EAGAIN on parisc
On Linux, only parisc uses a different value for EWOULDBLOCK which
causes a lot of troubles for applications not checking for both values.
Since the hpux compat is long dead, make EWOULDBLOCK behave the same as
all other architectures.

Signed-off-by: Guy Martin <gmsoft@tuxicoman.be>
Signed-off-by: Helge Deller <deller@gmx.de>
2014-02-02 20:57:42 +01:00
Helge Deller
9391bc777b parisc: convert uapi/asm/stat.h to use native types only
The stat.h header file is exported to userspace. Some userspace
applications failed to compile due to missing/unknown types, so we
better convert it to use native types only (like it's done on other
architectures too).

Signed-off-by: Helge Deller <deller@gmx.de>
2014-02-02 20:57:33 +01:00
Helge Deller
998bbb2fc0 parisc: wire up sched_setattr and sched_getattr
Signed-off-by: Helge Deller <deller@gmx.de>
2014-02-02 20:57:25 +01:00
Helge Deller
57737c49dd parisc: fix cache-flushing
This commit:
f8dae00684: parisc: Ensure full cache coherency for kmap/kunmap
caused negative caching side-effects, e.g. hanging processes with expect and
too many inequivalent alias messages from flush_dcache_page() on Debian 5 systems.

This patch now partly reverts it and has been in production use on our debian buildd
makeservers since a week without any major problems.

Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: John David Anglin <dave.anglin@bell.net>
Cc: stable@vger.kernel.org # v3.9+
Signed-off-by: Helge Deller <deller@gmx.de>
2014-02-02 20:57:16 +01:00
Helge Deller
8a10bc9d27 parisc/sti_console: prefer Linux fonts over built-in ROM fonts
The built-in ROM fonts lack many necessary ASCII characters, which is
why it makes sens to prefer the Linux fonts instead if they are
available.  This makes consoles on STI graphics cards which are not
supported by the stifb driver (e.g. Visualize FXe) looks much nicer.

Signed-off-by: Helge Deller <deller@gmx.de>
Cc: stable@vger.kernel.org # v3.13
2014-02-02 20:56:47 +01:00
Linus Torvalds
602456bf16 Merge branch 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging
Pull hwmon kconfig fixes from Jean Delvare.

* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
  hwmon: Fix SENSORS_TMP102 dependencies to eliminate build errors
  hwmon: Fix SENSORS_LM75 dependencies to eliminate build errors
2014-02-02 11:30:57 -08:00
Linus Torvalds
7b383bef25 Merge branch 'slab/next' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/linux
Pull SLAB changes from Pekka Enberg:
 "Random bug fixes that have accumulated in my inbox over the past few
  months"

* 'slab/next' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/linux:
  mm: Fix warning on make htmldocs caused by slab.c
  mm: slub: work around unneeded lockdep warning
  mm: sl[uo]b: fix misleading comments
  slub: Fix possible format string bug.
  slub: use lockdep_assert_held
  slub: Fix calculation of cpu slabs
  slab.h: remove duplicate kmalloc declaration and fix kernel-doc warnings
2014-02-02 11:30:08 -08:00