Commit Graph

8299 Commits

Author SHA1 Message Date
be86c606b5 drm/amdgpu: cleanup amdgpu_sync_rings V2
No longer needed now that semaphores are gone.

V2: remove the first amdgpu_sync_wait in amdgpu_ib_schedule

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Ken Wang  <Qingqing.Wang@amd.com> (V1)
Reviewed-by: Monk Liu <monk.liu@amd.com> (V2)
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:50 -05:00
2f4b940033 drm/amdgpu: clean up hw semaphore support in driver
No longer used.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Ken Wang  <Qingqing.Wang@amd.com>
Reviewed-by: Monk Liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:49 -05:00
a8480309df drm/amdgpu: try to find BO VAs only for the BOs in the list
The other ones don't have any VAs assigned anyway or are uninteresting to us.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:49 -05:00
15486fd20c drm/amdgpu: search only the BO list for VM mappings
Make UVD/VCE VM emulation more efficient.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:48 -05:00
1ea863fd73 drm/amdgpu: keep the prefered/allowed domains in the BO
Stop copying that to the bo list entry, it doesn't change anyway.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:47 -05:00
636ce25c30 drm/amdgpu: cleanup bo list bucket handling
Move that into the BO list. No functional change.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:47 -05:00
2a7d9bdabe drm/amdgpu: cleanup amdgpu_cs_parser_relocs
Rename it to amdgpu_cs_parser_bos and move validation and bo list init there.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:46 -05:00
c3cca41e62 drm/amdgpu: cleanup amdgpu_cs_parser structure
Remove unused user_ptr field, group fields by usage.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:46 -05:00
25cfc3c27e drm/amdgpu: group VM mapping tree with its lock (v2)
And also update the comment.

v2: agd: rebase on usptream.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:45 -05:00
36409d122c drm/amdgpu: cleanup amdgpu_cs_list_validate
No need to actually check the current placement. Just use the allowed domains
when the threshold is reached.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:45 -05:00
f69f90a113 drm/amdgpu: fix amdgpu_cs_get_threshold_for_moves handling
The threshold should only be computed once.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:44 -05:00
d7006964d4 drm/amdgpu: fix issue with overlapping userptrs
Otherwise we could try to evict overlapping userptr BOs in get_user_pages(),
leading to a possible circular locking dependency.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-10 14:16:43 -05:00
cc1de6e800 drm/amdgpu: fix issue with overlapping userptrs
Otherwise we could try to evict overlapping userptr BOs in get_user_pages(),
leading to a possible circular locking dependency.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-02-10 14:07:52 -05:00
b19763d0d8 drm/amdgpu: remove unnecessary forward declaration
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
2016-02-10 14:07:38 -05:00
a8d81b3626 drm/amdgpu: hold reference to fences in amdgpu_sa_bo_new (v2)
An arbitrary amount of time can pass between spin_unlock and
fence_wait_any_timeout, so we need to ensure that nobody frees the
fences from under us.

A stress test (rapidly starting and killing hundreds of glxgears
instances) ran into a deadlock in fence_wait_any_timeout after
about an hour, and this race condition appears to be a plausible
cause.

v2: agd: rebase on upstream

Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
2016-02-10 14:07:31 -05:00
ca19852884 drm/amdgpu: fix s4 resume
No need to re-init asic if it's already been initialized.
Skip IB tests since kernel processes are frozen in thaw.

Signed-off-by: Flora Cui <Flora.Cui@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-02-10 14:07:13 -05:00
156d7d4120 vga_switcheroo: Add handler flags infrastructure
Allow handlers to declare their capabilities and allow clients to
obtain that information. So far we have these use cases:

* If the handler is able to switch DDC separately, clients need to
  probe EDID with drm_get_edid_switcheroo(). We should allow them
  to detect a capable handler to ensure this function only gets
  called when needed.

* Likewise if the handler is unable to switch AUX separately, the active
  client needs to communicate link training parameters to the inactive
  client, which may then skip the AUX handshake and set up its output
  with these pre-calibrated values (DisplayPort specification v1.1a,
  section 2.5.3.3). Clients need a way to recognize such a situation.

The flags for the radeon_atpx_handler and amdgpu_atpx_handler are
initially set to 0, this can later on be amended with
  handler_flags |= VGA_SWITCHEROO_CAN_SWITCH_DDC;
when a ->switch_ddc callback is added.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=88861
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=61115
Tested-by: Lukas Wunner <lukas@wunner.de>
    [MBP  9,1 2012  intel IVB + nvidia GK107  pre-retina  15"]
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/2b0d93ed6e511ca09e95e45e0b35627f330fabce.1452525860.git.lukas@wunner.de
2016-02-09 11:21:07 +01:00
e3b04bc790 drma/dmgpu: move cg and pg flags into shared headers
So they can be used by powerplay.

Reviewed-by: Eric Huang <JinHuiEric.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-08 10:37:47 -05:00
b118af7012 drm/amdgpu: remove unused cg defines
Leftover from radeon.

Reviewed-by: Eric Huang <JinHuiEric.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-08 10:37:47 -05:00
08d3340876 drm/amdgpu: add a cgs interface to fetch cg and pg flags
Needed to pass the cg and pg info to powerplay.

Reviewed-by: Eric Huang <JinHuiEric.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-08 10:37:46 -05:00
35e5912d08 drm/amdgpu: be consistent with uvd cg flags
Don't do anything if the uvd cg flags are not set.

Reviewed-by: Eric Huang <JinHuiEric.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-08 10:37:44 -05:00
0fd4af9e32 drm/amdgpu: clean up vce pg flags for cz/st
It was already disabled elsewhere, make it offical.

Reviewed-by: Eric Huang <JinHuiEric.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-08 10:37:43 -05:00
808a934fd4 drm/amdgpu: handle vce pg flags properly
Don't attempt to start/stop the vce block if pg is disabled.

Reviewed-by: Eric Huang <JinHuiEric.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-08 10:37:43 -05:00
b6df77fc5c drm/amdgpu: handle uvd pg flags properly
Don't attempt to start/stop the uvd block if pg is disabled.

Reviewed-by: Eric Huang <JinHuiEric.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-08 10:37:42 -05:00
50171ebecf drm/amdgpu/dpm/ci: switch over to the common pcie caps interface
We already query this at driver init, so use that info.  Also
handles virtualization cases.

Reviewed-by: monk liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-08 10:37:42 -05:00
76ecb2c75b drm/amdgpu/cik: don't mess with aspm if gpu is root bus
Pcie registers may not be available in a virtualized
environment.

Reviewed-by: monk liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-08 10:37:41 -05:00
cd474ba0d6 drm/amdgpu: add pcie cap module parameters (v2)
Allows the user to force the supported pcie gen and lane
config on both the asic and the chipset.
Useful for debugging pcie problems and for virtualization
where we may not be able to query the pcie bridge caps.

Default to:
gen: chipset 1/2, asic 1/2/3
lanes: 1/2/4/8/16

v2: fix bare metal case

Reviewed-by: monk liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-08 10:37:41 -05:00
c745884b30 Merge branch 'drm-fixes-4.5' of git://people.freedesktop.org/~agd5f/linux into drm-fixes
- fix and enable iceland/topaz support
- handle WC on platforms that don't support it

* 'drm-fixes-4.5' of git://people.freedesktop.org/~agd5f/linux:
  drm/amdgpu: disable uvd and vce clockgating on Fiji
  drm/amdgpu: remove exp hardware support from iceland
  drm/amdgpu: load MEC ucode manually on iceland
  drm/amdgpu: don't load MEC2 on topaz
  drm/amdgpu: drop topaz support from gmc8 module
  drm/amdgpu: pull topaz gmc bits into gmc_v7
  drm/amdgpu: The VI specific EXE bit should only apply to GMC v8.0 above
  drm/amdgpu: iceland use CI based MC IP
  drm/amdgpu: move gmc7 support out of CIK dependency
  drm/amdgpu/gfx7: enable cp inst/reg error interrupts
  drm/amdgpu/gfx8: enable cp inst/reg error interrupts
  drm/amdgpu: mask out WC from BO on unsupported arches
  drm/radeon: mask out WC from BO on unsupported arches
  drm: add helper to check for wc memory support
  drm/amdgpu: no need to load MC firmware on fiji
2016-02-05 14:48:36 +10:00
ec71f1284a Merge tag 'drm-amdkfd-fixes-2016-01-28' of git://people.freedesktop.org/~gabbayo/linux into drm-fixes
two static checker fixes.

* tag 'drm-amdkfd-fixes-2016-01-28' of git://people.freedesktop.org/~gabbayo/linux:
  drm/amdkfd: Remove unnecessary cast in kfree
  drm/amdgpu: fix non-ANSI declaration of amdgpu_amdkfd_gfx_*_get_functions()
2016-02-05 14:47:24 +10:00
6357b75a5c drm/amdgpu: disable uvd and vce clockgating on Fiji
Doesn't work properly yet.

Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Sonny Jiang <sonny.jiang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-03 13:49:51 -05:00
dba280b20b drm/amdgpu: remove exp hardware support from iceland
It's working now.

bug:
https://bugs.freedesktop.org/show_bug.cgi?id=92270

Reviewed-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-02-02 22:54:39 -05:00
951e09624a drm/amdgpu: load MEC ucode manually on iceland
The smc doesn't handle it.

Reviewed-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-02-02 22:54:32 -05:00
97dde76a30 drm/amdgpu: don't load MEC2 on topaz
Not validated.

Reviewed-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-02-02 22:54:25 -05:00
8878d8548a drm/amdgpu: drop topaz support from gmc8 module
topaz is actually gmc7.

Reviewed-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-02-02 22:54:18 -05:00
72b459c8f7 drm/amdgpu: pull topaz gmc bits into gmc_v7
Add the topaz golden settings into the gmc7 module.

Reviewed-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-02-02 22:54:12 -05:00
8f3c162961 drm/amdgpu: The VI specific EXE bit should only apply to GMC v8.0 above
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Cc: stable@vger.kernel.org
2016-02-02 22:54:11 -05:00
429c45deae drm/amdgpu: iceland use CI based MC IP
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Cc: stable@vger.kernel.org
2016-02-02 22:54:10 -05:00
e42d852616 drm/amdgpu: move gmc7 support out of CIK dependency
It's used by iceland which is VI.

Reviewed-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-02-02 22:52:48 -05:00
ef720532ec drm/amdgpu/gfx7: enable cp inst/reg error interrupts
Enable CP register/instruction error interrupts. Useful
for debugging command stream problems.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-02 16:26:10 -05:00
1d22a454ec drm/amdgpu/gfx8: enable cp inst/reg error interrupts
Enable CP register/instruction error interrupts. Useful
for debugging command stream problems.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-02-02 16:25:40 -05:00
a187f17f0e drm/amdgpu: mask out WC from BO on unsupported arches
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-02-02 10:09:11 -05:00
ad32152eb2 drm/amdgpu: no need to load MC firmware on fiji
Vbios does this for us on asic_init.

Reviewed-by: Ken Wang >Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-01-29 11:10:40 -05:00
28b8d66e0c drm/amdgpu: only move pt bos in LRU list on success
This fixes a race condition in the error case: since the pt bos have not
necessarily been reserved in case of an error, we could move a pt bo that
is currently in the middle of being evicted/moved by another process,
which then resulted in a BUG_ON in ttm_bo_add_to_lru.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-01-28 11:46:45 -05:00
f785d98711 drm/amdgpu: fix non-ANSI declaration of amdgpu_amdkfd_gfx_*_get_functions()
amdgpu_amdkfd_gfx_7_get_functions and amdgpu_amdkfd_gfx_8_0_get_functions
have no parameters, so use the normal void parameter convention to make
them  match their prototypes in the header file
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
2016-01-28 14:39:34 +02:00
f49d45c973 drm/amdgpu: don't init fbdev if we don't have any connectors
Don't init fbdev if we don't have connectors.  E.g., if you have
a PX laptop with the displays attached to an IGP with no driver
support, you may end up with a blank screen rather than falling
back to vesa, etc.

Based on a similar radeon patch from Rob Clark.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-01-26 00:35:56 -05:00
41f2d99056 drm/amdgpu: fix next_rptr handling for debugfs
That somehow got lost.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
2016-01-22 14:44:08 -05:00
78d0e182b6 drm/amdgpu: fix amdgpu_bo_pin_restricted VRAM placing v2
We could pin BOs into invisible VRAM otherwise.

v2: make logic more readable as suggested by Michel

Cc: stable@vger.kernel.org
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1)
Reviewed-by: Rex Zhu <Rex.Zhu@amd.com> (v1)
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2016-01-22 10:47:48 -05:00
4b5ece24ce drm/amd/amdgpu: Improve amdgpu_dpm* macros to avoid unexpected result (v2)
The two macros returns are values which probably are used
in the expression of calculation. Without the brackets
the result of the expression may be wrong.

v2: agd: squash both patches together

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Eric Huang <JinHuiEric.Huang@amd.com>
2016-01-22 10:47:48 -05:00
3466904d38 drm/amdgpu: Allow the driver to load if amdgpu.powerplay=1 on asics without powerplay support
Avoid setting pp_enabled if there is no powerplay implementation.

Signed-off-by: Jordan Lazare <Jordan.Lazare@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2016-01-19 12:42:30 -05:00
9571e1d840 drm/amdgpu: Use drm_calloc_large for VM page_tables array
It can be big, depending on the VM address space size, which is tunable
via the vm_size module parameter.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93721
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2016-01-19 12:25:09 -05:00