Skip to content

L
linux
Switch branch/tag
  1. 15 Nov, 2016 1 commit
  3. 11 Nov, 2016 4 commits
    • Mario Kleiner's avatar
      drm/amdgpu: Attach exclusive fence to prime exported bo's. (v5) · 8e94a46c
      Mario Kleiner authored 
      External clients which import our bo's wait only
for exclusive dmabuf-fences, not on shared ones,
ditto for bo's which we import from external
providers and write to.

Therefore attach exclusive fences on prime shared buffers
if our exported buffer gets imported by an external
client, or if we import a buffer from an external
exporter.

See discussion in thread:
https://lists.freedesktop.org/archives/dri-devel/2016-October/122370.html

Prime export tested on Intel iGPU + AMD Tonga dGPU as
DRI3/Present Prime render offload, and with the Tonga
standalone as primary gpu.

v2: Add a wait for all shared fences before prime export,
    as suggested by Christian Koenig.

v3: - Mark buffer prime_exported in amdgpu_gem_prime_pin,
    so we only use the exclusive fence when exporting a
    bo to external clients like a separate iGPU, but not
    when exporting/importing from/to ourselves as part of
    regular DRI3 fd passing.

    - Propagate failure of reservation_object_wait_rcu back
    to caller.

v4: - Switch to a prime_shared_count counter instead of a
      flag, which gets in/decremented on prime_pin/unpin, so
      we can switch back to shared fences if all clients
      detach from our exported bo.

    - Also switch to exclusive fence for prime imported bo's.

v5: - Drop lret, instead use int ret -> long ret, as proposed
      by Christian.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=95472
Tested-by: Mike Lothian <mike@fireburn.co.uk> (v1)
Signed-off-by: default avatarMario Kleiner <mario.kleiner.de@gmail.com>
Reviewed-by: Christian König <christian.koenig@amd.com>.
Cc: Christian König <christian.koenig@amd.com>
Cc: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: default avatarAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
      8e94a46c
    • Dave Airlie's avatar
      Merge branch 'topic-arcpgu-fixes' of... · e2384535
      Dave Airlie authored 
      Merge branch 'topic-arcpgu-fixes' of https://github.com/foss-for-synopsys-dwc-arc-processors/linux into drm-fixes

* 'topic-arcpgu-fixes' of https://github.com/foss-for-synopsys-dwc-arc-processors/linux:
  drm/arcpgu: Accommodate adv7511 switch to DRM bridge
      e2384535
    • Eugeniy Paltsev's avatar
      drm/arcpgu: Accommodate adv7511 switch to DRM bridge · 7bc61cc5
      Eugeniy Paltsev authored 
      ARC PGU driver starts crashing on initialization after
'commit e12c2f64 ("drm/i2c: adv7511: Convert to drm_bridge")'
This happenes because in "arcpgu_drm_hdmi_init" function we get pointer
of "drm_i2c_encoder_driver" structure, which doesn't exist after
adv7511 hdmi encoder interface changed from slave encoder to drm bridge.
So, when we call "encoder_init" function from this structure driver
crashes.

Bootlog:
------------------------------------->8--------------------------------
[drm] Initialized drm 1.1.0 20060810
arcpgu e0017000.pgu: arc_pgu ID: 0xabbabaab
arcpgu e0017000.pgu: assigned reserved memory node frame_buffer@9e000000
Path: (null)
CPU: 0 PID: 1 Comm: swapper Not tainted 4.8.0-00001-gb5642252fa01-dirty #8
task: 9a058000 task.stack: 9a032000

[ECR   ]: 0x00220100 => Invalid Read @ 0x00000004 by insn @ 0x803934e8
[EFA   ]: 0x00000004
[BLINK ]: drm_atomic_helper_connector_dpms+0xa6/0x230
[ERET  ]: drm_atomic_helper_connector_dpms+0xa4/0x230
[STAT32]: 0x00000846 : K DE       E2 E1
BTA: 0x8016d949  SP: 0x9a033e34  FP: 0x00000000
LPS: 0x8036f6fc LPE: 0x8036f700 LPC: 0x00000000
r00: 0x8063c118 r01: 0x805b98ac r02: 0x00000b11
r03: 0x00000000 r04: 0x9a010f54 r05: 0x00000000
r06: 0x00000001 r07: 0x00000000 r08: 0x00000028
r09: 0x00000001 r10: 0x00000007 r11: 0x00000054
r12: 0x720a3033

Stack Trace:
  drm_atomic_helper_connector_dpms+0xa4/0x230
  arcpgu_drm_hdmi_init+0xbc/0x228
  arcpgu_probe+0x168/0x244
  platform_drv_probe+0x26/0x64
  really_probe+0x1f0/0x32c
  __driver_attach+0xa8/0xd0
  bus_for_each_dev+0x3c/0x74
  bus_add_driver+0xc2/0x184
  driver_register+0x50/0xec
  do_one_initcall+0x3a/0x120
  kernel_init_freeable+0x108/0x1a0
------------------------------------->8--------------------------------

Fix ARC PGU driver to be able work with drm bridge hdmi encoder
interface. The hdmi connector code isn't needed anymore as we expect
the adv7511 bridge driver to create/manage the connector.
Signed-off-by: default avatarEugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: default avatarArchit Taneja <architt@codeaurora.org>
Signed-off-by: default avatarAlexey Brodkin <abrodkin@synopsys.com>
      7bc61cc5
    • Dave Airlie's avatar
      Merge branch 'fixes-for-v4.9-rc5' of http://git.agner.ch/git/linux-drm-fsl-dcu into drm-fixes · 76af753b
      Dave Airlie authored 
      Yet another small batch of fixes. Two of the patches I had prepared
since quite some time, but they did not seem to affect operation in
a visible manner so far. Until recently, when I discovered the third
issue (disable planes before disabling CRTC), which made the two
previous fixes necessary.

* 'fixes-for-v4.9-rc5' of http://git.agner.ch/git/linux-drm-fsl-dcu:
  drm/fsl-dcu: disable planes before disabling CRTC
  drm/fsl-dcu: update all registers on flush
  drm/fsl-dcu: do not update when modifying irq registers
      76af753b
  5. 10 Nov, 2016 7 commits
  7. 09 Nov, 2016 6 commits
  9. 08 Nov, 2016 2 commits
    • Andrew Shadura's avatar
      drm/amd/powerplay: return false instead of -EINVAL · f20024d8
      Andrew Shadura authored 
      Returning -EINVAL from a bool-returning function
phm_check_smc_update_required_for_display_configuration has an unexpected
effect of returning true, which is probably not what was intended.
Replace -EINVAL by false.

The only place this function is called from is
psm_adjust_power_state_dynamic in
drivers/gpu/drm/amd/powerplay/eventmgr/psm.c:106:

	if (!equal || phm_check_smc_update_required_for_display_configuration(hwmgr)) {
		phm_apply_state_adjust_rules(hwmgr, requested, pcurrent);
		phm_set_power_state(hwmgr, &pcurrent->hardware, &requested->hardware);
		hwmgr->current_ps = requested;
	}

It seems to expect a boolean value here.

This issue has been found using the following Coccinelle semantic patch
written by Peter Senna Tschudin:
<smpl>
@@
identifier f;
constant C;
typedef bool;
@@
bool f (...){
<+...
* return -C;
...+>
}
</smpl>
Reviewed-by: default avatarAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: default avatarAndrew Shadura <andrew.shadura@collabora.co.uk>
Signed-off-by: default avatarAlex Deucher <alexander.deucher@amd.com>
      f20024d8
    • Arnd Bergmann's avatar
      drm/amdgpu/powerplay/smu7: fix unintialized data usage · a29d1260
      Arnd Bergmann authored 
      A recent bugfix replaced an out-of-bounds access with direct
use of unintialized data:

drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c: In function 'smu7_patch_limits_vddc':
drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:2033:6: error: 'vddc' may be used uninitialized in this function [-Werror=maybe-uninitialized]
drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:2146:11: note: 'vddc' was declared here
drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:2033:6: error: 'vddci' may be used uninitialized in this function [-Werror=maybe-uninitialized]
drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:2146:17: note: 'vddci' was declared here
  uint32_t vddc, vddci;

This initializes the data as before using the correct type.

Fixes: 77f7f71f ("drm/amdgpu/powerplay/smu7: fix static checker warning")
Signed-off-by: default avatarArnd Bergmann <arnd@arndb.de>
Signed-off-by: default avatarAlex Deucher <alexander.deucher@amd.com>
      a29d1260
  11. 07 Nov, 2016 7 commits
  13. 06 Nov, 2016 2 commits
  15. 05 Nov, 2016 11 commits
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7