RE: [PATCH] dma-buf: handle NULL src_list in reservation_object_copy_fences

> -Original Message- > From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf > Of Christian König > Sent: Thursday, August 10, 2017 9:42 AM > To: amd-gfx@lists.freedesktop.org; dri-de...@lists.freedesktop.org > Subject: [PATCH] dma-buf: handle NULL src_list in >

Re: dix/dispatch.c:4049: AttachOffloadGPU: Assertion `new->current_master == pScreen' failed.

Dear Linux folks, On 08/10/17 15:14, Paul Menzel wrote: Getting Dell Precision 3620 systems with an Intel Skylake system i7-6700K, and an external AMD graphics card, the X.Org server does not start and hits an assert. The software stack is Linux 4.9.38 and 4.12.5, X.Org server 1.19.3 and

Re: [PATCH] drm/amdgpu: potential shift wrapping bug

Am 10.08.2017 um 14:53 schrieb Dan Carpenter: On Thu, Aug 10, 2017 at 02:30:15PM +0200, Christian König wrote: Am 10.08.2017 um 14:16 schrieb Dan Carpenter: "frag_align" is a u64, so presumably we want to use the high bits as well instead of shift wrapping. Fixes: 6be7adb37d9b ("drm/amdgpu:

RE: [PATCH] dma-buf: handle NULL src_list in reservation_object_copy_fences

> -Original Message- > From: Christian König [mailto:deathsim...@vodafone.de] > Sent: Thursday, August 10, 2017 10:44 AM > To: Deucher, Alexander; amd-gfx@lists.freedesktop.org; dri- > de...@lists.freedesktop.org > Subject: Re: [PATCH] dma-buf: handle NULL src_list in >

[PATCH 4/8] drm/amdgpu: drop the extra VM huge page flag

From: Christian König Just add the flags to the addr field as well. Signed-off-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 15 ++- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 1 - 2 files changed, 6

[PATCH 6/8] drm/amdgpu: rename VM invalidated to moved

From: Christian König That better describes what happens here with the BO. Signed-off-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 16

[PATCH 2/8] drm/amdgpu: cleanup static CSA handling

From: Christian König Move the CSA bo_va from the VM to the fpriv structure. Signed-off-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 3 ++-

[PATCH 3/8] drm/amdgpu: remove superflous amdgpu_bo_kmap in the VM

From: Christian König We now properly kmap all BOs after validation. Signed-off-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 28 +--- 1 file changed, 9 insertions(+), 19 deletions(-) diff --git

[PATCH 1/8] drm/amdgpu: SHADOW and VRAM_CONTIGUOUS flags shouldn't be used by userspace

From: Christian König The shadow handling isn't implemented completely for userspace BOs and the kernel sets the VRAM_CONTIGUOUS as necessary. Signed-off-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 4 +--- 1 file

[PATCH] dma-buf: handle NULL src_list in reservation_object_copy_fences

From: Christian König The list of shared fences can be NULL and that needs to be handled as well. Signed-off-by: Christian König --- drivers/dma-buf/reservation.c | 28 +++- 1 file changed, 15 insertions(+), 13

RE: [PATCH 8/8] drm/amdgpu: rework page directory filling

> -Original Message- > From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf > Of Christian König > Sent: Thursday, August 10, 2017 10:51 AM > To: amd-gfx@lists.freedesktop.org > Subject: [PATCH 8/8] drm/amdgpu: rework page directory filling > > From: Christian König

Re: [PATCH] dma-buf: handle NULL src_list in reservation_object_copy_fences

Am 10.08.2017 um 16:40 schrieb Deucher, Alexander: -Original Message- From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf Of Christian König Sent: Thursday, August 10, 2017 9:42 AM To: amd-gfx@lists.freedesktop.org; dri-de...@lists.freedesktop.org Subject: [PATCH]

RE: [PATCH 4/8] drm/amdgpu: drop the extra VM huge page flag

> -Original Message- > From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf > Of Christian König > Sent: Thursday, August 10, 2017 10:51 AM > To: amd-gfx@lists.freedesktop.org > Subject: [PATCH 4/8] drm/amdgpu: drop the extra VM huge page flag > > From: Christian König

[PATCH] drm/amdgpu: fix missing endian-safe guard

Change-Id: I84a838d2d14318c1514f5146ae81ab58560b767b Signed-off-by: Evan Quan --- drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c

Re: [PATCH] drm/amdgpu: Uninitialized variable in amdgpu_ttm_backend_bind()

Am 10.08.2017 um 07:34 schrieb Alex Deucher: On Wed, Aug 9, 2017 at 10:16 PM, Michel Dänzer wrote: On 09/08/17 07:30 PM, Dan Carpenter wrote: My static checker complains that it's possible for "r" to be uninitialized. It used to be set to zero so this returns it to the

[PATCH] drm/amdgpu: jt_size was wrongly counted twice

Change-Id: Ic6c031f160cb5ed5744ed302f0d819505ece427e Signed-off-by: Evan Quan Signed-off-by: Huang Rui --- drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c

Re: [Intel-gfx] [PATCH libdrm] drm: Remove create_handle() drm_framebuffer "virtual".

On Wed, Aug 09, 2017 at 04:13:05PM -0700, Joe Kniss wrote: > On Wed, Aug 9, 2017 at 12:14 PM, Noralf Trønnes wrote: > > > > Den 09.08.2017 01.42, skrev Joe Kniss: > >> > >> Because all drivers currently use gem objects for framebuffer planes, > >> the virtual create_handle()

Re: [PATCH] drm/amdgpu: potential shift wrapping bug

On Thu, Aug 10, 2017 at 03:02:53PM +0200, Christian König wrote: > Am 10.08.2017 um 14:53 schrieb Dan Carpenter: > > On Thu, Aug 10, 2017 at 02:30:15PM +0200, Christian König wrote: > > > Am 10.08.2017 um 14:16 schrieb Dan Carpenter: > > > > "frag_align" is a u64, so presumably we want to use the

Re: [PATCH libdrm] drm/amdgpu: add new low overhead command submission API. (v2)

Hi Dave, On 18 July 2017 at 04:52, Dave Airlie wrote: > +int amdgpu_cs_submit_raw(amdgpu_device_handle dev, > +amdgpu_context_handle context, > +amdgpu_bo_list_handle bo_list_handle, > +int num_chunks, >

Re: [PATCH] drm/amdgpu: potential shift wrapping bug

Am 10.08.2017 15:02, schrieb Christian König: > Am 10.08.2017 um 14:53 schrieb Dan Carpenter: >> On Thu, Aug 10, 2017 at 02:30:15PM +0200, Christian König wrote: >>> Am 10.08.2017 um 14:16 schrieb Dan Carpenter: "frag_align" is a u64, so presumably we want to use the high bits as well

dix/dispatch.c:4049: AttachOffloadGPU: Assertion `new->current_master == pScreen' failed.

Dear Linux folks, Getting Dell Precision 3620 systems with an Intel Skylake system i7-6700K, and an external AMD graphics card, the X.Org server does not start and hits an assert. The software stack is Linux 4.9.38 and 4.12.5, X.Org server 1.19.3 and xf86-video-amdgpu 1.3.0. ``` $ lspci

[PATCH v2 20/29] drm/radeon: switch to drm_*_get(), drm_*_put() helpers

Use drm_*_get() and drm_*_put() helpers instead of drm_*_reference() and drm_*_unreference() helpers. drm_*_reference() and drm_*_unreference() functions are just compatibility alias for drm_*_get() and drm_*_put() and should not be used by new code. So convert all users of compatibility

[PATCH v2 01/29] drm/amdgpu: switch to drm_*_get(), drm_*_put() helpers

Use drm_*_get() and drm_*_put() helpers instead of drm_*_reference() and drm_*_unreference() helpers. drm_*_reference() and drm_*_unreference() functions are just compatibility alias for drm_*_get() and drm_*_put() and should not be used by new code. So convert all users of compatibility

[PATCH v2 00/29] DRM API Conversions

Changes since v1: - This time patches were generated with coccinelle instead of my own script, as suggested by Daniel Vetter. - Fixed the typo in commit messages. s/adn/and Note: I've included r-b, a-b tags, as these patches are identical to v1 except for the file:

Re: [PATCH] drm/amd/amdgpu: fix bug fail to remove debugfs when rmmod

On Thu, Aug 10, 2017 at 5:12 AM, Wang Hongcheng wrote: > Some debug files are forgotten to remove at fini. Remove them > all in pci_remove. > > BUG: SWDEV-129297 What files are failing to get removed? I'd prefer to remove them in the relevant places in the code (to mirror

[PATCH 1/2] dma-buf: add reservation_object_copy_fences (v2)

From: Christian König Allows us to copy all the fences in a reservation object to another one. v2: handle NULL src_list Signed-off-by: Christian König Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher

[PATCH 0/3] dma-buf changes for ttm and amdgpu

We have some changes in ttm and amdgpu that depend on these patches. Sumit, can you pull these in via dma-buf or should I roll them up through my amdgpu tree? Christian König (3): dma-buf: dma_fence_put is NULL safe dma-buf: add reservation_object_copy_fences dma-buf: fix

[PATCH 2/2] dma-buf: fix reservation_object_wait_timeout_rcu to wait correctly v2

From: Christian König With hardware resets in mind it is possible that all shared fences are signaled, but the exlusive isn't. Fix waiting for everything in this situation. v2: make sure we always wait for the exclusive fence Signed-off-by: Christian König

Re: [Intel-gfx] [PATCH libdrm] drm: Remove create_handle() drm_framebuffer "virtual".

On 10 August 2017 at 20:29, Joe Kniss wrote: > On Wed, Aug 9, 2017 at 4:13 PM, Joe Kniss wrote: >> On Wed, Aug 9, 2017 at 12:14 PM, Noralf Trønnes wrote: >>> >>> Den 09.08.2017 01.42, skrev Joe Kniss: Because all drivers currently

[PATCH 2/2] drm/amd/display/dc: make dce120_link_encoder_create static

Only used in dce120_resource.c. Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c

[PATCH 1/2] drm/amd/display/dc: Make dce110_validate_bandwidth static (v2)

Not used outside of the dce110_resource.c. v2: make a bunch of other functions static as well. Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c | 8 1 file changed, 4 insertions(+), 4 deletions(-) diff --git

Re: [PATCH 8/8] drm/amdgpu: rework page directory filling

On 2017-08-10 10:50 AM, Christian König wrote: > @@ -1185,9 +1185,31 @@ int amdgpu_vm_update_directories(struct amdgpu_device > *adev, > { > int r; > > - r = amdgpu_vm_update_level(adev, vm, >root, 0); > - if (r) > - amdgpu_vm_invalidate_level(>root); > +

Re: [PATCH 2/2] dma-buf: fix reservation_object_wait_timeout_rcu to wait correctly v2

Quoting Christian König (2017-08-10 19:19:52) > Am 10.08.2017 um 19:11 schrieb Chris Wilson: > > Quoting Alex Deucher (2017-08-10 18:01:49) > >> From: Christian König > >> > >> With hardware resets in mind it is possible that all shared fences are > >> signaled, but the

RE: [PATCH] drm/amdgpu: fix missing endian-safe guard

> -Original Message- > From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf > Of Evan Quan > Sent: Thursday, August 10, 2017 3:20 AM > To: amd-gfx@lists.freedesktop.org > Cc: Quan, Evan > Subject: [PATCH] drm/amdgpu: fix missing endian-safe guard > > Change-Id:

[PATCH 5/8] drm/amdgpu: separate bo_va structure

From: Christian König Split that into vm_bo_base and bo_va to allow other uses as well. Signed-off-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 4 +- drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c| 4 +-

[PATCH 7/8] drm/amdgpu: track evicted page tables

From: Christian König Instead of validating all page tables when one was evicted, track which one needs a validation. Signed-off-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 7 +-

[PATCH 8/8] drm/amdgpu: rework page directory filling

From: Christian König Keep track off relocated PDs/PTs instead of walking and checking all PDs. Signed-off-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 85 ++

RE: [PATCH] drm/amdgpu: jt_size was wrongly counted twice

> -Original Message- > From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf > Of Evan Quan > Sent: Thursday, August 10, 2017 4:23 AM > To: amd-gfx@lists.freedesktop.org > Cc: Quan, Evan; Huang, Ray > Subject: [PATCH] drm/amdgpu: jt_size was wrongly counted twice > >

Re: [PATCH 2/2] dma-buf: fix reservation_object_wait_timeout_rcu to wait correctly v2

Am 10.08.2017 um 19:11 schrieb Chris Wilson: Quoting Alex Deucher (2017-08-10 18:01:49) From: Christian König With hardware resets in mind it is possible that all shared fences are signaled, but the exlusive isn't. Fix waiting for everything in this situation. I'm

[PATCH 2/6] drm/amdgpu: add DCE8 APUs to dc_supported check

Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index f455d9d..8a1092e 100644 ---

[PATCH 3/6] drm/amd/display/dc: add DIGG for KV

KV has 7 possible DIG blocks. Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c | 6 ++ drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c | 3 ++- 2 files changed, 8 insertions(+), 1 deletion(-) diff --git

[PATCH 1/6] drm/amd/display/dm: add KV, KB, ML (v2)

Add DCE8 APUs to display manager. v2: rebase changes Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 16 1 file changed, 16 insertions(+) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c

[PATCH 5/6] drm/amd/disply/dc: add resource support for DCE8 APUs (v2)

Add the appropriate resources for APUs: KV:4 pipes, 7 dig, 3 PPLLs KB/ML: 2 pipes, 6 dig, 2 PPLLs v2: rebase changes Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 8 +- .../gpu/drm/amd/display/dc/dce80/dce80_resource.c |

[PATCH 4/6] drm/amd/display/dc: add DCE_VERSION for DCE8 APUs

DCE 8.1 = Kaveri DCE 8.3 = Kabini/Mullins Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/bios/command_table_helper.c | 2 ++ drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c | 2 ++ drivers/gpu/drm/amd/display/dc/core/dc_resource.c

[PATCH 6/6] drm/amdgpu/cik: add IP modules for DC for APUs

Enable DC for DCE8 APUs. Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/cik.c | 8 1 file changed, 8 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c index af3b66f..793b1470 100644 ---

Re: [PATCH 2/2] dma-buf: fix reservation_object_wait_timeout_rcu to wait correctly v2

Am 10.08.2017 um 20:50 schrieb Chris Wilson: Quoting Christian König (2017-08-10 19:19:52) Am 10.08.2017 um 19:11 schrieb Chris Wilson: Quoting Alex Deucher (2017-08-10 18:01:49) From: Christian König With hardware resets in mind it is possible that all shared

Re: [PATCH 7/8] drm/amdgpu: track evicted page tables

On 2017-08-10 10:50 AM, Christian König wrote: > @@ -2384,11 +2360,24 @@ void amdgpu_vm_bo_rmv(struct amdgpu_device *adev, > * Mark @bo as invalid. > */ > void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev, > - struct amdgpu_bo *bo) > +

[PATCH] drm/amd/amdgpu: fix bug fail to remove debugfs when rmmod

Some debug files are forgotten to remove at fini. Remove them all in pci_remove. BUG: SWDEV-129297 Signed-off-by: Wang Hongcheng --- drivers/gpu/drm/amd/amdgpu/amdgpu.h| 2 ++ drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 19 +++

[PATCH] drm/amdgpu: potential shift wrapping bug

"frag_align" is a u64, so presumably we want to use the high bits as well instead of shift wrapping. Fixes: 6be7adb37d9b ("drm/amdgpu: increase fragmentation size for Vega10 v2") Signed-off-by: Dan Carpenter diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c

Re: [PATCH 2/2] dma-buf: fix reservation_object_wait_timeout_rcu to wait correctly v2

Quoting Alex Deucher (2017-08-10 18:01:49) > From: Christian König > > With hardware resets in mind it is possible that all shared fences are > signaled, but the exlusive isn't. Fix waiting for everything in this > situation. I'm still puzzling over this one. Setting

Re: [PATCH libdrm] drm: Remove create_handle() drm_framebuffer "virtual".

On Wed, Aug 9, 2017 at 4:13 PM, Joe Kniss wrote: > On Wed, Aug 9, 2017 at 12:14 PM, Noralf Trønnes wrote: >> >> Den 09.08.2017 01.42, skrev Joe Kniss: >>> >>> Because all drivers currently use gem objects for framebuffer planes, >>> the virtual

RE: [PATCH] drm/amd/amdgpu: fix bug fail to remove debugfs when rmmod

Hi Alex, The following files are all left. amdgpu_fence_info amdgpu_firmware_info amdgpu_gem_info amdgpu_gpu_reset amdgpu_gtt_mm amdgpu_pm_info amdgpu_sa_info amdgpu_test_ib amdgpu_vram_mm ttm_dma_page_pool ttm_page_pool Instead of fini them separately, how about implement the callback

Re: [PATCH libdrm] drm/amdgpu: add new low overhead command submission API. (v2)

On 10 August 2017 at 23:56, Emil Velikov wrote: > Hi Dave, > > On 18 July 2017 at 04:52, Dave Airlie wrote: > >> +int amdgpu_cs_submit_raw(amdgpu_device_handle dev, >> +amdgpu_context_handle context, >> +

Re: [PATCH 2/2] drm/amd/display/dc: make dce120_link_encoder_create static

Both patches are Reviewed-by: Harry Wentland Harry On 2017-08-10 02:22 PM, Alex Deucher wrote: > Only used in dce120_resource.c. > > Signed-off-by: Alex Deucher > --- > drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c | 2 +- > 1 file

Re: [PATCH 6/6] drm/amdgpu/cik: add IP modules for DC for APUs

All patches are Reviewed-by: Harry Wentland Harry On 2017-08-10 02:39 PM, Alex Deucher wrote: > Enable DC for DCE8 APUs. > > Signed-off-by: Alex Deucher > --- > drivers/gpu/drm/amd/amdgpu/cik.c | 8 > 1 file changed, 8 insertions(+)

Re: [PATCH] drm/amdgpu: potential shift wrapping bug

Am 10.08.2017 um 14:16 schrieb Dan Carpenter: "frag_align" is a u64, so presumably we want to use the high bits as well instead of shift wrapping. Fixes: 6be7adb37d9b ("drm/amdgpu: increase fragmentation size for Vega10 v2") Signed-off-by: Dan Carpenter The fragment

Re: [PATCH] drm/amdgpu: potential shift wrapping bug

On Thu, Aug 10, 2017 at 02:30:15PM +0200, Christian König wrote: > Am 10.08.2017 um 14:16 schrieb Dan Carpenter: > > "frag_align" is a u64, so presumably we want to use the high bits as > > well instead of shift wrapping. > > > > Fixes: 6be7adb37d9b ("drm/amdgpu: increase fragmentation size for

Re: [PATCH] drm/amdgpu: potential shift wrapping bug

On Thu, Aug 10, 2017 at 02:30:15PM +0200, Christian König wrote: > Am 10.08.2017 um 14:16 schrieb Dan Carpenter: > > "frag_align" is a u64, so presumably we want to use the high bits as > > well instead of shift wrapping. > > > > Fixes: 6be7adb37d9b ("drm/amdgpu: increase fragmentation size for