On Wed, Jul 11, 2018 at 1:27 PM, Michel Dänzer wrote:
> From: Michel Dänzer
>
> These are ports of the outstanding applicable changes from the amdgpu
> driver.
Series is:
Acked-by: Alex Deucher
>
> Emil Velikov (1):
> Do not export the DriverRec RADEON
>
> Jim Qu (1):
> Wait for pending
Hi Dave,
A few display and GPUVM fixes for 4.18.
The following changes since commit dc81aab1be9fac2e11f31fe7538a50705eba08cf:
MAINTAINERS: update drm tree (2018-07-10 10:59:58 +1000)
are available in the git repository at:
git://people.freedesktop.org/~agd5f/linux drm-fixes-4.18
for you
With more thinking for you performance reason, Can we go further more
not to create temp bo list at all? directly add them into
parser->validated list?
In fact, if bo array is very long, then overhead of bo list creation in
CS is considerable, which will double iterate all BOs compared to
On 2018年07月12日 11:09, Zhou, David(ChunMing) wrote:
Hi Andrey,
Could you add compatibility flag or increase kms driver version? So that user
space can keep old path when using new one.
Sorry for noise, it's already at end of path.
Regards,
David Zhou
Regards,
David Zhou
-Original
On Wed, Jul 11, 2018 at 10:58 PM, Quan, Evan wrote:
>> You could simplify this to:
>> (clocks.data[i].clocks_in_khz / 10 == now) ? "*" : "");
>>
> It cannot be simplified as this. Since we need to do comparation in Mhz.
> Otherwise we will miss the 'equal' situation.
Ah I see,
Reviewed-by: Alex
Hi Andrey,
Could you add compatibility flag or increase kms driver version? So that user
space can keep old path when using new one.
Regards,
David Zhou
-Original Message-
From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf Of
zhoucm1
Sent: Thursday, July 12, 2018
> You could simplify this to:
> (clocks.data[i].clocks_in_khz / 10 == now) ? "*" : "");
>
It cannot be simplified as this. Since we need to do comparation in Mhz.
Otherwise we will miss the 'equal' situation.
Regards,
Evan
> -Original Message-
> From: Alex Deucher
On Wed, Jul 11, 2018 at 10:09 PM, zhoucm1 wrote:
>
>
> On 2018年07月12日 08:47, Marek Olšák wrote:
>>
>> From: Marek Olšák
>
> less patch comment to describe why amdgpu_bo_handle_type_kms doesn't meet
> requriement and what patch does.
The comment in amdgpu.h describes it well. (at least I hope
From: Shaoyun Liu
Compute contexts cannot keep going after a GPU reset. Currently the process
must terminate. In the future a process may be able recreate its context
from scratch. Either way, there is no need to restore the GPUVM page table
from shadow BOs.
Signed-off-by: Shaoyun Liu
From: Shaoyun Liu
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
include/uapi/linux/kfd_ioctl.h | 20 +++-
1 file changed, 19 insertions(+), 1 deletion(-)
diff --git a/include/uapi/linux/kfd_ioctl.h b/include/uapi/linux/kfd_ioctl.h
From: Shaoyun Liu
Lock KFD and evict existing queues on reset. Notify user mode by
signaling hw_exception events.
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 3 +++
From: Yong Zhao
This will make reading code much easier.
Signed-off-by: Yong Zhao
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
.../gpu/drm/amd/amdkfd/kfd_device_queue_manager.c | 105 +++--
.../gpu/drm/amd/amdkfd/kfd_device_queue_manager.h | 2 +-
From: Shaoyun Liu
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_device.c | 10 ++
drivers/gpu/drm/amd/amdkfd/kfd_module.c | 2 ++
drivers/gpu/drm/amd/amdkfd/kfd_priv.h | 4
3 files changed, 16
From: Shaoyun Liu
The reset will be performed in a new hw_exception work thread to
handle HWS hang without blocking the thread that detected the hang.
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
.../gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
From: Shaoyun Liu
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/include/kgd_kfd_interface.h | 10 ++
1 file changed, 10 insertions(+)
diff --git a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h
This is needed to prevent deadlocks when MMU notifiers run in
reclaim-FS context and take the DQM lock for userptr evictions.
Previously this was done by making all memory allocations under
DQM locks GFP_NOIO. This is error prone. Using
memalloc_nofs_save/restore will reliably affect all memory
From: Shaoyun Liu
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c | 48 ++
drivers/gpu/drm/amd/amdkfd/kfd_device.c| 23 +++
From: Lan Xiao
Upon VM Fault, the VMID and PASID written by HW are zeros in
Hawaii. Instead of reading from ih_ring_entry, read directly
from the registers. This workaround fix the soft hang issues
caused by mishandled VM Fault in Hawaii.
Signed-off-by: Lan Xiao
Signed-off-by: Felix Kuehling
This is no longer needed with the memalloc_nofs_save/restore in
dqm_lock/unlock.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_device.c | 4 ++--
drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c | 2 +-
drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c | 4 ++--
From: Yong Zhao
Signed-off-by: Yong Zhao
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/cik_int.h | 5 ++-
drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.c| 1 -
drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.h| 37
From: Moses Reuben
Signed-off-by: Moses Reuben
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_events.c | 7 +++
1 file changed, 7 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_events.c
b/drivers/gpu/drm/amd/amdkfd/kfd_events.c
From: shaoyunl
amdgpu save the vm fault related information for KFD usage and keep the
copy until KFD read it.
Signed-off-by: shaoyun liu
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h| 3 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c | 1 +
From: Shaoyun Liu
Hook up the gpu_recover callback from KFD to amdgpu to enable
handling of GPU hangs detected by KFD.
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c| 7 +++
From: Shaoyun Liu
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 20
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 4
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 5 +
3
From: Yong Zhao
This makes all module parameters use the same form. Meanwhile clean up
the surrounding code.
Signed-off-by: Yong Zhao
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c | 2 +-
This series fixes some KFD issues, adds robustness enhancements and
finally a few cleanups.
Patches 1-4 are important fixes.
Patches 5-8 add handling of GPU VM faults
Patches 9-22 add handling of GPU resets and detection of HWS hangs
Patches 23-25 are various cleanups
Felix Kuehling (2):
From: Yong Zhao
This avoids triggering a GPU reset or otherwise changing the HW
state. Instead KFD will hang, which allows HW debugging tools to
analyze the problem.
Signed-off-by: Yong Zhao
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
From: Shaoyun Liu
job could be NULL when amdgpu_device_gpu_recover is called
Signed-off-by: Shaoyun Liu
Reviewed-by: Andrey Grodzovsky
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
From: Shaoyun Liu
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c | 3 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c | 3 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c | 3 +++
3 files
From: shaoyunl
1. Pre-GFX9 the amdgpu ISR saves the vm-fault status and address per
per-vmid. amdkfd needs to get the information from amdgpu through the
new get_vm_fault_info interface. On GFX9 and later, all the required
information is in the IH ring
2. amdkfd unmaps all queues from
From: Shaoyun Liu
The bitmap index calculation should reverse the logic used on allocation
so it will clear the same bit used on allocation
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c | 9 +
1
From: Wei Lu
Return ERR_PTR(-EINVAL) if kfd_get_process fails to find the process.
This fixes kernel oopses when a child process calls KFD ioctls with
a file descriptor inherited from the parent process.
Signed-off-by: Wei Lu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
From: Shaoyun Liu
Signed-off-by: Shaoyun Liu
Reviewed-by: Felix Kuehling
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c | 3 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c | 3 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c | 3 +++
3 files
On 2018年07月12日 04:57, Andrey Grodzovsky wrote:
This change is to support MESA performace optimization.
Modify CS IOCTL to allow its input as command buffer and an array of
buffer handles to create a temporay bo list and then destroy it
when IOCTL completes.
This saves on calling for BO_LIST
On 2018年07月12日 05:27, Luís Mendes wrote:
Hi Jim,
I followed your suggestion and was able to bisect the kernel patches.
The offending patch is: drm/amdgpu: defer test IBs on the rings at
boot (V3)
commit:
2c773de2ecb8c327f2448bd1eecad224e9227087
On 2018年07月12日 08:47, Marek Olšák wrote:
From: Marek Olšák
less patch comment to describe why amdgpu_bo_handle_type_kms doesn't
meet requriement and what patch does.
less Signed-off-by.
---
amdgpu/amdgpu.h| 7 ++-
amdgpu/amdgpu_bo.c | 4
2 files changed, 10
On 07/12/2018 09:14 AM, Zhang, Jerry (Junwei) wrote:
On 07/12/2018 08:47 AM, Marek Olšák wrote:
From: Marek Olšák
---
amdgpu/amdgpu.h| 7 ++-
amdgpu/amdgpu_bo.c | 4
2 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/amdgpu/amdgpu.h b/amdgpu/amdgpu.h
index
On Wed, Jul 11, 2018 at 9:14 PM, Zhang, Jerry (Junwei)
wrote:
> On 07/12/2018 08:47 AM, Marek Olšák wrote:
>>
>> From: Marek Olšák
>>
>> ---
>> amdgpu/amdgpu.h| 7 ++-
>> amdgpu/amdgpu_bo.c | 4
>> 2 files changed, 10 insertions(+), 1 deletion(-)
>>
>> diff --git
On 07/12/2018 08:47 AM, Marek Olšák wrote:
From: Marek Olšák
---
amdgpu/amdgpu.h| 7 ++-
amdgpu/amdgpu_bo.c | 4
2 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/amdgpu/amdgpu.h b/amdgpu/amdgpu.h
index 36f91058..be83b457 100644
--- a/amdgpu/amdgpu.h
+++
From: Marek Olšák
---
amdgpu/amdgpu.h| 7 ++-
amdgpu/amdgpu_bo.c | 4
2 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/amdgpu/amdgpu.h b/amdgpu/amdgpu.h
index 36f91058..be83b457 100644
--- a/amdgpu/amdgpu.h
+++ b/amdgpu/amdgpu.h
@@ -77,21 +77,26 @@ struct
Kent just caught a similar backtrace in one of our KFD pre-submission
tests (see below)
Neither KFD nor AMDGPU are implied in the backtrace. Is this a
regression in the kernel itself? amd-kfd-staging is currently based on
4.18-rc1.
Regards,
Felix
[ 19.435544] [ cut here
Signed-off-by: Andrey Grodzovsky
Reviewed-by: Christian König
Reviewed-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c| 10 ++
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c| 6 --
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 7 +--
Stop using 'magic numbers' when registering interrupt sources.
v2: Switch to kernel style comments.
v3:
Rebase.
Signed-off-by: Andrey Grodzovsky
Acked-by: Christian König
---
.../drm/amd/include/ivsrcid/gfx/irqsrcs_gfx_9_0.h | 55 ++
v2: Rebase
v3: Use defines for CP_SQ and CP_ECC_ERROR interrupts.
Signed-off-by: Andrey Grodzovsky
Reviewed-by: Christian König
Reviewed-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/dce_v10_0.c | 6 --
drivers/gpu/drm/amd/amdgpu/dce_v11_0.c | 6 --
Stop using 'magic numbers' when registering interrupt sources.
v2:
Clean redundant comments.
Switch to kernel style comments.
v3:
Add CP_ECC_ERROR define
Signed-off-by: Andrey Grodzovsky
Acked-by: Christian König
Signed-off-by: Andrey Grodzovsky
---
Hi Jim,
I followed your suggestion and was able to bisect the kernel patches.
The offending patch is: drm/amdgpu: defer test IBs on the rings at boot (V3)
commit:
2c773de2ecb8c327f2448bd1eecad224e9227087
This change is to support MESA performace optimization.
Modify CS IOCTL to allow its input as command buffer and an array of
buffer handles to create a temporay bo list and then destroy it
when IOCTL completes.
This saves on calling for BO_LIST create and destry IOCTLs in MESA
and by this improves
Reviewed-by: Leo Liu
On 07/11/2018 02:55 PM, boyuan.zh...@amd.com wrote:
From: Boyuan Zhang
The emit_reg_write_reg_wait function was not assigned for vcn jpeg.
This patch add it back.
Signed-off-by: Boyuan Zhang
---
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 1 +
1 file changed, 1
From: Boyuan Zhang
The emit_reg_write_reg_wait function was not assigned for vcn jpeg.
This patch add it back.
Signed-off-by: Boyuan Zhang
---
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
This was to avoid extra code in amdgpu_cs_parser_bos when
amdgpu_bo_list_get is called. But I see your argument about random id
guess (practically unrealistic to happen but still a valid argument). I
will do some minor alterations to avoid this and respin the patch.
Andrey
On 07/11/2018
From: Michel Dänzer
Use our own BoxRec for the extents, and RegionEmpty for clearing the
scanout damage region.
(Ported from amdgpu commit 72c3e9c7308fbcdf85708b72f9be14a5f2f8e7b5)
Signed-off-by: Michel Dänzer
---
src/drmmode_display.c | 16
1 file changed, 4 insertions(+),
From: Michel Dänzer
When enabling a secondary GPU output, Xorg can try resizing the screen
beyond the limit advertised by the driver, leading to drmModeAddFB
failing and primary GPU outputs turning off. Check for this and bail
instead.
(Ported from amdgpu commit
From: Michel Dänzer
Preparation for next change, no behaviour change intended.
(Ported from amdgpu commit 04a5c5f7cfacad8d9ccffe81e388cc3da2036cb5)
Signed-off-by: Michel Dänzer
---
src/drmmode_display.h | 2 +-
src/radeon_kms.c | 16
2 files changed, 9 insertions(+), 9
From: Michel Dänzer
This allows a following change to be slightly simpler.
(Ported from amdgpu commit 8fcc3a9b43d3907052a83a96e5a2423afab5ad3f)
Signed-off-by: Michel Dänzer
---
src/radeon_drm_queue.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/radeon_drm_queue.c
From: Michel Dänzer
Shorter and sweeter. :)
(Ported from amdgpu commit e8e688f3852fb06b0c34ed5bce47c9493bcd1613)
Signed-off-by: Michel Dänzer
---
src/ati.c| 4 ++--
src/drmmode_display.c| 8
src/evergreen_exa.c | 10 +-
src/r600_exa.c
From: Michel Dänzer
Set crtc->driverIsPerformingTransform for any case we can handle before
calling xf86CrtcRotate. We already clear it afterwards when the latter
clears crtc->transform_in_use.
This should allow our separate scanout buffer mechanism to be used in
more cases.
(Cherry picked
From: Michel Dänzer
We don't need to wait for a non-TearFree scanout update before scanning
out from the screen pixmap or before flipping, as the scanout update
won't be visible anyway. Instead, just abort it.
(Ported from amdgpu commit 36d01989cd842588f12fdae5b2cba5fdcf9c91dd)
Signed-off-by:
From: Michel Dänzer
This has always been disabled, no need to keep it.
(Ported from amdgpu commit 19a40758be04e1d451a030f452efb49e8aaad541)
Signed-off-by: Michel Dänzer
---
src/drmmode_display.c | 12
1 file changed, 12 deletions(-)
diff --git a/src/drmmode_display.c
From: Michel Dänzer
These are ports of the outstanding applicable changes from the amdgpu
driver.
Emil Velikov (1):
Do not export the DriverRec RADEON
Jim Qu (1):
Wait for pending scanout update before calling
drmmode_crtc_scanout_free
Keith Packard (3):
modesetting: Record
From: Michel Dänzer
This prevents a nested call to drmHandleEvent, which would hang.
Fixes hangs when disabling TearFree on a CRTC while a DRI3 client is
page flipping.
(Ported from amdgpu commit 04947b83cce3a7782e59dece2c7797cc396c1e05)
Signed-off-by: Michel Dänzer
---
From: Michel Dänzer
Instead of from drmmode_set_mode_major. There's no need to re-set the
gamma LUT on every modeset, the kernel should preserve it.
(Ported from amdgpu commit baea4fa492f635cdfe746a84be2e337d9aeae8a9)
Signed-off-by: Michel Dänzer
---
src/drmmode_display.c | 26
From: Michel Dänzer
When disabling TearFree, drmmode_crtc->scanout_id could remain as 1,
but drmmode_set_mode_major would destroy drmmode_crtc->scanout[1], so
scanout_do_update() would keep bailing, and the scanout buffer would
stop being updated.
Fixes freeze after disabling TearFree on a CRTC
From: Keith Packard
Save any value of the kernel non-desktop property in the xf86Output
structure to avoid non-desktop outputs in the default configuration.
[Also bump randrproto requirement to a version that defines
RR_PROPERTY_NON_DESKTOP - ajax]
Signed-off-by: Keith Packard
(Ported from
From: Keith Packard
Signed-off-by: Keith Packard
(Ported from xserver commit e4e3447603b5fd3a38a92c3f972396d1f81168ad)
Signed-off-by: Michel Dänzer
(Ported from amdgpu commit 61040bdfa360975614fb47aa7ea1b3a1abac3427)
---
configure.ac | 2 +-
src/drmmode_display.c | 159
From: Keith Packard
This lets a DRM client map between X outputs and kernel connectors.
v2:
Change CONNECTOR_ID to enum -- Adam Jackson
Signed-off-by: Keith Packard
(Ported from xserver commit 023d4aba8d45e9e3630b944ecfb650c081799b96)
Signed-off-by: Michel Dänzer
(Ported from amdgpu
From: Michel Dänzer
Leo pointed out that drmmode_do_crtc_dpms wasn't getting called when
turning off an output with
xrandr --output --off
This meant that the vblank sequence number and timestamp wouldn't be
saved before turning off the CRTC in this case.
Reported-by: Leo (Sunpeng) Li
From: Michel Dänzer
Without this, RandR would report the CRTC and its outputs as enabled,
even though they were actually off due to the failure.
(Cherry picked from amdgpu commit 4dcda0b48d62944c841cd9540f4ad4c7ac8dee47)
Signed-off-by: Michel Dänzer
---
src/drmmode_display.c | 2 ++
1 file
From: Emil Velikov
Unused externally and should not be exported.
Signed-off-by: Emil Velikov
(Ported from amdgpu commit 7fb8b49895e225b3908c8bd186539de23afe91d1)
Signed-off-by: Michel Dänzer
---
src/radeon_probe.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
From: Michel Dänzer
Simplifies the latter slightly.
(Ported from amdgpu commit 74124f2c17dbb4b752707bb7eee398ae099e8a2c)
Signed-off-by: Michel Dänzer
---
src/drmmode_display.c | 9 +
1 file changed, 1 insertion(+), 8 deletions(-)
diff --git a/src/drmmode_display.c
From: Michel Dänzer
Preparation for the following fix, no functional change intended.
(Ported from amdgpu commit fa30f4601de7a44edfb4a95873bd648946fd4292)
Signed-off-by: Michel Dänzer
---
src/drmmode_display.c | 28 ++--
1 file changed, 18 insertions(+), 10
From: Michel Dänzer
It means we are not using our own scanout buffers.
Fixes crash when TearFree is supposed to be enabled, but
drmmode_handle_transform doesn't set crtc->driverIsPerformingTransform.
Bugzilla: https://bugs.freedesktop.org/105736
(Ported from commit
From: Jim Qu
There is a case that when set screen from reverse to normal, the old
scanout damage is freed in modesetting before scanout update handler,
so it causes segment fault issue.
Signed-off-by: Jim Qu
[ Michel Dänzer: Only call drmmode_crtc_wait_pending_event before
From: Michel Dänzer
Instead of CPU invisible VRAM. Preparation for the following, no
functional change intended.
Cc: sta...@vger.kernel.org
Signed-off-by: Michel Dänzer
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c| 5 ++---
From: Michel Dänzer
This shouldn't happen, but if it does, we'll get a backtrace of the
caller, and update the pin_size values as needed.
Signed-off-by: Michel Dänzer
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 37 +-
1 file changed, 29 insertions(+), 8 deletions(-)
From: Michel Dänzer
Finally got confirmation from the debugging patch in the bug report
that these values need to be atomic, or they can get corrupted.
Patch 1 is so that it's not necessary to use two atomic values for
computing the amount of non-pinned CPU visible VRAM, as that could
still
From: Michel Dänzer
Concurrent execution of the non-atomic arithmetic could result in
completely bogus values.
Cc: sta...@vger.kernel.org
Bugzilla: https://bugs.freedesktop.org/106872
Signed-off-by: Michel Dänzer
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 6 +++---
I've been occasionally getting the debugging warnings seen in the
attached kernel log excerpt. Only for piglit amd_pinned_memory and for
libdrm amdgpu_test, so I suspect it's pointing at a userptr related
issue. Christian, any ideas?
--
Earthling Michel Dänzer |
Am 11.07.2018 um 17:37 schrieb Andrey Grodzovsky:
On 07/11/2018 11:05 AM, Christian König wrote:
Am 11.07.2018 um 16:51 schrieb Andrey Grodzovsky:
[SNIP]
+
+ bool destroy_bo_list;
I think you don't need this. IIRC the bo_list structure was reference
counted.
So all you need to do is
On 07/11/2018 11:05 AM, Christian König wrote:
Am 11.07.2018 um 16:51 schrieb Andrey Grodzovsky:
This change is to support MESA performace optimization.
Modify CS IOCTL to allow its input as command buffer and an array of
buffer handles to create a temporay bo list and then destroy it
when
On Wed, Jul 11, 2018 at 4:26 AM, Evan Quan wrote:
> Convert the clocks into right Mhz unit. Otherwise, it will miss
> the equal situation.
>
> Change-Id: I80e1af975715712f5ddef4ca6a6e356d90932613
> Signed-off-by: Evan Quan
> ---
> drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c | 4 ++--
> 1
Am 11.07.2018 um 16:51 schrieb Andrey Grodzovsky:
This change is to support MESA performace optimization.
Modify CS IOCTL to allow its input as command buffer and an array of
buffer handles to create a temporay bo list and then destroy it
when IOCTL completes.
This saves on calling for BO_LIST
This change is to support MESA performace optimization.
Modify CS IOCTL to allow its input as command buffer and an array of
buffer handles to create a temporay bo list and then destroy it
when IOCTL completes.
This saves on calling for BO_LIST create and destry IOCTLs in MESA
and by this improves
getrawmonotonic64() and get_monotonic_boottime64() are deprecated
because of the nonstandard naming.
The replacement functions ktime_get_raw_ns() and ktime_get_boot_ns()
also simplify the callers.
Reviewed-by: Felix Kuehling .
Signed-off-by: Arnd Bergmann
---
Patch was originall sent on Jun 18
On Wed, Jul 11, 2018 at 01:13:18PM +0200, Michal Hocko wrote:
> On Wed 11-07-18 13:14:47, Leon Romanovsky wrote:
> > On Wed, Jul 11, 2018 at 11:03:53AM +0200, Michal Hocko wrote:
> > > On Tue 10-07-18 19:20:20, Leon Romanovsky wrote:
> > > > On Tue, Jul 10, 2018 at 04:14:10PM +0200, Michal Hocko
On Wed 11-07-18 13:14:47, Leon Romanovsky wrote:
> On Wed, Jul 11, 2018 at 11:03:53AM +0200, Michal Hocko wrote:
> > On Tue 10-07-18 19:20:20, Leon Romanovsky wrote:
> > > On Tue, Jul 10, 2018 at 04:14:10PM +0200, Michal Hocko wrote:
> > > > On Tue 10-07-18 16:40:40, Leon Romanovsky wrote:
> > > >
On Wed, 11 Jul 2018 12:21:03 +0200,
jimqu wrote:
>
>
>
> On 2018年07月11日 17:53, Takashi Iwai wrote:
> > On Wed, 11 Jul 2018 11:26:11 +0200,
> > jimqu wrote:
> >>
> >>
> >> On 2018年07月11日 17:04, Takashi Iwai wrote:
> >>> On Wed, 11 Jul 2018 10:41:38 +0200,
> >>> jimqu wrote:
>
> On
On 2018年07月11日 17:53, Takashi Iwai wrote:
On Wed, 11 Jul 2018 11:26:11 +0200,
jimqu wrote:
On 2018年07月11日 17:04, Takashi Iwai wrote:
On Wed, 11 Jul 2018 10:41:38 +0200,
jimqu wrote:
On 2018年07月11日 15:19, Takashi Iwai wrote:
On Tue, 10 Jul 2018 13:21:00 +0200,
Takashi Iwai wrote:
revert
On Wed, Jul 11, 2018 at 11:03:53AM +0200, Michal Hocko wrote:
> On Tue 10-07-18 19:20:20, Leon Romanovsky wrote:
> > On Tue, Jul 10, 2018 at 04:14:10PM +0200, Michal Hocko wrote:
> > > On Tue 10-07-18 16:40:40, Leon Romanovsky wrote:
> > > > On Mon, Jul 09, 2018 at 02:29:08PM +0200, Michal Hocko
On Wed, Jul 11, 2018 at 04:25:04PM +0800, Evan Quan wrote:
> The save restore list initialization does not have to be pg guarded.
> And for some asic(e.g. Vega12), it does not have cntl/gpm/srm lists.
>
> Change-Id: I85c0e3525ca7fb385c3d0b9e5abc13708c91e795
> Signed-off-by: Evan Quan
Thanks for
HI Luis,
Let us trace the issue one by one.
IB test fail:
This should be regression issue on 4.18, you can bisect the kernel patches.
GPU hang:
Fix IB test fail first.
Thanks
JimQu
On 2018年07月11日 17:34, Luís Mendes wrote:
Hi Jim,
Thanks for your interest in this issue. Actually
On Wed, 11 Jul 2018 11:26:11 +0200,
jimqu wrote:
>
>
>
> On 2018年07月11日 17:04, Takashi Iwai wrote:
> > On Wed, 11 Jul 2018 10:41:38 +0200,
> > jimqu wrote:
> >>
> >>
> >> On 2018年07月11日 15:19, Takashi Iwai wrote:
> >>> On Tue, 10 Jul 2018 13:21:00 +0200,
> >>> Takashi Iwai wrote:
> > revert
On 2018年07月11日 17:04, Takashi Iwai wrote:
On Wed, 11 Jul 2018 10:41:38 +0200,
jimqu wrote:
On 2018年07月11日 15:19, Takashi Iwai wrote:
On Tue, 10 Jul 2018 13:21:00 +0200,
Takashi Iwai wrote:
revert the fix of amdgpu suspend issue, audio issue also can be observed.
Did you check the
On Wed, 11 Jul 2018 10:41:38 +0200,
jimqu wrote:
>
>
>
> On 2018年07月11日 15:19, Takashi Iwai wrote:
> > On Tue, 10 Jul 2018 13:21:00 +0200,
> > Takashi Iwai wrote:
> >>> revert the fix of amdgpu suspend issue, audio issue also can be observed.
> >> Did you check the behavior with the single AMD
On Tue 10-07-18 19:20:20, Leon Romanovsky wrote:
> On Tue, Jul 10, 2018 at 04:14:10PM +0200, Michal Hocko wrote:
> > On Tue 10-07-18 16:40:40, Leon Romanovsky wrote:
> > > On Mon, Jul 09, 2018 at 02:29:08PM +0200, Michal Hocko wrote:
> > > > On Wed 27-06-18 09:44:21, Michal Hocko wrote:
> > > > >
On 2018年07月11日 15:19, Takashi Iwai wrote:
On Tue, 10 Jul 2018 13:21:00 +0200,
Takashi Iwai wrote:
revert the fix of amdgpu suspend issue, audio issue also can be observed.
Did you check the behavior with the single AMD GPU hardware?
If confirmed, we can forget about vga_switcheroo.
... and
On Tue, Jul 10, 2018 at 04:53:39PM -0500, Alex Deucher wrote:
> Use the PP_STATE_SUPPORT_* rather than AMD_CG_SUPPORT_*
> when communicating with the SMU.
>
> Signed-off-by: Alex Deucher
Reviewed-by: Huang Rui
> ---
> drivers/gpu/drm/amd/amdgpu/vi.c | 12 ++--
> 1 file changed, 6
Convert the clocks into right Mhz unit. Otherwise, it will miss
the equal situation.
Change-Id: I80e1af975715712f5ddef4ca6a6e356d90932613
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git
The save restore list initialization does not have to be pg guarded.
And for some asic(e.g. Vega12), it does not have cntl/gpm/srm lists.
Change-Id: I85c0e3525ca7fb385c3d0b9e5abc13708c91e795
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 18 --
1 file
On Tue, 10 Jul 2018 13:21:00 +0200,
Takashi Iwai wrote:
>
> > revert the fix of amdgpu suspend issue, audio issue also can be observed.
>
> Did you check the behavior with the single AMD GPU hardware?
> If confirmed, we can forget about vga_switcheroo.
... and taking a look back at the recent
Yes, Rv gfxoff is not affected.
Regards,
Evan
> -Original Message-
> From: Alex Deucher [mailto:alexdeuc...@gmail.com]
> Sent: Wednesday, July 11, 2018 1:20 AM
> To: Quan, Evan
> Cc: amd-gfx@lists.freedesktop.org
> Subject: Re: [PATCH 07/10] drm/amdgpu: reduce the idle period that RLC
100 matches
Mail list logo