From: Jay Cornwall
Gfx8 HW incorrectly clamps CP_HQD_EOP_CONTROL.EOP_SIZE, which can
lead to scheduling deadlock due to SE EOP done counter overflow.
Enforce a EOP queue size limit which prevents the CP from sending
more than 0xFF events at a time.
Signed-off-by: Jay Cornwall
Reviewed-by: Feli
Remove hard-coded assumption that the first compute pipe is
reserved for amdgpu. Pipe 0 actually means pipe 0 now.
Signed-off-by: Felix Kuehling
Reviewed-by: Oded Gabbay
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c | 4 ++--
2 f
From: Kent Russell
Consolidate log commands so that dev_info(NULL, "Error...") uses the more
accurate pr_err, remove the module name from the log (can be seen via
dynamic debugging with +m), and the function name (can be seen via
dynamic debugging with +f). We also don't need debug messages sayin
From: Moses Reuben
v2:
* Renamed ALLOC_MEMORY_OF_SCRATCH to SET_SCRATCH_BACKING_VA
* Removed size parameter from the ioctl, it was unused
* Removed hole in ioctl number space
* No more call to write_config_static_mem
* Return correct error code from ioctl
Signed-off-by: Moses Reuben
Signed-off-
From: Moses Reuben
v2:
* Shortened headline
* Removed write_config_static_mem, it gets initialized by gfx_v?_0_gpu_init
* Renamed alloc_memory_of_scratch to set_scratch_backing_va
* Made set_scratch_backing_va a void function
* Documented set_scratch_backing in kgd_kfd_interface.h
Signed-off-by:
From: Yong Zhao
v2:
* Removed amdgpu_amdkfd prefix from static functions
* Documented get_tile_config in kgd_kfd_interface.h
Signed-off-by: Yong Zhao
Signed-off-by: Felix Kuehling
Acked-by: Oded Gabbay
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c | 26 +++
driver
Signed-off-by: Felix Kuehling
Reviewed-by: Oded Gabbay
---
drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.c
b/drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.c
index d5e19b5..8b14a4e 100644
--- a/drivers/gp
Various bug fixes and improvements that accumulated over the last two
years.
Signed-off-by: Felix Kuehling
Acked-by: Oded Gabbay
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 16 ++
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c | 130 +---
drivers/gpu/drm/amd/amdgpu
gtt_sa_bitmap is accessed by bitmap functions, which operate on longs.
Therefore the array should be allocated in long units. Also round up
in case the number of bits is not a multiple of BITS_PER_LONG.
Signed-off-by: Felix Kuehling
Reviewed-by: Oded Gabbay
---
drivers/gpu/drm/amd/amdkfd/kfd_de
This register only has a single instance in the hardware. Its value
applies to all VMIDS.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c | 2 +-
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 3 ++-
2 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/
It's causing problems with user mode queues and the HIQ, and can
lead to hard hangs during boot after programming RLC_CP_SCHEDULERS.
Signed-off-by: Felix Kuehling
Reviewed-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/vi.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/
From: Kent Russell
Upstream prefers the !x notation to x==NULL or x==false. Along those lines
change the ==true or !=NULL references as well. Also make the references
to !x the same, excluding () for readability.
Signed-off-by: Kent Russell
Signed-off-by: Felix Kuehling
Reviewed-by: Oded Gabba
Handle errors in doorbell aperture initialization instead of BUG_ON.
iounmap doorbell aperture during finalization.
Signed-off-by: Felix Kuehling
Reviewed-by: Oded Gabbay
---
drivers/gpu/drm/amd/amdkfd/kfd_device.c | 9 -
drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c | 13 +++--
Use shared_resources.queue_bitmap to determine the queues available
for KFD in each pipe.
Signed-off-by: Felix Kuehling
Reviewed-by: Oded Gabbay
---
drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c | 12 +---
1 file changed, 9 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu
In most cases, BUG_ONs can be replaced with WARN_ON with an error
return. In some void functions just turn them into a WARN_ON and
possibly an early exit.
v2:
* Cleaned up error handling in pm_send_unmap_queue
* Removed redundant WARN_ON in kfd_process_destroy_delayed
Signed-off-by: Felix Kuehlin
Hi Alex,
How does firmware get published for the upstream driver? Where can I
check the currently published version of both CZ and KV firmware for
upstream?
Do you publish firmware updates at the same time as patches that depend
on them?
Thanks,
Felix
On 2017-08-13 04:49 AM, Oded Gabbay wrot
I'll turn it into a dev_warn and I'll make the error message more
helpful. I think this happens when a new DID has been added to amdgpu,
but not to KFD yet. That's probably worth a warning.
Regards,
Felix
On 2017-08-14 11:50 PM, Zhao, Yong wrote:
>
> Oded, I agree with you. When I made the ch
On 16/08/17 09:35 AM, Felix Kuehling wrote:
> On 2017-08-15 06:20 AM, Oded Gabbay wrote:
>> I prefer to do it incrementally, to avoid very large patch-sets which
>> usually end up in longer cycles of review-fix, which causes you more
>> pain because internal development continues during this time a
On 16/08/17 05:12 AM, Leo Liu wrote:
> Signed-off-by: Leo Liu
There's a typo ("phycial") in the shortlog, and acronyms should always
be written in upper case. So it should probably be:
drm/amdgpu: Add MJPEG check for UVD physical mode msg buffer
--
Earthling Michel Dänzer |
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Leo Liu
> Sent: Tuesday, August 15, 2017 4:13 PM
> To: amd-gfx@lists.freedesktop.org
> Cc: Liu, Leo
> Subject: [PATCH] drm/amdgpu: add mjpeg check for uvd phycial mode msg
> buffer
>
> Signed
Signed-off-by: Leo Liu
---
drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c | 4
1 file changed, 4 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
index ff8ae50..b464f62 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
+++ b/drivers/g
Update UVD 7.0 enc test according to firmware interface changes for session
info ib.
Please review.
Regards,
Boyuan
From 41a3084055bb070b0541315cef5cfa7356624afa Mon Sep 17 00:00:00 2001
From: Boyuan Zhang
Date: Tue, 15 Aug 2017 16:29:37 -0400
Subject: [PATCH] drm/amdgpu: update uvd enc tes
Am 15.08.2017 um 18:03 schrieb Felix Kuehling:
On 2017-08-15 04:24 AM, Christian König wrote:
Am 14.08.2017 um 17:31 schrieb Felix Kuehling:
[SNIP]
Repeating the same argument I made on another email:
Commented on that in the other mail, let's keep the discussion on this
topic there.
BTW: Wh
Am 15.08.2017 um 17:33 schrieb Boyuan Zhang:
Signed-off-by: Boyuan Zhang
Acked-by: Alex Deucher
Acked-by: Christian König
---
tests/amdgpu/Makefile.am | 1 +
tests/amdgpu/amdgpu_test.c | 6 +
tests/amdgpu/amdgpu_test.h | 15 ++
tests/amdgpu/frame.h | 2 +-
test
Signed-off-by: Boyuan Zhang
Acked-by: Alex Deucher
---
tests/amdgpu/Makefile.am | 1 +
tests/amdgpu/amdgpu_test.c | 6 +
tests/amdgpu/amdgpu_test.h | 15 ++
tests/amdgpu/frame.h | 2 +-
tests/amdgpu/uvd_enc_tests.c | 500
tests/amdg
Am 15.08.2017 um 17:17 schrieb Harry Wentland:
On 2017-08-15 10:36 AM, Alex Deucher wrote:
Was using the wrong prefix (gmc rather than gfx). The function
is related to the gfx hw, not gmc. This also makes it consistent
with the naming in gfx8.
Signed-off-by: Alex Deucher
Reviewed-by: Harry
On 2017-08-15 10:36 AM, Alex Deucher wrote:
> Was using the wrong prefix (gmc rather than gfx). The function
> is related to the gfx hw, not gmc. This also makes it consistent
> with the naming in gfx8.
>
> Signed-off-by: Alex Deucher
Reviewed-by: Harry Wentland
Harry
> ---
> drivers/gpu/d
> -Original Message-
> From: Kuehling, Felix
> Sent: Monday, August 14, 2017 9:13 PM
> To: Oded Gabbay; Marek Olšák; Deucher, Alexander
> Cc: amd-gfx list
> Subject: Re: [PATCH 1/4] drm/amdgpu: Adding new kgd/kfd interface
> functions to support scratch memory
>
> [+Marek, Alex for comment
On Tue, Aug 15, 2017 at 4:26 AM, Christian König
wrote:
> Ok in this case I don't think we should do anything here.
>
> That we can't rmmod the driver on pre 4.11 is a bit annoying, but doesn't
> affect the upstream driver.
This patch can be applied to the KCL for dkms.
Alex
>
> Christian.
>
>
Was using the wrong prefix (gmc rather than gfx). The function
is related to the gfx hw, not gmc. This also makes it consistent
with the naming in gfx8.
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git
On Tue, Aug 15, 2017 at 4:24 AM, Christian König
wrote:
> Am 14.08.2017 um 17:31 schrieb Felix Kuehling:
>>
>> [SNIP]
>> Repeating the same argument I made on another email:
>
>
> Commented on that in the other mail, let's keep the discussion on this topic
> there.
>
>>> BTW: What exactly this thi
On 15/08/17 08:16 AM, Tom St Denis wrote:
On 15/08/17 08:11 AM, Christian König wrote:
Am 15.08.2017 um 13:33 schrieb Tom St Denis:
On 15/08/17 07:22 AM, Christian König wrote:
Am 14.08.2017 um 13:37 schrieb Tom St Denis:
This patch adds trace points to the amdgpu bo_kmap/bo/kunmap functions
On 15/08/17 08:11 AM, Christian König wrote:
Am 15.08.2017 um 13:33 schrieb Tom St Denis:
On 15/08/17 07:22 AM, Christian König wrote:
Am 14.08.2017 um 13:37 schrieb Tom St Denis:
This patch adds trace points to the amdgpu bo_kmap/bo/kunmap functions
which capture internal allocations.
(v2):
Am 15.08.2017 um 13:33 schrieb Tom St Denis:
On 15/08/17 07:22 AM, Christian König wrote:
Am 14.08.2017 um 13:37 schrieb Tom St Denis:
This patch adds trace points to the amdgpu bo_kmap/bo/kunmap functions
which capture internal allocations.
(v2): Simply add a sleep to the init. Users can use
On 15/08/17 07:22 AM, Christian König wrote:
Am 14.08.2017 um 13:37 schrieb Tom St Denis:
This patch adds trace points to the amdgpu bo_kmap/bo/kunmap functions
which capture internal allocations.
(v2): Simply add a sleep to the init. Users can use this script to
load with map debugging enable
Am 14.08.2017 um 13:37 schrieb Tom St Denis:
This patch adds trace points to the amdgpu bo_kmap/bo/kunmap functions
which capture internal allocations.
(v2): Simply add a sleep to the init. Users can use this script to
load with map debugging enabled:
#!/bin/bash
modprobe amdgpu map_de
Am 15.08.2017 um 12:10 schrieb Roger He:
Change-Id: I70e4ea94b8520e19cfee5ba6c9a0ecf1ee3f5f1a
Signed-off-by: Roger He
Reviewed-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 6 ++
drivers/gpu/drm/amd/amdgpu/am
Am 15.08.2017 um 11:53 schrieb He, Roger:
[SNIP]
Don't adjust the global parameter here.
Better use the default handling to amdgpu_vm_adjust_size() like you had in the
last version of the patch.
For Vega10 fragment handling works a bit differently, the L1 works like
GFX8 but the L2 is currentl
Change-Id: If01e32baa903c8c35991b1517c6d8bde98f5dae2
Signed-off-by: Ken Wang
---
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 13 +
1 file changed, 13 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 68a0d40..f47ee5c 100644
--
On Tue, Aug 15, 2017 at 1:55 AM, Felix Kuehling wrote:
> OK, I was hoping to keep the ABI unchanged during upstreaming, but I
> realize that may not be realistic. And I feel I'm getting valuable
> feedback, so I don't want to limit myself with ABI requirements that are
> not necessary.
I think tha
Change-Id: If01e32baa903c8c35991b1517c6d8bde98f5dae2
Signed-off-by: Ken Wang
---
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 12
1 file changed, 12 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 68a0d40..66312d9 100644
---
Change-Id: I70e4ea94b8520e19cfee5ba6c9a0ecf1ee3f5f1a
Signed-off-by: Roger He
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 6 ++
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c| 4
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 25 +++
From: Ken Wang
Change-Id: If01e32baa903c8c35991b1517c6d8bde98f5dae2
Signed-off-by: Ken Wang
---
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 12
1 file changed, 12 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 68a0d40..6
-Original Message-
From: Koenig, Christian
Sent: Tuesday, August 15, 2017 4:48 PM
To: He, Roger ; amd-gfx@lists.freedesktop.org
Subject: Re: [PATCH 2/2] drm/amd/amdgpu: expose fragment size as module
parameter
Am 15.08.2017 um 10:36 schrieb Roger He:
> Change-Id: I70e4ea94b8520e19cfee5ba
Change-Id: If8de884538b8eca2214f21242925d854e16e63b7
Signed-off-by: Roger He
---
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 5 +
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 4 +---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 6 +-
drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c | 5 +++--
drivers/
Change-Id: I70e4ea94b8520e19cfee5ba6c9a0ecf1ee3f5f1a
Signed-off-by: Roger He
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 8
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c| 4
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 7 +--
Am 15.08.2017 um 10:36 schrieb Roger He:
Change-Id: I70e4ea94b8520e19cfee5ba6c9a0ecf1ee3f5f1a
Signed-off-by: Roger He
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 8
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c| 4
drivers/g
Only two nit picks, the first one is that we need a better commit message.
Something like: "This adds the fragment_size in the vm_manager structure
and implements hardware setup for it."
For the second see below.
Am 15.08.2017 um 10:35 schrieb Roger He:
Change-Id: If8de884538b8eca2214f212429
On 15/08/17 05:26 PM, Michel Dänzer wrote:
> From: Michel Dänzer
>
> Signed-off-by: Michel Dänzer
> ---
I meant to mention here that these patches adapt to
https://patchwork.freedesktop.org/patch/150938/ , which I'm about to push.
--
Earthling Michel Dänzer | http
From: Michel Dänzer
Signed-off-by: Michel Dänzer
---
src/amdgpu_drv.h | 26 ++
src/amdgpu_kms.c | 20 ++--
src/drmmode_display.c | 10 +++---
3 files changed, 43 insertions(+), 13 deletions(-)
diff --git a/src/amdgpu_drv.h b/src/amdgpu_drv.
Am 14.08.2017 um 17:31 schrieb Felix Kuehling:
[SNIP]
Repeating the same argument I made on another email:
Commented on that in the other mail, let's keep the discussion on this
topic there.
BTW: What exactly this this good for?
Scratch memory is private memory per work-item. It's used whe
Am 15.08.2017 um 00:55 schrieb Felix Kuehling:
OK, I was hoping to keep the ABI unchanged during upstreaming, but I
realize that may not be realistic. And I feel I'm getting valuable
feedback, so I don't want to limit myself with ABI requirements that are
not necessary.
Yeah, agree. On the GFX
52 matches
Mail list logo