Hello Tao Zhou,
The patch f9317014ea51: "drm/amdgpu: add function to creat all ras
debugfs node" from Mar 6, 2020, leads to the following static checker
warning:
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c:1132
amdgpu_ras_debugfs_create_all()
warn: variable dereferenced before check
From: Randy Dunlap
Fix the Kconfig dependencies so that the menu is presented
correctly by adding a dependency on DRM_AMDGPU to the "menu"
Kconfig statement. This makes a continuous dependency on
DRM_AMDGPU in the DRM AMD menus and eliminates a broken menu
structure.
Fixes: a8fe58cec351 ("drm/a
Am 11.03.20 um 21:55 schrieb Nirmoy:
On 3/11/20 9:35 PM, Andrey Grodzovsky wrote:
On 3/11/20 4:32 PM, Nirmoy wrote:
On 3/11/20 9:02 PM, Andrey Grodzovsky wrote:
On 3/11/20 4:00 PM, Andrey Grodzovsky wrote:
On 3/11/20 4:00 PM, Nirmoy Das wrote:
[SNIP]
@@ -1257,6 +1258,9 @@ static int amd
On 11/03/2020 18:35, Jason Gunthorpe wrote:
From: Jason Gunthorpe
This eventually calls into handle_mm_fault() which is a sleeping function.
Release the lock first.
hmm_vma_walk_hole() does not touch the contents of the PUD, so it does not
need the lock.
Fixes: 3afc423632a1 ("mm: pagewalk: ad
[AMD Official Use Only - Internal Distribution Only]
Hi Carpenter,
Thanks for your report and advice, I will update the code.
Regards,
Stanley
-Original Message-
From: amd-gfx On Behalf Of Dan Carpenter
Sent: Thursday, March 12, 2020 3:34 PM
To: Zhou1, Tao
Cc: amd-gfx@lists.freedeskto
clean up unused header in swsmu driver stack:
1. pp_debug.h
2. amd_pcie.h
3. soc15_common.h
Signed-off-by: Kevin Wang
---
drivers/gpu/drm/amd/powerplay/amdgpu_smu.c | 3 ---
drivers/gpu/drm/amd/powerplay/arcturus_ppt.c | 1 -
drivers/gpu/drm/amd/powerplay/navi10_ppt.c | 2 --
drivers/gpu/drm
By refactoring to deal with the !pud_huge(pud) || !pud_devmap(pud)
condition early it's possible to remove the 'ret' variable and remove a
level of indentation from half the function making the code easier to
read.
No functional change.
Signed-off-by: Steven Price
---
Thanks to Jason's changes t
Fix the warning
"warn: variable dereferenced before check 'obj' (see line 1131)"
by removing unnecessary checks as amdgpu_ras_debugfs_create_all()
is only called from amdgpu_debugfs_init() where obj member in
con->head list is not NULL.
Use list_for_each_entry() instead list_for_each_entry_safe() a
On 3/12/20 9:50 AM, Christian König wrote:
Am 11.03.20 um 21:55 schrieb Nirmoy:
On 3/11/20 9:35 PM, Andrey Grodzovsky wrote:
On 3/11/20 4:32 PM, Nirmoy wrote:
On 3/11/20 9:02 PM, Andrey Grodzovsky wrote:
On 3/11/20 4:00 PM, Andrey Grodzovsky wrote:
On 3/11/20 4:00 PM, Nirmoy Das wrote:
[AMD Official Use Only - Internal Distribution Only]
Reviewed-by: Hawking Zhang
Regards,
Hawking
-Original Message-
From: Stanley.Yang
Sent: Thursday, March 12, 2020 18:43
To: amd-gfx@lists.freedesktop.org
Cc: Zhang, Hawking ; Chen, Guchun ;
Li, Dennis ; Clements, John ; Zhou1,
Tao ;
The problem is that we can't add the clear fence to the BO
when there is an exclusive fence on it since we can't
guarantee the the clear fence will complete after the
exclusive one.
To fix this refactor the function and wait for any potential
exclusive fence with a small timeout before adding the
Can you give more details about " we can't guarantee the the clear fence will
complete after the exclusive one." ?
Thanks
_
Monk Liu|GPU Virtualization Team |AMD
-Original Message-
From: amd-gfx On Behalf Of Christian
K?nig
Sent: Thursday, March 1
We can remove the NULL check for "res_ctx" and
"res_ctx->pipe_ctx[i].stream->link". Also it's nicer to align the
conditions using spaces so I re-indented a bit.
Longer explanation: The "res_ctx" pointer points to an address in the
middle of a struct so it can't be NULL. For
"res_ctx->pipe_ctx[i]
From the semantic the dma_resv object contains a single exclusive and
multiple shared fences and it is mandatory that the shared fences
complete after the exclusive one.
Now what happens is that clearing the VM page tables runs asynchronously
to the exclusive fence which moves the buffer aroun
without your patch, the clear fence is also hooked in the shared list of bo's
reservation obj, no matter the exclusive fence of that BO signaled before
clear fence or not
since the clear fence is always kept in the bo's resv object, can you tell me
what's the problem than ? are we going to lo
This patch is
Reviewed-by: Leo Liu
-Original Message-
From: amd-gfx On Behalf Of James Zhu
Sent: March 11, 2020 4:50 PM
To: amd-gfx@lists.freedesktop.org
Cc: Zhu, James
Subject: [PATCH v5 2/4] drm/amdgpu/vcn: fix race condition issue for dpg
unpause mode switch
Couldn't only rely on e
>> Now what happens is that clearing the VM page tables runs asynchronously to
>> the exclusive fence which moves the buffer around.
The amdgpu_vm_clear_freed is already kicked off before you wait for the
exclusive fence signaled, why you can avoid clearing PT not overlap with the
"move" acti
[AMD Public Use]
Reviewed-by: Alex Deucher
From: Wang, Kevin(Yang)
Sent: Thursday, March 12, 2020 5:50 AM
To: amd-gfx@lists.freedesktop.org
Cc: Feng, Kenneth ; Deucher, Alexander
; Wang, Kevin(Yang)
Subject: [PATCH] drm/amdgpu/swsmu: clean up unused header in
On Mon, Mar 09, 2020 at 08:24:04AM +, Walter Harms wrote:
>
>
> Von: kernel-janitors-ow...@vger.kernel.org
> im Auftrag von Christophe JAILLET
>
> Gesendet: Sonntag, 8. März 2020 10:26
> An: harry.wentl...@amd.com; sunpeng...@amd.com; alexander.deuc
The problem is that dma_resv_test_signaled_rcu() tests only the shared
fence if one is present.
Now what happened is that the clear fence completed before the exclusive
one, and that in turn caused TTM to think that the BO is unused and
freed it.
Christian.
Am 12.03.20 um 14:25 schrieb Liu,
AccVGPRs are newly added in arcturus. Before reading these
registers, they should be initialized. Otherwise edc error
happens, when RAS is enabled.
Change-Id: I4ed384f0cc4b781a10cfd6ad1e3a132445bdc261
Signed-off-by: Dennis Li
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
b/drivers/gpu/drm/
Applied. thanks! (link fixed locally).
Alex
On Wed, Mar 11, 2020 at 1:07 AM Joe Perches wrote:
>
> Convert the various uses of fallthrough comments to fallthrough;
>
> Done via script
> Link:
> https://lore.kernel.org/lkml/b56602fcf79f849e733e7b521bb0e17895d390fa.1582230379.git.joe.com/
>
> Si
Applied. thanks! (link fixed locally).
Alex
On Wed, Mar 11, 2020 at 1:07 AM Joe Perches wrote:
>
> Convert the various uses of fallthrough comments to fallthrough;
>
> Done via script
> Link:
> https://lore.kernel.org/lkml/b56602fcf79f849e733e7b521bb0e17895d390fa.1582230379.git.joe.com/
>
> Si
[AMD Official Use Only - Internal Distribution Only]
Since we separated MI100 vgpr shader program configuration register list from
Vega20 ones, it's better to remove SE4 ~ SE7 thread registers from vega20 list.
Other than that, the patch is
Reviewed-by: Hawking Zhang
Regards,
Hawking
-Ori
Applied. Thanks. Link fixed locally.
Alex
On Wed, Mar 11, 2020 at 6:11 PM Joe Perches wrote:
>
> On Wed, 2020-03-11 at 17:50 -0400, Felix Kuehling wrote:
> > On 2020-03-11 12:51 a.m., Joe Perches wrote:
> > > Convert the various uses of fallthrough comments to fallthrough;
> > >
> > > Done vi
Applied. thanks!
Alex
On Thu, Mar 12, 2020 at 4:09 AM Randy Dunlap wrote:
>
> From: Randy Dunlap
>
> Fix the Kconfig dependencies so that the menu is presented
> correctly by adding a dependency on DRM_AMDGPU to the "menu"
> Kconfig statement. This makes a continuous dependency on
> DRM_AMDGP
On Thu, Mar 12, 2020 at 7:32 AM Dan Carpenter wrote:
>
> We can remove the NULL check for "res_ctx" and
> "res_ctx->pipe_ctx[i].stream->link". Also it's nicer to align the
> conditions using spaces so I re-indented a bit.
>
> Longer explanation: The "res_ctx" pointer points to an address in the
>
On Mon, Feb 24, 2020 at 5:31 AM Dan Carpenter wrote:
>
> These lines were accidentally indented 4 spaces more than they should
> be.
>
> Signed-off-by: Dan Carpenter
Applied. thanks!
Alex
> ---
> drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 8
> 1 file changed, 4 insertions(+
[AMD Official Use Only - Internal Distribution Only]
Per offline discussion, we can re-use the existing logical, gpr_reg_size =
compute_dim_x / 16 + 6; For vega20, totally 10 configuration registers, while
for Arcturus, totally 14 configuration registers. That's fine.
We just need to create a s
Applied. Thanks!
Alex
On Thu, Jan 30, 2020 at 11:58 PM Dan Carpenter wrote:
>
> If "speed" is zero then we use it as a divisor to find "prescale". It's
> better to move the check for zero to the very start of the function.
>
> Fixes: 9eeec26a1339 ("drm/amd/display: Refine i2c frequency calcula
Applied. thanks!
Alex
On Fri, Jan 24, 2020 at 9:48 AM Harry Wentland wrote:
>
> On 2020-01-23 8:07 p.m., Lyude Paul wrote:
> > These are some very loud debug statements that get printed on every
> > vblank when driver level debug printing is enabled in DRM, and doesn't
> > really tell us anythi
Applied. Thanks!
Alex
On Fri, Feb 28, 2020 at 4:36 PM Mario Kleiner
wrote:
>
> This fixes a problem found on the MacBookPro 2017 Retina panel:
>
> The panel reports 10 bpc color depth in its EDID, and the
> firmware chooses link settings at boot which support enough
> bandwidth for 10 bpc (3240
On Wed, Mar 11, 2020 at 06:28:30PM -0700, Ralph Campbell wrote:
> > mm/hmm.c | 8 ++--
> > 1 file changed, 6 insertions(+), 2 deletions(-)
> >
> > diff --git a/mm/hmm.c b/mm/hmm.c
> > index 72e5a6d9a41756..35f85424176d14 100644
> > +++ b/mm/hmm.c
> > @@ -325,6 +325,7 @@ static int hmm_vma_h
On Thu, Mar 5, 2020 at 4:21 PM Mario Kleiner wrote:
>
> Commit '16f17eda8bad ("drm/amd/display: Send vblank and user
> events at vsartup for DCN")' introduces a new way of pageflip
> completion handling for DCN, and some trouble.
>
> The current implementation introduces a race condition, which
>
On Thu, Mar 12, 2020 at 10:28:13AM +, Steven Price wrote:
> By refactoring to deal with the !pud_huge(pud) || !pud_devmap(pud)
> condition early it's possible to remove the 'ret' variable and remove a
> level of indentation from half the function making the code easier to
> read.
>
> No functi
On Wed, Mar 11, 2020 at 06:36:47PM -0700, Ralph Campbell wrote:
> > @@ -390,8 +384,15 @@ static int hmm_vma_walk_pmd(pmd_t *pmdp,
> > return -EBUSY;
> > }
> > return hmm_pfns_fill(start, end, range, HMM_PFN_NONE);
> > - } else if (!pmd_present(pmd))
> >
On 12/03/2020 14:27, Jason Gunthorpe wrote:
On Thu, Mar 12, 2020 at 10:28:13AM +, Steven Price wrote:
By refactoring to deal with the !pud_huge(pud) || !pud_devmap(pud)
condition early it's possible to remove the 'ret' variable and remove a
level of indentation from half the function making
>>> The problem is that dma_resv_test_signaled_rcu() tests only the shared
>>> fence if one is present.
Okay I got the point now, but why we cannot modify dma_resv_test_signaled_rcu()
to let it wait for both exclusive and shared lists ?
Ack-by: Monk Liu
_
M
On Thu, Mar 12, 2020 at 02:40:08PM +, Steven Price wrote:
> On 12/03/2020 14:27, Jason Gunthorpe wrote:
> > On Thu, Mar 12, 2020 at 10:28:13AM +, Steven Price wrote:
> > > By refactoring to deal with the !pud_huge(pud) || !pud_devmap(pud)
> > > condition early it's possible to remove the 'r
Am 12.03.20 um 16:03 schrieb Liu, Monk:
The problem is that dma_resv_test_signaled_rcu() tests only the shared fence if
one is present.
Okay I got the point now, but why we cannot modify dma_resv_test_signaled_rcu()
to let it wait for both exclusive and shared lists ?
That is exactly what I
On 12/03/2020 15:11, Jason Gunthorpe wrote:
On Thu, Mar 12, 2020 at 02:40:08PM +, Steven Price wrote:
On 12/03/2020 14:27, Jason Gunthorpe wrote:
On Thu, Mar 12, 2020 at 10:28:13AM +, Steven Price wrote:
By refactoring to deal with the !pud_huge(pud) || !pud_devmap(pud)
condition early
On 12/03/2020 16:37, Jason Gunthorpe wrote:
On Thu, Mar 12, 2020 at 04:16:33PM +, Steven Price wrote:
Actually, while you are looking at this, do you think we should be
adding at least READ_ONCE in the pagewalk.c walk_* functions? The
multiple references of pmd, pud, etc without locking seem
On Thu, Mar 12, 2020 at 04:16:33PM +, Steven Price wrote:
> > Actually, while you are looking at this, do you think we should be
> > adding at least READ_ONCE in the pagewalk.c walk_* functions? The
> > multiple references of pmd, pud, etc without locking seems sketchy to
> > me.
>
> I agree i
On Thu, Mar 12, 2020 at 05:02:18PM +, Steven Price wrote:
> > Having the walker deref the pointer and pass the value it into the ops
> > for use rather than repeatedly de-refing an unlocked value seems like
> > a much safer design to me.
>
> Yeah that sounds like a good idea.
Ok.. let see wh
pmd_to_hmm_pfn_flags() already checks it and makes the cpu flags 0. If no
fault is requested then the pfns should be returned with the not valid
flags.
It should not unconditionally fault if faulting is not requested.
Fixes: 2aee09d8c116 ("mm/hmm: change hmm_vma_fault() to allow write fault on
p
On 3/12/20 12:33 PM, Jason Gunthorpe wrote:
pmd_to_hmm_pfn_flags() already checks it and makes the cpu flags 0. If no
fault is requested then the pfns should be returned with the not valid
flags.
It should not unconditionally fault if faulting is not requested.
Fixes: 2aee09d8c116 ("mm/hmm: c
AccVGPRs are newly added in arcturus. Before reading these
registers, they should be initialized. Otherwise edc error
happens, when RAS is enabled.
v2: reuse the existing logical to calculate register size
Change-Id: I4ed384f0cc4b781a10cfd6ad1e3a132445bdc261
Signed-off-by: Dennis Li
diff --git
[AMD Official Use Only - Internal Distribution Only]
Reviewed-by: Hawking Zhang
Regards,
Hawking
-Original Message-
From: Dennis Li
Sent: Friday, March 13, 2020 11:22
To: amd-gfx@lists.freedesktop.org; Deucher, Alexander
; Zhou1, Tao ; Zhang, Hawking
; Chen, Guchun
Cc: Li, Dennis
[AMD Official Use Only - Internal Distribution Only]
Looks file mode is changed.
old mode 100644
new mode 100755
Regards,
Guchun
-Original Message-
From: Zhang, Hawking
Sent: Friday, March 13, 2020 11:35 AM
To: Li, Dennis ; amd-gfx@lists.freedesktop.org; Deucher,
Alexander ; Zhou1, Ta
49 matches
Mail list logo
