Re: [git pull] drm fixes for 6.1-rc1
Thanks Arunpravin, your patch applied to the 6.1-rc1 code built a kernel that loaded the amdgpu module on my pc with Cape Verde GPU card with no problems. Regards, Arthur. On 18 October 2022 7:10:45 am ACDT, Arunpravin Paneer Selvam wrote: >Hi Christian, > >Looks like we have to exit the loop if there are no blocks to compare. >May be that's why the function returns false. > >@Arthur Marsh Could you please test the attached patch. > >Thanks, >Arun > >On 10/17/2022 1:39 PM, Christian König wrote: >> Am 17.10.22 um 10:01 schrieb Dave Airlie: >>> On Mon, 17 Oct 2022 at 17:07, Christian König >>> wrote: Hi Arun, the hw generation doesn't matter. This error message here: amdgpu: Move buffer fallback to memcpy unavailable indicates that the detection of linear buffers still doesn't work as expected or that we have a bug somewhere else. Maybe the limiting when SDMA moves are not available isn't working correctly? >>> It is a CAPE_VERDE, so maybe something with the SI UVD memory limitations? >> >> Yeah, good point. Could be that we try to move something into the UVD memory >> window and that something isn't allocated linearly. >> >> Arun can you trace the allocation and make sure that all kernel allocations >> have the CONTIGUOUS flag set? >> >> Thanks, >> Christian. >> >>> >>> Dave. >> -- Sent from my Android device with K-9 Mail. Please excuse my brevity.
Re: [git pull] drm fixes for 6.1-rc1
Hi Christian,
Looks like we have to exit the loop if there are no blocks to compare.
May be that's why the function returns false.
@Arthur Marsh Could you please test the attached patch.
Thanks,
Arun
On 10/17/2022 1:39 PM, Christian König wrote:
Am 17.10.22 um 10:01 schrieb Dave Airlie:
On Mon, 17 Oct 2022 at 17:07, Christian König
wrote:
Hi Arun,
the hw generation doesn't matter. This error message here:
amdgpu: Move buffer fallback to memcpy unavailable
indicates that the detection of linear buffers still doesn't work as
expected or that we have a bug somewhere else.
Maybe the limiting when SDMA moves are not available isn't working
correctly?
It is a CAPE_VERDE, so maybe something with the SI UVD memory
limitations?
Yeah, good point. Could be that we try to move something into the UVD
memory window and that something isn't allocated linearly.
Arun can you trace the allocation and make sure that all kernel
allocations have the CONTIGUOUS flag set?
Thanks,
Christian.
Dave.
From 132ce83f893eaea743fb7f41a9dc72afea52cdaa Mon Sep 17 00:00:00 2001
From: Arunpravin Paneer Selvam
Date: Mon, 17 Oct 2022 13:15:21 -0700
Subject: [PATCH] drm/amdgpu: Fix for BO move issue
If there are no blocks to compare then exit
the loop.
Signed-off-by: Arunpravin Paneer Selvam
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index dc262d2c2925..57277b1cf183 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -439,6 +439,9 @@ static bool amdgpu_mem_visible(struct amdgpu_device *adev,
while (cursor.remaining) {
amdgpu_res_next(&cursor, cursor.size);
+ if (!cursor.remaining)
+ break;
+
/* ttm_resource_ioremap only supports contiguous memory */
if (end != cursor.start)
return false;
--
2.25.1
Re: [git pull] drm fixes for 6.1-rc1
Am 17.10.22 um 10:01 schrieb Dave Airlie: On Mon, 17 Oct 2022 at 17:07, Christian König wrote: Hi Arun, the hw generation doesn't matter. This error message here: amdgpu: Move buffer fallback to memcpy unavailable indicates that the detection of linear buffers still doesn't work as expected or that we have a bug somewhere else. Maybe the limiting when SDMA moves are not available isn't working correctly? It is a CAPE_VERDE, so maybe something with the SI UVD memory limitations? Yeah, good point. Could be that we try to move something into the UVD memory window and that something isn't allocated linearly. Arun can you trace the allocation and make sure that all kernel allocations have the CONTIGUOUS flag set? Thanks, Christian. Dave.
Re: [git pull] drm fixes for 6.1-rc1
On Mon, 17 Oct 2022 at 17:07, Christian König wrote: > > Hi Arun, > > the hw generation doesn't matter. This error message here: > > amdgpu: Move buffer fallback to memcpy unavailable > > indicates that the detection of linear buffers still doesn't work as > expected or that we have a bug somewhere else. > > Maybe the limiting when SDMA moves are not available isn't working > correctly? It is a CAPE_VERDE, so maybe something with the SI UVD memory limitations? Dave.
Re: [git pull] drm fixes for 6.1-rc1
Hi Arun,
the hw generation doesn't matter. This error message here:
amdgpu: Move buffer fallback to memcpy unavailable
indicates that the detection of linear buffers still doesn't work as
expected or that we have a bug somewhere else.
Maybe the limiting when SDMA moves are not available isn't working
correctly?
Regards,
Christian.
Am 17.10.22 um 08:54 schrieb Arunpravin Paneer Selvam:
Hi Arthur,
Is this old radeon card?
Thanks,
Arun
On 10/17/2022 11:50 AM, Christian König wrote:
Arun please take a look into this ASAP.
Thanks,
Christian.
Am 17.10.22 um 03:13 schrieb Arthur Marsh:
Thanks Dave, I reverted patch
312b4dc11d4f74bfe03ea25ffe04c1f2fdd13cb9 against 6.1-rc1 and the
resulting kernel loaded amdgpu fine on my pc with Cape Verde GPU.
Regards,
Arthur.
On 17 October 2022 8:14:18 am ACDT, Dave Airlie
wrote:
On Sun, 16 Oct 2022 at 18:09, Arthur Marsh
wrote:
From: Arthur Marsh
Hi, the "drm fixes for 6.1-rc1" commit caused the amdgpu module to
fail
with my Cape Verde radeonsi card.
I haven't been able to bisect the problem to an individual commit,
but
attach a dmesg extract below.
I'm happy to supply any other configuration information and test
patches.
Can you try reverting: it's the only think I can spot that might
affect a card that old since most changes in that request were for
display hw you don't have.
ommit 312b4dc11d4f74bfe03ea25ffe04c1f2fdd13cb9
Author: Arunpravin Paneer Selvam
Date: Tue Oct 4 07:33:39 2022 -0700
drm/amdgpu: Fix VRAM BO swap issue
DRM buddy manager allocates the contiguous memory requests in
a single block or multiple blocks. So for the ttm move operation
(incase of low vram memory) we should consider all the blocks to
compute the total memory size which compared with the struct
ttm_resource num_pages in order to verify that the blocks are
contiguous for the eviction process.
v2: Added a Fixes tag
v3: Rewrite the code to save a bit of calculations and
variables (Christian)
Fixes: c9cad937c0c5 ("drm/amdgpu: add drm buddy support to
amdgpu")
Signed-off-by: Arunpravin Paneer Selvam
Reviewed-by: Christian König
Signed-off-by: Alex Deucher
Thanks,
Dave.
Arthur.
Linux version 6.0.0+ (root@am64) (gcc-12 (Debian 12.2.0-5)
12.2.0, GNU ld (GNU Binutils for Debian) 2.39) #5179 SMP
PREEMPT_DYNAMIC Fri Oct 14 17:00:40 ACDT 2022
Command line: BOOT_IMAGE=/vmlinuz-6.0.0+
root=UUID=39706f53-7c27-4310-b22a-36c7b042d1a1 ro single
amdgpu.audio=1 amdgpu.si_support=1 radeon.si_support=0
page_owner=on amdgpu.gpu_recovery=1
...
[drm] amdgpu kernel modesetting enabled.
amdgpu :01:00.0: vgaarb: deactivate vga console
Console: switching to colour dummy device 80x25
[drm] initializing kernel modesetting (VERDE 0x1002:0x682B
0x1458:0x22CA 0x87).
[drm] register mmio base: 0xFE8C
[drm] register mmio size: 262144
[drm] add ip block number 0
[drm] add ip block number 1
[drm] add ip block number 2
[drm] add ip block number 3
[drm] add ip block number 4
[drm] add ip block number 5
[drm] add ip block number 6
[drm] add ip block number 7
[drm] BIOS signature incorrect 5b 7
resource sanity check: requesting [mem 0x000c-0x000d],
which spans more than PCI Bus :00 [mem 0x000d-0x000d
window]
caller pci_map_rom+0x68/0x1b0 mapping multiple BARs
amdgpu :01:00.0: No more image in the PCI ROM
amdgpu :01:00.0: amdgpu: Fetched VBIOS from ROM BAR
amdgpu: ATOM BIOS: xxx-xxx-xxx
amdgpu :01:00.0: amdgpu: Trusted Memory Zone (TMZ) feature
not supported
amdgpu :01:00.0: amdgpu: PCIE atomic ops is not supported
[drm] PCIE gen 2 link speeds already enabled
[drm] vm size is 64 GB, 2 levels, block size is 10-bit, fragment
size is 9-bit
RTL8211B Gigabit Ethernet r8169-0-300:00: attached PHY driver
(mii_bus:phy_addr=r8169-0-300:00, irq=MAC)
r8169 :03:00.0 eth0: Link is Down
amdgpu :01:00.0: amdgpu: VRAM: 2048M 0x00F4 -
0x00F47FFF (2048M used)
amdgpu :01:00.0: amdgpu: GART: 1024M 0x00FF -
0x00FF3FFF
[drm] Detected VRAM RAM=2048M, BAR=256M
[drm] RAM width 128bits DDR3
[drm] amdgpu: 2048M of VRAM memory ready
[drm] amdgpu: 3979M of GTT memory ready.
[drm] GART: num cpu pages 262144, num gpu pages 262144
amdgpu :01:00.0: amdgpu: PCIE GART of 1024M enabled (table
at 0x00F400A0).
[drm] Internal thermal controller with fan control
[drm] amdgpu: dpm initialized
[drm] AMDGPU Display Connectors
[drm] Connector 0:
[drm] HDMI-A-1
[drm] HPD1
[drm] DDC: 0x194c 0x194c 0x194d 0x194d 0x194e 0x194e 0x194f
0x194f
[drm] Encoders:
[drm] DFP1: INTERNAL_UNIPHY
[drm] Connector 1:
[drm] DVI-D-1
[drm] HPD2
[drm] DDC: 0x1950 0x1950 0x1951 0x1951 0x1952 0x1952 0x1953
0x1953
[drm] Encoders:
[drm] DFP2: INTERNAL_UNIPHY
[drm] Connector 2:
[drm] VGA-1
[drm] DDC: 0x1970 0x1970 0x1971
Re: [git pull] drm fixes for 6.1-rc1
Hi Arthur,
Is this old radeon card?
Thanks,
Arun
On 10/17/2022 11:50 AM, Christian König wrote:
Arun please take a look into this ASAP.
Thanks,
Christian.
Am 17.10.22 um 03:13 schrieb Arthur Marsh:
Thanks Dave, I reverted patch
312b4dc11d4f74bfe03ea25ffe04c1f2fdd13cb9 against 6.1-rc1 and the
resulting kernel loaded amdgpu fine on my pc with Cape Verde GPU.
Regards,
Arthur.
On 17 October 2022 8:14:18 am ACDT, Dave Airlie
wrote:
On Sun, 16 Oct 2022 at 18:09, Arthur Marsh
wrote:
From: Arthur Marsh
Hi, the "drm fixes for 6.1-rc1" commit caused the amdgpu module to
fail
with my Cape Verde radeonsi card.
I haven't been able to bisect the problem to an individual commit, but
attach a dmesg extract below.
I'm happy to supply any other configuration information and test
patches.
Can you try reverting: it's the only think I can spot that might
affect a card that old since most changes in that request were for
display hw you don't have.
ommit 312b4dc11d4f74bfe03ea25ffe04c1f2fdd13cb9
Author: Arunpravin Paneer Selvam
Date: Tue Oct 4 07:33:39 2022 -0700
drm/amdgpu: Fix VRAM BO swap issue
DRM buddy manager allocates the contiguous memory requests in
a single block or multiple blocks. So for the ttm move operation
(incase of low vram memory) we should consider all the blocks to
compute the total memory size which compared with the struct
ttm_resource num_pages in order to verify that the blocks are
contiguous for the eviction process.
v2: Added a Fixes tag
v3: Rewrite the code to save a bit of calculations and
variables (Christian)
Fixes: c9cad937c0c5 ("drm/amdgpu: add drm buddy support to amdgpu")
Signed-off-by: Arunpravin Paneer Selvam
Reviewed-by: Christian König
Signed-off-by: Alex Deucher
Thanks,
Dave.
Arthur.
Linux version 6.0.0+ (root@am64) (gcc-12 (Debian 12.2.0-5)
12.2.0, GNU ld (GNU Binutils for Debian) 2.39) #5179 SMP
PREEMPT_DYNAMIC Fri Oct 14 17:00:40 ACDT 2022
Command line: BOOT_IMAGE=/vmlinuz-6.0.0+
root=UUID=39706f53-7c27-4310-b22a-36c7b042d1a1 ro single
amdgpu.audio=1 amdgpu.si_support=1 radeon.si_support=0
page_owner=on amdgpu.gpu_recovery=1
...
[drm] amdgpu kernel modesetting enabled.
amdgpu :01:00.0: vgaarb: deactivate vga console
Console: switching to colour dummy device 80x25
[drm] initializing kernel modesetting (VERDE 0x1002:0x682B
0x1458:0x22CA 0x87).
[drm] register mmio base: 0xFE8C
[drm] register mmio size: 262144
[drm] add ip block number 0
[drm] add ip block number 1
[drm] add ip block number 2
[drm] add ip block number 3
[drm] add ip block number 4
[drm] add ip block number 5
[drm] add ip block number 6
[drm] add ip block number 7
[drm] BIOS signature incorrect 5b 7
resource sanity check: requesting [mem 0x000c-0x000d],
which spans more than PCI Bus :00 [mem 0x000d-0x000d
window]
caller pci_map_rom+0x68/0x1b0 mapping multiple BARs
amdgpu :01:00.0: No more image in the PCI ROM
amdgpu :01:00.0: amdgpu: Fetched VBIOS from ROM BAR
amdgpu: ATOM BIOS: xxx-xxx-xxx
amdgpu :01:00.0: amdgpu: Trusted Memory Zone (TMZ) feature
not supported
amdgpu :01:00.0: amdgpu: PCIE atomic ops is not supported
[drm] PCIE gen 2 link speeds already enabled
[drm] vm size is 64 GB, 2 levels, block size is 10-bit, fragment
size is 9-bit
RTL8211B Gigabit Ethernet r8169-0-300:00: attached PHY driver
(mii_bus:phy_addr=r8169-0-300:00, irq=MAC)
r8169 :03:00.0 eth0: Link is Down
amdgpu :01:00.0: amdgpu: VRAM: 2048M 0x00F4 -
0x00F47FFF (2048M used)
amdgpu :01:00.0: amdgpu: GART: 1024M 0x00FF -
0x00FF3FFF
[drm] Detected VRAM RAM=2048M, BAR=256M
[drm] RAM width 128bits DDR3
[drm] amdgpu: 2048M of VRAM memory ready
[drm] amdgpu: 3979M of GTT memory ready.
[drm] GART: num cpu pages 262144, num gpu pages 262144
amdgpu :01:00.0: amdgpu: PCIE GART of 1024M enabled (table at
0x00F400A0).
[drm] Internal thermal controller with fan control
[drm] amdgpu: dpm initialized
[drm] AMDGPU Display Connectors
[drm] Connector 0:
[drm] HDMI-A-1
[drm] HPD1
[drm] DDC: 0x194c 0x194c 0x194d 0x194d 0x194e 0x194e 0x194f 0x194f
[drm] Encoders:
[drm] DFP1: INTERNAL_UNIPHY
[drm] Connector 1:
[drm] DVI-D-1
[drm] HPD2
[drm] DDC: 0x1950 0x1950 0x1951 0x1951 0x1952 0x1952 0x1953 0x1953
[drm] Encoders:
[drm] DFP2: INTERNAL_UNIPHY
[drm] Connector 2:
[drm] VGA-1
[drm] DDC: 0x1970 0x1970 0x1971 0x1971 0x1972 0x1972 0x1973 0x1973
[drm] Encoders:
[drm] CRT1: INTERNAL_KLDSCP_DAC1
[drm] Found UVD firmware Version: 64.0 Family ID: 13
amdgpu: Move buffer fallback to memcpy unavailable
[drm:amdgpu_device_init.cold [amdgpu]] *ERROR* sw_init of IP
block failed -19
amdgpu :01:00.0: amdgpu: amdgpu_device_ip_init failed
amdgpu :01:00.0: amdgpu: Fatal error during GPU init
Re: [git pull] drm fixes for 6.1-rc1
Arun please take a look into this ASAP.
Thanks,
Christian.
Am 17.10.22 um 03:13 schrieb Arthur Marsh:
Thanks Dave, I reverted patch 312b4dc11d4f74bfe03ea25ffe04c1f2fdd13cb9 against
6.1-rc1 and the resulting kernel loaded amdgpu fine on my pc with Cape Verde
GPU.
Regards,
Arthur.
On 17 October 2022 8:14:18 am ACDT, Dave Airlie wrote:
On Sun, 16 Oct 2022 at 18:09, Arthur Marsh
wrote:
From: Arthur Marsh
Hi, the "drm fixes for 6.1-rc1" commit caused the amdgpu module to fail
with my Cape Verde radeonsi card.
I haven't been able to bisect the problem to an individual commit, but
attach a dmesg extract below.
I'm happy to supply any other configuration information and test patches.
Can you try reverting: it's the only think I can spot that might
affect a card that old since most changes in that request were for
display hw you don't have.
ommit 312b4dc11d4f74bfe03ea25ffe04c1f2fdd13cb9
Author: Arunpravin Paneer Selvam
Date: Tue Oct 4 07:33:39 2022 -0700
drm/amdgpu: Fix VRAM BO swap issue
DRM buddy manager allocates the contiguous memory requests in
a single block or multiple blocks. So for the ttm move operation
(incase of low vram memory) we should consider all the blocks to
compute the total memory size which compared with the struct
ttm_resource num_pages in order to verify that the blocks are
contiguous for the eviction process.
v2: Added a Fixes tag
v3: Rewrite the code to save a bit of calculations and
variables (Christian)
Fixes: c9cad937c0c5 ("drm/amdgpu: add drm buddy support to amdgpu")
Signed-off-by: Arunpravin Paneer Selvam
Reviewed-by: Christian König
Signed-off-by: Alex Deucher
Thanks,
Dave.
Arthur.
Linux version 6.0.0+ (root@am64) (gcc-12 (Debian 12.2.0-5) 12.2.0, GNU ld
(GNU Binutils for Debian) 2.39) #5179 SMP PREEMPT_DYNAMIC Fri Oct 14 17:00:40
ACDT 2022
Command line: BOOT_IMAGE=/vmlinuz-6.0.0+
root=UUID=39706f53-7c27-4310-b22a-36c7b042d1a1 ro single amdgpu.audio=1
amdgpu.si_support=1 radeon.si_support=0 page_owner=on amdgpu.gpu_recovery=1
...
[drm] amdgpu kernel modesetting enabled.
amdgpu :01:00.0: vgaarb: deactivate vga console
Console: switching to colour dummy device 80x25
[drm] initializing kernel modesetting (VERDE 0x1002:0x682B 0x1458:0x22CA
0x87).
[drm] register mmio base: 0xFE8C
[drm] register mmio size: 262144
[drm] add ip block number 0
[drm] add ip block number 1
[drm] add ip block number 2
[drm] add ip block number 3
[drm] add ip block number 4
[drm] add ip block number 5
[drm] add ip block number 6
[drm] add ip block number 7
[drm] BIOS signature incorrect 5b 7
resource sanity check: requesting [mem 0x000c-0x000d], which spans
more than PCI Bus :00 [mem 0x000d-0x000d window]
caller pci_map_rom+0x68/0x1b0 mapping multiple BARs
amdgpu :01:00.0: No more image in the PCI ROM
amdgpu :01:00.0: amdgpu: Fetched VBIOS from ROM BAR
amdgpu: ATOM BIOS: xxx-xxx-xxx
amdgpu :01:00.0: amdgpu: Trusted Memory Zone (TMZ) feature not supported
amdgpu :01:00.0: amdgpu: PCIE atomic ops is not supported
[drm] PCIE gen 2 link speeds already enabled
[drm] vm size is 64 GB, 2 levels, block size is 10-bit, fragment size is 9-bit
RTL8211B Gigabit Ethernet r8169-0-300:00: attached PHY driver
(mii_bus:phy_addr=r8169-0-300:00, irq=MAC)
r8169 :03:00.0 eth0: Link is Down
amdgpu :01:00.0: amdgpu: VRAM: 2048M 0x00F4 -
0x00F47FFF (2048M used)
amdgpu :01:00.0: amdgpu: GART: 1024M 0x00FF -
0x00FF3FFF
[drm] Detected VRAM RAM=2048M, BAR=256M
[drm] RAM width 128bits DDR3
[drm] amdgpu: 2048M of VRAM memory ready
[drm] amdgpu: 3979M of GTT memory ready.
[drm] GART: num cpu pages 262144, num gpu pages 262144
amdgpu :01:00.0: amdgpu: PCIE GART of 1024M enabled (table at
0x00F400A0).
[drm] Internal thermal controller with fan control
[drm] amdgpu: dpm initialized
[drm] AMDGPU Display Connectors
[drm] Connector 0:
[drm] HDMI-A-1
[drm] HPD1
[drm] DDC: 0x194c 0x194c 0x194d 0x194d 0x194e 0x194e 0x194f 0x194f
[drm] Encoders:
[drm] DFP1: INTERNAL_UNIPHY
[drm] Connector 1:
[drm] DVI-D-1
[drm] HPD2
[drm] DDC: 0x1950 0x1950 0x1951 0x1951 0x1952 0x1952 0x1953 0x1953
[drm] Encoders:
[drm] DFP2: INTERNAL_UNIPHY
[drm] Connector 2:
[drm] VGA-1
[drm] DDC: 0x1970 0x1970 0x1971 0x1971 0x1972 0x1972 0x1973 0x1973
[drm] Encoders:
[drm] CRT1: INTERNAL_KLDSCP_DAC1
[drm] Found UVD firmware Version: 64.0 Family ID: 13
amdgpu: Move buffer fallback to memcpy unavailable
[drm:amdgpu_device_init.cold [amdgpu]] *ERROR* sw_init of IP block
failed -19
amdgpu :01:00.0: amdgpu: amdgpu_device_ip_init failed
amdgpu :01:00.0: amdgpu: Fatal error during GPU init
amdgpu :01:00.0: amdgpu: amdgpu: finishing device.
BUG: kernel NULL pointer dereference, address: 009
Re: [git pull] drm fixes for 6.1-rc1
Thanks Dave, I reverted patch 312b4dc11d4f74bfe03ea25ffe04c1f2fdd13cb9 against
6.1-rc1 and the resulting kernel loaded amdgpu fine on my pc with Cape Verde
GPU.
Regards,
Arthur.
On 17 October 2022 8:14:18 am ACDT, Dave Airlie wrote:
>On Sun, 16 Oct 2022 at 18:09, Arthur Marsh
> wrote:
>>
>> From: Arthur Marsh
>>
>> Hi, the "drm fixes for 6.1-rc1" commit caused the amdgpu module to fail
>> with my Cape Verde radeonsi card.
>>
>> I haven't been able to bisect the problem to an individual commit, but
>> attach a dmesg extract below.
>>
>> I'm happy to supply any other configuration information and test patches.
>>
>
>Can you try reverting: it's the only think I can spot that might
>affect a card that old since most changes in that request were for
>display hw you don't have.
>
>ommit 312b4dc11d4f74bfe03ea25ffe04c1f2fdd13cb9
>Author: Arunpravin Paneer Selvam
>Date: Tue Oct 4 07:33:39 2022 -0700
>
>drm/amdgpu: Fix VRAM BO swap issue
>
>DRM buddy manager allocates the contiguous memory requests in
>a single block or multiple blocks. So for the ttm move operation
>(incase of low vram memory) we should consider all the blocks to
>compute the total memory size which compared with the struct
>ttm_resource num_pages in order to verify that the blocks are
>contiguous for the eviction process.
>
>v2: Added a Fixes tag
>v3: Rewrite the code to save a bit of calculations and
>variables (Christian)
>
>Fixes: c9cad937c0c5 ("drm/amdgpu: add drm buddy support to amdgpu")
>Signed-off-by: Arunpravin Paneer Selvam
>Reviewed-by: Christian König
>Signed-off-by: Alex Deucher
>
>
>Thanks,
>Dave.
>
>> Arthur.
>>
>> Linux version 6.0.0+ (root@am64) (gcc-12 (Debian 12.2.0-5) 12.2.0, GNU ld
>> (GNU Binutils for Debian) 2.39) #5179 SMP PREEMPT_DYNAMIC Fri Oct 14
>> 17:00:40 ACDT 2022
>> Command line: BOOT_IMAGE=/vmlinuz-6.0.0+
>> root=UUID=39706f53-7c27-4310-b22a-36c7b042d1a1 ro single amdgpu.audio=1
>> amdgpu.si_support=1 radeon.si_support=0 page_owner=on amdgpu.gpu_recovery=1
>> ...
>>
>> [drm] amdgpu kernel modesetting enabled.
>> amdgpu :01:00.0: vgaarb: deactivate vga console
>> Console: switching to colour dummy device 80x25
>> [drm] initializing kernel modesetting (VERDE 0x1002:0x682B 0x1458:0x22CA
>> 0x87).
>> [drm] register mmio base: 0xFE8C
>> [drm] register mmio size: 262144
>> [drm] add ip block number 0
>> [drm] add ip block number 1
>> [drm] add ip block number 2
>> [drm] add ip block number 3
>> [drm] add ip block number 4
>> [drm] add ip block number 5
>> [drm] add ip block number 6
>> [drm] add ip block number 7
>> [drm] BIOS signature incorrect 5b 7
>> resource sanity check: requesting [mem 0x000c-0x000d], which spans
>> more than PCI Bus :00 [mem 0x000d-0x000d window]
>> caller pci_map_rom+0x68/0x1b0 mapping multiple BARs
>> amdgpu :01:00.0: No more image in the PCI ROM
>> amdgpu :01:00.0: amdgpu: Fetched VBIOS from ROM BAR
>> amdgpu: ATOM BIOS: xxx-xxx-xxx
>> amdgpu :01:00.0: amdgpu: Trusted Memory Zone (TMZ) feature not supported
>> amdgpu :01:00.0: amdgpu: PCIE atomic ops is not supported
>> [drm] PCIE gen 2 link speeds already enabled
>> [drm] vm size is 64 GB, 2 levels, block size is 10-bit, fragment size is
>> 9-bit
>> RTL8211B Gigabit Ethernet r8169-0-300:00: attached PHY driver
>> (mii_bus:phy_addr=r8169-0-300:00, irq=MAC)
>> r8169 :03:00.0 eth0: Link is Down
>> amdgpu :01:00.0: amdgpu: VRAM: 2048M 0x00F4 -
>> 0x00F47FFF (2048M used)
>> amdgpu :01:00.0: amdgpu: GART: 1024M 0x00FF -
>> 0x00FF3FFF
>> [drm] Detected VRAM RAM=2048M, BAR=256M
>> [drm] RAM width 128bits DDR3
>> [drm] amdgpu: 2048M of VRAM memory ready
>> [drm] amdgpu: 3979M of GTT memory ready.
>> [drm] GART: num cpu pages 262144, num gpu pages 262144
>> amdgpu :01:00.0: amdgpu: PCIE GART of 1024M enabled (table at
>> 0x00F400A0).
>> [drm] Internal thermal controller with fan control
>> [drm] amdgpu: dpm initialized
>> [drm] AMDGPU Display Connectors
>> [drm] Connector 0:
>> [drm] HDMI-A-1
>> [drm] HPD1
>> [drm] DDC: 0x194c 0x194c 0x194d 0x194d 0x194e 0x194e 0x194f 0x194f
>> [drm] Encoders:
>> [drm] DFP1: INTERNAL_UNIPHY
>> [drm] Connector 1:
>> [drm] DVI-D-1
>> [drm] HPD2
>> [drm] DDC: 0x1950 0x1950 0x1951 0x1951 0x1952 0x1952 0x1953 0x1953
>> [drm] Encoders:
>> [drm] DFP2: INTERNAL_UNIPHY
>> [drm] Connector 2:
>> [drm] VGA-1
>> [drm] DDC: 0x1970 0x1970 0x1971 0x1971 0x1972 0x1972 0x1973 0x1973
>> [drm] Encoders:
>> [drm] CRT1: INTERNAL_KLDSCP_DAC1
>> [drm] Found UVD firmware Version: 64.0 Family ID: 13
>> amdgpu: Move buffer fallback to memcpy unavailable
>> [drm:amdgpu_device_init.cold [amdgpu]] *ERROR* sw_init of IP block
>> failed -19
>> amdgpu :01:00.0: amdgpu: amdgpu_device_ip_init failed
>> amdgpu :01:00.0: amdgpu: Fatal error during GPU ini
Re: [git pull] drm fixes for 6.1-rc1
On Sun, 16 Oct 2022 at 18:09, Arthur Marsh
wrote:
>
> From: Arthur Marsh
>
> Hi, the "drm fixes for 6.1-rc1" commit caused the amdgpu module to fail
> with my Cape Verde radeonsi card.
>
> I haven't been able to bisect the problem to an individual commit, but
> attach a dmesg extract below.
>
> I'm happy to supply any other configuration information and test patches.
>
Can you try reverting: it's the only think I can spot that might
affect a card that old since most changes in that request were for
display hw you don't have.
ommit 312b4dc11d4f74bfe03ea25ffe04c1f2fdd13cb9
Author: Arunpravin Paneer Selvam
Date: Tue Oct 4 07:33:39 2022 -0700
drm/amdgpu: Fix VRAM BO swap issue
DRM buddy manager allocates the contiguous memory requests in
a single block or multiple blocks. So for the ttm move operation
(incase of low vram memory) we should consider all the blocks to
compute the total memory size which compared with the struct
ttm_resource num_pages in order to verify that the blocks are
contiguous for the eviction process.
v2: Added a Fixes tag
v3: Rewrite the code to save a bit of calculations and
variables (Christian)
Fixes: c9cad937c0c5 ("drm/amdgpu: add drm buddy support to amdgpu")
Signed-off-by: Arunpravin Paneer Selvam
Reviewed-by: Christian König
Signed-off-by: Alex Deucher
Thanks,
Dave.
> Arthur.
>
> Linux version 6.0.0+ (root@am64) (gcc-12 (Debian 12.2.0-5) 12.2.0, GNU ld
> (GNU Binutils for Debian) 2.39) #5179 SMP PREEMPT_DYNAMIC Fri Oct 14 17:00:40
> ACDT 2022
> Command line: BOOT_IMAGE=/vmlinuz-6.0.0+
> root=UUID=39706f53-7c27-4310-b22a-36c7b042d1a1 ro single amdgpu.audio=1
> amdgpu.si_support=1 radeon.si_support=0 page_owner=on amdgpu.gpu_recovery=1
> ...
>
> [drm] amdgpu kernel modesetting enabled.
> amdgpu :01:00.0: vgaarb: deactivate vga console
> Console: switching to colour dummy device 80x25
> [drm] initializing kernel modesetting (VERDE 0x1002:0x682B 0x1458:0x22CA
> 0x87).
> [drm] register mmio base: 0xFE8C
> [drm] register mmio size: 262144
> [drm] add ip block number 0
> [drm] add ip block number 1
> [drm] add ip block number 2
> [drm] add ip block number 3
> [drm] add ip block number 4
> [drm] add ip block number 5
> [drm] add ip block number 6
> [drm] add ip block number 7
> [drm] BIOS signature incorrect 5b 7
> resource sanity check: requesting [mem 0x000c-0x000d], which spans
> more than PCI Bus :00 [mem 0x000d-0x000d window]
> caller pci_map_rom+0x68/0x1b0 mapping multiple BARs
> amdgpu :01:00.0: No more image in the PCI ROM
> amdgpu :01:00.0: amdgpu: Fetched VBIOS from ROM BAR
> amdgpu: ATOM BIOS: xxx-xxx-xxx
> amdgpu :01:00.0: amdgpu: Trusted Memory Zone (TMZ) feature not supported
> amdgpu :01:00.0: amdgpu: PCIE atomic ops is not supported
> [drm] PCIE gen 2 link speeds already enabled
> [drm] vm size is 64 GB, 2 levels, block size is 10-bit, fragment size is
> 9-bit
> RTL8211B Gigabit Ethernet r8169-0-300:00: attached PHY driver
> (mii_bus:phy_addr=r8169-0-300:00, irq=MAC)
> r8169 :03:00.0 eth0: Link is Down
> amdgpu :01:00.0: amdgpu: VRAM: 2048M 0x00F4 -
> 0x00F47FFF (2048M used)
> amdgpu :01:00.0: amdgpu: GART: 1024M 0x00FF -
> 0x00FF3FFF
> [drm] Detected VRAM RAM=2048M, BAR=256M
> [drm] RAM width 128bits DDR3
> [drm] amdgpu: 2048M of VRAM memory ready
> [drm] amdgpu: 3979M of GTT memory ready.
> [drm] GART: num cpu pages 262144, num gpu pages 262144
> amdgpu :01:00.0: amdgpu: PCIE GART of 1024M enabled (table at
> 0x00F400A0).
> [drm] Internal thermal controller with fan control
> [drm] amdgpu: dpm initialized
> [drm] AMDGPU Display Connectors
> [drm] Connector 0:
> [drm] HDMI-A-1
> [drm] HPD1
> [drm] DDC: 0x194c 0x194c 0x194d 0x194d 0x194e 0x194e 0x194f 0x194f
> [drm] Encoders:
> [drm] DFP1: INTERNAL_UNIPHY
> [drm] Connector 1:
> [drm] DVI-D-1
> [drm] HPD2
> [drm] DDC: 0x1950 0x1950 0x1951 0x1951 0x1952 0x1952 0x1953 0x1953
> [drm] Encoders:
> [drm] DFP2: INTERNAL_UNIPHY
> [drm] Connector 2:
> [drm] VGA-1
> [drm] DDC: 0x1970 0x1970 0x1971 0x1971 0x1972 0x1972 0x1973 0x1973
> [drm] Encoders:
> [drm] CRT1: INTERNAL_KLDSCP_DAC1
> [drm] Found UVD firmware Version: 64.0 Family ID: 13
> amdgpu: Move buffer fallback to memcpy unavailable
> [drm:amdgpu_device_init.cold [amdgpu]] *ERROR* sw_init of IP block
> failed -19
> amdgpu :01:00.0: amdgpu: amdgpu_device_ip_init failed
> amdgpu :01:00.0: amdgpu: Fatal error during GPU init
> amdgpu :01:00.0: amdgpu: amdgpu: finishing device.
> BUG: kernel NULL pointer dereference, address: 0090
> #PF: supervisor write access in kernel mode
> #PF: error_code(0x0002) - not-present page
> PGD 0 P4D 0
> Oops: 0002 [#1] PREEMPT SMP NOPTI
> CPU: 3 PID: 447 Comm: udevd Not tainted 6.0.0+ #5179
> Hardware name: System manufact
Re: [git pull] drm fixes for 6.1-rc1
From: Arthur Marsh Hi, the "drm fixes for 6.1-rc1" commit caused the amdgpu module to fail with my Cape Verde radeonsi card. I haven't been able to bisect the problem to an individual commit, but attach a dmesg extract below. I'm happy to supply any other configuration information and test patches. Arthur. Linux version 6.0.0+ (root@am64) (gcc-12 (Debian 12.2.0-5) 12.2.0, GNU ld (GNU Binutils for Debian) 2.39) #5179 SMP PREEMPT_DYNAMIC Fri Oct 14 17:00:40 ACDT 2022 Command line: BOOT_IMAGE=/vmlinuz-6.0.0+ root=UUID=39706f53-7c27-4310-b22a-36c7b042d1a1 ro single amdgpu.audio=1 amdgpu.si_support=1 radeon.si_support=0 page_owner=on amdgpu.gpu_recovery=1 ... [drm] amdgpu kernel modesetting enabled. amdgpu :01:00.0: vgaarb: deactivate vga console Console: switching to colour dummy device 80x25 [drm] initializing kernel modesetting (VERDE 0x1002:0x682B 0x1458:0x22CA 0x87). [drm] register mmio base: 0xFE8C [drm] register mmio size: 262144 [drm] add ip block number 0 [drm] add ip block number 1 [drm] add ip block number 2 [drm] add ip block number 3 [drm] add ip block number 4 [drm] add ip block number 5 [drm] add ip block number 6 [drm] add ip block number 7 [drm] BIOS signature incorrect 5b 7 resource sanity check: requesting [mem 0x000c-0x000d], which spans more than PCI Bus :00 [mem 0x000d-0x000d window] caller pci_map_rom+0x68/0x1b0 mapping multiple BARs amdgpu :01:00.0: No more image in the PCI ROM amdgpu :01:00.0: amdgpu: Fetched VBIOS from ROM BAR amdgpu: ATOM BIOS: xxx-xxx-xxx amdgpu :01:00.0: amdgpu: Trusted Memory Zone (TMZ) feature not supported amdgpu :01:00.0: amdgpu: PCIE atomic ops is not supported [drm] PCIE gen 2 link speeds already enabled [drm] vm size is 64 GB, 2 levels, block size is 10-bit, fragment size is 9-bit RTL8211B Gigabit Ethernet r8169-0-300:00: attached PHY driver (mii_bus:phy_addr=r8169-0-300:00, irq=MAC) r8169 :03:00.0 eth0: Link is Down amdgpu :01:00.0: amdgpu: VRAM: 2048M 0x00F4 - 0x00F47FFF (2048M used) amdgpu :01:00.0: amdgpu: GART: 1024M 0x00FF - 0x00FF3FFF [drm] Detected VRAM RAM=2048M, BAR=256M [drm] RAM width 128bits DDR3 [drm] amdgpu: 2048M of VRAM memory ready [drm] amdgpu: 3979M of GTT memory ready. [drm] GART: num cpu pages 262144, num gpu pages 262144 amdgpu :01:00.0: amdgpu: PCIE GART of 1024M enabled (table at 0x00F400A0). [drm] Internal thermal controller with fan control [drm] amdgpu: dpm initialized [drm] AMDGPU Display Connectors [drm] Connector 0: [drm] HDMI-A-1 [drm] HPD1 [drm] DDC: 0x194c 0x194c 0x194d 0x194d 0x194e 0x194e 0x194f 0x194f [drm] Encoders: [drm] DFP1: INTERNAL_UNIPHY [drm] Connector 1: [drm] DVI-D-1 [drm] HPD2 [drm] DDC: 0x1950 0x1950 0x1951 0x1951 0x1952 0x1952 0x1953 0x1953 [drm] Encoders: [drm] DFP2: INTERNAL_UNIPHY [drm] Connector 2: [drm] VGA-1 [drm] DDC: 0x1970 0x1970 0x1971 0x1971 0x1972 0x1972 0x1973 0x1973 [drm] Encoders: [drm] CRT1: INTERNAL_KLDSCP_DAC1 [drm] Found UVD firmware Version: 64.0 Family ID: 13 amdgpu: Move buffer fallback to memcpy unavailable [drm:amdgpu_device_init.cold [amdgpu]] *ERROR* sw_init of IP block failed -19 amdgpu :01:00.0: amdgpu: amdgpu_device_ip_init failed amdgpu :01:00.0: amdgpu: Fatal error during GPU init amdgpu :01:00.0: amdgpu: amdgpu: finishing device. BUG: kernel NULL pointer dereference, address: 0090 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 0 P4D 0 Oops: 0002 [#1] PREEMPT SMP NOPTI CPU: 3 PID: 447 Comm: udevd Not tainted 6.0.0+ #5179 Hardware name: System manufacturer System Product Name/M3A78 PRO, BIOS 1701 01/27/2011 RIP: 0010:drm_sched_fini+0x80/0xa0 [gpu_sched] Code: 76 83 0e c4 c6 85 8c 01 00 00 00 5b 5d 41 5c 41 5d c3 cc cc cc cc 4c 8d 63 f0 4c 89 e7 e8 08 99 8e c4 48 8b 03 48 39 d8 74 0f 80 90 00 00 00 01 48 8b 00 48 39 d8 75 f1 4c 89 e7 e8 c9 99 8e RSP: 0018:beb3c06bfbb8 EFLAGS: 00010213 RAX: RBX: 99bae8269a98 RCX: 99bab703afc0 RDX: 0001 RSI: 99bab703afe8 RDI: RBP: 99bae82699f0 R08: 85cd0bc2 R09: 0010 R10: 0035 R11: 99bb594806c0 R12: 99bae8269a88 R13: 99bae82699f8 R14: 99bae82665e8 R15: FS: 7fd81fcd9840() GS:99bb67cc() knlGS: CS: 0010 DS: ES: CR0: 80050033 CR2: 0090 CR3: 000111822000 CR4: 06e0 Call Trace: amdgpu_fence_driver_sw_fini+0xc2/0xd0 [amdgpu] amdgpu_device_fini_sw+0x17/0x3c0 [amdgpu] amdgpu_driver_release_kms+0x12/0x30 [amdgpu] devm_drm_dev_init_release+0x4a/0x70 [drm] release_nodes+0x40/0xb0 devres_release_all+0x89/0xc0 device_unbind_cleanup+0xe/0x70 really_probe+0x245/0x3a0 ? pm_runtime_barrier+0x61/0xb0 __driver_probe_de
Re: [git pull] drm fixes for 6.1-rc1
The pull request you sent on Fri, 14 Oct 2022 10:29:19 +1000: > git://anongit.freedesktop.org/drm/drm tags/drm-next-2022-10-14 has been merged into torvalds/linux.git: https://git.kernel.org/torvalds/c/9c9155a3509a2ebdb06d77c7a621e9685c802eac Thank you! -- Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/prtracker.html
Re: [git pull] drm fixes for 6.1-rc1
On Thu, Oct 13, 2022 at 5:29 PM Dave Airlie wrote: > > Round of fixes for the merge window stuff, bunch of amdgpu and i915 > changes, this should have the gcc11 warning fix, amongst other > changes. Some of those amd changes aren't "fixes". They are some major code changes. We're still in the merge window, so I'm letting it slide, but calling then "fixes" really stretches things. They are fixes exactly the same way completely new development can "fix" things. Linus
[git pull] drm fixes for 6.1-rc1
Hi Linus, Round of fixes for the merge window stuff, bunch of amdgpu and i915 changes, this should have the gcc11 warning fix, amongst other changes. Dave. drm-next-2022-10-14: drm fixes for 6.1-rc1 amdgpu: - DC mutex fix - DC SubVP fixes - DCN 3.2.x fixes - DCN 3.1.x fixes - SDMA 6.x fixes - Enable DPIA for 3.1.4 - VRR fixes - VRAM BO swapping fix - Revert dirty fb helper change - SR-IOV suspend/resume fixes - Work around GCC array bounds check fail warning - UMC 8.10 fixes - Misc fixes and cleanups i915: - Round to closest in g4x+ HDMI clock readout - Update MOCS table for EHL - Fix PSR_IMR/IIR field handling - Fix watermark calculations for gen12+/DG2 modifiers - Reject excessive dotclocks early - Fix revocation of non-persistent contexts - Handle migration for dpt - Fix display problems after resume - Allow control over the flags when migrating - Consider DG2_RC_CCS_CC when migrating buffers The following changes since commit bafaf67c42f4b547bf4fb329ac6dcb28b05de15e: Revert "drm/sched: Use parent fence instead of finished" (2022-10-07 12:58:39 +1000) are available in the Git repository at: git://anongit.freedesktop.org/drm/drm tags/drm-next-2022-10-14 for you to fetch changes up to fc3523a833c9c109e68209f1ecdd15864373e66a: Merge tag 'amd-drm-fixes-6.1-2022-10-12' of https://gitlab.freedesktop.org/agd5f/linux into drm-next (2022-10-14 07:47:25 +1000) drm fixes for 6.1-rc1 amdgpu: - DC mutex fix - DC SubVP fixes - DCN 3.2.x fixes - DCN 3.1.x fixes - SDMA 6.x fixes - Enable DPIA for 3.1.4 - VRR fixes - VRAM BO swapping fix - Revert dirty fb helper change - SR-IOV suspend/resume fixes - Work around GCC array bounds check fail warning - UMC 8.10 fixes - Misc fixes and cleanups i915: - Round to closest in g4x+ HDMI clock readout - Update MOCS table for EHL - Fix PSR_IMR/IIR field handling - Fix watermark calculations for gen12+/DG2 modifiers - Reject excessive dotclocks early - Fix revocation of non-persistent contexts - Handle migration for dpt - Fix display problems after resume - Allow control over the flags when migrating - Consider DG2_RC_CCS_CC when migrating buffers Alex Deucher (7): drm/amdgpu: switch sdma buffer function tear down to a helper drm/amdgpu: fix SDMA suspend/resume on SR-IOV drm/amd/display: make dcn32_split_stream_for_mpc_or_odm static drm/amd/display: fix indentation in dc.c drm/amd/display: make virtual_disable_link_output static drm/amd/display: add a license to cursor_reg_cache.h drm/amd/display: fix transfer function passed to build_coefficients() Alexey Kodanev (2): drm/amd/pm: vega10_hwmgr: fix potential off-by-one overflow in 'performance_levels' drm/amd/pm: smu7_hwmgr: fix potential off-by-one overflow in 'performance_levels' Alvin Lee (5): drm/amd/display: Only commit SubVP state after pipe programming drm/amd/display: Block SubVP if rotation being used drm/amd/display: Disable GSL when enabling phantom pipe drm/amd/display: For SubVP pipe split case use min transition into MPO drm/amd/display: Fix watermark calculation Aric Cyr (4): Revert "drm/amd/display: correct hostvm flag" drm/amd/display: Fix vupdate and vline position calculation drm/amd/display: 3.2.206 drm/amd/display: 3.2.207 Arunpravin Paneer Selvam (1): drm/amdgpu: Fix VRAM BO swap issue Aurabindo Pillai (2): drm/amd/display: Do not trigger timing sync for phantom pipes drm/amd/display: Add HUBP surface flip interrupt handler Bokun Zhang (1): drm/amdgpu: Fix SDMA engine resume issue under SRIOV Candice Li (2): drm/amdgpu: Update umc v8_10_0 headers drm/amdgpu: Add poison mode query for umc v8_10_0 Charlene Liu (1): drm/amd/display: prevent S4 test from failing Daniel Gomez (1): drm/amd/display: Fix mutex lock in dcn10 Dave Airlie (3): Merge tag 'drm-intel-next-fixes-2022-10-06-1' of git://anongit.freedesktop.org/drm/drm-intel into drm-next Merge tag 'drm-intel-next-fixes-2022-10-13' of git://anongit.freedesktop.org/drm/drm-intel into drm-next Merge tag 'amd-drm-fixes-6.1-2022-10-12' of https://gitlab.freedesktop.org/agd5f/linux into drm-next Dillon Varone (8): drm/amd/display: Program SubVP in dc_commit_state_no_check drm/amd/display: Reorder FCLK P-state switch sequence for DCN32 drm/amd/display: Increase compbuf size prior to updating clocks drm/amd/display: Fix merging dynamic ODM+MPO configs on DCN32 Revert "drm/amd/display: skip commit minimal transition state" drm/amd/display: Use correct pixel clock to program DTBCLK DTO's drm/amd/display: Acquire FCLK DPM levels on DCN32 drm/amd/display: Fix bug preventing FCLK Pstate allow message being sent Dmytro Laktyushkin (3): drm/amd/display: fix dcn315 dml detile overestimation
