amd-staging-4.6 for Fedora 24 (AMDGPU)

[Mystro256]

Nope Kaveri APU is CI and another offloading GPU is SI... there is no SI APU, etc...

Are you sure Radeon R7 M265 is the right model name? I was under the impression that was GCN 1st gen only. I'm not exactly familiar with the APU's to be honest.

Either way, finalzone, was the bug introduced with 4.7.2-902 or does it exist for 4.7.2-901?
Also what version of linux-firmware are you using?
I've notices some issues with the CIK stuff and I want to see if I could debug it or at least report it to the developers.

[dungeon]

Are you sure Radeon R7 M265 is the right model name? I was under the impression that was GCN 1st gen only. I'm not exactly familiar with the APU's to be honest.

That is AMD's Dual Graphics, so DX branding which has different branding from alone....

So right model, but point is that he missed to mention that DX for APUs GPU so that is incorrect - and that is catch 22 So both GPUs together there are branded as R7 M265DX

Here you can see it (tab Notebook), if he has R5 M230 offload GPU and if whole thing is brended as R7 M265DX then APU is A10-7400P

http://www.amd.com/en-us/innovations/software-technologies/dual-graphics

Anyway APU has CI GPU for sure, as there was no any SI APU in existence

[bridgman]

Anyway APU has CI GPU for sure, as there was no any SI APU in existence

Right... Trinity/Richland was VLIW4, then we jumped straight to CI for Kaveri.

[dungeon]

Well even whole year before Kaveri appeared cats APUs Temash and Kabini was GCN 1.1 too... so GCN 1.0 (SI) APU never heppened

[Mystro256]

That is AMD's Dual Graphics, so DX branding which has different branding from alone....

So right model, but point is that he missed to mention that DX for APUs GPU so that is incorrect - and that is catch 22 So both GPUs together there are branded as R7 M265DX

Here you can see it (tab Notebook), if he has R5 M230 offload GPU and if whole thing is brended as R7 M265DX then APU is A10-7400P

http://www.amd.com/en-us/innovations/software-technologies/dual-graphics

Anyway APU has CI GPU for sure, as there was no any SI APU in existence

Neat, I'll have to read up on it.

[finalzone]

Either way, finalzone, was the bug introduced with 4.7.2-902 or does it exist for 4.7.2-901?
Also what version of linux-firmware are you using?
I've notices some issues with the CIK stuff and I want to see if I could debug it or at least report it to the developers.

Linux-firmware is 20160816 67.git7c3dfc0b.fc24
Lockup on radeon driver happened

radeon 0000:00:01.0: GPU lockup (current fence id 0x0000000000000003 last fence id 0x0000000000000004 on ring 0)
Aug 29 16:51:37  kernel: radeon 0000:00:01.0: ring 0 stalled for more than 10581msec

It is also happening on kernel 4.8.0 suggesting radeon driver bug.

[Mystro256]

Linux-firmware is 20160816 67.git7c3dfc0b.fc24
Lockup on radeon driver happened

radeon 0000:00:01.0: GPU lockup (current fence id 0x0000000000000003 last fence id 0x0000000000000004 on ring 0)
Aug 29 16:51:37 kernel: radeon 0000:00:01.0: ring 0 stalled for more than 10581msec

It is also happening on kernel 4.8.0 suggesting radeon driver bug.

Can you reproduce this with linux-firmware-20160609-66.gita4bbc811?

I have a feeling there was an update to the firmware that's not playing nicely with the latest driver.

[finalzone]

Can you reproduce this with linux-firmware-20160609-66.gita4bbc811?

I have a feeling there was an update to the firmware that's not playing nicely with the latest driver.

It looks like your suspicion is correct. Installing linux-firmware-20160609-66.gita4bbc811 resolved the issue. The latest kernel from amd-staged repository is working fine now.

$ uname -r
4.7.2-902.amd.25082016.fc24.x86_64

$ rpm -q linux-firmware
linux-firmware-20160609-66.gita4bbc811.fc24.noarch

Here is resulted boot

$ dmesg | grep amdgpu
[    0.000000] Command line: BOOT_IMAGE=/vmlinuz-4.7.2-902.amd.25082016.fc24.x86_64 root=UUID=232c0ccf-375a-4e4d-a080-9986ea920e1c ro rhgb quiet LANG=en_CA.UTF-8 modprobe.blacklist=sp5100_tco amdgpu.exp_hw_support=1 amdgpu.powerplay=1 modprobe.blacklist=radeon acpi.trace_state=enable "dyndbg=file ec.c +p"
[    0.000000] Kernel command line: BOOT_IMAGE=/vmlinuz-4.7.2-902.amd.25082016.fc24.x86_64 root=UUID=232c0ccf-375a-4e4d-a080-9986ea920e1c ro rhgb quiet LANG=en_CA.UTF-8 modprobe.blacklist=sp5100_tco amdgpu.exp_hw_support=1 amdgpu.powerplay=1 modprobe.blacklist=radeon acpi.trace_state=enable "dyndbg=file ec.c +p"
[    2.067194] [drm] amdgpu kernel modesetting enabled.
[    2.083884] fb: switching to amdgpudrmfb from EFI VGA
[    2.084859] amdgpu 0000:00:01.0: VRAM: 1024M 0x0000000000000000 - 0x000000003FFFFFFF (1024M used)
[    2.084862] amdgpu 0000:00:01.0: GTT: 7493M 0x0000000040000000 - 0x000000021456DFFF
[    2.084888] [drm] amdgpu: 1024M of VRAM memory ready
[    2.084890] [drm] amdgpu: 7493M of GTT memory ready.
[    2.186220] amdgpu 0000:00:01.0: amdgpu: using MSI.
[    2.186249] [drm] amdgpu: irq initialized.
[    2.186257] [drm] amdgpu: dpm initialized
[    2.190290] [drm] amdgpu atom DIG backlight initialized
[    2.190627] amdgpu 0000:00:01.0: fence driver on ring 0 use gpu addr 0x0000000040000010, cpu addr 0xffff88006cb5e010
[    2.190672] amdgpu 0000:00:01.0: fence driver on ring 1 use gpu addr 0x0000000040000020, cpu addr 0xffff88006cb5e020
[    2.190712] amdgpu 0000:00:01.0: fence driver on ring 2 use gpu addr 0x0000000040000030, cpu addr 0xffff88006cb5e030
[    2.190748] amdgpu 0000:00:01.0: fence driver on ring 3 use gpu addr 0x0000000040000040, cpu addr 0xffff88006cb5e040
[    2.190800] amdgpu 0000:00:01.0: fence driver on ring 4 use gpu addr 0x0000000040000050, cpu addr 0xffff88006cb5e050
[    2.190873] amdgpu 0000:00:01.0: fence driver on ring 5 use gpu addr 0x0000000040000060, cpu addr 0xffff88006cb5e060
[    2.190928] amdgpu 0000:00:01.0: fence driver on ring 6 use gpu addr 0x0000000040000070, cpu addr 0xffff88006cb5e070
[    2.191137] amdgpu 0000:00:01.0: fence driver on ring 7 use gpu addr 0x0000000040000080, cpu addr 0xffff88006cb5e080
[    2.191197] amdgpu 0000:00:01.0: fence driver on ring 8 use gpu addr 0x0000000040000090, cpu addr 0xffff88006cb5e090
[    2.191306] amdgpu 0000:00:01.0: fence driver on ring 9 use gpu addr 0x00000000400000a0, cpu addr 0xffff88006cb5e0a0
[    2.191806] amdgpu 0000:00:01.0: fence driver on ring 10 use gpu addr 0x00000000400000b0, cpu addr 0xffff88006cb5e0b0
[    2.192412] amdgpu 0000:00:01.0: fence driver on ring 11 use gpu addr 0x0000000000f2fd30, cpu addr 0xffffc90001e38d30
[    2.192567] amdgpu 0000:00:01.0: fence driver on ring 12 use gpu addr 0x00000000400000d0, cpu addr 0xffff88006cb5e0d0
[    2.192606] amdgpu 0000:00:01.0: fence driver on ring 13 use gpu addr 0x00000000400000e0, cpu addr 0xffff88006cb5e0e0
[    3.418049] fbcon: amdgpudrmfb (fb0) is primary device
[    6.103885] amdgpu 0000:00:01.0: fb0: amdgpudrmfb frame buffer device
[    6.265117] [drm] Initialized amdgpu 3.5.0 20150101 for 0000:00:01.0 on minor 0
[   23.222265] audit: type=1130 audit(1472603981.565:63): pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=systemd-backlight@backlight:amdgpu_bl0 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'

Adding *amdgpu.powerplay=1* on kernel boot parameter seems reducing the fan noise.
As dungeon explained, my hardware uses R7 M265DX graphic card.

[Mystro256]

It looks like your suspicion is correct. Installing linux-firmware-20160609-66.gita4bbc811 resolved the issue. The latest kernel from amd-staged repository is working fine now.

$ uname -r
4.7.2-902.amd.25082016.fc24.x86_64

$ rpm -q linux-firmware
linux-firmware-20160609-66.gita4bbc811.fc24.noarch

Here is resulted boot

$ dmesg | grep amdgpu
[ 0.000000] Command line: BOOT_IMAGE=/vmlinuz-4.7.2-902.amd.25082016.fc24.x86_64 root=UUID=232c0ccf-375a-4e4d-a080-9986ea920e1c ro rhgb quiet LANG=en_CA.UTF-8 modprobe.blacklist=sp5100_tco amdgpu.exp_hw_support=1 amdgpu.powerplay=1 modprobe.blacklist=radeon acpi.trace_state=enable "dyndbg=file ec.c +p"
[ 0.000000] Kernel command line: BOOT_IMAGE=/vmlinuz-4.7.2-902.amd.25082016.fc24.x86_64 root=UUID=232c0ccf-375a-4e4d-a080-9986ea920e1c ro rhgb quiet LANG=en_CA.UTF-8 modprobe.blacklist=sp5100_tco amdgpu.exp_hw_support=1 amdgpu.powerplay=1 modprobe.blacklist=radeon acpi.trace_state=enable "dyndbg=file ec.c +p"
[ 2.067194] [drm] amdgpu kernel modesetting enabled.
[ 2.083884] fb: switching to amdgpudrmfb from EFI VGA
[ 2.084859] amdgpu 0000:00:01.0: VRAM: 1024M 0x0000000000000000 - 0x000000003FFFFFFF (1024M used)
[ 2.084862] amdgpu 0000:00:01.0: GTT: 7493M 0x0000000040000000 - 0x000000021456DFFF
[ 2.084888] [drm] amdgpu: 1024M of VRAM memory ready
[ 2.084890] [drm] amdgpu: 7493M of GTT memory ready.
[ 2.186220] amdgpu 0000:00:01.0: amdgpu: using MSI.
[ 2.186249] [drm] amdgpu: irq initialized.
[ 2.186257] [drm] amdgpu: dpm initialized
[ 2.190290] [drm] amdgpu atom DIG backlight initialized
[ 2.190627] amdgpu 0000:00:01.0: fence driver on ring 0 use gpu addr 0x0000000040000010, cpu addr 0xffff88006cb5e010
[ 2.190672] amdgpu 0000:00:01.0: fence driver on ring 1 use gpu addr 0x0000000040000020, cpu addr 0xffff88006cb5e020
[ 2.190712] amdgpu 0000:00:01.0: fence driver on ring 2 use gpu addr 0x0000000040000030, cpu addr 0xffff88006cb5e030
[ 2.190748] amdgpu 0000:00:01.0: fence driver on ring 3 use gpu addr 0x0000000040000040, cpu addr 0xffff88006cb5e040
[ 2.190800] amdgpu 0000:00:01.0: fence driver on ring 4 use gpu addr 0x0000000040000050, cpu addr 0xffff88006cb5e050
[ 2.190873] amdgpu 0000:00:01.0: fence driver on ring 5 use gpu addr 0x0000000040000060, cpu addr 0xffff88006cb5e060
[ 2.190928] amdgpu 0000:00:01.0: fence driver on ring 6 use gpu addr 0x0000000040000070, cpu addr 0xffff88006cb5e070
[ 2.191137] amdgpu 0000:00:01.0: fence driver on ring 7 use gpu addr 0x0000000040000080, cpu addr 0xffff88006cb5e080
[ 2.191197] amdgpu 0000:00:01.0: fence driver on ring 8 use gpu addr 0x0000000040000090, cpu addr 0xffff88006cb5e090
[ 2.191306] amdgpu 0000:00:01.0: fence driver on ring 9 use gpu addr 0x00000000400000a0, cpu addr 0xffff88006cb5e0a0
[ 2.191806] amdgpu 0000:00:01.0: fence driver on ring 10 use gpu addr 0x00000000400000b0, cpu addr 0xffff88006cb5e0b0
[ 2.192412] amdgpu 0000:00:01.0: fence driver on ring 11 use gpu addr 0x0000000000f2fd30, cpu addr 0xffffc90001e38d30
[ 2.192567] amdgpu 0000:00:01.0: fence driver on ring 12 use gpu addr 0x00000000400000d0, cpu addr 0xffff88006cb5e0d0
[ 2.192606] amdgpu 0000:00:01.0: fence driver on ring 13 use gpu addr 0x00000000400000e0, cpu addr 0xffff88006cb5e0e0
[ 3.418049] fbcon: amdgpudrmfb (fb0) is primary device
[ 6.103885] amdgpu 0000:00:01.0: fb0: amdgpudrmfb frame buffer device
[ 6.265117] [drm] Initialized amdgpu 3.5.0 20150101 for 0000:00:01.0 on minor 0
[ 23.222265] audit: type=1130 audit(1472603981.565:63): pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=systemd-backlight@backlight:amdgpu_bl0 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'

Adding *amdgpu.powerplay=1* on kernel boot parameter seems reducing the fan noise.
As dungeon explained, my hardware uses R7 M265DX graphic card.

Could you report this report this upstream? The mailing list that bridgman gave should be sufficient:

https://lists.freedesktop.org/mailman/listinfo/amd-gfx

I would specify this happens when updating your linux-firmware snapshot from 20160609 (gita4bbc811) to 20160816 (git7c3dfc0b), and affects kernel 4.8+ and the staging kernel branches.

If you're able to bisect or figure out what file caused the issue, the commit is likely one of these two:

http://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git/commit/?id=1dc99be22222fbf88adc98d85f8ffb7529c6c4e3

http://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git/commit/?id=6bc2c609988823744d1a45d33ee370c8310ec305

agd5f tends to author the amd firmware commits, so he may have some insight on what could be causing it as well.

[finalzone]

Could you report this report this upstream? The mailing list that bridgman gave should be sufficient:

https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Done.