[Bug 107296] WARNING: CPU: 0 PID: 370 at drivers/gpu/drm/amd/amdgpu/../display/dc/calcs/dcn_calcs.c:1355 dcn_bw_update_from_pplib+0x16b/0x280 [amdgpu]

[bugzilla-daemon@freedesktop.org]

[-- Attachment #1.1: Type: text/plain, Size: 2881 bytes --]

https://bugs.freedesktop.org/show_bug.cgi?id=107296

--- Comment #13 from John Pathy <pathyj@kwic.com> ---
I have the same problem with my 2400G.

My kernel is as follows

[    0.000000] Linux version 4.19.0-5-amd64 (debian-kernel@lists.debian.org)
(gcc version 8.3.0 (Debian 8.3.0-7)) #1 SMP Debian 4.19.37-3 (2019-05-15)

When my system boots I get a RIP with a trace as above

[    1.532390] amdgpu: [powerplay] dpm has been enabled
[    1.532453] [drm] DM_PPLIB: values for Invalid clock
[    1.532454] [drm] DM_PPLIB:   0 in kHz
[    1.532456] [drm] DM_PPLIB:   400000 in kHz
[    1.532457] [drm] DM_PPLIB:   933000 in kHz
[    1.532459] [drm] DM_PPLIB:   1067000 in kHz
[    1.532537] WARNING: CPU: 6 PID: 134 at
drivers/gpu/drm/amd/amdgpu/../display/dc/calcs/dcn_calcs.c:1372
dcn_bw_update_from_pplib+0x171/0x290 [amdgpu]
[    1.532540] Modules linked in: amdkfd amdgpu(+) crc32c_intel mxm_wmi chash
gpu_sched ttm aesni_intel ahci drm_kms_helper aes_x86_64 crypto_simd libahci
cryptd glue_helper xhci_pci libata drm xhci_hcd igb i2c_piix4 scsi_mod usbcore
dca i2c_algo_bit usb_common video wmi gpio_amdpt gpio_generic button
[    1.532558] CPU: 6 PID: 134 Comm: systemd-udevd Not tainted 4.19.0-5-amd64
#1 Debian 4.19.37-3
[    1.532561] Hardware name: Gigabyte Technology Co., Ltd. X470 AORUS ULTRA
GAMING/X470 AORUS ULTRA GAMING-CF, BIOS F30 04/16/2019
[    1.532620] RIP: 0010:dcn_bw_update_from_pplib+0x171/0x290 [amdgpu]

So I looked into this and the problem I see is that the fclks are not
initializing properly is routine

res = dm_pp_get_clock_levels_by_type_with_voltage(
                        ctx, DM_PP_CLOCK_TYPE_FCLK, &fclks);

When the sanity check in 

res = verify_clock_values(&fclks);

is completed, it detects the 0 in index [0] and initiated the RIP. This is the
data that initiated the RIP.

[    1.532454] [drm] DM_PPLIB:   0 in kHz

The fclks in my log appear to be shifted down. I seems it should look like
this, which is what a 2200G reports on boot. This was from somebody else's
system.

[    3.774892] [drm] DM_PPLIB: values for F clock
[    3.774894] [drm] DM_PPLIB:   400000 in kHz
[    3.774894] [drm] DM_PPLIB:   933000 in kHz
[    3.774895] [drm] DM_PPLIB:   1067000 in kHz
[    3.774895] [drm] DM_PPLIB:   1200000 in kHz
[    3.774896] [drm] DM_PPLIB: values for DCF clock
[    3.774896] [drm] DM_PPLIB:   300000 in kHz
[    3.774897] [drm] DM_PPLIB:   600000 in kHz
[    3.774897] [drm] DM_PPLIB:   626000 in kHz
[    3.774897] [drm] DM_PPLIB:   654000 in kHz

Also to note is that for my kernel (4.19.0-5-amd64) the F and DCF are reported
as invalid, but for subsequent kernels 5.1.3 and 5.2-rc1, this appears to be
corrected.

My whole kern.log is located in the attachment for comment 12.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[-- Attachment #1.2: Type: text/html, Size: 3944 bytes --]

[-- Attachment #2: Type: text/plain, Size: 159 bytes --]

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel