linux/drivers/gpu/drm/amd/amdgpu, branch v6.3

drm/amdgpu: Fix desktop freezed after gpu-reset

2023-04-18T21:14:13Z

[Why] After gpu-reset, sometimes the driver fails to enable vblank irq, causing flip_done timed out and the desktop freezed. During gpu-reset, we disable and enable vblank irq in dm_suspend() and dm_resume(). Later on in amdgpu_irq_gpu_reset_resume_helper(), we check irqs' refcount and decide to enable or disable the irqs again. However, we have 2 sets of API for controling vblank irq, one is dm_vblank_get/put() and another is amdgpu_irq_get/put(). Each API has its own refcount and flag to store the state of vblank irq, and they are not synchronized. In drm we use the first API to control vblank irq but in amdgpu_irq_gpu_reset_resume_helper() we use the second set of API. The failure happens when vblank irq was enabled by dm_vblank_get() before gpu-reset, we have vblank->enabled true. However, during gpu-reset, in amdgpu_irq_gpu_reset_resume_helper() vblank irq's state checked from amdgpu_irq_update() is DISABLED. So finally it disables vblank irq again. After gpu-reset, if there is a cursor plane commit, the driver will try to enable vblank irq by calling drm_vblank_enable(), but the vblank->enabled is still true, so it fails to turn on vblank irq and causes flip_done can't be completed in vblank irq handler and desktop become freezed. [How] Combining the 2 vblank control APIs by letting drm's API finally calls amdgpu_irq's API, so the irq's refcount and state of both APIs can be synchronized. Also add a check to prevent refcount from being less then 0 in amdgpu_irq_put(). v2: - Add warning in amdgpu_irq_enable() if the irq is already disabled. - Call dc_interrupt_set() in dm_set_vblank() to avoid refcount change if it is in gpu-reset. v3: - Improve commit message and code comments. Signed-off-by: Alan Liu Reviewed-by: Christian König Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org

drm/amdgpu: allow more APUs to do mode2 reset when go to S4

2023-03-30T15:23:58Z

Skip mode2 reset only for IMU enabled APUs when do S4. This patch is to fix the regression issue https://gitlab.freedesktop.org/drm/amd/-/issues/2483 It is generated by commit b589626674de ("drm/amdgpu: skip ASIC reset for APUs when go to S4"). Fixes: b589626674de ("drm/amdgpu: skip ASIC reset for APUs when go to S4") Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2483 Tested-by: Yuan Perry Signed-off-by: Tim Huang Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.1.x

drm/amdgpu/nv: Apply ASPM quirk on Intel ADL + AMD Navi

2023-03-23T13:34:35Z

S2idle resume freeze can be observed on Intel ADL + AMD WX5500. This is caused by commit 0064b0ce85bb ("drm/amd/pm: enable ASPM by default"). The root cause is still not clear for now. So extend and apply the ASPM quirk from commit e02fe3bc7aba ("drm/amdgpu: vi: disable ASPM on Intel Alder Lake based systems"), to workaround the issue on Navi cards too. Fixes: 0064b0ce85bb ("drm/amd/pm: enable ASPM by default") Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2458 Reviewed-by: Alex Deucher Signed-off-by: Kai-Heng Feng Reviewed-by: Mario Limonciello Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org

drm/amdgpu/gfx: set cg flags to enter/exit safe mode

2023-03-23T13:32:01Z

sriov needs to enter/exit safe mode in update umd p state add the cg flag to let it enter or exit while needed Signed-off-by: Jane Jian Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher

drm/amdgpu: Force signal hw_fences that are embedded in non-sched jobs

2023-03-23T13:31:30Z

[Why] For engines not supporting soft reset, i.e. VCN, there will be a failed ib test before mode 1 reset during asic reset. The fences in this case are never signaled and next time when we try to free the sa_bo, kernel will hang. [How] During pre_asic_reset, driver will clear job fences and afterwards the fences' refcount will be reduced to 1. For drm_sched_jobs it will be released in job_free_cb, and for non-sched jobs like ib_test, it's meant to be released in sa_bo_free but only when the fences are signaled. So we have to force signal the non_sched bad job's fence during pre_asic_reset or the clear is not complete. Signed-off-by: YuBiao Wang Acked-by: Luben Tuikov Signed-off-by: Alex Deucher

drm/amdgpu: add mes resume when do gfx post soft reset

2023-03-23T13:30:59Z

[why] when gfx do soft reset, mes will also do reset, if mes is not resumed when do recover from soft reset, mes is unable to respond in later sequence [how] resume mes when do gfx post soft reset Signed-off-by: Tong Liu01 Acked-by: Alex Deucher Signed-off-by: Alex Deucher

drm/amdgpu: skip ASIC reset for APUs when go to S4

2023-03-23T13:29:55Z

For GC IP v11.0.4/11, PSP TMR need to be reserved for ASIC mode2 reset. But for S4, when psp suspend, it will destroy the TMR that fails the ASIC reset. [ 96.006101] amdgpu 0000:62:00.0: amdgpu: MODE2 reset [ 100.409717] amdgpu 0000:62:00.0: amdgpu: SMU: I'm not done with your previous command: SMN_C2PMSG_66:0x00000011 SMN_C2PMSG_82:0x00000002 [ 100.411593] amdgpu 0000:62:00.0: amdgpu: Mode2 reset failed! [ 100.412470] amdgpu 0000:62:00.0: PM: pci_pm_freeze(): amdgpu_pmops_freeze+0x0/0x50 [amdgpu] returns -62 [ 100.414020] amdgpu 0000:62:00.0: PM: dpm_run_callback(): pci_pm_freeze+0x0/0xd0 returns -62 [ 100.415311] amdgpu 0000:62:00.0: PM: pci_pm_freeze+0x0/0xd0 returned -62 after 4623202 usecs [ 100.416608] amdgpu 0000:62:00.0: PM: failed to freeze async: error -62 We can skip the reset on APUs, assuming we can resume them properly. Verified on some GFX11, GFX10 and old GFX9 APUs. Signed-off-by: Tim Huang Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.1.x

drm/amdgpu: reposition the gpu reset checking for reuse

2023-03-23T13:29:32Z

Move the amdgpu_acpi_should_gpu_reset out of CONFIG_SUSPEND to share it with hibernate case. Signed-off-by: Tim Huang Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.1.x

drm/amdgpu/vcn: Disable indirect SRAM on Vangogh broken BIOSes

2023-03-14T14:40:48Z

The VCN firmware loading path enables the indirect SRAM mode if it's advertised as supported. We might have some cases of FW issues that prevents this mode to working properly though, ending-up in a failed probe. An example below, observed in the Steam Deck: [...] [drm] failed to load ucode VCN0_RAM(0x3A) [drm] psp gfx command LOAD_IP_FW(0x6) failed and response status is (0xFFFF0000) amdgpu 0000:04:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring vcn_dec_0 test failed (-110) [drm:amdgpu_device_init.cold [amdgpu]] *ERROR* hw_init of IP block failed -110 amdgpu 0000:04:00.0: amdgpu: amdgpu_device_ip_init failed amdgpu 0000:04:00.0: amdgpu: Fatal error during GPU init [...] Disabling the VCN block circumvents this, but it's a very invasive workaround that turns off the entire feature. So, let's add a quirk on VCN loading that checks for known problematic BIOSes on Vangogh, so we can proactively disable the indirect SRAM mode and allow the HW proper probe and VCN IP block to work fine. Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2385 Fixes: 82132ecc5432 ("drm/amdgpu: enable Vangogh VCN indirect sram mode") Cc: stable@vger.kernel.org Cc: James Zhu Cc: Leo Liu Signed-off-by: Guilherme G. Piccoli Signed-off-by: Alex Deucher

drm/amdgpu/nv: fix codec array for SR_IOV

2023-03-14T14:40:41Z

Copy paste error. Fixes: 384334120b66 ("drm/amdgpu/nv: don't expose AV1 if VCN0 is harvested") Reported-by: Abaci Robot Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=4454 Cc: Jiapeng Chong Acked-by: Guchun Chen Signed-off-by: Alex Deucher