amdgpu crash with 3.5.11

unclejack commented 11 months ago

Your system information

Steam client version: 1702667398
SteamOS version: 3.5.11
Opted into Steam client beta?: Yes
Opted into SteamOS beta?: Yes
Have you checked for updates in Settings > System?: Yes

Please describe your issue in as much detail as possible:

I expected gamescope and the gpu driver to not crash.

What happened:

Dec 00 00:44:53 steamdeck fancontrol.py[577]: Warning: CPU temperature of 94.0 greater than max 90! Setting fan to max speed.
Dec 00 00:44:54 steamdeck fancontrol.py[577]: Warning: CPU temperature of 92.2 greater than max 90! Setting fan to max speed.
Dec 00 00:44:55 steamdeck fancontrol.py[577]: Warning: CPU temperature of 91.0 greater than max 90! Setting fan to max speed.
Dec 00 00:45:21 steamdeck fancontrol.py[577]: Warning: CPU temperature of 90.6 greater than max 90! Setting fan to max speed.
Dec 00 00:45:45 steamdeck dbus-daemon[572]: [system] Activating via systemd: service name='org.freedesktop.home1' unit='dbus-org.freedesktop.home1.service' requested by ':1.172' (uid=0 pid=5783 comm="sudo -s")
Dec 00 00:45:45 steamdeck dbus-daemon[572]: [system] Activation via systemd failed for unit 'dbus-org.freedesktop.home1.service': Unit dbus-org.freedesktop.home1.service not found.
Dec 00 00:48:19 steamdeck kernel: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring sdma0 timeout, signaled seq=35171, emitted seq=35175
Dec 00 00:48:19 steamdeck kernel: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process  pid 0 thread  pid 0
Dec 00 00:48:19 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: GPU reset begin!
Dec 00 00:48:19 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: MODE2 reset
Dec 00 00:48:19 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: GPU reset succeeded, trying to resume
Dec 00 00:48:19 steamdeck kernel: [drm] PCIE GART of 1024M enabled (table at 0x000000F43FC00000).
Dec 00 00:48:19 steamdeck kernel: [drm] PSP is resuming...
Dec 00 00:48:19 steamdeck (udev-worker)[5882]: devcd1: Process 'cat /sys/devices/virtual/devcoredump/devcd1/data > /var/lib/steamos-log-submitter/pending/devcoredump/4785' failed with exit code 1.
Dec 00 00:48:19 steamdeck kernel: [drm] reserve 0xa00000 from 0xf43e000000 for PSP TMR
Dec 00 00:48:19 steamdeck fancontrol.py[577]: Traceback (most recent call last):
Dec 00 00:48:19 steamdeck fancontrol.py[577]:   File "/usr/share/jupiter-fan-control/fancontrol.py", line 542, in <module>
Dec 00 00:48:19 steamdeck fancontrol.py[577]:     controller.loop_control()
Dec 00 00:48:19 steamdeck fancontrol.py[577]:   File "/usr/share/jupiter-fan-control/fancontrol.py", line 486, in loop_control
Dec 00 00:48:19 steamdeck fancontrol.py[577]:     self.loop_read_sensors()
Dec 00 00:48:19 steamdeck fancontrol.py[577]:   File "/usr/share/jupiter-fan-control/fancontrol.py", line 452, in loop_read_sensors
Dec 00 00:48:19 steamdeck fancontrol.py[577]:     self.power_sensor.get_avg_value()
Dec 00 00:48:19 steamdeck fancontrol.py[577]:   File "/usr/share/jupiter-fan-control/fancontrol.py", line 356, in get_avg_value
Dec 00 00:48:19 steamdeck fancontrol.py[577]:     self.values.append(self.get_value())
Dec 00 00:48:19 steamdeck fancontrol.py[577]:                        ^^^^^^^^^^^^^^^^
Dec 00 00:48:19 steamdeck fancontrol.py[577]:   File "/usr/share/jupiter-fan-control/fancontrol.py", line 351, in get_value
Dec 00 00:48:19 steamdeck fancontrol.py[577]:     self.value = int(f.read().strip()) / 1000000
Dec 00 00:48:19 steamdeck fancontrol.py[577]:                      ^^^^^^^^
Dec 00 00:48:19 steamdeck fancontrol.py[577]: PermissionError: [Errno 1] Operation not permitted
Dec 00 00:48:19 steamdeck systemd[1]: jupiter-fan-control.service: Main process exited, code=exited, status=1/FAILURE
Dec 00 00:48:19 steamdeck fancontrol.py[5887]: loaded critical temp from SSD hwmon: 79.85
Dec 00 00:48:19 steamdeck fancontrol.py[5887]: returning fan to EC control loop
Dec 00 00:48:19 steamdeck systemd[1]: jupiter-fan-control.service: Failed with result 'exit-code'.
Dec 00 00:48:19 steamdeck systemd[1]: jupiter-fan-control.service: Consumed 15.142s CPU time.
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: SMU is resuming...
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: SMU is resumed successfully!
Dec 00 00:48:20 steamdeck kernel: [drm] DMUB hardware initialized: version=0x0300000A
Dec 00 00:48:20 steamdeck kernel: [drm] Failed to add display topology, DTM TA is not initialized.
Dec 00 00:48:20 steamdeck kernel: [drm] kiq ring mec 2 pipe 1 q 0
Dec 00 00:48:20 steamdeck kernel: [drm] VCN decode and encode initialized successfully(under DPG Mode).
Dec 00 00:48:20 steamdeck kernel: [drm] JPEG decode initialized successfully.
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 5 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 6 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 7 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 8 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 9 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 10 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring kiq_0.2.1.0 uses VM inv eng 11 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring vcn_dec_0 uses VM inv eng 0 on hub 8
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring vcn_enc_0.0 uses VM inv eng 1 on hub 8
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring vcn_enc_0.1 uses VM inv eng 4 on hub 8
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: ring jpeg_dec uses VM inv eng 5 on hub 8
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: recover vram bo from shadow start
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: recover vram bo from shadow done
Dec 00 00:48:20 steamdeck kernel: amdgpu 0000:04:00.0: amdgpu: GPU reset(1) succeeded!
Dec 00 00:48:20 steamdeck systemd[1]: jupiter-fan-control.service: Scheduled restart job, restart counter is at 1.
Dec 00 00:48:20 steamdeck systemd[1]: Stopped Jupiter fan control.
Dec 00 00:48:20 steamdeck systemd[1]: jupiter-fan-control.service: Consumed 15.142s CPU time.
Dec 00 00:48:20 steamdeck systemd[1]: Started Jupiter fan control.
Dec 00 00:48:21 steamdeck fancontrol.py[5897]: loaded critical temp from SSD hwmon: 79.85
Dec 00 00:48:21 steamdeck fancontrol.py[5897]: jupiter-fan-control started successfully.
Dec 00 00:49:25 steamdeck dbus-daemon[572]: [system] Activating via systemd: service name='org.freedesktop.timedate1' unit='dbus-org.freedesktop.timedate1.service' requested by ':1.174' (uid=1000 pid=5944 comm="timedatectl status")
Dec 00 00:49:25 steamdeck systemd[1]: Starting Time & Date Service...
Dec 00 00:49:25 steamdeck dbus-daemon[572]: [system] Successfully activated service 'org.freedesktop.timedate1'
Dec 00 00:49:25 steamdeck systemd[1]: Started Time & Date Service.
Dec 00 00:49:35 steamdeck systemd[1]: Created slice Slice /system/systemd-coredump.
Dec 00 00:49:35 steamdeck systemd[1]: Started Process Core Dump (PID 5980/UID 0).
Dec 00 00:49:35 steamdeck core_handler[5981]: Minidump generated at /var/lib/steamos-log-submitter/pending/minidump/.staging-1702676974-gamescope-xwm-3325-None.dmp
Dec 00 00:49:35 steamdeck kernel: input: Steam Deck as /devices/pci0000:00/0000:00:08.1/0000:04:00.4/usb3/3-3/3-3:1.2/0003:28DE:1205.0003/input/input35
Dec 00 00:49:35 steamdeck systemd-coredump[5982]: Process 3325 (gamescope-wl) of user 1000 dumped core.

                                                  Stack trace of thread 3360:
                                                  #0  0x00007f9d5589f26c n/a (libc.so.6 + 0x8926c)
                                                  #1  0x00007f9d5584fa08 raise (libc.so.6 + 0x39a08)
                                                  #2  0x00007f9d55838538 abort (libc.so.6 + 0x22538)
                                                  #3  0x00007f9d5583845c n/a (libc.so.6 + 0x2245c)
                                                  #4  0x00007f9d558483d6 __assert_fail (libc.so.6 + 0x323d6)
                                                  #5  0x0000561db0f8cd97 n/a (gamescope + 0x7fd97)
                                                  #6  0x0000561db0f960ca n/a (gamescope + 0x890ca)
                                                  #7  0x0000561db0f658a0 n/a (gamescope + 0x588a0)
                                                  #8  0x0000561db0f67b3f n/a (gamescope + 0x5ab3f)
                                                  #9  0x0000561db0f82fac n/a (gamescope + 0x75fac)
                                                  #10 0x00007f9d55ae1943 execute_native_thread_routine (libstdc++.so.6 + 0xe1943)
                                                  #11 0x00007f9d5589d44b n/a (libc.so.6 + 0x8744b)
                                                  #12 0x00007f9d55920e40 n/a (libc.so.6 + 0x10ae40)

                                                  Stack trace of thread 3325:
                                                  #0  0x00007f9d55913c0f __poll (libc.so.6 + 0xfdc0f)
                                                  #1  0x0000561db0f8555f n/a (gamescope + 0x7855f)
                                                  #2  0x0000561db0f2f446 n/a (gamescope + 0x22446)
                                                  #3  0x00007f9d55839850 n/a (libc.so.6 + 0x23850)
                                                  #4  0x00007f9d5583990a __libc_start_main (libc.so.6 + 0x2390a)
                                                  #5  0x0000561db0f51555 n/a (gamescope + 0x44555)

                                                  Stack trace of thread 3326:
                                                  #0  0x00007f9d55921266 epoll_wait (libc.so.6 + 0x10b266)
                                                  #1  0x0000561db0f73bcf n/a (gamescope + 0x66bcf)
                                                  #2  0x0000561db0f77424 n/a (gamescope + 0x6a424)
                                                  #3  0x00007f9d55ae1943 execute_native_thread_routine (libstdc++.so.6 + 0xe1943)
                                                  #4  0x00007f9d5589d44b n/a (libc.so.6 + 0x8744b)
                                                  #5  0x00007f9d55920e40 n/a (libc.so.6 + 0x10ae40)

                                                  Stack trace of thread 3328:
                                                  #0  0x00007f9d55913c0f __poll (libc.so.6 + 0xfdc0f)
                                                  #1  0x0000561db0f84987 n/a (gamescope + 0x77987)
                                                  #2  0x00007f9d55ae1943 execute_native_thread_routine (libstdc++.so.6 + 0xe1943)
                                                  #3  0x00007f9d5589d44b n/a (libc.so.6 + 0x8744b)
                                                  #4  0x00007f9d55920e40 n/a (libc.so.6 + 0x10ae40)

                                                  Stack trace of thread 3330:
                                                  #0  0x00007f9d558e59e5 clock_nanosleep (libc.so.6 + 0xcf9e5)
                                                  #1  0x00007f9d558ea5e7 __nanosleep (libc.so.6 + 0xd45e7)
                                                  #2  0x00007f9d54100455 n/a (libvulkan_radeon.so + 0x100455)
                                                  #3  0x00007f9d5425c7cc n/a (libvulkan_radeon.so + 0x25c7cc)
                                                  #4  0x00007f9d5589d44b n/a (libc.so.6 + 0x8744b)
                                                  #5  0x00007f9d55920e40 n/a (libc.so.6 + 0x10ae40)

                                                  Stack trace of thread 3359:
                                                  #0  0x00007f9d55913c0f __poll (libc.so.6 + 0xfdc0f)
                                                  #1  0x0000561db0fa99b2 n/a (gamescope + 0x9c9b2)
                                                  #2  0x00007f9d55ae1943 execute_native_thread_routine (libstdc++.so.6 + 0xe1943)
                                                  #3  0x00007f9d5589d44b n/a (libc.so.6 + 0x8744b)
                                                  #4  0x00007f9d55920e40 n/a (libc.so.6 + 0x10ae40)

                                                  Stack trace of thread 3362:
                                                  #0  0x00007f9d558e59e5 clock_nanosleep (libc.so.6 + 0xcf9e5)
                                                  #1  0x00007f9d558ea5e7 __nanosleep (libc.so.6 + 0xd45e7)
                                                  #2  0x0000561db0f85037 n/a (gamescope + 0x78037)
                                                  #3  0x00007f9d55ae1943 execute_native_thread_routine (libstdc++.so.6 + 0xe1943)
                                                  #4  0x00007f9d5589d44b n/a (libc.so.6 + 0x8744b)
                                                  #5  0x00007f9d55920e40 n/a (libc.so.6 + 0x10ae40)

                                                  Stack trace of thread 3358:
                                                  #0  0x00007f9d55921266 epoll_wait (libc.so.6 + 0x10b266)
                                                  #1  0x00007f9d48148579 n/a (libspa-support.so + 0x13579)
                                                  #2  0x00007f9d4813bbe3 n/a (libspa-support.so + 0x6be3)
                                                  #3  0x00007f9d55eb026f n/a (libpipewire-0.3.so.0 + 0x4126f)
                                                  #4  0x00007f9d5589d44b n/a (libc.so.6 + 0x8744b)
                                                  #5  0x00007f9d55920e40 n/a (libc.so.6 + 0x10ae40)

                                                  Stack trace of thread 3329:
                                                  #0  0x00007f9d55899f0e n/a (libc.so.6 + 0x83f0e)
                                                  #1  0x00007f9d5589c7a0 pthread_cond_wait (libc.so.6 + 0x867a0)
                                                  #2  0x00007f9d5425c89e n/a (libvulkan_radeon.so + 0x25c89e)
                                                  #3  0x00007f9d54239e0c n/a (libvulkan_radeon.so + 0x239e0c)
                                                  #4  0x00007f9d5425c7cc n/a (libvulkan_radeon.so + 0x25c7cc)
                                                  #5  0x00007f9d5589d44b n/a (libc.so.6 + 0x8744b)
                                                  #6  0x00007f9d55920e40 n/a (libc.so.6 + 0x10ae40)

                                                  Stack trace of thread 3361:
                                                  #0  0x00007f9d5590f900 __open64 (libc.so.6 + 0xf9900)
                                                  #1  0x0000561db0f5dbe5 n/a (gamescope + 0x50be5)
                                                  #2  0x00007f9d55ae1943 execute_native_thread_routine (libstdc++.so.6 + 0xe1943)
                                                  #3  0x00007f9d5589d44b n/a (libc.so.6 + 0x8744b)
                                                  #4  0x00007f9d55920e40 n/a (libc.so.6 + 0x10ae40)
                                                  ELF object binary architecture: AMD x86-64
Dec 00 00:49:35 steamdeck systemd[1]: systemd-coredump@0-5980-0.service: Deactivated successfully.

I'll retrieve the dumps to provide them.

Steps for reproducing this issue:

Play a game for a while on 3.5.11
See the driver crash at some point with the image freezing on the screen
The screen goes black after a while and the frozen image returns
Gamescope recovers somewhat with the steam menu being visible under the frozen image
Crashes again to a black screen
The gamescope session restarts after a timeout

RodoMa92 commented 10 months ago

Replying to https://github.com/ValveSoftware/SteamOS/issues/1312#issuecomment-1880004592

Feedback this to the AMD bugzilla above then. Their engineers are struggling to repro this issue it seems. The more reproducer they get the quicker we can get a fix from them.

hungrymonkey commented 10 months ago

https://www.phoronix.com/news/Radeon-Gallium3D-SDMA-Dropped

https://www.phoronix.com/news/RadeonSI-Disables-Polaris-SDMA

https://www.phoronix.com/news/RadeonSI-SDMA-CIK-CZ-Again

https://www.phoronix.com/news/AMDGPU-LSDMA-Light-SDMA

https://www.phoronix.com/news/RadeonSI-New-SDMA-Tex-Copy

I decided to google SDMA for the heck of it. The features has so many issues

https://gitlab.freedesktop.org/mesa/mesa/-/issues/1889

AMD_DEBUG=nodma I cannot reproduce this bug reliably, but I wonder whether this flag will help.

https://docs.kernel.org/gpu/amdgpu/driver-core.html

RodoMa92 commented 10 months ago

If anyone that can reproduce this easily enough can test this, that would be useful:

https://gitlab.freedesktop.org/drm/amd/-/issues/2220#note_2229270

I doubt it will change much, but you can always try.

RodoMa92 commented 10 months ago

Replying to https://github.com/ValveSoftware/SteamOS/issues/1312#issuecomment-1880143149

Additional details found here: https://rocm.docs.amd.com/en/develop/conceptual/gpu-memory.html#system-direct-memory-access

It seems that my hyphotesis on the function based on the name wasn't that far off.

RodoMa92 commented 10 months ago

Completely forgot about my desktop crashes on idle, but it was still a sdma0 ring crash again.

@unclejack if you can easily test stuff, can you also try this test patch here? I never managed to test this especially on an immutable distro, but it might be worth to test with people that can repro this much quicker.

Thanks,

Marco.

RodoMa92 commented 10 months ago

Another data point I've found roaming around is that seemingly forcing amdgpu.vm_update_mode=3 seems to resolve the issue, and according to the driver docs:

vm_update_mode (int)

Override VM update mode. VM updated by using CPU (0 = never, 1 = Graphics only, 2 = Compute only, 3 = Both). The default is -1 (Only in large BAR(LB) systems Compute VM tables will be updated by CPU, otherwise 0, never).

This always force the CPU to do virtual memory updates (which likely basically disable the sdma ring to do the job). It has a performance hit, but for testing it might be worth it temporarily while AMD wakes up. The patch in my previous post should be tested before this tho.

unclejack commented 10 months ago

@RodoMa92: I don't have a simple way to reproduce the crash. It happens randomly when I'm in game. Perhaps it might be a good idea to find a way to reproduce the crash in a reliable way. That's likely to be a better idea.

Regardless, I'll build a new kernel with all the stab-in-the-dark kernel patches later.

update: I've managed to crash the driver again. Starting the same game which plays video before starting is what crashed it on a fresh boot. The Steam Deck was on battery. The newly built kernel is based on Valve's 6.1.52-valve14 kernel tree. It has the patch https://gitlab.freedesktop.org/drm/amd/uploads/ecfb67b0ae46e95d7ab30c49c932c95f/0001-drm-amdgpu-add-wmb-barrier-for-sdma-timeout-issue-te.patch applied on top. The patched kernel hasn't crashed yet. That probably doesn't mean much since it doesn't always crash.

RodoMa92 commented 10 months ago

Replying to https://github.com/ValveSoftware/SteamOS/issues/1312#issuecomment-1881479445

You got a gpu crash and it recovered or only the game crashed? Does the kernel log looks identical as before?

Sadly mine has been quite decent since I've got it, so repro this on my end is extremely difficult.

unclejack commented 10 months ago

Ok, this is what I've done so far:

reproduced the crashes more reliably after reboot with the unpatched kernel 6.1.52-valve14-1 from Valve's SteamOS 3.5.13
built a patched kernel based on the tree of the kernel from 3.5.13 & switched to it
managed to reproduce the crashes with the patched kernel based on 6.1.52-valve14 & the patch https://gitlab.freedesktop.org/drm/amd/uploads/ecfb67b0ae46e95d7ab30c49c932c95f/0001-drm-amdgpu-add-wmb-barrier-for-sdma-timeout-issue-te.patch

I've also posted a comment on GitLab to let the people from AMD know. I have a semi-reliable way to trigger this bug now.

update: Since the amdgpu driver still crashes with the patch which attempts to work around potential cache coherence issues and the patches are provided on GitLab, I'll only post updates there. There have been no updates from Valve and no guidance was provided either after testing. I'll stop testing and debugging once the full set of existing patches is tested.

RoyalBrook commented 10 months ago

Can confirm mine also has the same sort of issues, logs are the same too save for the temp warnings. Setting the UMA frame buffer to 4gb seems to offer longer times between crashes, but doesn't stop them entirely.

edit: I did have some success, had a starfield save in a shop where firing would instantly cause the crash, after doing an APU reset which I didn't even realize it was an option, it didn't crash and I was able to follow through. I'll update if the issue seems to return, but in the meantime if anyone finds this post and is out of options otherwise, power off the device. Hold down the vol - button and the quick access (...) button, and then hold the power button, after it beeps let go of the power button and wait for it to boot, it'll take a while

edit 2: Unfortunately still got it it just took a while :(

unclejack commented 9 months ago

https://store.steampowered.com/news/app/1675200/view/4064004735511926127 has become available. It includes some changes which may help avoid crashes in some cases. The relevant commits added since valve14 are here https://gitlab.com/evlaV/linux-integration/-/commits/6.1.52-valve16?ref_type=tags.

Testing that 3.5.15 preview release might be a good idea if your Steam Deck crashes.

Voltage offsets for undervolting should be disabled in the firmware or in any tool if you have something like that. The silicon might not be good enough to work properly with those voltage offsets or may cause instability under load. It's something to rule out anyway.

Terrance commented 7 months ago

I believe I'm seeing the same issues described here (with similar GPU / gamescope logs and without any temperature warnings).

Testing that 3.5.15 preview release might be a good idea if your Steam Deck crashes.

Am I right in thinking these changes have made it to the stable channel now? I'm on OS Version 3.5.17, Kernel Version 6.1.52-valve16, where it sounds like the potential fixes were released, so it seems they haven't helped in my case.

fejnartal commented 5 months ago

I'm also experiencing this error in several games. I even have a support ticket open in steamppwered.com (HT-6PJV-6T7D-XJD6).

They asked me to send the Steam Deck to the Repair Center, even after I told them about this GitHub issue. Although, in my opinion, if this issue is in fact caused by a driver, sending it to the Repair Center is both a waste of time and resources for both Steam and myself.

But since they didn't provide me with an alternative I will send it anyway. I'm writing this in case it can help with fixing the issue.

To easily reproduce the issue try playing "Headsnatchers" in single player mode (aka Zombie Castle). Everytime I tried it crashed in less than 15min (even if you leave the game open without playing, it tends to end up crashing, eventually).

I have always used stable channel, my Steam Deck specs are as follows:

OS Name: "SteamOS Holo" OS Codename: holo OS Variant: steamdeck OS Version: 3.5.19 OS Build: 20240422.1 Kernel Version: 6.1.52-valve16-1-neptune-61 Steam Deck Controller FW Build Date: Sun, Mar 3 11:54 PM UTC +01:00 BIOS Version: F7A0120

Steam Version: 1716584667 Steam Client Build Date: Fri, May 24 10:48 PM UTC +01:00 Steam Web Build Date: Fri, May 24 10:31 PM UTC +01:00 Steam API Version: SteamClient021

CPU Vendor: AuthenticAMD CPU Name: AMD Custom APU 0405 CPU Frequency: 2.8 GHz CPU Physical Cores: 4 CPU Logical Cores: 8 RAM Size: 14.47 GB Video Card: AMD AMD Custom GPU 0405 (vangogh, LLVM 15.0.7, DRM 3.5.4, 6.1.52-valve16-1-neptune-61) Video Driver: 4.6 (Compatibility Profile) Mesa 23.1.3 (git-87ebaf765d) VRAM Size: 1,024 MB

It's the LCD 512GB SSD model. I bought it refurbished directly from Steam on November 2023. It has been presenting this issue from day 1.

In case this helps, I have been able to play the 3D game "Prey" withouth any issues.

unclejack commented 5 months ago

Those who still run into this issue should send their Steam Deck's serial number to Mario Limonciello from AMD: https://gitlab.freedesktop.org/drm/amd/-/issues/3111#note_2438007. The goal is to figure out whether all of the affected units are from the same batch. This could help sort out this issue.

emcy849 commented 5 months ago

https://github.com/ValveSoftware/SteamOS/issues/1312#issuecomment-2148121815

He didnt say what his email is?

unclejack commented 1 month ago

For those who can still reproduce this issue, please post your details here https://gitlab.freedesktop.org/drm/amd/-/issues/3111. This will help avoid having the issue closed there. The issue should stay open since the root cause of the issue hasn't been discovered. My unit doesn't seem to exhibit the issue anymore.

@emcy849: You can find that easily on the Internet. It might not be a good idea to post it to avoid spam.

emcy849 commented 1 month ago

Just obfuscate the email and post it?

ValveSoftware / SteamOS