12:36MrCooper: nice, got 100% pass rate with deqp-gles31 on Navi 14 today
15:33mareko: Venemo: FYI, DISPATCH_TASKMESH_GFX and DISPATCH_MESH_INDIRECT_MULTI should set RESET_FILTER_CAM in the packet header according to PAL
17:49Venemo: mareko: what is RESET_FILTER_CAM? I am not familiar with that
18:49KitsuWhooa: A dvb stream with bad signal just crashed my 6900XT and amdgpu had to reset it. Is there anything I can do to collect information about this so that there's any hope of it being fixed?
18:55mareko: Venemo: it's a bit in the packer header
18:59mareko: I don't know the exact meaning
19:32Venemo: mareko: is this for GFX11 or also 10.3? I don't recall it from last time I looked at PAL
20:06iive: KitsuWhooa, do you have recording of the stream? Can you reproduce the issue?
20:06KitsuWhooa: unfortunately I do not have a recording of it, but I assume I can reproduce it again (if I can figure out how to dump the stream to disk)
20:09iive: KitsuWhooa, i'm using tvheadend under linux.
20:10KitsuWhooa: Does it let you write the raw ts stream to disk?
20:10iive: yes
20:10KitsuWhooa: Awesome, I'll give it a try, thanks!
20:11iive: it has passthrough mode
21:03dok: Hi, I am having a strange issue with my graphics card (pitcairn) amdgpu driver no longer works since yesterday, since a crash while playing a game
21:04dok: right now the only solution I have is to disable both radeon and amdgpu drivers, and only use the linux console
21:05dok: I didn't find any messages in journalctl at the time of the crash
21:05agd5f: dok, try unplugging your PC for a little while and try again. Could also be that the crash was due to some hardware failing on the board
21:05dok: but when I tried booting with amdgpu i had the following error:
21:06dok: [drm:uvd_v3_1_hw_init [amdgpu]] *ERROR* amdgpu: UVD Firmware validate fail (-22).
21:06dok: agd5f: I left my PC ~20h since the crash but not without power
21:10agd5f: dok, you can try skipping UVD by setting amdgpu.ip_block_mask=0x7f
21:10dok: okay I'll try
21:11agd5f: but try removing power
21:11dok: how long is "a little while" ?
21:12agd5f: 30 seconds. maybe a minutes. enough time to let the capacitors drain
21:18dok: what's the correct way to pass more than one module parameter on the linux boot cmd line ?
21:18dok: is it okay if I do: amdgpu.si_support=1 followed by amdgpu.ip_block_mask=0x7f ?
21:20agd5f: dok, yes
21:23dok: no luck, i still have a black screen
21:25dok: and I don't know in which state the kernel is, but I think it is in a shitty state as even the capslock led on my keyboard isn't responding
21:25dok: (or numlock)
21:42dok: if I set amdgpu.ip_block_mask=0x3f I don't a graphics but the kernel doesn't seems to crash
21:43dok: same with 0xbf
21:46dok: and I have a stack trace with 0x40
21:48dok: meh it's a null pointer dereference in amdgpu_device_need_post+0x4d/0xd0
21:59dok: this feels like an hardware failure, do you think a software bug could have bricked my gpu ? or maybe caused by "wear"
22:04agd5f: sounds like the hw failure caused the GPU to hang in the first place
22:05agd5f: check the caps on the GPU. see if any of them are leaking. Could also try reseating the board, clearing out the dust, etc.
22:07dok: i've cleaned the card ~1 month ago... first time since maybe 2years ^^
22:08dok: but this card is almost 13years old
22:26dok: all caps seems good, magic smoke hasn't escaped either
22:27airlied: could also be a psu esp if it's 13 years old
22:30dok: the powersupply is more recent... but winning a lot
22:37dok: i am going to stop for tonight, but yeah this look like the card is dead :/
22:38dok: I was already tired, it's not 13 years old but almost 11
22:39dok: bought it during summer 2012
23:53mareko: Venemo: it applies to both gfx10.3 and gfx11