13:55 <bl4ckb0ne> austriancoder: got the stream(s) dumped and decoded

13:57 <austriancoder> bl4ckb0ne: great

13:58 <bl4ckb0ne> what should I be looking for in them

14:00 <austriancoder> If you have a devcoredump you should see where the FE is stuck. Last DMA address. This should point you somewhere into the command stream.

14:02 <bl4ckb0ne> still not getting those devcoredump, ill work on that

14:19 <bl4ckb0ne> ha, had `N` in /sys/module/etnaviv/parameters/dump_core

15:34 <bl4ckb0ne> > 00000004 = 7fffbf76 Idle: FE- DE+ PE+ SH- PA+ SE+ RA+ TX- VG+ IM+ FP+ TS+

15:34 <bl4ckb0ne> is that the address i have to look for in the cmd stream dump?

15:36 <austriancoder> That's the register offset of the idle register (mmio) and it's value

15:37 <austriancoder> There should be a Command DMA address line or something like that

15:37 <bl4ckb0ne> > 00000660 = 00000001 Cmd: [dec DMA: idle Fetch: idle] Req idle Cal idle

15:38 <austriancoder> 00000664 is the one you want

15:40 <bl4ckb0ne> no hit in all 16 dums

15:40 <bl4ckb0ne> dumps*

15:44 <austriancoder> Maybe you can share a zip with all the information so I might have a look

15:45 <bl4ckb0ne> sure

15:46 <austriancoder> Will have time in about 3 hours. Dinner time here

15:48 <lynxeye> bl4ckb0ne: Note that you want 20faf2005ec8 ("drm/etnaviv: fix dumping of active MMU context") on your kernel, else the addresses for the cmdbuffers might not match what the FE is executing at the time of the hang

15:49 <bl4ckb0ne> https://litter.catbox.moe/agmpcv.gz tarball of the stream + devcoredump

15:49 <bl4ckb0ne> lynxeye: ill grab that one thanks

17:57 <bl4ckb0ne> i think the linux-fslc-imx repo alreay has that one

17:58 <bl4ckb0ne> https://github.com/Freescale/linux-fslc/commit/1dd43e1e54d50277d590633e9dc1a205ddb279c9 yup

20:43 <bl4ckb0ne> > 00000664 = 00000905 Command DMA address

20:43 <bl4ckb0ne> the `00000905` value is consistant across runs