Bug 105663

From IRC:

<marta_> Adrinael, igt@tools_test@tools_test is 100% incomplete in drmtip runs on fi-cnl-drrs, I did a new bug for this https://bugs.freedesktop.org/show_bug.cgi?id=105663
<mlankhorst> vsyrjala: could we unify skl sprite mod supported and primary mod supported?
* jsaa has quit (Quit: Leaving.)
<mlankhorst> the identical name probably gives a clue that they should be unified
<vsyrjala> there is a patch for that on the list as well
<vsyrjala> though it needs a rebase after this one
-SaslServ- <Unknown user (via SASL):host217-101-static.9-188-b.business.telecomitalia.it> failed to login to marta.  There have been 334 failed login attempts since your last successful login.
* [Enrico] (~chiccoroc@gentoo/contributor/Enrico) has joined
-NickServ- marta!~Thunderbi@host217-101-static.9-188-b.business.telecomitalia.it failed to login to marta.  There have been 335 failed login attempts since your last successful login.
* jeffmc (~jeff.mcge@134.134.139.76) has joined
<mlankhorst> we should probably do that first
<Adrinael> marta_, when the machine is idle, can you sneak in and run these commands:   intel_reg read 0x4030,   intel_reg dump
<Adrinael> That's all tools_test@tools_test is doing
<Adrinael> marta_, works on other cnls?
<marta_> Adrinael, yes only the drrs, so probably BIOS again...
<Adrinael> Executing those commands will tell whether the issue is real (or test is f00ked), or fallout from other things
<marta_> Adrinael, I don't think I dare to do that tsa, might be mad if I hang the machine...
<Adrinael> We can possibly convince tsa to do that himself :P
<vsyrjala> mlankhorst: i don't want to push all that to drm-misc-next, which is where we need this thing to land
<marta_> Adrinael, yeah better ask tsa, maybe I don't know which machine it is ;)
<Adrinael> What register is 0x4030 anyway, is that kosher on cnl
<vsyrjala> why isn't elk included in bat runs?
* vsyrjala has probably asked before and forgot
<vsyrjala> oh because it is
* vsyrjala just cliked th wrong link
<mlankhorst> ah
<ickle> 0x4030 was replace by 0x4a08
<ickle> who know what now lies behind box 0x4030
<tsa> Adrinael: cnl-drrs: (0x00004030): 0x00004000
<tsa> ickle: (0x00004a08): 0x00000000
<Adrinael> So it didn't hang?
* Notify: jsaa is online (Ubuntu Servers (freenode))
* jsaa (~jsaa@134.134.139.76) has joined
<Adrinael> What about dump?
* kumarmah (~mahesh@nat/intel/x-qhldvlnfzhfrbanj) has joined
<tsa> Adrinael: dump whatnow
<tsa> reg dump hangs, yes
<vsyrjala> hmm. whatis gdr?
* pekkari has quit (Quit: Konversation terminated!)
* krogovin (~kevin@91-150-49-145.customer.karistelefon.fi) has joined
<Adrinael> If I read the output correctly, it reads PFC_CTL_1 successfully, and hangs reading PFC_WIN_POS
* vbelgaum (Thunderbir@nat/intel/x-tdnndqaffrkqfnkc) has joined
<vsyrjala> those register offsets are bogus
<vsyrjala> hmm. or c, maybe not.
<vsyrjala> well, they are bogus for skl+

I put this back to drm/intel since there is no conclusion if it is a driver or igt issue yet. Change to IGT if/when a conclusion is clear.

Sorry for the spam, it was difficult to read:

<marta_> Adrinael, igt@tools_test@tools_test is 100% incomplete in drmtip runs on fi-cnl-drrs, I did a new bug for this https://bugs.freedesktop.org/show_bug.cgi?id=105663
<Adrinael> marta_, when the machine is idle, can you sneak in and run these commands:   intel_reg read 0x4030,   intel_reg dump
<Adrinael> That's all tools_test@tools_test is doing
<Adrinael> marta_, works on other cnls?
<marta_> Adrinael, yes only the drrs, so probably BIOS again...
<Adrinael> Executing those commands will tell whether the issue is real (or test is f00ked), or fallout from other things
....
<Adrinael> What register is 0x4030 anyway, is that kosher on cnl
<ickle> 0x4030 was replace by 0x4a08
<ickle> who know what now lies behind box 0x4030
<tsa> Adrinael: cnl-drrs: (0x00004030): 0x00004000
<tsa> ickle: (0x00004a08): 0x00000000
<Adrinael> So it didn't hang?
<Adrinael> What about dump?
<tsa> Adrinael: dump whatnow
<tsa> reg dump hangs, yes
<Adrinael> If I read the output correctly, it reads PFC_CTL_1 successfully, and hangs reading PFC_WIN_POS

I'm marking as GPU HANG, please if I mistook change to correct feature. Thank you.

This machine is no longer in the lab, I will close and archive