alyssa changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard & Bifrost - Logs https://freenode.irclog.whitequark.org/panfrost - <daniels> avoiding X is a huge feature
<enunes> I'm not aware of it being an issue in practice
<enunes> HdkR: at least on my pinebook (non-pro) I have used this including grub2 for a while, if you're actually going to use it I might recommend the Workstation variant rather than the Server, as it includes the desktop stack and defaults to ext4 instead of xfs
<HdkR> Fancy
<enunes> heh :) yeah
<HdkR> I presume it's quite a bit easier to build grub on a native machine rather than cross-compiling is the issue?
<anarsoul> yes
<anarsoul> so far I cross compile only kernel, u-boot and atf
<anarsoul> not everyone cares about cross-compilation, so it's usually broken here and there
<anarsoul> and life is too short to fix it all
* HdkR tries native build
<HdkR> It's not like I don't have ARMv8 devices around
<stikonas> well, atf on RK3399 does not have fuly native compilation, part of it needs 32bit arm compiler
<anarsoul> yeah, that's fine
<HdkR> yea, built natively fine. That's annoying
raster has quit [Quit: Gettin' stinky!]
chewitt has joined #panfrost
<HdkR> Hm, that grub version boots but then resets the device
kaspter has quit [Ping timeout: 240 seconds]
camus has joined #panfrost
camus is now known as kaspter
<HdkR> Also curiously it hecks up the screen resolution while the kernel was fine with it
warpme_ has quit [Quit: Connection closed for inactivity]
kaspter has quit [Ping timeout: 276 seconds]
kaspter has joined #panfrost
pH5 has quit [Quit: Lost terminal]
vstehle has quit [Ping timeout: 265 seconds]
stikonas has quit [Remote host closed the connection]
<alyssa> 3
<HdkR> 2
megi has quit [Ping timeout: 276 seconds]
<alyssa> 1
<alyssa> Okay what's the best way to model presolved nodes in lcra. since what I have now is not great.
<alyssa> Actually it's fine just need some reordering.
_whitelogger has joined #panfrost
<alyssa> Although I can still iterate out-of-priority for the tests themselves, those are order-independent. So cache perf is still fine as long as it's in the right row-major/column-major form (not sure which atm, let's see)
TheKit has quit [Ping timeout: 276 seconds]
krh has quit [Ping timeout: 245 seconds]
TheKit has joined #panfrost
_whitelogger has joined #panfrost
davidlt has joined #panfrost
_whitelogger has joined #panfrost
vstehle has joined #panfrost
_whitelogger has joined #panfrost
NeuroScr has quit [Ping timeout: 265 seconds]
_whitelogger has joined #panfrost
_whitelogger has joined #panfrost
fysa has joined #panfrost
fysa has quit [Ping timeout: 265 seconds]
guillaume_g has joined #panfrost
yann has quit [Ping timeout: 268 seconds]
davidlt has quit [Ping timeout: 265 seconds]
Kwiboo has quit [Ping timeout: 246 seconds]
fysa has joined #panfrost
fysa has quit [Ping timeout: 240 seconds]
yann has joined #panfrost
warpme_ has joined #panfrost
megi has joined #panfrost
karolherbst has quit [Quit: duh 🐧]
fysa has joined #panfrost
<warpme_> i see ongoing work with t720 which is fantastic. As progress on t720 - now i started to see the same white screen issue on t720 like I have since some time on t820 (t820 was working ok but at some time it started to withe-screen). This smells to me like the same issue which makes on t720/t820 white screen. As it is regression (in past I had working t820 (for GLES UI)) - maybe it is time to look on white screen issue
<warpme_> more deeply?
fysa has quit [Ping timeout: 240 seconds]
raster has joined #panfrost
raster has quit [Remote host closed the connection]
maccraft123 has joined #panfrost
<tomeu> warpme_: what's this white screen?
raster has joined #panfrost
<warpme_> it is just whole monitor white screen of mythtv UI (it uses GLES from mythtv Qt app). All works nicely on mali450 and also was worked ok on t820 in past - but stopped. Month ago or so I bisected bbrezillin commit responsible for regression - so t820 started woking OK at that time - but after sometime it again returns to white screen. Browsing UI gives sometimes black squares so it looks like polygon issue (just
<warpme_> guessing :-p)
karolherbst has joined #panfrost
<raster> warpme_: was this the commit that was optimizing the wallpapering for partial update stuff? ore related?
<warpme_> ough - don't remember - it was month ago....
<raster> :|
<warpme_> at that time revert of c9bebae2877e55cdcd94f9f9f3f6805238caeb28 was solution
<tomeu> surprising
<tomeu> warpme_: what is this SoC?
<warpme_> t820: s912; t720:H6
<tomeu> ok, in a week from now I think we'll have improved substantially t720 support
<tomeu> warpme_: if in the meantime you could bisect the second breakage, that would be great
<bbrezillon> raster: no it wasn't
<bbrezillon> which reminds me that this problem is still to be solved
<raster> bbrezillon: actualy interestingly your revert didnt fix the issue for me.. i havent gotten time to dig into why though
<warpme_> fantastic. will it propagate also to t820 white-screen (rather not hopping for that...)
<bbrezillon> warpme_: the commit you're mentioning has been reverted, but as I stated last time, I don't think it was the root cause of your issue
<bbrezillon> raster: my revert has been reverted :)
<warpme_> bbrezillon: right. Also I think so!
<warpme_> :-)
<bbrezillon> because lima started to use the partial_update() infra in the meantime
<raster> bbrezillon: oh... so we're back to being broken then?
<bbrezillon> and it broke the mali build
<raster> ugh
<raster> :(
<bbrezillon> I pinged them several times
<bbrezillon> I think the next step is to simply disable the feature on panfrost
<raster> lima might be broken too then
<raster> i don't have any mali 400 hw around to check with though... well not at hand
<warpme_> just FYI: mali450 works perfectly fine for me (on exactly the same SD card)
<bbrezillon> raster: I think it's broken too
<bbrezillon> warpme_: mali450 does not use the same driver :P
<raster> bbrezillon: that patch is already in mesa or i should apply it?
<bbrezillon> raster: you should apply it
<bbrezillon> and ideally post it :)
<raster> cool
<raster> post it? like submit it to mesa mainline?
<bbrezillon> yes
<raster> shouldn't you? :) as the guy who wrote it?
<raster> or... you mean me because i can see the problem easily and confirm it fixes it?
<bbrezillon> or I can do it, but I need you to test :)
<raster> let me un-ddk my rk3399
<warpme_> guys:tell me when I should I give test tun on mali450/t720/t820
<warpme_> btw: is it possible to unblack-list t720?
<raster> warpme_: i'd love to know if e+efl work on the 450
<raster> (e in wayland mode specifically)
<warpme_> raster: do you mean current meas git on 450?
<raster> well yeah
<raster> mesa git and possible with the above patch bbrezillon pasted a link to above
<warpme_> ok: just building :-p
<raster> and enlightenment+efl
<bbrezillon> warpme_: I don't think that will fix your problem
<raster> enlightenment shows off the partial update issue in mesa pretty easily
<bbrezillon> those are 2 different issues
maccraft123 has quit [Ping timeout: 240 seconds]
<raster> hurry up rockpro... build that mesa!
<bbrezillon> raster: can you join #lima
<bbrezillon> ?
<raster> sure
mearon has quit [Ping timeout: 244 seconds]
<raster> bbrezillon: confirmed - your patch gets my screen back to rendering right :)
<raster> yay!
<warpme_> current mesa git: H6: still white screen; mali450:works ok
<raster> no-more-black-rect-junk.com
fysa has joined #panfrost
<tomeu> warpme_: I hope to be able to unbacklist in a 1-4 weeks
<warpme_> super. one patch less for me at testing then :-)
fysa has quit [Ping timeout: 276 seconds]
mearon has joined #panfrost
maccraft123 has joined #panfrost
maccraft123 has quit [Quit: WeeChat 2.6]
vstehle has quit [Ping timeout: 265 seconds]
davidlt has joined #panfrost
maccraft123 has joined #panfrost
maccraft123 has quit [Ping timeout: 245 seconds]
fysa has joined #panfrost
fysa has quit [Ping timeout: 246 seconds]
<raster> 👍 :)
fysa has joined #panfrost
megi has quit [Ping timeout: 268 seconds]
<paulk-leonov> ooh, so RK3308 has no GPU, interesting
fysa has quit [Ping timeout: 264 seconds]
Kwiboo has joined #panfrost
gnufan has joined #panfrost
<alyssa> paulk-leonov: Looks like it's for audio-only stuff mostly?
fysa has joined #panfrost
megi has joined #panfrost
fysa has quit [Ping timeout: 276 seconds]
gnufan has quit [Ping timeout: 276 seconds]
karolherbst has quit [Ping timeout: 276 seconds]
fysa has joined #panfrost
guillaume_g has quit [Quit: Konversation terminated!]
ph5 has joined #panfrost
fysa has quit [Read error: Connection reset by peer]
fysa has joined #panfrost
<mmind00> paulk-leonov alyssa: or vacuum cleaners https://twitter.com/IloveRockchip/status/1191991278176624640
<mmind00> although I can't really understand who would build a vacuum cleaner without gpu :-D
<paulk-leonov> haha :)
<mmind00> paulk-leonov: rk3326 (not 3328) does have a G31 bifrost GPU though
<paulk-leonov> that's px30 isn't it?
<paulk-leonov> mmind00: very nice to see your work on that one btw ;)
<mmind00> yep ... nearly same build ... I think one of them is missing things like the second display controller
<mmind00> paulk-leonov: now if only someone would dive into bifrost support ;-)
<paulk-leonov> fingers crossed, it'll happen :)
<mmind00> but I guess not "fast" enough and it looks like I'll need to look into that binary-monster at some point
<paulk-leonov> mmind00: truth be told, it's on my tasks list for an upcoming project too :(
<paulk-leonov> alyssa please don't take this as a betrayal :p
<mmind00> "mali-binary-users-anonymous" or so
<paulk-leonov> yeah maybe we'll need some group therapy to get over the trauma
<paulk-leonov> well, could be worse I guess
<paulk-leonov> could be powervr :p
* mmind00 looks to his rk3368 board
karolherbst has joined #panfrost
* urjaman looks at the pandora on his table ...
<paulk-leonov> urjaman: the omap3 one?
<urjaman> affirmative
<paulk-leonov> pretty cool!
<paulk-leonov> mmind00: didn't know rockchip ever produced a soc with powervr
<paulk-leonov> happened to allwinner too
<paulk-leonov> sad story
<mmind00> yep ... sadly they did ... rk3368 is pretty cool, except for the gpu selection
<urjaman> i mean, even intel did that :P
<mmind00> thankfully it seems to have stayed an exception
<paulk-leonov> urjaman: tell me about it, I've been working on gma500 recently :(
<paulk-leonov> but rk3368 has this weird ddr resume blob
yann has quit [Ping timeout: 240 seconds]
vstehle has joined #panfrost
<alyssa> Pfff
megi has quit [Quit: WeeChat 2.6]
yann has joined #panfrost
stikonas has joined #panfrost
raster has quit [Quit: Gettin' stinky!]
gcl has quit [Ping timeout: 240 seconds]
gcl has joined #panfrost
NeuroScr has joined #panfrost
maccraft123 has joined #panfrost
adjtm has quit [Read error: Connection reset by peer]
adjtm_ has joined #panfrost
davidlt has quit [Remote host closed the connection]
davidlt has joined #panfrost
davidlt has quit [Ping timeout: 268 seconds]
ph5 has quit [Ping timeout: 276 seconds]
maccraft123 has quit [Ping timeout: 245 seconds]
Kwiboo has quit [Quit: .]
fysa has quit [Remote host closed the connection]
Kwiboo has joined #panfrost