alyssa changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard & Bifrost - https://gitlab.freedesktop.org/panfrost - Logs https://freenode.irclog.whitequark.org/panfrost - Discord Discard
<alyssa> :eyes:
<Lyude> i have a powerful feeling about this patch
<alyssa> Merged
<marcodiego> hmmm maybe rounding errors are breaking things?
<alyssa> Powerful?
<Lyude> alyssa: i don't know, rounding errors usually mean something terrible
<Lyude> so not having them is probably fixing something somewhere :)
<Lyude> marcodiego: unfortunately that isn't going to fix my current issues getting kmscube to run :p
<alyssa> Lyude: Pff, fair :)
<Lyude> btw alyssa, just a side commebtn
<Lyude> we should probably get rid of like, most of the #ifdef stuff
<Lyude> when we upstream this I mean
<alyssa> Mm
<alyssa> It feels professional
<alyssa> I learned it from the old kbase
<alyssa> ;)
<Lyude> oh lord
<Lyude> TIL aarch64 does not appear to have hw watchpoints
<Lyude> as enabling a watchpoint in gdb slows everything down /tremendously/
<HdkR> Lyude: That's fixed with MTE or PAC, I forget which
<Lyude> HdkR: what are those?
<HdkR> Extensions that'll appear in ARMv8.5 or so
<HdkR> There was an LLVM dev meeting talk about it
<Lyude> ah
marcodiego has quit [Quit: Leaving]
<Lyude> tomeu: on the initial context creation in https://gitlab.freedesktop.org/tomeu/mesa/blob/winsys/src/gallium/drivers/panfrost/pan_context.c#L2929 I'm getting a segfault at https://paste.fedoraproject.org/paste/~3cRLsaOL~50yM~nB4ZoEg . Looking it seems that the problem is https://paste.fedoraproject.org/paste/SCeG5dWTYIy53e~I0gvWig that there's no nr_cbufs when panfrost_new_frag_framebuffer() is called, but
<Lyude> for the life of me I can't seem to figure out where ctx->pipe_framebuffer is supposed to be set at all o_O
<Lyude> am I missing something here?
<alyssa> Maybe tomeu chanegd it while changing winsys stuff?
<Lyude> alyssa: line in question was changed by tomeu which is why I'm asking
<Lyude> according to git
<tomeu> Lyude: in panfrost_set_framebuffer_state
<tomeu> narmstrong: very cool, I actually get a big part of the screen covered with black rectangles
<tomeu> also with weston
<tomeu> narmstrong: have you tried with weston as well?
<narmstrong> tomeu: I’ll try today !
<tomeu> awesome!
<tomeu> narmstrong: what did you have to do to get that? only changes to the DT or to the kernel outside kbase?
<tomeu> Lyude: I'm quite surprised at me not getting a SIGSEGV there when trying to dereference 0x8
<narmstrong> tomeu: only switched kbase to always-on
<tomeu> cool, makes sense as a workaround
<narmstrong> I’ll have a look at the amlogic version, they maybe have some code fixing that
pH5 has quit [Quit: bye]
pH5 has joined #panfrost
rhyskidd has quit [Quit: rhyskidd]
rhyskidd has joined #panfrost
<narmstrong> tomeu: weston launches !
<tomeu> narmstrong: any visual glitches?
<tomeu> narmstrong: and, what about SHM clients such as weston-flowers?
<narmstrong> tomeu: I failed to have a mouse, I will test further now
<narmstrong> I will try all these right now if I get a mouse and keyboard
<tomeu> you can run one from the console if you set XDG_RUNTIME to the appropriate value
<tomeu> narmstrong: ^
<narmstrong> yeah moving a mouse breaks everything !
<tomeu> narmstrong: same thing here
<tomeu> my next project is improving error reporting, btw
<tomeu> it should also fix command flushing which may be related to this
rhyskidd has quit [Quit: rhyskidd]
rhyskidd has joined #panfrost
rhyskidd has quit [Quit: rhyskidd]
rhyskidd has joined #panfrost
rhyskidd has quit [Quit: rhyskidd]
rhyskidd has joined #panfrost
<narmstrong> Yep, I managed to fix the power error, on the Amlogic Platform it needs a specific power_on sequence and a specific reset sequence with some external reset controls
<narmstrong> now I only have the `DATA_INVALID_FAULT` error when rendering
rhyskidd has quit [Quit: rhyskidd]
rhyskidd has joined #panfrost
tgall_foo has joined #panfrost
afaerber has joined #panfrost
<tomeu> Lyude: maybe we should consolidate our repos and branches a bit: https://gitlab.freedesktop.org/lyudess/panfrost-mesa/merge_requests/1
rhyskidd has quit [Quit: rhyskidd]
rhyskidd has joined #panfrost
<narmstrong> tomeu: yep, since the devicetree platform runs ok, i will cleanup and start from a fresh one
<narmstrong> tomeu: but the tricky point is that the amlogic platform need a custom power on and reset code and this define
<narmstrong> #define PLATFORM_POWER_DOWN_ONLY (1) instead of 0
<narmstrong> And it’s a compile time switch
<narmstrong> Maybe this could go as a config
<tomeu> makes sense
rhyskidd has quit [Quit: rhyskidd]
<narmstrong> I'll refresh the mali_kbase in the next days, I'll ping you when I have something you could take !
rhyskidd has joined #panfrost
rhyskidd has quit [Remote host closed the connection]
<tomeu> narmstrong, Lyude: my thinking right now is that we can keep submitting MRs to Lyude's repo until the winsys is mature enough so alyssa can switch to it
<tomeu> at which point we review and merge into the panfrost repo
<tomeu> and from now on we submit MRs and review each other's code
<narmstrong> tomeu: seems good for me, I'll push the meson drm winsys renderonly driver as MR
<tomeu> awesome
<tomeu> alyssa: what do you think? I'm not sure what test apps you need to run to continue your work
indy has quit [Read error: Connection reset by peer]
indy has joined #panfrost
<alyssa> Lyude: gotcha
<alyssa> tomeu: Not sure. As I've said, I've been distracted from Panfrost for a myriad of reasons as of late, but the good news is that you don't have to worry about breaking my setup since mine will be outdated by the time I have time to frost pans again ;)
<alyssa> (Real-life stuff piling up and a Secret Project)
<tomeu> alyssa: ok, hopefully you will find time for panfrost later :)
<alyssa> For sure :)
<tomeu> alyssa: is Lyude maintainer in gitlab?
<alyssa> She should be?
<alyssa> Just remember, last time I was working on a Secret Project, it was Panfrost, so... ;)
<tomeu> hehe, will try hard to not guess :)
<alyssa> tomeu: Guesses welcome :P
<alyssa> Though I should add it's not really secret, I just haven't been making noise about it yet until it's ready for people to test
<tomeu> alyssa: hopefully not a IMG driver :p
<alyssa> Ew.
<alyssa> No RE'ing involved :p
<alyssa> (Well, not directly anyway)
rhyskidd has joined #panfrost
BenG83 has joined #panfrost
rhyskidd has quit [Quit: rhyskidd]
rhyskidd has joined #panfrost
<Lyude> tomeu: seems like a good idea to me
<Lyude> (re: 09:03 <freenode#panfrost> <tomeu> Lyude: maybe we should consolidate our repos and branches a bit: https://gitlab.freedesktop.org/lyudess/panfrost-mesa/merge_requests/1 )
<Lyude> also will look at that when I get to work
BenG83 has quit [Quit: Leaving]
BenG83 has joined #panfrost
rhyskidd has quit [Quit: rhyskidd]
rhyskidd has joined #panfrost
BenG83 has quit [Quit: Leaving]
pH5 has quit [Quit: bye]
BenG83 has joined #panfrost
pH5 has joined #panfrost
rhyskidd has quit [Quit: rhyskidd]
rhyskidd has joined #panfrost
cwabbott_ has joined #panfrost
cwabbott has quit [Ping timeout: 240 seconds]
cwabbott_ is now known as cwabbott
anarsoul|2 has joined #panfrost
cwabbott_ has joined #panfrost
cwabbott has quit [Ping timeout: 264 seconds]
cwabbott_ is now known as cwabbott
rhyskidd has quit [Quit: rhyskidd]
rhyskidd has joined #panfrost
rhyskidd has quit [Quit: rhyskidd]
<Lyude> btw tomeu and narmstrong: I wouldn't mind just moving wip/mali-uapi-update to the main panfrost group on gitlab
rhyskidd has joined #panfrost
mjourdan has left #panfrost [#panfrost]