alyssa changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard & Bifrost - - Logs - <daniels> avoiding X is a huge feature
<Lyude> any recent bifrost progress btw?
<HdkR> Today and this weekend is more compiler work and attempting to get a newer kernel running on my N2
<Lyude> ahh, cool
<alyssa> I'm eyeing MRT
<alyssa> The fragment writeout is interesting. Basically, we do a brx.write.always block for each RT, with r1.z = (target number) * 256. And just one hang at the bottom with r1.z clear.
<alyssa> Weird branch offsets, tho
<HdkR> hm
<HdkR> Backporting panfrost drm to that branch of v5.1 doesn't work due to missing shmem helper
* HdkR pulls the commit that adds that as well
<HdkR> c201?
<urjaman> that has (among other mess) all the commits that i identified panfrost to depend on back when i was rebasing it to 5.1 :P
<HdkR> I see
<urjaman> just thought it could be useful
<HdkR> Could probably cherry-pick more easily with that
<urjaman> I just basically did a rebase -i from drm-misc to 5.1 and removed everything but panfrost, but it did take like 3 tries to pick up everything that it needed
<urjaman> (drm-misc-next? or whatever... it's 4am, my memory isnt the best)
<alyssa> What's the rationale of going back to 5.1 rather than pure mainline?
<HdkR> Because I don't know what works and what doesn't
<urjaman> (for me? I'm making a patchset that i can just keep on top of a stable release... and i did the decision before i noticed that there's an unusual amount of love for the rk3288 going on right now...)
<HdkR> Looks like cherry-picking all that appliedk cleanly on to that branch. Will find out in an hour if it compiles
<urjaman> it seems google decided to update their veyron kernel (to 4.19, but anyways they're sending stuff upstream)
<HdkR> Also running in to storage limits on this 16GB eMMC
<alyssa> urjaman: Wow, huh, thought they might be stuck on 3.14 forever :p
<HdkR> Oh, it's linking, nice
vstehle has quit [Ping timeout: 245 seconds]
* HdkR crosses fingers and hope it boots
<HdkR> Dang
_whitelogger has joined #panfrost
<HdkR> I'll get a USB flash drive in an hour or so and do a rebuild
tlwoerner has quit [Quit: Leaving]
tlwoerner has joined #panfrost
daniels has quit [Ping timeout: 250 seconds]
daniels has joined #panfrost
<HdkR> Fresh new 32GB USB flash drive for abusing
<HdkR> Time to mash `sudo apt-get remove unattended-upgrades` until that process goes away
vstehle has joined #panfrost
TheKit has quit [Ping timeout: 246 seconds]
TheKit has joined #panfrost
<HdkR> Good to know
<HdkR> I'll grab that branch and cherry-pick the bits required for panfrost drm on to that one instead
paulk-leonov has quit [Excess Flood]
paulk-leonov has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
paulk-leonov has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
mifritscher_ has quit [Quit: Quit]
paulk-leonov has joined #panfrost
mifritscher has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
paulk-leonov has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
paulk-leonov has joined #panfrost
<HdkR> gah. /boot partition filled up
<hanetzer> HdkR: heh. device?
<HdkR> N2's eMMC
<HdkR> resizing and fixing
<HdkR> 128MB just isn't large enough
<HdkR> er, guess it was uboot partition, not actually /boot
<hanetzer> I wonder if you can craft coreboot to run from emmc...
<HdkR> beh, it just keeps faultingand restarting in uboot
<HdkR> Wonder if this is a defconfig issue
<HdkR> Waiting for the kernel to compile bums me out
<HdkR> Takes 37 seconds for a defconfig on my desktop
<hanetzer> heh.
<hanetzer> I use distcc heavily myself :)
<hanetzer> basically I do an emerge pass with distcc-pump, with --keep-going, then distcc alone, then a final emerge without to catch the packages that failed because they don't support distcc in some manner
<HdkR> Tried that before and it seemed like the preprocessor bottlenecked pretty hard
<bbrezillon> anarsoul: yes, I know, was just pointing it as a quick fix for ezequiel
stikonas has joined #panfrost
<HdkR> Finally linked
<HdkR> So a bit over a half hour to compile?
<HdkR> bleh. Another fail
<HdkR> What a waste of time
<HdkR> Well I was going to do more work once I got the kernel booting but since that isn't the case, I'm just going to stop for the night
cwabbott has quit [Ping timeout: 255 seconds]
cwabbott has joined #panfrost
travankor1 has quit [Ping timeout: 264 seconds]
travankor1 has joined #panfrost
cwabbott_ has joined #panfrost
cwabbott_ is now known as cwabbott
cwabbott has quit [Ping timeout: 258 seconds]
MoeIcenowy has quit [Quit: ZNC 1.6.5+deb1+deb9u1 -]
MoeIcenowy has joined #panfrost
narmstrong has quit []
narmstrong has joined #panfrost
_whitelogger has quit [Remote host closed the connection]
_whitelogger has joined #panfrost
<Lyude> HdkR: let me know if you make progress with the mainline kernel, would be nice to get my odroid n2 working as well
Elpaulo has joined #panfrost
<Lyude> (be back in a bit, updating server to f30)
Lyude has quit [Quit: WeeChat 2.2]
<alyssa> narmstrong: Any timeline on mainline support (for both kernel + u-boot)? I'm excited! :)
<narmstrong> alyssa: hopefully 5.3 and a little bit later for u-boot, to sync the DT
<narmstrong> but it won't need a lot of off-tree patches on top of 5.2
<narmstrong> most of it went into 5.2
<alyssa> narmstrong: That's wonderful to hear, thank you :)
Lyude has joined #panfrost
belgin has joined #panfrost
<HdkR> narmstrong: Not sure what the failure was. I was mostly just copy and pasting a bunch of commands from a post from memeka
belgin has quit [Quit: Leaving]
pendarsoheil has quit [Remote host closed the connection]
anarsoul|c has joined #panfrost
<alyssa> I think I need to implement a safety mode to debug these DEs.
<HdkR> pfft, nobody needs a DE
<alyssa> HdkR: Well, WM
<hanetzer> HdkR: I assume u-boot is in the emmc?
<HdkR> Correct
<hanetzer> any spi flash at all?
<HdkR> Nop. I'm using a USB flash drive as external storage
<hanetzer> gotcher.
* hanetzer once again wonders about getting coreboot to run from emmc
<hanetzer> won't try it with the hardware I currently have, would semi-brick it if something goes wrong
<hanetzer> I once semi-bricked a c201/speedy by trying u-boot on the emmc
NeuroScr has joined #panfrost
<alyssa> Just did a nice refactor wrt blending
<alyssa> So on Midgard, we now understand how blending works with MRT
<alyssa> And on Bifrost, we now have blending that doesn't totally break Midgard
<alyssa> One HUGE source of compile-time differences: eliminated!
mateo` has quit [Ping timeout: 246 seconds]
mateo` has joined #panfrost
unoccupied has joined #panfrost
stikonas has quit [Remote host closed the connection]
jolan has quit [Quit: leaving]