05:26 <zenmetsu> sjoerd: which mali binary version should i use with your branch? r4p0-02rel0 gives me egl initialization failure, while r4p1-00rel0 causes a kernel NULL pointer panic

06:16 <zenmetsu> sjoerd: actually i had that backwards, i get the null pointer panic with r4p0-02rel0. the kernel sources appear to be built against r4p0-02rel0 as well :/

06:45 <sjoerd> zenmetsu: I'm using the mali binaries from chrome os

06:45 <sjoerd> which should be 0.r4p0+02rel0

06:46 <sjoerd> which X driver?

06:48 <zenmetsu> i'm following the malideveloper guide more or less, using the armsoc driver that i compiled

06:50 <zenmetsu> really it is just libmali.so that changes, so i can md5sum the one on my chromeos partition to see if it is the same i suppose

06:52 <sjoerd> you build armsoc with --with-drmmode=exynos right?

06:52 <sjoerd> I guess you get to work out why EglInit fails, i can't really think of anything obvious ofhad

06:53 <zenmetsu> 7110ce3a54c149557a98fcc46855f84fe /usr/lib/opengl/xorg-x11/lib/libmali.so

06:53 <zenmetsu> and yes, build with --with-drmmode=exynos

06:54 <zenmetsu> i'll have to boot to chromeos since i cannot mount the partition from linux, but that is the md5sum of my libmali.so

06:54 <zenmetsu> if yours differs, then i'll have to snag the one from chromeos :)

07:09 <zenmetsu> should i give it a shot without CONFIG_MALI_PLATFORM_FAKE=y?

08:21 <zenmetsu> sjoerd: EglInit happened when i tried using r4p1-00rel0 userland binaries. The r4p0-02rel0 driver from chromeos, as well as the one from the malideveloper site, both cause a kernel panic when EglInit is attempted

12:19 <zenmetsu> sjoerd: was config problem, i used exynos defconfig and userspace gles works well. i've got a permissions issue somewhere as it only works for root, but this is enough to get me going. thanks :D

12:19 <sjoerd> ooh right

12:19 <sjoerd> yeah /dev/mali0 needs tobe accessible by your user

12:20 <zenmetsu> sweet, that did it. ownership was right, mode was not rw for group :)

12:21 <zenmetsu> derp moment