00:49 vv222: I am currently trying to use a 32-bit Mesa build on a 64-bit host, using LD_LIBRARY_PATH and LIBGL_DRIVERS_PATH to point to my build custom path.
00:49 vv222: glxinfo fails with "libGL error: MESA-LOADER: failed to open radeonsi"
00:49 vv222: But radeonsi_dri.so is present in LIBGL_DRIVERS_PATH.
00:50 vv222: Any idea about what I am doing wrong here?
00:51 linkmauve: vv222, can you do `file` on it, to see if it’s really an i686 ELF?
00:51 vv222: Did it, got "ELF 32-bit LSB shared object" ;)
00:54 HdkR: strace to see what it is trying to open and failing on?
01:00 vv222: Well done HdkR, it was actually failing to load libunwind.so.8
01:00 vv222: Now that I installed it, glxinfo works as expected.
01:00 linkmauve: \o/
01:00 linkmauve: Good night vv222. \o_
01:01 vv222: Thanks ;)
02:27 vv222: Trying to build old (17.x) Mesa in a 32-bit chroot on a 64-bit host, I get a lot of assembler errors, mainly the three following ones repeating a lot of times:
02:27 vv222: Error: bad register name `%rax)'
02:27 vv222: Error: bad register name `%r11)'
02:27 vv222: Error: bad register name `%rip)'
02:27 vv222: This is the first time I get this kind of errors, so I don’t know how I should investigate.
02:28 anarsoul: x86 doesn't have r11, that's x86_64
02:28 vv222: Does that means the build process is trying to generate 64-bit binaries?
02:30 anarsoul: that likely means that you're trying to compile x86_64 assembly with 32-bit toolchain
02:32 vv222: Then I guess I’m missing some configure option to explicitely ask for a 32-bit build.
02:32 vv222: Right, it seems to be better with --build=i686-pc-linux-gnu
02:32 vv222: Thanks anarsoul ;)
02:33 anarsoul: np
11:10 austriancoder: are there some good tests for glDepthRange(..)? pilgit has some but with interaction with some exts
12:14 dj-death: are there naming requirements for piglit tests?
12:14 dj-death: like no space
13:21 arora: Hey jekstrand, I have been reading about layers, and I also installed the necessary drivers and configured my hardware so it works with vulkan.
13:24 arora: Shall I send you the GSoC application for a review?
17:02 imirkin: is the hdmi audio without video use-case covered by any drivers?
17:03 imirkin: vsyrjala: agd5f --^
17:19 arora: jekstrand: Shall I send you the GSoC application for a review?
17:28 imirkin: austriancoder: like what interactions did you have in mind?
17:33 arora:wonders if he'll get a reply
17:34 imirkin: my guess is you will, on monday
17:37 arora: Any idea around what time?
17:37 imirkin: i would guess morning PST
17:38 arora: Ok, I will try then, thanks
17:39 anarsoul: imirkin: or PDT? :)
17:39 imirkin: US/Pacific :p
17:40 anarsoul: unfortunately west coast is still practising daylight savings time, I'm not sure about the rest of US
17:41 imirkin: depends on the place
17:41 imirkin: arizona doesn't do it
17:42 imirkin: and like half of pennsylvania (or is it indiana) doesn't either
17:42 anarsoul: so much envy :(
17:48 arora: What does iowa follow?
17:48 imirkin: dunno, search for "time in iowa" on google
17:49 anarsoul: arora: anyway, be patient, people aren't available 24/7 on irc
17:50 arora: anarsoul: yea, I understand, that's why I tried at different times assuming he might see one of those
17:54 imirkin: arora: his client is always on
17:54 imirkin: he'll see a ping
17:59 arora: ok
19:12 austriancoder: imirkin: at the moment no interactions.. just bare glDepthRange(..) tests.
19:47 imirkin: austriancoder: i thought we had those
19:47 imirkin: in piglit, like you said
19:48 imirkin: oh, i see, you're saying the piglit ones are interactions with exts (like the one that adds 0/-z flip thing)
19:48 imirkin: austriancoder: stuff like tests/general/depth-clamp-range.c is probably good
19:49 imirkin: and tests/general/depthrange-clear.c
22:45 austriancoder: imirkin: thx will have a look
23:33 anarsoul: any ideas why x11perf may use XYImage? It hits sw fallback in glamor_put_image_gl() since apparently it can't handle it
23:33 anarsoul: s/XYImage/XYPixmap
23:37 airlied: because the test is testing xyimages?
23:37 airlied: xypixmaps rather
23:38 anarsoul: right
23:38 anarsoul: shmgetxy500 :(
23:40 airlied: run a different test :)
23:49 anarsoul: shmput500 is better, but still 640.0/sec with glamor vs 693.0/sec without glamor
23:51 anarsoul: shmget500 is 63.1/sec with glamor vs 63.5/sec without glamor
23:57 airlied: might be just hiting memory bw limits