| Summary: | media-libs/mesa-17.0.0 fails on (compile phase): | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | tman <cornicx> |
| Component: | Current packages | Assignee: | Gentoo X packagers <x11> |
| Status: | RESOLVED OBSOLETE | ||
| Severity: | normal | CC: | jstein |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: |
complete build log
emerge --info '=media-libs/mesa-17.0.0::gentoo' |
||
|
Description
tman
2017-02-20 10:24:17 UTC
Attach full build log. i have a issue with mesa build too /tmp/portage/media-libs/mesa-17.0.0/work/mesa-17.0.0/src/compiler/nir/nir_loop_analyze.c: In function 'find_trip_count': /tmp/portage/media-libs/mesa-17.0.0/work/mesa-17.0.0/src/compiler/nir/nir_loop_analyze.c:631:1: internal compiler error: Segmentation fault } ^ Created attachment 464770 [details]
complete build log
Created attachment 464772 [details]
emerge --info '=media-libs/mesa-17.0.0::gentoo'
builds ok with gcc-4.8.5 and fails with gcc-4.9.4 and gcc-5.4.0-r3 I cannot reproduce. The initial report seems different from comment #2. Is it possible that the compile is exhausting the memory in your system, and causing the out-of-memory failures? (In reply to Matt Turner from comment #6) > I cannot reproduce. > > The initial report seems different from comment #2. > > Is it possible that the compile is exhausting the memory in your system, and > causing the out-of-memory failures? i have 7gb free memory, which is much more consumptions on mesa compile time. also i successfully build qtwebkit or qtwebengine, which required much more free memory. do you have any ideas about initial report? (In reply to sango from comment #7) > do you have any ideas about initial report? I just realized *how* different your report is from the original. *Please* do not hijack bugs like this in the future. It's not just annoying for the original reporter to have their bug taken over, but it's confusing for the person (i.e., me) trying to determine what the problem is. I'm very surprised you didn't lead with "hey, I'm using some strange options in CFLAGS, maybe that's it?" Yes, I can reproduce your failure with gcc-4.9.4 and CFLAGS="-O2 -pipe -fomit-frame-pointer -march=i686 -msse -msse2 -msse3 -mssse3 -m8bit-idiv -mmmx" The good news is that gcc-6.3.0 successfully compiles Mesa with those CFLAGS. I'll see if I can narrow down the problem a bit more. First guess was correct: -m8bit-idiv is the culprit. I can reproduce the failure with just CFLAGS="-O2 -pipe -march=i686 -m8bit-idiv" |