Bug 18983
Summary: | Gcc crashes when compiling glibc | ||
---|---|---|---|
Product: | [Retired] Red Hat Linux | Reporter: | jfm2 |
Component: | gcc | Assignee: | Jakub Jelinek <jakub> |
Status: | CLOSED NOTABUG | QA Contact: | David Lawrence <dkl> |
Severity: | medium | Docs Contact: | |
Priority: | medium | ||
Version: | 7.0 | ||
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | i386 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2000-10-13 08:20:52 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
jfm2
2000-10-12 18:00:46 UTC
Details please, which file does it crash on, which options passed to it, preprocessed file (best run rpm -ba glibc.spec 2>&1 | tee glibc.log, then cut'n'paste the command which caused the gcc to crash, add -save-temps -v to it and send me its output and the .i file). I cannot reproduce this myself and I'm building glibc with these flags (both i386 and i686 targets) almost daily for a couple of months. Actually, why are you building i686 glibc on K6/2? K6/2 AFAIK does not have conditional move instructions, so if the SIGILL comes when glibc tries to run rpcgen or similar program, it can crash on K6/2 if there are any cmoves generated (which gcc can legally use if -march=i686). During glibc bootstrap after the main library is compiled glibc executes binaries using the newly built dynamic linker and library. It turned to be a mistake. Somewhere in the the building of glibc an auxilary program is byuilt and executed. Since that executable was generated with the PPro instruction set and I was on a K6/2 it core dumped and I thought it was the compiler who had core dumped. |