Bug 20532
| Summary: | optimisations bug | ||
|---|---|---|---|
| Product: | [Retired] Red Hat Raw Hide | Reporter: | Chmouel Boudjnah <cboudjnah> |
| Component: | gcc | Assignee: | Jakub Jelinek <jakub> |
| Status: | CLOSED RAWHIDE | QA Contact: | David Lawrence <dkl> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 1.0 | ||
| 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: | 2000-11-13 09:40:12 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: | |||
Will be fixed in gcc-2.96-64, fix was already commited into CVS head. |
The following is a very simplified pb that appeared in ruby. Any attempt to simplify the example makes the bug disappear. Minimal optimization for the bug to appear: -O2 -fomit-frame-pointer The bug is "x" being put in the stack for "f" as the second argument, not the first one ("f" is a tail-call) -------------------------------------------------------------------------------- void raise(long) __attribute__ ((noreturn)); long f(double); long g(long i, double x) { if (x < 0) raise(1); return f(x); } -------------------------------------------------------------------------------- The test prints "0" when it should print "2" -------------------------------------------------------------------------------- void raise (long e) { exit(0); } long f(double i) { printf("%g\n", i); return 0; } long g(long i, double x); main() { g(0, 2.0); }