Summary: | sys-devel/gcc-4.5.0 create invalid binaries with -fdelayed-branch on parisc | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Guy Martin (RETIRED) <gmsoft> |
Component: | [OLD] GCC Porting | Assignee: | HPPA Porters <hppa> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | toolchain |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | HPPA | ||
OS: | Linux | ||
URL: | http://thread.gmane.org/gmane.linux.ports.parisc/2903 | ||
See Also: | https://gcc.gnu.org/bugzilla/show_bug.cgi?id=44469 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | 365935 | ||
Bug Blocks: | 915000 | ||
Attachments: | Stripped down version of gengenrtl which triggers the issue |
Description
Guy Martin (RETIRED)
2010-06-05 11:41:17 UTC
Created attachment 234185 [details]
Stripped down version of gengenrtl which triggers the issue
Added stripped down version of gengenrtl.c to repro the issue.
Observed behavior :
gcc -O2 -> Segv
gcc -O2 -fno-delayed-branch -> OK
gcc -O1 -> OK
gcc -O1 -freorder-blocks -> Segv
gcc -O1 -freorder-blocks -fno-delayed-branch -> OK
gcc -O0 -freorder-blocks -fdelayed-branch -> OK
This seems to be similar to PR 44469 : http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44469 The testcase is quite similar to the one I've attached but the workaround doesn't seem to help. I'm doublechecking right now. What's the verdict? According to the upstream bug report, it should be in 4.5.3, which was released months ago but isn't in the tree yet. Fixed in 4.5.3. |