| Summary: | dev-util/leaktracer-2.4-r1: now contains x86 asm, does not compile on ppc | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Robin Johnson <robbat2> |
| Component: | [OLD] Development | Assignee: | C++ Team [disbanded] <cpp+disabled> |
| Status: | RESOLVED FIXED | ||
| Severity: | critical | CC: | dev-zero |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
|
Description
Robin Johnson
2007-12-27 23:46:59 UTC
and even if it builds on amd64, it seems to be broken there as well because of the multistack-patch. The reason is that the programmer of the patch assumed that the return-address is at 4(%ebp), but on x86-64 it's at 8(%ebp), thus it's very likely to generate a segfault... Fixed by dropping the multi-stack patch. There's no sane way to fix it (the patch just dereferences the stack frame pointer (resp. the return address pointer of the outer-most frame) without checking and therefore generates a segfault, but the problem is that the outer-most frame should have the address 0x0 and here it has 0x1, which is kind of weird. I already fixed the assembler-code to work on x86-64 and ppc). So, if someone else can come up with a good idea on how to make this segfault away, please reopen and attach a corrected patch... |