[Bug gprof/32224] Call graph not displayed for coff-go32 (no function-flag on symbols)
https://sourceware.org/bugzilla/show_bug.cgi?id=32224 --- Comment #1 from jwjagersma at gmail dot com --- Created attachment 15724 --> https://sourceware.org/bugzilla/attachment.cgi?id=15724&action=edit proposed patch Attached patch restores the call graph. But maybe these function-symbols are expected elsewhere too, not sure. I see the associated flag for coff is DT_FCN, but this is only used in gas/config/tc-ppc.c. -- You are receiving this mail because: You are on the CC list for the bug.
[Bug binutils/18973] Support arbitrary DOS stubs in coff-go32-exe and pei-* formats
https://sourceware.org/bugzilla/show_bug.cgi?id=18973 jwjagersma at gmail dot com changed: What|Removed |Added CC||jwjagersma at gmail dot com --- Comment #1 from jwjagersma at gmail dot com --- Old report but I just stumbled upon it. For the record: binutils/bfd supports arbitrary-size stubs now for coff-go32-exe (since v2.35). I don't know the status for PE. -- You are receiving this mail because: You are on the CC list for the bug.
[Bug gprof/32224] New: Call graph not displayed for coff-go32 (no function-flag on symbols)
https://sourceware.org/bugzilla/show_bug.cgi?id=32224 Bug ID: 32224 Summary: Call graph not displayed for coff-go32 (no function-flag on symbols) Product: binutils Version: 2.43.1 Status: UNCONFIRMED Severity: normal Priority: P2 Component: gprof Assignee: unassigned at sourceware dot org Reporter: jwjagersma at gmail dot com Target Milestone: --- I'm looking to fix djgpp's profiler support, which is suffering from some bit-rot. I fixed some issues with mcount in libc, but still don't get a call graph (all samples are ). Stepping through gprof, I find that none of the symbols have the 'is_func' flag set, and it makes cg_tally() drop all the arc entries. >From what I can tell then, 'core_has_func_syms' (in core_create_function_syms(), gprof/corefile.c) should be 0 for coff-go32. Or, the assembler should emit a function flag on .text symbols. -- You are receiving this mail because: You are on the CC list for the bug.