[mips] relative FDE relocations without DW_EH_PE_signed
Richard Sandiford
rdsandiford@googlemail.com
Wed Oct 13 21:11:00 GMT 2010
More information about the Binutils mailing list
Wed Oct 13 21:11:00 GMT 2010
- Previous message (by thread): [mips] relative FDE relocations without DW_EH_PE_signed
- Next message (by thread): [PATCH] Fix oopsie in ld-plugin tests.
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Mingjie Xing <mingjie.xing@gmail.com> writes: > 2010/10/13 Alan Modra <amodra@gmail.com>: >> On Wed, Oct 13, 2010 at 11:28:47AM +0800, Mingjie Xing wrote: >>> When bfd starts to write section in _bfd_elf_write_section_eh_frame, >>> the addresses will be converted as negative values, for example >>> 0xfffeffe4. But the fde encoding is not DW_EH_PE_signed at this time. >>> This will cause the address be dumped as an unsigned value, which >>> looks like pc=100020000. >>> >>> Is it a bug? >> >> I think your eh_frame data is OK. The bug is in readelf. When >> readelf is built on a 64-bit host, it can display more address bits >> than are relevant for a 32-bit target. > > Indeed. My host is ia64. The date type of the address is 'long' which > means 64-bit on it. > > I have no idea about that if this is a bug or not. :-) Yeah, it's a readelf bug, like Alan says. I see the same problem on x86_64, but I'm afraid I've never got around to fixing it. Richard (shame-faced)
- Previous message (by thread): [mips] relative FDE relocations without DW_EH_PE_signed
- Next message (by thread): [PATCH] Fix oopsie in ld-plugin tests.
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Binutils mailing list