--- Additional Comments From amodra at gmail dot com 2010-05-13 10:25
---
Created an attachment (id=4790)
-- (http://sourceware.org/bugzilla/attachment.cgi?id=4790action=view)
Alternate patch to avoid the abort
Nick, I prefer to leave in the abort to catch other cases we may not have
--- Additional Comments From amodra at gmail dot com 2010-05-13 10:22
---
No, what you're trying to do is not supported, and I think it would not be easy
to implement.
--
What|Removed |Added
Dear binutils maintainers,
I was recently profiling some algorithm improvements in the Bos
Wars game, using gcc -pg and gprof on amd64. To my surprise,
when I ran the same input through the game several times, gprof
displayed different timings _and_ call counts each time, even
though section 6.1
Hi Alan,
Nick, I prefer to leave in the abort to catch other cases we may not have
considered. Instead, undo the change in type to bfd_link_hash_new we made in
bfd_elf_record_link_assignment.
Ok that makes a lot more sense.
Cheers
Nick
___
bra.s to next instruction is invalid because it would be the same opcode as
bra.
$ cat bug.s
bra.s next
next: nop
This is correctly rejected for ELF targets:
$ m68k-elf-as bug.s -o bug.o
bug.s: Assembler messages:
bug.s:1: Error: invalid byte branch offset
However it is accepted
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2010-05-14
03:15 ---
Subject: Bug 11583
CVSROOT:/cvs/src
Module name:src
Changes by: amo...@sourceware.org 2010-05-14 03:15:20
Modified files:
ld : ChangeLog ldexp.c
Log message:
--- Additional Comments From amodra at gmail dot com 2010-05-14 03:54
---
Take a look at vmas shown by objdump -h. They are different to those shown by
your program. Find out why by looking at the source to objdump.
--
What|Removed |Added