This is the mail archive of the
gdb@sourceware.org
mailing list for the GDB project.
Re: CORE_ADDR representation
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: gdb at sourceware dot org
- Date: Thu, 18 Feb 2010 11:34:04 +0100
- Subject: Re: CORE_ADDR representation
- References: <20100218044416.GA19485@caradoc.them.org>
On Thu, 18 Feb 2010 05:44:19 +0100, Daniel Jacobowitz wrote:
> struct section_addr_info *
> build_section_addr_info_from_objfile (const struct objfile *objfile)
> {
> ...
> CORE_ADDR mask = CORE_ADDR_MAX;
>
> if (addr_bit < (sizeof (CORE_ADDR) * HOST_CHAR_BIT))
> mask = ((CORE_ADDR) 1 << addr_bit) - 1;
> ...
> sap->other[i].addr = (bfd_get_section_vma (objfile->obfd, sec)
> + objfile->section_offsets->offsets[i]) & mask;
>
> This truncates the high bits. MIPS sign-extends pointers, even
> internally in CORE_ADDR, and this results in separate debug info files
> for MIPS executables being relocated off to la-la land.
If we follow Mark Kettenis's suggestion on 64bit arithmetics for 32bit
inferiors.
Re: [patch] bfd/: bfd_elf_bfd_from_remote_memory 32bit &= 0xffffffff
http://sourceware.org/ml/gdb-patches/2010-02/msg00286.html
which should work now with checked-in
[patch] Fix PIE for 64bit gdb -> 32bit inferior
http://sourceware.org/ml/gdb-patches/2010-02/msg00289.html
(+ a similar fix may be needed even elsewhere)
these "& mask" parts can be removed. This masking was there already before
start of the PIE(+OSX) patches.
I thought about their removal only just as a simplification in future but it
looks to be required for mips*. Downloading some mips .iso if it will run in
qemu to test it.
Thanks,
Jan