This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [LTO] PATCH: Make it build
- From: "Mark Mitchell" <mark at codesourcery dot com>
- To: <dberlin at dberlin dot org>
- Cc: <pinskia at gmail dot com>,<gcc-patches at gcc dot gnu dot org>
- Date: 24 Jun 2007 11:03:00 -0600
- Subject: Re: [LTO] PATCH: Make it build
An alternative attribute, with compression, would be swell. But, LTO can't mangle names itself, so it does need them. In some cases we might be able to pull the information out of the ELF symbol table, but probably not in all cases.
Thanks,
--
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-6685 x713
-----Original Message-----
From: "Daniel Berlin" <dberlin@dberlin.org>
Date: Sunday, Jun 24, 2007 8:36 am
Subject: Re: [LTO] PATCH: Make it build
On 6/24/07, Mark Mitchell <mark@codesourcery.com> wrote:
Andrew Pinski wrote:
> On 6/23/07, Mark Mitchell <mark@codesourcery.com> wrote:
>> 4. If no DW_AT_MIPS_linkage_name attribute is available, we need to
>> fall back to DW_AT_name.
>
> As mentioned a couple times before, GCC should really not needing to
> emit DW_AT_MIPS_linkage_name at all. It only does because gdb is not
> able to reconstruct the mangled name.
> For LTO, we need the mangled name. Therefore, independently of what GDB
should or should not do, having GCC emit this attribute is useful and
necessary for LTO.
You should be aware that these names take up the majority of binary space in a lot of cases (this is why we wanted to get rid of them in
the first place).
debug_str does not help here.
We really should be doing compression of them, even if it means not using DW_AT_mips_linkage_name as the attribute name.