Fwd: PATCH: Extend .zdebug section support to binutils and ld

[Cary Coutant <ccoutant at google dot com>]

[this time with the right email for Alan...]


---------- Forwarded message ----------
From: Cary Coutant <ccoutant@google.com>
Date: Tue, Oct 26, 2010 at 1:03 PM
Subject: Re: PATCH: Extend .zdebug section support to binutils and ld
To: binutils@sourceware.org
Cc: "H.J. Lu" <hjl.tools@gmail.com>, Alan Modra <amodra@bigpond.net.au>


Ping? I'd like to see binutils and gnu ld supporting compressed debug sections.

-cary


On Fri, Aug 6, 2010 at 1:37 PM, Cary Coutant <ccoutant@google.com> wrote:
> On Thu, Jul 22, 2010 at 12:59 PM, Daniel Jacobowitz
> <dan@codesourcery.com> wrote:
>> That doesn't tell me anything. ?What is the right thing to do, now
>> that you've made binutils do it? ?Do you have to uncompress the
>> section and then apply relocations to it?
>>
>> If you do, there's no separate issue - don't check this in unless
>> objects are marked in such a way that other tools will immediately
>> reject them.
>
> Yes, it would be an improvement to mark compressed sections with a new
> section type and SHF_OS_NONCONFORMING, but the ".zdebug" mechanism is
> already there right now, and it would be nice to have this patch to
> make the utilities work with what the assembler is producing today.
> The renaming convention was established a couple of years ago as a
> cheap way of preventing unaware tools of choking on these sections --
> this patch fixes a bug that was introduced then (trying to apply
> relocations before decompressing), but not actually exposed until gas
> started generating relocatable compressed sections.
>
> I'm not sure I understand why this patch should be held up until we
> implement a more foolproof mechanism. Adding this patch shouldn't
> prevent us from pursuing the better solution, but it will improve
> things in the meantime.
>
> -cary
>