From: Jeffrey Yasskin <jyasskin@google.com>
To: binutils@sourceware.org
Subject: Re: [PATCH] Loosen the Gold ODR checker to only compare filenames
Date: Thu, 03 Feb 2011 22:11:00 -0000 [thread overview]
Message-ID: <AANLkTik0NX-6HpG95Y2F2QpiiSN4t-nMOVWXmR_fF7en@mail.gmail.com> (raw)
In-Reply-To: <AANLkTinwc9hNW7xcYD02zX4poN+8_daWDe02yO5e6VBU@mail.gmail.com>
Whoops, please don't submit this. The strings that Source_location
points at don't live long enough to use them in the way I'm using
them. I'll send a working patch shortly.
On Wed, Feb 2, 2011 at 10:44 PM, Jeffrey Yasskin <jyasskin@google.com> wrote:
> Code compiled with different flags, especially -O, may have a
> different line number for the first instruction in a function. This
> produces false positives in the ODR checker when linking object files
> that should be ABI-compatible.
>
> The best fixes would be to A) look at the DW_AT_decl_line of the
> DW_TAG_subprogram for the function, but this would require gold to
> parse a whole new debug section, or B) hash the ODR-relevant aspects
> of each function into a new dwarf attribute, but this would require
> gcc to produce the hash and gold to parse a whole new debug section.
>
> Instead, loosening the ODR check to allow a function's definitions to
> be from anywhere within the same file removes the false positives with
> much less work, and would have caused very few extra false negatives
> in Google's codebase.
>
> 2011-02-02 Jeffrey Yasskin <jyasskin@google.com>
>
> * dwarf_reader.h: Add a Source_location type, and change the
> addr2line functions to return it.
> * dwarf_reader.cc: Implement Source_location, and change the
> addr2line functions to return it.
> * symtab.cc: Sort by just the filename.
> * object.cc: Convert a Source_location return to a std::string.
>
next prev parent reply other threads:[~2011-02-03 22:11 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-02-03 6:45 Jeffrey Yasskin
2011-02-03 22:11 ` Jeffrey Yasskin [this message]
2011-02-03 23:09 ` Jeffrey Yasskin
2011-02-04 19:24 ` Jeffrey Yasskin
2011-02-04 22:08 ` Ian Lance Taylor
2011-02-05 1:56 ` Ian Lance Taylor
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=AANLkTik0NX-6HpG95Y2F2QpiiSN4t-nMOVWXmR_fF7en@mail.gmail.com \
--to=jyasskin@google.com \
--cc=binutils@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).