From: "Martin Liška" <mliska@suse.cz>
To: Alexander Monakov <amonakov@ispras.ru>
Cc: Richard Biener <richard.guenther@gmail.com>,
GCC Patches <gcc-patches@gcc.gnu.org>,
Jan Hubicka <hubicka@ucw.cz>,
rui314@gmail.com
Subject: Re: [PATCH] lto-plugin: add support for feature detection
Date: Mon, 9 May 2022 11:05:41 +0200 [thread overview]
Message-ID: <62cd2acb-7072-5f9e-a787-ae02a605fbd5@suse.cz> (raw)
In-Reply-To: <80f37f2-efdf-673-a8f4-69f2d5842ea2@ispras.ru>
CCing mold's author.
On 5/6/22 16:46, Alexander Monakov wrote:
>
>
> On Thu, 5 May 2022, Martin Liška wrote:
>
>> On 5/5/22 12:52, Alexander Monakov wrote:
>>> Feels a bit weird to ask, but before entertaining such an API extension,
>>> can we step back and understand the v3 variant of get_symbols? It is not
>>> documented, and from what little I saw I did not get the "motivation" for
>>> its existence (what it is doing that couldn't be done with the v2 api).
>>
>> Please see here:
>> https://github.com/rui314/mold/issues/181#issuecomment-1037927757
>
> Thanks. I've also re-read [1] and [2] which provided some relevant ideas.
>
> [1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86490
> [2] https://sourceware.org/bugzilla/show_bug.cgi?id=23411
>
>
> OK, so the crux of the issue is that sometimes the linker needs to feed the
> compiler plugin with LTO .o files extracted from static archives. This is
> not really obvious, because normally .a archives have an index that enumerates
> symbols defined/used by its .o files, and even during LTO the linker can simply
> consult the index to find out which members to extract. In theory, at least.
>
> The theory breaks in the following cases:
>
> - ld.bfd and common symbols (I wonder if weak/comdat code is also affected?):
> archive index does not indicate which definitions are common, so ld.bfd
> extracts the member and feeds it to the plugin to find out;
>
> - ld.gold and emulated archives via --start-lib a.o b.o ... --end-lib: here
> there's no index to consult and ld.gold feeds each .o to the plugin.
>
> In those cases it may happen that the linker extracts an .o file that would
> not be extracted during non-LTO link, and if that happens, the linker needs to
> inform the plugin. This is not the same as marking each symbol from spuriously
> extracted .o file as PREEMPTED when the .o file has constructors (the plugin
> will assume the constructors are kept while the linker needs to discard them).
>
> So get_symbols_v3 allows the linker to discard an LTO .o file to solve this.
>
> In absence of get_symbols_v3 mold tries to ensure correctness by restarting
> itself while appending a list of .o files to be discarded to its command line.
>
> I wonder if mold can invoke plugin cleanup callback to solve this without
> restarting.
>
> (also, hm, it seems to confirm my idea that LTO .o files should have had the
> correct symbol table so normal linker algorithms would work)
>
> Hopefully this was useful.
> Alexander
next prev parent reply other threads:[~2022-05-09 9:05 UTC|newest]
Thread overview: 76+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-02 7:51 [PATCH] Support LDPT_GET_SYMBOLS_V3 Martin Liška
2022-05-04 12:20 ` [PATCH] lto-plugin: add support for feature detection Martin Liška
2022-05-04 12:32 ` Alexander Monakov
2022-05-04 12:41 ` Martin Liška
2022-05-04 13:10 ` Alexander Monakov
2022-05-04 13:31 ` Martin Liška
2022-05-04 15:06 ` Bernhard Reutner-Fischer
2022-05-05 6:15 ` Richard Biener
2022-05-05 6:31 ` Richard Biener
2022-05-05 10:52 ` Alexander Monakov
2022-05-05 12:50 ` Martin Liška
2022-05-06 14:46 ` Alexander Monakov
2022-05-09 9:05 ` Martin Liška [this message]
2022-05-15 6:57 ` Rui Ueyama
2022-05-15 7:53 ` Alexander Monakov
2022-05-15 8:07 ` Rui Ueyama
2022-05-15 8:50 ` Alexander Monakov
2022-05-15 10:01 ` Rui Ueyama
2022-05-15 10:09 ` Alexander Monakov
2022-05-15 10:32 ` Rui Ueyama
2022-05-15 11:37 ` Alexander Monakov
2022-05-15 11:52 ` Rui Ueyama
2022-05-15 12:07 ` Alexander Monakov
2022-05-16 2:41 ` Rui Ueyama
2022-05-16 6:38 ` Alexander Monakov
2022-05-16 8:37 ` Rui Ueyama
2022-05-16 9:10 ` Richard Biener
2022-05-16 9:15 ` Alexander Monakov
2022-05-16 9:25 ` Jan Hubicka
2022-05-16 9:38 ` Martin Liška
2022-05-16 9:50 ` Jan Hubicka
2022-05-16 10:22 ` Richard Biener
2022-05-16 9:58 ` Rui Ueyama
2022-05-16 10:28 ` Richard Biener
2022-05-16 10:44 ` Rui Ueyama
2022-05-16 12:04 ` Martin Liška
2022-05-16 13:07 ` Rui Ueyama
2022-05-16 13:38 ` Alexander Monakov
2022-05-16 15:16 ` Alexander Monakov
2022-05-17 6:20 ` Richard Biener
2022-05-17 13:44 ` Martin Liška
2022-06-16 6:59 ` [PATCH 1/3] lto-plugin: support LDPT_GET_SYMBOLS_V3 Martin Liška
2022-06-20 9:23 ` Richard Biener
2022-06-16 7:01 ` [PATCH 2/3] lto-plugin: make claim_file_handler thread-safe Martin Liška
2022-06-20 9:32 ` Richard Biener
2022-06-20 10:20 ` Martin Liška
2022-06-21 7:56 ` Richard Biener
2022-06-21 8:43 ` Martin Liška
2022-06-24 8:37 ` Richard Biener
2022-06-16 7:01 ` [PATCH 3/3] lto-plugin: implement LDPT_GET_API_VERSION Martin Liška
2022-06-16 8:00 ` Alexander Monakov
2022-06-16 12:25 ` Martin Liška
2022-06-20 9:35 ` Richard Biener
2022-06-20 13:01 ` Martin Liška
2022-06-30 6:43 ` Rui Ueyama
2022-06-30 8:42 ` Martin Liška
2022-07-01 6:36 ` Richard Biener
2022-07-04 14:17 ` Martin Liška
2022-07-07 2:19 ` Rui Ueyama
2022-07-08 8:42 ` Martin Liška
2022-07-08 12:41 ` Alexander Monakov
2022-07-11 7:23 ` Rui Ueyama
2022-07-11 9:16 ` Alexander Monakov
2022-07-11 9:55 ` Richard Biener
2022-07-11 10:51 ` Martin Liška
2022-07-11 12:24 ` Rui Ueyama
2022-07-11 12:38 ` Alexander Monakov
2022-07-11 12:51 ` Martin Liška
2022-07-12 1:36 ` Rui Ueyama
2022-07-11 16:35 ` Alexander Monakov
2022-07-12 6:28 ` Richard Biener
2022-07-12 7:36 ` Martin Liška
2022-07-12 11:50 ` Rui Ueyama
2022-07-12 13:21 ` Richard Biener
2022-07-12 13:31 ` Martin Liška
2022-07-13 7:44 ` Rui Ueyama
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=62cd2acb-7072-5f9e-a787-ae02a605fbd5@suse.cz \
--to=mliska@suse.cz \
--cc=amonakov@ispras.ru \
--cc=gcc-patches@gcc.gnu.org \
--cc=hubicka@ucw.cz \
--cc=richard.guenther@gmail.com \
--cc=rui314@gmail.com \
/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).