From: Qing Zhao <qing.zhao@oracle.com>
To: Alexander Monakov <amonakov@ispras.ru>
Cc: Jose Marchesi <jose.marchesi@oracle.com>,
"gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>
Subject: Re: [PATCH V2] Disable sched1 in functions that call setjmp
Date: Thu, 22 Dec 2022 19:28:28 +0000 [thread overview]
Message-ID: <35CA4F79-CA10-4560-8920-CDA65810E308@oracle.com> (raw)
In-Reply-To: <53b93d7e-a157-9116-d07a-4d51cd43d205@ispras.ru>
> On Dec 22, 2022, at 12:56 PM, Alexander Monakov <amonakov@ispras.ru> wrote:
>
>
> On Thu, 22 Dec 2022, Jose E. Marchesi via Gcc-patches wrote:
>
>> The first instruction scheduler pass reorders instructions in the TRY
>> block in a way `b=true' gets executed before the call to the function
>> `f'. This optimization is wrong, because `main' calls setjmp and `f'
>> is known to call longjmp.
>>
>> As discussed in BZ 57067, the root cause for this is the fact that
>> setjmp is not properly modeled in RTL, and therefore the backend
>> passes have no normalized way to handle this situation.
>>
>> As Alexander Monakov noted in the BZ, many RTL passes refuse to touch
>> functions that call setjmp. This includes for example gcse,
>> store_motion and cprop. This patch adds the sched1 pass to that list.
>>
>> Note that the other instruction scheduling passes are still allowed to
>> run on these functions, since they reorder instructions within basic
>> blocks, and therefore they cannot cross function calls.
>>
>> This doesn't fix the fundamental issue, but at least assures that
>> sched1 wont perform invalid transformation in correct C programs.
>
> I think scheduling across calls in the pre-RA scheduler is simply an oversight,
> we do not look at dataflow information and with 50% chance risk extending
> lifetime of a pseudoregister across a call, causing higher register pressure at
> the point of the call, and potentially an extra spill.
I am a little confused, you mean pre-RA scheduler does not look at the data flow
information at all when scheduling insns across calls currently?
Qing
>
> Therefore I would suggest to indeed solve the root cause, with (untested):
>
> diff --git a/gcc/sched-deps.cc b/gcc/sched-deps.cc
> index 948aa0c3b..343fe2bfa 100644
> --- a/gcc/sched-deps.cc
> +++ b/gcc/sched-deps.cc
> @@ -3688,7 +3688,13 @@ deps_analyze_insn (class deps_desc *deps, rtx_insn *insn)
>
> CANT_MOVE (insn) = 1;
>
> - if (find_reg_note (insn, REG_SETJMP, NULL))
> + if (!reload_completed)
> + {
> + /* Do not schedule across calls, this is prone to extending lifetime
> + of a pseudo and causing extra spill later on. */
> + reg_pending_barrier = MOVE_BARRIER;
> + }
> + else if (find_reg_note (insn, REG_SETJMP, NULL))
> {
> /* This is setjmp. Assume that all registers, not just
> hard registers, may be clobbered by this call. */
>
> Alexander
next prev parent reply other threads:[~2022-12-22 19:28 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-12-22 17:32 Jose E. Marchesi
2022-12-22 17:56 ` Alexander Monakov
2022-12-22 19:28 ` Qing Zhao [this message]
2022-12-23 7:33 ` Alexander Monakov
2022-12-23 14:34 ` Qing Zhao
2022-12-23 15:03 ` Alexander Monakov
2022-12-23 17:27 ` Jose E. Marchesi
2022-12-23 17:31 ` Alexander Monakov
2022-12-23 17:45 ` Jose E. Marchesi
2022-12-23 19:21 ` Qing Zhao
2022-12-23 19:36 ` Alexander Monakov
2022-12-23 19:57 ` Qing Zhao
2022-12-24 8:10 ` Alexander Monakov
2022-12-24 9:26 ` Richard Biener
2022-12-24 9:58 ` Jose E. Marchesi
2022-12-24 10:06 ` Richard Biener
2022-12-24 10:18 ` Alexander Monakov
2022-12-26 0:29 ` Segher Boessenkool
2023-01-05 18:11 ` Qing Zhao
2022-12-23 10:05 ` Richard Sandiford
2022-12-23 10:42 ` Jose E. Marchesi
2023-01-13 18:20 ` [PATCH] sched-deps: do not schedule pseudos across calls [PR108117] Alexander Monakov
2023-01-13 18:23 ` Richard Sandiford
2023-01-13 18:33 ` Jose E. Marchesi
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=35CA4F79-CA10-4560-8920-CDA65810E308@oracle.com \
--to=qing.zhao@oracle.com \
--cc=amonakov@ispras.ru \
--cc=gcc-patches@gcc.gnu.org \
--cc=jose.marchesi@oracle.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).