From: Andy Lutomirski <luto@kernel.org>
To: Andy Lutomirski <luto@kernel.org>
Cc: X86 ML <x86@kernel.org>,
Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
LKML <linux-kernel@vger.kernel.org>,
Nicholas Piggin <npiggin@gmail.com>,
Arnd Bergmann <arnd@arndb.de>, Anton Blanchard <anton@ozlabs.org>,
stable <stable@vger.kernel.org>
Subject: Re: [PATCH v3 4/4] membarrier: Execute SYNC_CORE on the calling thread
Date: Tue, 8 Dec 2020 20:15:17 -0800 [thread overview]
Message-ID: <CALCETrX1TDAhJypd04iMoNSo=aNK5dbxCu_kNj3i=3AdeWNz6Q@mail.gmail.com> (raw)
In-Reply-To: <250ded637696d490c69bef1877148db86066881c.1607058304.git.luto@kernel.org>
On Thu, Dec 3, 2020 at 9:07 PM Andy Lutomirski <luto@kernel.org> wrote:
>
> membarrier()'s MEMBARRIER_CMD_PRIVATE_EXPEDITED_SYNC_CORE is documented
> as syncing the core on all sibling threads but not necessarily the
> calling thread. This behavior is fundamentally buggy and cannot be used
> safely. Suppose a user program has two threads. Thread A is on CPU 0
> and thread B is on CPU 1. Thread A modifies some text and calls
> membarrier(MEMBARRIER_CMD_PRIVATE_EXPEDITED_SYNC_CORE). Then thread B
> executes the modified code. If, at any point after membarrier() decides
> which CPUs to target, thread A could be preempted and replaced by thread
> B on CPU 0. This could even happen on exit from the membarrier()
> syscall. If this happens, thread B will end up running on CPU 0 without
> having synced.
>
> In principle, this could be fixed by arranging for the scheduler to
> sync_core_before_usermode() whenever switching between two threads in
> the same mm if there is any possibility of a concurrent membarrier()
> call, but this would have considerable overhead. Instead, make
> membarrier() sync the calling CPU as well.
>
> As an optimization, this avoids an extra smp_mb() in the default
> barrier-only mode.
Fixes: 70216e18e519 ("membarrier: Provide core serializing command,
*_SYNC_CORE")
also:
> + /*
> + * For regular membarrier, we can save a few cycles by
> + * skipping the current cpu -- we're about to do smp_mb()
> + * below, and if we migrate to a different cpu, this cpu
> + * and the new cpu will execute a full barrier in the
> + * scheduler.
> + *
> + * For CORE_SYNC, we do need a barrier on the current cpu --
s/CORE_SYNC/SYNC_CORE/
--Andy
next prev parent reply other threads:[~2020-12-09 4:16 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-12-04 5:07 [PATCH v3 0/4] membarrier fixes Andy Lutomirski
2020-12-04 5:07 ` [PATCH v3 1/4] x86/membarrier: Get rid of a dubious optimization Andy Lutomirski
2020-12-09 4:10 ` Andy Lutomirski
2020-12-09 8:42 ` [tip: x86/urgent] " tip-bot2 for Andy Lutomirski
2020-12-04 5:07 ` [PATCH v3 2/4] membarrier: Add an actual barrier before rseq_preempt() Andy Lutomirski
2020-12-09 4:12 ` Andy Lutomirski
2020-12-09 8:42 ` [tip: x86/urgent] " tip-bot2 for Andy Lutomirski
2020-12-14 18:05 ` Andy Lutomirski
2020-12-04 5:07 ` [PATCH v3 3/4] membarrier: Explicitly sync remote cores when SYNC_CORE is requested Andy Lutomirski
2020-12-09 4:13 ` Andy Lutomirski
2020-12-09 8:42 ` [tip: x86/urgent] " tip-bot2 for Andy Lutomirski
2020-12-04 5:07 ` [PATCH v3 4/4] membarrier: Execute SYNC_CORE on the calling thread Andy Lutomirski
2020-12-04 19:35 ` Mathieu Desnoyers
2020-12-09 4:15 ` Andy Lutomirski [this message]
2020-12-09 8:42 ` [tip: x86/urgent] " tip-bot2 for Andy Lutomirski
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='CALCETrX1TDAhJypd04iMoNSo=aNK5dbxCu_kNj3i=3AdeWNz6Q@mail.gmail.com' \
--to=luto@kernel.org \
--cc=anton@ozlabs.org \
--cc=arnd@arndb.de \
--cc=linux-kernel@vger.kernel.org \
--cc=mathieu.desnoyers@efficios.com \
--cc=npiggin@gmail.com \
--cc=stable@vger.kernel.org \
--cc=x86@kernel.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 NNTP newsgroup(s).