linux-toolchains.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Florian Weimer <fweimer@redhat.com>
To: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: will@kernel.org, paulmck@kernel.org,
	Peter Zijlstra <peterz@infradead.org>,
	Segher Boessenkool <segher@kernel.crashing.org>,
	linux-kernel@vger.kernel.org,
	Linus Torvalds <torvalds@linux-foundation.org>,
	stern@rowland.harvard.edu, parri.andrea@gmail.com,
	boqun.feng@gmail.com, npiggin@gmail.com, dhowells@redhat.com,
	j.alglave@ucl.ac.uk, luc.maranget@inria.fr, akiyks@gmail.com,
	linux-toolchains@vger.kernel.org, linux-arch@vger.kernel.org
Subject: Re: [RFC PATCH] LKMM: Add ctrl_dep() macro for control dependency
Date: Wed, 29 Sep 2021 14:28:37 +0200	[thread overview]
Message-ID: <87lf3f7eh6.fsf@oldenburg.str.redhat.com> (raw)
In-Reply-To: <20210928211507.20335-1-mathieu.desnoyers@efficios.com> (Mathieu Desnoyers's message of "Tue, 28 Sep 2021 17:15:07 -0400")

* Mathieu Desnoyers:

> + * will ensure that the STORE to B happens after the LOAD of A. Normally a
> + * control dependency relies on a conditional branch having a data dependency
> + * on the LOAD and an architecture's inability to speculate STOREs. IOW, this
> + * provides a LOAD->STORE order.
> + *
> + * Due to optimizing compilers, extra care is needed; as per the example above
> + * the LOAD must be 'volatile' qualified in order to ensure the compiler
> + * actually emits the load, such that the data-dependency to the conditional
> + * branch can be formed.
> + *
> + * Secondly, the compiler must be prohibited from lifting anything out of the
> + * selection statement, as this would obviously also break the ordering.
> + *
> + * Thirdly, architectures that allow the LOAD->STORE reorder must ensure
> + * the compiler actually emits the conditional branch instruction.

If you need a specific instruction emitted, you need a compiler
intrinsic or inline assembly.

So something like this:

#define control_dep(x)                          \
  ({                                            \
    __typeof(x) x__ = (x);                      \
    __asm__("test $0, %0\n\t"                   \
            "jnz 1f\n\t"                        \
            "1:"                                \
            :: "r"(x__) : "cc");                \
  })

with an appropriate instruction sequence for each architecture.

I don't think it's possible to piggy-back this on something else.

Thanks,
Florian


  parent reply	other threads:[~2021-09-29 12:28 UTC|newest]

Thread overview: 35+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-28 21:15 [RFC PATCH] LKMM: Add ctrl_dep() macro for control dependency Mathieu Desnoyers
2021-09-29 12:06 ` Marco Elver
2021-10-01 15:45   ` Mathieu Desnoyers
2021-10-01 16:20     ` Linus Torvalds
2021-10-01 17:28       ` Mathieu Desnoyers
2021-10-01 18:18         ` Linus Torvalds
2021-09-29 12:28 ` Florian Weimer [this message]
2021-09-29 17:41   ` Segher Boessenkool
2021-09-29 19:46     ` Florian Weimer
2021-10-01 16:13     ` Mathieu Desnoyers
2021-10-01 16:26       ` Florian Weimer
2021-10-01 16:35         ` Linus Torvalds
2021-10-10 14:02           ` Florian Weimer
2021-10-14  0:01             ` Paul E. McKenney
2021-10-14  2:14               ` Alan Stern
2021-10-14 16:14                 ` Paul E. McKenney
2021-10-14 15:58               ` Florian Weimer
2021-10-14 16:23                 ` Paul E. McKenney
2021-10-14 18:19                   ` Florian Weimer
2021-10-14 21:09                     ` Paul E. McKenney
2021-10-14 22:36                       ` Linus Torvalds
2021-09-30 13:28   ` Mathieu Desnoyers
2021-09-29 14:47 ` Linus Torvalds
2021-09-29 14:54   ` Linus Torvalds
2021-09-29 19:50     ` Mathieu Desnoyers
2021-09-29 20:13       ` Mathieu Desnoyers
2021-09-29 19:27   ` Mathieu Desnoyers
2021-09-29 22:14     ` Linus Torvalds
2021-09-29 21:47 ` Segher Boessenkool
2021-09-29 23:57   ` Paul E. McKenney
2021-10-01 15:28     ` Mathieu Desnoyers
2021-10-01 22:53       ` Paul E. McKenney
2021-10-01 19:10     ` Segher Boessenkool
2021-10-01 22:50       ` Paul E. McKenney
2021-10-02 14:29       ` Alan Stern

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=87lf3f7eh6.fsf@oldenburg.str.redhat.com \
    --to=fweimer@redhat.com \
    --cc=akiyks@gmail.com \
    --cc=boqun.feng@gmail.com \
    --cc=dhowells@redhat.com \
    --cc=j.alglave@ucl.ac.uk \
    --cc=linux-arch@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-toolchains@vger.kernel.org \
    --cc=luc.maranget@inria.fr \
    --cc=mathieu.desnoyers@efficios.com \
    --cc=npiggin@gmail.com \
    --cc=parri.andrea@gmail.com \
    --cc=paulmck@kernel.org \
    --cc=peterz@infradead.org \
    --cc=segher@kernel.crashing.org \
    --cc=stern@rowland.harvard.edu \
    --cc=torvalds@linux-foundation.org \
    --cc=will@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).