linux-kernel - Re: [tip:x86/asm] x86: Speed up ___preempt

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20141003235145.GA4548@redhat.com>
Date:	Sat, 4 Oct 2014 01:51:45 +0200
From:	Oleg Nesterov <oleg@...hat.com>
To:	Andy Lutomirski <luto@...capital.net>
Cc:	Chuck Ebbert <cebbert.lkml@...il.com>,
	Sasha Levin <sasha.levin@...cle.com>,
	Ingo Molnar <mingo@...nel.org>,
	"H. Peter Anvin" <hpa@...or.com>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	Peter Zijlstra <peterz@...radead.org>,
	Denys Vlasenko <dvlasenk@...hat.com>,
	Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [tip:x86/asm] x86: Speed up ___preempt_schedule*() by using
	THUNK helpers

On 10/03, Andy Lutomirski wrote:
>
> Is this thing missing a flags or cc clobber:
>
> # define __preempt_schedule() asm ("call ___preempt_schedule")

Well, calling.h says "rflags is clobbered", so probably yes?


> Also, I'm at a loss as to wtf all this code is doing.
>
> There's preempt_schedule, which appears to be a normal C function.
>
> There's ___preempt_schedule, which is written in assembly and calls
> preempt_schedule.
>
> Then there's __preempt_schedule, which is an inline assembler function
> that calls ___preempt_schedule.
>
> Is this all just to make the call sequence for preempt_schedule shorter?

Yes, please look at 1a338ac32ca630f67df25b4a16436cccc314e997

    sched, x86: Optimize the preempt_schedule() call

    Remove the bloat of the C calling convention out of the
    preempt_enable() sites by creating an ASM wrapper which allows us to
    do an asm("call ___preempt_schedule") instead.

Oleg.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/