[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <2077369633.12794.1624037192994.JavaMail.zimbra@efficios.com>
Date: Fri, 18 Jun 2021 13:26:32 -0400 (EDT)
From: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
To: Christophe Leroy <christophe.leroy@...roup.eu>
Cc: Ingo Molnar <mingo@...hat.com>,
Peter Zijlstra <peterz@...radead.org>,
Thomas Gleixner <tglx@...utronix.de>,
maged michael <maged.michael@...il.com>,
Dave Watson <davejwatson@...com>,
Will Deacon <will.deacon@....com>,
"Russell King, ARM Linux" <linux@...linux.org.uk>,
David Sehr <sehr@...gle.com>,
Paul Mackerras <paulus@...ba.org>,
"H. Peter Anvin" <hpa@...or.com>,
linux-arch <linux-arch@...r.kernel.org>, x86 <x86@...nel.org>,
Andrew Hunter <ahh@...gle.com>,
Greg Hackmann <ghackmann@...gle.com>,
Alan Stern <stern@...land.harvard.edu>,
Paul <paulmck@...ux.vnet.ibm.com>,
Andrea Parri <parri.andrea@...il.com>,
Avi Kivity <avi@...lladb.com>,
Boqun Feng <boqun.feng@...il.com>,
linuxppc-dev <linuxppc-dev@...ts.ozlabs.org>,
Nicholas Piggin <npiggin@...il.com>,
Alexander Viro <viro@...iv.linux.org.uk>,
Andy Lutomirski <luto@...nel.org>,
linux-api <linux-api@...r.kernel.org>,
linux-kernel <linux-kernel@...r.kernel.org>,
Linus Torvalds <torvalds@...ux-foundation.org>
Subject: Re: [PATCH for 4.16 v7 02/11] powerpc: membarrier: Skip memory
barrier in switch_mm()
----- On Jun 18, 2021, at 1:13 PM, Christophe Leroy christophe.leroy@...roup.eu wrote:
[...]
>
> I don't understand all that complexity to just replace a simple
> 'smp_mb__after_unlock_lock()'.
>
> #define smp_mb__after_unlock_lock() smp_mb()
> #define smp_mb() barrier()
> # define barrier() __asm__ __volatile__("": : :"memory")
>
>
> Am I missing some subtility ?
On powerpc CONFIG_SMP, smp_mb() is actually defined as:
#define smp_mb() __smp_mb()
#define __smp_mb() mb()
#define mb() __asm__ __volatile__ ("sync" : : : "memory")
So the original motivation here was to skip a "sync" instruction whenever
switching between threads which are part of the same process. But based on
recent discussions, I suspect my implementation may be inaccurately doing
so though.
Thanks,
Mathieu
>
> Thanks
> Christophe
--
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com
Powered by blists - more mailing lists