[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <alpine.LFD.2.00.1005312102040.3637@i5.linux-foundation.org>
Date: Mon, 31 May 2010 21:05:55 -0700 (PDT)
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Rusty Russell <rusty@...tcorp.com.au>
cc: Andrew Morton <akpm@...ux-foundation.org>,
Brandon Philips <brandon@...p.org>,
"Rafael J. Wysocki" <rjw@...k.pl>,
LKML <linux-kernel@...r.kernel.org>,
Jon Masters <jonathan@...masters.org>,
Tejun Heo <htejun@...il.com>,
Masami Hiramatsu <mhiramat@...hat.com>,
Kay Sievers <kay.sievers@...y.org>
Subject: Re: [PATCH 1/2] Make the module 'usage' lists be two-way
On Mon, 31 May 2010, Linus Torvalds wrote:
>
> Oh. And back then, it was inside a "modlist_lock". And that lock is long
> gone, but the GFP_ATOMIC remains.
Actually, keep it that way. Because I think that with your locking
cleanup, we could actually turn the module_mutex back into a spinlock.
And the thing is, we don't necessarily want it to be a spinlock per se,
but at the same time, I think that would be a good sanity test. If we can
turn it into a spinlock without triggering the "might_sleep()" debugging
code, that means that we are only holding it over real critical regions.
So while I don't think it needs to be a spinlock, it would be a nice added
sanity check if it were one. We have all this helper debugging code to
verify that spinlocked regions don't do "bad" things.
Linus
--
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/
Powered by blists - more mailing lists