[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20090602000850.GA31064@redhat.com>
Date: Tue, 2 Jun 2009 02:08:50 +0200
From: Oleg Nesterov <oleg@...hat.com>
To: Roland McGrath <roland@...hat.com>
Cc: Alan Cox <alan@...rguk.ukuu.org.uk>, paul@...-scientist.net,
linux-kernel@...r.kernel.org, stable@...nel.org,
Andrew Morton <akpm@...ux-foundation.org>,
Andi Kleen <andi@...stfloor.org>
Subject: Re: [PATCH] coredump: Retry writes where appropriate
On 06/01, Roland McGrath wrote:
>
> That is almost a separate subject, really. Having i/o calls' waits wrongly
> interrupted and then clearing TIF_SIGPENDING just seems goofy to me.
Yes, agreed. The patch I sent make the coredumping task invisible to all
signals except SIGKILL.
> But there is the possibility of recalc_sigpending_and_wake
> via cancel_freezing, at least. Seems safer to make recalc_sigpending_tsk
> robust in this case.
Oh, I forgot about freezer...
Well, not good to complicate recalc_sigpending_tsk() for this unlikely case.
And this can't help, freezer does signal_wake_up() unconditionally.
So in fact this is another argument to check signal_pending() and clear it
in dump_write/seek.
But since the coredumping task is not freezable anyway, perhaps we should
change fake_signal_wake_up() to ignore SIGNAL_GROUP_DUMPING task.
Or we should make the coredumping freezable. This means dump_write/seek
and exit_mm() should do try_to_freeze().
In any case, the coredumping is special. If ->write() returns -ERESTART/EINTR
it assumes the return to ths user-space, this is not true for the coredump.
This means that handling the spurious signals in coredump_file_write() is
not so bad if we can't avoid this.
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/
Powered by blists - more mailing lists