Re: [PATCH v3 1/2] pstore: Avoid deadlock in panic andemergency-restart path

From: Tony Luck
Date: Thu Jan 10 2013 - 13:20:46 EST


On Thu, Dec 20, 2012 at 7:12 AM, Seiji Aguchi <seiji.aguchi@xxxxxxx> wrote:
> + if (pstore_cannot_block_path(reason)) {
> + is_locked = spin_trylock_irqsave(&psinfo->buf_lock, flags);
> + if (!is_locked) {
> + pr_err("pstore dump routine blocked in %s path, may corrupt error record\n"
> + , in_nmi() ? "NMI" : why);
> + }

My only quibble with this patchset is this message. The sentiment is
nice, but nobody will see it. kmsg_dump has already picked the
pieces of log_buf that will be saved to pstore - so this new message
won't be included. I suppose it will show up on a serial console - but
if a user has a serial console, they don't need to use pstore.

But I don't think it is likely to hurt us (to get this far in a panic we already
printed a bunch of stuff to the console and I can't think of a credible
scenario where a few extra bytes would run into a problem that the
earlier messages didn't).

So:

Acked-by: Tony Luck <tony.luck@xxxxxxxxx>

[Also Ack for part2 the touches efivars.c]

-Tony

[Or are you asking me to apply these rather than just Ack them??]
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/