[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Yrw5f8GN2fh2orid@zx2c4.com>
Date: Wed, 29 Jun 2022 13:37:35 +0200
From: "Jason A. Donenfeld" <Jason@...c4.com>
To: Gregory Erwin <gregerwin256@...il.com>
Cc: Herbert Xu <herbert@...dor.apana.org.au>,
linux-kernel@...r.kernel.org, linux-wireless@...r.kernel.org,
Toke Høiland-Jørgensen <toke@...hat.com>,
Kalle Valo <kvalo@...nel.org>,
Rui Salvaterra <rsalvaterra@...il.com>, stable@...r.kernel.org
Subject: Re: [PATCH v7] ath9k: let sleep be interrupted when unregistering
hwrng
Hi Gregory,
On Tue, Jun 28, 2022 at 08:41:44PM -0700, Gregory Erwin wrote:
> Hmm, set_notify_signal() calls wake_up_state() in kernel/sched/core.c, which
> is not currently exported. Only by including EXPORT_SYMBOL(wake_up_state) and
> rebuilding vmlinux was I able to build rng-core.ko and load it successfully.
>
> That said, this patch allows 'ip link set wlan0 down' to wake a blocked process
> reading from /dev/hwrng, eliminating the delay as described. I'll give my
> sign-off with the EXPORT_SYMBOL sorted out.
Thanks for testing, and thanks for the note about EXPORT_SYMBOL(). I'll
send a v+1 with that fixed. And then it sounds like this patch finally
addresses all the issues you were seeing. Pfiew!
Jason
Powered by blists - more mailing lists