linux-kernel - Re: [PATCH] spi: a3700: fix hang caused by a3700_spi_transfer_one

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20200630111137.GI5272@sirena.org.uk>
Date:   Tue, 30 Jun 2020 12:11:37 +0100
From:   Mark Brown <broonie@...nel.org>
To:     Daisuke Yamane <yamane07ynct@...il.com>
Cc:     linux-spi@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] spi: a3700: fix hang caused by
 a3700_spi_transfer_one_fifo()

On Tue, Jun 30, 2020 at 02:44:21AM +0900, Daisuke Yamane wrote:
> transfer_one() must call spi_finalize_current_transfer() before
> returning to inform current transfer has finished. Otherwise spi driver
> doesn't issue next transfer, and hang.

To be clear it can also return a positive value and then finalize later,
there's no need to finalize before returning (otherwise finalizing would
be a bit redundant) and if the driver doesn't return a positive value
there should be no need to finalize at all.

> However a3700_spi_transfer_one_fifo() doesn't call it if waiting for
> "wfifo empty" or "xfer ready" has timed out.
> Thus, this patch corrects error handling of them.

The core shouldn't be waiting at all if the driver returned an error, we
only wait if the return value was positive.  Looking at the code it's
not clear to me how we manage to end up waiting - it looks like the
driver passes back the error correctly and the core looks like it does
the right thing.  Have you seen hangs in operation?

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)