linux-kernel - Re: [PATCH 4/5] spi: zynqmp-gqspi: Allow interrupting operations

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: <87y0z9j0ha.fsf@bootlin.com>
Date: Fri, 17 Jan 2025 09:41:53 +0100
From: Miquel Raynal <miquel.raynal@...tlin.com>
To: Sean Anderson <sean.anderson@...ux.dev>
Cc: Mark Brown <broonie@...nel.org>,  Michal Simek <michal.simek@....com>,
  linux-spi@...r.kernel.org,  linux-kernel@...r.kernel.org,  Jinjie Ruan
 <ruanjinjie@...wei.com>,  linux-arm-kernel@...ts.infradead.org,  Amit
 Kumar Mahapatra <amit.kumar-mahapatra@....com>
Subject: Re: [PATCH 4/5] spi: zynqmp-gqspi: Allow interrupting operations

Hello Sean,

On 16/01/2025 at 17:55:20 -05, Sean Anderson <sean.anderson@...ux.dev> wrote:

> Some operations (such as reading several megabytes of data from a flash)
> can take several seconds or more. Users may want to cancel such
> operations. Allow them to do so now that we have a way to recover.

I fully agree with the observation, I tried myself interrupting too long
transfers with another spi controller:

e0205d6203c2c ("spi: atmel: Prevent false timeouts on long transfers")

But there were issues reported, so we limited the signals to SIGKILLs:

1ca2761a77349 ("spi: atmel: Do not cancel a transfer upon any signal")

But jffs2 plays with sigkills, so for spi memories it does not work
well, we had to revert:

890188d2d7e4a ("spi: atmel: Prevent spi transfers from being killed")

Same thing was also observed on Zynq7000:

26cfc0dbe43aa ("spi: spi-zynq-qspi: use wait_for_completion_timeout to make zynq_qspi_exec_mem_op not interruptible")

I would however hint to use a specific helper for deriving your timeouts
if you play with spi memories, because it is interesting to adapt the
values nevertheless:

d8e4ebf870187 ("spi: Create a helper to derive adaptive timeouts")

Thanks,
Miquèl