Re: [PATCH v2 2/2] spi: spl022: switch to use default spi_transfer_one_message()

From: Linus Walleij <linus.walleij@linaro.org>
To: Nam Cao <namcao@linutronix.de>
Cc: broonie@kernel.org, linux-arm-kernel@lists.infradead.org,
	linux-spi@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 2/2] spi: spl022: switch to use default spi_transfer_one_message()
Date: Thu, 14 Dec 2023 01:19:04 +0100	[thread overview]
Message-ID: <CACRpkdZfuKmg8gY1J+prF0L0=S7PSW+0ZuqM3HhNrB1sXcNG5g@mail.gmail.com> (raw)
In-Reply-To: <ae1940abd6ff6a9e77b4373cff60007c641a0c6c.1701274975.git.namcao@linutronix.de>

On Wed, Nov 29, 2023 at 5:32 PM Nam Cao <namcao@linutronix.de> wrote:

> Except for polling mode, this driver's transfer_one_message() makes use
> of interrupt handler and tasklet. This is problematic because
> spi_transfer_delay_exec(), who may sleep, is called in interrupt handler
> and tasklet. This causes the following warning:
> BUG: sleeping function called from invalid context at drivers/spi/spi.c:1428
>
> Switch to use the default spi_transfer_one_message() instead, which
> calls spi_transfer_delay_exec() appropriately.
>
> Signed-off-by: Nam Cao <namcao@linutronix.de>
> Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> Tested with polling mode and interrupt mode, NOT with DMA mode.
> Support with testing very appreciated!

FWIW I tested this now on a device using DMA for the transfers
and everything works fine like before.

Yours,
Linus Walleij