Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:1 as permitted sender) client-ip=2620:137:e000::3:1;
Date:   Fri, 22 Sep 2023 11:50:32 +0200
From:   Sascha Hauer <s.hauer@pengutronix.de>
To:     Tim van der Staaij | Zign <Tim.vanderstaaij@zigngroup.com>
Cc:     Shawn Guo <shawnguo@kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Vinod Koul <vkoul@kernel.org>,
        Pengutronix Kernel Team <kernel@pengutronix.de>,
        "dmaengine@vger.kernel.org" <dmaengine@vger.kernel.org>,
        Fabio Estevam <festevam@gmail.com>,
        "linux-arm-kernel@lists.infradead.org" 
        <linux-arm-kernel@lists.infradead.org>
Subject: Re: [PATCH] dmaengine: imx-sdma: fix deadlock in interrupt handler
Message-ID: <20230922095032.GU637806@pengutronix.de>
References: <AM0PR08MB30897429213E8DB9BCC1D6C880F8A@AM0PR08MB3089.eurprd08.prod.outlook.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <AM0PR08MB30897429213E8DB9BCC1D6C880F8A@AM0PR08MB3089.eurprd08.prod.outlook.com>
User-Agent: Mutt/1.10.1 (2018-07-13)
Precedence: bulk

Hi Tim,

On Thu, Sep 21, 2023 at 09:57:11AM +0000, Tim van der Staaij | Zign wrote:
> dev_warn internally acquires the lock that is already held when
> sdma_update_channel_loop is called. Therefore it is acquired twice and
> this is detected as a deadlock. Temporarily release the lock while
> logging to avoid this.
> 
> Signed-off-by: Tim van der Staaij <tim.vanderstaaij@zigngroup.com>
> Link: https://lore.kernel.org/all/AM0PR08MB308979EC3A8A53AE6E2D3408802CA@AM0PR08MB3089.eurprd08.prod.outlook.com/
> ---
>  drivers/dma/imx-sdma.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
> index 51012bd39900..3a7cd783a567 100644
> --- a/drivers/dma/imx-sdma.c
> +++ b/drivers/dma/imx-sdma.c
> @@ -904,7 +904,10 @@ static void sdma_update_channel_loop(struct sdma_channel *sdmac)
>  	 * owned buffer is available (i.e. BD_DONE was set too late).
>  	 */
>  	if (sdmac->desc && !is_sdma_channel_enabled(sdmac->sdma, sdmac->channel)) {
> +		spin_unlock(&sdmac->vc.lock);
>  		dev_warn(sdmac->sdma->dev, "restart cyclic channel %d\n", sdmac->channel);
> +		spin_lock(&sdmac->vc.lock);

This is strange. Why and how does dev_warn() call back into the SDMA
driver?

We shouldn't merge this without having a clue what exactly goes wrong
here. Please provide the corresponding lockdep output.

Sascha

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |