linux-kernel - Re: [PATCH] dmaengine: ti: edma: Check return value of of_dma_controller

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: <5a25044d-b8e6-4260-ad90-95e8f0c7c5c5@gmail.com>
Date: Wed, 25 Sep 2024 19:14:46 +0300
From: Péter Ujfalusi <peter.ujfalusi@...il.com>
To: Mikhail Arkhipov <m.arhipov@...a.ru>
Cc: Vinod Koul <vkoul@...nel.org>, dmaengine@...r.kernel.org,
 linux-kernel@...r.kernel.org, lvc-project@...uxtesting.org
Subject: Re: [PATCH] dmaengine: ti: edma: Check return value of
 of_dma_controller_register



On 23/09/2024 22:37, Mikhail Arkhipov wrote:
> If of_dma_controller_register() fails within the edma_probe() function,
> the driver does not check the return value or log the failure. This
> oversight can cause other drivers that rely on this DMA controller to fail
> during their probe phase. Specifically, when other drivers call
> of_dma_request_slave_channel(), they may receive an error code
> -EPROBE_DEFER, causing their initialization to be delayed or fail without
> clear logging of the root cause.
> 
> Add a check for the return value of of_dma_controller_register() in the
> edma_probe() function. If the function returns an error, log an appropriate
> error message and handle the failure by cleaning up resources and returning
> the error code. This ensures that the failure is properly reported, which
> aids in debugging and maintains system stability.
> 
> Found by Linux Verification Center (linuxtesting.org) with SVACE.

Acked-by: Peter Ujfalusi <peter.ujfalusi@...il.com>

> 
> Fixes: dc9b60552f6a ("ARM/dmaengine: edma: Move of_dma_controller_register to the dmaengine driver")
> Signed-off-by: Mikhail Arkhipov <m.arhipov@...a.ru>
> ---
>  drivers/dma/ti/edma.c | 17 +++++++++++++----
>  1 file changed, 13 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/dma/ti/edma.c b/drivers/dma/ti/edma.c
> index 5f8d2e93ff3f..9fcbd97d346f 100644
> --- a/drivers/dma/ti/edma.c
> +++ b/drivers/dma/ti/edma.c
> @@ -2529,18 +2529,27 @@ static int edma_probe(struct platform_device *pdev)
>  		if (ret) {
>  			dev_err(dev, "memcpy ddev registration failed (%d)\n",
>  				ret);
> -			dma_async_device_unregister(&ecc->dma_slave);
> -			goto err_reg1;
> +			goto err_unregister_dma_slave;
>  		}
>  	}
>  
> -	if (node)
> -		of_dma_controller_register(node, of_edma_xlate, ecc);
> +	if (node) {
> +		ret = of_dma_controller_register(node, of_edma_xlate, ecc);
> +		if (ret) {
> +			dev_err(dev, "Failed to register DMA controller (%d)\n", ret);
> +			goto err_unregister_dma_memcpy;
> +		}
> +	}
> 
>  	dev_info(dev, "TI EDMA DMA engine driver\n");
>  
>  	return 0;
>  
> +err_unregister_dma_memcpy:
> +	if (ecc->dma_memcpy)
> +		dma_async_device_unregister(ecc->dma_memcpy);
> +err_unregister_dma_slave:
> +	dma_async_device_unregister(&ecc->dma_slave);
>  err_reg1:
>  	edma_free_slot(ecc, ecc->dummy_slot);
>  err_disable_pm:

-- 
Péter