linux-kernel - Re: [PATCH 09/14] dmaengine: dma350: Support dma-channel-mask

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: <40f46471-850c-4700-8076-914ebeb68b40@kernel.org>
Date: Sat, 23 Aug 2025 18:10:43 +0200
From: Krzysztof Kozlowski <krzk@...nel.org>
To: Jisheng Zhang <jszhang@...nel.org>, Vinod Koul <vkoul@...nel.org>,
 Rob Herring <robh@...nel.org>, Krzysztof Kozlowski <krzk+dt@...nel.org>,
 Conor Dooley <conor+dt@...nel.org>, Robin Murphy <robin.murphy@....com>
Cc: dmaengine@...r.kernel.org, devicetree@...r.kernel.org,
 linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 09/14] dmaengine: dma350: Support dma-channel-mask

On 23/08/2025 17:40, Jisheng Zhang wrote:
> Not all channels are available to kernel, we need to support
> dma-channel-mask.
> 
> Signed-off-by: Jisheng Zhang <jszhang@...nel.org>
> ---
>  drivers/dma/arm-dma350.c | 20 ++++++++++++++++++--
>  1 file changed, 18 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/dma/arm-dma350.c b/drivers/dma/arm-dma350.c
> index 6a6d1c2a3ee6..72067518799e 100644
> --- a/drivers/dma/arm-dma350.c
> +++ b/drivers/dma/arm-dma350.c
> @@ -534,7 +534,7 @@ static int d350_probe(struct platform_device *pdev)
>  	struct device *dev = &pdev->dev;
>  	struct d350 *dmac;
>  	void __iomem *base;
> -	u32 reg;
> +	u32 reg, dma_chan_mask;
>  	int ret, nchan, dw, aw, r, p;
>  	bool coherent, memset;
>  
> @@ -563,6 +563,15 @@ static int d350_probe(struct platform_device *pdev)
>  
>  	dmac->nchan = nchan;
>  
> +	/* Enable all channels by default */
> +	dma_chan_mask = nchan - 1;
> +
> +	ret = of_property_read_u32(dev->of_node, "dma-channel-mask", &dma_chan_mask);
> +	if (ret < 0 && (ret != -EINVAL)) {
> +		dev_err(&pdev->dev, "dma-channel-mask is not complete.\n");
> +		return ret;
> +	}
> +
>  	reg = readl_relaxed(base + DMAINFO + DMA_BUILDCFG1);
>  	dmac->nreq = FIELD_GET(DMA_CFG_NUM_TRIGGER_IN, reg);
>  
> @@ -592,6 +601,11 @@ static int d350_probe(struct platform_device *pdev)
>  	memset = true;
>  	for (int i = 0; i < nchan; i++) {
>  		struct d350_chan *dch = &dmac->channels[i];
> +		char ch_irqname[8];
> +
> +		/* skip for reserved channels */
> +		if (!test_bit(i, (unsigned long *)&dma_chan_mask))
> +			continue;
>  
>  		dch->coherent = coherent;
>  		dch->base = base + DMACH(i);
> @@ -602,7 +616,9 @@ static int d350_probe(struct platform_device *pdev)
>  			dev_warn(dev, "No command link support on channel %d\n", i);
>  			continue;
>  		}
> -		dch->irq = platform_get_irq(pdev, i);
> +
> +		snprintf(ch_irqname, sizeof(ch_irqname), "ch%d", i);
> +		dch->irq = platform_get_irq_byname(pdev, ch_irqname);

Actual ABI break.

That's a no-go, sorry. You cannot decide to break all users just because
"Not all channels are available to the kernel". That's really, really
incomplete ABI breakage reasoning.

See also writing bindings doc.

Best regards,
Krzysztof