linux-kernel - Re: [PATCH] reset: mchp: sparx5: Fix for lan966x

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] [day] [month] [year] [list]

Message-ID: <20250224140443.42ddc57c@bootlin.com>
Date: Mon, 24 Feb 2025 14:04:43 +0100
From: Herve Codina <herve.codina@...tlin.com>
To: Horatiu Vultur <horatiu.vultur@...rochip.com>
Cc: <p.zabel@...gutronix.de>, <Steen.Hegelund@...rochip.com>,
 <daniel.machon@...rochip.com>, <UNGLinuxDriver@...rochip.com>,
 <linux-arm-kernel@...ts.infradead.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] reset: mchp: sparx5: Fix for lan966x

Hi Horatiu,

On Mon, 24 Feb 2025 10:29:23 +0100
Horatiu Vultur <horatiu.vultur@...rochip.com> wrote:

> With the blamed commit it seems that lan966x doesn't seem to boot
> anymore when the internal CPU is used.
> The reason seems to be the usage of the devm_of_iomap, if we replace
> this with of_iomap, this seems to fix the issue as we use the same
> region also for other devices.
> 
> Fixes: 0426a920d6269c ("reset: mchp: sparx5: Map cpu-syscon locally in case of LAN966x")
> Signed-off-by: Horatiu Vultur <horatiu.vultur@...rochip.com>
> ---
>  drivers/reset/reset-microchip-sparx5.c | 14 +++++++++-----
>  1 file changed, 9 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/reset/reset-microchip-sparx5.c b/drivers/reset/reset-microchip-sparx5.c
> index aa5464be7053b..5a75f9833a91a 100644
> --- a/drivers/reset/reset-microchip-sparx5.c
> +++ b/drivers/reset/reset-microchip-sparx5.c
> @@ -8,6 +8,7 @@
>   */
>  #include <linux/mfd/syscon.h>
>  #include <linux/of.h>
> +#include <linux/of_address.h>
>  #include <linux/module.h>
>  #include <linux/platform_device.h>
>  #include <linux/property.h>
> @@ -72,14 +73,17 @@ static struct regmap *mchp_lan966x_syscon_to_regmap(struct device *dev,
>  						    struct device_node *syscon_np)
>  {
>  	struct regmap_config regmap_config = mchp_lan966x_syscon_regmap_config;
> -	resource_size_t size;
> +	struct resource res;
>  	void __iomem *base;
>  
> -	base = devm_of_iomap(dev, syscon_np, 0, &size);
> -	if (IS_ERR(base))
> -		return ERR_CAST(base);
> +	if (of_address_to_resource(syscon_np, 0, &res))
> +		return ERR_PTR(-ENOMEM);
>  
> -	regmap_config.max_register = size - 4;
> +	base = of_iomap(syscon_np, 0);
> +	if (!base)
> +		return ERR_PTR(-ENOMEM);
> +
> +	regmap_config.max_register =  resource_size(&res) - 4;
>  
>  	return devm_regmap_init_mmio(dev, base, &regmap_config);
>  }

In the Lan966x PCI device use case, the reset driver can be loaded, unloaded
and re-loaded.

When the driver is unloaded, resources have to be released and so with a
call to of_iomap(), a call to iounmap() is needed.

Maybe .remove() function in this driver should handle the needed iounmap()
call.

Best regards,
Hervé