Re: [PATCH 2/2] crypto: atmel-ecc - add support for atecc608a

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Use-After-Free due to early return in `atmel_ecc_remove()` missing list deletion.
--

commit da00b1f7ab58791bf49d80087ba781e30c9433d0
Author: Thorsten Blum <blum@kernel.org>

crypto: atmel-ecc - add support for atecc608a
    
The driver already supports the ATECC508A and ATECC608B and uses only
commands shared by these devices.

Microchip documents the ATECC608B as a backwards-compatible, functional
drop-in replacement for the ATECC608A, with the same device structure
and command interface [1]. Therefore, add support for the ATECC608A.

ATECC608A hardware was not available for testing.

Link: https://ww1.microchip.com/downloads/en/Appnotes/Migrating-from-the-ATECC608A-to-the-ATECC608B-DS40002237A.pdf [1]

> diff --git a/drivers/crypto/atmel-ecc.c b/drivers/crypto/atmel-ecc.c
> index 075f38a67935c..0904bb47fd0a9 100644
> --- a/drivers/crypto/atmel-ecc.c
> +++ b/drivers/crypto/atmel-ecc.c
> @@ -348,6 +348,7 @@ static void atmel_ecc_remove(struct i2c_client *client)

[Severity: High]
This is a pre-existing issue, but can the early return in atmel_ecc_remove()
cause a use-after-free?

If a user unbinds the ATECC device via sysfs while a crypto operation is in
flight (tfm_count > 0):

drivers/crypto/atmel-ecc.c:atmel_ecc_remove() {
	...
	if (atomic_read(&i2c_priv->tfm_count)) {
		dev_emerg(&client->dev, "Device is busy, expect memory corruption.\n");
		return;
	}
	...
	list_del(&i2c_priv->i2c_client_list_node);
	...
}

The early return bypasses list_del(&i2c_priv->i2c_client_list_node). The
driver core then frees the devm_kmalloc-allocated i2c_priv, leaving a freed
node on the global driver_data.i2c_client_list.

The next time any crypto operation is requested, atmel_ecc_i2c_client_alloc()
will iterate over the list and access the freed i2c_priv node, causing a
use-after-free.

Does it make sense to always remove the node from the list before returning?

>  
>  static const struct of_device_id atmel_ecc_dt_ids[] = {
>  	{ .compatible = "atmel,atecc508a", },
> +	{ .compatible = "atmel,atecc608a", },
>  	{ .compatible = "atmel,atecc608b", },
>  	{ }
>  };

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260830220151.21469-3-blum@kernel.org?part=2