Re: [PATCH 1/6] remoteproc: amd_mbv_bram: Bound device name formatting

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] remoteproc: carveout memory leak on rproc_parse_fw() failure
--

commit bb5382f10341b400baaa921ca1112dc728aa259e
Author: Ben Levinsky <ben.levinsky@amd.com>

remoteproc: amd_mbv_bram: Bound device name formatting

Pass the device name through a constant format string to prevent percent
conversions in the device name from consuming nonexistent variadic arguments,
and bound it to 31 characters.

> diff --git a/drivers/remoteproc/amd_mbv_bram_rproc.c b/drivers/remoteproc/amd_mbv_bram_rproc.c
> index e4a103cf84558..6cf33c7b7e641 100644
> --- a/drivers/remoteproc/amd_mbv_bram_rproc.c
> +++ b/drivers/remoteproc/amd_mbv_bram_rproc.c
> @@ -72,7 +72,7 @@ static int amd_bram_rproc_prepare(struct rproc *rproc)
>  				   resource_size(&res), da,
>  				   rproc_mem_entry_ioremap_wc,
>  				   rproc_mem_entry_iounmap,
> -				   dev_name(priv->dev));
> +				   "%.31s", dev_name(priv->dev));
>  	if (!mem)
>  		return -ENOMEM;
>  

[Severity: Medium]
Could this introduce a regression where the carveout memory leaks?

If rproc_parse_fw() fails during boot (for instance, due to a memory
allocation failure in rproc_elf_load_rsc_table()), the error path in
rproc_fw_boot() bypasses rproc_resource_cleanup():

drivers/remoteproc/remoteproc_core.c:rproc_fw_boot() {
    ...
	ret = rproc_parse_fw(rproc, fw);
	if (ret)
		goto unprepare_rproc;
    ...
clean_up_resources:
	rproc_resource_cleanup(rproc);
    ...
unprepare_rproc:
	rproc_unprepare_device(rproc);
    ...
}

Because the driver adds the carveout to rproc->carveouts shortly after this
allocation:

	mem = rproc_mem_entry_init(...);
	if (!mem)
		return -ENOMEM;

	rproc_add_carveout(rproc, mem);

And relies on the core's rproc_resource_cleanup() for cleanup, the carveout
would remain on the list if parsing fails.

Subsequent rproc_boot() attempts would then call prepare() again, potentially
adding duplicate carveouts and causing unbounded list growth.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261014213426.2957480-1-ben.levinsky@amd.com?part=1