Re: [PATCH 5/9] PCI: tegra: Explicitly specify PMC instance to use

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Unassigned error code in tegra_pcie_get_resources() when devm_tegra_pmc_get() fails causes a kernel panic.
--

commit 03b6ac6d72e8b9fdb4c4e55b15fd0c26dfc93236
Author: Thierry Reding <treding@nvidia.com>

PCI: tegra: Explicitly specify PMC instance to use

Currently the kernel relies on a global variable to reference the PMC
context. This commit updates the driver to use an explicit lookup for
the PMC and passes that to the public PMC APIs.

> diff --git a/drivers/pci/controller/pci-tegra.c b/drivers/pci/controller/pci-tegra.c
> index 512309763d1fb..2c6c521e69017 100644
> --- a/drivers/pci/controller/pci-tegra.c
> +++ b/drivers/pci/controller/pci-tegra.c

[ ... ]

> @@ -1432,6 +1436,12 @@ static int tegra_pcie_get_resources(struct tegra_pcie *pcie)
>  		return err;
>  	}
>  
> +	pcie->pmc = devm_tegra_pmc_get(dev);
> +	if (IS_ERR(pcie->pmc)) {
> +		dev_err_probe(dev, PTR_ERR(pcie->pmc), "failed to get PMC\n");
> +		return err;

Does this code accidentally return success when devm_tegra_pmc_get() fails?

If devm_tegra_pmc_get() returns an error pointer, the error path is taken,
but err is not updated. Looking earlier in tegra_pcie_get_resources(), err
is 0 from the successful call to tegra_pcie_resets_get().

If the function returns 0, probe will continue with an error pointer
stored in pcie->pmc.

Later, when tegra_pcie_power_on() calls tegra_pmc_powergate_power_off(),
will it dereference this error pointer and cause a kernel panic?

Should this return the result of dev_err_probe() instead?

> +	}
> +
>  	if (soc->program_uphy) {
>  		err = tegra_pcie_phys_get(pcie);
>  		if (err < 0) {

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260506-pmc-v1-0-a6de5da7216b@nvidia.com?part=5