Re: [PATCH v2 2/2] PCI: dwc: rcar-gen4: Use .post_deinit to handle dw_pcie_ep_init() failures

[sashiko-bot@kernel.org]

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

Pre-existing issues:
- [High] Missing early return in dw_pcie_ep_init_registers() error path causes EPC to be incorrectly notified as ready.
--

--- Patch [2]: [PATCH v2 2/2] PCI: dwc: rcar-gen4: Use .post_deinit to handle dw_pcie_ep_init() failures ---
commit 6071cb241e93cb3a6592fddfd5caeb02f9e49fac
Author: Marek Vasut <marek.vasut+renesas@mailbox.org>

PCI: dwc: rcar-gen4: Use .post_deinit to handle dw_pcie_ep_init() failures

Implement the .post_deinit() callback in R-Car Gen4 to cleanly stop clocks
and assert resets if endpoint initialization fails. This prevents duplicate
cleanup attempts by removing explicit deinit calls from the error paths.

> diff --git a/drivers/pci/controller/dwc/pcie-rcar-gen4.c b/drivers/pci/controller/dwc/pcie-rcar-gen4.c
> --- a/drivers/pci/controller/dwc/pcie-rcar-gen4.c
> +++ b/drivers/pci/controller/dwc/pcie-rcar-gen4.c
[ ... ]
> @@ -570,16 +576,13 @@ static int rcar_gen4_add_dw_pcie_ep(struct rcar_gen4_pcie *rcar)
>  	ep->ops = &pcie_ep_ops;
>  
>  	ret = dw_pcie_ep_init(ep);
> -	if (ret) {
> -		rcar_gen4_pcie_ep_deinit(rcar);
> +	if (ret)
>  		return ret;
> -	}
>  
>  	ret = dw_pcie_ep_init_registers(ep);
>  	if (ret) {
>  		dev_err(dev, "Failed to initialize DWC endpoint registers\n");
>  		dw_pcie_ep_deinit(ep);
> -		rcar_gen4_pcie_ep_deinit(rcar);
>  	}

[Severity: High]
This is a pre-existing issue, but does this error path fall through and
incorrectly notify the EPC framework?

If dw_pcie_ep_init_registers() fails in rcar_gen4_add_dw_pcie_ep(), the
code de-initializes the endpoint via dw_pcie_ep_deinit(), but seems to miss
a return statement. It will unconditionally fall through and call
pci_epc_init_notify(ep->epc).

Could this cause bound Endpoint Function (EPF) drivers to receive an
initialization complete notification on a disabled controller, potentially
leading to an asynchronous external abort (SError) or kernel panic when they
attempt memory mapping and register accesses?

>  
>  	pci_epc_init_notify(ep->epc);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260820035944.163472-1-marek.vasut+renesas@mailbox.org?part=2