Re: [PATCH v2 1/2] PCI: Add pci=noaer_recovery kernel boot option

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] The `pci=noaer_recovery` boot option fails to disable error recovery for CXL RCH internal errors.
--

--- Patch [1]: [PATCH v2 1/2] PCI: Add pci=noaer_recovery kernel boot option ---
commit 055898e8a2a1640801c835036f570a27cacd734b
Author: Yury Murashka <yurypm@arista.com>

PCI: Add pci=noaer_recovery kernel boot option

This patch adds a pci=noaer_recovery kernel boot option to disable AER error
recovery when an uncorrectable error is reported. This allows userspace to
inspect the unmodified error state without the kernel taking recovery actions.

> diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c
> index c4fd9c0b2a548..fb79990ce8fff 100644
> --- a/drivers/pci/pcie/aer.c
> +++ b/drivers/pci/pcie/aer.c
[ ... ]
> @@ -1187,10 +1193,12 @@ static void pci_aer_handle_error(struct pci_dev *dev, struct aer_err_info *info)
>  				pdrv->err_handler->cor_error_detected(dev);
>  			pcie_clear_device_status(dev);
>  		}
> -	} else if (info->severity == AER_NONFATAL)
> -		pcie_do_recovery(dev, pci_channel_io_normal, aer_root_reset);
> -	else if (info->severity == AER_FATAL)
> -		pcie_do_recovery(dev, pci_channel_io_frozen, aer_root_reset);
> +	} else if (!pcie_aer_recovery_disable) {
> +		if (info->severity == AER_NONFATAL)
> +			pcie_do_recovery(dev, pci_channel_io_normal, aer_root_reset);
> +		else if (info->severity == AER_FATAL)
> +			pcie_do_recovery(dev, pci_channel_io_frozen, aer_root_reset);
> +	}
>  }
>  
>  static void handle_error_source(struct pci_dev *dev, struct aer_err_info *info)

[Severity: High]
Does this fail to disable error recovery for CXL RCH internal errors?

If a CXL RCH device reports an internal uncorrectable error,
handle_error_source() calls cxl_rch_handle_error() before calling
pci_aer_handle_error():

drivers/pci/pcie/aer.c:handle_error_source() {
	cxl_rch_handle_error(dev, info);
	pci_aer_handle_error(dev, info);
	pci_dev_put(dev);
}

The CXL error handler unconditionally walks the RCEC and invokes the device
driver's error_detected callback, completely bypassing the
pcie_aer_recovery_disable check:

drivers/pci/pcie/aer_cxl_rch.c:cxl_rch_handle_error_iter() {
	...
		if (info->severity == AER_NONFATAL)
			err_handler->error_detected(dev, pci_channel_io_normal);
		else if (info->severity == AER_FATAL)
			err_handler->error_detected(dev, pci_channel_io_frozen);
	...
}

Should the recovery disable check also apply to cxl_rch_handle_error() so
that unexpected recovery actions aren't executed for CXL devices when
pci=noaer_recovery is used?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260709185429.627968-1-yurypm@arista.com?part=1