Re: [Intel-wired-lan] [PATCH iwl-net V4, 2/2] i40e: Fully suspend and resume IO operations in EEH case

[Lukas Wunner <lukas@wunner.de>]

On Wed, May 15, 2024 at 04:07:05PM -0500, Thinh Tran wrote:
> When EEH events occurs, the callback functions in the i40e, which are
> managed by the EEH driver, will completely suspend and resume all IO
> operations.
> 
> - In the PCI error detected callback, replaced i40e_prep_for_reset()
>   with i40e_io_suspend(). The change is to fully suspend all I/O
>   operations
> - In the PCI error slot reset callback, replaced pci_enable_device_mem()
>   with pci_enable_device(). This change enables both I/O and memory of
>   the device.
> - In the PCI error resume callback, replaced i40e_handle_reset_warning()
>   with i40e_io_resume(). This change allows the system to resume I/O
>   operations

The above was applied as commit c80b6538d35a.

> @@ -16481,7 +16483,8 @@ static pci_ers_result_t i40e_pci_error_slot_reset(struct pci_dev *pdev)
>  	u32 reg;
>  
>  	dev_dbg(&pdev->dev, "%s\n", __func__);
> -	if (pci_enable_device_mem(pdev)) {
> +	/* enable I/O and memory of the device  */
> +	if (pci_enable_device(pdev)) {
>  		dev_info(&pdev->dev,
>  			 "Cannot re-enable PCI device after reset.\n");
>  		result = PCI_ERS_RESULT_DISCONNECT;

Why was this change made?

The driver calls pci_enable_device_mem() in i40e_probe(),
so calling pci_enable_device() here doesn't seem to make any sense.

The difference between pci_enable_device() and pci_enable_device_mem()
is that the former also enables access to the I/O Space of the device.
However I/O Space access is usually not used outside of x86.
And your patch targets powerpc because you seek to support EEH,
a powerpc-specific mechanism.

Unfortunately the commit message is not helpful at all because it
merely lists the code changes in prose form but doesn't explain
the *reason* for the change.

Thanks,

Lukas