Re: [PATCH] serial: 8250_pci: Don't specify conflicting values to pci_device_id members

[Andy Shevchenko]

On Wed, Jun 03, 2026 at 01:45:37PM +0300, Andy Shevchenko wrote:
> On Wed, Jun 03, 2026 at 11:56:16AM +0200, Uwe Kleine-König (The Capable Hub) wrote:
> > The PCI_VDEVICE macro assigns 0 to .class and .class_mask to allow the
> > next value in the initializer to define the value for .driver_data.
> > 
> > So the construct
> > 
> > 	{
> > 		PCI_VDEVICE(INTASHIELD, 0x0D21),
> > 		.class = PCI_CLASS_COMMUNICATION_MULTISERIAL << 8,
> > 		.class_mask = 0xffff00,
> > 		.driver_data = pbn_b2_4_115200,
> > 	},
> > 
> > introduced in commit 44e55f1f3088 ("serial: 8250_pci: Consistently
> > define pci_device_ids using named initializers") has conflicting
> > assignments. In only some configurations (i.e. W=1 for me) that makes
> > the compiler unhappy.
> > 
> > So convert the two affected items to PCI_DEVICE which doesn't have that
> > hidden assigment to .class and .class_mask.
> 
> Reported-by? Closes?
> 
> > Fixes: 44e55f1f3088 ("serial: 8250_pci: Consistently define pci_device_ids using named initializers")
> > Signed-off-by: Uwe Kleine-König (The Capable Hub) <u.kleine-koenig@xxxxxxxxxxxx>
> 
> I will test it right away.

Reported-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
Closes: https://lore.kernel.org/linux-serial/ah_5qVKOf8LXG1Xo@ashevche-desk.local/T/#ma6eab90ca801b4292639f5c255a89b4033b33d21
Tested-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>

-- 
With Best Regards,
Andy Shevchenko