Thanks for the patch.  I've committed a different version that
sets NOCSS for all AMD devices.

On Tue, Aug 05, 2025 at 08:41:43PM -0700, Renato Aguiar wrote:
> ping
> 
> On Sun, Jul 27 2025, Mike Larkin wrote:
> 
> > On Sun, Jul 27, 2025 at 04:37:37PM -0700, Renato Aguiar wrote:
> >> All USB ports stop working after a suspend/resume cycle on my Framework
> >> Laptop 13 (AMD Ryzen 7040 Series). This diff fixes the problem for me.
> >>
> >> --- sys/dev/pci/xhci_pci.c
> >> +++ sys/dev/pci/xhci_pci.c
> >> @@ -159,7 +159,11 @@ xhci_pci_attach(struct device *parent, struct device *
> >>  		    PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_AMD_19_4X_XHCI_2 ||
> >>  		    PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_AMD_19_4X_XHCI_3 ||
> >>  		    PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_AMD_19_4X_XHCI_4 ||
> >> -		    PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_AMD_19_4X_XHCI_5)
> >> +		    PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_AMD_19_4X_XHCI_5 ||
> >> +		    PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_AMD_19_7X_XHCI_1 ||
> >> +		    PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_AMD_19_7X_XHCI_2 ||
> >> +		    PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_AMD_19_7X_XHCI_3 ||
> >> +		    PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_AMD_19_7X_XHCI_4)
> >>  			psc->sc.sc_flags |= XHCI_NOCSS;
> >>  		break;
> >>  	}
> >>
> >
> > ok mlarkin
> >
> > but as kettenis pointed out a few times, we should probably just enable NOCSS
> > on all these rather than building up a long list.
> 
> -- 
> Renato Aguiar
> 
>