On Tue, Apr 25, 2017 at 01:00:06PM +0100, Julien Grall wrote: > Hi Roger, > > On 25/04/17 12:49, Roger Pau Monne wrote: > > On Mon, Apr 24, 2017 at 04:31:57PM +0100, Julien Grall wrote: > > > > +static int vpci_init_msi(struct pci_dev *pdev) > > > > +{ > > > > + uint8_t seg = pdev->seg, bus = pdev->bus; > > > > + uint8_t slot = PCI_SLOT(pdev->devfn), func = PCI_FUNC(pdev->devfn); > > > > + struct vpci_msi *msi = NULL; > > > > + unsigned int msi_offset; > > > > + uint16_t control; > > > > + int rc; > > > > + > > > > + msi_offset = pci_find_cap_offset(seg, bus, slot, func, > > > > PCI_CAP_ID_MSI); > > > > + if ( !msi_offset ) > > > > + return 0; > > > > + > > > > + if ( !dom0_msi ) > > > > > > I would introduce an helper to allow per-architecture decision. Likely on > > > ARM MSI will be enabled by default. > > > > dom0_msi is also enabled by default on x86. > > Sorry by default I meant that they will never be disabled on ARM. So you > could introduce a helper similar to is_domain_direct_mapped avoid the > introduction of dom0_msi for ARM.
OK, no problem. I've added two vpci_msi{x}_enabled macros that you can replace with 'true' if you wish. Roger. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org https://lists.xen.org/xen-devel