From: Jake Oshins <jakeo@xxxxxxxxxxxxx> Sent: Friday, June 21, 2024 9:51 AM > [...] >On Fri, Jun 21, 2024 at 06:19:05AM +0000, Wei Liu wrote: > On Fri, Jun 21, 2024 at 03:15:19AM +0000, Michael Kelley wrote: > > From: Wei Liu <mailto:wei.liu@xxxxxxxxxx> Sent: Thursday, June 20, 2024 6:48 PM > > > > > > The intent of the code snippet is to always return 0 for both fields. > > > The check is wrong though. Fix that. > > > > > > This is discovered by this call in VFIO: > > > > > > pci_read_config_byte(vdev->pdev, PCI_INTERRUPT_PIN, &pin); > > > > > > The old code does not set *val to 0 because the second half of the check is > > > incorrect. Hi Wei, so you got a non-zero 'pin' value returned by the host when the guest reads from the MMIO config page. What's the consequence? Will VFIO try to use the legacy INTx rather than MSI/MSI-X? I'm curious how you noticed the bug. I'm also curious why the host doesn't return 0 for the 'PIN' register when the guest reads it from the config page. > I believe that this fix is correct. (And I'm frankly surprised that this bug didn't > cause a problem before this. It's been there since I first wrote the code.) > -- Jake Oshins I suppose it didn't cause any issue because the PCI device drivers use MSI/MSI-X, so they don't care about the values of the 'PIN' and 'LINE' registers. Thanks, Dexuan
