On Tue, Apr 21, 2026 at 12:02:51PM -0700, David Matlack wrote: > > > I think a call to pci_clear_master() is needed here to so the device > > > stops issuing memory transactions while VFIO is trying to resetting > > > it. > > > > That doesn't sound right, reset should always work.. > > pcie_flr() calls pci_wait_for_pending_transaction() which times out > and prints an error message. The reset seems to be working fine though > after that.
Wow, that doesn't sound right in more ways :) pci_reset_function() just calls straight down to that and it is directly tied into the sysfs reset file. Meaning the sysfs doesn't work if the device is actively doing DMA? Feels wrong! However it has been like that since 2008 so I probably wouldn't touch it and agree that disabling busmastering appears to be an undocumented requirement. Jason
