On 12/15/22 2:15 AM, Christoph Hellwig wrote: > The conversion itself looks good, but how the f**k did we end up with > a LLDD calling unquiry below the SCSI midlayer? I don't know how it happened. It looks like a hack around scsi_scan_host not removing devices. Going forward, it looks like we can remove the inquiry code by having scsi_scan_host be able to remove devices that are no longer returned. However, there's an extra catch because they have that DID_BAD_TARGET case where qemu returns VIRTIO_SCSI_S_BAD_TARGET if there are no devices on the host, but other drivers uses that error code for a bunch of different reasons. I was thinking to handle the DID_BAD_TARGET use case above and this type of issue: https://lore.kernel.org/linux-scsi/CA+PODjqrRzyJnOKoabMOV4EPByNnL1LgTi+QAKENP3NwUq5YCw@xxxxxxxxxxxxxx/ maybe we want to have a driver level BLIST like: static struct { char *driver; blist_flags_t flags; } scsi_static_driver_list[] __initdata = { {"virtio_scsi", BLIST_REMOVE_ON_BAD_TARGET}, {"aacard", BLIST_SHOW_DEV_PQ1}, One other question, can I do this work after the patchset in this email, the scsi_cmnd retry patches and the actual PR ones? I keep going off track on these side adventures.
