Damien, > That said, it is weird that scsi_get_vpd_page() does not call > scsi_device_supports_vpd(). The first patch in the series already makes that change. I noticed because the allocation for sd_read_cpr() is fairly big so it stuck out in my test runs while reworking scsi_get_vpd_page(). I didn't remove the conditional in sd_revalidate_disk(). While it is superfluous, I do like that the "fancy" protocol features are grouped. Guess we could switch it to a comment instead. I'll think about it... -- Martin K. Petersen Oracle Linux Engineering
