On Thu, Feb 09, 2006 at 02:55:15PM -0500, Alan Stern wrote: > A fair number of SCSI BLIST bits have duplicate entries in the scsi_device > structure (in one instance the meaning is inverted): > > BLIST_BORKEN borken > BLIST_SINGLELUN single_lun > BLIST_USE_10_BYTE_MS use_10_for_ms > BLIST_MS_SKIP_PAGE_08 skip_ms_page_8 > BLIST_MS_SKIP_PAGE_3F skip_ms_page_3f > BLIST_MS_192_BYTES_FOR_3F use_192_bytes_for_3f > BLIST_RETRY_HWERROR retry_hwerror > BLIST_NOT_LOCKABLE lockable > BLIST_NOSTARTONADD no_start_on_add > > There's no good reason for this duplication. It's largely historical; at > the time the device flags were added there was no way to alter the bflags > value. Now there is, since sdev_bflags is stored in the scsi_device. > > Should we keep this duplication? Should we remove the device bits and use > the sdev_bflags bitmasks instead? I presume we don't want to remove the > BLIST entries because they are visible to userspace. (Although probably > some of them are not used anywhere at all.) > > What about things like BLIST_REPORTLUN2 and BLIST_NOREPORTLUN? They don't > mean anything for scsi_devices, only for scsi_targets. Nevertheless they > are part of the same BLIST entries. Again, I don't see any way around > that without affecting userspace. Do we need per-device blist flags at all? I suspect just having them in the target should be enough. - : send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
