On Wed, 3 Dec 2008, FUJITA Tomonori wrote: > On Tue, 2 Dec 2008 21:08:19 +0200 (EET) > Kai Makisara <Kai.Makisara@xxxxxxxxxxx> wrote: > ... > > I added simple conversion (at the end > > of this message) but after this, the driver did not pass my simple tests. > > Looking at st_scsi_kern_execute(), it seems that it does not copy the > > sense data to struct st_request. > > I think that blk_execute_rq copies the sense data for us. So we don't > need to worry about it. > Now that I looked more carefully, it does that. > Without your patch, that is, with only my patchset, the driver passed > your tests? > Yes. > If so, can you try this patch to convert st_int_ioctl? > I replaced my patch with yours. Now the tests pass. I did some tests with debugging enabled and those showed that the sense data is returned correctly. > > Do you prefer me to change st_scsi_kern_execute to set -EBUSY to > syscall_result if st_scsi_kern_execute internally fails (that is, > blk_rq_map_kern or blk_get_request fails)? > I am not sure that -EBUSY is a valid error return any more. Should the error be -ENOMEM if blk_get_request fails and otherwise return what blk_rq_map_kern returns? > Note that scsi_execute_async (which st_do_scsi uses) could fail > internally. scsi_execute_async and st_scsi_kern_execute uses the same > functions that could fail, blk_rq_map_kern or blk_get_request > Yes. I am not sure that -EBUSY is correct return value there. Thanks, Kai -- To unsubscribe from this list: 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
