On Thursday 12 June 2008, Borislav Petkov wrote: > This is the first one in the series attempting to phase out ide_atapi_pc and use > block request structs. Also, the pc request type is now REQ_TYPE_BLOCK_PC. As a result, > idefloppy_blockpc_cmd becomes unused and can go. Do not attack too many dragons at once or they will slay you... ;) IOW this patch mixes too many changes (some _really_ non-trivial) in one go which results in rather bad outcome... [...] > diff --git a/drivers/ide/ide-floppy.c b/drivers/ide/ide-floppy.c > index b368943..ffebf83 100644 > --- a/drivers/ide/ide-floppy.c > +++ b/drivers/ide/ide-floppy.c > @@ -217,7 +217,7 @@ static int idefloppy_end_request(ide_drive_t *drive, int uptodate, int nsecs) > /* Why does this happen? */ > if (!rq) > return 0; > - if (!blk_special_request(rq)) { > + if (!blk_pc_request(rq)) { > /* our real local end request function */ > ide_end_request(drive, uptodate, nsecs); > return 0; > @@ -282,13 +282,14 @@ static void idefloppy_update_buffers(ide_drive_t *drive, > * pass through the driver. > */ > static void idefloppy_queue_pc_head(ide_drive_t *drive, struct ide_atapi_pc *pc, > - struct request *rq) > + struct request *rq, unsigned char *cmd) > { > struct ide_floppy_obj *floppy = drive->driver_data; > > blk_rq_init(NULL, rq); > rq->buffer = (char *) pc; > - rq->cmd_type = REQ_TYPE_SPECIAL; > + rq->cmd_type = REQ_TYPE_BLOCK_PC; > + memcpy(rq->cmd, cmd, 12); > rq->cmd_flags |= REQ_PREEMPT; > rq->rq_disk = floppy->disk; > ide_do_drive_cmd(drive, rq); REQUEST SENSE command is switched from REQ_TYPE_SPECIAL to REQ_TYPE_BLOCK_PC which should belong to a separate patch and is premature IMHO (the other ATAPI drivers seem to still use REQ_TYPE_SPECIAL so probably it makes sense to unify/move REQUEST_SENSE handling to ide-atapi.c first). > @@ -323,10 +324,10 @@ static void ide_floppy_callback(ide_drive_t *drive) > if (floppy->failed_pc == pc) > floppy->failed_pc = NULL; > > - if (pc->c[0] == GPCMD_READ_10 || pc->c[0] == GPCMD_WRITE_10 || > - (pc->rq && blk_pc_request(pc->rq))) > + if (pc->rq->cmd[0] == GPCMD_READ_10 || pc->rq->cmd[0] == GPCMD_WRITE_10 > + || (pc->rq && blk_pc_request(pc->rq))) > uptodate = 1; /* FIXME */ This actually seems to break REQUEST SENSE handling (please note the blk_pc_request() check above)... :( > - else if (pc->c[0] == GPCMD_REQUEST_SENSE) { > + else if (pc->rq->cmd[0] == GPCMD_REQUEST_SENSE) { [...] > @@ -592,25 +601,6 @@ static void idefloppy_create_rw_cmd(idefloppy_floppy_t *floppy, > pc->flags |= PC_FLAG_DMA_OK; > } > > -static void idefloppy_blockpc_cmd(idefloppy_floppy_t *floppy, > - struct ide_atapi_pc *pc, struct request *rq) > -{ > - idefloppy_init_pc(pc); > - memcpy(pc->c, rq->cmd, sizeof(pc->c)); > - pc->rq = rq; > - pc->b_count = rq->data_len; > - if (rq->data_len && rq_data_dir(rq) == WRITE) > - pc->flags |= PC_FLAG_WRITING; > - pc->buf = rq->data; > - if (rq->bio) > - pc->flags |= PC_FLAG_DMA_OK; > - /* > - * possibly problematic, doesn't look like ide-floppy correctly > - * handled scattered requests if dma fails... > - */ > - pc->req_xfer = pc->buf_size = rq->data_len; > -} [...] > @@ -644,12 +634,9 @@ static ide_startstop_t idefloppy_do_request(ide_drive_t *drive, > } > pc = idefloppy_next_pc_storage(drive); > idefloppy_create_rw_cmd(floppy, pc, rq, block); > - } else if (blk_special_request(rq)) { > + } else if (blk_pc_request(rq)) > pc = (struct ide_atapi_pc *) rq->buffer; > - } else if (blk_pc_request(rq)) { > - pc = idefloppy_next_pc_storage(drive); > - idefloppy_blockpc_cmd(floppy, pc, rq); > - } else { > + else { > blk_dump_rq_flags(rq, > "ide-floppy: unsupported command in queue"); > idefloppy_end_request(drive, 0, 0); Also the above changes seem to break handling of blk_pc_request() requests (i.e. SG_IO ioctl) because they will be no longer initialized properly. OTOH the main change (pc->c to rq->cmd) looks OK (only minor complaint is that 'u8' is both shorter and more readable than 'unsigned short')... [ I had to also skip patches #16-17 because of skipping this patch. ] -- To unsubscribe from this list: send the line "unsubscribe linux-ide" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html