On 10/25/2021 10:59 AM, Michael S. Tsirkin wrote: I think it will disappear after your patch with the returned status values. >> sparse warnings: (new ones prefixed by >>) >>>> drivers/block/virtio_blk.c:238:24: sparse: sparse: incorrect type in return expression (different base types) @@ expected int @@ got restricted blk_status_t [usertype] @@ >> drivers/block/virtio_blk.c:238:24: sparse: expected int >> drivers/block/virtio_blk.c:238:24: sparse: got restricted blk_status_t [usertype] >> drivers/block/virtio_blk.c:246:32: sparse: sparse: incorrect type in return expression (different base types) @@ expected int @@ got restricted blk_status_t [usertype] @@ >> drivers/block/virtio_blk.c:246:32: sparse: expected int >> drivers/block/virtio_blk.c:246:32: sparse: got restricted blk_status_t [usertype] >>>> drivers/block/virtio_blk.c:320:24: sparse: sparse: incorrect type in return expression (different base types) @@ expected restricted blk_status_t @@ got int [assigned] err @@ >> drivers/block/virtio_blk.c:320:24: sparse: expected restricted blk_status_t >> drivers/block/virtio_blk.c:320:24: sparse: got int [assigned] err >> >> vim +238 drivers/block/virtio_blk.c >> >> 203 >> 204 static int virtblk_setup_cmd(struct virtio_device *vdev, struct request *req, >> 205 struct virtblk_req *vbr) >> 206 { >> 207 bool unmap = false; >> 208 u32 type; >> 209 >> 210 vbr->out_hdr.sector = 0; >> 211 >> 212 switch (req_op(req)) { >> 213 case REQ_OP_READ: >> 214 type = VIRTIO_BLK_T_IN; >> 215 vbr->out_hdr.sector = cpu_to_virtio64(vdev, >> 216 blk_rq_pos(req)); >> 217 break; >> 218 case REQ_OP_WRITE: >> 219 type = VIRTIO_BLK_T_OUT; >> 220 vbr->out_hdr.sector = cpu_to_virtio64(vdev, >> 221 blk_rq_pos(req)); >> 222 break; >> 223 case REQ_OP_FLUSH: >> 224 type = VIRTIO_BLK_T_FLUSH; >> 225 break; >> 226 case REQ_OP_DISCARD: >> 227 type = VIRTIO_BLK_T_DISCARD; >> 228 break; >> 229 case REQ_OP_WRITE_ZEROES: >> 230 type = VIRTIO_BLK_T_WRITE_ZEROES; >> 231 unmap = !(req->cmd_flags & REQ_NOUNMAP); >> 232 break; >> 233 case REQ_OP_DRV_IN: >> 234 type = VIRTIO_BLK_T_GET_ID; >> 235 break; >> 236 default: >> 237 WARN_ON_ONCE(1); >> > 238 return BLK_STS_IOERR; >> 239 } >> 240 >> 241 vbr->out_hdr.type = cpu_to_virtio32(vdev, type); >> 242 vbr->out_hdr.ioprio = cpu_to_virtio32(vdev, req_get_ioprio(req)); >> 243 >> 244 if (type == VIRTIO_BLK_T_DISCARD || type == VIRTIO_BLK_T_WRITE_ZEROES) { >> 245 if (virtblk_setup_discard_write_zeroes(req, unmap)) >> 246 return BLK_STS_RESOURCE; >> 247 } >> 248 >> 249 return 0; >> 250 } >> 251 >> 252 static inline void virtblk_request_done(struct request *req) >> 253 { >> 254 struct virtblk_req *vbr = blk_mq_rq_to_pdu(req); >> 255 >> 256 virtblk_unmap_data(req, vbr); >> 257 virtblk_cleanup_cmd(req); >> 258 blk_mq_end_request(req, virtblk_result(vbr)); >> 259 } >> 260 >> 261 static void virtblk_done(struct virtqueue *vq) >> 262 { >> 263 struct virtio_blk *vblk = vq->vdev->priv; >> 264 bool req_done = false; >> 265 int qid = vq->index; >> 266 struct virtblk_req *vbr; >> 267 unsigned long flags; >> 268 unsigned int len; >> 269 >> 270 spin_lock_irqsave(&vblk->vqs[qid].lock, flags); >> 271 do { >> 272 virtqueue_disable_cb(vq); >> 273 while ((vbr = virtqueue_get_buf(vblk->vqs[qid].vq, &len)) != NULL) { >> 274 struct request *req = blk_mq_rq_from_pdu(vbr); >> 275 >> 276 if (likely(!blk_should_fake_timeout(req->q))) >> 277 blk_mq_complete_request(req); >> 278 req_done = true; >> 279 } >> 280 if (unlikely(virtqueue_is_broken(vq))) >> 281 break; >> 282 } while (!virtqueue_enable_cb(vq)); >> 283 >> 284 /* In case queue is stopped waiting for more buffers. */ >> 285 if (req_done) >> 286 blk_mq_start_stopped_hw_queues(vblk->disk->queue, true); >> 287 spin_unlock_irqrestore(&vblk->vqs[qid].lock, flags); >> 288 } >> 289 >> 290 static void virtio_commit_rqs(struct blk_mq_hw_ctx *hctx) >> 291 { >> 292 struct virtio_blk *vblk = hctx->queue->queuedata; >> 293 struct virtio_blk_vq *vq = &vblk->vqs[hctx->queue_num]; >> 294 bool kick; >> 295 >> 296 spin_lock_irq(&vq->lock); >> 297 kick = virtqueue_kick_prepare(vq->vq); >> 298 spin_unlock_irq(&vq->lock); >> 299 >> 300 if (kick) >> 301 virtqueue_notify(vq->vq); >> 302 } >> 303 >> 304 static blk_status_t virtio_queue_rq(struct blk_mq_hw_ctx *hctx, >> 305 const struct blk_mq_queue_data *bd) >> 306 { >> 307 struct virtio_blk *vblk = hctx->queue->queuedata; >> 308 struct request *req = bd->rq; >> 309 struct virtblk_req *vbr = blk_mq_rq_to_pdu(req); >> 310 unsigned long flags; >> 311 unsigned int num; >> 312 int qid = hctx->queue_num; >> 313 int err; >> 314 bool notify = false; >> 315 >> 316 BUG_ON(req->nr_phys_segments + 2 > vblk->sg_elems); >> 317 >> 318 err = virtblk_setup_cmd(vblk->vdev, req, vbr); >> 319 if (unlikely(err)) >> > 320 return err; >> 321 >> 322 blk_mq_start_request(req); >> 323 >> 324 num = virtblk_map_data(hctx, req, vbr); >> 325 if (unlikely(num < 0)) { >> 326 virtblk_cleanup_cmd(req); >> 327 return BLK_STS_RESOURCE; >> 328 } >> 329 >> 330 spin_lock_irqsave(&vblk->vqs[qid].lock, flags); >> 331 err = virtblk_add_req(vblk->vqs[qid].vq, vbr, vbr->sg_table.sgl, num); >> 332 if (err) { >> 333 virtqueue_kick(vblk->vqs[qid].vq); >> 334 /* Don't stop the queue if -ENOMEM: we may have failed to >> 335 * bounce the buffer due to global resource outage. >> 336 */ >> 337 if (err == -ENOSPC) >> 338 blk_mq_stop_hw_queue(hctx); >> 339 spin_unlock_irqrestore(&vblk->vqs[qid].lock, flags); >> 340 virtblk_unmap_data(req, vbr); >> 341 virtblk_cleanup_cmd(req); >> 342 switch (err) { >> 343 case -ENOSPC: >> 344 return BLK_STS_DEV_RESOURCE; >> 345 case -ENOMEM: >> 346 return BLK_STS_RESOURCE; >> 347 default: >> 348 return BLK_STS_IOERR; >> 349 } >> 350 } >> 351 >> 352 if (bd->last && virtqueue_kick_prepare(vblk->vqs[qid].vq)) >> 353 notify = true; >> 354 spin_unlock_irqrestore(&vblk->vqs[qid].lock, flags); >> 355 >> 356 if (notify) >> 357 virtqueue_notify(vblk->vqs[qid].vq); >> 358 return BLK_STS_OK; >> 359 } >> 360 >> >> --- >> 0-DAY CI Kernel Test Service, Intel Corporation >> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.0... >