When we end I/O struct request with error, we need to pass obj_request->length as @nr_bytes so that the entire obj_request worth of bytes is completed. Otherwise block layer ends up confused and we trip on rbd_assert(more ^ (which == img_request->obj_request_count)); in rbd_img_obj_callback() due to more being true no matter what. We already do it in most cases but we are missing some, in particular those where we don't even get a chance to submit any obj_requests, due to an early -ENOMEM for example. A number of obj_request->xferred assignments seem to be redundant but I haven't touched any of obj_request->xferred stuff to keep this small and isolated.
Unified · Split
| @@ -2264,6 +2264,11 @@ static bool rbd_img_obj_end_request(struct rbd_obj_request *obj_request) · 2263 unmodified lines | |||
| 2264 | result, xferred); | 2264 | result, xferred); |
| 2265 | if (!img_request->result) | 2265 | if (!img_request->result) |
| 2266 | img_request->result = result; | 2266 | img_request->result = result; |
| 2267 | /* | ||
| 2268 | * Need to end I/O on the entire obj_request worth of | ||
| 2269 | * bytes in case of error. | ||
| 2270 | */ | ||
| 2271 | xferred = obj_request->length; | ||
| 2267 | } | 2272 | } |
| 2268 | 2273 | ||
| 2269 | /* Image object requests don't own their page array */ | 2274 | /* Image object requests don't own their page array */ |