A previous commit fixed a poll race that can occur, but it's only applicable for multishot requests. For a multishot request, we can safely ignore a spurious wakeup, as we never leave the waitqueue to begin with. A blunt reissue of a multishot armed request can cause us to leak a buffer, if they are ring provided. While this seems like a bug in itself, it's not really defined behavior to reissue a multishot request directly. It's less efficient to do so as well, and not required to rearm anything like it is for singleshot poll requests.
Unified · Split
| @@ -283,8 +283,12 @@ static int io_poll_check_events(struct io_kiocb *req, bool *locked) · 282 unmodified lines | ||
| 283 | 283 | * to the waitqueue, so if we get nothing back, we |
| 284 | 284 | * should be safe and attempt a reissue. |
| 285 | 285 | */ |
| 286 | − if (unlikely(!req->cqe.res)) | |
| 286 | + if (unlikely(!req->cqe.res)) { | |
| 287 | + /* Multishot armed need not reissue */ | |
| 288 | + if (!(req->apoll_events & EPOLLONESHOT)) | |
| 289 | + continue; | |
| 287 | 290 | return IOU_POLL_REISSUE; |
| 291 | + } | |
| 288 | 292 | } |
| 289 | 293 | if (req->apoll_events & EPOLLONESHOT) |
| 290 | 294 | return IOU_POLL_DONE; |