When stacking devices, a request_queue is not always available. This forced us to have a no_cluster flag in the queue_limits that could be used as a carrier until the request_queue had been set up for a metadevice. There were several problems with that approach. First of all it was up to the stacking device to remember to set queue flag after stacking had completed. Also, the queue flag and the queue limits had to be kept in sync at all times. We got that wrong, which could lead to us issuing commands that went beyond the max scatterlist limit set by the driver. The proper fix is to avoid having two flags for tracking the same thing. We deprecate QUEUE_FLAG_CLUSTER and use the queue limit directly in the block layer merging functions. The queue_limit 'no_cluster' is turned into 'cluster' to avoid double negatives and to ease stacking. Clustering defaults to being enabled as before. The queue flag logic is removed from the stacking function, and explicitly setting the cluster flag is no longer necessary in DM and MD.
Showing the changes to include/linux (1 of 7 files). Show all
Unified · Split
| @@ -250,7 +250,7 @@ struct queue_limits { · 249 unmodified lines | ||
| 250 | 250 | |
| 251 | 251 | unsigned char misaligned; |
| 252 | 252 | unsigned char discard_misaligned; |
| 253 | − unsigned char no_cluster; | |
| 253 | + unsigned char cluster; | |
| 254 | 254 | signed char discard_zeroes_data; |
| 255 | 255 | }; |
| 256 | 256 | |
| @@ -380,7 +380,6 @@ struct request_queue · 123 unmodified lines | ||
| 380 | 380 | #endif |
| 381 | 381 | }; |
| 382 | 382 | |
| 383 | −#define QUEUE_FLAG_CLUSTER 0 /* cluster several segments into 1 */ | |
| 384 | 383 | #define QUEUE_FLAG_QUEUED 1 /* uses generic tag queueing */ |
| 385 | 384 | #define QUEUE_FLAG_STOPPED 2 /* queue is stopped */ |
| 386 | 385 | #define QUEUE_FLAG_SYNCFULL 3 /* read queue has been filled */ |
| @@ -403,7 +402,6 @@ struct request_queue · 16 unmodified lines | ||
| 403 | 402 | #define QUEUE_FLAG_SECDISCARD 19 /* supports SECDISCARD */ |
| 404 | 403 | |
| 405 | 404 | #define QUEUE_FLAG_DEFAULT ((1 << QUEUE_FLAG_IO_STAT) | \ |
| 406 | − (1 << QUEUE_FLAG_CLUSTER) | \ | |
| 407 | 405 | (1 << QUEUE_FLAG_STACKABLE) | \ |
| 408 | 406 | (1 << QUEUE_FLAG_SAME_COMP) | \ |
| 409 | 407 | (1 << QUEUE_FLAG_ADD_RANDOM)) |
| @@ -510,6 +508,11 @@ static inline void queue_flag_clear(unsigned int flag, struct request_queue *q) · 100 unmodified lines | ||
| 510 | 508 | |
| 511 | 509 | #define rq_data_dir(rq) ((rq)->cmd_flags & 1) |
| 512 | 510 | |
| 511 | +static inline unsigned int blk_queue_cluster(struct request_queue *q) | |
| 512 | +{ | |
| 513 | + return q->limits.cluster; | |
| 514 | +} | |
| 515 | + | |
| 513 | 516 | /* |
| 514 | 517 | * We regard a request as sync, if either a read or a sync write |
| 515 | 518 | */ |