Commit e5863d9ad ("dm: allocate requests in target when stacking on blk-mq devices") served as the first step toward fully utilizing blk-mq in request-based DM -- it enabled stacking an old-style (request_fn) request_queue ontop of the underlying blk-mq device(s). That first step didn't improve performance of DM multipath ontop of fast blk-mq devices (e.g. NVMe) because the top-level old-style request_queue was severely limited by the queue_lock. The second step offered here enables stacking a blk-mq request_queue ontop of the underlying blk-mq device(s). This unlocks significant performance gains on fast blk-mq devices, Keith Busch tested on his NVMe testbed and offered this really positive news: "Just providing a performance update. All my fio tests are getting roughly equal performance whether accessed through the raw block device or the multipath device mapper (~470k IOPS). I could only push ~20% of the raw iops through dm before this conversion, so this latest tree is looking really solid from a performance standpoint."
Showing the changes to include/uapi/linux (1 of 4 files). Show all
Unified · Split
| @@ -267,9 +267,9 @@ enum { · 266 unmodified lines | |||
| 267 | #define DM_DEV_SET_GEOMETRY _IOWR(DM_IOCTL, DM_DEV_SET_GEOMETRY_CMD, struct dm_ioctl) | 267 | #define DM_DEV_SET_GEOMETRY _IOWR(DM_IOCTL, DM_DEV_SET_GEOMETRY_CMD, struct dm_ioctl) |
| 268 | 268 | ||
| 269 | #define DM_VERSION_MAJOR 4 | 269 | #define DM_VERSION_MAJOR 4 |
| 270 | #define DM_VERSION_MINOR 30 | 270 | #define DM_VERSION_MINOR 31 |
| 271 | #define DM_VERSION_PATCHLEVEL 0 | 271 | #define DM_VERSION_PATCHLEVEL 0 |
| 272 | #define DM_VERSION_EXTRA "-ioctl (2014-12-22)" | 272 | #define DM_VERSION_EXTRA "-ioctl (2015-3-12)" |
| 273 | 273 | ||
| 274 | /* Status bits */ | 274 | /* Status bits */ |
| 275 | #define DM_READONLY_FLAG (1 << 0) /* In/Out */ | 275 | #define DM_READONLY_FLAG (1 << 0) /* In/Out */ |