block-queue: Delay and batch metadata write
Get series with b4
b4 am [email protected]Versions newest first
- v1 Latest4 patches 1 replyIn review
Patches in v1 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/4 | Make vm_stop available for block layercpus.c qemu-common.h qemu-tool.c +1 more | +13 | 0 replies | |
| 2/4 | Add block-queueMakefile.objs block-queue.c block-queue.h | +937 | 0 replies | |
| 3/4 | Test cases for block-queueMakefile check-block-queue.c | +403 | 0 replies | |
| 4/4 | qcow2: Use block-queueblock/ block/ block/ +2 more | +440 | 1 reply |
Reviews trailers per patch
No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.
Cover letter [Qemu-devel] [PATCH 0/4] block-queue: Delay and batch metadata write
Differences to RFC v3 include proper conversion of qcow2, addressing Stefan's comments and fixing some error cases in which two write requests to the same location might conflict.
Also worth noting is that bdrv_aio_pwrite is dropped. It was unsafe with respect to multiple concurrent requests on the same sector and it's impossible to safely emulate byte-wise access with bdrv_aio_readv/writev without introducing yet another queue. Instead we fall back to synchronous bdrv_pwrite now with unaligned requests in block-queue (they are rare).
Kevin Wolf (4):
Make vm_stop available for block layer
Add block-queue
Test cases for block-queue
qcow2: Use block-queue
Makefile | 1 + Makefile.objs | 2 +- block-queue.c | 875 ++++++++++++++++++++++++++++++++++++++++++++++++ block-queue.h | 61 ++++ block/qcow2-cluster.c | 139 +++++---- block/qcow2-refcount.c | 217 +++++++----- block/qcow2-snapshot.c | 106 +++++-- block/qcow2.c | 144 +++++++- block/qcow2.h | 33 ++- check-block-queue.c | 402 ++++++++++++++++++++++ cpus.c | 8 +- qemu-common.h | 3 + qemu-tool.c | 5 + sysemu.h | 1 - 14 files changed, 1793 insertions(+), 204 deletions(-) create mode 100644 block-queue.c create mode 100644 block-queue.h create mode 100644 check-block-queue.c
Signature · 1 line
1.7.2.3