Dylan Reid 503c5abef6 devices: Add an asynchronous block device ... This enables the use of basic disk images with async IO. A new block_async.rs is added which mostly mirrors block, except that all IO operations are asynchronous allowing for multiple virt queues to be used. The old block remains unchanged and is still used for qcow, android sparse, and composite disks. Those should be converted to as time allows, but this dual approach will have to do for now so ARCVM disk performance can be properly evaluated. fio --ioengine=libaio --randrepeat=1 --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randrw --rwmixread=75 desktop with nvme: before: READ: bw=36.2MiB/s (37.9MB/s), 36.2MiB/s-36.2MiB/s (37.9MB/s-37.9MB/s), io=3070MiB (3219MB), run=84871-84871msec WRITE: bw=12.1MiB/s (12.7MB/s), 12.1MiB/s-12.1MiB/s (12.7MB/s-12.7MB/s), io=1026MiB (1076MB), run=84871-84871msec after: READ: bw=257MiB/s (269MB/s), 257MiB/s-257MiB/s (269MB/s-269MB/s), io=3070MiB (3219MB), run=11964-11964msec WRITE: bw=85.8MiB/s (89.9MB/s), 85.8MiB/s-85.8MiB/s (89.9MB/s-89.9MB/s), io=1026MiB (1076MB), run=11964-11964msec samus with 5.6 kernel before: READ: bw=55.3MiB/s (57.9MB/s), 55.3MiB/s-55.3MiB/s (57.9MB/s-57.9MB/s), io=768MiB (805MB), run=13890-13890msec WRITE: bw=18.5MiB/s (19.4MB/s), 18.5MiB/s-18.5MiB/s (19.4MB/s-19.4MB/s), io=256MiB (269MB), run=13890-13890msec after: READ: bw=71.2MiB/s (74.7MB/s), 71.2MiB/s-71.2MiB/s (74.7MB/s-74.7MB/s), io=3070MiB (3219MB), run=43096-43096msec WRITE: bw=23.8MiB/s (24.0MB/s), 23.8MiB/s-23.8MiB/s (24.0MB/s-24.0MB/s), io=1026MiB (1076MB), run=43096-43096msec kevin with 5.6 kernel before: READ: bw=12.9MiB/s (13.5MB/s), 12.9MiB/s-12.9MiB/s (13.5MB/s-13.5MB/s), io=1534MiB (1609MB), run=118963-118963msec WRITE: bw=4424KiB/s (4530kB/s), 4424KiB/s-4424KiB/s (4530kB/s-4530kB/s), io=514MiB (539MB), run=118963-118963msec after: READ: bw=12.9MiB/s (13.5MB/s), 12.9MiB/s-12.9MiB/s (13.5MB/s-13.5MB/s), io=1534MiB (1609MB), run=119364-119364msec WRITE: bw=4409KiB/s (4515kB/s), 4409KiB/s-4409KiB/s (4515kB/s-4515kB/s), io=514MiB (539MB), run=119364-119364msec eve with nvme and 5.7 kernel before: READ: bw=49.4MiB/s (51.8MB/s), 49.4MiB/s-49.4MiB/s (51.8MB/s-51.8MB/s), io=3070MiB (3219MB), run=62195-62195msec WRITE: bw=16.5MiB/s (17.3MB/s), 16.5MiB/s-16.5MiB/s (17.3MB/s-17.3MB/s), io=1026MiB (1076MB), run=62195-62195msec after READ: bw=125MiB/s (131MB/s), 125MiB/s-125MiB/s (131MB/s-131MB/s), io=3070MiB (3219MB), run=24593-24593msec WRITE: bw=41.7MiB/s (43.7MB/s), 41.7MiB/s-41.7MiB/s (43.7MB/s-43.7MB/s), io=1026MiB (1076MB), run=24593-24593msec rammus with 5.10 kernel before: READ: bw=6927KiB/s (7093kB/s), 6927KiB/s-6927KiB/s (7093kB/s-7093kB/s), io=3070MiB (3219MB), run=453822-453822msec WRITE: bw=2315KiB/s (2371kB/s), 2315KiB/s-2315KiB/s (2371kB/s-2371kB/s), io=1026MiB (1076MB), run=453822-453822msec after: Run status group 0 (all jobs): READ: bw=10.0MiB/s (11.5MB/s), 10.0MiB/s-10.0MiB/s (11.5MB/s-11.5MB/s), io=3070MiB (3219MB), run=279111-279111msec WRITE: bw=3764KiB/s (3855kB/s), 3764KiB/s-3764KiB/s (3855kB/s-3855kB/s), io=1026MiB (1076MB), run=279111-279111msec BUG=chromium:901139 TEST=unitests TEST=boot a test image and run fio tests from the guest to measure speed. TEST=start ARCVM TEST=tast run $DUT crostini.ResizeOk.dlc_stretch_stable Change-Id: Idb63628871d0352bd18501a69d9c1c887c37607b Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2306786 Tested-by: Keiichi Watanabe <keiichiw@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Reviewed-by: Dylan Reid <dgreid@chromium.org> Reviewed-by: Chirantan Ekbote <chirantan@chromium.org> Reviewed-by: Daniel Verkamp <dverkamp@chromium.org> Commit-Queue: Keiichi Watanabe <keiichiw@chromium.org>		2021-02-17 04:11:55 +00:00
..
src	devices: Add an asynchronous block device	2021-02-17 04:11:55 +00:00
.build_test_skip	build_test: misc options, improvements, amd bug fixes	2020-10-15 13:40:32 +00:00
Cargo.toml	cros_async: Refactor executors	2021-01-25 08:52:27 +00:00