On Sun, 2026-04-26 at 05:28 -0700, Andrew Morton wrote: > Naive questions... > > On Sun, 26 Apr 2026 07:56:08 -0400 Jeff Layton <[email protected]> wrote: > > > The IOCB_DONTCACHE writeback path in generic_write_sync() calls > > filemap_flush_range() on every write, submitting writeback inline in > > the writer's context. Perf lock contention profiling shows the > > performance problem is not lock contention but the writeback submission > > work itself — walking the page tree and submitting I/O blocks the writer > > for milliseconds, inflating p99.9 latency from 23ms (buffered) to 93ms > > (dontcache). > > So in the current case, when generic_write_sync() returns, all that > memory is written back and clean&reclaimable (or freed?), yes? >
No. Before returning, it submits the I/Os for the portion that it wrote rather than leaving it to the flusher to take care of things, but it doesn't wait for the I/Os to complete. > > Replace the inline filemap_flush_range() call with a flusher kick that > > drains dirty pages in the background. This moves writeback submission > > completely off the writer's hot path. > > Whereas after this change, that pagecache is probably still dirty, > unreclaimable, waiting for the flusher to do its thing? > Correct, but that's sort of the case today too since DONTCACHE I/Os don't wait for the completion. With this change we're just deferring the I/O submission to the flusher thread (which should hopefully soon wake and take care of business). If the flusher thread can't keep up, then eventually balance_dirty_pages() will kick in and start slowing things down. > So is there potential that the system will get all gummed up with > dirty, to-be-written-soon pagecache? Is there something which limits > this buildup? > Today in this situation, the writers are limited by the backing device throughput. Once the I/O submission queues are full, then the DONTCACHE writers end up stacking up on those. With this change, the writers will be more limited by traditional VM limits in this situation. In the test runs I did, the peak pagecache with DONTCACHE writes was higher than with the unpatched version but still considerably less than with normal buffered I/O. That's the cost of deferring the I/O submission to the flusher. One thing we could consider is going back to submitting the writes inline when the number of dirty pages is high. But, that could have a detrimental effect on performance too. > > ... > > > > dontcache-bench results on dual-socket Xeon Gold 6138 (80 CPUs, 256 GB > > RAM, Samsung MZ1LB1T9HALS 1.7 TB NVMe, local XFS, io_uring, file size > > ~503 GB, compared to a v6.19-ish baseline): > > > > Single-client sequential write (MB/s): > > baseline patched change > > buffered 1449.8 1440.1 -0.7% > > dontcache 1347.9 1461.5 +8.4% > > direct 1450.0 1440.1 -0.7% > > > > Single-client sequential write latency (us): > > baseline patched change > > dontcache p50 3031.0 10551.3 +248.1% > > dontcache p99 74973.2 21626.9 -71.2% > > dontcache p99.9 85459.0 23199.7 -72.9% > > > > Single-client random write (MB/s): > > baseline patched change > > dontcache 284.2 295.4 +3.9% > > > > Single-client random write p99.9 latency (us): > > baseline patched change > > dontcache 2277.4 872.4 -61.7% > > > > Multi-writer aggregate throughput (MB/s): > > baseline patched change > > buffered 1619.5 1611.2 -0.5% > > dontcache 1281.1 1629.4 +27.2% > > direct 1545.4 1609.4 +4.1% > > > > Mixed-mode noisy neighbor (dontcache writer + buffered readers): > > baseline patched change > > writer (MB/s) 1297.6 1471.1 +13.4% > > readers avg (MB/s) 855.0 462.4 -45.9% > > These results look ambiguous. Sometimes better, sometimes worse? > > > nfsd-io-bench results on same hardware (XFS on NVMe, NFSv3 via fio > > NFS engine with libnfs, 1024 NFSD threads, pool_mode=pernode, > > file size ~502 GB, compared to v6.19-ish baseline): > > > > Single-client sequential write (MB/s): > > baseline patched change > > buffered 4844.2 4653.4 -3.9% > > dontcache 3028.3 3723.1 +22.9% > > direct 957.6 987.8 +3.2% > > > > Single-client sequential write p99.9 latency (us): > > baseline patched change > > dontcache 759169.0 175112.2 -76.9% > > > > Single-client random write (MB/s): > > baseline patched change > > dontcache 590.0 1561.0 +164.6% > > > > Multi-writer aggregate throughput (MB/s): > > baseline patched change > > buffered 9636.3 9422.9 -2.2% > > dontcache 1894.9 9442.6 +398.3% > > direct 809.6 975.1 +20.4% > > > > Noisy neighbor (dontcache writer + random readers): > > baseline patched change > > writer (MB/s) 1854.5 4063.6 +119.1% > > readers avg (MB/s) 131.2 101.6 -22.5% > > Ditto but less so. > > > The NFS results show even larger improvements than the local benchmarks. > > Multi-writer dontcache throughput improves nearly 5x, matching buffered > > I/O. Dirty page footprint drops 85-95% in sequential workloads vs. > > buffered. > > It sounds that you like the results, so OK ;) -- Jeff Layton <[email protected]>
