summaryrefslogtreecommitdiff
path: root/drivers
diff options
context:
space:
mode:
authorNathan Huckleberry <nhuck@google.com>2023-02-02 04:23:48 +0300
committerMike Snitzer <snitzer@kernel.org>2023-02-02 22:26:09 +0300
commitc25da5b7baf1d243e6612ba2b97e2a2c4a1376f6 (patch)
tree0e6b8982c450d3046619e4c22e26b10b003d0aac /drivers
parent5cd6d1d53a1f74222e73d8b42ab7ecf28ee2f34f (diff)
downloadlinux-c25da5b7baf1d243e6612ba2b97e2a2c4a1376f6.tar.xz
dm verity: stop using WQ_UNBOUND for verify_wq
Setting WQ_UNBOUND increases scheduler latency on ARM64. This is likely due to the asymmetric architecture of ARM64 processors. I've been unable to reproduce the results that claim WQ_UNBOUND gives a performance boost on x86-64. This flag is causing performance issues for multiple subsystems within Android. Notably, the same slowdown exists for decompression with EROFS. | open-prebuilt-camera | WQ_UNBOUND | ~WQ_UNBOUND | |-----------------------|------------|---------------| | verity wait time (us) | 11746 | 119 (-98%) | | erofs wait time (us) | 357805 | 174205 (-51%) | | sha256 ramdisk random read | WQ_UNBOUND | ~WQ_UNBOUND | |----------------------------|-----------=---|-------------| | arm64 (accelerated) | bw=42.4MiB/s | bw=212MiB/s | | arm64 (generic) | bw=16.5MiB/s | bw=48MiB/s | | x86_64 (generic) | bw=233MiB/s | bw=230MiB/s | Using a alloc_workqueue() @max_active arg of num_online_cpus() only made sense with WQ_UNBOUND. Switch the @max_active arg to 0 (aka default, which is 256 per-cpu). Also, eliminate 'wq_flags' since it really doesn't serve a purpose. Cc: Sami Tolvanen <samitolvanen@google.com> Cc: Eric Biggers <ebiggers@kernel.org> Signed-off-by: Nathan Huckleberry <nhuck@google.com> Reviewed-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Mike Snitzer <snitzer@kernel.org>
Diffstat (limited to 'drivers')
-rw-r--r--drivers/md/dm-verity-target.c6
1 files changed, 1 insertions, 5 deletions
diff --git a/drivers/md/dm-verity-target.c b/drivers/md/dm-verity-target.c
index ccf5b852fbf7..4c52b056ecf0 100644
--- a/drivers/md/dm-verity-target.c
+++ b/drivers/md/dm-verity-target.c
@@ -1162,7 +1162,6 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv)
struct dm_verity_sig_opts verify_args = {0};
struct dm_arg_set as;
unsigned int num;
- unsigned int wq_flags;
unsigned long long num_ll;
int r;
int i;
@@ -1399,8 +1398,6 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv)
goto bad;
}
- /* WQ_UNBOUND greatly improves performance when running on ramdisk */
- wq_flags = WQ_MEM_RECLAIM | WQ_UNBOUND;
/*
* Using WQ_HIGHPRI improves throughput and completion latency by
* reducing wait times when reading from a dm-verity device.
@@ -1410,8 +1407,7 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv)
* will fall-back to using it for error handling (or if the bufio cache
* doesn't have required hashes).
*/
- wq_flags |= WQ_HIGHPRI;
- v->verify_wq = alloc_workqueue("kverityd", wq_flags, num_online_cpus());
+ v->verify_wq = alloc_workqueue("kverityd", WQ_MEM_RECLAIM | WQ_HIGHPRI, 0);
if (!v->verify_wq) {
ti->error = "Cannot allocate workqueue";
r = -ENOMEM;