On Wed, 9 Nov 2016, Russell King - ARM Linux wrote: > Please include Thomas in this. Thanks! > On Wed, Nov 09, 2016 at 10:46:21AM +0200, Horia Geantă wrote: > > This reverts commit 66d2e2028091a074aa1290d2eeda5ddb1a6c329c. > > > > Quoting from Russell's findings: > > https://www.mail-archive.com/linux-crypto@xxxxxxxxxxxxxxx/msg21136.html > > > > [quote] > > Okay, I've re-tested, using a different way of measuring, because using > > openssl speed is impractical for off-loaded engines. I've decided to > > use this way to measure the performance: > > > > dd if=/dev/zero bs=1048576 count=128 | /usr/bin/time openssl dgst -md5 > > > > For the threaded IRQs case gives: > > > > 0.05user 2.74system 0:05.30elapsed 52%CPU (0avgtext+0avgdata 2400maxresident)k > > 0.06user 2.52system 0:05.18elapsed 49%CPU (0avgtext+0avgdata 2404maxresident)k > > 0.12user 2.60system 0:05.61elapsed 48%CPU (0avgtext+0avgdata 2460maxresident)k > > => 5.36s => 25.0MB/s > > > > and the tasklet case: > > > > 0.08user 2.53system 0:04.83elapsed 54%CPU (0avgtext+0avgdata 2468maxresident)k > > 0.09user 2.47system 0:05.16elapsed 49%CPU (0avgtext+0avgdata 2368maxresident)k > > 0.10user 2.51system 0:04.87elapsed 53%CPU (0avgtext+0avgdata 2460maxresident)k > > => 4.95 => 27.1MB/s > > > > which corresponds to an 8% slowdown for the threaded IRQ case. So, > > tasklets are indeed faster than threaded IRQs. Each operation involves: submit job hard irq handler wakeup irq thread context switch wakeup openssl context switch and that repeats over and over. Russell has provided me traces and the extra wakeup/context switch is what causes the slowdown vs. the tasklet case which has only one wakeup/context switch, unless it gets outsourced to ksoftirqd which might get even slower than the threaded handler. You might add something like that to the changelog so it's not just a number comparison which does not explain why this happens. Thanks, tglx
