But I've noticed something odd.
To quickly test ZSWAP, I use eatmemory tool to allocate 7GB of 8GB on my system.
Allocating worked and ZSWAP activated but freeing 7GB of memory took ~45s (system was barely responsive) instead of usual 1-2s.
I tested on freshly booted machine on native text console (no Graphics DE) and I didn't see that behavior on other kernels even with very busy system.
Looks like 6.8 needs some work or maybe there were some commits that altered swap/zswap behavior after 6.7.1.
So if you are planning on backporting patches to 6.5 I'd base those on 6.7.1 instead of 6.8.
I am back to using Noble's 6.6.0-14.14 which works ok for now.
P.S.
When testing mainline 6.6.13 I've had dead keyboard after suspend on my XPS13-9350, not sure if Nobles 6.6 has some XPS related fixes or I just was lucky to not catch the bug again ;-)
Please let me know if I need to open a separate bug for that.
Hello Mathew,
Thanks for responding so promptly.
I did test 6.8-rc1 and ZSWAP works as expected.
But I've noticed something odd.
To quickly test ZSWAP, I use eatmemory tool to allocate 7GB of 8GB on my system.
Allocating worked and ZSWAP activated but freeing 7GB of memory took ~45s (system was barely responsive) instead of usual 1-2s.
I tested on freshly booted machine on native text console (no Graphics DE) and I didn't see that behavior on other kernels even with very busy system.
Looks like 6.8 needs some work or maybe there were some commits that altered swap/zswap behavior after 6.7.1.
So if you are planning on backporting patches to 6.5 I'd base those on 6.7.1 instead of 6.8.
I am back to using Noble's 6.6.0-14.14 which works ok for now.
P.S.
When testing mainline 6.6.13 I've had dead keyboard after suspend on my XPS13-9350, not sure if Nobles 6.6 has some XPS related fixes or I just was lucky to not catch the bug again ;-)
Please let me know if I need to open a separate bug for that.
Thanks,
Eugene.