Alexander Motin
Alexander Motin
As I see, in most of cases (I've spotted only one different) when you are taking `db_rwlock`, you also take `db_mtx`. It makes no sense to me, unless the only...
@asomers Let me rephrase the key points: - Indirects and L0 dnode dbufs are special in having only one data copy ever. They are always decompressed in memory, and if...
@asomers Despite my comments on many of the changes here, IIRC there were some that could be useful. Do you plan to clean this up, document, etc, or I'll have...
Seems this was duplicated by #17546, that just landed. I'm sorry this PR was lost.
> not very nice, especially given that the merged fix is in my view incomplete What would be nice is to have more of regular reviewers, so that things would...
> I could think of an implementation of this which would be nice to use and maintain, both, that's why I'd go for a module option + envvar Aside of...
It was brought up on yesterday leadership call, and discussed that it would be nice instead of a global switch (which would be a permanent "holy war") to somehow identify...
I've just manually run `./raidz_test -S -e -t 300` on FreeBSD and observed it gradually consuming >106GB of RAM before completing successfully. I bet something is leaking there inside the...
I think it should start from a different side. First we should separate case when L2ARC is empty from cases of overwrite. In the first case we can write write...
With writing faster I did not mean it is CPU-bound, but that it should be able to scan multiple sublists evenly and may be use markers similar to eviction code...