Rob Norris

I compiled 6.1.83 with `CONFIG_INIT_STACK_ALL_PATTERN=y`, with OpenZFS master 21bc066ec. Creating a new dataset with `encryption=aes-256-gcm`, it completed a write cycle, scrub, export and reimport without issue. So, more info required....

@rincebrain > A counter for `fpu_enter`/`fpu_exit` (or just more general critical section enter/exit) calls in the taskq might also be informative, as would if your taskq triggers kmem cache growth,...

Another option would be to define those two if they're missing: ``` #ifndef SEEK_DATA #define SEEK_DATA 3 #endif #ifndef SEEK_HOLE #define SEEK_HOLE 4 #endif ``` That particular value gets passed...

Don't know this code super well, but I've been reading it and @amotin seems right. > is there anything to free zap_leaf_t associated with deleted leafs before dbuf eviction code...

Confirmed. Quick fix: revert 71216b91d281e7e58f5e29ca4d4553945e080fe9 (#16126 #16129). (My Python skills are nonexistent, so I'm not much use on a fix). I have the same output as above. Note also `dist-packages`...

The difference seems to be whether or not the architecture has support for "sg chaining". I think any fix is a little more complicated than that, but yeah, we could...

@tonyhutter do you have some specific use cases in mind for this? Any design notes? I'm not sure yet what to make of this as a concept, and I'm not...

Yeah, my first thought before I read the entire description was "`import -m` but for cache devices". Seems straightforward and worthwhile.

@osleg can you post `/proc/spl/kmem/slab` from before and after the OOM event? Doesn't need to be exact, but I'd like to see what happens as more files are deleted, into...

> As the reporter for #15930 , I am obligated to note that running OpenZFS on _not yet officially supported kernels_ is exactly how I find bugs to report. >...