Rob Norris
Rob Norris
Ok, last commit is the final form I believe. Rebased to master, now that #17420 and #17435 are merged. Here I add `zil_commit_flags()`, following the trend set by `dmu_tx_assign_flags()` and...
Last push fixes the two comments mentioned, and a handful of other crimes against English. @behlendorf thank you for your time!
> "its complicated" 😰
@amotin asked me what would happen if in `zil_lwb_write_done()`, we call `list_next(&zilog->zl_lwb_list, lwb)` while `lwb` is not on `zl_lwb_list` but rather, on `zl_lwb_crash_list`. I said "nothing good, I bet" and...
Last push rebases and resolves the significant conflicts. I'm not entirely sure if the async/sync counters were actually handling it before, or if they were hiding some details from view,...
Thank you both for your thoughtful review. I know time and effort it takes and I really appreciate it.
Last push just rebasing to master in preparation for more work, nothing to see here.
Experimental work from yesterday and today pushed. First, `zfs_slack_compress` is converted to operate on ABDs directly, rather than linearising. In simple tests this makes very little difference, as the IO...
> May be I've missed something, but what will make transactions move and more metaslabs flushed if the pool is idle? > > Also, if the pool is idle, will...