| mav | |
| avg |
As ZFS can request up to SPA_MAXBLOCKSIZE memory block e.g. during zfs recv, update the threshold at which we start agressive reclamation to use SPA_MAXBLOCKSIZE (16M) instead of the lower zfs_max_recordsize which defaults to 1M.
This was reported here:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=194513
| Automatic diff as part of commit; lint not applicable. |
| Automatic diff as part of commit; unit tests not applicable. |
I have no objections. It may be too aggressive if fragmentation was somehow caused not by ZFS, but any way I don't know any other place in system which would bother to do this, while I know other places that suffer quietly from fragmentation.