On Mon, Mar 24, 2025 at 01:00:11AM -0700, Gleb Smirnoff wrote: T> This is an automated email to inform you that the March 2025 stabilization week T> started with FreeBSD/main at main-n276073-67c1c4dfd1cc, which was tagged as T> main-stabweek-2025-Mar.
At Netflix testing we didn't find any stability or performance regressions with the stabweek tag. Several people (including myself) updated their laptops and workstations and confirmed everything went smooth. This time I got more feedback than usual, several people emailed me privately or on a media outside of the freebsd- current@ mailing list. Very glad and grateful seeing more participation! Thanks everyone! The stabweek is over, advisory freeze is thawed! Note 1: those running on AMD GPUs are adviced to use 718d1928f874 hash for update, or cherry-pick just this hash on top of the stabweek hash. It fixes a long standing bug (not a recent regression!) and unfortunately was pushed after the stabweek was tagged. Note 2: There is a new LOR between tcpinpcb and socket buffer lock if you use kTLS and kernel is compiled with INVARIANTS. We assume this LOR safe, and not having any chance of a deadlock. Nevertheless, I'll try to figure out how to get rid of it. -- Gleb Smirnoff
