On Mon, Apr 21, 2025 at 01:00:13AM -0700, Gleb Smirnoff wrote: T> This is an automated email to inform you that the April 2025 stabilization week T> started with FreeBSD/main at main-n276625-d4763484f911, which was tagged as T> main-stabweek-2025-Apr.
The A/B testing at Netflix didn't find any stability or performance regressions compared to the March stabweek. My personal experience with a home router and desktop was also successful. The users of drm-kmod (mostly laptops) faced a problem. The packages available at the beginning of the stabweek refused to load on main-n276625-d4763484f911 due to broken KPI. Wednesday night required fixes were added to the drm-kmod family of ports. The drm-know packages available right now in the official repo will still refuse to load on main-n276625-d4763484f911. However, manually built drm-kmod from fresh ports will work. Hopefully, we will see updated packages pretty soon, too. There is also a report of regression in iichid(4), also hitting some laptops: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=286045 With these two known issues we are ending the stabweek advisory code freeze. P.S. By the way, speaking of code freeze. This time it was not really held by a number of developers. Don't want to single any one out. Fortunately the stability/performance testing was successful, which made the stabweek fairly easy. If it happened to be the opposite, then the flow of functional check-ins during beginning of the week would really disrupt efforts to quickly diagnose and fix those regressions. Given that I was on vacation, and Warner was running the week for a first time, that ignoring of advisory freeze would not be a helpful thing at all. -- Gleb Smirnoff
