On 2020-03-30 20:08, Graham Perrin wrote:
Theron, thank you! That might explain a recent sense of randomness with the symptoms. I've been testing in three different boot environments, only one of which is above the 358629 at <https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=194727#c22>:
I don't know what would be attaching to mixer before login. For me the problem is predictable: suspending after I've opened a player reliably hangs kernel, and killing every user of /dev/dsp* /dev/mixer* before sleep reliably avoids the problem. Maybe something is attaching on startup.
(In retrospect, most remarkable to me was that the bug could bite before logging in; taking the 'Sleep' option in sddm did not lead to a suspend of the system. Hopefully fixed for FreeBSD-CURRENT by r358629; I'll not test now, maybe later in the week.)
As I understand it, r358629 adds a mechanism that applications can use to "do the right thing" but does not fix the underlying kernel vulnerability.
Theron _______________________________________________ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
