Date: Tue, 22 Jun 2021 22:57:38 +0000 From: bugzilla-noreply@freebsd.org To: fs@FreeBSD.org Subject: [Bug 256511] UFS assertion failure when shutting down from single-user mode Message-ID: <bug-256511-3630-Cq1nP3Kww3@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-256511-3630@https.bugs.freebsd.org/bugzilla/> References: <bug-256511-3630@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D256511 --- Comment #6 from Robert Wing <rew@FreeBSD.org> --- (In reply to Konstantin Belousov from comment #5) Hmm, I have a question. Can the namei() call go before the MNT_UPDATE {} bl= ock and either return early or cache the result from namei() and skip initialization based on that? Here's why I ask that: softdep_mount() is what sets up um_softdep and softdep_mount() is only call= ed in the (MNT_UPDATE && upgrade from ro->rw) {} block. Also in that block of code, the superblock gets updated with new values for fs_mtime and fs_ronly. The geom access counters get incremented for exclusive write access. Seems = like those would also need to be cleaned up? I'll put the above idea (patch) on phabricator for comment: https://reviews.freebsd.org/D30870 --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-256511-3630-Cq1nP3Kww3>