Date: Sat, 18 May 2024 22:59:13 +0000 From: bugzilla-noreply@freebsd.org To: bugs@FreeBSD.org Subject: [Bug 279138] NFS and NFSUPG and BUFWAIT Message-ID: <bug-279138-227-Osv91KLJOU@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-279138-227@https.bugs.freebsd.org/bugzilla/> References: <bug-279138-227@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=3D279138 --- Comment #3 from Konstantin Belousov <kib@FreeBSD.org> --- The first LoR looks as a false positive, mostly. It is readdirplus doing scan of the directory, and then querying the attributes of each found entry' node, except dotdot. So the lock order is parent vnode -> (it's buffers) -> child vnode, as expected by VFS. I said "mostly" because server can do the directory move like from A->B to B->A while client is not aware, and I am not completely sure that our clien= t- side invalidations would avoid tricking in this situation. But this is very unlikely. For the second LoR, nfsupg->bufwait, the reported order is right. I wonder where the reversed order was recorded. That said, to diagnose the problem, you need to gather the information list= ed in the developer's handbook, debugging kernels, debugging deadlocks: https://docs.freebsd.org/en/books/developers-handbook/kerneldebug/#kernelde= bug-deadlocks --=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-279138-227-Osv91KLJOU>