Date: Fri, 20 Sep 2024 15:41:42 +0000 From: bugzilla-noreply@freebsd.org To: net@FreeBSD.org Subject: [Bug 281560] gve (4) uma deadlock during high tcp throughput Message-ID: <bug-281560-7501-A0GDKmZzy2@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-281560-7501@https.bugs.freebsd.org/bugzilla/> References: <bug-281560-7501@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=3D281560 Andrew Gallatin <gallatin@FreeBSD.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |gallatin@FreeBSD.org --- Comment #5 from Andrew Gallatin <gallatin@FreeBSD.org> --- (In reply to Mark Johnston from comment #4) To expand on this, the pattern of stack traces look very much like a degene= rate behavior in the UMA memory allocator when the per-cpu caches are empty, and= the allocation must go back to the zone, and you wind up taking a lock on every allocation. I'd check to see if there is a mbuf leak. The output of vmstat -z would be helpful here. --=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-281560-7501-A0GDKmZzy2>