Date: Mon, 3 Oct 2011 10:43:27 +0400 From: Lev Serebryakov <lev@FreeBSD.org> To: Lawrence Stewart <lstewart@freebsd.org> Cc: hackers@FreeBSD.org Subject: Re: alq_open_flags() panics in _mtx_lock_flags() Message-ID: <535056585.20111003104327@serebryakov.spb.ru> In-Reply-To: <4E89051D.30908@freebsd.org> References: <1258376930.20111002193223@serebryakov.spb.ru> <4E89051D.30908@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Hello, Lawrence. You wrote 3 =EE=EA=F2=FF=E1=F0=FF 2011 =E3., 4:43:09: > I don't see anything obviously wrong with your call to alq_open_flags(), > other than that I think you want to set your buffer size to be a=20 > multiple of "sizeof(struct g_log_entry)". The small queue size wouldn't > cause the panic you're seeing though. > It's hard to draw any real conclusions without seeing your code. Are you > able to share it? As I wrote to current@, fs@ and geom@, it seems, that vnode API could not be called from geom threads (g_events), at least with topology lock held. So, it is not ALQ-specific problem, but some mis-interaction between GEOM and VFS levels. --=20 // Black Lion AKA Lev Serebryakov <lev@FreeBSD.org>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?535056585.20111003104327>