From owner-freebsd-bugs@FreeBSD.ORG Sat Mar 8 21:02:26 2008 Return-Path: Delivered-To: freebsd-bugs@hub.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id B6D7D1065676; Sat, 8 Mar 2008 21:02:26 +0000 (UTC) (envelope-from rwatson@FreeBSD.org) Received: from freefall.freebsd.org (freefall.freebsd.org [IPv6:2001:4f8:fff6::28]) by mx1.freebsd.org (Postfix) with ESMTP id A697B8FC12; Sat, 8 Mar 2008 21:02:26 +0000 (UTC) (envelope-from rwatson@FreeBSD.org) Received: from freefall.freebsd.org (rwatson@localhost [127.0.0.1]) by freefall.freebsd.org (8.14.2/8.14.2) with ESMTP id m28L2QfY065152; Sat, 8 Mar 2008 21:02:26 GMT (envelope-from rwatson@freefall.freebsd.org) Received: (from rwatson@localhost) by freefall.freebsd.org (8.14.2/8.14.1/Submit) id m28L2Q3s065148; Sat, 8 Mar 2008 21:02:26 GMT (envelope-from rwatson) Date: Sat, 8 Mar 2008 21:02:26 GMT Message-Id: <200803082102.m28L2Q3s065148@freefall.freebsd.org> To: dev@26th.net, rwatson@FreeBSD.org, freebsd-bugs@FreeBSD.org From: rwatson@FreeBSD.org Cc: Subject: Re: kern/94269: [nullfs] procfs shows wrong data if executable is running from nullfs X-BeenThere: freebsd-bugs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Bug reports List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 08 Mar 2008 21:02:26 -0000 Synopsis: [nullfs] procfs shows wrong data if executable is running from nullfs State-Changed-From-To: open->suspended State-Changed-By: rwatson State-Changed-When: Sat Mar 8 20:59:03 UTC 2008 State-Changed-Why: This is an accurate bug report -- procfs and other related tools rely on the name cache to generate names for file system objects on demand. However, nullfs doesn't use the name cache, as the underlying file system layer will already be performing its own caching on the name space. Transition this to suspended as it's not clear what the right fix for this is. From a performance perspective, it doesn't make a lot of sense to use the name cache for nullfs, and there is otherwise no good source of name data with which to build a name for an arbitrary vnode. There is also no cache invalidation of "higher layers" with file system stacking for the name cache, so quite a lot of work would have to be done in order to implement it with proper name cache invalidation. Thanks for the report; hopefully we can interest someone in further investigating this in the future. http://www.freebsd.org/cgi/query-pr.cgi?pr=94269