From mboxrd@z Thu Jan 1 00:00:00 1970 From: Joel Becker Date: Fri, 28 Aug 2009 14:49:43 -0700 Subject: [Ocfs2-devel] [PATCH] ocfs2: adds mlogs to aops.c -V3 In-Reply-To: <4A9789E5.9090205@oracle.com> References: <200908140604.n7E64rhS031970@acsinet15.oracle.com> <20090828011018.GG18962@mail.oracle.com> <4A973D1D.70500@oracle.com> <20090828053911.GB19658@mail.oracle.com> <4A978098.60306@oracle.com> <20090828073359.GC19658@mail.oracle.com> <4A9789E5.9090205@oracle.com> Message-ID: <20090828214943.GB29312@mail.oracle.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ocfs2-devel@oss.oracle.com On Fri, Aug 28, 2009 at 03:40:21PM +0800, Wengang Wang wrote: > Ok. Then I will pick up the mlog_error()s and mlog(ERR...)s and re-post > later... Sounds good. If the area of logging interests you, would you take a stab at adding tracepoints to ocfs2? Mainline has a generic infrastructure for "tracepoints" that can be dropped in nicely in code and don't have to be printk()d. Debugfs files can be used to enable them or disable them as a group or individually, so we could even just say "only trace ocfs2_new_inode". The tracing can be filtered by pid or by function, too. See Documentation/trace/tracepoints.txt for a little overview, but really look at include/trace/events/ext4.h and the associated calls in fs/ext4. I'd love to see a small include/trace/events/ocfs2.h to start. Maybe just give aio_write and aio_read callers or something. Just so there is a nice example for others to follow. include/trace/events/o2dlm.h would be cool too, we could trace locking calls. Feel free to ask me any questions. Joel -- Life's Little Instruction Book #314 "Never underestimate the power of forgiveness." Joel Becker Principal Software Developer Oracle E-mail: joel.becker at oracle.com Phone: (650) 506-8127