Andrew Morton
2012-Feb-23 09:57 UTC
Re: [PATCH 1/2] treewide: fix memory corruptions when TASK_COMM_LEN != 16
On Thu, 23 Feb 2012 10:09:33 +0100 (CET) Jan Engelhardt <jengelh@medozas.de> wrote:> On Wednesday 2012-02-22 21:58, Andrew Morton wrote: > > >On Wed, 22 Feb 2012 13:48:08 +0100 (CET) > >Jan Engelhardt <jengelh@medozas.de> wrote: > > > >> task: provide a larger task command buffer > > > ><scratches head> > > > >Why are we bothering ourselves about this? > > Some prefer to know what''s going on in the system. Every other or > so kernel release there are some new happy kthreads, such as > > 24930 ? S 0:00 \_ [btrfs-endio-1] > 24931 ? S 0:00 \_ [btrfs-endio-met] > 24932 ? S 0:00 \_ [btrfs-endio-met] > 24933 ? S 0:00 \_ [btrfs-endio-wri] > 24934 ? S 0:00 \_ [btrfs-freespace] > > at which point one is curious to find out the rest of the met and why > there are two of them. If expanded one actually sees they are different > kthreads (rather than just per-cpu instances for a WQ, for example) > > $ grep Name /proc/{29431,29432}/stat* > /proc/29431/status:Name: btrfs-endio-meta-1 > /proc/29432/status:Name: btrfs-endio-meta-write-1 > > That''s all.doh. The fix for that is to have less clueless btrfs developers.
Jan Engelhardt
2012-Feb-23 11:19 UTC
Re: [PATCH 1/2] treewide: fix memory corruptions when TASK_COMM_LEN != 16
On Thursday 2012-02-23 10:57, Andrew Morton wrote:>>But there''s more,>> >> 24931 ? S 0:00 \_ [btrfs-endio-met]\_ [kconservative/5] \_ [ext4-dio-unwrit]>> >> [with a wondersome patch:] $ grep Name /proc/{29431,29432}/stat* >> /proc/29431/status:Name: btrfs-endio-meta-1 >> /proc/29432/status:Name: btrfs-endio-meta-write-1Name: kconservative/512 Name: ext4-dio-unwritten> >doh. The fix for that is to have less clueless btrfs developers.And truncate their names to SUNWbtfs, ORCLintg and EXT4diou? I think not :)
Andrew Morton
2012-Feb-23 17:30 UTC
Re: [PATCH 1/2] treewide: fix memory corruptions when TASK_COMM_LEN != 16
On Thu, 23 Feb 2012 12:19:28 +0100 (CET) Jan Engelhardt <jengelh@medozas.de> wrote:> > On Thursday 2012-02-23 10:57, Andrew Morton wrote: > >> > But there''s more, > >> > >> 24931 ? S 0:00 \_ [btrfs-endio-met] > \_ [kconservative/5] > \_ [ext4-dio-unwrit] > >> > >> [with a wondersome patch:] $ grep Name /proc/{29431,29432}/stat* > >> /proc/29431/status:Name: btrfs-endio-meta-1 > >> /proc/29432/status:Name: btrfs-endio-meta-write-1 > Name: kconservative/512 > Name: ext4-dio-unwritten > > > >doh. The fix for that is to have less clueless btrfs developers. > > And truncate their names to SUNWbtfs, ORCLintg and EXT4diou? > I think not :)Teach ps(1) to look in /proc/pid/status for kernel threads? -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Jan Engelhardt
2012-Feb-23 21:59 UTC
Re: [PATCH 1/2] treewide: fix memory corruptions when TASK_COMM_LEN != 16
On Thursday 2012-02-23 18:30, Andrew Morton wrote:> >Teach ps(1) to look in /proc/pid/status for kernel threads?To what end? The name in /proc/pid/status was also limited to TASK_COMM_LEN.