Messages in this thread | | | Date | Wed, 28 Sep 2011 07:03:52 -0700 (PDT) | From | Dan Magenheimer <> | Subject | RE: [PATCH V10 5/6] mm: cleancache: update to match akpm frontswap feedback |
| |
> From: KAMEZAWA Hiroyuki [mailto:kamezawa.hiroyu@jp.fujitsu.com] > Subject: Re: [PATCH V10 5/6] mm: cleancache: update to match akpm frontswap feedback > > On Thu, 15 Sep 2011 14:34:46 -0700 > Dan Magenheimer <dan.magenheimer@oracle.com> wrote: > > > From: Dan Magenheimer <dan.magenheimer@oracle.com> > > Subject: [PATCH V10 5/6] mm: cleancache: update to match akpm frontswap feedback > err = sysfs_create_group(mm_kobj, &cleancache_attr_group); > > -#endif /* CONFIG_SYSFS */ > > +#ifdef CONFIG_DEBUG_FS > > + struct dentry *root = debugfs_create_dir("cleancache", NULL); > > + if (root == NULL) > > + return -ENXIO; > > + debugfs_create_u64("succ_gets", S_IRUGO, root, &cleancache_succ_gets); > > + debugfs_create_u64("failed_gets", S_IRUGO, > > + root, &cleancache_failed_gets); > > + debugfs_create_u64("puts", S_IRUGO, root, &cleancache_puts); > > + debugfs_create_u64("invalidates", S_IRUGO, > > + root, &cleancache_invalidates); > > +#endif
Hi Kame --
Thanks for the review!
> No exisiting userlands are affected by this change of flush->invalidates ?
Not that I'm aware of. As required by Andrew Morton, the frontswap patchset now exposes ALL statistics through debugfs instead of sysfs. For consistency, all cleancache statistics are also moved from sysfs into debugfs, so this is a good time to also do the name change. (The name change was also required by Andrew Morton, and previously suggested by Minchan Kim.)
Thanks, Dan
| |