From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jaegeuk Kim Subject: Re: [PATCH] f2fs: invalidate inode and data pages if inode is no longer used Date: Thu, 25 Aug 2016 10:00:37 -0700 Message-ID: <20160825170037.GB84318@jaegeuk> References: <20160825023419.82566-1-jaegeuk@kernel.org> <79f3d1c0-63b6-c0ba-3e50-e75598083cb2@huawei.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from sog-mx-3.v43.ch3.sourceforge.com ([172.29.43.193] helo=mx.sourceforge.net) by sfs-ml-4.v29.ch3.sourceforge.com with esmtp (Exim 4.76) (envelope-from ) id 1bcy1H-0006tv-MC for linux-f2fs-devel@lists.sourceforge.net; Thu, 25 Aug 2016 17:00:47 +0000 Received: from mail.kernel.org ([198.145.29.136]) by sog-mx-3.v43.ch3.sourceforge.com with esmtps (TLSv1:AES256-SHA:256) (Exim 4.76) id 1bcy1F-0001Vi-U8 for linux-f2fs-devel@lists.sourceforge.net; Thu, 25 Aug 2016 17:00:47 +0000 Content-Disposition: inline In-Reply-To: <79f3d1c0-63b6-c0ba-3e50-e75598083cb2@huawei.com> List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: linux-f2fs-devel-bounces@lists.sourceforge.net To: Chao Yu Cc: linux-f2fs-devel@lists.sourceforge.net On Thu, Aug 25, 2016 at 07:59:45PM +0800, Chao Yu wrote: > Hi Jaegeuk, > > On 2016/8/25 10:34, Jaegeuk Kim wrote: > > When a file is closed, let's deactivate inode page to mitigate further memory > > pressure. We can do data pages as well in the corner case of f2fs_drop_inode. > > > > Signed-off-by: Jaegeuk Kim > > --- > > fs/f2fs/file.c | 4 ++++ > > fs/f2fs/super.c | 1 + > > 2 files changed, 5 insertions(+) > > > > diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c > > index e460211..5f9a6dc 100644 > > --- a/fs/f2fs/file.c > > +++ b/fs/f2fs/file.c > > @@ -1465,6 +1465,10 @@ static int f2fs_release_file(struct inode *inode, struct file *filp) > > filemap_fdatawrite(inode->i_mapping); > > clear_inode_flag(inode, FI_DROP_CACHE); > > } > > + > > + /* deactivate written inode page */ > > + invalidate_mapping_pages(NODE_MAPPING(F2FS_I_SB(inode)), > > + inode->i_ino, inode->i_ino); > > Seems if the file was opened by two processes, one process closes this file, > another one will loss page cache of inode. I'm not sure whether this predication > is friendly to user. No, this only calls when it's the last writer. BTW, I need to add reader case as well. :) > > > return 0; > > } > > > > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c > > index 7f863a6..f84696d 100644 > > --- a/fs/f2fs/super.c > > +++ b/fs/f2fs/super.c > > @@ -618,6 +618,7 @@ static int f2fs_drop_inode(struct inode *inode) > > sb_end_intwrite(inode->i_sb); > > > > fscrypt_put_encryption_info(inode, NULL); > > + invalidate_mapping_pages(inode->i_mapping, 0, -1); > > Hmm, shouldn't we truncate page cache before f2fs_truncate() like we do in > f2fs_evict? Then, it needs to be done under i_lock, and IMO, we don't need to wait for writeback here. Thanks, > > Thanks, > > > spin_lock(&inode->i_lock); > > atomic_dec(&inode->i_count); > > } > > > > > ------------------------------------------------------------------------------ > _______________________________________________ > Linux-f2fs-devel mailing list > Linux-f2fs-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel ------------------------------------------------------------------------------