From mboxrd@z Thu Jan 1 00:00:00 1970 From: Linus Torvalds Subject: Re: [PATCH 04/21] fs: Replace CURRENT_TIME with current_fs_time() for inode timestamps Date: Thu, 9 Jun 2016 12:08:04 -0700 Message-ID: References: <1465448705-25055-1-git-send-email-deepa.kernel@gmail.com> <1465448705-25055-5-git-send-email-deepa.kernel@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Return-path: In-Reply-To: <1465448705-25055-5-git-send-email-deepa.kernel@gmail.com> Sender: owner-linux-mm@kvack.org To: Deepa Dinamani Cc: linux-fsdevel , Linux Kernel Mailing List , Arnd Bergmann , Thomas Gleixner , Al Viro , y2038@lists.linaro.org, Steve French , "linux-cifs@vger.kernel.org" , samba-technical@lists.samba.org, Joern Engel , Prasad Joshi , logfs@logfs.org, Andrew Morton , Julia Lawall , David Howells , Firo Yang , Jaegeuk Kim , Changman Lee , Chao Yu , "Linux F2FS DEV, Mailing List" , Michal Hocko , Konstantin Khlebnikov , Naoya Horiguchi List-Id: linux-f2fs-devel.lists.sourceforge.net On Wed, Jun 8, 2016 at 10:04 PM, Deepa Dinamani wrote: > CURRENT_TIME macro is not appropriate for filesystems as it > doesn't use the right granularity for filesystem timestamps. > Use current_fs_time() instead. Again - using the inode instead fo the syuperblock in tghis patch would have made the patch much more obvious (it could have been 99% generated with the sed-script I sent out a week or two ago), and it would have made it unnecessary to add these kinds of things: > diff --git a/drivers/usb/core/devio.c b/drivers/usb/core/devio.c > index e9f5043..85c12f0 100644 > --- a/drivers/usb/core/devio.c > +++ b/drivers/usb/core/devio.c > @@ -2359,6 +2359,7 @@ static long usbdev_do_ioctl(struct file *file, unsigned int cmd, > { > struct usb_dev_state *ps = file->private_data; > struct inode *inode = file_inode(file); > + struct super_block *sb = inode->i_sb; > struct usb_device *dev = ps->dev; > int ret = -ENOTTY; where we add a new variable just because the calling convention was wrong. It's not even 100% obvious that a filesystem has to have one single time representation, so making the time function about the entity whose time is set is also conceptually a much better model, never mind that it is just what every single user seems to want anyway. So I'd *much* rather see + inode->i_atime = inode->i_mtime = inode->i_ctime = current_fs_time(inode); over seeing either of these two variants:: + inode->i_atime = inode->i_mtime = inode->i_ctime = current_fs_time(inode->i_sb); + ret->i_atime = ret->i_mtime = ret->i_ctime = current_fs_time(sb); because the first of those variants (grep for current_fs_time() in the current git tree, and notice that it's the common one) we have the pointless "let's chase a pointer in every caller" And while it's true that the second variant is natural for *some* situations, I've yet to find one where it wasn't equally sane to just pass in the inode instead. Linus -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org