[PATCH] fs: relatime update - Match comment with behavior

[PATCH] fs: relatime update - Match comment with behavior

[Aurelien Thierry]

2 comments right before code wrongly state that if (c|m)time is younger
than atime, then atime is updated (behavior is the other way around).

Fix aligns comments with actual behavior, function description and
documentation (man mount).

Signed-off-by: Aurelien Thierry <aurelien.thierry@quoscient.io>
---
 fs/inode.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/inode.c b/fs/inode.c
index df6542ec3b88..1e5489288aa9 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -1620,12 +1620,12 @@ static int relatime_need_update(struct vfsmount *mnt, struct inode *inode,
 	if (!(mnt->mnt_flags & MNT_RELATIME))
 		return 1;
 	/*
-	 * Is mtime younger than atime? If yes, update atime:
+	 * Is atime earlier than mtime? If yes, update atime:
 	 */
 	if (timespec64_compare(&inode->i_mtime, &inode->i_atime) >= 0)
 		return 1;
 	/*
-	 * Is ctime younger than atime? If yes, update atime:
+	 * Is atime earlier than ctime? If yes, update atime:
 	 */
 	if (timespec64_compare(&inode->i_ctime, &inode->i_atime) >= 0)
 		return 1;
-- 
2.17.1

Re: [PATCH] fs: relatime update - Match comment with behavior

[Al Viro]

On Tue, Jun 11, 2019 at 07:45:40AM +0200, Aurelien Thierry wrote:
> 2 comments right before code wrongly state that if (c|m)time is younger
> than atime, then atime is updated (behavior is the other way around).
> 
> Fix aligns comments with actual behavior, function description and
> documentation (man mount).

Huh?  "mtime is younger than atime" means that mtime refers to the moment
later than that refered to by atime, i.e that atime refers to the moment
earlier than that refered to by mtime.  What is the problem you are trying
to fix?

Both the original and changed comments mean exact same thing.  And yes,
the changed comment does match the actual behaviour.  Just as the original
one does...

Re: [PATCH] fs: relatime update - Match comment with behavior

[Aurelien Thierry]

You are right.

As atime and mtime are timestamps and not "live" objects, I understood 
their age as being the actual timestamp and thus I understood "mtime is
younger than atime" the other way around.

So this can be a fix for better clarity ("earlier" is used in other
places rather than "younger") but the current comments are not wrong.


On 6/18/19 3:12 AM, Al Viro wrote:
> On Tue, Jun 11, 2019 at 07:45:40AM +0200, Aurelien Thierry wrote:
>> 2 comments right before code wrongly state that if (c|m)time is younger
>> than atime, then atime is updated (behavior is the other way around).
>>
>> Fix aligns comments with actual behavior, function description and
>> documentation (man mount).
> 
> Huh?  "mtime is younger than atime" means that mtime refers to the moment
> later than that refered to by atime, i.e that atime refers to the moment
> earlier than that refered to by mtime.  What is the problem you are trying
> to fix?
> 
> Both the original and changed comments mean exact same thing.  And yes,
> the changed comment does match the actual behaviour.  Just as the original
> one does...
>