Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
MIME-Version: 1.0
References: <20220415094518.380543-1-cccheng@synology.com> <20220415094518.380543-3-cccheng@synology.com>
 <87h76utrj2.fsf@mail.parknet.co.jp>
In-Reply-To: <87h76utrj2.fsf@mail.parknet.co.jp>
From:   Chung-Chiang Cheng <shepjeng@gmail.com>
Date:   Fri, 22 Apr 2022 16:20:06 +0800
Message-ID: <CAHuHWt=xYOBhRp_KJrSF5-xDBW+x_O4gnQSsX2M5u2D8zSXwTg@mail.gmail.com>
Subject: Re: [PATCH v3 3/3] fat: report creation time in statx
To:     OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Cc:     Chung-Chiang Cheng <cccheng@synology.com>,
        linux-kernel@vger.kernel.org, kernel@cccheng.net
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

> > +     } else {
> >               fat_truncate_atime(sbi, &inode->i_mtime, &inode->i_atime);
> > +             fat_truncate_crtime(sbi, &inode->i_mtime, &MSDOS_I(inode)->i_crtime);
> > +     }
>
> Probably, nothing worth to update i_crtime here, right?
>

That's right. I also miss truncating the create-time for newly created
files in vfat. I will add it back in the next patch. It will be the only
location that calls fat_truncate_crtime().

@@ -781,11 +782,12 @@ static int vfat_create(struct user_namespace
*mnt_userns, struct inode *dir,
        }
        inode_inc_iversion(inode);
        fat_truncate_time(inode, &ts, S_ATIME|S_CTIME|S_MTIME);
+       fat_truncate_crtime(...)


I also have a question. If msdos doesn't support create-time, why does the
original fat_truncate_time() truncate it to a 2-sec granularity?

if (flags & S_CTIME) {
        if (sbi->options.isvfat)
                inode->i_ctime = fat_timespec64_trunc_10ms(*now);
        else
                inode->i_ctime = fat_timespec64_trunc_2secs(*now);
}

This logic was introduced in 6bb885ecd746 (fat: add functions to update
and truncate timestamps appropriately"), and seems not for change-time.
Should I remove it?

Thanks.