linux-kernel - Re: 回复: [PATCH 2/2] hfs: fix to update ctime after rename

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <d83f9ef1820c52004ce82ce69e58333ba83af85d.camel@dubeyko.com>
Date: Fri, 09 May 2025 12:02:29 -0700
From: Viacheslav Dubeyko <slava@...eyko.com>
To: 李扬韬 <frank.li@...o.com>,  Viacheslav
 Dubeyko <Slava.Dubeyko@....com>, "glaubitz@...sik.fu-berlin.de"
 <glaubitz@...sik.fu-berlin.de>
Cc: "linux-fsdevel@...r.kernel.org" <linux-fsdevel@...r.kernel.org>, 
 "linux-kernel@...r.kernel.org"
	 <linux-kernel@...r.kernel.org>
Subject: Re: 回复:  [PATCH 2/2] hfs: fix to
 update ctime after rename

On Fri, 2025-05-09 at 10:51 -0700, Viacheslav Dubeyko wrote:
> On Wed, 2025-05-07 at 14:22 +0000, 李扬韬 wrote:
> > Hi Slava,
> > 
> > >    +ERROR: access time has changed for file1 after remount
> > >    +ERROR: access time has changed after modifying file1
> > >    +ERROR: access time has changed for file in read-only
> > > filesystem
> > 
> > > It looks like that it is not the whole fix of the issue for HFS
> > > case.
> > 
> > The test cases that failed after applying this patch are all
> > related
> > to the atime not being updated,
> 
> If I understood correctly "ERROR: access time has changed for file1
> after remount" means atime has been changed.
> 
> > but hfs actually does not have atime. 
> > 
> 
> But how the test detects that atime has been updated? If HFS hasn't
> atime, then test cannot detect such update, from my point of view.
> 

As far as I can see, HFS code operates by atime [1 - 3]. As a result,
generic_fillattr [4] can retrieve the atime value. If we don't store
atime in the on-disk metadata, then the test could see inconsistent
atime value. I believe that dirMdDat/filMdDat (mtime) needs to be
considered as atime too. Because, modification cannot be done without
access action. So, atime == mtime.

Thanks,
Slava.

> > So the current fix is sufficient, should we modify the 003 test
> > case?
> > 
> 
> I don't think so. Probably, something is wrong in HFS code. We need
> to
> double check it.
> 
> Thanks,
> Slava.
> 
> >    dirCrDat:      LongInt;    {date and time of creation}
> >    dirMdDat:      LongInt;    {date and time of last modification}
> >    dirBkDat:      LongInt;    {date and time of last backup}
> > 
> >    filCrDat:      LongInt;    {date and time of creation}
> >    filMdDat:      LongInt;    {date and time of last modification}
> >    filBkDat:      LongInt;    {date and time of last backup}
> > 
> > Thanks,
> > Yangtao

[1] https://elixir.bootlin.com/linux/v6.12.6/source/fs/hfs/sysdep.c#L35
[2] https://elixir.bootlin.com/linux/v6.12.6/source/fs/hfs/inode.c#L356
[3] https://elixir.bootlin.com/linux/v6.12.6/source/fs/hfs/inode.c#L367
[4] https://elixir.bootlin.com/linux/v6.12.6/source/fs/stat.c#L60