DMARC-Filter: OpenDMARC Filter v1.4.2 delorie.com 5ALBLUmg3704549
Authentication-Results: delorie.com; dmarc=pass (p=none dis=none) header.from=cygwin.com
Authentication-Results: delorie.com; spf=pass smtp.mailfrom=cygwin.com
DKIM-Filter: OpenDKIM Filter v2.11.0 delorie.com 5ALBLUmg3704549
Authentication-Results: delorie.com;
	dkim=pass (1024-bit key, unprotected) header.d=cygwin.com header.i=@cygwin.com header.a=rsa-sha256 header.s=default header.b=OA//rN+W
X-Recipient: archive-cygwin AT delorie DOT com
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 38319385B52B
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=cygwin.com;
	s=default; t=1763724089;
	bh=ebZhU3H+EFZ/DmW/hm+SMSHYBBC522Bo+vkkJo0svHY=;
	h=Date:To:Subject:References:In-Reply-To:List-Id:List-Unsubscribe:
	 List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To:Cc:
	 From;
	b=OA//rN+WAyDcMY7VON5W2yQSVUApZCgpgsZp0EkPUIfUrhMSl7Kjmotu+qreQFIG5
	 6BkET6g/ATua6EpnbFA+j20W0SJ/75j/dbDXZ14V6IV9VYxYyBa9chObOMu1qG7Wjp
	 BBtHTpiUenW0vAnRgkxxX/5+rKMnxRBPjnPN5nAY=
X-Original-To: cygwin AT cygwin DOT com
Delivered-To: cygwin AT cygwin DOT com
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org EB9B43857037
Date: Fri, 21 Nov 2025 12:21:04 +0100
To: cygwin AT cygwin DOT com
Subject: Re: flock/open random error
Message-ID: <aSBLIOHmUlbblgaG@calimero.vinschen.de>
Mail-Followup-To: cygwin AT cygwin DOT com
References: <CA+1R0VjcBajGpLMJ_0Waie0g_5S15_kPfzpT2=GUyN+39RWrMw AT mail DOT gmail DOT com>
 <20251112182412 DOT ba3a65f36838b9b5fd7d3f9b AT nifty DOT ne DOT jp>
 <CA+1R0VjW5rbKAVBb_vAFqKcKmE0yfvOFi6i0-GB=2-mjOhCY7A AT mail DOT gmail DOT com>
 <20251121190009 DOT f08a3229007bbbf101ad1463 AT nifty DOT ne DOT jp>
MIME-Version: 1.0
Content-Disposition: inline
In-Reply-To: <20251121190009.f08a3229007bbbf101ad1463@nifty.ne.jp>
X-BeenThere: cygwin AT cygwin DOT com
X-Mailman-Version: 2.1.30
Precedence: list
List-Id: General Cygwin discussions and problem reports <cygwin.cygwin.com>
List-Unsubscribe: <https://cygwin.com/mailman/options/cygwin>,
 <mailto:cygwin-request AT cygwin DOT com?subject=unsubscribe>
List-Archive: <https://cygwin.com/pipermail/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-request AT cygwin DOT com?subject=help>
List-Subscribe: <https://cygwin.com/mailman/listinfo/cygwin>,
 <mailto:cygwin-request AT cygwin DOT com?subject=subscribe>
From: Corinna Vinschen via Cygwin <cygwin AT cygwin DOT com>
Reply-To: cygwin AT cygwin DOT com
Cc: Corinna Vinschen <corinna-cygwin AT cygwin DOT com>
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Errors-To: cygwin-bounces~archive-cygwin=delorie DOT com AT cygwin DOT com
Sender: "Cygwin" <cygwin-bounces~archive-cygwin=delorie DOT com AT cygwin DOT com>

On Nov 21 19:00, Takashi Yano via Cygwin wrote:
> On Fri, 14 Nov 2025 11:27:04 -0800
> Nahor wrote:
> > If `flock()` was used on the same file descriptor, then this might
> > have been a valid point. However, each thread has its own file
> > descriptor in this case, so this would be very surprising if it wasn't
> > thread-safe.
> 
> IIUC, flock() locks file itself, but not file descriptor. Usually,
> flock() is used for inter-process file protection, isn't it?
> 
> > Moreover, it's not just `flock()` failing, it's also (and mostly!)
> > `open()` that fails. And it's the `open()` for a completely different
> > file than the one being locked. So that would suggest that `open()` is
> > not also not MT-safe. And not safe when using different files. And not
> > safe across multiple different functions (flock+open).
> 
> Indeed, this is really weird. I looked into this, and found 'upath' in
> path.cc is destroyed after 'NtCreateFile()' call at the following line.
> 
> I added assertion as follows:
> 
> diff --git a/winsup/cygwin/path.cc b/winsup/cygwin/path.cc
> index 710775e38..562100161 100644
> --- a/winsup/cygwin/path.cc
> +++ b/winsup/cygwin/path.cc
> @@ -3189,6 +3189,8 @@ restart:
>  	 symlink (which would spoil the task of this method quite a bit).
>  	 Fortunately it's ignored on most other file systems so we don't have
>  	 to special case NFS too much. */
> +      wchar_t c;
> +      c = upath.Buffer[0];
>        status = NtCreateFile (&h,
>  			     READ_CONTROL | FILE_READ_ATTRIBUTES | FILE_READ_EA,
>  			     &attr, &io, NULL, 0, FILE_SHARE_VALID_FLAGS,
> @@ -3196,6 +3198,7 @@ restart:
>  			     FILE_OPEN_REPARSE_POINT
>  			     | FILE_OPEN_FOR_BACKUP_INTENT,
>  			     eabuf, easize);
> +      assert (upath.Buffer[0] == c);
>        debug_printf ("%y = NtCreateFile (%S)", status, &upath);
>        /* No right to access EAs or EAs not supported? */
>        if (!NT_SUCCESS (status)
> 
> then, the assertion fails for your test case like:
> tmp_dir: /tmp/flockAQ4Hbb
> assertion "upath.Buffer[0] == c" failed: file "../../.././winsup/cygwin/path.cc", line 3201, function: int symlink_info::check(char*, const suffix_info*, fs_info&, path_conv_handle&)
>                       Abort
> 
> Does another thread destroy the puthbuf? But pathbuf is thread local, IIUC.
> Corinna, have you noticed anything?

No, I haven't. The tmp_pathbuf buffers are malloced and reused, but they
are only ever used in the same thread. So afaics, either the buffer gets
incorrect stored in a global datastructure and overwritten, or there's
a buffer overflow in the allocation preceeding the upath.Buffer.  That
could be an application allocation just as well as a DLL allocation.


Corinna

-- 
Problem reports:      https://cygwin.com/problems.html
FAQ:                  https://cygwin.com/faq/
Documentation:        https://cygwin.com/docs.html
Unsubscribe info:     https://cygwin.com/ml/#unsubscribe-simple