linux-kernel - Re: [PATCH 4/4] Return -EINVAL if userspace passes bogus flags to open()

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: <ce981281-bb3a-a539-8d5a-b2d88107d4c0@gmail.com>
Date:   Tue, 12 Feb 2019 18:00:44 -0500
From:   "Demi M. Obenour" <demiobenour@...il.com>
To:     Al Viro <viro@...iv.linux.org.uk>
Cc:     "open list:FILESYSTEMS (VFS and infrastructure)" 
        <linux-fsdevel@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 4/4] Return -EINVAL if userspace passes bogus flags to
 open()



On 2/12/19 3:38 PM, Al Viro wrote:
> On Tue, Feb 12, 2019 at 09:54:47AM -0500, demiobenour@...il.com wrote:
>> From: "Demi M. Obenour" <demiobenour@...il.com>
>>
>> While testing the O_PATHSTATIC patch, I discovered that Linux does not
>> return any error if an invalid flag is passed to open(2).  This prevents
>> adding new flags without a (minor) risk of breaking userspace.
>> Therefore, add a check for invalid flags, and return -EINVAL if any are
>> found.
> 
> ... which would qualtify as userland ABI breakage all by itself.
> 
I suspect that very few (if any) programs pass invalid flags to open().
Additionally, O_DIRECT was added in Linux 2.4, even though it had
previously been ignored.  If someone knows of a real program that does,
I can instead create an open2() syscall, but I would prefer to avoid that.



Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)