Re: [PATCH v2] fs/exec: require argv[0] presence in do_execveat_common()

[Ariadne Conill <ariadne@dereferenced.org>]

Hi,

On Wed, 26 Jan 2022, Matthew Wilcox wrote:

> On Wed, Jan 26, 2022 at 10:57:29AM -0600, Eric W. Biederman wrote:
>> Matthew Wilcox <willy@infradead.org> writes:
>>
>>> On Wed, Jan 26, 2022 at 11:44:47AM +0000, Ariadne Conill wrote:
>>>> Interestingly, Michael Kerrisk opened an issue about this in 2008[1],
>>>> but there was no consensus to support fixing this issue then.
>>>> Hopefully now that CVE-2021-4034 shows practical exploitative use
>>>> of this bug in a shellcode, we can reconsider.
>>>>
>>>> [0]: https://pubs.opengroup.org/onlinepubs/9699919799/functions/exec.html
>>>> [1]: https://bugzilla.kernel.org/show_bug.cgi?id=8408
>>>
>>> Having now read 8408 ... if ABI change is a concern (and I really doubt
>>> it is), we could treat calling execve() with a NULL argv as if the
>>> caller had passed an array of length 1 with the first element set to
>>> NULL.  Just like we reopen fds 0,1,2 for suid execs if they were
>>> closed.
>>
>> Where do we reopen fds 0,1,2 for suid execs?  I feel silly but I looked
>> through the code fs/exec.c quickly and I could not see it.
>
> I'm wondering if I misremembered and it's being done in ld.so
> rather than in the kernel?  That might be the right place to put
> this fix too.
>
>> I am attracted to the notion of converting an empty argv array passed
>> to the kernel into something we can safely pass to userspace.
>>
>> I think it would need to be having the first entry point to "" instead
>> of the first entry being NULL.  That would maintain the invariant that you
>> can always dereference a pointer in the argv array.
>
> Yes, I like that better than NULL.

If we are doing {"", NULL}, then I think it makes sense that we could just 
say argc == 1 at that point, which probably sidesteps the concern Jann 
raised with the {NULL, NULL} patch, no?

Ariadne