Re: [PATCH v4] fs: clear file privilege bits when mmap writing
From: Kees Cook
Date: Wed Dec 09 2015 - 12:53:42 EST
On Wed, Dec 9, 2015 at 4:49 AM, Jan Kara <jack@xxxxxxx> wrote:
> On Tue 08-12-15 15:28:18, Kees Cook wrote:
>> Normally, when a user can modify a file that has setuid or setgid bits,
>> those bits are cleared when they are not the file owner or a member
>> of the group. This is enforced when using write and truncate but not
>> when writing to a shared mmap on the file. This could allow the file
>> writer to gain privileges by changing a binary without losing the
>> setuid/setgid/caps bits.
>>
>> Changing the bits requires holding inode->i_mutex, so it cannot be done
>> during the page fault (due to mmap_sem being held during the fault). We
>> could do this during vm_mmap_pgoff, but that would need coverage in
>> mprotect as well, but to check for MAP_SHARED, we'd need to hold mmap_sem
>> again.
>>
>> Instead, detect the need to clear the bits during the page fault, and
>> actually remove the bits during final fput. Since the file was open for
>> writing, it wouldn't have been possible to execute it yet.
>>
>> Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>
>> ---
>> Here's another way? I wonder which of these will actually work. I
>> wish we could reject writes if file_remove_privs() fails.
>
> Yeah, the fact that we cannot do anything with file_remove_privs() failure
> is rather unfortunate. So open for writing may be the best choice for
> file_remove_privs() in the end? It's not perfect but it looks like the
> least problematic solution.
Yeah, back to just the open itself. I can't even delay this to the mmap. :(
I will do a v5. :)
-Kees
>
> Frankly writeable files that have SUID / SGID bits set are IMHO problems on
> its own, with IMA attrs which are handled by file_remove_privs() as well
> things may be somewhat different.
>
>> diff --git a/fs/file_table.c b/fs/file_table.c
>> index ad17e05ebf95..abb537ef4344 100644
>> --- a/fs/file_table.c
>> +++ b/fs/file_table.c
>> @@ -191,6 +191,14 @@ static void __fput(struct file *file)
>>
>> might_sleep();
>>
>> + /*
>> + * XXX: While avoiding mmap_sem, we've already been written to.
>> + * We must ignore the return value, since we can't reject the
>> + * write.
>> + */
>> + if (unlikely(file->f_remove_privs))
>> + file_remove_privs(file);
>> +
>
> You're missing i_mutex locking again ;).
>
> Honza
> --
> Jan Kara <jack@xxxxxxxx>
> SUSE Labs, CR
--
Kees Cook
Chrome OS & Brillo Security
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/