lkml.org 
[lkml]   [2015]   [Dec]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH v4] fs: clear file privilege bits when mmap writing
From
On Wed, Dec 9, 2015 at 4:49 AM, Jan Kara <jack@suse.cz> 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@chromium.org>
>> ---
>> 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@suse.com>
> SUSE Labs, CR



--
Kees Cook
Chrome OS & Brillo Security


\
 
 \ /
  Last update: 2015-12-09 19:21    [W:0.058 / U:0.148 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site