On 1/25/24 08:38, Mickaël Salaün wrote:
> On Wed, Jan 24, 2024 at 01:32:02PM -0800, Kees Cook wrote:
>> On Wed, Jan 24, 2024 at 12:47:34PM -0800, Linus Torvalds wrote:
>>> On Wed, 24 Jan 2024 at 12:15, Kees Cook <keescook@chromium.org> wrote:
>>>>
>>>> Hmpf, and frustratingly Ubuntu (and Debian) still builds with
>>>> CONFIG_USELIB, even though it was reported[2] to them almost 4 years ago.
>>
>> For completeness, Fedora hasn't had CONFIG_USELIB for a while now.
>>
>>> Well, we could just remove the __FMODE_EXEC from uselib.
>>>
>>> It's kind of wrong anyway.
>>
>> Yeah.
>>
>>> So I think just removing __FMODE_EXEC would just do the
>>> RightThing(tm), and changes nothing for any sane situation.
>>
>> Agreed about these:
>>
>> - fs/fcntl.c is just doing a bitfield sanity check.
>>
>> - nfs_open_permission_mask(), as you say, is only checking for
>>    unreadable case.
>>
>> - fsnotify would also see uselib() as a read, but afaict,
>>    that's what it would see for an mmap(), so this should
>>    be functionally safe.
>>
>> This one, though, I need some more time to examine:
>>
>> - AppArmor, TOMOYO, and LandLock will see uselib() as an
>>    open-for-read, so that might still be a problem? As you
>>    say, it's more of a mmap() call, but that would mean
>>    adding something a call like security_mmap_file() into
>>    uselib()...
> 
> If user space can emulate uselib() without opening a file with
> __FMODE_EXEC, then there is no security reason to keep __FMODE_EXEC for
> uselib().
> 
agreed

> Removing __FMODE_EXEC from uselib() looks OK for Landlock.  We use
> __FMODE_EXEC to infer if a file is being open for execution i.e., by
> execve(2).
> 

apparmor the hint should be to avoid doing permission work again that we
are doing in exec. That it regressed anything more than performance here
is a bug, that will get fixed.


> If __FMODE_EXEC is removed from uselib(), I think it should also be
> backported to all stable kernels for consistency though.
> 
hrmmm, I am not opposed to it being backported but I don't know that
it should be backported. Consistency is good but its not a serious
bug fix either

> 
>>
>> The issue isn't an insane "support uselib() under AppArmor" case, but
>> rather "Can uselib() be used to bypass exec/mmap checks?"
>>
>> This totally untested patch might give appropriate coverage:
>>
>> diff --git a/fs/exec.c b/fs/exec.c
>> index d179abb78a1c..0c9265312c8d 100644
>> --- a/fs/exec.c
>> +++ b/fs/exec.c
>> @@ -143,6 +143,10 @@ SYSCALL_DEFINE1(uselib, const char __user *, library)
>>   	if (IS_ERR(file))
>>   		goto out;
>> 
>> +	error = security_mmap_file(file, PROT_READ | PROT_EXEC, MAP_FIXED | MAP_SHARED);
>> +	if (error)
>> +		goto exit;
>> +
>>   	/*
>>   	 * may_open() has already checked for this, so it should be
>>   	 * impossible to trip now. But we need to be extra cautious
>>
>>> Of course, as you say, not having CONFIG_USELIB enabled at all is the
>>> _truly_ sane thing, but the only thing that used the FMODE_EXEC bit
>>> were landlock and some special-case nfs stuff.
>>
>> Do we want to attempt deprecation again? This was suggested last time:
>> https://lore.kernel.org/lkml/20200518130251.zih2s32q2rxhxg6f@wittgenstein/
>>
>> -Kees
>>
>> -- 
>> Kees Cook
>>