Re: [PATCH v2] drivers/crypto: Mark QCE as BROKEN

[Demi Marie Obenour]

On 7/13/26 15:49, Demi Marie Obenour wrote:
> On 7/13/26 11:31, Greg KH wrote:
>> On Mon, Jul 13, 2026 at 10:42:04AM -0400, Demi Marie Obenour wrote:
>>> On 7/13/26 09:07, Eric Biggers wrote:
>>>> On Mon, Jul 13, 2026 at 06:47:07AM +0200, Greg KH wrote:
>>>>> On Sun, Jul 12, 2026 at 05:31:31PM -0400, Demi Marie Obenour via B4 Relay wrote:
>>>>>> From: Demi Marie Obenour <demiobenour@xxxxxxxxx>
>>>>>>
>>>>>> This driver is harmful:
>>>>>>
>>>>>> - It is much slower than the CPU [1] [2].
>>>>>> - It Has a history of bugs [2] [3].
>>>>>> - It does not have exclusive access to the hardware [4], causing races
>>>>>>   with the secure world.
>>>>>> - It register its implementations with too low a cra_priority for them
>>>>>>   to be actually used [5].
>>>>>>
>>>>>> Therefore, disable it to ensure that nobody builds it into kernels they
>>>>>> intend to ship.
>>>>>>
>>>>>> In the future, the driver will be used for processing restricted media
>>>>>> content.  However, the kernel does not currently support this.  Since
>>>>>> the driver will have future uses, allow building it if COMPILE_TEST is
>>>>>> enabled.
>>>>>
>>>>> Why not just delete it now, and then bring it back when it is needed in
>>>>> the future?  Otherwise this will just trip up the static code checkers
>>>>> who will attempt to "fix" things in it.
>>>>
>>>> That makes sense to me, but Qualcomm pushed back on deletion:
>>>> https://lore.kernel.org/linux-crypto/20260602-qcom-qce-broken-v1-1-a4ef756089e0@xxxxxxxxxxxxxxxx/
>>>>
>>>> But I've still not seen any evidence that this driver is useful for
>>>> anything or has any users.  Even Qualcomm seems to be unwilling to make
>>>> such claims; they only claim that the IP is used (i.e., not in Linux)
>>>> and that new features are planned.
>>>>
>>>> - Eric
>>>
>>> Here is my reading of Qualcomm's statements:
>>>
>>> QCE is currently used by the Arm secure world.  In the future, QCE
>>> will be used by the kernel as part of restricted content playback.
>>> Qualcomm wants to add the needed features to the existing driver.
>>> This will not use the crypto API.
>>
>> Why is a crypto driver not going to use the crypto API?
> 
> I suspect that the restricted content playback use-case involves
> QCE decrypting data stored in kernel memory, using keys inaccessible
> to the kernel, to memory inaccessible to the kernel.  The crypto API
> can't express this.
> 
>>> I would be fine with the driver being removed, but not if it means
>>> another out-of-tree Android driver.
>>
>> It's not another out-of-tree Android driver if nothing in Android
>> actually uses it :)
> 
> Nothing uses it *yet*.

That said, deleting the driver is also an option.  I don't have a
problem with that myself, but Qualcomm does.  I figured that
marking the driver BROKEN would be less objectionable.
-- 
Sincerely,
Demi Marie Obenour (she/her/hers)

Attachment: OpenPGP_signature.asc
Description: OpenPGP digital signature