Re: [RFC 1/3] fbdev: hyperv_fb: Remove hyperv_fb driver

[Thomas Zimmermann <tzimmermann@suse.de>]

Hi

Am 12.09.25 um 05:24 schrieb Michael Kelley:
> From: Prasanna Kumar T S M <ptsm@linux.microsoft.com> Sent: Thursday, September 11, 2025 9:28 AM
>> Hi Michael,
>>
>> On 10-09-2025 20:55, Michael Kelley wrote:
>>> From: Thomas Zimmermann <tzimmermann@suse.de> Sent: Wednesday, September
>> 10, 2025 2:36 AM
>>>> Hi
>>>>
>>>> Am 09.09.25 um 18:58 schrieb Prasanna Kumar T S M:
>>>>> The Hyper-V DRM driver is available since kernel version 5.14 and
>>>>> provides full KMS support along with fbdev emulation via the DRM fbdev
>>>>> helpers. This makes the hyperv_fb driver redundant, remove it.
>>>> I'm all for removing obsolete drivers. But hyperv_drm likely first needs
>>>> to merge the patch at
>>>> https://lore.kernel.org/dri-devel/20250904145806.430568-5-tzimmermann@suse.de/
>>>> It's been tested and works well. If maintainers from Microsoft have a
>>>> look at the patch first, we could possibly land it fairly soon.
>>> Thomas --
>>>
>>> My testing of your v3 patch series for vblank timers ended up getting a
>>> WARN_ON after about 3 days of usage. See [1]. So I don't think it's 100%
>>> ready yet.
>>>
>>> But I agree we need your synthetic vblank timer support to address the
>>> Hyper-V DRM driver performance issue, before removing the Hyper-V
>>> fbdev driver. (See [2] for a description of the performance issue.)
>>>
>>> Second, isn't it customary to mark a driver as deprecated for a period
>>> of time, before removing it entirely? I don't see any documentation
>>> on the deprecation process, but I've seen it done in other cases. If you
>>> grep through all the kernel Kconfig files, you'll see entries tagged with
>>> DEPRECATED. Also the driver should be updated to output a deprecated
>>> message when it loads.
>> Is deprecating the driver a mandatory step?
>>
> I'm not aware of a mandatory requirement, at least not in the sense
> of it being documented in Documentation/process like other aspects
> of the Linux kernel development process. So I would say it's up to
> the Maintainers for Hyper-V and FBDEV as to whether the Hyper-V
> FB driver should go through a deprecation phase before being
> removed.
>
> Of course, the purpose of the deprecation phase is to be "nice"
> to users of the driver by giving them some warning that it is going
> away. That gives them an opportunity to raise objections, and/or
> to do any necessary migration to the replacement driver. I suspect
> there aren't many (or any?) users of Hyper-V FB that can't just move
> to the Hyper-V DRM driver, but who knows. We might be surprised.

Let's mark the driver as unmaintained now. There should be a kernel LTS 
release around December [1], after which hyperv-fb could be removed. 
Anyone with hard dependencies on hyperv-fb can remain in the LTS for a 
few more years.

Best regards
Thomas

[1] https://www.kernel.org/category/releases.html

>
> Michael

-- 
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nuernberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman
HRB 36809 (AG Nuernberg)