Regression in 6.1.35 / 6.3.9

Regression in 6.1.35 / 6.3.9

[Limonciello, Mario]

[Public]

Hi,
 A regression was reported in 6.4-rc6 that monitor resolutions are no longer present for anything but native resolution on eDP panels.  This specific change backported into stable at 6.1.35 and 6.3.9:
e749dd10e5f29 ("drm/amd/display: edp do not add non-edid timings")

After discussing it with the original author, they submitted a revert commit for review:
https://patchwork.freedesktop.org/patch/544273/

I suggested the revert also CC stable, and I expect this will go up in 6.5-rc1, but given the timing of the merge window and the original issue hit the stable trees, can we revert it sooner in the stable
trees to avoid exposing the regression to more people?

Thanks!

Re: Regression in 6.1.35 / 6.3.9

[Greg KH]

On Wed, Jun 28, 2023 at 05:56:01PM +0000, Limonciello, Mario wrote:
> [Public]
> 
> Hi,
>  A regression was reported in 6.4-rc6 that monitor resolutions are no longer present for anything but native resolution on eDP panels.  This specific change backported into stable at 6.1.35 and 6.3.9:
> e749dd10e5f29 ("drm/amd/display: edp do not add non-edid timings")
> 
> After discussing it with the original author, they submitted a revert commit for review:
> https://patchwork.freedesktop.org/patch/544273/
> 
> I suggested the revert also CC stable, and I expect this will go up in 6.5-rc1, but given the timing of the merge window and the original issue hit the stable trees, can we revert it sooner in the stable
> trees to avoid exposing the regression to more people?

As the submitted patch had the wrong git id, it might be good to be able
to take a real one?  I can take it once it shows up in linux-next if
it's really going to be going into 6.5, but I need a stable git id for
it.

thanks,

greg k-h

Re: Regression in 6.1.35 / 6.3.9

[Salvatore Bonaccorso]

Hi Mario,

On Wed, Jun 28, 2023 at 08:16:25PM +0200, Greg KH wrote:
> On Wed, Jun 28, 2023 at 05:56:01PM +0000, Limonciello, Mario wrote:
> > [Public]
> > 
> > Hi,
> >  A regression was reported in 6.4-rc6 that monitor resolutions are no longer present for anything but native resolution on eDP panels.  This specific change backported into stable at 6.1.35 and 6.3.9:
> > e749dd10e5f29 ("drm/amd/display: edp do not add non-edid timings")
> > 
> > After discussing it with the original author, they submitted a revert commit for review:
> > https://patchwork.freedesktop.org/patch/544273/
> > 
> > I suggested the revert also CC stable, and I expect this will go up in 6.5-rc1, but given the timing of the merge window and the original issue hit the stable trees, can we revert it sooner in the stable
> > trees to avoid exposing the regression to more people?
> 
> As the submitted patch had the wrong git id, it might be good to be able
> to take a real one?  I can take it once it shows up in linux-next if
> it's really going to be going into 6.5, but I need a stable git id for
> it.

Do you know, did that felt trough the cracks or is it still planned to
do the revert?

Regards,
Salvatore

Re: Regression in 6.1.35 / 6.3.9

[Mario Limonciello]

On 7/3/23 08:57, Salvatore Bonaccorso wrote:
> Hi Mario,
> 
> On Wed, Jun 28, 2023 at 08:16:25PM +0200, Greg KH wrote:
>> On Wed, Jun 28, 2023 at 05:56:01PM +0000, Limonciello, Mario wrote:
>>> [Public]
>>>
>>> Hi,
>>>   A regression was reported in 6.4-rc6 that monitor resolutions are no longer present for anything but native resolution on eDP panels.  This specific change backported into stable at 6.1.35 and 6.3.9:
>>> e749dd10e5f29 ("drm/amd/display: edp do not add non-edid timings")
>>>
>>> After discussing it with the original author, they submitted a revert commit for review:
>>> https://patchwork.freedesktop.org/patch/544273/
>>>
>>> I suggested the revert also CC stable, and I expect this will go up in 6.5-rc1, but given the timing of the merge window and the original issue hit the stable trees, can we revert it sooner in the stable
>>> trees to avoid exposing the regression to more people?
>>
>> As the submitted patch had the wrong git id, it might be good to be able
>> to take a real one?  I can take it once it shows up in linux-next if
>> it's really going to be going into 6.5, but I need a stable git id for
>> it.
> 
> Do you know, did that felt trough the cracks or is it still planned to
> do the revert?
> 
> Regards,
> Salvatore

Hi,

It's part of the PR that was sent for 6.5-rc1 [1]. Unfortunately it's 
not yet merged AFAICT to drm-next yet nor Linus' tree.

d6149086b45e [2] is the specific commit ID.

[1] https://patchwork.freedesktop.org/patch/545125/
[2] 
https://gitlab.freedesktop.org/agd5f/linux/-/commit/d6149086b45e150c170beaa4546495fd1880724c

Thanks,

Re: Regression in 6.1.35 / 6.3.9

[Salvatore Bonaccorso]

Hi Mario,

On Mon, Jul 03, 2023 at 12:43:06PM -0500, Mario Limonciello wrote:
> On 7/3/23 08:57, Salvatore Bonaccorso wrote:
> > Hi Mario,
> > 
> > On Wed, Jun 28, 2023 at 08:16:25PM +0200, Greg KH wrote:
> > > On Wed, Jun 28, 2023 at 05:56:01PM +0000, Limonciello, Mario wrote:
> > > > [Public]
> > > > 
> > > > Hi,
> > > >   A regression was reported in 6.4-rc6 that monitor resolutions are no longer present for anything but native resolution on eDP panels.  This specific change backported into stable at 6.1.35 and 6.3.9:
> > > > e749dd10e5f29 ("drm/amd/display: edp do not add non-edid timings")
> > > > 
> > > > After discussing it with the original author, they submitted a revert commit for review:
> > > > https://patchwork.freedesktop.org/patch/544273/
> > > > 
> > > > I suggested the revert also CC stable, and I expect this will go up in 6.5-rc1, but given the timing of the merge window and the original issue hit the stable trees, can we revert it sooner in the stable
> > > > trees to avoid exposing the regression to more people?
> > > 
> > > As the submitted patch had the wrong git id, it might be good to be able
> > > to take a real one?  I can take it once it shows up in linux-next if
> > > it's really going to be going into 6.5, but I need a stable git id for
> > > it.
> > 
> > Do you know, did that felt trough the cracks or is it still planned to
> > do the revert?
> > 
> > Regards,
> > Salvatore
> 
> Hi,
> 
> It's part of the PR that was sent for 6.5-rc1 [1]. Unfortunately it's not
> yet merged AFAICT to drm-next yet nor Linus' tree.
> 
> d6149086b45e [2] is the specific commit ID.
> 
> [1] https://patchwork.freedesktop.org/patch/545125/
> [2] https://gitlab.freedesktop.org/agd5f/linux/-/commit/d6149086b45e150c170beaa4546495fd1880724c

Ack, thanks! 

Regards,
Salvatore

Re: Regression in 6.1.35 / 6.3.9

[Limonciello, Mario]

On 7/3/2023 14:06, Salvatore Bonaccorso wrote:
> Hi Mario,
> 
> On Mon, Jul 03, 2023 at 12:43:06PM -0500, Mario Limonciello wrote:
>> On 7/3/23 08:57, Salvatore Bonaccorso wrote:
>>> Hi Mario,
>>>
>>> On Wed, Jun 28, 2023 at 08:16:25PM +0200, Greg KH wrote:
>>>> On Wed, Jun 28, 2023 at 05:56:01PM +0000, Limonciello, Mario wrote:
>>>>> [Public]
>>>>>
>>>>> Hi,
>>>>>    A regression was reported in 6.4-rc6 that monitor resolutions are no longer present for anything but native resolution on eDP panels.  This specific change backported into stable at 6.1.35 and 6.3.9:
>>>>> e749dd10e5f29 ("drm/amd/display: edp do not add non-edid timings")
>>>>>
>>>>> After discussing it with the original author, they submitted a revert commit for review:
>>>>> https://patchwork.freedesktop.org/patch/544273/
>>>>>
>>>>> I suggested the revert also CC stable, and I expect this will go up in 6.5-rc1, but given the timing of the merge window and the original issue hit the stable trees, can we revert it sooner in the stable
>>>>> trees to avoid exposing the regression to more people?
>>>>
>>>> As the submitted patch had the wrong git id, it might be good to be able
>>>> to take a real one?  I can take it once it shows up in linux-next if
>>>> it's really going to be going into 6.5, but I need a stable git id for
>>>> it.
>>>
>>> Do you know, did that felt trough the cracks or is it still planned to
>>> do the revert?
>>>
>>> Regards,
>>> Salvatore
>>
>> Hi,
>>
>> It's part of the PR that was sent for 6.5-rc1 [1]. Unfortunately it's not
>> yet merged AFAICT to drm-next yet nor Linus' tree.
>>
>> d6149086b45e [2] is the specific commit ID.
>>
>> [1] https://patchwork.freedesktop.org/patch/545125/
>> [2] https://gitlab.freedesktop.org/agd5f/linux/-/commit/d6149086b45e150c170beaa4546495fd1880724c
> 
> Ack, thanks!
> 
> Regards,
> Salvatore

The revert is in Linus' tree as of this morning.  Greg, can you take 
this back now?

d6149086b45e1 ("Revert "drm/amd/display: edp do not add non-edid timings"")

Re: Regression in 6.1.35 / 6.3.9

[Greg KH]

On Fri, Jul 07, 2023 at 11:45:57AM -0500, Limonciello, Mario wrote:
> On 7/3/2023 14:06, Salvatore Bonaccorso wrote:
> > Hi Mario,
> > 
> > On Mon, Jul 03, 2023 at 12:43:06PM -0500, Mario Limonciello wrote:
> > > On 7/3/23 08:57, Salvatore Bonaccorso wrote:
> > > > Hi Mario,
> > > > 
> > > > On Wed, Jun 28, 2023 at 08:16:25PM +0200, Greg KH wrote:
> > > > > On Wed, Jun 28, 2023 at 05:56:01PM +0000, Limonciello, Mario wrote:
> > > > > > [Public]
> > > > > > 
> > > > > > Hi,
> > > > > >    A regression was reported in 6.4-rc6 that monitor resolutions are no longer present for anything but native resolution on eDP panels.  This specific change backported into stable at 6.1.35 and 6.3.9:
> > > > > > e749dd10e5f29 ("drm/amd/display: edp do not add non-edid timings")
> > > > > > 
> > > > > > After discussing it with the original author, they submitted a revert commit for review:
> > > > > > https://patchwork.freedesktop.org/patch/544273/
> > > > > > 
> > > > > > I suggested the revert also CC stable, and I expect this will go up in 6.5-rc1, but given the timing of the merge window and the original issue hit the stable trees, can we revert it sooner in the stable
> > > > > > trees to avoid exposing the regression to more people?
> > > > > 
> > > > > As the submitted patch had the wrong git id, it might be good to be able
> > > > > to take a real one?  I can take it once it shows up in linux-next if
> > > > > it's really going to be going into 6.5, but I need a stable git id for
> > > > > it.
> > > > 
> > > > Do you know, did that felt trough the cracks or is it still planned to
> > > > do the revert?
> > > > 
> > > > Regards,
> > > > Salvatore
> > > 
> > > Hi,
> > > 
> > > It's part of the PR that was sent for 6.5-rc1 [1]. Unfortunately it's not
> > > yet merged AFAICT to drm-next yet nor Linus' tree.
> > > 
> > > d6149086b45e [2] is the specific commit ID.
> > > 
> > > [1] https://patchwork.freedesktop.org/patch/545125/
> > > [2] https://gitlab.freedesktop.org/agd5f/linux/-/commit/d6149086b45e150c170beaa4546495fd1880724c
> > 
> > Ack, thanks!
> > 
> > Regards,
> > Salvatore
> 
> The revert is in Linus' tree as of this morning.  Greg, can you take this
> back now?
> 
> d6149086b45e1 ("Revert "drm/amd/display: edp do not add non-edid timings"")

Now queued up, thanks.

greg k-h