4.19-y-cip missing backport?

4.19-y-cip missing backport?

[Yann Sionneau]

Hello CIP people,

I noticed that 4.19.325 (as well as the latest -cip126 release) has this 
backport:

commit 2b8f2afa311c722a90f00fb2960e6deb4f5100a5
Author: Marek Szyprowski <m.szyprowski@samsung.com>
Date:   Thu Aug 29 07:30:28 2019 +0200

     usb: dwc3: remove generic PHY calibrate() calls

     [ Upstream commit a0a465569b45e3690be155c96fb54603d6904f41 ]


But it does not have this:

commit 34c7ed72f4f04e73901ca6ec2e1db6c040495935
Author: Marek Szyprowski <m.szyprowski@samsung.com>
Date:   Thu Aug 29 07:30:27 2019 +0200

     usb: core: phy: add support for PHY calibration


So it seems nobody calls phy_calibrate() anymore.
I must add to this report that I didn't notice any crash or debug anything.
I was just skimming through the 4.19.323 ChangeLog when I noticed this.

So maybe someone needs to confirm that this is a real issue, but it's 
suspicious at least.

Enjoy end-of-year breaks if you have those!

Regards,

-- 


--
Yann Sionneau | Vates XCP-ng Developer

XCP-ng & Xen Orchestra - Vates solutions

web: https://vates.tech

Re: [cip-dev] 4.19-y-cip missing backport?

[Ulrich Hecht]

> On 12/23/2025 3:19 PM CET Yann Sionneau via lists.cip-project.org <yann.sionneau=vates.tech@lists.cip-project.org> wrote:
> I noticed that 4.19.325 (as well as the latest -cip126 release) has this 
> backport:
> 
> commit 2b8f2afa311c722a90f00fb2960e6deb4f5100a5
> Author: Marek Szyprowski <m.szyprowski@samsung.com>
> Date:   Thu Aug 29 07:30:28 2019 +0200
> 
>      usb: dwc3: remove generic PHY calibrate() calls
> 
>      [ Upstream commit a0a465569b45e3690be155c96fb54603d6904f41 ]
> 
> 
> But it does not have this:
> 
> commit 34c7ed72f4f04e73901ca6ec2e1db6c040495935
> Author: Marek Szyprowski <m.szyprowski@samsung.com>
> Date:   Thu Aug 29 07:30:27 2019 +0200
> 
>      usb: core: phy: add support for PHY calibration
> 
> 
> So it seems nobody calls phy_calibrate() anymore.
> I must add to this report that I didn't notice any crash or debug anything.
> I was just skimming through the 4.19.323 ChangeLog when I noticed this.
> 
> So maybe someone needs to confirm that this is a real issue, but it's 
> suspicious at least.

Thank you for reporting this. I'm not quite sure what to do about it, though. The options are:

1. Backport "usb: core: phy: add support for PHY calibration". I like this one the least because it potentially affects all drivers, none of which have any issues ATM.
2. Revert "usb: dwc3: remove generic PHY calibrate() calls". Better in terms of regression risk, but still may have unwanted side effects for dwc3.
3. Do nothing. Least likely to introduce regressions but leaves the code in a known-bad state, although one that may not cause any real-life issue.

Any opinions, anyone?

CU
Uli

Re: [cip-dev] 4.19-y-cip missing backport?

[Yann Sionneau]

On 12/27/25 12:20, Ulrich Hecht via lists.cip-project.org wrote:
> 
>> On 12/23/2025 3:19 PM CET Yann Sionneau via lists.cip-project.org <yann.sionneau=vates.tech@lists.cip-project.org> wrote:
>> I noticed that 4.19.325 (as well as the latest -cip126 release) has this
>> backport:
>>
>> commit 2b8f2afa311c722a90f00fb2960e6deb4f5100a5
>> Author: Marek Szyprowski <m.szyprowski@samsung.com>
>> Date:   Thu Aug 29 07:30:28 2019 +0200
>>
>>       usb: dwc3: remove generic PHY calibrate() calls
>>
>>       [ Upstream commit a0a465569b45e3690be155c96fb54603d6904f41 ]
>>
>>
>> But it does not have this:
>>
>> commit 34c7ed72f4f04e73901ca6ec2e1db6c040495935
>> Author: Marek Szyprowski <m.szyprowski@samsung.com>
>> Date:   Thu Aug 29 07:30:27 2019 +0200
>>
>>       usb: core: phy: add support for PHY calibration
>>
>>
>> So it seems nobody calls phy_calibrate() anymore.
>> I must add to this report that I didn't notice any crash or debug anything.
>> I was just skimming through the 4.19.323 ChangeLog when I noticed this.
>>
>> So maybe someone needs to confirm that this is a real issue, but it's
>> suspicious at least.
> 
> Thank you for reporting this. I'm not quite sure what to do about it, though. The options are:
> 
> 1. Backport "usb: core: phy: add support for PHY calibration". I like this one the least because it potentially affects all drivers, none of which have any issues ATM.
> 2. Revert "usb: dwc3: remove generic PHY calibrate() calls". Better in terms of regression risk, but still may have unwanted side effects for dwc3.
> 3. Do nothing. Least likely to introduce regressions but leaves the code in a known-bad state, although one that may not cause any real-life issue.
> 
> Any opinions, anyone?
> 
> CU
> Uli
> 

Hello Uli,

Thanks for your answer, I think you're right about the 3 options.
Let's pull Marek Szyprowski into this so that he can give us his opinion 
since he's the Author of those patches.
I'm putting him in CC.

Regards,

-- 


--
Yann Sionneau | Vates XCP-ng Developer

XCP-ng & Xen Orchestra - Vates solutions

web: https://vates.tech

Re: [cip-dev] 4.19-y-cip missing backport?

[Marek Szyprowski]

On 28.12.2025 22:23, Yann Sionneau wrote:
> On 12/27/25 12:20, Ulrich Hecht via lists.cip-project.org wrote:
>>> On 12/23/2025 3:19 PM CET Yann Sionneau via lists.cip-project.org <yann.sionneau=vates.tech@lists.cip-project.org> wrote:
>>> I noticed that 4.19.325 (as well as the latest -cip126 release) has this
>>> backport:
>>>
>>> commit 2b8f2afa311c722a90f00fb2960e6deb4f5100a5
>>> Author: Marek Szyprowski <m.szyprowski@samsung.com>
>>> Date:   Thu Aug 29 07:30:28 2019 +0200
>>>
>>>        usb: dwc3: remove generic PHY calibrate() calls
>>>
>>>        [ Upstream commit a0a465569b45e3690be155c96fb54603d6904f41 ]
>>>
>>>
>>> But it does not have this:
>>>
>>> commit 34c7ed72f4f04e73901ca6ec2e1db6c040495935
>>> Author: Marek Szyprowski <m.szyprowski@samsung.com>
>>> Date:   Thu Aug 29 07:30:27 2019 +0200
>>>
>>>        usb: core: phy: add support for PHY calibration
>>>
>>>
>>> So it seems nobody calls phy_calibrate() anymore.
>>> I must add to this report that I didn't notice any crash or debug anything.
>>> I was just skimming through the 4.19.323 ChangeLog when I noticed this.
>>>
>>> So maybe someone needs to confirm that this is a real issue, but it's
>>> suspicious at least.
>> Thank you for reporting this. I'm not quite sure what to do about it, though. The options are:
>>
>> 1. Backport "usb: core: phy: add support for PHY calibration". I like this one the least because it potentially affects all drivers, none of which have any issues ATM.
>> 2. Revert "usb: dwc3: remove generic PHY calibrate() calls". Better in terms of regression risk, but still may have unwanted side effects for dwc3.
>> 3. Do nothing. Least likely to introduce regressions but leaves the code in a known-bad state, although one that may not cause any real-life issue.
>>
>> Any opinions, anyone?
>>
>> CU
>> Uli
>>
> Hello Uli,
>
> Thanks for your answer, I think you're right about the 3 options.
> Let's pull Marek Szyprowski into this so that he can give us his opinion
> since he's the Author of those patches.
> I'm putting him in CC.

I think You can go with 3rd option. In 4.19 kernel this phy->calibrate 
call affects only one legacy Exynos based board, which I don't think is 
or will be ever used with this kernel branch. You can always apply a 
revert (2nd option) if one complains.

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland

Re: [cip-dev] 4.19-y-cip missing backport?

[Ulrich Hecht]

> On 12/31/2025 8:23 AM CET Marek Szyprowski <m.szyprowski@samsung.com> wrote:
> >> 3. Do nothing. Least likely to introduce regressions but leaves the code in a known-bad state, although one that may not cause any real-life issue.
> 
> I think You can go with 3rd option. In 4.19 kernel this phy->calibrate 
> call affects only one legacy Exynos based board, which I don't think is 
> or will be ever used with this kernel branch. You can always apply a 
> revert (2nd option) if one complains.

I'll leave it as is then. Thanks for your help!

CU
Uli