Ok
I now tested the whole thing also with some of the v5.4.y kernels. Just to avoid misunderstandings I listed all my tests below:
Kernel | Hash | Bug | Comment
Version | (date) | present |
----------|--------------------------------------------|---------|--------------
v4.19.72 | ee809c7e08956d737cb66454f5b6ca32cc0d9f26 | No | This is the old system we are trying to replace
| (10. September 2019) | | with the new Kernel 5.15
----------|--------------------------------------------|---------|--------------
v4.20 | ae8a2ca8a2215c7e31e6d874f7303801bb15fbbc | ? | This kernel has never been tested as our build
| (20. September 2018) | | can not build this kernel, but I mentioned it
| | | because it is the first kernel where the tcpm
| | | module has been moved to its own subdirectory
----------|--------------------------------------------|---------|--------------
v5.15.74 | f0bee94053065c7cb8eacadfdd6bf739a2042b35 | Yes | This is the Kernel we would like to use for the
| (19. October 2022) | | new system.
| | | It is working with with a USB-PD capable power-
| | | supply but connecting a USB-A to USB-C cable
| | | provokes the bug.
----------|--------------------------------------------|---------|--------------
v5.11 | b5206275b46c30a8236feb34a1dc247fa3683d83 | Yes | This kernel behaves exactly like 5.15.74
| (11. December 2020) | |
----------|--------------------------------------------|---------|--------------
v5.10-rc1 | 3650b228f83adda7e5ee532e2b90429c03f7b9ec | Yes | This kernel behaves exactly like 5.15.74
| (25. October 2020) | |
----------|--------------------------------------------|---------|--------------
v5.4.219 | 35826e154ee014b64ccfa0d1f12d36b8f8a75939 | Yes | This kernel performs segnificantly worse in
| (19. October 2022) | worse | nagotiating with the USP-PD power-supply and
| | | even crashes about 50% of the tries. It then
| | | gets stuck in the ISR just as it does 100%
| | | of the time when connecting the USB-A to USB-C
| | | cable. (and even when disconnecting the cable)
----------|--------------------------------------------|---------|--------------
v5.4.1 | 79438f37a69a80822333c86acb06a71abce1bce3 | ? | The build system was unable to compile the kernel
| (29. November 2019) | | scripts/dtc/dtc-parser.tab.o:(.bss+0x20):
| | | multiple definition of `yylloc';
| | | scripts/dtc/dtc-lexer.lex.o:(.bss+0x0)
----------|--------------------------------------------|---------|--------------
v5.4.110 | 59c8e332926875d337f426fde14fec986faab414 | Yes | Same behaviour as 5.4.219
| (7. April 2021) | |
----------|--------------------------------------------|---------|--------------
v5.4.55 | 169b93899c7dfb93a2b57da8e3505da9b2afcf5c | Yes | Same behaviour as 5.4.219
| (31. July 2020) | |
----------|--------------------------------------------|---------|--------------
v5.4.28 | 462afcd6e7ea94a7027a96a3bb12d0140b0b4216 | ? | The build system was unable to compile the kernel
| (25. April 2020) | | scripts/dtc/dtc-parser.tab.o:(.bss+0x20):
| | | multiple definition of `yylloc';
| | | scripts/dtc/dtc-lexer.lex.o:(.bss+0x0)
I would love to send some kernel logs but I can not see any entry in dmesg. Can you instruct me how to get the corresponding kernel logs?
-----Ursprüngliche Nachricht-----
Von: Guenter Roeck <groeck7@xxxxxxxxx> Im Auftrag von Guenter Roeck
Gesendet: Dienstag, 25. Oktober 2022 15:37
An: Christian Bach <christian.bach@xxxxxx>; Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx>
Cc: stable@xxxxxxxxxxxxxxx; regressions@xxxxxxxxxxxxxxx; linux-usb@xxxxxxxxxxxxxxx
Betreff: Re: AW: tcpci module in Kernel 5.15.74 with PTN5110 not working correctly
Hi,
On 10/25/22 05:48, Christian Bach wrote:
> Thank you for answering. I did try a Kernel from 1 year ago (11. December 2020 - with the Hash b5206275b46c30a8236feb34a1dc247fa3683d83). But this Kernel had the exact same behavior.
> I even wanted to go back further the when the tcpm module got it's own subdirectory (v4.20 - 20. September 2018 - Hash ae8a2ca8a2215c7e31e6d874f7303801bb15fbbc) to see if it still worked at that time but my build system was not able to build it.
>
Greg asked for you to test with a v5.4.y kernel. ae8a2ca8a221..b5206275b46c30a82 is again a pretty large step with more than 100 commits in the drivers/usb/typec/tcpm/ directory.
Also, it might be useful to provide the respective kernel logs.
Thanks,
Guenter
> -----Ursprüngliche Nachricht-----
> Von: Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx>
> Gesendet: Dienstag, 25. Oktober 2022 14:38
> An: Christian Bach <christian.bach@xxxxxx>
> Cc: stable@xxxxxxxxxxxxxxx; regressions@xxxxxxxxxxxxxxx;
> linux@xxxxxxxxxxxx; linux-usb@xxxxxxxxxxxxxxx
> Betreff: Re: tcpci module in Kernel 5.15.74 with PTN5110 not working
> correctly
>
> [You don't often get email from gregkh@xxxxxxxxxxxxxxxxxxx. Learn why
> this is important at https://aka.ms/LearnAboutSenderIdentification ]
>
> On Tue, Oct 25, 2022 at 12:19:39PM +0000, Christian Bach wrote:
>> Hello
>>
>> For a few weeks now I am trying to make the PTN5110 chip work with the new Kernel 5.15.74. The same hardware setup was working with the 4.19.72 Kernel. The steps I took so far are as follows:
>
> That is a huge jump. Why not use 'git bisect'?
>
> Or start with a smaller jump. Why not go to 5.4.y first, that's only a year's worth of changes, instead of 4 years of changes.
>
> thanks,
>
> greg k-h
