Re: arm64 system corruption on linux-next?

[John Garry <john.garry@xxxxxxxxxx>]

On 11/03/2020 15:04, Robin Murphy wrote:
> Hi John,
>
> On 10/03/2020 6:38 pm, John Garry wrote:
> > Hi all,
> >
> > On my arm64 Huawei D06 dev board, I see this on sometimes on 
> > linux-next 20200310 just as the boot completes:
> >
> > [   48.452674] pcieport 0000:b4:01.0: Adding to iommu group 40
> > [   48.473426] rtc-efi rtc-efi.0: setting system clock to 
> > 2020-03-10T18:31:29 UTC (1583865089)
> > [   48.473426] rtc-efi rtc-efi.0: setting system clock to 
> > 2020-03-10T18:31:29 UTC (1583865089)
> > [   48.486755] hid-generic 0003:12D1:0003.0001: input: USB HID v1.10 
> > Keyboard [Keyboard/Mouse KVM 1.1.0] on usb-0000:7a:01.0-1.1/input0
> > [   48.486755] hid-generic 0003:12D1:0003.0001: input: USB HID v1.10 
> > Keyboard [Keyboard/Mouse KVM 1.1.0] on usb-0000:7a:01.0-1.1/input0
> > [   48.491033] ALSA device list:
> > [   48A device list:
> > [   48.522304]   No soundcards found.
> > [   48.522304]   No soundcards found.
> > [   48.526319] input: Keyboard/Mouse KVM 1.1.0 as 
> > /devices/pci0000:7a/0000:7a:01.0/usb1/1-1/1-1.1/1-1.1:1.1/0003:12D1:0003.0002/input/input2 
> >
> > [   48.526319] input: Keyboard/Mouse KVM 1.1.0 as 
> > /devices/pci0000:7a/0000:7a:01.0/usb1/1-1/1-1.1/1-1.1:1.1/0003:12D1:0003.0002/input/input2 
> >
> > [|Uz
> > ���p X4n0Jj ЬH,p~wv�^;�~Tt�k�v��"��±�
> >                                        �rZ�,� � ��Hk��
> >                                                        v7�C (���� 
> > չ�E^�- z
> >                                                                           
> > �
> > ±�  � ��� �\ ���p X4n0Jj p�H,p~wv�^;+~Tt�k�v� "�
> > w�b���P�k�h�P��.�p���N�P P�X��9� 0� �k�h�P�i.�p���N��&�����.¬�jP�V�v 
> > t�Y��*�v���� �F9�
> >
> >
> > The system is still alive, as keypresses respond with garbage.

Hi Robin,

> FWIW, that looks every bit like something changing the UART baud rate 
> under the console's feet, either explicitly, or implicitly by messing 
> with the input clock behind the driver's back.

Maybe, but I didn't think that the kernel would know about the UART 
clock for ACPI-based FW.

Anyway, today's linux-next does not seem to have the issue from my 
limited testing. I'll keep an eye out.

Cheers,
John

> Robin.
>
> > Full dmesg:
> > https://pastebin.com/C2Xy0yUW
> >
> >
> > v5.6-rc4 is ok from my attempts.
> >
> > Anyone else see such an issue?
> >
> > Thanks,
> > John
> >
> > _______________________________________________
> > linux-arm-kernel mailing list
> > linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
> > http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
> .