Re: kernel 6.2 stuck at boot (efi_call_rts) on arm64

[Darren Hart]

On Thu, Apr 13, 2023 at 10:24:38PM +0200, Andrea Righi wrote:
> 
> Not sure if it's a similar issue, but I have found another Ampere box
> that is booting fine with your fixes, but the eifvars.sh kselftest is
> failing with some I/O errors, specifically:

Thanks for reporting. Can you confirm this worked reliably for you prior
to v6.1?

--
Darren

> 
> $ sudo ./efivarfs.sh
> --------------------
> running test_create
> --------------------
> ./efivarfs.sh: line 58: printf: write error: Input/output error
> /sys/firmware/efi/efivars/test_create-210be57c-9849-4fc7-a635-e6382d1aec27 has invalid size
>   [FAIL]
> --------------------
> running test_create_empty
> --------------------
>   [PASS]
> --------------------
> running test_create_read
> --------------------
>   [PASS]
> --------------------
> running test_delete
> --------------------
> ./efivarfs.sh: line 103: printf: write error: Input/output error
>   [PASS]
> --------------------
> running test_zero_size_delete
> --------------------
> ./efivarfs.sh: line 126: printf: write error: Input/output error
> ./efivarfs.sh: line 134: printf: write error: Input/output error
> /sys/firmware/efi/efivars/test_zero_size_delete-210be57c-9849-4fc7-a635-e6382d1aec27 should have been deleted
>   [FAIL]
> --------------------
> running test_open_unlink
> --------------------
> open(O_WRONLY): Operation not permitted
>   [FAIL]
> --------------------
> running test_valid_filenames
> --------------------
> ./efivarfs.sh: line 158: printf: write error: Input/output error
> ./efivarfs.sh: line 158: printf: write error: Input/output error
> ./efivarfs.sh: line 158: printf: write error: Input/output error
> ./efivarfs.sh: line 158: printf: write error: Input/output error
>   [PASS]
> --------------------
> running test_invalid_filenames
> --------------------
>   [PASS]
> 
> If it helps:
> 
> $ sudo hexdump -C /sys/firmware/dmi/entries/4-0/raw
> 00000000  04 30 04 00 01 03 fe 02  c1 d0 3f 41 00 00 00 00  |.0........?A....|
> 00000010  03 8a 72 06 b8 0b f0 0a  41 06 05 00 06 00 07 00  |..r.....A.......|
> 00000020  04 05 06 50 50 50 04 00  01 01 01 00 01 00 01 00  |...PPP..........|
> 00000030  43 50 55 20 31 00 41 6d  70 65 72 65 28 52 29 00  |CPU 1.Ampere(R).|
> 00000040  41 6d 70 65 72 65 28 52  29 20 41 6c 74 72 61 28  |Ampere(R) Altra(|
> 00000050  52 29 20 50 72 6f 63 65  73 73 6f 72 00 30 30 30  |R) Processor.000|
> 00000060  30 30 30 30 30 30 30 30  30 30 30 30 30 30 32 35  |0000000000000025|
> 00000070  35 30 32 30 39 30 33 33  38 36 35 42 34 00 30 30  |50209033865B4.00|
> 00000080  30 30 30 30 30 31 00 51  38 30 2d 33 30 00 00     |000001.Q80-30..|
> 0000008f
> 
> I guess EFI is not very reliable here...
> 
> -Andrea

-- 
Darren Hart
Ampere Computing / OS and Kernel