On Wed, Aug 01, 2018 at 11:07:17AM +0100, James Morse wrote:

Hi Geoff,

On 31/07/18 16:54, Geoff Levand wrote:

quoted

This m400 HEST problem is blocking me from enabling APEI support in
Debian [1], so I'd like to come to a conclusion on it.

In summary from [2], Riku and Graeme mention they still have some m400
systems in use.  Mark did some investigation and found the cause to be
generated by the m400's firmware, appearing after the EFI stub's call
to ExitBootServices.

I did manage to borrow of one of these boxes to discover one additional piece of
information: this firmware bug doesn't exist in the latest released firmware.

The box was running the most recent firmware, and quotes its System ROM version as:
| U02 v1.01 (10/02/2015)

This thing doesn't have a HEST at all.


Why is the the latest? A current Moonshot component pack[0]'s release notes
quote m400 as removed, and say 'MCP 2017.06.0' was the latest with support. I
can't find the Release Notes for '2017.07.0', but '2017.04.0' is here[1], and it
contains:
| ProLiant_m400_Server_ROM_U02_2015_10_02.HPb


It looks like we've been talking about a bug in firmware that was never
released. Can anyone shed any light on where "U02 v1.10 (08/19/2016)" comes from?

quoted

Do we want to have an arm64 specific in-kernel fix for this, or should
it be left to the distro maintainers and end users to decide if and
how they want to handle it?

Leave it to the distro maintainers to decide. Any user capable of installing
unofficial firmware should be able to add a string to the kernel command line.

I think most people are running the firmware provided from HPe support
but was never put on release site. NDA prevents us talking about it
though!

Graeme

Thanks,

James

[0]
https://support.hpe.com/hpsc/swd/public/detail?sp4ts.oid=5409784&swItemId=MTX_7ad3cb148b4a444f96470839d8&swEnvOid=4184#tab4
[1] https://support.hpe.com/hpsc/doc/public/display?docId=c05335771