[edk2-devel] [PATCH 0/9] ArmVirtPkg: support two PL011 UARTs

[Ard Biesheuvel]

On Thu, 26 Oct 2023 at 17:19, Laszlo Ersek <lersek at redhat.com> wrote:
>
> On 10/26/23 16:21, Peter Maydell wrote:
> > On Tue, 10 Oct 2023 at 16:33, Laszlo Ersek <lersek at redhat.com> wrote:
> >> On 10/10/23 09:43, Ard Biesheuvel wrote:
> >>> Thanks for looking into this - a cleanup was overdue here.
> >>>
> >>> I will take a look in more detail later, but one thing that occurred
> >>> to me when reading this overview is that having a separate DEBUG
> >>> serial port would permit us to
> >>>
> >>> a) remove it from the DT
> >>
> >> ... as in, hide it from Linux, I assume?
> >>
> >>> b) add a runtime mapping for it
> >>> c) keep using it after ExitBootServices
> >>>
> >>> This could be useful for debugging issues with the variable store etc.
> >>>
> >>> Not saying this is something to address in this series, but I'd like
> >>> to hear your take on this.
> >>>
> >>
> >> Sounds like a useful feature.
> >>
> >> I see four challenges:
> >>
> >>
> >> (1) We'd have to coordinate it with Peter. If we hide any one of the
> >> serial ports from Linux, that may not be what QEMU intends for Linux to
> >> happen. Linux currently ties getties to all serial ports -- via the
> >> serial* aliases, IIUC. Thus, some "positive identification" in the DT
> >> could be necessary (i.e., that edk2 was welcome to hide that port from
> >> Linux).
> >
> > The potential awkwardness here is that what the guest thinks about
> > the serial ports depends on the ACPI table fragments which QEMU
> > provides. EDK2 would need to edit the table fragment to remove any
> > mention of the second UART if it wanted to hide it from the kernel.
> > I don't know how hard that would be in EDK2.
> >
> > (As far as I'm aware usually a boot via EDK2 doesn't pass the
> > dtb on to Linux, though I guess there's no reason it can't.)
> >
> > From QEMU's point of view, we provide two UARTs to the guest, and we
> > don't really care whether that means one is used by EDK2 and one by
> > Linux, or both are used as getty terminals by Linux, or whether the
> > Linux guest uses one serial as a terminal and leaves the other to its
> > userspace programs  -- it's all just guest software to us :-)
> >
> > [snip other technical stuff]
>
> Thanks, good point -- I wasn't aware of the ACPI impact.
>
> We don't edit / patch QEMU's ACPI tables, ever. (Beyond obeying the ACPI
> linker/loader script.) That's a principle we've upheld many times.
> Whenever ACPI content needs to change, that implies a QEMU patch.
>
> So, for this purpose, only the following could have a chance of working:
>
> - Expose a new config option on the QEMU command line to the user,
> regarding the intended use of the serial port(s). This could be of any
> tolerable form (machine property, front-end (device) property, whatever
> -- anything that QEMU reviewers can accept).
>
> - In QEMU, generate both the DT and the ACPI tables accordingly. The
> ACPI tables would have to immediately *not* contain the UART-to-hide (so
> as to keep it secret from the guest OS). The DT at the same time would
> still have to expose the "runtime DEBUG UART", because edk2 would have
> to know where that UART was (and that it was meant specifically for OS
> runtime debug output).
>
> - Edk2 would have to patch the DT (we tend to do that already), because
> (in some configs) we do forward the DT to the guest OS. This need for
> patching could be lifted if QEMU adopted such a form of expression for
> the "runtime DEBUG UART" that would be ignored by Linux out of the box.
>
> >
> >> All in all, I think the implementation would be quite a steep divergence
> >> from, or on top of, this patch set. :)
> >
> > I agree with this and with Ard's "not something to address in this
> > series" comment above; it doesn't sound like this is something that
> > needs to hold up the patchset we have currently.
>
> Right; I'd like to flush this one. The runtime debug UART seems to need
> more joint pondering.
>
> >
> > Does anybody have time to review Laszlo's code? It would be nice
> > to be able to get this into the next EDK2 release.
>

I'm happy for this to go in if it covers our needs.

Acked-by: Ard Biesheuvel <ardb at kernel.org>


-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#110115): https://edk2.groups.io/g/devel/message/110115
Mute This Topic: https://groups.io/mt/101834880/1813853
Group Owner: devel+owner at edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [edk2-devel-archive at redhat.com]
-=-=-=-=-=-=-=-=-=-=-=-