Re: Discussion: Patch series that adds disable-tcg option for ppc targets

From: David Gibson <david@gibson.dropbear.id.au>
To: Bruno Piazera Larsen <bruno.larsen@eldorado.org.br>
Cc: "qemu-devel@nongnu.org" <qemu-devel@nongnu.org>,
	Lucas Mateus Martins Araujo e Castro
	<lucas.araujo@eldorado.org.br>,
	Fernando Eckhardt Valle <fernando.valle@eldorado.org.br>,
	"qemu-ppc@nongnu.org" <qemu-ppc@nongnu.org>,
	Andre Fernando da Silva <andre.silva@eldorado.org.br>,
	Matheus Kowalczuk Ferst <matheus.ferst@eldorado.org.br>,
	Luis Fernando Fujita Pires <luis.pires@eldorado.org.br>
Subject: Re: Discussion: Patch series that adds disable-tcg option for ppc targets
Date: Tue, 6 Apr 2021 16:11:30 +1000	[thread overview]
Message-ID: <YGv7ku+6XNMW1QFM@yekko.fritz.box> (raw)
In-Reply-To: <CP2PR80MB44990766F1D104E630ADFDDAC7779@CP2PR80MB4499.lamprd80.prod.outlook.com>

[-- Attachment #1: Type: text/plain, Size: 2585 bytes --]

On Mon, Apr 05, 2021 at 01:32:18PM +0000, Bruno Piazera Larsen wrote:
> Hi all,
> 
> The team I'm working on started to work on adding support for
> building the ppc target with the disable-tcg option. However, we're
> not quite sure on where to start with such a big patch.
> 
>     * Should we focus first on changing the .c files, so that it
>       will build when we finally patch the meson.build everything
>       build correctly? Or should we start by changing the meson
>       file, so that everyone agrees that the excluded files can
>       indeed be excluded?

It's usually best to focus on logical changes, rather than
file-by-file.  That said, I'd probably suggest changing the .c files
first, then changing the meson file.

>     * Should we first finish the whole series to then mail it, or
>       should we send one file at a time, as soon as we have them
>       ready?

I'd lean towards building a whole series, since I think the individual
file changes won't make a lot of sense on their own.  That said, it's
ok and encouraged to post a relatively early draft of the series as an
RFC, so that the overall idea can be reviewed, even if it has obvious
omissions (just mention them in the cover letter).

> So far we're thinking we'll need to:
>     * change 6 files (arch_dump.c, machine.c, mmu-hash32.c,
>       mmu-hash64.c, mmu-radix64.c and meson.build);

I'd expect  mmu-*.c to be excluded.  Those are softmmu implementations
which shouldn't be used with KVM.  It's possible there are a few
helpers that will need to be extracted, though.

You'll probably also need changes in hw/ppc/spapr_hcall.c and maybe
some other parts of the spapr code: there are a number of hypercalls
that we implement in qemu for TCG, but which are (and must be)
implemented in KVM when KVM is in use.  So, I expect you'll need to
suppress compilation of h_enter, h_remove, h_protect, h_read and
h_bulk_remove at least in the !TCG case.

>     * exclude 8 files from the build (dfp_helper.c, excp_helper.c,
>       fpu_helper.c, int_helper.c, mem_helper.c, misc_helper.c, *
>       translate.c, timebase_helper.c);

That looks about right.

>     * create a new one that holds stubs.

Sounds reasonable.

> Does this sound about right? This is mostly guesswork based on how
> x86, s390x and ARM are doing it.

Sounds fine so far.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]