All of lore.kernel.org
 help / color / mirror / Atom feed
From: Bertrand Marquis <Bertrand.Marquis@arm.com>
To: Julien Grall <julien@xen.org>
Cc: Dongjiu Geng <gengdongjiu1@gmail.com>,
	"sstabellini@kernel.org" <sstabellini@kernel.org>,
	"Volodymyr_Babchuk@epam.com" <Volodymyr_Babchuk@epam.com>,
	"xen-devel@lists.xenproject.org" <xen-devel@lists.xenproject.org>
Subject: Re: [PATCH v4] xen/arm: Allow QEMU platform to be built with GICv2
Date: Tue, 18 Jan 2022 08:32:19 +0000	[thread overview]
Message-ID: <C744D1DD-9AF1-4CBA-96F1-1338A44F7C53@arm.com> (raw)
In-Reply-To: <2a1a0dd5-b8a7-8072-2961-b1fbe9ec746a@xen.org>

Hi,

> On 17 Jan 2022, at 14:15, Julien Grall <julien@xen.org> wrote:
> 
> Hi,
> 
> On 17/01/2022 10:40, Dongjiu Geng wrote:
>> It turns out that QEMU has been supporting GICv2 virtualization since
>> v3.1.0. So remove the dependencies on GICv3.
> 
> 
> Technically, the current form of CONFIG_QEMU allows the same binary to boot on QEMU with GICv2 or GICv3.
> 
>> If we want to use GICv3,
>> we can select the QEMU_LEGACY configuration.
> 
> AFAIK, GICv3 is not a legacy feature... So it feels a bit odd to name it like that (see more below).
> 
>> Signed-off-by: Dongjiu Geng <gengdongjiu1@gmail.com>
>> ---
>>  xen/arch/arm/platforms/Kconfig | 10 +++++++++-
>>  1 file changed, 9 insertions(+), 1 deletion(-)
>> diff --git a/xen/arch/arm/platforms/Kconfig b/xen/arch/arm/platforms/Kconfig
>> index c93a6b2756..41e82a42ee 100644
>> --- a/xen/arch/arm/platforms/Kconfig
>> +++ b/xen/arch/arm/platforms/Kconfig
>> @@ -13,7 +13,15 @@ config ALL_PLAT
>>  	automatically select any of the related drivers.
>>    config QEMU
>> -	bool "QEMU aarch virt machine support"
>> +	bool "QEMU aarch virt machine support >= v3.1.0"
> 
> This is a bit misleading. A user may select this thinking that this will select GICv3. However, this will not.
> 
> This also raises the question of what is the default GIC version in QEMU (i.e. if you don't pass anything on the command line)? If this is GICv3, then I am afraid that this patch would be a no-go for me.
> 
> Looking at overall discussion, you seem to push the patch only to allow building a tiny Xen for QEMU and the new vGIC.
> 
> The default Xen (i.e. make defconfig) will also work on QEMU. Given that the new vGIC is a still in development, I am seriously considering to say that if you want to try it then you have to use the default configuration.
> 
> @Dongjiu, is there any reason why you want to use the tiny QEMU config rather than the default configuration?
> 
> @Bertrand, @Stefano, what do you think?

I do think that the new Legacy option is very misleading and users will tend to select Qemu instead of the legacy one where having GICv3 makes more sense.
Using the new VGIC which is unsupported is not a standard use case so maybe it would make more sense to create a specific Qemu entry for GICv2 using new VGIC instead and keep the default as is.
To sum up I am no in favour of a solution changing the existing QEMU. A solution creating a new one for the wanted use case would be better for me.

Cheers
Bertrand

> 
> Cheers,
> 
> -- 
> Julien Grall



  reply	other threads:[~2022-01-18  8:33 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-01-17  6:40 [PATCH v4] xen/arm: Allow QEMU platform to be built with GICv2 Dongjiu Geng
2022-01-17  9:00 ` Bertrand Marquis
2022-01-17 11:12   ` Dongjiu Geng
2022-01-17 11:38     ` Bertrand Marquis
2022-01-17 11:49       ` Dongjiu Geng
2022-01-17 14:15 ` Julien Grall
2022-01-18  8:32   ` Bertrand Marquis [this message]
2022-01-18  8:45   ` Dongjiu Geng
2022-01-18  8:48     ` Bertrand Marquis
2022-01-18  8:58       ` Dongjiu Geng
2022-01-18  9:17         ` Bertrand Marquis
2022-01-19 11:30           ` Dongjiu Geng
2022-01-23 10:31             ` Julien Grall
2022-01-23 10:42         ` Julien Grall

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=C744D1DD-9AF1-4CBA-96F1-1338A44F7C53@arm.com \
    --to=bertrand.marquis@arm.com \
    --cc=Volodymyr_Babchuk@epam.com \
    --cc=gengdongjiu1@gmail.com \
    --cc=julien@xen.org \
    --cc=sstabellini@kernel.org \
    --cc=xen-devel@lists.xenproject.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.