Re: [Qemu-devel] [PATCH RFC V2 4/4] Add virtv2 machine that uses GIC-500

From: Shlomo Pongratz <shlomo.pongratz@huawei.com>
To: Eric Auger <eric.auger@linaro.org>,
	"shlomopongratz@gmail.com" <shlomopongratz@gmail.com>,
	"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>
Cc: "peter.maydell@linaro.org" <peter.maydell@linaro.org>,
	Pavel Fedin <p.fedin@samsung.com>,
	"ashoks@broadcom.com" <ashoks@broadcom.com>
Subject: Re: [Qemu-devel] [PATCH RFC V2 4/4] Add virtv2 machine that uses GIC-500
Date: Tue, 19 May 2015 15:07:58 +0000	[thread overview]
Message-ID: <B5C21809C482C148A58516F588CD431467ACA3@szxeml557-mbx.china.huawei.com> (raw)
In-Reply-To: <555B4B0C.1070506@linaro.org>

> -----Original Message-----
> From: Eric Auger [mailto:eric.auger@linaro.org]
> Sent: Tuesday, 19 May, 2015 5:39 PM
> To: shlomopongratz@gmail.com; qemu-devel@nongnu.org
> Cc: peter.maydell@linaro.org; Shlomo Pongratz
> Subject: Re: [Qemu-devel] [PATCH RFC V2 4/4] Add virtv2 machine that uses
> GIC-500
> 
> Hi Shlomo,
> On 05/06/2015 04:04 PM, shlomopongratz@gmail.com wrote:
> > From: Shlomo Pongratz <shlomo.pongratz@huawei.com>
> >
> > There is a need to support flexible clusters size. The GIC-500 can
> > support up to 128 cores, up to 32 clusters and up to 8 cores is a cluster.
> > So for example, if one wishes to have 16 cores, the options are:
> > 2 clusters of 8 cores each, 4 clusters with 4 cores each Currently
> > only the first option is supported.
> > There is an issue of passing clock affinity to via the dtb. In the dtb
> >
> > interrupt section there are only 24 bit left to affinity since the
> > variable is a 32 bit entity and 8 bits are reserved for flags.
> > See Documentation/devicetree/bindings/arm/arch_timer.txt.
> > Note that this issue is not seems to be critical as when checking
> > /proc/irq/3/smp_affinity with 32 cores all 32 bits are one.
> 
> As a general comment your series needs a good rebase in virt.c. This would
> definitively ease the review.
> 
> I have the impression that the number of modifications related to the
> GICv3 addition do not require to create a new virt.c file. My personal feeling
> is using Ashok's approach based on machine properties may be worth trying.
> Obviously this will urge you to fill the memory map holes but that's should be
> feasible.
> 

Hi Eric,

Thank you.

I have just pulled from git and I'm currently working on rebasing my 1-3 patches to commit 62bf3d , mainly adding interrupt groups and fiq.
I accept that virtv2 is not needed and I'm currently using Pavel's patch https://lists.gnu.org/archive/html/qemu-devel/2015-05/msg02930.html with small modifications.
However If there is a consensus I'll move to Ashok's virt instead of Pavel's one.

I'm looking forward for a decision regarding which virt is to be used.

Best regards,

S.P. 