[PATCH 4/4] arm64: gicv3: its: Increase FORCE_MAX_ZONEORDER for Cavium ThunderX

[robert.richter@caviumnetworks.com (Robert Richter)]

On 12.05.15 18:24:16, Will Deacon wrote:
> On Tue, May 12, 2015 at 05:20:49PM +0100, Robert Richter wrote:
> > For allocation of 16MB cont. phys mem of a defconfig kernel (4KB
> > default pagesize) I see this different approaches:
> 
> 16MB sounds like an awful lot. Is this because you have tonnes of MSIs or
> a sparse DeviceID space or both?
> 
> >  * set FORCE_MAX_ZONEORDER to 13 as default,
> > 
> >  * set FORCE_MAX_ZONEORDER to 13 if ARM_GIC_V3 is set,
> > 
> >  * set FORCE_MAX_ZONEORDER to 13 if ARCH_THUNDER is set (this patch),
> 
> I'm not hugely fond of these suggestions, as there's still no guarantee
> that such a huge allocation is going to succeed and we end up bumping
> MAX_ORDER for all platforms in defconfig if we enable THUNDER there.

I actually was expecting this...

> >  * use hugepages if enabled (defconfig has the following options
> >    enable: CGROUP_HUGETLB, TRANSPARENT_HUGEPAGE, HUGETLBFS, this might
> >    work with current default kernel without changing defconfig
> >    options),
> 
> I don't think hugepages help with DMA.
> 
> >  * use devicetree to reserve mem for gicv3 (need to check ACPI).

I am quite a bit concerned letting firmware handle this. But if that
would solve it, fine.

> Using a carveout like this might be the best bet. I assume the memory used
> by the ITS can never be reclaimed by the syste (and therefore there's no
> issue with wastage)?
> 
> > Do you see any direction?
> 
> Dunno, does CMA also require the MAX_ORDER bump?

Looks promising at the first glance. Will look into it.

Thanks,

-Robert