[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] XEN[ARM] Master not working on Allwinner A20



On Sun, 5 Jan 2014, karim.allah.ahmed@xxxxxxxxx wrote:
> On Sun, Jan 5, 2014 at 7:28 PM, Stefano Stabellini
> <stefano.stabellini@xxxxxxxxxxxxx> wrote:
> > Could you use plain text for emails, please?
> Sorry about that :)
> 
> >
> > On Sun, 5 Jan 2014, karim.allah.ahmed@xxxxxxxxx wrote:
> >> On Sun, Jan 5, 2014 at 5:39 PM, Stefano Stabellini 
> >> <stefano.stabellini@xxxxxxxxxxxxx> wrote:
> >>       On Sun, 5 Jan 2014, karim.allah.ahmed@xxxxxxxxx wrote:
> >>       > Hi Peter,
> >>       >
> >>       > If you still can't boot with any memory bigger than 128M, as a 
> >> fast workaround you can apply this patch.
> >>       >
> >>       > diff --git a/xen/arch/arm/domain_build.c 
> >> b/xen/arch/arm/domain_build.c
> >>       > index faff88e..849df3f 100644
> >>       > --- a/xen/arch/arm/domain_build.c
> >>       > +++ b/xen/arch/arm/domain_build.c
> >>       > @@ -22,7 +22,7 @@
> >>       >  static unsigned int __initdata opt_dom0_max_vcpus;
> >>       >  integer_param("dom0_max_vcpus", opt_dom0_max_vcpus);
> >>       >
> >>       > -static int dom0_11_mapping = 1;
> >>       > +static int dom0_11_mapping = 0;
> >>       >
> >>       >  #define DOM0_MEM_DEFAULT 0x8000000 /* 128 MiB */
> >>       >  static u64 __initdata dom0_mem = DOM0_MEM_DEFAULT;
> >>       >
> >>       >
> >>       > It's failing because none of the zones has a contiguous memory 
> >> block with an order bigger than 15 ( 128M ). I think
> >>       this is due
> >>       > to the alignment of the phys_start with buddy system in 
> >> cubieboard, I'll look further and let you know if there's a
> >>       cleaner
> >>       > approach to fix that.
> >>       >
> >>       > It used to work before because the 11_mapping wasn't forced to 
> >> "true" for all platforms and there was a quirk
> >>       exposed by the
> >>       > platform that used to express that. I think Julien removed that 
> >> quirk and defaulted to 11_mapping in commit
> >>       > "71952bfcbe9187765cf4010b1479af86def4fb1f"
> >>
> >> Unfortunately dom0_11_mapping is needed if at least one device driver
> >> for the Allwinner uses DMA.
> >> For example, if you disable dom0_11_mapping, can you still access the
> >> network? On the other hand if all device drivers do not use DMA we can
> >> set dom0_11_mapping to false for this platform.
> >>
> >>
> >> I'm not quite sure about all devices in cubieboard, but at least for the 
> >> network case I think it'll still work ( well, it's
> >> working for me ) . Besides, Cubieboard didn't have this quirk to begin 
> >> with before defaulting to the 11_mapping
> >
> > What is the linux device driver that you are using for the network? And
> > the one for the disk/sdcard?
> 
> For network "sun4i-emac", and currently I'm mounting my rootfs through
> nfs, so no disk/sdcard

Ah, that explains it! AFAICT sun4i-emac does not use DMA.
So yes, if this is your use case, you can safely disable the 1:1
workaround.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.