Re: [Xen-devel] [PATCH RFC 34/35] arm : acpi workarounds for firmware/linux dependencies

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 10.02.15 at 10:38, <julien.grall@xxxxxxxxxx> wrote:
> Why it's working on x86? This big allocation is done via the boot 
> allocator memory (because the system state is early boot). Hopefully, we 
> never have to resize it.
> 
> On ARM64, ACPI is initialized after the boot allocator has ended, so we 
> have to use xmalloc which will return a page-align pointer.
> 
> As ACPI on ARM64 will never use the boot allocator,

How come you're so certain? The reason why on x86 it needs to
be done this way is because of NUMA initialization. Are you not
expecting NUMA to become of interest to ARM? Or if you do, are
you sure you can get away without the same ordering that x86
uses?

> maybe the 
> xmalloc_is_ptr is acceptable? It would require a big explanation which 
> it's safe.

Indeed, it _might_ be acceptable, but certainly not without a proper
comment.

Jan


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