Re: [Xen-devel] [v6][PATCH 3/7] tools/libxc: check if mmio BAR is out of reserved device memory maps

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 11.09.14 at 17:38, <JBeulich@xxxxxxxx> wrote:
>>>> On 10.09.14 at 07:49, <tiejun.chen@xxxxxxxxx> wrote:
>> +                             mmio_start, mmio_size) )
> 
> Can you really use check_mmio_hole() here? I'm not certain that
> its two one-offs are intentional (to allow for a gap between MMIO
> and RAM), but these are certainly harmful to you.

Oops - I was sure I deleted this after closer looking. Please
disregard this.

However, the while thing here seems to get done in the wrong
place anyway: The (mmio_start,mmio_size) tuple here doesn't
represent any actual MMIO assignments, just the hole below
4G that's being left available for such. Instead you'd need to
avoid the specific regions in hvmloader's pci_setup().

Plus, considering migration, all of this isn't going to help if on
the destination host (in order for subsequent device hotplug to
work) different RMRR regions need to be taken care of.

All in all I continue thinking that we'd be best off disallowing
assignment of devices associated with RMRR regions when the
1:1 mappings needed collide with something already there,
perhaps with a 2-level operator override to allow USB ones
(possibly making the current USB-specific code conditional) or
anything at their own risk.

Jan


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