On 2016/4/12 11:35, Zhenzhong Duan wrote:
On 2016/4/11 19:27, Wei Liu wrote:
On Mon, Apr 11, 2016 at 09:42:57AM +0800, Zhenzhong Duan wrote:
It's tool's duty to pass a correct cpumap to XEN. On a host with
less than
64
CPUS, it just shows below error.
[root@localhost /]# xm vcpu-pin 3 all all
Error: Cannot pin vcpu: 0 to cpu: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10,
11, 12,
13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29,
30, 31,
32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48,
49, 50,
51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63] - (22, 'Invalid
argument')
This error clearly shows that the upper layer is not passing down the
right cpumap. So this example along won't convince me to that this is a
bug in python binding. It's free to reject input that it deems wrong.
I guess the question is more or less what should it do in such
situation. So ...
The fix make it same as in xl code.
Is there reference to this? I guess the commit message or mail thread
will shed more light on the expected behaviour.
Calling path:
main_vcpupin->cpurange_parse->update_cpumap_range->libxl_bitmap_set
void libxl_bitmap_set(libxl_bitmap *bitmap, int bit)
{
if (bit >= bitmap->size * 8)
return;
bitmap->map[bit / 8] |= 1 << (bit & 7);
}
I referenced above code. It just ignore the cpu number beyond the max
CPU count rather than raise an error.
Any further comments?
thanks
zduan
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
