Re: [Xen-devel] [PATCH] VPMU issue on Nehalem cpus

[Keir Fraser <keir@xxxxxxx>]

On 22/11/2010 09:19, "Jan Beulich" <JBeulich@xxxxxxxxxx> wrote:

>>>> +    val = msr_content & ((1 << num_gen_pmc) - 1);
>>> 
>>> What's the point of masking if the subsequent loop looks at the
>>> bottom so many bits only anyway?
>> 
>> Bits 0-31 flag the overflow of the general counters (currently max 4) and
>> 32-63
>> flag the overflow of the fixed counter (currently max 3).
>> Yes the first mask is not necessary, maybe a comment would be better?
> 
> Neither is the second mask (below) - the shift is all that's really
> needed. Afaic, a comment doesn't seem necessary, but Keir
> may by of different opinion here.

It's clear from the code that the mask operation is unnecessary. No code
comment required.

 -- Keir



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel