Re: [Qemu-devel] [PATCH v14 04/13] Add cache handling functions

[Eric Blake <eblake@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 825 bytes --]

On 07/03/2012 01:23 PM, Blue Swirl wrote:

>> +
>> +static inline int64_t round2pow2(int64_t value)

round up or down?

>> +{
>> +    while (!is_power_of_2(value)) {
>> +        value &=  ~(1 << (ffs(value) - 1));
> 
> ffs() only uses 'int', not int64_t.  ffsl() is not universally available.
> 
>> +    }
>> +    return value;
>> +}

Not to mention that iterating one bit at a time is inefficient.  We
already gave you several more efficient solutions; my favorite being:

static inline int64_t pow2floor(int64_t value)
{
    if (!is_power_of_2(value)) {
        value = 0x8000000000000000ULL >> clz64(value);
    }
    return value;
}

since clz64() is already part of qemu sources.

-- 
Eric Blake   eblake@redhat.com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 620 bytes --]