On 02.05.2016 16:19, Erik Skultety wrote:
>>> diff --git a/src/rpc/virnetserver.c b/src/rpc/virnetserver.c
>>> index 57bd95c..8216da6 100644
>>> --- a/src/rpc/virnetserver.c
>>> +++ b/src/rpc/virnetserver.c
>>> @@ -65,6 +65,7 @@ struct _virNetServer {
>>>
>>> size_t nclients; /* Current clients count */
>>> virNetServerClientPtr *clients; /* Clients */
>>> + unsigned long long next_client_id; /* next client ID */
>>
>> Any reason this shouldn't be size_t just like the rest below?
>>
>>> size_t nclients_max; /* Max allowed clients count */
>>> size_t nclients_unauth; /* Unauthenticated clients count */
>>> size_t nclients_unauth_max; /* Max allowed unauth clients count */
>>
>
> Well, it can't...according POSIX, size_t isn't guaranteed to be always
> at least 64-bits, it just has to represent the largest unsigned type
> supported by an implementation and also according to
> asm-generic/posix_types.h most 32-bit archs use unsigned int for
> size_t...which isn't what we actually want, since clients come and go
> and we just increment the counter till the daemon's end of life and imho
> we do want the largest type possible here....the same applies below...
Interesting. I thought that size_t is always the biggest possible type
(not by definition, but by actual code). In that case, we should stick
with unsigned long long.
ACK to your version.
Michal
--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list