On Fri, 22 Mar 2013 14:44:05 -0600 Eric Blake <[email protected]> wrote:
> On 03/22/2013 02:31 PM, Luiz Capitulino wrote: > > Today we reset fd_counter if it wraps, but it's better to abort() > > instead, as fd_counter should never reach INT64_MAX. > > > > Signed-off-by: Luiz Capitulino <[email protected]> > > --- > > qga/main.c | 8 ++++++-- > > 1 file changed, 6 insertions(+), 2 deletions(-) > > > > > > diff --git a/qga/main.c b/qga/main.c > > index 74ef788..5f505a2 100644 > > --- a/qga/main.c > > +++ b/qga/main.c > > @@ -889,9 +889,13 @@ int64_t ga_get_fd_handle(GAState *s, Error **errp) > > g_assert(!ga_is_frozen(s)); > > > > handle = s->pstate.fd_counter++; > > - if (s->pstate.fd_counter < 0) { > > - s->pstate.fd_counter = 0; > > + > > + /* This should never happen on a resonable timeframe, as > > guest-file-open > > s/resonable/reasonable/ Michael, can you fix it for me? > > > + * would have to be issued 2^63 times */ > > + if (s->pstate.fd_counter == INT64_MAX) { > > + abort(); > > Fix the typo, and you can add: > > Reviewed-by: Eric Blake <[email protected]> Thanks!
