On Fri, 14 Dec 2012 13:09:17 +0100
Markus Armbruster <arm...@redhat.com> wrote:
> liguang <lig.f...@cn.fujitsu.com> writes:
>
> > if value to be translated is larger than INT64_MAX,
> > this function will not be convenient for caller to
> > be aware of it, so change a little for this.
> >
> > Signed-off-by: liguang <lig.f...@cn.fujitsu.com>
> > ---
> > cutils.c | 3 ++-
> > 1 files changed, 2 insertions(+), 1 deletions(-)
> >
> > diff --git a/cutils.c b/cutils.c
> > index 4f0692f..da05c9e 100644
> > --- a/cutils.c
> > +++ b/cutils.c
> /*
> * Convert string to bytes, allowing either B/b for bytes, K/k for KB,
> * M/m for MB, G/g for GB or T/t for TB. End pointer will be returned
> * in *end, if not NULL. Return -1 on error.
> */
Size is not the only user of it, this function is used to convert KHz,... in
target-i386/cpu.c, i.e. unit might be something else than 1024. That's why
there is question about generalizing strtosz_suffix_unit() in future instead of
duplicating suffixed int parsing. And specialization for size with
unit=1024 and enforcing limits could be made in strtosz_suffix() which is
used by current size users.
> > @@ -219,7 +219,7 @@ static int64_t suffix_mul(char suffix, int64_t unit)
> > int64_t strtosz_suffix_unit(const char *nptr, char **end,
> > const char default_suffix, int64_t unit)
> > {
> > - int64_t retval = -1;
> > + int64_t retval = EINVAL;
> > char *endptr;
> > unsigned char c;
> > int mul_required = 0;
> > @@ -246,6 +246,7 @@ int64_t strtosz_suffix_unit(const char *nptr, char
> > **end,
> > goto fail;
> > }
> > if ((val * mul >= INT64_MAX) || val < 0) {
> > + retval = ERANGE;
> > goto fail;
> > }
> > retval = val * mul;
>
> Your error codes aren't negative, and you failed to update the function
> comment!
Returning negative here is fine for now form target-i386/cpu.c pov.
--
Regards,
Igor
