On 10/26/2015 04:44 PM, Eric Blake wrote: > On 10/26/2015 04:06 PM, John Snow wrote: >> Make sure there's not trailing garbage, e.g. >> "64k-whatever-i-want-here" >> >> Reported-by: Max Reitz <mre...@redhat.com> >> Signed-off-by: John Snow <js...@redhat.com> >> --- >> qemu-io-cmds.c | 9 ++++++++- >> 1 file changed, 8 insertions(+), 1 deletion(-)
>> + ret = qemu_strtosz_suffix(s, &end, QEMU_STRTOSZ_DEFSUFFIX_B);
>> + if (*end != '\0') {
>> + /* Detritus at the end of the string */
>> + return -EINVAL;
>> + }
>> + return ret;
>> }
>
> Eww. This mixes up two return types, negative errno, and negative
> input. User input of -22 shouldn't behave differently than -21, just
> because it happens to match -EINVAL.
>
> Do we ever want to allow a negative return from cvtnum(), or should we
> just blindly map a negative int64_t into -ERANGE for a contract that we
> only accept 63-bit numbers?
Uggh. Maybe I should read qemu_strtosz_suffix() before making bogus
claims (and assuming that it is merely sugar for strtoll).
I stand corrected - the only time you return negative values is if
qemu_strtosz_suffx() populated an errno.
Reviewed-by: Eric Blake <ebl...@redhat.com>
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
signature.asc
Description: OpenPGP digital signature
