> On Feb 24, 2017, at 07:36, Ian Lepore <i...@freebsd.org> wrote: > > On Fri, 2017-02-24 at 06:24 +0000, Ngie Cooper wrote: >> Author: ngie >> Date: Fri Feb 24 06:24:39 2017 >> New Revision: 314189 >> URL: https://svnweb.freebsd.org/changeset/base/314189 >> >> Log: >> Fix up NULL/'\0' uses and fix 2 derefs after NULL >> >> MFC after: 2 weeks >> Reported by: Coverity >> CID: 1018898, 1018899 >> Differential Revision: https://reviews.freebsd.org/D6142 >> Sponsored by: Dell EMC Isilon >> >> Modified: >> head/lib/libcam/scsi_cmdparse.c >> >> Modified: head/lib/libcam/scsi_cmdparse.c >> ===================================================================== >> ========= >> --- head/lib/libcam/scsi_cmdparse.c Fri Feb 24 01:39:12 2017 >> (r314188) >> +++ head/lib/libcam/scsi_cmdparse.c Fri Feb 24 06:24:39 2017 >> (r314189) >> @@ -128,14 +128,14 @@ do_buff_decode(u_int8_t *buff, size_t le >> *(va_arg(*ap, int *)) = (ARG); \ >> assigned++; \ >> } \ >> - field_name[0] = 0; \ >> + field_name[0] = '\0'; \ > > Is coverity seriously complaining about using 0 instead of '\0'? What > possible *real* benefit (not just shutting up coverity) does this churn > bring?
Warner made the same comment. It was actually a combination of issues brought up by Coverity and coccinelle (badzero.cocci). The forward-NULL CIDs noted were pointed out by Coverity. The NULL/‘\0’ use issues were pointed out by coccinelle. Thanks! -Ngie
signature.asc
Description: Message signed with OpenPGP using GPGMail