On Wed, 22 Nov 2023 07:23:36 +0000
Hemant Agrawal <hemant.agra...@nxp.com> wrote:
> > -----Original Message-----
> > From: Tyler Retzlaff <roret...@linux.microsoft.com>
> > Sent: Tuesday, November 21, 2023 10:49 PM
> > To: Stephen Hemminger <step...@networkplumber.org>
> > Cc: Hemant Agrawal <hemant.agra...@nxp.com>; dev@dpdk.org; Sachin
> > Saxena <sachin.sax...@nxp.com>; Gagandeep Singh <g.si...@nxp.com>
> > Subject: Re: [PATCH v4 10/10] common/dpaxx: remove zero length array
> > Importance: High
> >
> > On Tue, Nov 21, 2023 at 08:46:41AM -0800, Stephen Hemminger wrote:
> > > On Tue, 21 Nov 2023 10:49:59 +0000
> > > Hemant Agrawal <hemant.agra...@nxp.com> wrote:
> > >
> > > > > struct ipsec_encap_pdb {
> > > > > uint32_t options;
> > > > > @@ -350,7 +349,6 @@ struct ipsec_encap_pdb {
> > > > > };
> > > > > uint32_t spi;
> > > > > uint32_t ip_hdr_len;
> > > > > - uint8_t ip_hdr[0];
> > > > [Hemant] This should be replaced with
> > > > uint8_t ip_hdr[];
> > > > > };
> > > > >
> > >
> > > That won't work because the structure is embedded in another struct
> > > and then clang will correctly report an error.
> >
> > Hemant if the field is not referenced what's the harm in removing it?
>
> It is being used indirectly, e.g. dpaa_sec.h usages it for ip address
> description preparation for the hardware.
>
> struct ipsec_encap_pdb encap_pdb;
> union {
> struct ip ip4_hdr;
> struct rte_ipv6_hdr ip6_hdr;
> };
>
> We need to change it, however in the interim, I will like to add
> "-Wno-gnu-variable-sized-type-not-at-end" for both caam_jr and dpaa_sec to
> avoid this warning.
Ok, there are several ways to fix that. Driver specific compiler flags are the
one
that is least favored.
