[PATCH 1/2] rtnl_route: decode rtmsg netlink attributes
JingPiao Chen
chenjingpiao at gmail.com
Wed Aug 23 01:57:37 UTC 2017
On Wed, Aug 23, 2017 at 04:30:28AM +0300, Dmitry V. Levin wrote:
> On Wed, Aug 23, 2017 at 08:45:18AM +0800, JingPiao Chen wrote:
> > * configure.ac (AC_CHECK_TYPES): Check for
> > rta_mfc_stats and rtvia structures in <linux/rtnetlink.h>.
> > * nlattr.c (decode_nla_ifindex): New function.
> > * nlattr.h (decode_nla_ifindex): New prototype.
> > * rtnl_route.c (decode_route_addr, decode_rta_metrics,
> > decode_rta_multipath, decode_rta_cacheinfo,
> > decode_rta_mfc_stats, decode_rtvia,
> > decode_rta_encap_type): New functions.
> > (rta_metrics_nla_decoders, rtmsg_nla_decoders): New arrays.
> > (decode_rtmsg): Use rtmsg_nla_decoders.
> > * xlat/lwtunnel_encap_types.in: New file.
> > * xlat/route_nexthop_flags.in: Likewise.
> > * xlat/rtnl_rta_metrics_attrs.in: Likewise.
> [...]
> > +static bool
> > +decode_rtvia(struct tcb *const tcp,
> > + const kernel_ulong_t addr,
> > + const unsigned int len,
> > + const void *const opaque_data)
> > +{
> > +#ifdef HAVE_STRUCT_RTVIA
> > + struct rtvia via;
> > +
> > + if (len < sizeof(via))
> > + return false;
> > + else if (!umove_or_printaddr(tcp, addr, &via)) {
> > + PRINT_FIELD_XVAL("{", via, rtvia_family, addrfams, "AF_???");
> > +
> > + if (len > sizeof(via)) {
> > + tprints(", rtvia_addr=");
> > + printstr_ex(tcp, addr + sizeof(via),
> > + len - sizeof(via), QUOTE_FORCE_HEX);
>
> Looks like decode_inet_addr would be more appropriate here than printstr_ex.
>
> We might end up with renaming decode_inet_addr to something more generic
> and extending it to handle AF_PACKET in addition to AF_INET and AF_INET6.
bool
decode_proto_address(struct tcb *const tcp,
const kernel_ulong_t addr,
const unsigned int len,
const int family,
const char *const var_name)
{
switch (family) {
case AF_INET:
case AF_INET6:
return decode_inet_addr(tcp, addr, len, family, var_name);
case AF_PACKET:
....
}
if (var_name)
tprintf("%s=", var_name);
printstr_ex(tcp, addr, len, QUOTE_FORCE_HEX);
return false;
}
How about this?
--
JingPiao Chen
More information about the Strace-devel
mailing list