Take 'include/uapi/linux/snmp.h' from current kernel v3.13
(commit d8ec26d7f8287f5788a494f56e8814210f0e64be).
The header file added new values for IPSTATS_MIB_* and ICMP6_MIB_*, but
more importantly, the kernel broke user space API by reordering enum values in
IPSTATS_MIB_*. Add a workaround when parsing IFLA_PROTINFO trying to
be compatible with both older and newer kernels.
Note that this workaround might fail for some specific kernel versions by
assuming the old enum value mapping, although the kernel version already
contains the API change. In this case rtnl_link_get_stat() mixes up
values.
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Thomas Haller <thaller@redhat.com>
Older kernel version might have fewer values defined, so they would send
netlink messages that got rejected. Only check that at least one value
got sent.
This is especially grave as libnl uses an internal copy of the
kernel header files. Thus not only it is bound to the installed kernel
headers but to the libnl internal header copies that might easily be out
of sync with the kernel.
This affects IFLA_PROTINFO, inet6_parse_protinfo():
- tb[IFLA_INET6_CONF], expecting DEVCONF_MAX
- tb[IFLA_INET6_STATS], expecting __IPSTATS_MIB_MAX
- tb[IFLA_INET6_ICMP6STATS], expecting __ICMP6_MIB_MAX
and IFLA_AF_SPEC, inet_parse_af():
- tb[IFLA_INET_CONF], expecting IPV4_DEVCONF_MAX
https://bugzilla.redhat.com/show_bug.cgi?id=1062533
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Thomas Haller <thaller@redhat.com>
On some architectures, uint64_t is defined as:
typedef unsigned long long int __u64;
on another architectures as:
typedef unsigned long int __u64;
So, according to man 3 printf,
uint64_t should be printed as "%llu" on some architectures, and as "%lu" on another. The same for scanf.
To eliminate that challenge, there is inttypes.h, in which appropriate constants
are defined for current architecture.
32-bit types (and even 16 and 8 bit types) should be printed using such constants if
printed variable defined as uint_XXXt or intXXXt type. But in reality 32-bit and less
types does not gain run-time error (except in scanf), because they pushed to stack as
32-bit values at least. So, I decide not to fix that.
This patch fixes an unaligned access for IPv6. On systems with strict alignment requirements, the unaligned access will either result in garbage data or a crash.
This feature isn't upstream yet. It's required to test a patch in
my local tree.
Makes the link parser understand IFLA_AF_SPEC and call the address
family specific parser.
