On 5/27/15, 1:08 PM, roopa wrote:
On 5/27/15, 12:59 PM, Robert Shearman wrote:
On 27/05/15 19:37, Roopa Prabhu wrote:
From: Roopa Prabhu <ro...@cumulusnetworks.com>
Kernel expects type RTN_UNICAST for mpls route/dels
Signed-off-by: Vivek Venkataraman <vi...@cumulusnetworks.com>
Signed-off-by: Roopa Prabhu <ro...@cumulusnetworks.com>
---
ip/iproute.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/ip/iproute.c b/ip/iproute.c
index 670a4c6..71c088b 100644
--- a/ip/iproute.c
+++ b/ip/iproute.c
@@ -803,6 +803,7 @@ static int iproute_modify(int cmd, unsigned
flags, int argc, char **argv)
int scope_ok = 0;
int table_ok = 0;
int raw = 0;
+ int type_ok = 0;
memset(&req, 0, sizeof(req));
@@ -1095,6 +1096,7 @@ static int iproute_modify(int cmd, unsigned
flags, int argc, char **argv)
rtnl_rtntype_a2n(&type, *argv) == 0) {
NEXT_ARG();
req.r.rtm_type = type;
+ type_ok = 1;
}
if (matches(*argv, "help") == 0)
@@ -1160,6 +1162,9 @@ static int iproute_modify(int cmd, unsigned
flags, int argc, char **argv)
}
}
+ if (!type_ok && req.r.rtm_family == AF_MPLS)
+ req.r.rtm_type = RTN_UNICAST;
+
if (req.r.rtm_family == AF_UNSPEC)
req.r.rtm_family = AF_INET;
There is this block of code near the start of iproute_modify that
sets req.r.rtm_type in the add/modify cases:
if (cmd != RTM_DELROUTE) {
req.r.rtm_protocol = RTPROT_BOOT;
req.r.rtm_scope = RT_SCOPE_UNIVERSE;
req.r.rtm_type = RTN_UNICAST;
}
How about doing similar for the mpls delete case? This would avoid
the need to track if the type has been set and would also make the
way rtm_type is set in the delete case as close as possible to that
in the add/modify cases.
sure that works too. There was already *_ok checks for the rest of the
attributes, ..so added it there.
v3 ...coming...
looking at the code again..now i remember why i have it this way. I will
have to add a check for family around
the code you point out above. And it some cases if the user has not
specified the family explicitly, we derive the msg family
in the while loop that parses the args...based on the other arguments
given by the user.
In the particular mpls case though, user explicitly specifies the family
and moving the patch to the code you point above should be ok.
But to be consistent with the rest of the code, it seems better to do
the check and set the defaults at the end after parsing all the args.
So, now i am inclined to keep the v2 patch as is...unless you have
strong reasons.
thanks.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
