summaryrefslogtreecommitdiff
path: root/sys/netmpls
AgeCommit message (Collapse)Author
2011-04-04de-guttenberg our stack a bitHenning Brauer
we don't need 7 f***ing copies of the same code to do the protocol checksums (or not, depending on hw capabilities). claudio ok
2011-04-04no need to check the csum flags before calling mpls_do_cksum whichHenning Brauer
immediately checks exactly those flags, claudio ok
2011-03-31- use nitems(); no binary changeJasper Lievisse Adriaanse
ok claudio@
2011-01-21make mpls compile w/o errors when inet6 is not definedMike Belopuhov
diff from form@ who's too busy drinking vodka with pelmeni to commit this; ok claudio
2010-10-15Enable multipath support on the MPLS routing table. ldpd needs this andClaudio Jeker
the priority support that comes together with it. Currently no balancing will happen, only the first path will be used.
2010-10-07Handle IMPLNULL labels (#3) directly in the kernel like we do with theClaudio Jeker
explicit NULL labels. IMPLNULL labels should never ever be seen on the wire but we're realists. michele@ agrees
2010-09-13First shot at ICMP error handling inside an MPLS path. Currently onlyClaudio Jeker
TTL exceeded errors for IPv4 are handled. This makes traceroute through MPLS tunnels work (including RFC 4950 MPLS extension header). "best to get it in." deraadt@
2010-09-08Map the TTL when building the initial MPLS stack from the encapsulatedClaudio Jeker
packet. For IP and IPv6 the corresponding mapttl sysctl needs to be set. This is needed to make traceroute work over MPLS. OK michele@
2010-09-03Use MPLS_INKERNEL_LOOP_MAX instead of 16 when initializing mpls_inkloop.Claudio Jeker
2010-07-15Correct function in panic string. From Christophe FillotClaudio Jeker
2010-07-07Use correct function name in panic()Claudio Jeker
2010-07-07Remove duplicate mtod(). No need to do it twice since the mbuf is notClaudio Jeker
modified in between.
2010-06-09Fix the pop operation to make PHP work again. When popping the last labelClaudio Jeker
by a pop operation we need to forward the packet to the specified nexthop as is. This is done by calling the interface output routine directly. This is different to the local operation since that one injects the packets into ip_input() via netisr. OK michele
2010-06-02Kill the mpls enable sysctl knob. Since MPLS needs to be enabled explicitlyClaudio Jeker
on each interface there is no need for yet another knob. OK michele@
2010-05-31Add a comment about the explicit NULL label handling. We follow RFC 4182Claudio Jeker
which relaxes the position of these labels. After discussion with Christophe Fillot (cf (at) utc fr)
2010-05-31Do not use IF_ENQUEUE() but IF_INPUT_ENQUEUE() instead. The first makroClaudio Jeker
does no overflow checking and does not set the congestion flag. Protocol input queues (inet, inet6, ...) should always use IF_INPUT_ENQUEUE(). OK henning@
2010-05-28Rework the way we handle MPLS in the kernel. Instead of fumbling MPLS intoClaudio Jeker
ether_output() and later on other L2 output functions use a trick and over- load the ifp->if_output() function pointer on MPLS enabled interfaces to go through mpls_output() which will then call the link level output function. By setting IFXF_MPLS on an interface the output pointers are switched. This now allows to cleanup the MPLS input and output pathes and fix mpe(4) so that the MPLS code now actually works for both P and PE systems. Tested by myself and michele (A custom kernel with MPLS and mpe enabled is still needed).
2010-05-07Start cleaning up the mess called rtalloc*. Kill rtalloc2, make rtalloc1Claudio Jeker
accept flags for report and nocloning. Move the rtableid into struct route (with a minor twist for now) and make a few more codepathes rdomain aware. Appart from the pf.c and route.c bits the diff is mostly mechanical. More to come... OK michele, henning
2010-04-22Unbreak build of MPLS enabled kernel. proc.h is needed.Michele Marchetto
ok claudio@
2009-07-09Remove commented variable.Michele Marchetto
2009-04-29Split up functionalities between mpls_input and mpls_output.Michele Marchetto
The former is responsible to handle packets destined to localhost, the latter handles packet that have to be forwarded or the ones originated from localhost. "i like it" claudio@
2009-04-28Introduce a new dummy operation to easily handle packets that have to leaveMichele Marchetto
the MPLS cloud and join ipv4/ipv6 stack. ok claudio@ laurent@
2009-04-17Correctly handle packets that are leaving the MPLS cloud andMichele Marchetto
joining the ipv4/ipv6 stack. Get also rid of the dependency on mpe(4) for packets not belonging to VPN traffic. ok laurent@, ok and input claudio@
2009-02-03Use our own flags to set the operations and not RTF_PROTOX.Michele Marchetto
Switch the padding field into a MPLS one in rt_msghdr to store relevant informations. OK claudio@ laurent@
2009-01-29mpls_operation have to be 2 bytes long to storeMichele Marchetto
operations correctly ok claudio@ laurent@
2009-01-28Get rid of the ugly rtentry hack.Michele Marchetto
We can now act as edge node and allow ipv4 packets to enter a Label Switched Path and not just forwarding MPLS packets. OK claudio@
2009-01-08Fix sockaddr_mpls structure.Michele Marchetto
Now it contains just the label as it must be. This introduces a ugly hack in rtentry that will be removed as soon as possible. OK claudio@
2008-12-15Initial fixup of sockaddr_mpls struct.Michele Marchetto
Remove smpls_in_iface as labels must be per platform unique. Remove smpls_out_iface. It is redudant and can be inferred from nexthop. Move smpls_operation to rt_flags and temporary remove smpls_out_exp. ok claudio@
2008-11-06Fix cvs id tag.Michele Marchetto
2008-11-01Introduced Uniform Model for TTL handling.Michele Marchetto
MPLS TTL is mapped into network layer one as the packet exits the LSP. Just IPv4 support for now. Added the relevant sysctls to enable this behaviour. Input and OK claudio@
2008-10-28Added mpls_output() used to output mpls packets originating from local host.Michele Marchetto
Strictly similar to mpls_input(). Input and OK claudio@, OK laurent@
2008-10-14Now mpls_input() handles ipv4 and ipv6 explicit null labels.Michele Marchetto
Added mpe_input6 to manage also ipv6 packets insted of just ipv4 ones. OK claudio@ laurent@
2008-05-23Forgot this one. Wrap a printf() into MPLS_DEBUG sinceThordur I. Bjornsson
MPLS_LABEL_GET() is only available #ifdef MPLS_DEBUG
2008-05-23Make this compile again, by adding a struct proc * argumentThordur I. Bjornsson
to the usrreq's. (Both MPLS and NETATALK are disabled in GENERIC).
2008-05-10Make MPLS code compile without pseudo device mpe. OK norbyClaudio Jeker
2008-05-09Kill some useless makros.Claudio Jeker
2008-05-08Hook mpe(4) correctly into mpls so that it is possible to tunnel packets overClaudio Jeker
MPLS. Still a bit hackish but getting closer. hai norby@
2008-05-07add mpe_input() to be able to be fed packets that will reenter ip.Pierre-Yves Ritschard
no consumers yet, they should come soon. ok norby@
2008-05-07Make mpe a point-to-point interface.Pierre-Yves Ritschard
Make the mtu user definable. ok claudio@
2008-05-07Correctly initialize labels by byte-swapping and inserting BoS and TTL.Pierre-Yves Ritschard
While there bump MTU to 1500 and set the receiving interface before calling mpls_input. mpestart will eventually call mpls_output when it exists. ``yaaaaayyy'' and ok claudio@ and norby@
2008-05-06bring in the mpe interface - for ``MPLS Provider Edge'' - this is a workPierre-Yves Ritschard
in progress and some bits need to be cleaned up but will be in-tree for convenience. ok claudio@, norby@
2008-05-06Drop all broadcast and multicast packets.Esben Norby
ok claudio@
2008-05-06get rid of rtalloc, use rtalloc1 instead.Esben Norby
ok claudio@
2008-05-06fix debug output.Esben Norby
ok claudio@
2008-05-02Fix the label swithing and forwarding logic a bit. OK norby@Claudio Jeker
2008-05-02There is no need to endian convert the label after swapping it, all label opsClaudio Jeker
are done with network byte ordering now. OK norby@
2008-04-30Drop all MPLS packets when MPLS is not enabled.Esben Norby
Discussed with claudio@ dlg@ laurent@ Remember to enable MPLS vith sysctl(8). sysctl -w net.mpls.enable=1 ok claudio@ dlg@ laurent@
2008-04-24the softnet intr handlers check if the input queue has packets onDavid Gwynne
it by reading the queues head pointer. if that pointer is not null then it takes splnet and dequeues a packet for handling. this is bad because the ifqueue head is modified at splnet and the sofnet handlers read it without holding splnet. this removes that check of the head pointer and simply checks if the dequeue gave us a packet or not before proceeding. found while reading mpls code. discussed with norby@ and henning@ ok mcbride@ henning@
2008-04-23unbreak compilationg with option MPLS but without option MPLS_DEBUGDavid Gwynne
2008-04-23add missing prototypesEsben Norby
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-24 21:26:38 +0000