1 $Id: BUGS.TODO,v 1.1.1.1 1999/08/08 23:30:57 itojun Exp $
4 THIS LIST IS FAR AWAY FROM BEING COMPLETE, so these are the few things
5 that came up at the right moment to be written down.
7 * Experimental kernel MFC (*,G) related:
8 If the (S,G) iif or oifs are different from the (*,G) or (*,*,RP)
9 iifs/oifs, the resp. (*,G) or (*,*,RP) will delete and disallow
10 creating (*,G) MFC. Only after all MRT (S,G) are deleted, the
11 corresponding (*,G) or (*,*,RP) will create (*,G) MFC.
13 * Experimental kernel MFC (*,G) related:
14 Right now when the MFC (*,G) total datarate is above the SPT switch
15 threshold, the (*,G) MFC will be deleted, and any further cache miss
16 will result in (S,G) MFC (the problem is that we must do (S,G)
17 monitoring for eventually high datagate sources). Only after all
18 (S,G) MFCs expire, the daemon's MRT will stop creating (S,G) MFCs
19 (i.e. the next cache miss will result in (*,G) kernel MFC).
20 A better selection should be applied to sort out the higher
21 datarate sources, and at the same time to have (*,G)MFC as well.
22 For example, create few (S,G), and after that create the (*,G). If some
23 of the created (S,G) MFC entries have very low datarate, delete them.
25 * Use NetBSD's definition for IPADDR (netinet/in.h):
27 #define __IPADDR(x) ((u_int32_t) htonl((u_int32_t)(x)))
29 #define __IPADDR(x) ((u_int32_t)(x))
33 * The (S,G)RPbit in the DR for the sender and the (S,G)SPT in the
34 downstream router won't timeout and will refresh each other even
35 if the sender is not active:
37 S--DR-----------------R1------------RP
41 * Check whether the kernel code sends CACHE_MISS and WRONG_IIF for
42 the LAN-scoped addresses
44 * If the RP for a group changes, the DR should cancel any PIM-register-stop
45 timers (XXX: not in the spec, but should be there)
47 * If a new interface is configured, include it automatically
49 * Don't create routing entries for local link scoped groups
51 * Implement adm. scoped filters
53 * Do precise check of the timer events to speed up the propagation of the
54 Cand-RP messages + Cand-BSR messages and the election of the BSR.
56 * Fix the bug for messing up the things when the receiver is on the
57 same host as the RP for the multicast group (probably was fixed with alpha6,
58 because I cannot reproduce it anymore)
60 * Do more precise error check for the received PIM messages. In most cases,
61 the whole message must be parsed completely before starting processing it.
63 * Clean up the debugging messages.
65 * Use Patricia tree to search the routing table
66 (There is a nice paper in Sigcomm '97 about fast routing tables
67 implementation, so need to check it as well)
69 * Do switch back to the Shared Tree by timing out the SPT if the rate
70 is too low (not in the spec, but Ahmed pointed out some complications if
73 * Change all countdown timers to events timeout (callout.c)
74 (The current implementation is very unefficient if the routing table becomes
77 * Send immediately Join/Prune, instead of relying of Join/Prune timer = 0
79 * Fix the code allowing interface UP/DOWN without restarting pimd.
81 * Do more testings for SPT switch, Join/Prune, asserts, etc...
83 * Test the (*,*,RP) code (need PIM/DVMRP border router to do so)
85 * Test the RSRR (RSVP support) code
87 * Send Initial_Reply RSRR message if the interfaces detected by pimd change
91 ===TODO by function name===
93 * accept_group_report():
94 - add a leaf if DR or forwarder (currently only if DR)???
95 * accept_leave_message():
96 - send immediately PIM prune message if the last member has left
100 - use a combination of time and hostid to initialize the random generator.
102 - check the implementation
106 - IF THE BORDER BIT IS SET, THEN FORWARD THE WHOLE PACKET FROM USER SPACE
107 AND AT THE SAME TIME IGNORE ANY CACHE_MISS SIGNALS FROM THE KERNEL.
109 - REGISTER_STOP rate limiting
112 * process_cache_miss()
113 - use negative cache.
116 * add_rp_grp_entry():
117 - FIX THE BUG when adding an RP for different prefix requires remapping
119 (Intentionally left, waiting to come up with an idea how to implement
120 it simple and efficient. If you configure all RPs to advertise the
121 same prefix, the bug won't "show up")
123 ================DONE=====================
125 * When receive PIM_REGISTER, check whether I am the chosen RP