From ashishkarpe at gmail.com Tue Oct 3 05:39:01 2006 From: ashishkarpe at gmail.com (Ashish Karpe) Date: Tue, 3 Oct 2006 18:09:01 +0530 Subject: [Xorp-hackers] Xorp Setup help !!! Message-ID: hi all, if we want to configure PIM-SM using Xorp what kind of minimum setup do we require to just check how it works in xorp. Can it be done using Virtual Network or physical network is necessary ? If yes which software is commonly used for this propose. please reply, Ashish From pavlin at icir.org Tue Oct 3 09:43:04 2006 From: pavlin at icir.org (Pavlin Radoslavov) Date: Tue, 03 Oct 2006 09:43:04 -0700 Subject: [Xorp-hackers] Xorp Setup help !!! In-Reply-To: Message from "Ashish Karpe" of "Tue, 03 Oct 2006 18:09:01 +0530." Message-ID: <200610031643.k93Gh4Xk079389@possum.icir.org> Ashish Karpe wrote: > hi all, > if we want to configure PIM-SM using Xorp what kind of minimum > setup do we require to just check how it works in xorp. Can it be done > using Virtual Network or physical network is necessary ? If yes which > software is commonly used for this propose. At minimum, you need one XORP router that is configured as a static RP and has a directly connected sender and a receiver (on different interfaces). The XORP configuration should have the following sections/protocols: interfaces, mfea4, igmp, pimsm4, fib2mrib. If you don't have a machine with two physical network interfaces you could use virtual interfaces for that purpose. You could use OpenVPN to configure them. However, the OpenVPN virtual interfaces must be configured before starting XORP (in the future XORP should have support to do that on its own), and make sure that inside the XORP "interfaces" section each such interface is configured like: interface tun0 { default-system-config } Also, if you have 2+ PIM-SM routers connected with VPNs, make sure that the RPF (Reverse Path Forwarding) information (as seen by PIM-SM) is using the VPNs. For that purpose you need either: (a) add unicast static routes that use the VPNs as appropriate (or just run a dynamic protocol like RIP or OSPF), or (b) add "mrib-route" static routes inside the XORP configuration. Hope that helps, Pavlin From santhosh at ku.edu Thu Oct 5 16:40:47 2006 From: santhosh at ku.edu (Santhosh Sundararaman) Date: Thu, 05 Oct 2006 18:40:47 -0500 Subject: [Xorp-hackers] spt.cc class definitions Message-ID: <452597FF.3030202@ku.edu> Hi All, I was going through the code for the SPT computation and realized that the definitions for the SPT class' member functions were inside a #if 0 -- #endif and there was also a comment stating that this file was obsolete. But ospf which uses SPT seems to be working fine. If spt.cc is obsoleted, where is the definition for the spt functions declared in spt.hh. Am i missing something?? Someone please clarify this issue. Thanks Santhosh From santhosh at ku.edu Thu Oct 5 17:17:15 2006 From: santhosh at ku.edu (Santhosh Sundararaman) Date: Thu, 05 Oct 2006 19:17:15 -0500 Subject: [Xorp-hackers] please disregard my previous mail - spt.cc Message-ID: <4525A08B.3010600@ku.edu> Hi All, I just realized that the definitions for the spt class was in spt.hh file. Im sorry for the hasty mail i sent out previously. Thanks Santhosh From atanu at ICSI.Berkeley.EDU Thu Oct 5 18:36:05 2006 From: atanu at ICSI.Berkeley.EDU (Atanu Ghosh) Date: Thu, 05 Oct 2006 18:36:05 -0700 Subject: [Xorp-hackers] spt.cc class definitions In-Reply-To: Message from Santhosh Sundararaman of "Thu, 05 Oct 2006 18:40:47 CDT." <452597FF.3030202@ku.edu> Message-ID: <98024.1160098565@tigger.icir.org> Hi, The definitions are all in spt.hh. Atanu. >>>>> "Santhosh" == Santhosh Sundararaman writes: Santhosh> Hi All, I was going through the code for the SPT Santhosh> computation and realized that the definitions for the SPT Santhosh> class' member functions were inside a #if 0 -- #endif and Santhosh> there was also a comment stating that this file was Santhosh> obsolete. But ospf which uses SPT seems to be working Santhosh> fine. If spt.cc is obsoleted, where is the definition for Santhosh> the spt functions declared in spt.hh. Am i missing Santhosh> something?? Someone please clarify this issue. Santhosh> Thanks Santhosh Santhosh> _______________________________________________ Santhosh> Xorp-hackers mailing list Xorp-hackers at icir.org Santhosh> http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-hackers From marat at vyatta.com Mon Oct 9 15:07:47 2006 From: marat at vyatta.com (Marat Nepomnyashy) Date: Mon, 9 Oct 2006 15:07:47 -0700 (PDT) Subject: [Xorp-hackers] Minor patch for slave_conf_tree Message-ID: <10293726.1321160431667999.JavaMail.root@mail.vyatta.com> Hi Pavlin, Back in 2005 I sent you guys a patch to split out the parent class XorpShellBase from XorpShell. It looks like in that initial patch I forgot to modify the methods SlaveConfigTree::save_phase4(...) and SlaveConfigTree::save_phase5(...) to use a pointer to XorpShellBase instead of XorpShell. It is a really minor change, please find the patch file attached. Thanks, Marat -------------- next part -------------- A non-text attachment was scrubbed... Name: patch-xsb-01 Type: application/octet-stream Size: 1353 bytes Desc: not available Url : http://mailman.ICSI.Berkeley.EDU/pipermail/xorp-hackers/attachments/20061009/888db688/attachment.obj From pavlin at icir.org Tue Oct 10 14:41:14 2006 From: pavlin at icir.org (Pavlin Radoslavov) Date: Tue, 10 Oct 2006 14:41:14 -0700 Subject: [Xorp-hackers] Minor patch for slave_conf_tree In-Reply-To: Message from Marat Nepomnyashy of "Mon, 09 Oct 2006 15:07:47 PDT." <10293726.1321160431667999.JavaMail.root@mail.vyatta.com> Message-ID: <200610102141.k9ALfEFA010376@possum.icir.org> > Back in 2005 I sent you guys a patch to split out the parent class > XorpShellBase from XorpShell. It looks like in that initial patch > I forgot to modify the methods SlaveConfigTree::save_phase4(...) > and SlaveConfigTree::save_phase5(...) to use a pointer to > XorpShellBase instead of XorpShell. It is a really minor change, > please find the patch file attached. Patch committed to CVS. Thanks, Pavlin From ashishkarpe at gmail.com Thu Oct 12 23:21:59 2006 From: ashishkarpe at gmail.com (Ashish Karpe) Date: Fri, 13 Oct 2006 11:51:59 +0530 Subject: [Xorp-hackers] PIM security !! Message-ID: Hi all , In PIM-SM implementation of Xorp what security mechanism has been implemented i.e has been IPsec used in PIM-SM Implementation of Xorp ?? thanks, ashish From ashishkarpe at gmail.com Fri Oct 13 11:22:17 2006 From: ashishkarpe at gmail.com (Ashish Karpe) Date: Fri, 13 Oct 2006 23:52:17 +0530 Subject: [Xorp-hackers] IPSEC ??? Message-ID: Hi all , has any implementation of any protocol supported in zebra/quagga uses/using IPsec mechanism ?? please reply !!! thanks, ashish From pavlin at icir.org Fri Oct 13 11:28:08 2006 From: pavlin at icir.org (Pavlin Radoslavov) Date: Fri, 13 Oct 2006 11:28:08 -0700 Subject: [Xorp-hackers] PIM security !! In-Reply-To: Message from "Ashish Karpe" of "Fri, 13 Oct 2006 11:51:59 +0530." Message-ID: <200610131828.k9DIS8h4063287@possum.icir.org> > Hi all , > In PIM-SM implementation of Xorp what security mechanism has > been implemented i.e has been IPsec used in PIM-SM Implementation of > Xorp ?? Currently the PIM-SM implementation doesn't have a security mechanism. To be precise, it has limited security checks as described in the PIM-SM spec; e.g., the link-local control messages must fit certain profile, a PIM Hello message must have been accepted before accepting other link-local control messages, etc. If IPsec is to be used, currently it has to be configured outside of XORP. Regards, Pavlin From pavlin at icir.org Fri Oct 13 11:34:28 2006 From: pavlin at icir.org (Pavlin Radoslavov) Date: Fri, 13 Oct 2006 11:34:28 -0700 Subject: [Xorp-hackers] IPSEC ??? In-Reply-To: Message from "Ashish Karpe" of "Fri, 13 Oct 2006 23:52:17 +0530." Message-ID: <200610131834.k9DIYS8s063441@possum.icir.org> > has any implementation of any protocol supported in zebra/quagga ~~~~~~~~~~~~ Wrong mailing list, I presume you mean XORP ;) > uses/using IPsec mechanism ?? please reply !!! No, currently the IPsec mechanism hasn't been integrated in XORP. Any IPsec related configuration has to be done outside XORP. Regards, Pavlin From ashishkarpe at gmail.com Fri Oct 13 14:40:14 2006 From: ashishkarpe at gmail.com (Ashish Karpe) Date: Sat, 14 Oct 2006 03:10:14 +0530 Subject: [Xorp-hackers] IPSEC ??? In-Reply-To: <200610131834.k9DIYS8s063441@possum.icir.org> References: <200610131834.k9DIYS8s063441@possum.icir.org> Message-ID: On 10/14/06, Pavlin Radoslavov wrote: > > has any implementation of any protocol supported in zebra/quagga > ~~~~~~~~~~~~ > > Wrong mailing list, I presume you mean XORP ; ) sorry :) for mistake > > > uses/using IPsec mechanism ?? please reply !!! > > No, currently the IPsec mechanism hasn't been integrated in XORP. > Any IPsec related configuration has to be done outside XORP. So what will be beneficial, implement IPsec within protocol or IPsec mechanism can bee integrated in Xorp ?? And also what should bee considered while doing this i.e Integrate openswan like IPsec tools witin Xorp or Implement our own IPsec mechanism ?? > > Regards, > Pavlin > From pavlin at icir.org Fri Oct 13 15:25:44 2006 From: pavlin at icir.org (Pavlin Radoslavov) Date: Fri, 13 Oct 2006 15:25:44 -0700 Subject: [Xorp-hackers] IPSEC ??? In-Reply-To: Message from "Ashish Karpe" of "Sat, 14 Oct 2006 03:10:14 +0530." Message-ID: <200610132225.k9DMPi3F007852@possum.icir.org> > So what will be beneficial, implement IPsec within protocol or > IPsec mechanism can bee integrated in Xorp ?? > > And also what should bee considered while doing this i.e Integrate > openswan like IPsec tools witin Xorp or Implement our own IPsec > mechanism ?? I would say it is better to integrate existing implementations. Though, I don't have enough experience with IPsec, so you should make your final decision based on opinion from folks who are quite familiar with it. Pavlin From kristian at spritelink.se Sat Oct 14 03:32:56 2006 From: kristian at spritelink.se (Kristian Larsson) Date: Sat, 14 Oct 2006 12:32:56 +0200 Subject: [Xorp-hackers] Anoncvs.xorp.org Message-ID: <20061014103255.GF12103@spritelink.se> Is it just me or is anoncvs.xorp.org incredibly slow? I'm checking out the xorp repository as I write this email, it is not done but has been at it for over ten minutes. I'd be delighted to help out with a european mirror of the cvs :) Regards, Kristian. -- Kristian Larsson KLL-RIPE Network Engineer Net at Once [AS35706] +46 704 910401 kristian at spritelink.se From elcinturapartida at yahoo.es Mon Oct 16 04:30:27 2006 From: elcinturapartida at yahoo.es (David H. Guerrero) Date: Mon, 16 Oct 2006 13:30:27 +0200 (CEST) Subject: [Xorp-hackers] Install without CD-R or CD-RW Message-ID: <20061016113027.95608.qmail@web26010.mail.ukl.yahoo.com> Hello, how do I install CD Live XORP 1.3 if I don't have a CD-R or CD-RW drive (or "I have no way to burn this image that I just downloaded.")? I wonder if it?s possible. Thanks in advance. ______________________________________________ LLama Gratis a cualquier PC del Mundo. Llamadas a fijos y m?viles desde 1 c?ntimo por minuto. http://es.voice.yahoo.com From pavlin at icir.org Mon Oct 16 10:21:51 2006 From: pavlin at icir.org (Pavlin Radoslavov) Date: Mon, 16 Oct 2006 10:21:51 -0700 Subject: [Xorp-hackers] [Xorp-users] Install without CD-R or CD-RW In-Reply-To: Message from Kristian Larsson of "Mon, 16 Oct 2006 14:27:04 +0200." <20061016122703.GK12103@spritelink.se> Message-ID: <200610161721.k9GHLptX040192@possum.icir.org> [Note: CC to xorp-hackers because the original question was posted to that list as well] Kristian Larsson wrote: > On Mon, Oct 16, 2006 at 01:45:04PM +0200, David H. Guerrero wrote: > > Hello, how do I install CD Live XORP 1.3 if I don't have a CD-R or CD-RW drive > > (or "I have no way to burn this image that I just downloaded.")? > You could install FreeBSD or Linux or some other > system first by PXE booting or whatever and then > just download XORP and install on top of that. > > XORP is merely a routing application suite and can > be run on top of several different operating > systems. Perhaps you already have a computer > installed and ready to go. To clarify few points: * The XORP LiveCD doesn't have the option to install the software on the hard disk. I.e., you can use the LiveCD to try the software or just to run a XORP router on a PC that doesn't have a hard disk. In both cases you need a CD drive. * If you have a floppy on your PC you could use it to install first FreeBSD. * If you have already an OS on your hard disk, you could download the source code, compile it, and then run "gmake install" as a root. This will install the software in the "/usr/local/xorp" directory. Though, first you should check file "xorp/BUILD_NOTES" that your OS is supported. * One possible solution to try the LiveCD without a CD drive is to create a VMware instance for FreeBSD or Linux, but don't install the OS itself. Then edit that instance and point the VMware drive for that instance to the LiveCD ISO image. When you boot the VMware instance you will actually boot the LiveCD ISO image. Regards, Pavlin > Another option would be to try out Vyatta which is > Linux tightly coupled with XORP and a bunch of > other programs to provide more of a uniform router > feeling. You can take a look at www.vyatta.com, > though I don't think there is an installation > option other that CD for Vyatta either. > > Regards, > Kristian. > > -- > Kristian Larsson KLL-RIPE > Network Engineer Net at Once [AS35706] > +46 704 910401 kristian at spritelink.se > > _______________________________________________ > Xorp-users mailing list > Xorp-users at xorp.org > http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-users From atanu at ICSI.Berkeley.EDU Mon Oct 16 12:44:14 2006 From: atanu at ICSI.Berkeley.EDU (Atanu Ghosh) Date: Mon, 16 Oct 2006 12:44:14 -0700 Subject: [Xorp-hackers] Anoncvs.xorp.org In-Reply-To: Message from Kristian Larsson of "Sat, 14 Oct 2006 12:32:56 +0200." <20061014103255.GF12103@spritelink.se> Message-ID: <19938.1161027854@tigger.icir.org> Hi, If you are willing to run an anonymous cvs server that would be great. Atanu. >>>>> "Kristian" == Kristian Larsson writes: Kristian> Is it just me or is anoncvs.xorp.org incredibly slow? Kristian> I'm checking out the xorp repository as I write this Kristian> email, it is not done but has been at it for over ten Kristian> minutes. Kristian> I'd be delighted to help out with a european mirror of the Kristian> cvs :) Kristian> Regards, Kristian. Kristian> -- Kristian Larsson KLL-RIPE Network Engineer Net at Once Kristian> [AS35706] +46 704 910401 kristian at spritelink.se Kristian> _______________________________________________ Kristian> Xorp-hackers mailing list Xorp-hackers at icir.org Kristian> http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-hackers From kristian at spritelink.se Mon Oct 16 13:07:47 2006 From: kristian at spritelink.se (Kristian Larsson) Date: Mon, 16 Oct 2006 22:07:47 +0200 Subject: [Xorp-hackers] Anoncvs.xorp.org In-Reply-To: <19938.1161027854@tigger.icir.org> References: <20061014103255.GF12103@spritelink.se> <19938.1161027854@tigger.icir.org> Message-ID: <20061016200747.GL12103@spritelink.se> On Mon, Oct 16, 2006 at 12:44:14PM -0700, Atanu Ghosh wrote: > Hi, > > If you are willing to run an anonymous cvs server that would be great. How do we proceed from here? Should I run rsync from one of your servers or will you be pushing out updates to me? Kristian. -- Kristian Larsson KLL-RIPE Network Engineer Net at Once [AS35706] +46 704 910401 kristian at spritelink.se From elcinturapartida at yahoo.es Mon Oct 16 13:20:45 2006 From: elcinturapartida at yahoo.es (David H. Guerrero) Date: Mon, 16 Oct 2006 22:20:45 +0200 (CEST) Subject: [Xorp-hackers] [Xorp-users] Install without CD-R or CD-RW In-Reply-To: <200610161721.k9GHLptX040192@possum.icir.org> Message-ID: <20061016202045.43892.qmail@web26003.mail.ukl.yahoo.com> --- Pavlin Radoslavov escribi?: Kristian, Pavlin thanks you for your responses. > [Note: CC to xorp-hackers because the original question was posted > to that list as well] > > Kristian Larsson wrote: > > > On Mon, Oct 16, 2006 at 01:45:04PM +0200, David H. Guerrero wrote: > > > Hello, how do I install CD Live XORP 1.3 if I don't have a CD-R or CD-RW > drive > > > (or "I have no way to burn this image that I just downloaded.")? > > You could install FreeBSD or Linux or some other > > system first by PXE booting or whatever and then > > just download XORP and install on top of that. I didn?t know that option. I had been explore with grub and itn't possible. > > > > XORP is merely a routing application suite and can > > be run on top of several different operating > > systems. Perhaps you already have a computer > > installed and ready to go. > > To clarify few points: > > * The XORP LiveCD doesn't have the option to install the software > on the hard disk. I.e., you can use the LiveCD to try the > software or just to run a XORP router on a PC that doesn't have a > hard disk. In both cases you need a CD drive. > > * If you have a floppy on your PC you could use it to install first > FreeBSD. > > * If you have already an OS on your hard disk, you could download > the source code, compile it, and then run "gmake install" as a > root. This will install the software in the "/usr/local/xorp" > directory. > Though, first you should check file "xorp/BUILD_NOTES" that your > OS is supported. > I have an old laptop (AMD 475 Mhz, 64Mb and 4Gb) with FreeBSD. I can build from source but it would be very slow =). There isn?t binary distribution of XORP in FreeBSD. > * One possible solution to try the LiveCD without a CD drive is to > create a VMware instance for FreeBSD or Linux, but don't install > the OS itself. Then edit that instance and point the VMware > drive for that instance to the LiveCD ISO image. > When you boot the VMware instance you will actually boot the > LiveCD ISO image. I didn?t know that option. > > Regards, > Pavlin > > > > Another option would be to try out Vyatta which is > > Linux tightly coupled with XORP and a bunch of > > other programs to provide more of a uniform router > > feeling. You can take a look at www.vyatta.com, > > though I don't think there is an installation > > option other that CD for Vyatta either. > > I had installed Vyatta on the hard disk =). > > Regards, > > Kristian. > > > > -- > > Kristian Larsson KLL-RIPE > > Network Engineer Net at Once [AS35706] > > +46 704 910401 kristian at spritelink.se > > > > _______________________________________________ > > Xorp-users mailing list > > Xorp-users at xorp.org > > http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-users > Regards, David ______________________________________________ LLama Gratis a cualquier PC del Mundo. Llamadas a fijos y m?viles desde 1 c?ntimo por minuto. http://es.voice.yahoo.com From atanu at ICSI.Berkeley.EDU Tue Oct 17 16:01:03 2006 From: atanu at ICSI.Berkeley.EDU (Atanu Ghosh) Date: Tue, 17 Oct 2006 16:01:03 -0700 Subject: [Xorp-hackers] XORP_SA_06:01.ospf: An LSA with invalid length will crash OSPFv2 Message-ID: <45609.1161126063@tigger.icir.org> ============================================================================= XORP_SA_06:01.ospf Security Advisory The XORP Project Topic: An LSA with invalid length will crash OSPFv2 Module: OSPF Announced: 2006-10-17 Credits: http://www.musecurity.com/ Affects: XORP 1.2 and XORP 1.3 Releases Corrected: 2006-10-16 06:50:04 UTC (Release 1.4-WIP) I. Background OSPFv2 is a link-state routing protocol defined in RFC 2328, implemented by the XORP project. II. Problem Description The OSPF protocol carries link state information in Link State Advertisements (LSAs). One or more LSAs can be carried in a Link State Update Packet. Each LSA has its own length field and checksum amongst other fields. One of the first checks made when processing an LSA is to verify the checksum. The checksum verification routine takes into account the LSA length field. If the length field has certain invalid values, then OSPF might crash. III. Impact An attacker sending specially crafted packets with certain invalid LSA length value will be able to terminate the XORP OSPF process. It should be noted that the attacker does not need to be on the same network segment as the XORP router. IV. Workaround One possible workaround is to filter all external IP packets with protocol number 89 (OSPF) at the border router. V. Solution Apply the relevant patch to your XORP system and restart OSPF. 1) To patch your present system: [XORP 1.2] # wget http://www.xorp.org/patches/SA-06:01/xorp_sa_06:01.ospf_1.2.patch [XORP 1.3] # wget http://www.xorp.org/patches/SA-06:01/xorp_sa_06:01.ospf_1.3.patch 2) Execute the following commands (only the last one has to be as root): # cd xorp # patch -p0 < /path/to/patch # gmake # cd ospf # gmake install 3) Restart OSPFv2 a) Save the current configuration to a file. # xorpsh Xorp> configure XORP# save /tmp/xorp.boot b) Delete ospf4 from the configuration and commit. OSPFv2 should no longer be running. XORP# delete protocols ospf4 XORP# commit c) Reload the saved configuration, which will restart OSPFv2 XORP# load /tmp/xorp.boot VI. Correction details The following list contains the revision numbers of each file that was corrected in XORP. Branch Revision Path ------------------------------------------------------------------------- HEAD xorp/ospf/lsa.cc 1.72 ------------------------------------------------------------------------- VII. References The latest revision of this advisory is available at: http://www.xorp.org/advisories/XORP_SA_06:01.ospf.txt ============================================================================= From sureshkannan at gmail.com Sat Oct 28 01:38:43 2006 From: sureshkannan at gmail.com (Suresh Kannan.D) Date: Sat, 28 Oct 2006 14:08:43 +0530 Subject: [Xorp-hackers] PIM MRE deletion pending In-Reply-To: <84f679e0610280136g3c11ea64yf66e4e9083057253@mail.gmail.com> References: <84f679e0610280136g3c11ea64yf66e4e9083057253@mail.gmail.com> Message-ID: <84f679e0610280138g5366d6fanfb366bddf0ae5a52@mail.gmail.com> Hi all, What is advantage of marking a pim mre for pending deletion rather than immediate delete?. Please let me know is there reason any behind it apart from saying design.. ---code snip..--------------- void PimMrt::add_task_delete_pim_mre(PimMre *pim_mre) { PimMreTask *pim_mre_task = NULL; PimMreTrackState::input_state_t input_state = PimMreTrackState::INPUT_STATE_MAX; if (pim_mre->is_task_delete_pending()) { // The entry is already pending deletion. // Shoudn't happen, but just in case... return; } ... ... } ---code snip..--------------- Thanks, Regards, Suresh kannan. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.ICSI.Berkeley.EDU/pipermail/xorp-hackers/attachments/20061028/80040100/attachment.html From pavlin at icir.org Sat Oct 28 15:29:29 2006 From: pavlin at icir.org (Pavlin Radoslavov) Date: Sat, 28 Oct 2006 15:29:29 -0700 Subject: [Xorp-hackers] PIM MRE deletion pending In-Reply-To: Message from "Suresh Kannan.D" of "Sat, 28 Oct 2006 14:08:43 +0530." <84f679e0610280138g5366d6fanfb366bddf0ae5a52@mail.gmail.com> Message-ID: <200610282229.k9SMTTxA043467@possum.icir.org> > What is advantage of marking a pim mre for pending deletion rather than > immediate delete?. Please let me know is there reason any behind it apart > from saying design.. > > ---code snip..--------------- > void > PimMrt::add_task_delete_pim_mre(PimMre *pim_mre) > { > PimMreTask *pim_mre_task = NULL; > PimMreTrackState::input_state_t input_state = > PimMreTrackState::INPUT_STATE_MAX; > > if (pim_mre->is_task_delete_pending()) { > // The entry is already pending deletion. > // Shoudn't happen, but just in case... > return; > } > ... > ... > } > ---code snip..--------------- The reason for this is design-related. PIM-SM uses the so-called dependency tracking mechanism. E.g., an event such as a MRIB update could trigger a number of operations (tasks) that need to be performed on a number of multicast routing entries. Some of those operations might trigger other events which translate in more operations (tasks), and so on. Before you ask, no the dependency tracking does not result in loops :) The operations are scheduled as a list of tasks, where each task is atomic and may contain a number of operations that need to be performed on a routing entry. To simplify the code and for optimization purpose, we lookup the appropriate multicast routing entry at the beginning of a task, and perform all operations until the end of the task. Hence, if one of the intermediate operations attempts to delete an entry, we mark the entry as a candidate for deletion, and continue the processing, so the rest of the operations and tasks can proceed. To be more accurate, we actually schedule a new task "delete the entry" which goes to the end of the list with tasks. This "delete the entry" task, when scheduled, will delete the routing entry or will be a no-op if an earlier task marked the entry as being in-use. Hope that helps, Pavlin From sureshkannan at gmail.com Tue Oct 31 05:46:01 2006 From: sureshkannan at gmail.com (Suresh Kannan.D) Date: Tue, 31 Oct 2006 19:16:01 +0530 Subject: [Xorp-hackers] PIM MRE deletion pending In-Reply-To: <200610282229.k9SMTTxA043467@possum.icir.org> References: <84f679e0610280138g5366d6fanfb366bddf0ae5a52@mail.gmail.com> <200610282229.k9SMTTxA043467@possum.icir.org> Message-ID: <84f679e0610310546r10cb22f9k32f64efb961b88a7@mail.gmail.com> Hi Pavlin, Thanks for your inputs. It makes few things easier and few places makes issues..But anyhow I thought another point that, by having pending deletion, we can prevent from cache miss due to PRUNE-PENDING state of upstream neighbour. Thanks, Regards, Suresh kannan. On 10/29/06, Pavlin Radoslavov wrote: > > > What is advantage of marking a pim mre for pending deletion rather than > > immediate delete?. Please let me know is there reason any behind it > apart > > from saying design.. > > > > ---code snip..--------------- > > void > > PimMrt::add_task_delete_pim_mre(PimMre *pim_mre) > > { > > PimMreTask *pim_mre_task = NULL; > > PimMreTrackState::input_state_t input_state = > > PimMreTrackState::INPUT_STATE_MAX; > > > > if (pim_mre->is_task_delete_pending()) { > > // The entry is already pending deletion. > > // Shoudn't happen, but just in case... > > return; > > } > > ... > > ... > > } > > ---code snip..--------------- > > The reason for this is design-related. > > PIM-SM uses the so-called dependency tracking mechanism. > E.g., an event such as a MRIB update could trigger a number of > operations (tasks) that need to be performed on a number of > multicast routing entries. Some of those operations might trigger > other events which translate in more operations (tasks), and so on. > Before you ask, no the dependency tracking does not result in loops :) > > The operations are scheduled as a list of tasks, where each task is > atomic and may contain a number of operations that need to be > performed on a routing entry. > To simplify the code and for optimization purpose, we lookup the > appropriate multicast routing entry at the beginning of a task, and > perform all operations until the end of the task. Hence, if one of > the intermediate operations attempts to delete an entry, we mark the > entry as a candidate for deletion, and continue the processing, so > the rest of the operations and tasks can proceed. > To be more accurate, we actually schedule a new task "delete the > entry" which goes to the end of the list with tasks. > This "delete the entry" task, when scheduled, will delete the > routing entry or will be a no-op if an earlier task marked the entry > as being in-use. > > Hope that helps, > Pavlin > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.ICSI.Berkeley.EDU/pipermail/xorp-hackers/attachments/20061031/d5993edf/attachment.html