December 2002 - ixp1200 - lists.cs.princeton.edu

RE: [ixp1200] Using PETH drivers with L3fwd16 on ENP-2505
by Arthur Berkovitch 12 Dec '02

12 Dec '02

pay attention to descriptor format: a PETH driver is built with RFC1812 uses a newer descriptor format but the reference design use the older packet descriptor. You need to remove -DRFC1812 from the peth/Makefile to set the PETH to older descriptor format. ... MODFLAGS = -D__KERNEL__ -DMODULE -DIOSTYLE=HARDWARE -DLINUX -DOS=UNIX \ -D__linux__ -DSPECTACLE_ISLAND -DRFC1812 ... -----Original Message----- From: Chong Poh Kit [mailto:thesoothsayer@pd.jaring.my] Sent: Wednesday, December 11, 2002 10:52 AM To: ixp1200(a)CS.Princeton.EDU Subject: [ixp1200] Using PETH drivers with L3fwd16 on ENP-2505 Hi all, I was wondering if anyone could tell me how to get the ENP-2505 to forward packets by using the peth driver and hardware debugging on the workbench and not through the MicroACE framework. After inserting the ue.o, octmac.o, giga_drv.o modules and running rs_udebug and rtmapp, I started a hardware debug session using the L3fwd16 project. I then loaded the peth driver and initialised the peth devices similar to the example in the "load" file. However, when I try to ping a computer connected to the ENP-2505 ethernet port, I get this message: Bad descriptor 0, dropping outgoing packet pethFreeDescriptor: Bad descriptor 0 I've tried defining RFC1812 in the code and also in build>settings for the project but the same problem occurs. Any help or pointers would be greatly appreciated. Thanks. Regards, P.K. Chong Multimedia University, Malaysia

2 1

Using PETH drivers with L3fwd16 on ENP-2505
by Chong Poh Kit 11 Dec '02

11 Dec '02

Hi all, I was wondering if anyone could tell me how to get the ENP-2505 to forward packets by using the peth driver and hardware debugging on the workbench and not through the MicroACE framework. After inserting the ue.o, octmac.o, giga_drv.o modules and running rs_udebug and rtmapp, I started a hardware debug session using the L3fwd16 project. I then loaded the peth driver and initialised the peth devices similar to the example in the "load" file. However, when I try to ping a computer connected to the ENP-2505 ethernet port, I get this message: Bad descriptor 0, dropping outgoing packet pethFreeDescriptor: Bad descriptor 0 I've tried defining RFC1812 in the code and also in build>settings for the project but the same problem occurs. Any help or pointers would be greatly appreciated. Thanks. Regards, P.K. Chong Multimedia University, Malaysia

1 0

Regarding running tcpdump in IXPENP2505 board
by NING WENG 10 Dec '02

10 Dec '02

Hi, Is there anyone who had experience to run tcpdump on IXPENP2505 board? How to install it? Thanks. -Ning

2 1

RE: [ixp1200] Threads
by Heaton, Eric D 08 Dec '02

08 Dec '02

Shruti, Why is the thread swapped out? That is, what signal is it waiting on? That will be the first clue to understand what's going on! To find out, you need to query the CTX_X_WAKEUP_EVENTS and CTX_X_SIG_EVENTS (X = thread number) family of CSRs. And the easiest way to do that if you've got the Developer's Workbench hooked up to the hardware, then you can simply open up the "Thread Status" windown in debug mode. If not, you'll have to write some StrongARM code to query the StrongARM memory map (see Appendix C of the IXP1200 PRM - Programmer's Reference Manual) or use your favorite StrongARM debugger to do this. If I had to guess, the thread is probably swapped out waiting for a packet indicator. Perhaps you haven't setup the global CSRs of the IXP1200 to match the media buses on your hardware? What code are you running on the StrongARM? It might be easiest to simply load the precompiled VxWorks image in the IXA SDK v2.01 first to get the hang to using the hardware before you try other configurations (ie. like the Linux code out there or something homegrown!). Good luck! --eric h. -----Original Message----- From: Shruti Gorappa [mailto:goraps@rpi.edu] Sent: Tuesday, November 26, 2002 8:05 PM To: ixp1200(a)CS.Princeton.EDU Subject: [ixp1200] Threads Hi I am using the IXP1200 Eval platform to run a routing and forwarding program. One of the threads simply forward packets and the other sends out generated packets. When I run the code on the hardware, I notice that the first thread gets swapped out and does not run. I also ran the l3fwd_16_RFC1812 example code on the hardware and noticed that the thread is swapped out in that too. The programs run great when I use the simulator and I am not able to find out what is different on the hardware. Could somebody help me out here? Thanks in advance. Shruti

2 2

compiler error "__mul_32 not found"
by Yan Luo 08 Dec '02

08 Dec '02

Hi, I got this compilation error in SDK 2.0 : (0): error : Function __mul_32 not found I use hash_48() in my micro C code. The parameters of hash_48 are sram xfer registers. And I declared and used other sram xfer registers too. Is the error because of running out of sram xfer registers ? Thanks, -- Luo, Yan yluo(a)cs.ucr.edu Dept. of CS, Univ. of California Riverside, CA 92507, USA http://www.cs.ucr.edu/~yluo

2 1

Intel IXA University Program Student Design Competition
by Schuetze, Joel D 06 Dec '02

06 Dec '02

Greetings, We are pleased to announce the Intel IXA University Program Student Design Competition. The Intel IXA University Program Student Design competition will allow undergraduate and graduate university students from U.S. Universities to present innovative work based on the Intel® IXA architecture, specifically the IXP family of network processors. There will be a web-based application process available through the IXAEDU.com website and all entries will be reviewed by an Intel panel of experts. Six (6) qualified selected entries will be awarded grants for the proposed projects. After the selected projects have been completed and submitted for judging, the top two (2) projects will be selected to receive an award presented at the Fall 2003 Intel Developer Forum (IDF). For complete information please visit http://ixaedu.com/news/DesignCompetition. Joel Schuetze IXA University Program Intel Corporation

1 0

Patch for SDK 2.0 to use all the available memory on bridalveil
by Abhijeet Joglekar 01 Dec '02

01 Dec '02

Hi, We are using the Bridalveil board with SDK 2.0 at Utah. SDK 2.0 sets up the board to use only 128 MB of Sdram and 4 MB of sram. Since we need more memory for packet buffers and other data structures, we patched some header and source files to use all the memory on the board (256 MB of sdram and 8 MB of sram) and to increase the memory pool allocated to ASL and RM. So our modified board configuration is: Sdram (total 256 MB) --> ASL (96 MB) --> RM (32 MB) --> Linux (126 MB) --> Vmon (2 MB) (Thus linux is loaded at an offset of 128 MB from the start of Sdram) Sram (total 8 MB) --> ASL (6 MB) --> RM ( 2MB) If you want to change the default memory configuration of SDK 2.0 on bridalveil for your application, you might find these changes useful. If you want a different configuration, you can look up the patch file and make the relevant changes in your code. The changes are available at www.cs.utah.edu/~abhijeet/ixp/utah_bv.tar.gz There are three files in it: - README (which explains how to patch your current source tree, which files to back up, which modules to recompile and how to test if the patch works) - bvpatch (the patch file for setting up the board to use all memory) - autobootbv (a script for automating boot procedure of bv. This is independent of the patch above, just bundled up in the same .tar.gz file, because it might be useful to automate your card bootup.) We have done minimal testing by running both the l2bridge and l3forward designs in workbench and non-workbench mode under the modified board configuration. They seem to be working fine. Feel free to mail abhijeet(a)cs.utah.edu if you have any questions or want to report any bugs in the patch. Thanks. -- abhijeet

1 0