June 2004 - ixp1200 - lists.cs.princeton.edu

RE: booting over PCI (ENP boards)...
by Ing. Ricardo Camarena Ramos 10 Jun '04

10 Jun '04

Hi Willem. Currently I'm trying to set-up five Linux (RedHat 7.2 and 7.3) workstations with ENP-2505 boards, but failed to accomplish the task. (Using the IXA EDU Setup guide 1.01 from Intel).I think the problem is on the pciDg.o and pciDgNet.o modules. Because the only problem I have is regarging loading these modules, as a result I cannot ping between Linux host and board. Linux and W2K can ping between each other. The main problem I have with Linux 7.2 (besides having a hard time to configure the video card, not finished yet either) is that although it seems pciDg.o and pciDgNet.o load correctly, eth1 is not created, I got this message: "SIOCSIFADDR: No such device eth1: unknown interface: no such device." With Linux 7.3 I have tried to recompile the modules following a guide from Laboratory of Advanced Systems Research at UCLA (I think)(http://webdiee.cem.itesm.mx/web2/comunicaciones/proyectos/otraguia.doc) but still I got a message telling me the modules "pciDg.o was compilled for version 2.4.18-4 while this kernel version is 2.4.18-3smp." So, I decided to follow my own re-compilation procedure. This is what I did: cd /usr/src/linux edit the Makefile and change the EXTRAVERSION=-4 make mrproper make oldconfig cd /mnt/cdrom/ENP-2505_Driver rpm -recompile boot_drv-2.0-1.src.rpm cd /usr/src/linux make clean make zImage To big, so I change to make bzImage make modules make modules_install cp /arch/i386/boot/bzImage /boot/vmlinuz-2.4.18-4 I modified the grub.conf file to add the new kernel reboot uname -r and I got 2.4.18-4 cd /opt/ixasdk/enp-2505/bin insmod pciDg.o pciDg.o was compiled for kernel ersion 2.4.17 while this kernel is version 2.4.18-4 Hope you or somebody can help me. Best regards. >-- Mensaje Original -- >Date: Thu, 10 Jun 2004 12:00:14 -0400 (EDT) >From: ixp1200-request(a)lists.cs.princeton.edu >Subject: ixp1200 Digest, Vol 7, Issue 1 >To: ixp1200(a)lists.cs.princeton.edu >Reply-To: ixp1200(a)lists.cs.princeton.edu > > >Send ixp1200 mailing list submissions to ixp1200(a)lists.cs.princeton.edu To subscribe or unsubscribe via the World Wide Web, visit https://lists.cs.princeton.edu/mailman/listinfo/ixp1200 or, via email, send a message with subject or body 'help' >o ixp1200-request(a)lists.cs.princeton.edu You can reach the person managing the list at ixp1200-owner(a)lists.cs.princeton.edu When replying, please edit your Subject line so it is more specific than "Re: Contents of ixp1200 digest..." >Today's Topics: 1. [WORKING] booting over PCI (ENP boards) from a 2.6 kernel installation (Willem de Bruijn) 2. How to duplicate a packet in memory? (?yvind Hvamstad) >Principio del mensaje anexo > >From: Willem de Bruijn <wdebruij(a)dds.nl> >Subject: [ixp1200] [WORKING] booting over PCI (ENP boards) from a 2.6 kernel > installation >Date: Thu, 10 Jun 2004 14:48:33 +0200 >To: ixp1200(a)lists.cs.princeton.edu >Reply-To: IXP1200 Developer's List <ixp1200(a)lists.cs.princeton.edu> > > >Hi, this is just a short message for people running Linux on an IXP board that connects to the host over PCI. As you probably know, most support applications are currently restricted to linux 2.4 kernels. For one of our projects I'm in the p >ocess of writing an alternate PciDg driver that will also work with newer (2.6) kernels. At the moment we are able to (1) map both ENP and EB memory over PCI to kernel- and userspace on the host. (2) boot the ENP boards over PCI using Radisys' o > Intel's bvloadapp utility while relying on our own GPLed mmapping module instead of PciDg. Note that pseudo-ethernet is not working yet. However, if you simply want to get the board up and running that's not an issue. Please let me know if yo >'re interested in an updated pseudo-ethernet driver, I may consider investing time in it then. If you want to try out the module, do the following: get the latest ffpf distribution from anonymous CVS via ffpf.sourceforge.net (I'll update the r >gular distribution download shortly): cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/ffpf login cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/ffpf co ffpf/src/v1.2 and then build the kernelspace modules in [dir]/kernelspac >, using 1) for 2.5.70+ and 2.6 (kbuild) infrastructure) ln -s Makefile.kernel.kbuild Makefile make -C [directory of your linux src, e.g. /usr/src/linux] SUBDIRS=${PWD} modules 2) for 2.4 and lower make -f Makefile.kernel.nokbuild this shoul > give you a module called genpci.ko. inserting this module with the proper parameters will ioremap all memory on an arbitrary PCI device to the host's kernel and export it to userspace via a file in /dev through both read/write/llseek and mmap mec >anisms. If the module is inserted without any options it will try to create the file /dev/genpci with all memory on an ENP board mapped. Since the SDRAM is in the 4th onboard memory region you'll have to add the length of the previous three me >ory reigons to the offsets used in bvloadapp.c to get bvloadapp running: #define LINUX_LOAD_ADDRESS 0x04008000 + 256 + 4096 + (1024*1024) #define RAMDISK_LOAD_ADDRESS 0x04800000 + 256 + 4096 + (1024*1024) also, bvloadapp asks for /dev/bv, so y >u should insmod genpci.ko using the following command: insmod genpci.ko name=/dev/bv or change the filename in bvloadapp.c. To get an EB booted, add the option device=0x1200 to the insmod parameter list and move the offsets around a bit. U >fortunately, my EB board kept resetting itself when put in a PCI slot, so I couldn't get it to work (I posted a question about this before). I hope that this can be of use to some of you. regards, Willem -- Willem de Bruijn wdebrui >_at_dds.nl http://www.wdebruij.dds.nl/ current project : Fairly Fast Packet Filter (FFPF) http://ffpf.sourceforge.net/ -- Dit bericht is gescand op virussen en andere gevaarlijke inhoud door ULCN MailScanner en het bericht lijkt schoon te z >jn. This message has been scanned for viruses and dangerous content by ULCN MailScanner, and is believed to be clean. >Fin del mensaje anexo > >Principio del mensaje anexo > >From: Øyvind Hvamstad <oyvindh(a)ifi.uio.no> >Subject: [ixp1200] How to duplicate a packet in memory? >Date: Thu, 10 Jun 2004 15:58:07 +0200 >To: ixp1200(a)lists.cs.princeton.edu >Reply-To: ixp1200(a)lists.cs.princeton.edu > > >Hi, I'm wondering what is the best way to duplicate a packet in SDRAM? The reason for this is that I want to send one copy of the packet to the StrongARM and the other copy will be forwarded to a different host after some modifications. 1. I coul > loop in microcode and use sdram[read] -> sdram[write] in sequence until I reach the end of the packet. 2. I can copy it using the StrongARM, but this would probably mean that the whole forwarding process will be stalled until the copying is fini >hed. 3. There is some 'magic' instruction that does this for me, that I'm unaware of. I'd much appreciate it if someone could give me a clue about what is the best practice here? Info: ENP-2505, embedded linux, Intel SDK 2.01 -- Øyvind Hv >mstad <oyvindh(a)ifi.uio.no> >Fin del mensaje anexo > >_______________________________________________ ixp1200 mailing list ixp1200(a)lists.cs.princeton.edu https://lists.cs.princeton.edu/mailman/listinfo/ixp1200 Ricardo Camarena Ramos Asistente de Investigación Departamento de Ingeniería Eléctrica y Electrónica División de Profesional y Graduados, Campus Estado de México Premio Nacional de Calidad 1998 Teléfono: (5255)5864.5555 x 3208, 5864.5660 camarena(a)itesm.mx enlace-intercampus: 80.236.3208, www.cem.itesm.mx Tecnológico de Monterrey, 60 AÑOS COMPROMETIDOS CON LA EXCELENCIA EDUCATIVA Hoy mejor que ayer, y mañana mejor que hoy.

1 0

How to duplicate a packet in memory?
by Øyvind Hvamstad 10 Jun '04

10 Jun '04

Hi, I'm wondering what is the best way to duplicate a packet in SDRAM? The reason for this is that I want to send one copy of the packet to the StrongARM and the other copy will be forwarded to a different host after some modifications. 1. I could loop in microcode and use sdram[read] -> sdram[write] in sequence until I reach the end of the packet. 2. I can copy it using the StrongARM, but this would probably mean that the whole forwarding process will be stalled until the copying is finished. 3. There is some 'magic' instruction that does this for me, that I'm unaware of. I'd much appreciate it if someone could give me a clue about what is the best practice here? Info: ENP-2505, embedded linux, Intel SDK 2.01 -- Øyvind Hvamstad <oyvindh(a)ifi.uio.no>

1 0

[WORKING] booting over PCI (ENP boards) from a 2.6 kernel installation
by Willem de Bruijn 10 Jun '04

10 Jun '04

Hi, this is just a short message for people running Linux on an IXP board that connects to the host over PCI. As you probably know, most support applications are currently restricted to linux 2.4 kernels. For one of our projects I'm in the process of writing an alternate PciDg driver that will also work with newer (2.6) kernels. At the moment we are able to (1) map both ENP and EB memory over PCI to kernel- and userspace on the host. (2) boot the ENP boards over PCI using Radisys' or Intel's bvloadapp utility while relying on our own GPLed mmapping module instead of PciDg. Note that pseudo-ethernet is not working yet. However, if you simply want to get the board up and running that's not an issue. Please let me know if you're interested in an updated pseudo-ethernet driver, I may consider investing time in it then. If you want to try out the module, do the following: get the latest ffpf distribution from anonymous CVS via ffpf.sourceforge.net (I'll update the regular distribution download shortly): cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/ffpf login cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/ffpf co ffpf/src/v1.2 and then build the kernelspace modules in [dir]/kernelspace, using 1) for 2.5.70+ and 2.6 (kbuild) infrastructure) ln -s Makefile.kernel.kbuild Makefile make -C [directory of your linux src, e.g. /usr/src/linux] SUBDIRS=${PWD} modules 2) for 2.4 and lower make -f Makefile.kernel.nokbuild this should give you a module called genpci.ko. inserting this module with the proper parameters will ioremap all memory on an arbitrary PCI device to the host's kernel and export it to userspace via a file in /dev through both read/write/llseek and mmap mechanisms. If the module is inserted without any options it will try to create the file /dev/genpci with all memory on an ENP board mapped. Since the SDRAM is in the 4th onboard memory region you'll have to add the length of the previous three memory reigons to the offsets used in bvloadapp.c to get bvloadapp running: #define LINUX_LOAD_ADDRESS 0x04008000 + 256 + 4096 + (1024*1024) #define RAMDISK_LOAD_ADDRESS 0x04800000 + 256 + 4096 + (1024*1024) also, bvloadapp asks for /dev/bv, so you should insmod genpci.ko using the following command: insmod genpci.ko name=/dev/bv or change the filename in bvloadapp.c. To get an EB booted, add the option device=0x1200 to the insmod parameter list and move the offsets around a bit. Unfortunately, my EB board kept resetting itself when put in a PCI slot, so I couldn't get it to work (I posted a question about this before). I hope that this can be of use to some of you. regards, Willem -- Willem de Bruijn wdebruij_at_dds.nl http://www.wdebruij.dds.nl/ current project : Fairly Fast Packet Filter (FFPF) http://ffpf.sourceforge.net/ -- Dit bericht is gescand op virussen en andere gevaarlijke inhoud door ULCN MailScanner en het bericht lijkt schoon te zijn. This message has been scanned for viruses and dangerous content by ULCN MailScanner, and is believed to be clean.

1 0