- ixp1200 - lists.cs.princeton.edu

Running Vera software on newer boards.
by Christopher Adam Telfer 19 Jan '02

19 Jan '02

Hello, We were considering trying to install the Vera software on ixp1200 boards we recieved here at Purdue. However, I noticed that the software was written for the IXP12EB evaluation board. We, however, are using Radisys ENP-2505 boards : http://www.radisys.com/oem_products/ds-page.cfm?ProductDatasheetsID=1055&ms… Has anyone managed to get Vera compiled and run for this board? Some notable differences between the two seem to include a lack of jumpers and LEDs on the Radisys board. This board seems to use the same BootMgr, although it seems to lack VxWorks. (actually bad things seem to happen when you try to boot VxWorks) Any information would be helpful. Thanks in advance, -- Chris Telfer Purdue University

3 3

Memory map question...
by Austen McDonald 15 Jan '02

15 Jan '02

Ive asked this question before and got no answer: according to the header files in asm/arch-ixp1200/ the way SDRAM packet memory (address 0 in the micorengines) is mapped to virt memory hasnt changed between BV the eval board. But, when I go to read from virt addr 0xd8000000 (where DRAM_BASE is in board.h) I dont get what I write to 0x0 in the uE. Where is the bottom of the uE's SDRAM mapped in linux? An additional question: how exactly is linux positioned in memory? What I really want to know is where can the uEs write packets to and how much space do they have on the BV boards? Austen McDonald austen(a)cc.gatech.edu God does not care about our mathematical difficulties. He integrates empirically. -- Albert Einstein

1 0

RE: [ixp1200] Network Cards' Compatibility with the Octal ports o f IXP1200
by Raghunandan, Makaram 11 Jan '02

11 Jan '02

The LEDs on the IXP1200 boards are dual colored. So, green is 100 Mbps and amber indicates 10 mbps... -----Original Message----- From: Tan Chee Wei [mailto:eng80386@nus.edu.sg] Sent: Wednesday, January 09, 2002 1:10 AM To: ''ixp1200(a)CS.Princeton.EDU ' ' Subject: [ixp1200] Network Cards' Compatibility with the Octal ports of IXP1200 hallo, I have some network cards that seem not compatible with the octal ports of the IXP1200 ie. the LEDs show an orange-colored signal when connected using those network cards and not the green-colored indication. I have 6 working 3-Com cards, and all have that same problems. I also have 2 similar working Xircom cards, one gives a green-colored signal and the other not. It has been traced down to be the problem of the network cards and not software or PCI planes. Does anyone have any idea, why ? Thank You and kind regards, cheewei

1 0

RE: [ixp1200] Problem making ixp1200 linux armtoolchain
by Brandewie, Dirk J 10 Jan '02

10 Jan '02

Look back in your logs to see which patch failed. Patch should have created a *.rej file for the hunk that failed. I am assuming you are using the tarballs from netwinder.org true? A far easier way to get a working toolchain is to use the toolchain that comes with the Intel IXA SDK 2.0. You can order your free copy of the SDK from http://developer.intel.com/design/network/products/npfamily/sdk2.htm or call 1-800-548-4725 and order SKU number for SDK V2.0 Vol II is A68978-001 This CD set contain an RPM to install a working toolchain. Hope this helps --Dirk > -----Original Message----- > From: lcard [mailto:lcard@win.trlabs.ca] > Sent: Thursday, January 10, 2002 2:33 PM > To: ixp1200(a)CS.Princeton.EDU > Subject: [ixp1200] Problem making ixp1200 linux armtoolchain > > > Hello, > > I've run into a bit of a snag trying to make the ixp1200 armtoolchain. > Make is happy until it exits trying to make the linux_patch target. > > (Lots of output ommitted) > patching file linux/include/asm-arm/arch-ixp1200/system.h > patching file linux/include/asm-arm/arch-ixp1200/time.h > patching file linux/include/asm-arm/arch-ixp1200/timex.h > patching file linux/include/asm-arm/arch-ixp1200/uncompress.h > patching file linux/include/asm-arm/arch-ixp1200/vmalloc.h > make: *** [linux_patch] Error 1 > > Has anyone had a similar problem? > > -Regards > > Loren Card > B.Sc. Computer Engineering > System Administrator, Graduate Student, > TRLabs, Winnipeg, Canada. > Tel: (204)488-5626 > Email: lcard(a)win.trlabs.ca > > > > >

1 0

Problem making ixp1200 linux armtoolchain
by lcard 10 Jan '02

10 Jan '02

Hello, I've run into a bit of a snag trying to make the ixp1200 armtoolchain. Make is happy until it exits trying to make the linux_patch target. (Lots of output ommitted) patching file linux/include/asm-arm/arch-ixp1200/system.h patching file linux/include/asm-arm/arch-ixp1200/time.h patching file linux/include/asm-arm/arch-ixp1200/timex.h patching file linux/include/asm-arm/arch-ixp1200/uncompress.h patching file linux/include/asm-arm/arch-ixp1200/vmalloc.h make: *** [linux_patch] Error 1 Has anyone had a similar problem? -Regards Loren Card B.Sc. Computer Engineering System Administrator, Graduate Student, TRLabs, Winnipeg, Canada. Tel: (204)488-5626 Email: lcard(a)win.trlabs.ca

1 0

RE: [ixp1200] Network Cards' Compatibility with the Octal ports o f IXP1200
by Brandewie, Dirk J 10 Jan '02

10 Jan '02

The single orange light denotes that the port is running at 10Mbps. When both led's are lit the port is running at 100Mbps --Dirk > -----Original Message----- > From: Tan Chee Wei [mailto:eng80386@nus.edu.sg] > Sent: Tuesday, January 08, 2002 10:10 PM > To: ''ixp1200(a)CS.Princeton.EDU ' ' > Subject: [ixp1200] Network Cards' Compatibility with the > Octal ports of > IXP1200 > > > hallo, > > I have some network cards that seem not compatible with the octal > ports of the IXP1200 ie. the LEDs show an orange-colored signal when > connected using those network cards and not the green-colored > indication. > I have 6 working 3-Com cards, and all have that same > problems. I also have > 2 similar working Xircom cards, one gives a green-colored > signal and the > other not. > It has been traced down to be the problem of the network cards and > not software or PCI planes. > Does anyone have any idea, why ? > > Thank You and kind regards, > cheewei >

1 0

RE: [ixp1200] Errors during installation
by Tan Chee Wei 10 Jan '02

10 Jan '02

Hallo Vishal, >Now however I am getting the following error when doing a install in the >"Runtime/linux-i386/driver" directory >[vanand@localhost driver]$ make install >kgcc -I. -I/usr/src/linux/include -DMODULE -D__KERNEL__ >-DIXP1200_SUPPORT >-O -Wall >-c drvVera.c -o drvVera.o >In file included from /usr/include/linux/module.h:21, > from drvVera.c:59: >/usr/include/linux/modversions.h:1: #error Modules should never use >kernel-headers system headers, >/usr/include/linux/modversions.h:2: #error but headers from an >appropriate >kernel-source >In file included from drvVera.c:70: >/usr/include/linux/modversions.h:1: #error Modules should never use >kernel-headers system headers, >/usr/include/linux/modversions.h:2: #error but headers from an >appropriate >kernel-source >make: *** [drvVera.o] Error 1 The above problems can be resolved in 2 methods. Untar a new linux tree higher than 2.4.5 in /usr/src/ and change the HDR_DIRS to /usr/src/linux-2.4/include in makefile in "Runtime/linux-i386/driver" directory. Or else you can create a soft link linux in /usr/src pointing to the link linux-2.4 kind regards, cheewei

1 0

Errors during installation
by Vishal Anand 09 Jan '02

09 Jan '02

>gcc: cannot specify -o with -c or -S and multiple compilations > >Somehow you have the string "export" in one of your environment >variables. What are the values of the environment variables >XCCROOT and VERAROOT? They should look something like this: > >[scott@vera i386]$ echo $XCCROOT >/usr/xcc >[scott@vera i386]$ echo $VERAROOT >/home/scott/Vera You were right, there was a missing quote in the .bashrc file which caused an extra "export" to appear in the env. var's. Now however I am getting the following error when doing a install in the "Runtime/linux-i386/driver" directory [vanand@localhost driver]$ make install kgcc -I. -I/usr/src/linux/include -DMODULE -D__KERNEL__ -DIXP1200_SUPPORT -O -Wall -c drvVera.c -o drvVera.o In file included from /usr/include/linux/module.h:21, from drvVera.c:59: /usr/include/linux/modversions.h:1: #error Modules should never use kernel-headers system headers, /usr/include/linux/modversions.h:2: #error but headers from an appropriate kernel-source In file included from drvVera.c:70: /usr/include/linux/modversions.h:1: #error Modules should never use kernel-headers system headers, /usr/include/linux/modversions.h:2: #error but headers from an appropriate kernel-source make: *** [drvVera.o] Error 1 The relevant lines in the /usr/include/linux/module.h file (starting from line #14) are #ifdef __GENKSYMS__ # define _set_ver(sym) sym # undef MODVERSIONS # define MODVERSIONS #else /* ! __GENKSYMS__ */ # if !defined(MODVERSIONS) && defined(EXPORT_SYMTAB) # define _set_ver(sym) sym # include <linux/modversions.h> # endif #endif /* __GENKSYMS__ */ Any clue what I am doing wrong now? Regards, -Vishal _________________________________________________________________ Join the worlds largest e-mail service with MSN Hotmail. http://www.hotmail.com

1 0

Network Cards' Compatibility with the Octal ports of IXP1200
by Tan Chee Wei 09 Jan '02

09 Jan '02

hallo, I have some network cards that seem not compatible with the octal ports of the IXP1200 ie. the LEDs show an orange-colored signal when connected using those network cards and not the green-colored indication. I have 6 working 3-Com cards, and all have that same problems. I also have 2 similar working Xircom cards, one gives a green-colored signal and the other not. It has been traced down to be the problem of the network cards and not software or PCI planes. Does anyone have any idea, why ? Thank You and kind regards, cheewei

1 0

Re: Plugging ixp into Pentium's backplane
by Tan Chee Wei 09 Jan '02

09 Jan '02

Dear Scott and all, The README and makefiles given in the release of Vera are clear, and useful enough to get the software and hardware up successfully. I noted some points below during my set-up which I think might be useful to some for the time being. I'm using linux kernel 2.4.7 and both microcode assembler 1.2.76 and 2.0.83. Software set-up: - Do create a soft link linux in /usr/src pointing to linux-2.4 if using linux kernel higher than 2.4.5. - In README.details.txt, after "make nodes", do a "make all" to create vera.o - For microcode with vrp compilation, the uwin's development package failed to install on my win98, so I used a native win32 ports of gnu instead: http://www.wzw.tu-muenchen.de/~syring/win32/UnxUtils.html - The microcode with novrp can be compiled directly using the Intel Developer Workbench. - Do replace #if USE_REGALLOC with #ifdef USE_REGALLOC in some uc source. - All microcodes are compatible with both assembler 1.2.76 and 2.0.83. - Yitzchak Gottlieb from Princeton provided the following definitions to compile IPmm and IPcache. Include them at $VERAROOT/Runtime/ixp1200/sa/src/Lib/mem_access_p.h if they are not there. /* BEGIN DEFINTIONS */ extern inline u16 memRead16(u32 address) { return *((volatile u16 *)address); } extern inline void memWrite16(u32 address, u16 value) { *((volatile u16 *)address) = value; } /* END DEFINITIONS */ Hardware set-up : - I was indeed not careful enough with the J27 jumper. Its factory setting seems to differ with that described in manual. That could cause the problem of BIOS not coming up as described in earlier mail. Follow the recommendations from Scott as in previous mail. - Do connect the serial port of the ixp1200 as the Firmware makes minimon call during start-up. - Do set 9600 8N1 for minimon kind regards, cheewei -----Original Message----- From: Scott C. Karlin To: ixp1200(a)CS.Princeton.EDU Cc: Tan Chee Wei; xiaolin(a)krdl.org.sg; biswas(a)krdl.org.sg; Tham Chen Khong Sent: 1/9/02 6:41 AM Subject: [ixp1200] Re: Plugging ixp into Pentium's backplane On Sat, Jan 05, 2002 at 12:12:29AM +0800, Tan Chee Wei wrote: > > >The test against bit 2 of IXP1200_GPIO_DATA > >is a check on jumper J27. The README.jumpers.txt says to set > >J27 to 2-3. Is J27 set correctly? > > Yes, J27 is set correctly. In fact, there isn't any need to change > J27 since the factory's setting is similar to Vera's setting. Some more thoughts... I just opened a factory-fresh development system and the factory setting for J27 was 1-2. Be sure to verify all jumper settings to ensure they are jumpered for the Vera setting. Also, pin 1 for the jumpers is denoted by a square solder pad on the back side of the board. I hope this helps, Scott

1 0

Re: Plugging ixp into Pentium's backplane
by Scott C. Karlin 08 Jan '02

08 Jan '02

On Sat, Jan 05, 2002 at 12:12:29AM +0800, Tan Chee Wei wrote: > > >The test against bit 2 of IXP1200_GPIO_DATA > >is a check on jumper J27. The README.jumpers.txt says to set > >J27 to 2-3. Is J27 set correctly? > > Yes, J27 is set correctly. In fact, there isn't any need to change > J27 since the factory's setting is similar to Vera's setting. Some more thoughts... I just opened a factory-fresh development system and the factory setting for J27 was 1-2. Be sure to verify all jumper settings to ensure they are jumpered for the Vera setting. Also, pin 1 for the jumpers is denoted by a square solder pad on the back side of the board. I hope this helps, Scott

1 0

Re: [ixp1200] Re: Plugging ixp into Pentium's backplane
by Vishal Anand 08 Jan '02

08 Jan '02

Hi, When doing th efollowing step in the installation process: cd $VERAROOT/Libsck ; make install I get the following errors: [vanand@localhost Libsck]$ make install make[1]: Entering directory `/home/vanand/study/Vera/Libsck/src' make[2]: Entering directory `/home/vanand/study/Vera/Libsck/src/i386' Updating dependency makefile strlenn.d from strlenn.c gcc: export: No such file or directory Updating dependency makefile strtoule.d from strtoule.c gcc: export: No such file or directory Updating dependency makefile strtoargs.d from strtoargs.c gcc: export: No such file or directory Updating dependency makefile stricmp.d from stricmp.c gcc: export: No such file or directory Updating dependency makefile srec.d from srec.c gcc: export: No such file or directory Updating dependency makefile rmgr.d from rmgr.c gcc: export: No such file or directory ../rmgr.c:19:38: warning: No include path in which to find bit.h ../rmgr.c:20:38: warning: No include path in which to find mem.h ../rmgr.c:21:38: warning: No include path in which to find table.h Updating dependency makefile mrm.d from mrm.c gcc: export: No such file or directory ../mrm.c:21:38: warning: No include path in which to find mem.h Updating dependency makefile lineproc.d from lineproc.c gcc: export: No such file or directory make[2]: Leaving directory `/home/vanand/study/Vera/Libsck/src/i386' make[2]: Entering directory `/home/vanand/study/Vera/Libsck/src/i386' gcc -I. -I.. -I/home/vanand/study/Vera export -IXCCROOT=/usr/xcc/Libcii/include -Wall -Wstrict-prototypes -O2 -c lineproc.c -o lineproc.o gcc: cannot specify -o with -c or -S and multiple compilations make[2]: *** [lineproc.o] Error 1 make[2]: Leaving directory `/home/vanand/study/Vera/Libsck/src/i386' make[2]: Entering directory `/home/vanand/study/Vera/Libsck/src/arm' Updating dependency makefile strlenn.d from strlenn.c arm-unknown-coff-gcc: export: No such file or directory arm-unknown-coff-gcc: export: No such file or directory Updating dependency makefile strtoule.d from strtoule.c arm-unknown-coff-gcc: export: No such file or directory arm-unknown-coff-gcc: export: No such file or directory Updating dependency makefile strtoargs.d from strtoargs.c arm-unknown-coff-gcc: export: No such file or directory arm-unknown-coff-gcc: export: No such file or directory Updating dependency makefile stricmp.d from stricmp.c arm-unknown-coff-gcc: export: No such file or directory arm-unknown-coff-gcc: export: No such file or directory Updating dependency makefile srec.d from srec.c arm-unknown-coff-gcc: export: No such file or directory arm-unknown-coff-gcc: export: No such file or directory Updating dependency makefile rmgr.d from rmgr.c arm-unknown-coff-gcc: export: No such file or directory arm-unknown-coff-gcc: export: No such file or directory ../rmgr.c:19: warning: No include path in which to find bit.h ../rmgr.c:20: warning: No include path in which to find mem.h ../rmgr.c:21: warning: No include path in which to find table.h Updating dependency makefile mrm.d from mrm.c arm-unknown-coff-gcc: export: No such file or directory arm-unknown-coff-gcc: export: No such file or directory ../mrm.c:21: warning: No include path in which to find mem.h Updating dependency makefile log2_floor.d from log2_floor.c arm-unknown-coff-gcc: export: No such file or directory arm-unknown-coff-gcc: export: No such file or directory make[2]: Leaving directory `/home/vanand/study/Vera/Libsck/src/arm' make[2]: Entering directory `/home/vanand/study/Vera/Libsck/src/arm' arm-unknown-coff-gcc -I. -I.. -I/home/vanand/study/Vera export -IXCCROOT=/usr/xcc/Libcii/include -I/home/vanand/study/Vera export -IXCCROOT=/usr/xcc/Libsck/include -nostartfiles -nodefaultlibs -fno-builtin -mcpu=strongarm110 -mapcs-32 -mno-sched-prolog -fvolatile -Wall -Wstrict-prototypes -O2 -c ../log2_floor.c -o log2_floor.o arm-unknown-coff-gcc: cannot specify -o with -c or -S and multiple compilations make[2]: *** [log2_floor.o] Error 1 make[2]: Leaving directory `/home/vanand/study/Vera/Libsck/src/arm' make[1]: *** [install] Error 2 make[1]: Leaving directory `/home/vanand/study/Vera/Libsck/src' make: *** [install] Error 2 ---------------------------------------------------------------------- I did not get any errors when installing in the Libcii, Libesc, Libesc2 dir's Any idea what I am missing in the installation steps. Regards, -Vishal _________________________________________________________________ MSN Photos is the easiest way to share and print your photos: http://photos.msn.com/support/worldwide.aspx

2 1

Re: Plugging ixp into Pentium's backplane
by Scott C. Karlin 08 Jan '02

08 Jan '02

On Sat, Jan 05, 2002 at 12:12:29AM +0800, Tan Chee Wei wrote: > > > The test against bit 2 of IXP1200_GPIO_DATA > > is a check on jumper J27. The README.jumpers.txt says to set > > J27 to 2-3. Is J27 set correctly? > > Yes, J27 is set correctly. In fact, there isn't any need to change > J27 since the factory's setting is similar to Vera's setting. > But I used the ixp board without using the serial cable connected because > I was concerned whether Pentium will get booted up or not. In that case, > I guess, it is a must to connect up the serial port of the ixp board. At this point I cannot think of another (simple) explanation as to why you would see "16" on the leds without GPIO bit 2 (0x4) being set. There is no requirement that the serial port of the IXP1200 board be connected. However, as you can see, it's more useful to connect it to a different machine so that you can see the output messages even if the host Pentium doesn't boot for some reason. > > If J27 is set to 1-2, the board enters minimon *before* > > initializing the PCI bus. The BIOS will hang because the > > IXP1200 is not responding to the BIOS query. You can verify > > this by checking the serial port output from the ixp1200 board. > > Can I confirm that the mini-monitor in $VERAROOT/Runtime/ixp1200/sa/src/Lib/ > will work with both Window's Hyperterminal as well as linux minimon? Minimon uses no special control codes and will work with any RS-232 terminal or terminal emulator set to 9600 baud, 8 data bits, 1 stop bit, no parity. Scott

1 0

Plugging ixp into Pentium's backplane
by Tan Chee Wei 04 Jan '02

04 Jan '02

Dear Scott, Thank You very much for your prompt reply. Currently, I have successfully compiled all Vera's codes , changed the Intel's firmware to a Princeton's firmware and reset the jumpers' setting according to README.jumpers.txt. But I encountered problems when I removed the ixp board from its factory chassis and plugged it into a Pentium machine. I tried out the same ixp board on 3 different Pentiums, two P2 and one P4. The common problem is, the Pentium cannot boot up. Nothing appears on the monitor,not even BIOS messages. Judging from the initial "noise" from the CPU, the Pentium did go into initial BIOS mode of memory test but after that it simply stopped. All the ixp board's LEDs flashed once upon power up and the LED indicators showed '1' and '6'. What does '1' and '6' mean ? I swopped the PCI slot of the ixp board in the active backplane, but to no success. Using any PCI slot on the Pentium is fine, right? Initially, there was the doubt of conflict with the graphics card, but surely not after 3 different Pentium machines were tested, right ? Next, I shorted the reset pins (J21)and retried all my procedures. The Pentium booted up successfully and then the reset pins were opened. The LED indicators changed from the initial '8''8' to '1''6'. I feel that my ixp board is not yet configured properly as a PCI board, despite the fact that the jumpers setting were double-checked to be that of Vera's. What could cause the conflicts ? Why is PCI slot 4 on the factory chassis's PCI plane enabled and the rest disabled ? Thank You for reading through my problems and I really appreciate all help now ! Please hit the "reply-all" button when replying. Thank You and kind regards, cheewei

2 1

RE: [ixp1200] Query
by Heaton, Eric D 03 Jan '02

03 Jan '02

KP, A Microengine can interupt the SA Core by performing a fast_wr instruction to the IREG register. Here is some microcode to do this: Sv_Ack#: ;Calculate mailbox address immed[thread_mb_base, UTHREAD_TO_CORE_MAILBOXES] alu[to_core_mb, thread_mb_base, +, my_thread_id, <<2] ;Format a message and write to shared memory location alu[reply, reply, OR, format, <<28] ; insert format alu[$xfer0, --, B, reply] alu[indirect_wr_len, format, OR, 0x10] ; setup word count alu[indirect_wr_len, --, B, indirect_wr_len, <<16] sram[write, $xfer0, to_core_mb, 0, 1], indirect_ref, ctx_swap ;Signal Core by writing to IREG fast_wr[1, ireg] For the Core code, there is only one interrupt routine registered to the operating system. Secondary interrupt handler functions are registered with this first layer function that handle specific interrupts. The Interrupt handler demux parses the interrupt vector to determine which of the secondary functions to call to actually do something with the interrupt. In the case above, the debug interrupt handler might be called when bit 24 is set in the IREG register or the thread interrupt handler is called when one of the bits 0:23 are set in IREG. I hope this helps! Eric Heaton Technical Marketing Engineer Intel Corporation - NPBU -----Original Message----- From: paramesh [mailto:paramesh@trinc.com] Sent: Thursday, January 03, 2002 11:28 AM To: ixp1200(a)CS.Princeton.EDU Subject: [ixp1200] Query Hi, Can anyone of you make me sense on the usage of interrupt handling on StrongArm(SA) side over Vxworks with IREG for FIQ interrupts from microengine plz.. Best Regards KP

1 0

Query
by paramesh 03 Jan '02

03 Jan '02

Hi, Can anyone of you make me sense on the usage of interrupt handling on StrongArm(SA) side over Vxworks with IREG for FIQ interrupts from microengine plz.. Best Regards KP

1 0

Re: FILE.bin in Vera
by Scott C. Karlin 03 Jan '02

03 Jan '02

On Thu, Jan 03, 2002 at 05:18:24PM +0800, Tan Chee Wei wrote: > > I am trying out the Vera's installation as described in README.flash. > What puzzles me currently is this file FILE.bin which you described > under the pre-requisites of changing the Flash from Intel firmware to > Princeton firmware. I believe the Princeton firmware refers FILE.bin. > But where should I get the Princeton firmware? Is it the file > FILE.bin refering to $VERAROOT/Flash/bin/flash? I can see flash is an > executable. Is this the Princeton firmware? If not, where can I get > it? FILE.bin refers to the flash image you build when you compile the Runtime code. It is $VERAROOT/Runtime/ixp1200/sa/src/Firmware/flashImage.bin $VERAROOT/Flash/linux-i386/bin/flash is the Linux program you run to update the flash on the ixp1200 board which is already running in a Linux PC. > The next question is the use of serial to access the message from > StrongARM. Do you connect the output of the serial from the IXP1200 > directly to the serial input of the Pentium host machine? You can if you like. However, if the Linux machine crashes, you won't be able to see the last bit of output or use the minimon mode to poke around memory. I'd recommend connecting the IXP1200 serial port to a separate machine. Scott

1 0

RE: [ixp1200] How to use large Boot ramdisk?
by Naik, Uday 02 Jan '02

02 Jan '02

The ramdisk for SDK 2.0 IS 16.0 MB. But if you want to change it --- Change linux/arch/arm/kernel/arch.c. Look for CONFIG_ARCH_IXP1200. Also you need to change the way you build the ramdisk. Look in utils/ramdisk/mkramdisk -----Original Message----- From: Victor [mailto:fei.xie@asu.edu] Sent: Monday, December 31, 2001 2:00 AM To: ixp1200(a)CS.Princeton.EDU Subject: [ixp1200] How to use large Boot ramdisk? Hi, there As I know the default ramdisk size of Linux version 2.3.99-pre3-rmk3-np2 of IXASDK 2.0 is 8MB, how can we recompile the kernel to make use of the larger ramdisk? I have tried to set "CONFIG_BLK_DEV_RAM_SIZE=16384" in the .config, but it didn't work. And because the kernel is downloaded from the host, we can modify the lilo.conf to make it. If somebody have experience on it, please give some advices. Thanks Victor

1 0

How to use large Boot ramdisk?
by Victor 31 Dec '01

31 Dec '01

Hi, there As I know the default ramdisk size of Linux version 2.3.99-pre3-rmk3-np2 of IXASDK 2.0 is 8MB, how can we recompile the kernel to make use of the larger ramdisk? I have tried to set "CONFIG_BLK_DEV_RAM_SIZE=16384" in the .config, but it didn't work. And because the kernel is downloaded from the host, we can modify the lilo.conf to make it. If somebody have experience on it, please give some advices. Thanks Victor

1 0

subscription
by Siew Mee Yong 28 Dec '01

28 Dec '01

===== SiEw MeE __________________________________________________ Do You Yahoo!? Yahoo! Shopping - Beat the crowds! Do your X'mas Shopping Online! http://shopping.yahoo.com.sg/

1 0

Query abt Interrupt handling on IXP1200!!
by param 27 Dec '01

27 Dec '01

Hi there, I did come across abt this IXP1200 mailing list thru websearch. As am new to this area, please do make me sense abt the usage of Interrupts on SA (Strong Arm) side from uEngine. Query : For interrupting a praticular thread in the uEngine from SA, am using INTER_THD_SIG register : *((unsigned int*) (0xB004 01C0))=thd_id // thd_id range : 0 to 23 Afore mentioned C statement would interrupt a thread in the uEngine based on thd_id. Where as in other case, Interrupting SA from uEgine, am not sure abt making use of IREG ... I would appreciate if some one make me sense in accomplishing the following required task. 1. Interrupting SA from uEngine (by IREG) 2. Capturing interrupt at SA which has been asserted by uEngine 3. Serving the Interrupt by making use of vector table, handler & etc... Sample snippet would be greatly appreciated for the afore mentioned request. Thanks & Kindest Regards Param

1 0

SDRAM Packet address on BV cards...
by Austen McDonald 20 Dec '01

20 Dec '01

The entry SDRAM_PACKET_BASE is gone from asm/arch-ixp1200/hardware.h---what is the address in linux that corresponds to the uE sdram addr 0? Austen McDonald austen(a)cc.gatech.edu God does not care about our mathematical difficulties. He integrates empirically. -- Albert Einstein

1 0

Bugs in octmac.c documentation?
by Austen McDonald 20 Dec '01

20 Dec '01

In ixasdk/src/microace/aces/interface_ace/source/driver/octdrv/octmac.c it says that the default names for the devices in /dev are octMac0,...octMac7 but the README file says mac0...mac7. I changed my code to say mac--will this affect anything (like the interface aces)? Austen McDonald austen(a)cc.gatech.edu God does not care about our mathematical difficulties. He integrates empirically. -- Albert Einstein

1 0

IXP1200 for TCP termination
by Vishal Anand 19 Dec '01

19 Dec '01

Hi, We have 2 IXP12EB evaluation kit for the IXP1200 processor. Was curious if anybody has ever done a feasibility study for doing TCP termination on the IXP. Theoretically it would seem that the Micro engines are an ideal choice for executing the TCP Fast Path code, while the ARM could run the Slow Path. Would appreciate any feedback if somebody did any work in this direction. Regards, -Vishal Anand (San Jose State University, CA) _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp

3 3

The PciDg Driver's problem
by Victor 13 Dec '01

13 Dec '01

Hi, Dirk How are you doing? I just tested some program on the StrongARM, which transfer the packets between the uEngine and the Host PC. But I found it seems the PciDg module( the Pseudo Network Driver between Host and StrongARM) will lost packets in the transmission. I tested my program on the real network driver, it works fine. It only lost packets on PciDg driver. So do you have any infomation about this problem? I read some source code of the PciDg, but can't figure out the problem yet. Have anyone encountered this problem before? Please give some advices. Thank you very much. Victor

1 0