Re: ml310 kernel2.6 booting problems
Hello All, I have a sidewise question related to this thread. We have several ML-310 and ML-410 boards. At some point, we may need PCI but right now it is not urgent so I am not ready to invest a lot of time yet either. I imagine that 2.6 + PCI would be relatively similiar for both ML-310 and ML-410... My question is this: Would it make sense to have a Kernel configuration something like "VIRTEX_PCI" and a platforms "XILINX_ML4xx" / "XILINX_ML3xx" ... or create four platforms: XILINX_ML300 XILINX_ML310 XILINX_ML40x XILINX_ML410 I ask now because we are working on some device drivers for the ML410 that may work on 40x and I wondering if I should introduce a new ML410 platform option... Ron > > Date: Thu, 17 Jan 2008 21:19:04 +0100 > > From: Joachim Meyer > > Hmmm... I'm not really ready to invest that much time into the PCI for ML= > -310. > > -= > --- > > http://ozlabs.org/pipermail/linuxppc-embedded/2007-December/029211.html > > > > How much work would it be (approximated), and do you think I can success = > (I'm a novice, like you know)? > > Would you help me (Tell me where to start & what to do)? > > Hi Joachim, > > I was also interested in getting the PCI to run on the ML-310. I did > spend an evening trying to integrate a patch sent over by Stephen > Neuendorffer from Xilinx. I managed to get everything to compile > (basically I fixed a few missing macros and had to search a bit to > ensure that the memory mapping was correctly ported). Unfortunately, > when I tried the new kernel, I think that I must have messed something > up with the interrupt mappings (I am a newbie to PCI) and I had to > remove some interrupt mapping from Stephen's patch. > > In any case, my ML-310 was booting the Linux kernel, but something was > not right and I did not get anymore printouts from the UART Lite. > Since I don't have the JTAG cable, I was stuck at this point with > little means to debug the system. One day I'll spoil myself with one > of those platform JTAG USB cables... > > I used Z-modem to transfer files to/from my workstation to my CF card > on the ML-310 and this was an acceptable compromise. > > I'd still be interested in having the PCI bus working on the ML-310, > but now that I got my BEE2 booting Linux 2.6, its not such a big > priority for me. Let me know if you have some success in that > direction. > > Regards, > > Jean-Samuel > -- = > > Integrated Microsystems Laboratory > McGill University, Montr=E9al, QC, CANADA > Web Page: http://chaos.ece.mcgill.ca > ___ > Linuxppc-embedded mailing list > Linuxppc-embedded@ozlabs.org > https://ozlabs.org/mailman/listinfo/linuxppc-embedded ___ Linuxppc-embedded mailing list Linuxppc-embedded@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-embedded
Re: ml310 kernel2.6 booting problems
Oops. I should have looked at git.xilinx.com. I've been using secretlab's for a while and I saw xilinx site come up but never took the time to investigate. Even if it not mainline suitable code and 2.6.10-era, I would be interested. I'll explore in my free time over the next couple of weeks and see what I can figure out. Thanks for the speedy response! Ron > > Ron, > > It might be better to start with the kernel from git.xilinx.com, which > already has an ML41x option. The current version of EDK will also > generate support for ML410 with the PCI driver, if you are still using a > 2.6.10-era kernel. =20 > > I've refrained putting the PCI stuff on git.xilinx.com, because any > non-trivial usage of it that I've tried has generated warnings. On top > of that, the code has some interrupt values hardcoded in the ALI > southbridge code. Since there is no way to have this code pushed into > mainline, I'm not terribly interested in encouraging any work on it, > either. Perhaps since there is so much interest in this code, someone > would like to get it working in ARCH=3Dpowerpc? I'm happy to provide = > some > handholding to get someone started. > > Steve > > > -Original Message- > > From: [EMAIL PROTECTED] > [mailto:linuxppc-embedded- > > [EMAIL PROTECTED] On Behalf Of Ron Sass > > Sent: Friday, January 18, 2008 4:11 PM > > To: Jean-Samuel Chenard > > Cc: linuxppc-embedded@ozlabs.org > > Subject: Re: ml310 kernel2.6 booting problems > >=20 > >=20 > > Hello All, > >=20 > > I have a sidewise question related to this thread. We have > > several ML-310 and ML-410 boards. At some point, we may need > > PCI but right now it is not urgent so I am not ready to invest > > a lot of time yet either. I imagine that 2.6 + PCI would be > > relatively similiar for both ML-310 and ML-410... > >=20 > > My question is this: Would it make sense to have a Kernel > > configuration something like "VIRTEX_PCI" and a platforms > > "XILINX_ML4xx" / "XILINX_ML3xx" ... or create four platforms: > >=20 > > XILINX_ML300 > > XILINX_ML310 > > XILINX_ML40x > > XILINX_ML410 > >=20 > > I ask now because we are working on some device drivers for > > the ML410 that may work on 40x and I wondering if I should > > introduce a new ML410 platform option... > >=20 > > Ron > >=20 > > > > Date: Thu, 17 Jan 2008 21:19:04 +0100 > > > > From: Joachim Meyer > > > > Hmmm... I'm not really ready to invest that much time into the PCI > for ML=3D > > > -310. > > > > > > -=3D > > > --- > > > > > http://ozlabs.org/pipermail/linuxppc-embedded/2007-December/029211.html > > > > > > > > How much work would it be (approximated), and do you think I can > success =3D > > > (I'm a novice, like you know)? > > > > Would you help me (Tell me where to start & what to do)? > > > > > > Hi Joachim, > > > > > > I was also interested in getting the PCI to run on the ML-310. I > did > > > spend an evening trying to integrate a patch sent over by Stephen > > > Neuendorffer from Xilinx. I managed to get everything to compile > > > (basically I fixed a few missing macros and had to search a bit to > > > ensure that the memory mapping was correctly ported). > Unfortunately, > > > when I tried the new kernel, I think that I must have messed > something > > > up with the interrupt mappings (I am a newbie to PCI) and I had to > > > remove some interrupt mapping from Stephen's patch. > > > > > > In any case, my ML-310 was booting the Linux kernel, but something > was > > > not right and I did not get anymore printouts from the UART Lite. > > > Since I don't have the JTAG cable, I was stuck at this point with > > > little means to debug the system. One day I'll spoil myself with > one > > > of those platform JTAG USB cables... > > > > > > I used Z-modem to transfer files to/from my workstation to my CF > card > > > on the ML-310 and this was an acceptable compromise. > > > > > > I'd still be interested in having the PCI bus working on the ML-310, > > > but now that I got my BEE2 booting Linux 2.6, its not such a big > > > priority for me. Let me know if you have some success in that > > > direction. > > > > > > Regards, > > > > > > Jean-Samuel > > > -- =3D > > > > > > Integrated Microsystems Laboratory > > > McGill University, Montr=3DE9al, QC, CANADA > > > Web Page: http://chaos.ece.mcgill.ca > > > ___ > > > Linuxppc-embedded mailing list > > > Linuxppc-embedded@ozlabs.org > > > https://ozlabs.org/mailman/listinfo/linuxppc-embedded > > ___ > > Linuxppc-embedded mailing list > > Linuxppc-embedded@ozlabs.org > > https://ozlabs.org/mailman/listinfo/linuxppc-embedded > > ___ Linuxppc-embedded mailing list Linuxppc-embedded@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-embedded
Re: looking for 3U system with FPGA support
I'm not sure why you are keen on a 3U form factor but the Xilinx ML-410 board is an ATX form factor and fits in various 1U chassis. We have a couple in off-the-shelf chassis and some in a chassis that Xilinx sells. http://www.xilinx.com/ml410 The chassis is part number HW-MLX10-RACK-G. Coincidentally, the board has all of the features you mentioned. Ron > > Hi ! > > hope this is ok for this list - after searching quite some time for a > 3U board with a PPC on it + more or less any FPGA - I only could find the > MEN F12 (5200) and F13 (8540) - does naybody know of other 3U systems ? > > Basically a FPGA based system like an ML405 would be of interest but I > could not locate any 3U products. > > Roughly: > > PPC roughly equivalent to a 500 MHz PIII/Celeron > 2++ Ethernets > 2 Seriell > 4++ GPIO (via FPGA) > Linux available > U-boot if posible > > Any pointers would be appreciated. > > thx! > hofrat > > > ___ > Linuxppc-embedded mailing list > Linuxppc-embedded@ozlabs.org > https://ozlabs.org/mailman/listinfo/linuxppc-embedded ___ Linuxppc-embedded mailing list Linuxppc-embedded@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-embedded
linux-headers (powerpc v. ppc)
I hope this is to the right list... the recent discussion of retiring ppc arch encouraged me to ask. Ultimately, I'm trying to build a native PPC 405 compiler (i.e Canadian cross). I've got a crosstool's generated cross-compiler and I have glibc and a native bintools on my 405 root filesystem. What I'd like to do is: make ARCH=ppc headers_check make ARCH=ppc INSTALL_HDR_PATH=dest headers_install but that's not supported in the ppc tree. So my question is, can I just ... make ARCH=powerpc headers_check make ARCH=powerpc INSTALL_HDR_PATH=dest headers_install and then switch to ppc when I build the kernel? I suspect it will work but just don't know if there is a subtle difference that will come back to haunt me at run-time. Any advice would be greatly appreciated! Ron ___ Linuxppc-embedded mailing list Linuxppc-embedded@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-embedded
Re: EABI
Unfortunately, it is not that easy. The Xilinx tool chain use newlib but your Linux kernel (and rootfs) is probably built around glibc. When we've wanted to do a quick-and-dirty test of something that relies on Xilinx's patches to gcc, we have used powerpc-eabi-gcc to compile to assembler and then used our own crosstools compilers to link and load. That doesn't solve your problem but hopefully it explains the issue. Ron > Is it possible to compile a Linux application using an EABI compiler > (specfically, Xilinx's EDK powerpc-eabi-gcc.exe)? > > The issue at hand is that we'd like for our customers to be able to use > the Xilinx EDK toolchain (which we know they will have) to compile Linx > apps without having to install another toolchain (crosstool, ELDK, etc). > > So, what I'm hoping is that the EDK toolchain can be configured to be > Linux compatible binaries, or that there is some kind of wrapper that > will handle the incompatible interfaces. Searching around, I've seen > some mention of Linux EABI compatibility (for Debian ARM releases), but > I haven't found any clear concensus... > > P.S. My apologies if this message appears on the mailing list twice... > > -- > Brian Silverman > Concept X, LLC ___ Linuxppc-embedded mailing list Linuxppc-embedded@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-embedded
Re: APU FPU in Virtex ppc405
Shan and I are going to sit down and rebuild our tool chain (a mix of crosstool, linux-from-scratch, and Xilinx patches) next week. Our aim is to quantify the gains/losses of frequency versus hardware acceleration since we have several large scientific apps we plan to run. Like he said, we're hoping to have a handle on the best approach over the next couple of weeks. Ron > My understanding is that with the fpu on the ppc405 (not ppc440), the > processor clock rate had to be reduced for the fpu to work. This ended > up negating much of the performance benefits of using the fpu. > > I don't know the specifics offhand, but Xilinx answer records can be > searched. Hopefully I'm wrong, and there can still be a benefit to using > the ppc405 fpu. > > - John > > > On Thu, 2008-05-22 at 19:06 -0400, David H. Lynch Jr. wrote: > >Were the issues associated with getting the Xilinx Floating point APU > > working with Linux completely resolved ? > > Is there a git tree or patch collection with the appropriate patches > > etc ? > > > > > > > ___ Linuxppc-embedded mailing list Linuxppc-embedded@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-embedded
Re: Booting ML405 (Kernel panic - not syncing: No init found)
In addition to what John wrote, I would also investigate your ramdisk. I would be sure to check that you have the permissions/owner set correctly on bin/busybox. Also, I would double check that, if you compiled busybox with shared libraries, the shared libraries are in the right place on your ramdisk. Ron > > Hi, > > Yes I am using ARCH=ppc (actual line is $make ARCH=ppc > CROSS_COMPILE=powerpc-405-linux-gnu- zImage.initrd ) for this I have > placed ramdisk.image.gz in arch/ppc/boot/images. In case of ARCH=powerpc > I cannot find processor type 405 , in make menuconfig. Thats why i am > using ARCH=ppc. > > And when I give kernel command string as init=/bin/sh , it says : > >Failed to execute /bin/sh. Attempting defaults... > [3.744035] Kernel panic - not syncing: No init found. Try passing > init= option to kernel. > [3.768073] Rebooting in 30 seconds.. > > --init is present in sbin/init which is a soft link to ../bin/busybox > > --I used powepc-405-gnu-readelf -h bin/sh to verify that it is compiled > for powerpc itself. > > Any other suggestions ? > ___ Linuxppc-embedded mailing list Linuxppc-embedded@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-embedded
Re: Booting ML405 (Kernel panic - not syncing: No init found)
Hmmm... I think you've got a problem with your ramdisk. There are a couple of issues but these alone don't explain your helloworld test. 1. busybox should be set to root not neeraj 2. you need to setuid busybox for some apps to run (chmod 4775 bin/busybox so that it is srwxrwxr-x-r) 3. I would be wary of Yaghmour's text... it is a little dated and if I remember correctly, he uses some regular expressions to copy "just" the libraries you need to lib; but with newer versions of gcc/glibc these regular expressions don't catch everything Two questions: Can you tell me exactly what your cross-compile command-line looks like? How are compiling helloworld? How are you creating the ramdisk? How do you go from directory structure to ramdisk.image.gz? Ron > > Hi, > > --Ron,Permissions are -rwxrwxr-x for all and owner is neeraj itself. And > shared libraries are present in lib/ . Now I have placed helloworld.elf > (using printf to print helloworld, and linked with -static option) in > bin/ and changed init=/bin/helloworld , again it says cannot execute > /bin/helloworld. > > --John, we are using our custom hardware board, not exactly ML405 but > its more or less similar to ML405, so I cannot use bit file provided for > ML405. Till now we were using xilkernel, but now onward we are planning > to use Linux. For serial console I have no option other than uartlite. > This is how I compiled kernel : > > 1) make ARCH=ppc ml405_defconfig > 2) patched kernel src with EDK(10.1) , so as to change xparameter.h, and > later made some changed in arch/ppc/boot/simple/embed_config.c and > xparameters.h file. > 3) make ARCH=ppc menuconfig , and selected uartlite to be console and > ramdisk as file system. > 4) Created ramdisk (as per Building embedded linux, Karim Yaghmour, > however major and minor number for device nodes is similar to ramdisk > provided at git.xilinx.com) > 5) I have placed this ramdisk in arch/ppc/boot/images > 6) And then issued $make ARCH=ppc CROSS_COMPILE=powerpc-405-linux-gnu- > zImge.initrd > > > --I have downloaded ELDK4.1 and installed it. And when I compile > simple helloworld.c using cross compiler, it says unresolved symbol > 'printf' . Is there anything else to install with ELDK ? > > > --- > Neeraj Garg > > > > In addition to what John wrote, I would also investigate your ramdisk. > I would be sure to check that you have the permissions/owner set correctly > on bin/busybox. Also, I would double check that, if you compiled busybox > with shared libraries, the shared libraries are in the right place > on your ramdisk. > > > Ron > > > > > Hi, > > > > Yes I am using ARCH=ppc (actual line is $make ARCH=ppc > > CROSS_COMPILE=powerpc-405-linux-gnu- zImage.initrd ) for this I have > > placed ramdisk.image.gz in arch/ppc/boot/images. In case of ARCH=powerpc > > I cannot find processor type 405 , in make menuconfig. Thats why i am > > using ARCH=ppc. > > > > --060307000300070002040802 > Content-Type: text/html; charset=ISO-8859-1 > Content-Transfer-Encoding: 7bit > > > > > > > > Hi, > > --Ron,Permissions are -rwxrwxr-x for all and owner is neeraj itself. > And shared libraries are present in lib/ . Now I have placed > helloworld.elf (using printf to print helloworld, and linked with > -static option) in bin/ and changed init=/bin/helloworld , again it > says cannot execute /bin/helloworld. > > --John, we are using our custom hardware board, not exactly ML405 but > its more or less similar to ML405, so I cannot use bit file provided > for ML405. Till now we were using xilkernel, but now onward we are > planning to use Linux. For serial console I have no option other than > uartlite. This is how I compiled kernel : > > 1) make ARCH=ppc ml405_defconfig > 2) patched kernel src with EDK(10.1) , so as to change xparameter.h, > and later made some changed in arch/ppc/boot/simple/embed_config.c and > xparameters.h file. > 3) make ARCH=ppc menuconfig , and selected uartlite to be console and > ramdisk as file system. > 4) Created ramdisk (as per Building embedded linux, Karim Yaghmour, > however major and minor number for device nodes is similar to ramdisk > provided at git.xilinx.com) > 5) I have placed this ramdisk in arch/ppc/boot/images > 6) And then issued $make ARCH=ppc CROSS_COMPILE=powerpc-405-linux-gnu- > zImge.initrd > > > --I have downloaded ELDK4.1 and installed it. And when I compile > simple helloworld.c using cross compiler, it says unresolved symbol > 'printf' . Is there anything else to install with ELDK ? > > > > --- > Neeraj Garg > > > > > In addition to what John wrote, I would also investigate your ramdisk. > > I would be sure to check that you have the permissions/owner set > correctly > > on bin/busybox. Also, I would double check that, if you compiled > bus
Re: How to create custom ramdisk
Under my home page, I have links to the courses I teach. http://rcs.uncc.edu/~rsass/ Follow the links to "Fundamentals of Reconfigurable Computing" specifically, http://rcs.uncc.edu/~rsass/courses/2008-Fall/6890/ There are labs and tutorials there. Under Labs, (Lab 3 in particular) we build Linux-based systems. Lab 3 is has a tutorial based on Secret Lab's git repository and some in-house scripts. The students use RHEL5 machines but in our research lab we Fedora Core and CentOS4 machines. Also we have a wiki with other tutorials. Sorry, it's not very organized (or up to date...) http://www.rcs.uncc.edu/wiki/index.php/Main_Page Ron > > Hi, > > I want to create a custom ramdisk (initrd which can be placed in > arch/ppc/boot/images). Can anybody guide me steps to built the same. My > host have RHEL-4. > > -- > > --- > Neeraj Garg > > > > --080207080004080602090709 > Content-Type: text/html; charset=ISO-8859-1 > Content-Transfer-Encoding: 7bit > > > > > > > > > Hi, > > I want to create a custom ramdisk (initrd which can be placed in > arch/ppc/boot/images). Can anybody guide me steps to built > the same. My host have RHEL-4. > > -- > > --- > Neeraj Garg > > > > > > > --080207080004080602090709-- > > > --===0773240910== > Content-Type: text/plain; charset="us-ascii" > MIME-Version: 1.0 > Content-Transfer-Encoding: 7bit > Content-Disposition: inline > > ___ > Linuxppc-embedded mailing list > Linuxppc-embedded@ozlabs.org > https://ozlabs.org/mailman/listinfo/linuxppc-embedded > --===0773240910==-- > ___ Linuxppc-embedded mailing list Linuxppc-embedded@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-embedded
