Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, uImage to. Itried 0x8100 and 0x80008000 and i am still unable to load the kernel off u-boot.? Thanks for all the help! I have got linux booting from u-boot. So it had something to do with one of the device drivers Doing a post mortem on the printk log buffer pointed me to the problem. Thanks, Alfred. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi 2009/5/4 alfred steele alfred.jaq...@gmail.com: Hi Magnus, uImage to. Itried 0x8100 and 0x80008000 and i am still unable to load the kernel off u-boot.? Thanks for all the help! I have got linux booting from u-boot. So it had something to do with one of the device drivers Doing a post mortem on the printk log buffer pointed me to the problem. Excellent! /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, I tend to load my kernels to 0x8100 and not 0x80008000 but I don't know if that really matters. Other than that I think 'run bootcmd_net' should work. Oops i realized i top posted:( I am re-posting again. i checked in the kernel code rpm/BUILD/linux/arch/arm/mach-mx3/Makefile.boot. which probably expects the uboot load the kernel at address 0x80008000(zreladdr) which is probably the kernel image relocated address zreladdr-y := 0x80008000 params_phys-y := 0x8100 (boot arg location which i verified doing a md on u-boot). initrd_phys-y := 0x8080 Are these the correct params or is there any other place you look at for verifying if the u-boot and kernel match on the values of loadaddr, bootarg location ? Thanks ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi 2009/5/1 alfred steele alfred.jaq...@gmail.com: Hi Magnus, I tend to load my kernels to 0x8100 and not 0x80008000 but I don't know if that really matters. Other than that I think 'run bootcmd_net' should work. Oops i realized i top posted:( I am re-posting again. i checked in the kernel code rpm/BUILD/linux/arch/arm/mach-mx3/Makefile.boot. which probably expects the uboot load the kernel at address 0x80008000(zreladdr) which is probably the kernel image relocated address zreladdr-y := 0x80008000 params_phys-y := 0x8100 (boot arg location which i verified doing a md on u-boot). initrd_phys-y := 0x8080 Are these the correct params or is there any other place you look at for verifying if the u-boot and kernel match on the values of loadaddr, bootarg location ? Yes, those values are correct. The boot_params address is also in arch/arm/mach-mx3/mx31pdk.c (mx3_3stack.c if you use Freescale's BSP). /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Yes, those values are correct. The boot_params address is also in arch/arm/mach-mx3/mx31pdk.c (mx3_3stack.c if you use Freescale's BSP). Where do i get the latest version of Freescale BSP? i know this is not the correct forum...but still Thanks, Alfred ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Dear alfred steele, In message 528f13590905010838g1a1dc360gfd5fd4a64e84d...@mail.gmail.com you wrote: Where do i get the latest version of Freescale BSP? i know this is not the correct forum...but still From Freescale? Try www.freescale.com ... Best regards, Wolfgang Denk -- DENX Software Engineering GmbH, MD: Wolfgang Denk Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: w...@denx.de When a program is being tested, it is too late to make design changes. -- Geoffrey James, The Tao of Programming ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
--- On Fri, 5/1/09, alfred steele alfred.jaq...@gmail.com wrote: Where do i get the latest version of Freescale BSP? i know this is not the correct forum...but still http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=i.MX31PDKnodeId=0162468rH311432973ZrDRfpsp=1tab=Design_Tools_Tab (Click on IMX31_PDK_14_LINUX_BSP_R14 ) Regards, Fabio Estevam ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
--- On Fri, 5/1/09, alfred steele alfred.jaq...@gmail.com wrote: Where do i get the latest version of Freescale BSP? i know this is not the correct forum...but still http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=i.MX31PDKnodeId=0162468rH311432973ZrDRfpsp=1tab=Design_Tools_Tab (Click on IMX31_PDK_14_LINUX_BSP_R14 ) Regards, Fabio Estevam ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
(Click on IMX31_PDK_14_LINUX_BSP_R14 ) Thanks Fabio. Do you know what the load address and entry point for creating a uImage are for the mx31 pdk board and where to load the uImage to. Itried 0x8100 and 0x80008000 and i am still unable to load the kernel off u-boot.? Best Regards, Alfred ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi 2009/4/30 alfred steele alfred.jaq...@gmail.com: Hi Magnus, The mainline u-boot + my patches from early April does not contain any driver for the i.MX31 NAND flash controller so that can't autodetect any flash part at all. How does it boot out of NAND then? I am getting confused. By detection, i meant the basic init of the NFC and the NFCBUF and the associated initiliazation to pool the SPL off the NAND which further hand over to the uboot. I realize I wasn't clear. U-boot NAND booting is done in several steps. For the i.MX31 the following steps are involved: 1. i.MX31 reads the first 2048 bytes from NAND. This is the SPL. 2. The SPL contains lowlevel-init and sets up the SDRAM (in assembler). SPL is compiled without SKIP_LOWLEVEL_INIT and without SKIP_RELOATE_UBOOT. 3. The SPL copies the rest of U-boot, a predefined number of blocks, from NAND to SDRAM (nand_boot_mx31.c). This is currently done without run-time detection of the NAND type.CFG_NAND_CHIP_SIZE/CFG_NAND_PAGE_SIZE and others control this. 4. After copying, SPL jumps to U-boot 5. U-boot starts executing (cpu/arm1136/start.S) 6. SKIP_LOWLEVEL_INIT and SKIP_RELOCATE_UBOOT are defined when compiling U-boot since the SDRAM has already been setup by SPL and SPL has already placed U-boot at the correct address (i.e. no relocation needed). 7. Eventually board/freescale/mx31pdk/mx31pdk.c:board_init is called. 8. U-boot main loop is entered. (I hope I got the above right) If U-boot had full NAND support for i.MX31, NAND would be detected somewhere between steps 7 and 8. Remember that there are two separate applications involved above, SPL which is built in nand_spl, and U-boot. The two applications are concatenated into one binary. What's may be confusing is that the same config-file is used both when compiling SPL and U-boot, but config.h is modified a bit. CONFIG_NAND_SPL is defined in nand_spl/board/freescale/mx31pdk/Makefile when it builds the second phase loader. CONFIG_NAND_SPL is fine but CONFIG_SKIP_LOWLEVEL_INIT and CONFIG_SKIP_RELOCATE_UBOOT should not be defined in case we are booting out of NAND, right? See my explanation above. CONFIG_SKIP* is defined when compiling U-boot, but not when compiling SPL. /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, If U-boot had full NAND support for i.MX31, NAND would be detected somewhere between steps 7 and 8. Thanks for the elaborate explanation. I did not find any detailed documentation for the same in the u-boot readme's. See my explanation above. CONFIG_SKIP* is defined when compiling U-boot, but not when compiling SPL. That explains why i swa the #warning spit out. So you mean the CONFIG_SKIP_LOWLEVEL_INIT and CONFIG_SKIP_RELOCATE_UBOOT get toggled during compile time only. I am still struggling to get the NAND boot working. I am not even seeing U-boot spit out the first line on the console U-boot version no when i power up subsequently. This is what i did to build and run. i) make mx31_pdk_nand_config ii) make Take U-boot-nand.bin and burn it at physical offset 0x0 of the NAND using Redboot loaded via JTAG. When i reset, i see nothing on the console. I may be missing something which applying the patches then. Remote gdb does not allow me to debug. If i put printfs in _start_armboot in board.c, looks like its not ven entering that function. Has the standard output and the serial console dev been initiliazed before this point i.e start_armboot? Have you got a chance to post the patches to the mainline yet? Thanks, Alfred. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Have you got a chance to post the patches to the mainline yet? No, but as I stated earlier the only merge problem I got was in the top Makefile and that was easy to fix. I need more time to post the patches since I'm taking care of some comments that were posted to Maxim's patches (which I'm using as a base). /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, I need more time to post the patches since I'm taking care of some comments that were posted to Maxim's patches (which I'm using as a base). That's fine. Which linker script gets actually used for the layout in in case of nand_spl. Is it u-boot-nand.lds or u-boot.lds.? Thanks. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi
That's fine.
Which linker script gets actually used for the layout in in case of
nand_spl. Is it u-boot-nand.lds or u-boot.lds.?
I am going back to the original intent of this thread, booting off the
linux kernel using TFTP. Here's is a dump of my U-boot env. Does
it look good enough for booting a linux kernel. WHat could be the
reason that the( md of the printk log does not show up the console
=ttymxc0,115200 correctly.
It looks like its. just ttymxc0,1152 as evident from the memory
dump i had sent ealier in this thread.
=printenv
baudrate=115200
netdev=eth0
uboot=u-boot.bin
nfsroot=/tools/rootfs/rootfs-2.6.24
bootargs_base=setenv bootargs console=ttymxc0,115200
bootargs_nfs=setenv bootargs ${bootargs} root=/dev/nfs ip=dhcp nfsroot=${serveri
p}:${nfsroot} rw
bootcmd=run bootcmd_net
bootcmd_net=run bootargs_base bootargs_nfs; tftpboot ${loadaddr} ${tftp_server}:
${kernel}; bootm
tftp_server=206.44.18.25
kernel=uImage_spin2
loadaddr=0x80008000
bootdelay=5
filesize=196828
fileaddr=8100
ipaddr=206.44.18.31
serverip=206.44.18.25
loadaddr=0x80008000
bootargs=root=/dev/nfs rw
stdin=serial
stdout=serial
stderr=serial
Thanks.
___
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi
2009/4/30 alfred steele alfred.jaq...@gmail.com:
Hi
That's fine.
Which linker script gets actually used for the layout in in case of
nand_spl. Is it u-boot-nand.lds or u-boot.lds.?
I am going back to the original intent of this thread, booting off the
linux kernel using TFTP. Here's is a dump of my U-boot env. Does
it look good enough for booting a linux kernel. WHat could be the
reason that the( md of the printk log does not show up the console
=ttymxc0,115200 correctly.
It looks like its. just ttymxc0,1152 as evident from the memory
dump i had sent ealier in this thread.
=printenv
baudrate=115200
netdev=eth0
uboot=u-boot.bin
nfsroot=/tools/rootfs/rootfs-2.6.24
bootargs_base=setenv bootargs console=ttymxc0,115200
bootargs_nfs=setenv bootargs ${bootargs} root=/dev/nfs ip=dhcp
nfsroot=${serveri
p}:${nfsroot} rw
bootcmd=run bootcmd_net
bootcmd_net=run bootargs_base bootargs_nfs; tftpboot ${loadaddr}
${tftp_server}:
${kernel}; bootm
tftp_server=206.44.18.25
kernel=uImage_spin2
loadaddr=0x80008000
bootdelay=5
filesize=196828
fileaddr=8100
ipaddr=206.44.18.31
serverip=206.44.18.25
loadaddr=0x80008000
bootargs=root=/dev/nfs rw
stdin=serial
stdout=serial
stderr=serial
I tend to load my kernels to 0x8100 and not 0x80008000 but I don't
know if that really matters. Other than that I think 'run bootcmd_net'
should work.
/Magnus
___
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi 2009/4/28 alfred steele alfred.jaq...@gmail.com: Hi Magnus, But when i tried burning u-boot-nand.bin to the base address and boot , i get only one line printed on the console which is U-Boot 2009.01 (Apr 28 2009 - 15:27:19) If it comes that far it looks like a lot is working. Don't know why the rest isn't though. Looks like somethings messed up in the NAND detection in the first place. I am using Redboot FIS to burn the u-boot-nand.bin . I know it ignores bad blocks but i had used successfully to burn in the past. I do not know the procedure for using u-boot to burn u-boot. Let me know if you want me to try it. I suppose Redboot should be fine to use, haven't tried it much myself though. I use a JTAG programmer for that. /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Alfred, --- On Tue, 4/28/09, alfred steele alfred.jaq...@gmail.com wrote: Looks like somethings messed up in the NAND detection in the first place. I am using Redboot FIS to burn the u-boot-nand.bin . I know it ignores bad blocks but i had used successfully to burn in the past. I do not know the procedure for using u-boot to burn u-boot. Let me know if you want me to try it. I use ATK tool to flash Redboot and U-boot into the MX31PDK NAND. You can get it from Freescale website. Regards, Fabio Estevam ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, Thanks. If it comes that far it looks like a lot is working. Don't know why the rest isn't though. At this point, it should be showing up the processor info. I am using a custom board with a different flash part. Samsung(KR series), and i guess it is part of nand_id table as i have used that to get my other version (patches form internal fresclae git) up and running. But i believe the overall flash detection happens automatically once you have the id in place. Correct me if i am wrong. Are there any specific files to look into from here. Another thing i noticed is that the config file in my build was somehow defining CONFIG_SKIP_LOWLEVEL_INIT and CONFIG_SKIP_RELOCATE_UBOOT which should not be the case. I put a #warning to identify that. What could be wrong? //#if defined(CONFIG_NAND_U_BOOT) !defined(CONFIG_NAND_SPL) //#warning defined(CONFIG_NAND_U_BOOT) !defined(CONFIG_NAND_SPL) //#define CONFIG_SKIP_LOWLEVEL_INIT //#define CONFIG_SKIP_RELOCATE_UBOOT //#endif Thanks, Alfred. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
At this point, it should be showing up the processor info. I am using a custom board with a different flash part. Samsung(KR series), and i guess it is part of nand_id table as i have used that to get my other version (patches form internal fresclae git) up and running. But i believe the overall flash detection happens automatically once you have the id in place. Correct me if i am wrong. A custom board with different flash and to me it's very confusing which patches/version of u-boot you're and referring to in different mails. The mainline u-boot + my patches from early April does not contain any driver for the i.MX31 NAND flash controller so that can't autodetect any flash part at all. Are there any specific files to look into from here. Follow the startup sequence in lib_arm/board.c:start_armboot and place printf at strategic locations (unless you have a debugger which you can single step in). Another thing i noticed is that the config file in my build was somehow defining CONFIG_SKIP_LOWLEVEL_INIT and CONFIG_SKIP_RELOCATE_UBOOT which should not be the case. I put a #warning to identify that. What could be wrong? //#if defined(CONFIG_NAND_U_BOOT) !defined(CONFIG_NAND_SPL) //#warning defined(CONFIG_NAND_U_BOOT) !defined(CONFIG_NAND_SPL) //#define CONFIG_SKIP_LOWLEVEL_INIT //#define CONFIG_SKIP_RELOCATE_UBOOT //#endif CONFIG_NAND_SPL is defined in nand_spl/board/freescale/mx31pdk/Makefile when it builds the second phase loader. /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, The mainline u-boot + my patches from early April does not contain any driver for the i.MX31 NAND flash controller so that can't autodetect any flash part at all. How does it boot out of NAND then? I am getting confused. By detection, i meant the basic init of the NFC and the NFCBUF and the associated initiliazation to pool the SPL off the NAND which further hand over to the uboot. Follow the startup sequence in lib_arm/board.c:start_armboot and place printf at strategic locations (unless you have a debugger which you can single step in). How do i debug after u-boot relocates itself in RAM. I am getting this on using u-boot as the ELF file with arm-elf-gdb. I remember when i had the BDI3000 handy, i was getting a segfault when doing a remote target IP:2001. Please suggest. CONFIG_NAND_SPL is defined in nand_spl/board/freescale/mx31pdk/Makefile when it builds the second phase loader. CONFIG_NAND_SPL is fine but CONFIG_SKIP_LOWLEVEL_INIT and CONFIG_SKIP_RELOCATE_UBOOT should not be defined in case we are booting out of NAND, right? Thanks. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi 2009/4/27 alfred steele alfred.jaq...@gmail.com: Dear Magnus, Thanks for the reply! And we need to know which U-boot patches you're using to boot the PDK board. I am using the internal git tree code supplied to me by freescale. The tarball is called uboot-imx-imx_v2009.01.tar.gz. I can boot uboot out of NAND successfully using that as the code base. I had started using this even before Magnus had posted the patches and i had to write a patch for the MMC/SD driver on top of that . To add to that , i tried downloading the mainline using GIT but it seems the corporate firewall/proxy we have blocks it. Due to the shortage of time, i continued to use the Freescale supplied code. It would be great if you can supply me the code in a tarball(mainline git) with the patches you posted if you want me to try it out. There's a 'snapshot' function in git-web (http://git.denx.de/?p=u-boot.git;a=tree) which will produce a tarball of the current tree. My patches are in the mailing list archives so you should be able to download and apply them as well. In your original post you use the go command to boot the kernel, I use the bootm command now. My sole intention of using go was to check whether or not zImage( instead of uImage) boots up as a generic application. I tried verify the bootags in the U-boot environment. I chacked on the machine id which gets registered with the kernel. They all seemed to be the expected values. Coming back to I am not sure about ,was when i tried dumping the address(0x8100) which supposedly stores the bootargs, i get the following : = md 0x8100 8100: c100 c100 c108 c108 8110: c110 c110 c118 c118 8120: 92492492 24924924 0149 736e6f63 .$I.$I.$I...cons 8130: 3d656c6f 6d797474 2c306378 32353131 ole=ttymxc0,1152 8140: c140 c140 c148 c148 �...@...@...H...H... It looks like the console is not set correctly to 115200, assuming its in ASCII. Any clues as to what could be gone wrong? No idea. Your printk dump from the first mail seems to indicate that the kernel was able to boot, but for some reason it doesn't print on the serial port. And I can't see any init of the serial port in that log either. Seems like a problem with your kernel or its configuration. /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, Thanks! Here's a 'snapshot' function in git-web (http://git.denx.de/?p=u-boot.git;a=tree) which will produce a tarball of the current tree. Before that, i actually managed to git clone to a windows machine(at my home) and export the u-boot tarball to my office linux host where its destined to build. I got quite a few HUNK failures when i applied your patch to that so I hand edited MAKEALL and Makefile and some other to restore parity. But then on doing make mx31_pdk_config, i got the following: $ make mx31ads_config : invalid option make: *** [mx31ads_config] Error 1 So looks like somewhere out there , some file has some weird windows inherited character/s e.g CRLF characters while doing a git cone although i am not sure I am unable to identify the file. Any idea. Thanks, Alfred. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi 2009/4/28 alfred steele alfred.jaq...@gmail.com: Hi Magnus, Thanks! But then on doing make mx31_pdk_config, i got the following: $ make mx31ads_config : invalid option make: *** [mx31ads_config] Error 1 So looks like somewhere out there , some file has some weird windows inherited character/s e.g CRLF characters while doing a git cone although i am not sure I am unable to identify the file. Any idea. No, but it seems that the problem is in the patched Makefile since it says mx31ads_config instead of mx31_pdk_config. /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, No, but it seems that the problem is in the patched Makefile since it says mx31ads_config instead of mx31_pdk_config. I tried this on all the RULES as i was getting the same results for the rule mx31_pdk_config. Apologize for for posting the wrong one. This problem got resolved when i useda tarball from the snapshot feature of the GIT instead. So , in order to boot out of NAND, all i need to do is use the image generated out mx31pdk_nand_config instead of the regular mx31_pdk_config. Right? Thanks. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, So , in order to boot out of NAND, all i need to do is use the image generated out mx31pdk_nand_config instead of the regular mx31_pdk_config. Right? Another thing which confuses me is the file nand_boot_mx31.c . Is it a symbolik link which points to the genric nand_boot.c or is it a file by itself. Thanks. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
So , in order to boot out of NAND, all i need to do is use the image generated out mx31pdk_nand_config instead of the regular mx31_pdk_config. Right? Yes. /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi 2009/4/28 alfred steele alfred.jaq...@gmail.com: Hi Magnus, So , in order to boot out of NAND, all i need to do is use the image generated out mx31pdk_nand_config instead of the regular mx31_pdk_config. Right? Another thing which confuses me is the file nand_boot_mx31.c . Is it a symbolik link which points to the genric nand_boot.c or is it a file by itself. The nand_spl/nand_boot_mx31.c is a regular file if that's the one you mean. /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, Thanks again! The nand_spl/nand_boot_mx31.c is a regular file if that's the one you mean. The reason i am asking is I am getting a build error related to this file after i apply your patches. When i do a ll(ls -l) on the file, i get $ll u-boot-head/u-boot/nand_spl/board/freescale/mx31pdk/nand_boot_mx31.c u-boot-head/u-boot/nand_spl/board/freescale/mx31pdk/nand_boot_mx31.c - u-boot-head/u-boot/nand_spl/board/freescale/mx31pdk/nand_boot_mx31.c And there is NO real nand_boot_mx31.c from the patches i picked up from the mail archives. Perhaps i am looking at your older post. Can you point me to the exact location for the basic support and the NAND spl support. As i am writing this email, . i found a new one http://www.mail-archive.com/u-boot@lists.denx.de/msg11732.html. IS that the correct one? Thanks. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
2009/4/28 alfred steele alfred.jaq...@gmail.com: Hi Magnus, Thanks again! The nand_spl/nand_boot_mx31.c is a regular file if that's the one you mean. The reason i am asking is I am getting a build error related to this file after i apply your patches. When i do a ll(ls -l) on the file, i get $ll u-boot-head/u-boot/nand_spl/board/freescale/mx31pdk/nand_boot_mx31.c u-boot-head/u-boot/nand_spl/board/freescale/mx31pdk/nand_boot_mx31.c - u-boot-head/u-boot/nand_spl/board/freescale/mx31pdk/nand_boot_mx31.c Hmm, I don't have my development machine up and running at the moment. I'll check later. And there is NO real nand_boot_mx31.c from the patches i picked up from the mail archives. Perhaps i am looking at your older post. Can you point me to the exact location for the basic support and the NAND spl support. As i am writing this email, . i found a new one http://www.mail-archive.com/u-boot@lists.denx.de/msg11732.html. IS that the correct one? Yes, that's the latest one I've posted (I've done minor updates locally but haven't tested those yet). /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Dear Magnus, Yes, that's the latest one I've posted (I've done minor updates locally but haven't tested those yet). I guess that's there are two patches for NAND_SPL, one which just adds the SPL framework and the other which creates the mx31pdk_nand.c and changes to start.S. Please correct me if i am wrong. Since the cpu/arm1136/start.S has changed in a major way since April 02, since this is a generic file, i am getting a BIG .rej file for the HUNKS which failed. I can hand edit start.S as i have done for the rest of the HUNK failures which were minor. But is there a way i can get the snapshot as of April2, so that i can apply this patch right away. This would greatly reduce the effort. Or you suggest sticking to the latest and greatest mainline. Thanks. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Or you suggest sticking to the latest and greatest mainline. Alternatively, will use the .rej to generate the new lot of patches. That seems the best route. Thanks. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi 2009/4/28 alfred steele alfred.jaq...@gmail.com: Dear Magnus, Yes, that's the latest one I've posted (I've done minor updates locally but haven't tested those yet). I guess that's there are two patches for NAND_SPL, one which just adds the SPL framework and the other which creates the mx31pdk_nand.c and changes to start.S. Something like that, in total there's four or five patches in that series I think. Please correct me if i am wrong. Since the cpu/arm1136/start.S has changed in a major way since April 02, since this is a generic file, i am getting a BIG .rej file for the HUNKS which failed. I can hand edit start.S as i have done for the rest of the HUNK failures which were minor. But is there a way i can get the snapshot as of April2, so that i can apply this patch right away. This would greatly reduce the effort. Or you suggest sticking to the latest and greatest mainline. If there are difficulties in applying the patches you could get a snapshot of commit dfc91c33957c95da34e3888dc87912d5c15a7603 which is the tree that I based my patches on. That's probably the quickest solution. I'll update the patches to the current u-boot tree, but I that can take a couple of days. Please put a question mark at the end of sentences that actually contains a question! /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Please correct me if i am wrong. Since the cpu/arm1136/start.S has changed in a major way since April 02, since this is a generic file, i am getting a BIG .rej file for the HUNKS which failed. I can hand edit start.S as i have done for the rest of the HUNK failures which were minor. But is there a way i can get the snapshot as of April2, so that i can apply this patch right away. This would greatly reduce the effort. Or you suggest sticking to the latest and greatest mainline. I just applied the patch series to the current u-boot tip and the only problem was the top Makefile but that was easy to fix. No warnings from start.S. /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Magnus, I just applied the patch series to the current u-boot tip and the only problem was the top Makefile but that was easy to fix. No warnings from start.S. I was able to build successfully after hand editing some of the hunks in start.S. Even after using the commit snapshot you had referred to. I don't know where he disconnect is. Perhaps i am still not using the latest set of patches. Towards the end of my build , i get cat nand_spl/u-boot-spl-aligned.bin u-boot.bin u-boot-nand.bin #cat nand_spl/u-boot-spl-16k.bin u-boot.bin u-boot-nand.bin But when i tried burning u-boot-nand.bin to the base address and boot , i get only one line printed on the console which is U-Boot 2009.01 (Apr 28 2009 - 15:27:19) Looks like somethings messed up in the NAND detection in the first place. I am using Redboot FIS to burn the u-boot-nand.bin . I know it ignores bad blocks but i had used successfully to burn in the past. I do not know the procedure for using u-boot to burn u-boot. Let me know if you want me to try it. Thanks. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
I think the first step would be to get MX31PDK into mainline U-boot. Are you willing to help on that? Of course, as long as i have it up and running and tested on my hardware. But in order to do that, i would need your help in answering my earlier question. Thanks for your help! -Alfred. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi 2009/4/27 alfred steele alfred.jaq...@gmail.com: I think the first step would be to get MX31PDK into mainline U-boot. Are you willing to help on that? Of course, as long as i have it up and running and tested on my hardware. But in order to do that, i would need your help in answering my earlier question. And we need to know which U-boot patches you're using to boot the PDK board. In your original post you use the go command to boot the kernel, I use the bootm command. Using my latest set of patches (posted here a couple of weeks ago) I can boot Linux on the PDK using ethernet+tftp. The kernel (uImage) was compiled using Linux v2.6.24 + Freescale's patches from one of their BSPs (don't remember which actually, 1.3 or 1.4). /Magnus ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Dear Magnus,
Thanks for the reply!
And we need to know which U-boot patches you're using to boot the PDK
board.
I am using the internal git tree code supplied to me by freescale. The
tarball is called uboot-imx-imx_v2009.01.tar.gz. I can boot uboot
out of NAND successfully using that as the code base. I had started
using this even before Magnus had posted the patches and i had to
write a patch for the MMC/SD driver on top of that . To add to that ,
i tried downloading the mainline using GIT but it seems the corporate
firewall/proxy we have blocks it. Due to the shortage of time, i
continued to use the Freescale supplied code. It would be great if you
can supply me the code in a tarball(mainline git) with the patches you
posted if you want me to try it out.
In your original post you use the go command to boot the
kernel,
I use the bootm command now. My sole intention of using go was to
check whether or not zImage( instead of uImage) boots up as a generic
application. I tried verify the bootags in the U-boot environment. I
chacked on the machine id which gets registered with the kernel. They
all seemed to be the expected values.
Coming back to I am not sure about ,was when i tried dumping the
address(0x8100) which supposedly stores the bootargs, i get the
following :
= md 0x8100
8100: c100 c100 c108 c108
8110: c110 c110 c118 c118
8120: 92492492 24924924 0149 736e6f63.$I.$I.$I...cons
8130: 3d656c6f 6d797474 2c306378 32353131ole=ttymxc0,1152
8140: c140 c140 c148 c148@@...h...h...
It looks like the console is not set correctly to 115200, assuming its
in ASCII. Any clues as to what could be gone wrong?
Another quick question is do i have to supply the mem=XX argument
if my CONFIG_SETUP_MEMORY_TAGS is set in include/configs/mx31_pdk.h
BTW, following is an extract from the uboot linker script -u-boot.lds
which i had edited. Please let me know if you think something's wrong
with it.
OUTPUT_ARCH(arm)
ENTRY(reset)
SECTIONS
{
. = 0x;
. = ALIGN(4);
.text :
{
/* WARNING - the following is hand-optimized to fit within*/
/* the sector layout of our flash chips! XXX FIXME XXX */
* (.text.head)/* arm reset handler */
* (.text.init)/* lowlevel initial */
* (.text.load)/* nand copy and load */
* (.text.setup)
board/freescale/mx31_3stack/libmx31_3stack.a (.text)
lib_arm/libarm.a (.text)
net/libnet.a (.text)
drivers/mtd/libmtd.a (.text)
. = DEFINED(env_offset) ? env_offset : .;
common/env_embedded.o(.text)
*(.text)
}
. = ALIGN(4);
-Alfred.
___
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi Fabio, Are the load address and other params same for PDK and the ADS board. I intend to do the same as what you intended to aith sucess. I saw your post her : http://www.nabble.com/Loading-a-kernel-on-MX31ADS-using-U-boot-td16642427.html Please let me know how were able to sucessfully load the kernel. My bootup sequence too hangs after Uncompressing kernel...done booting the kernel. Thanks. On Fri, Apr 24, 2009 at 1:23 PM, Fabio Estevam fabioeste...@yahoo.com wrote: Hi Alfred, --- On Thu, 4/23/09, alfred steele alfred.jaq...@gmail.com wrote: I have attached the printk circular buffer log. smc911x: initializing smc911x: detected NULL controller smc911x: phy initialized smc911x: MAC 92:92:92:bb:bb:bb TFTP from server 206.44.18.25; our IP address is 206.44.18.31 Filename 'zImage_spin2_23march'. Is this really an uImage? The name of the file starts with zImage, so it looks confusing. Can you confirm? How are you generating the uImage? Assuming you are using MX31PDK Freescale BSP, you can do as follows: cd rpm/BUILD/linux make ARCH=arm uImage (You need to have mkimage installed) Which U-boot version are you using? Regards, Fabio Estevam ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Dear alfred steele, A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? (see http://www.netmeister.org/news/learn2quote.html) In message 528f13590904262208s66b17bdekac218fba7994d...@mail.gmail.com you wrote: Hi Fabio, Are the load address and other params same for PDK and the ADS board. I intend to do the same as what you intended to aith sucess. I saw your post her : http://www.nabble.com/Loading-a-kernel-on-MX31ADS-using-U-boot-td16642427.html Please let me know how were able to sucessfully load the kernel. My bootup sequence too hangs after Uncompressing kernel...done booting the kernel. ... [Full quote deleted.] You top post. You full quote. And you post off topic stuff (Linux questions are off topic on the U-Boot list). Note that you cannot win any price for the maximum number of netiquette violations in a single posting. Wolfgang Denk -- DENX Software Engineering GmbH, MD: Wolfgang Denk Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: w...@denx.de It is common sense to take a method and try it. If it fails, admit it frankly and try another. But above all, try something. - Franklin D. Roosevelt ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] U-boot -printk kernel crash dump using md on PDK
Dear Wolfgang, You top post. You full quote. Sincere Apologies. I better watch out before i hit the send button. Best Regards. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
[U-Boot] U-boot -printk kernel crash dump using md on PDK
Hi All, I am using mx31 pdk and uImage generated by LTIB. Basically, my problem is my kernel gets stuck after printing the Done...Booting the kernel. Please the responses below. Upon inspecting the printk log by doing a memory dump of the _buf_log symbol in System.map(i translated it to the physical address before dumping it ). It seems before the console got initialized, the bootup was successful to a certain extent, but i am unable to interpret anything conclusive from the logs. ANy hints as to what might be happening? I am guessing that unexpected RAM regions are getting overwritten after uncompression happens or the sdram hasn't been initialized fully by U-boot. I have attached the printk circular buffer log. smc911x: initializing smc911x: detected NULL controller smc911x: phy initialized smc911x: MAC 92:92:92:bb:bb:bb TFTP from server 206.44.18.25; our IP address is 206.44.18.31 Filename 'zImage_spin2_23march'. Load address: 0x8100 Loading: * # # done Bytes transferred = 1665000 (1967e8 hex) = go 0x8100 ## Starting application at 0x8100 ... Uncompressing Linux. ... done, booting the kernel. PRINTK log == md 80350f68 5000 50 80350f68: 4c3e353c 78756e69 72657620 6e6f69735Linux version 80350f78: 362e3220 2d34322e 2d303431 65383667 2.6.24-140-g68e 80350f88: 34623462 79742820 406f7561 656c7954b4b4 (ty...@tyle 80350f98: 4c575372 694c6261 3278756e 2e4d412erSWLabLinux2.AM. 80350fa8: 50524f43 4952502e 28202956 20636367CORP.PRIV) (gcc 80350fb8: 73726576 206e6f69 2e312e34 23202932version 4.1.2) # 80350fc8: 52502033 504d4545 68542054 704120753 PREEMPT Thu Ap 80350fd8: 33322072 3a353120 323a3131 44432039r 23 15:11:29 CD 80350fe8: 30322054 3c0a3930 50433e34 41203a55T 2009.4CPU: A 80350ff8: 36764d52 6d6f632d 69746170 20656c62RMv6-compatible 80351008: 636f7270 6f737365 345b2072 62373031processor [4107b 80351018: 5d343633 76657220 6f697369 2034206e364] revision 4 80351028: 4d524128 45543676 202c294a 303d7263(ARMv6TEJ), cr=0 80351038: 33356530 0a663738 4d3e343c 696863610e5387f.4Machi 80351048: 203a656e 65657246 6c616373 584d2065ne: Freescale MX 80351058: 4d2f3133 20323358 74532d33 206b636131/MX32 3-Stack 80351068: 72616f42 343c0a64 6d654d3e 2079726fBoard.4Memory 80351078: 696c6f70 203a7963 20434345 61736964policy: ECC disa 80351088: 64656c62 6144202c 63206174 65686361bled, Data cache 80351098: 69727720 61626574 3c0a6b63 6e4f3e37 writeback.7On 803510a8: 646f6e20 20302065 61746f74 6761706c node 0 totalpag 803510b8: 203a7365 36373233 373c0a38 4420203ees: 32768.7 D 803510c8: 7a20414d 3a656e6f 20383420 65676170MA zone: 48 page 803510d8: 73752073 66206465 6d20726f 616d6d65s used for memma 803510e8: 373c0a70 4420203e 7a20414d 3a656e6fp.7 DMA zone: 803510f8: 70203020 73656761 73657220 65767265 0 pages reserve 80351108: 373c0a64 4420203e 7a20414d 3a656e6fd.7 DMA zone: 80351118: 39303620 61702036 2c736567 46494c20 6096 pages, LIF 80351128: 6162204f 3a686374 373c0a30 4e20203eO batch:0.7 N 80351138: 616d726f 6f7a206c 203a656e 20383032ormal zone: 208 80351148: 65676170 73752073 66206465 6d20726fpages used for m 80351158: 616d6d65 373c0a70 4e20203e 616d726femmap.7 Norma 80351168: 6f7a206c 203a656e 31343632 61702036l zone: 26416 pa 80351178: 2c736567 46494c20 6162204f 3a686374ges, LIFO batch: 80351188: 373c0a37 4d20203e 6261766f 7a20656c7.7 Movable z 80351198: 3a656e6f 70203020 73656761 65737520one: 0 pages use 803511a8: 6f662064 656d2072 70616d6d 3e343c0ad for memmap.4 803511b8: 30555043 2044203a 54504956 69727720CPU0: D VIPT wri 803511c8: 622d6574 206b6361 68636163 343c0a65te-back cache.4 803511d8: 5550433e 49203a30 63616320 203a6568CPU0: I cache: 803511e8: 38333631 79622034 2c736574 7373612016384 bytes, ass 803511f8: 6169636f 69766974 34207974 3233202cociativity 4, 32 80351208: 74796220 696c2065 2c73656e 38323120 byte lines, 128 80351218: 74657320 343c0a73 5550433e 44203a30 sets.4CPU0: D 80351228: 63616320 203a6568 38333631 79622034 cache: 16384 by 80351238: 2c736574 73736120 6169636f 69766974tes, associativi 80351248: 34207974 3233202c 74796220 696c2065ty 4, 32 byte li 80351258: 2c73656e 38323120 74657320 343c0a73nes, 128 sets.4 80351268: 6975423e 3120746c 6e6f7a20 73696c65Built 1 zonelis 80351278: 69207374 6f5a206e 6f20656e 72656472ts in Zone order 80351288: 6f6d202c 696c6962 67207974 70756f72, mobility group 80351298: 20676e69 202e6e6f 746f5420 70206c61ing on. Total p 803512a8: 73656761 3233203a 0a323135 4b3e353cages: 32512.5K 803512b8: 656e7265 6f63206c 6e616d6d 696c2064ernel command li 803512c8: 203a656e 6e696f6e 64727469 6e6f6320ne: noinitrd con
