Hi All,
I’m testing L4re on a customized I.MX6 Solo-X-4 board. I found that the fiasco kernel (the simple hello world example built into an uimage) cannot be load by the U-boot. It shows “data abort". The reason I guess it is due to different loading address. The U-boot’s default $loadaddr is 0x80800000, but the uimage I build using the L4re buildsystem has:
Load Address: 11000000 Entry Point: 11000000 ==> Installing bootstrap.uimage in image directory Image size(s) in bytes: bootstrap_hello-cfg.elf: 1470844 bootstrap.raw: 1437696 bootstrap.uimage: 1437760 Start address: 0x11000000
I used the RAM_BASE=0x8080000 option to change the starting address when I made the image (as shown below), the boot did not crash but the kernel hangs at “Starting kernel..."
Load Address: 01000000 Entry Point: 01000000 ==> Installing bootstrap.uimage in image directory Image size(s) in bytes: bootstrap_hello-cfg.elf: 1470844 bootstrap.raw: 1437696 bootstrap.uimage: 1437760 Start address: 0x801000000
I think it is still related to the different load address. But I do not know how to fix it. I would really appreciate if someone can point me to a right direction.
Thanks -Dan
On Friday 21. July 2017 20.19.09 Daniel (Xiaolong) Wang wrote:
I’m testing L4re on a customized I.MX6 Solo-X-4 board. I found that the fiasco kernel (the simple hello world example built into an uimage) cannot be load by the U-boot. It shows “data abort". The reason I guess it is due to different loading address.
My own experiences with U-Boot are that it should be able to deduce the correct addresses for loading and execution from the image. However, you need to be aware of the memory that U-Boot is using itself because overwriting that is obviously going to cause problems.
I imagine that it is also worthwhile to know exactly what U-Boot is doing to initialise the machine. My experiences, admittedly with another architecture, involved U-Boot in different forms that did different levels of initialisation depending on where it had been installed (NAND, SD, and so on).
[...]
Start address: 0x801000000
I think it is still related to the different load address. But I do not know how to fix it. I would really appreciate if someone can point me to a right direction.
The above start address looks too "long": it has more than eight digits and won't be a 32-bit value any more. Maybe that doesn't make a difference on your platform, though.
Paul
Hi Daniel,
On 07/21/2017 08:19 PM, Daniel (Xiaolong) Wang wrote:
Hi All,
I’m testing L4re on a customized I.MX6 Solo-X-4 board. I found that the fiasco kernel (the simple hello world example built into an uimage) cannot be load by the U-boot. It shows “data abort". The reason I guess it is due to different loading address. The U-boot’s default $loadaddr is 0x80800000, but the uimage I build using the L4re buildsystem has:
Load Address: 11000000 Entry Point: 11000000 ==> Installing bootstrap.uimage in image directory Image size(s) in bytes: bootstrap_hello-cfg.elf: 1470844 bootstrap.raw: 1437696 bootstrap.uimage: 1437760 Start address: 0x11000000
I used the RAM_BASE=0x8080000 option to change the starting address when I made the image (as shown below), the boot did not crash but the kernel hangs at “Starting kernel..."
Did you provide the correct platform type when you built the image? You should do something like this
make PLATFORM_TYPE=imx6 RAM_SIZE_MB=1024 UART_NR=1 E=hello uimage
Change the values of RAM_SIZE_MB and UART_NR according to your platform.
Best, Matthias.
Load Address: 01000000 Entry Point: 01000000 ==> Installing bootstrap.uimage in image directory Image size(s) in bytes: bootstrap_hello-cfg.elf: 1470844 bootstrap.raw: 1437696 bootstrap.uimage: 1437760 Start address: 0x801000000
I think it is still related to the different load address. But I do not know how to fix it. I would really appreciate if someone can point me to a right direction.
Thanks -Dan
l4-hackers mailing list l4-hackers@os.inf.tu-dresden.de http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers
Hi Matthias and Paul,
Thank you for your response. The board I was working on is a customized IMX6. Turn out the memory is mapped into different region. Thanks to Adam’s help. the problem is solved by editing
l4/mk/platforms/imx6.conf: the UART numbers (as seen from the Linux output): PLATFORM_NAME = "Freescale i.MX6" PLATFORM_ARCH = arm PLATFORM_RAM_BASE = 0x80000000 PLATFORM_RAM_SIZE_MB = 256 PLATFORM_UART_NR = 1
Thank you very much! -Dan
On Jul 23, 2017, at 3:26 PM, Matthias Lange matthias.lange@kernkonzept.com wrote:
Hi Daniel,
On 07/21/2017 08:19 PM, Daniel (Xiaolong) Wang wrote:
Hi All,
I’m testing L4re on a customized I.MX6 Solo-X-4 board. I found that the fiasco kernel (the simple hello world example built into an uimage) cannot be load by the U-boot. It shows “data abort". The reason I guess it is due to different loading address. The U-boot’s default $loadaddr is 0x80800000, but the uimage I build using the L4re buildsystem has:
Load Address: 11000000 Entry Point: 11000000 ==> Installing bootstrap.uimage in image directory Image size(s) in bytes: bootstrap_hello-cfg.elf: 1470844 bootstrap.raw: 1437696 bootstrap.uimage: 1437760 Start address: 0x11000000
I used the RAM_BASE=0x8080000 option to change the starting address when I made the image (as shown below), the boot did not crash but the kernel hangs at “Starting kernel..."
Did you provide the correct platform type when you built the image? You should do something like this
make PLATFORM_TYPE=imx6 RAM_SIZE_MB=1024 UART_NR=1 E=hello uimage
Change the values of RAM_SIZE_MB and UART_NR according to your platform.
Best, Matthias.
Load Address: 01000000 Entry Point: 01000000 ==> Installing bootstrap.uimage in image directory Image size(s) in bytes: bootstrap_hello-cfg.elf: 1470844 bootstrap.raw: 1437696 bootstrap.uimage: 1437760 Start address: 0x801000000
I think it is still related to the different load address. But I do not know how to fix it. I would really appreciate if someone can point me to a right direction.
Thanks -Dan
l4-hackers mailing list l4-hackers@os.inf.tu-dresden.de mailto:l4-hackers@os.inf.tu-dresden.de http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers
-- Matthias Lange, matthias.lange@kernkonzept.com mailto:matthias.lange@kernkonzept.com, +49-351-41 888 614
Kernkonzept GmbH. Sitz: Dresden. Amtsgericht Dresden, HRB 31129. Geschäftsführer: Dr.-Ing. Michael Hohmuth
l4-hackers@os.inf.tu-dresden.de
-
Daniel (Xiaolong) Wang -
Matthias Lange -
Paul Boddie