Updates to u-boot script for Raspberry Pi 4b devices to properly configure the serial console and to choose the proper boot address to be used on newer board revisions. With these changes in place, I'm able to fully boot into a serial console in u-boot. The serial console configuration was derived from the Grub configuration, while the boot address changes were taken from the IPFire Raspberry 4b documentation page: https://www.ipfire.org/docs/hardware/arm/rpi/four Per the conversation in the forums, I don't know why the boot address changes are required. I was simply trying to automate those changes such that downloaded builds will work out of the box. https://community.ipfire.org/t/hang-with-kernel-starting-on-raspberry-pi/115... NOTE: The use of the different parameter options to the test command to minimize the need for deeply nested conditionals. I did try more bash-shell like syntax, but u-boot didn't seem to like that. The operators used here were found in the u-boot source. Tested-by: Craig Setera craigjunk@setera.org Signed-off-by: Craig Setera craigjunk@setera.org --- config/u-boot/boot.cmd | 27 ++++++++++++++++++++++++--- 1 file changed, 24 insertions(+), 3 deletions(-) diff --git a/config/u-boot/boot.cmd b/config/u-boot/boot.cmd index a27996780..eedd5776d 100644 --- a/config/u-boot/boot.cmd +++ b/config/u-boot/boot.cmd @@-51,7 +51,11 @@if test "${SERIAL-CONSOLE}" = "ON"; then if test "${fdtfile}" = "broadcom/bcm2837-rpi-3-b.dtb"; then setenv console ttyS1,115200n8; else -setenv console ttyAMA0,115200n8; +if test "${fdtfile}" = "broadcom/bcm2711-rpi-4-b.dtb"; then +setenv console ttyS0,115200n8; +else +setenv console ttyAMA0,115200n8; +fi; fi; fi; else @@-95,11 +99,28 @@else setenv ramdisk_addr -; fi; +# +# Handle Raspberry Pi 4 address differences +# https://www.ipfire.org/docs/hardware/arm/rpi/four +# +setenv booti_fdt_addr ${fdt_addr_r}; +if test "${board}" -eq "rpi" -a ${cpu} -eq "armv8" -a "${fdtfile}" -eq "broadcom/bcm2711-rpi-4-b.dtb"; then +# Hardware revision 1.4 +if test ${board_revision} -eq 0xB03114 -o ${board_revision} -eq 0xC03114 -o ${board_revision} -eq 0xD03114; then +setenv booti_fdt_addr ${fdt_addr}; +fi + +# Hardware revision 1.5 +if test ${board_revision} -eq 0xB03115 -o ${board_revision} -eq 0xC03115 -o ${board_revision} -eq 0xD03115; then +setenv booti_fdt_addr ${fdt_addr}; +fi +fi; + bootz ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; -booti ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; +booti ${kernel_addr_r} ${ramdisk_addr} ${booti_fdt_addr}; bootz ${kernel_addr_r} - ${fdt_addr_r}; -booti ${kernel_addr_r} - ${fdt_addr_r}; +booti ${kernel_addr_r} - ${booti_fdt_addr}; # Recompile with: # mkimage -C none -A arm -T script -d /boot/boot.cmd /boot/boot.scr
Apologies. I just realized that I forgot to prefix the subject with [Patch]. Should I resend with a new subject or can this be considered without that step?
On 5/15/24 7:17 PM, Craig Setera wrote:
Updates to u-boot script for Raspberry Pi 4b devices to properly configure the serial console and to choose the proper boot address to be used on newer board revisions. With these changes in place, I'm able to fully boot into a serial console in u-boot. The serial console configuration was derived from the Grub configuration, while the boot address changes were taken from the IPFire Raspberry 4b documentation page: https://www.ipfire.org/docs/hardware/arm/rpi/four Per the conversation in the forums, I don't know why the boot address changes are required. I was simply trying to automate those changes such that downloaded builds will work out of the box. https://community.ipfire.org/t/hang-with-kernel-starting-on-raspberry-pi/115...
NOTE: The use of the different parameter options to the test command to minimize the need for deeply nested conditionals. I did try more bash-shell like syntax, but u-boot didn't seem to like that. The operators used here were found in the u-boot source. Tested-by: Craig Setera craigjunk@setera.org Signed-off-by: Craig Setera craigjunk@setera.org
config/u-boot/boot.cmd | 27 ++++++++++++++++++++++++--- 1 file changed, 24 insertions(+), 3 deletions(-) diff --git a/config/u-boot/boot.cmd b/config/u-boot/boot.cmd index a27996780..eedd5776d 100644 --- a/config/u-boot/boot.cmd +++ b/config/u-boot/boot.cmd @@-51,7 +51,11 @@if test "${SERIAL-CONSOLE}" = "ON"; then if test "${fdtfile}" = "broadcom/bcm2837-rpi-3-b.dtb"; then setenv console ttyS1,115200n8; else -setenv console ttyAMA0,115200n8; +if test "${fdtfile}" = "broadcom/bcm2711-rpi-4-b.dtb"; then +setenv console ttyS0,115200n8; +else +setenv console ttyAMA0,115200n8; +fi; fi; fi; else @@-95,11 +99,28 @@else setenv ramdisk_addr -; fi; +# +# Handle Raspberry Pi 4 address differences +# https://www.ipfire.org/docs/hardware/arm/rpi/four +# +setenv booti_fdt_addr ${fdt_addr_r}; +if test "${board}" -eq "rpi" -a ${cpu} -eq "armv8" -a "${fdtfile}" -eq "broadcom/bcm2711-rpi-4-b.dtb"; then +# Hardware revision 1.4 +if test ${board_revision} -eq 0xB03114 -o ${board_revision} -eq 0xC03114 -o ${board_revision} -eq 0xD03114; then +setenv booti_fdt_addr ${fdt_addr}; +fi
+# Hardware revision 1.5 +if test ${board_revision} -eq 0xB03115 -o ${board_revision} -eq 0xC03115 -o ${board_revision} -eq 0xD03115; then +setenv booti_fdt_addr ${fdt_addr}; +fi +fi;
bootz ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; -booti ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; +booti ${kernel_addr_r} ${ramdisk_addr} ${booti_fdt_addr}; bootz ${kernel_addr_r} - ${fdt_addr_r}; -booti ${kernel_addr_r} - ${fdt_addr_r}; +booti ${kernel_addr_r} - ${booti_fdt_addr}; # Recompile with: # mkimage -C none -A arm -T script -d /boot/boot.cmd /boot/boot.scr
Hello Craig,
Thanks again for the patch. The tag in the subject line isn’t that important, but the email you sent has again removed all whitespace which makes it impossible to read or apply the patch.
-Michael
On 16 May 2024, at 18:38, Craig Setera craigjunk@setera.org wrote:
Apologies. I just realized that I forgot to prefix the subject with [Patch]. Should I resend with a new subject or can this be considered without that step?
On 5/15/24 7:17 PM, Craig Setera wrote:
Updates to u-boot script for Raspberry Pi 4b devices to properly configure the serial console and to choose the proper boot address to be used on newer board revisions. With these changes in place, I'm able to fully boot into a serial console in u-boot. The serial console configuration was derived from the Grub configuration, while the boot address changes were taken from the IPFire Raspberry 4b documentation page: https://www.ipfire.org/docs/hardware/arm/rpi/four Per the conversation in the forums, I don't know why the boot address changes are required. I was simply trying to automate those changes such that downloaded builds will work out of the box. https://community.ipfire.org/t/hang-with-kernel-starting-on-raspberry-pi/115... NOTE: The use of the different parameter options to the test command to minimize the need for deeply nested conditionals. I did try more bash-shell like syntax, but u-boot didn't seem to like that. The operators used here were found in the u-boot source. Tested-by: Craig Setera craigjunk@setera.org Signed-off-by: Craig Setera craigjunk@setera.org
config/u-boot/boot.cmd | 27 ++++++++++++++++++++++++--- 1 file changed, 24 insertions(+), 3 deletions(-) diff --git a/config/u-boot/boot.cmd b/config/u-boot/boot.cmd index a27996780..eedd5776d 100644 --- a/config/u-boot/boot.cmd +++ b/config/u-boot/boot.cmd @@-51,7 +51,11 @@if test "${SERIAL-CONSOLE}" = "ON"; then if test "${fdtfile}" = "broadcom/bcm2837-rpi-3-b.dtb"; then setenv console ttyS1,115200n8; else -setenv console ttyAMA0,115200n8; +if test "${fdtfile}" = "broadcom/bcm2711-rpi-4-b.dtb"; then +setenv console ttyS0,115200n8; +else +setenv console ttyAMA0,115200n8; +fi; fi; fi; else @@-95,11 +99,28 @@else setenv ramdisk_addr -; fi; +# +# Handle Raspberry Pi 4 address differences +# https://www.ipfire.org/docs/hardware/arm/rpi/four +# +setenv booti_fdt_addr ${fdt_addr_r}; +if test "${board}" -eq "rpi" -a ${cpu} -eq "armv8" -a "${fdtfile}" -eq "broadcom/bcm2711-rpi-4-b.dtb"; then +# Hardware revision 1.4 +if test ${board_revision} -eq 0xB03114 -o ${board_revision} -eq 0xC03114 -o ${board_revision} -eq 0xD03114; then +setenv booti_fdt_addr ${fdt_addr}; +fi
+# Hardware revision 1.5 +if test ${board_revision} -eq 0xB03115 -o ${board_revision} -eq 0xC03115 -o ${board_revision} -eq 0xD03115; then +setenv booti_fdt_addr ${fdt_addr}; +fi +fi;
bootz ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; -booti ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; +booti ${kernel_addr_r} ${ramdisk_addr} ${booti_fdt_addr}; bootz ${kernel_addr_r} - ${fdt_addr_r}; -booti ${kernel_addr_r} - ${fdt_addr_r}; +booti ${kernel_addr_r} - ${booti_fdt_addr}; # Recompile with: # mkimage -C none -A arm -T script -d /boot/boot.cmd /boot/boot.scr
Thanks Michael
I will try to find time this weekend to try again. I may have to see if I can get this done outside of Thunderbird.
To make sure I'm understanding the issue. It seems like the patch contents are probably ok, but everything up to the contents of the patch are smashed together?
Craig
On 5/17/24 4:27 AM, Michael Tremer wrote:
Hello Craig,
Thanks again for the patch. The tag in the subject line isn’t that important, but the email you sent has again removed all whitespace which makes it impossible to read or apply the patch.
-Michael
On 16 May 2024, at 18:38, Craig Setera craigjunk@setera.org wrote:
Apologies. I just realized that I forgot to prefix the subject with [Patch]. Should I resend with a new subject or can this be considered without that step?
On 5/15/24 7:17 PM, Craig Setera wrote:
Updates to u-boot script for Raspberry Pi 4b devices to properly configure the serial console and to choose the proper boot address to be used on newer board revisions. With these changes in place, I'm ablemay to fully boot into a serial console in u-boot. The serial console configuration was derived from the Grub configuration, while the boot address changes were taken from the IPFire Raspberry 4b documentation page: https://www.ipfire.org/docs/hardware/arm/rpi/four Per the conversation in the forums, I don't know why the boot address changes are required. I was simply trying to automate those changes such that downloaded builds will work out of the box. https://community.ipfire.org/t/hang-with-kernel-starting-on-raspberry-pi/115... NOTE: The use of the different parameter options to the test command to minimize the need for deeply nested conditionals. I did try more bash-shell like syntax, but u-boot didn't seem to like that. The operators used here were found in the u-boot source. Tested-by: Craig Setera craigjunk@setera.org Signed-off-by: Craig Setera craigjunk@setera.org
config/u-boot/boot.cmd | 27 ++++++++++++++++++++++++--- 1 file changed, 24 insertions(+), 3 deletions(-) diff --git a/config/u-boot/boot.cmd b/config/u-boot/boot.cmd index a27996780..eedd5776d 100644 --- a/config/u-boot/boot.cmd +++ b/config/u-boot/boot.cmd @@-51,7 +51,11 @@if test "${SERIAL-CONSOLE}" = "ON"; then if test "${fdtfile}" = "broadcom/bcm2837-rpi-3-b.dtb"; then setenv console ttyS1,115200n8; else -setenv console ttyAMA0,115200n8; +if test "${fdtfile}" = "broadcom/bcm2711-rpi-4-b.dtb"; then +setenv console ttyS0,115200n8; +else +setenv console ttyAMA0,115200n8; +fi; fi; fi; else @@-95,11 +99,28 @@else setenv ramdisk_addr -; fi; +# +# Handle Raspberry Pi 4 address differences +# https://www.ipfire.org/docs/hardware/arm/rpi/four +# +setenv booti_fdt_addr ${fdt_addr_r}; +if test "${board}" -eq "rpi" -a ${cpu} -eq "armv8" -a "${fdtfile}" -eq "broadcom/bcm2711-rpi-4-b.dtb"; then +# Hardware revision 1.4 +if test ${board_revision} -eq 0xB03114 -o ${board_revision} -eq 0xC03114 -o ${board_revision} -eq 0xD03114; then +setenv booti_fdt_addr ${fdt_addr}; +fi
+# Hardware revision 1.5 +if test ${board_revision} -eq 0xB03115 -o ${board_revision} -eq 0xC03115 -o ${board_revision} -eq 0xD03115; then +setenv booti_fdt_addr ${fdt_addr}; +fi +fi;
bootz ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; -booti ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; +booti ${kernel_addr_r} ${ramdisk_addr} ${booti_fdt_addr}; bootz ${kernel_addr_r} - ${fdt_addr_r}; -booti ${kernel_addr_r} - ${fdt_addr_r}; +booti ${kernel_addr_r} - ${booti_fdt_addr}; # Recompile with: # mkimage -C none -A arm -T script -d /boot/boot.cmd /boot/boot.scr
Craig,
I believe Michael is referring to the leading white space or the indents.
Take a look at the existing code and that is easy to read with the indents.
Here is the existing code below:
I am guessing he has not looked at the content yet because the patch is difficult to read.
Jon
On May 17, 2024, at 6:48 AM, Craig Setera craigjunk@setera.org wrote:
Thanks Michael
I will try to find time this weekend to try again. I may have to see if I can get this done outside of Thunderbird.
To make sure I'm understanding the issue. It seems like the patch contents are probably ok, but everything up to the contents of the patch are smashed together?
Craig
On 5/17/24 4:27 AM, Michael Tremer wrote:
Hello Craig,
Thanks again for the patch. The tag in the subject line isn’t that important, but the email you sent has again removed all whitespace which makes it impossible to read or apply the patch.
-Michael
On 16 May 2024, at 18:38, Craig Setera craigjunk@setera.org wrote:
Apologies. I just realized that I forgot to prefix the subject with [Patch]. Should I resend with a new subject or can this be considered without that step?
On 5/15/24 7:17 PM, Craig Setera wrote:
Updates to u-boot script for Raspberry Pi 4b devices to properly configure the serial console and to choose the proper boot address to be used on newer board revisions. With these changes in place, I'm ablemay to fully boot into a serial console in u-boot. The serial console configuration was derived from the Grub configuration, while the boot address changes were taken from the IPFire Raspberry 4b documentation page: https://www.ipfire.org/docs/hardware/arm/rpi/four Per the conversation in the forums, I don't know why the boot address changes are required. I was simply trying to automate those changes such that downloaded builds will work out of the box. https://community.ipfire.org/t/hang-with-kernel-starting-on-raspberry-pi/115... NOTE: The use of the different parameter options to the test command to minimize the need for deeply nested conditionals. I did try more bash-shell like syntax, but u-boot didn't seem to like that. The operators used here were found in the u-boot source. Tested-by: Craig Setera craigjunk@setera.org Signed-off-by: Craig Setera craigjunk@setera.org
config/u-boot/boot.cmd | 27 ++++++++++++++++++++++++--- 1 file changed, 24 insertions(+), 3 deletions(-) diff --git a/config/u-boot/boot.cmd b/config/u-boot/boot.cmd index a27996780..eedd5776d 100644 --- a/config/u-boot/boot.cmd +++ b/config/u-boot/boot.cmd @@-51,7 +51,11 @@if test "${SERIAL-CONSOLE}" = "ON"; then if test "${fdtfile}" = "broadcom/bcm2837-rpi-3-b.dtb"; then setenv console ttyS1,115200n8; else -setenv console ttyAMA0,115200n8; +if test "${fdtfile}" = "broadcom/bcm2711-rpi-4-b.dtb"; then +setenv console ttyS0,115200n8; +else +setenv console ttyAMA0,115200n8; +fi; fi; fi; else @@-95,11 +99,28 @@else setenv ramdisk_addr -; fi; +# +# Handle Raspberry Pi 4 address differences +# https://www.ipfire.org/docs/hardware/arm/rpi/four +# +setenv booti_fdt_addr ${fdt_addr_r}; +if test "${board}" -eq "rpi" -a ${cpu} -eq "armv8" -a "${fdtfile}" -eq "broadcom/bcm2711-rpi-4-b.dtb"; then +# Hardware revision 1.4 +if test ${board_revision} -eq 0xB03114 -o ${board_revision} -eq 0xC03114 -o ${board_revision} -eq 0xD03114; then +setenv booti_fdt_addr ${fdt_addr}; +fi
+# Hardware revision 1.5 +if test ${board_revision} -eq 0xB03115 -o ${board_revision} -eq 0xC03115 -o ${board_revision} -eq 0xD03115; then +setenv booti_fdt_addr ${fdt_addr}; +fi +fi;
bootz ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; -booti ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; +booti ${kernel_addr_r} ${ramdisk_addr} ${booti_fdt_addr}; bootz ${kernel_addr_r} - ${fdt_addr_r}; -booti ${kernel_addr_r} - ${fdt_addr_r}; +booti ${kernel_addr_r} - ${booti_fdt_addr}; # Recompile with: # mkimage -C none -A arm -T script -d /boot/boot.cmd /boot/boot.scr
Makes sense. Not sure how it ended up that way because I definitely had it formatted correctly in my local source. I will need to try to get a better patch out of Git.
Thanks for the clarification!
On 5/17/24 10:54 AM, jon wrote:
Craig,
I believe Michael is referring to the leading white space or the indents.
Take a look at the existing code and that is easy to read with the indents.
Here is the existing code below:
I am guessing he has not looked at the content yet because the patch is difficult to read.
Jon
On May 17, 2024, at 6:48 AM, Craig Setera craigjunk@setera.org wrote:
Thanks Michael
I will try to find time this weekend to try again. I may have to see if I can get this done outside of Thunderbird.
To make sure I'm understanding the issue. It seems like the patch contents are probably ok, but everything up to the contents of the patch are smashed together?
Craig
On 5/17/24 4:27 AM, Michael Tremer wrote:
Hello Craig,
Thanks again for the patch. The tag in the subject line isn’t that important, but the email you sent has again removed all whitespace which makes it impossible to read or apply the patch.
-Michael
On 16 May 2024, at 18:38, Craig Setera craigjunk@setera.org wrote:
Apologies. I just realized that I forgot to prefix the subject with [Patch]. Should I resend with a new subject or can this be considered without that step?
On 5/15/24 7:17 PM, Craig Setera wrote:
Updates to u-boot script for Raspberry Pi 4b devices to properly configure the serial console and to choose the proper boot address to be used on newer board revisions. With these changes in place, I'm ablemay to fully boot into a serial console in u-boot. The serial console configuration was derived from the Grub configuration, while the boot address changes were taken from the IPFire Raspberry 4b documentation page: https://www.ipfire.org/docs/hardware/arm/rpi/four Per the conversation in the forums, I don't know why the boot address changes are required. I was simply trying to automate those changes such that downloaded builds will work out of the box. https://community.ipfire.org/t/hang-with-kernel-starting-on-raspberry-pi/115... NOTE: The use of the different parameter options to the test command to minimize the need for deeply nested conditionals. I did try more bash-shell like syntax, but u-boot didn't seem to like that. The operators used here were found in the u-boot source. Tested-by: Craig Setera craigjunk@setera.org Signed-off-by: Craig Setera craigjunk@setera.org
config/u-boot/boot.cmd | 27 ++++++++++++++++++++++++--- 1 file changed, 24 insertions(+), 3 deletions(-) diff --git a/config/u-boot/boot.cmd b/config/u-boot/boot.cmd index a27996780..eedd5776d 100644 --- a/config/u-boot/boot.cmd +++ b/config/u-boot/boot.cmd @@-51,7 +51,11 @@if test "${SERIAL-CONSOLE}" = "ON"; then if test "${fdtfile}" = "broadcom/bcm2837-rpi-3-b.dtb"; then setenv console ttyS1,115200n8; else -setenv console ttyAMA0,115200n8; +if test "${fdtfile}" = "broadcom/bcm2711-rpi-4-b.dtb"; then +setenv console ttyS0,115200n8; +else +setenv console ttyAMA0,115200n8; +fi; fi; fi; else @@-95,11 +99,28 @@else setenv ramdisk_addr -; fi; +# +# Handle Raspberry Pi 4 address differences +# https://www.ipfire.org/docs/hardware/arm/rpi/four +# +setenv booti_fdt_addr ${fdt_addr_r}; +if test "${board}" -eq "rpi" -a ${cpu} -eq "armv8" -a "${fdtfile}" -eq "broadcom/bcm2711-rpi-4-b.dtb"; then +# Hardware revision 1.4 +if test ${board_revision} -eq 0xB03114 -o ${board_revision} -eq 0xC03114 -o ${board_revision} -eq 0xD03114; then +setenv booti_fdt_addr ${fdt_addr}; +fi
+# Hardware revision 1.5 +if test ${board_revision} -eq 0xB03115 -o ${board_revision} -eq 0xC03115 -o ${board_revision} -eq 0xD03115; then +setenv booti_fdt_addr ${fdt_addr}; +fi +fi;
bootz ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; -booti ${kernel_addr_r} ${ramdisk_addr} ${fdt_addr_r}; +booti ${kernel_addr_r} ${ramdisk_addr} ${booti_fdt_addr}; bootz ${kernel_addr_r} - ${fdt_addr_r}; -booti ${kernel_addr_r} - ${fdt_addr_r}; +booti ${kernel_addr_r} - ${booti_fdt_addr}; # Recompile with: # mkimage -C none -A arm -T script -d /boot/boot.cmd /boot/boot.scr
-
Craig Setera -
jon -
Michael Tremer