[U-Boot] [PATCH 1/2] board/db410c: add missing linker map entries for efi
Otherwise the loaded image would miss the efi_runtime sections, and fall over hard when grub (for example) tried to call runtime services located in this section.
Signed-off-by: Rob Clark robdclark@gmail.com --- board/qualcomm/dragonboard410c/u-boot.lds | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)
diff --git a/board/qualcomm/dragonboard410c/u-boot.lds b/board/qualcomm/dragonboard410c/u-boot.lds index 6e1c5a8..62ac4d7 100644 --- a/board/qualcomm/dragonboard410c/u-boot.lds +++ b/board/qualcomm/dragonboard410c/u-boot.lds @@ -43,6 +43,22 @@ SECTIONS
. = ALIGN(8);
+ .efi_runtime : { + __efi_runtime_start = .; + *(efi_runtime_text) + *(efi_runtime_data) + __efi_runtime_stop = .; + } + + .efi_runtime_rel : { + __efi_runtime_rel_start = .; + *(.relaefi_runtime_text) + *(.relaefi_runtime_data) + __efi_runtime_rel_stop = .; + } + + . = ALIGN(8); + .image_copy_end : { *(.__image_copy_end)
Signed-off-by: Rob Clark robdclark@gmail.com --- Maybe there is a better way to not hardcode this? But at least with the build of lk that I have, the fdt table is at 0x81e00000. I guess there must be a more robust way to do this, since presumably lk when booting the linux kernel directly somehow passes the fdt address.
include/configs/dragonboard410c.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/configs/dragonboard410c.h b/include/configs/dragonboard410c.h index 11c842d..3b9932d 100644 --- a/include/configs/dragonboard410c.h +++ b/include/configs/dragonboard410c.h @@ -105,7 +105,7 @@ REFLASH(dragonboard/u-boot.img, 8)\ "linux_image=Image\0" \ "kernel_addr_r=0x81000000\0"\ "fdtfile=apq8016-sbc.dtb\0" \ - "fdt_addr_r=0x83000000\0"\ + "fdt_addr_r=0x81e00000\0"\ "ramdisk_addr_r=0x84000000\0"\ "scriptaddr=0x90000000\0"\ "pxefile_addr_r=0x90100000\0"\
On Tue, Jun 20, 2017 at 05:55:25PM -0400, Rob Clark wrote:
Signed-off-by: Rob Clark robdclark@gmail.com
Maybe there is a better way to not hardcode this? But at least with the build of lk that I have, the fdt table is at 0x81e00000. I guess there must be a more robust way to do this, since presumably lk when booting the linux kernel directly somehow passes the fdt address.
I would assume that lk does what Documentation/arm64/booting.txt describes and places the physical address in x0, so you might be able to implement a save_boot_params that saves this information for later use? Perhaps even make this somewhat generic for armv8 as there's probably other cases where U-Boot is being called in this manner? Thanks!
On Fri, Jun 23, 2017 at 10:32 AM, Tom Rini trini@konsulko.com wrote:
On Tue, Jun 20, 2017 at 05:55:25PM -0400, Rob Clark wrote:
Signed-off-by: Rob Clark robdclark@gmail.com
Maybe there is a better way to not hardcode this? But at least with the build of lk that I have, the fdt table is at 0x81e00000. I guess there must be a more robust way to do this, since presumably lk when booting the linux kernel directly somehow passes the fdt address.
I would assume that lk does what Documentation/arm64/booting.txt describes and places the physical address in x0, so you might be able to implement a save_boot_params that saves this information for later use? Perhaps even make this somewhat generic for armv8 as there's probably other cases where U-Boot is being called in this manner? Thanks!
yup.. figured this out.. I have a WIP patch to actually use the fdt that the fw passes to u-boot (instead of appending the fdt to u-boot).. haven't wired it up to setenv_hex() yet, but that should be trivial.
fwiw, I have a WIP u-boot equiv to linux's simplefb display driver that can inherit the scanout setup by fw (and some related patches) plus lk patches to create a chosen/framebuffer simple-framebuffer node.. need to clean up and send out some of my pending stack of patches, but been buried in figuring out u-boot reloc stuff to figure out how to not get the vaddr space associated w/ fw configured framebuffer reloc'd.
BR, -R
On Fri, Jun 23, 2017 at 4:24 PM, Rob Clark robdclark@gmail.com wrote:
On Fri, Jun 23, 2017 at 10:32 AM, Tom Rini trini@konsulko.com wrote:
On Tue, Jun 20, 2017 at 05:55:25PM -0400, Rob Clark wrote:
Signed-off-by: Rob Clark robdclark@gmail.com
Maybe there is a better way to not hardcode this? But at least with the build of lk that I have, the fdt table is at 0x81e00000. I guess there must be a more robust way to do this, since presumably lk when booting the linux kernel directly somehow passes the fdt address.
I would assume that lk does what Documentation/arm64/booting.txt describes and places the physical address in x0, so you might be able to implement a save_boot_params that saves this information for later use? Perhaps even make this somewhat generic for armv8 as there's probably other cases where U-Boot is being called in this manner? Thanks!
yup.. figured this out.. I have a WIP patch to actually use the fdt that the fw passes to u-boot (instead of appending the fdt to u-boot).. haven't wired it up to setenv_hex() yet, but that should be trivial.
fwiw, I have a WIP u-boot equiv to linux's simplefb display driver that can inherit the scanout setup by fw (and some related patches) plus lk patches to create a chosen/framebuffer simple-framebuffer node.. need to clean up and send out some of my pending stack of patches, but been buried in figuring out u-boot reloc stuff to figure out how to not get the vaddr space associated w/ fw configured framebuffer reloc'd.
(and just in-case that wasn't clear, ignore patch 2/2.. but 1/2 is still valid)
BR, -R
On Fri, Jun 23, 2017 at 04:35:43PM -0400, Rob Clark wrote:
On Fri, Jun 23, 2017 at 4:24 PM, Rob Clark robdclark@gmail.com wrote:
On Fri, Jun 23, 2017 at 10:32 AM, Tom Rini trini@konsulko.com wrote:
On Tue, Jun 20, 2017 at 05:55:25PM -0400, Rob Clark wrote:
Signed-off-by: Rob Clark robdclark@gmail.com
Maybe there is a better way to not hardcode this? But at least with the build of lk that I have, the fdt table is at 0x81e00000. I guess there must be a more robust way to do this, since presumably lk when booting the linux kernel directly somehow passes the fdt address.
I would assume that lk does what Documentation/arm64/booting.txt describes and places the physical address in x0, so you might be able to implement a save_boot_params that saves this information for later use? Perhaps even make this somewhat generic for armv8 as there's probably other cases where U-Boot is being called in this manner? Thanks!
yup.. figured this out.. I have a WIP patch to actually use the fdt that the fw passes to u-boot (instead of appending the fdt to u-boot).. haven't wired it up to setenv_hex() yet, but that should be trivial.
fwiw, I have a WIP u-boot equiv to linux's simplefb display driver that can inherit the scanout setup by fw (and some related patches) plus lk patches to create a chosen/framebuffer simple-framebuffer node.. need to clean up and send out some of my pending stack of patches, but been buried in figuring out u-boot reloc stuff to figure out how to not get the vaddr space associated w/ fw configured framebuffer reloc'd.
(and just in-case that wasn't clear, ignore patch 2/2.. but 1/2 is still valid)
OK, cool. I'll move 2/2 to Rejected and 1/2 is currently in my testing queue today. Thanks!
On Tue, Jun 20, 2017 at 05:55:24PM -0400, Rob Clark wrote:
Otherwise the loaded image would miss the efi_runtime sections, and fall over hard when grub (for example) tried to call runtime services located in this section.
Signed-off-by: Rob Clark robdclark@gmail.com
Applied to u-boot/master, thanks!
participants (2)
-
Rob Clark -
Tom Rini