On 16.3.2016 01:50, Tom Rini wrote:

On Wed, Mar 16, 2016 at 01:25:14AM +0100, Michal Simek wrote:

...

On 15.3.2016 23:29, Tom Rini wrote:

...

Starting with 96e5b03 we use a linker list for partition table information. However since we use this in SPL we need to make sure that the SPL linker scripts include these as well.

Cc: Michal Simek michal.simek@xilinx.com Reviewed-by: Simon Glass sjg@chromium.org Reported-by: Nishanth Menon nm@ti.com Tested-by: Nishanth Menon nm@ti.com Signed-off-by: Tom Rini trini@konsulko.com

---

Changes in v2:

• Drop the CONFIG_SPL_DM tests for zynq as DM is the only case for zynq and while in here drop a now useless line.

---

arch/arm/cpu/armv7/omap-common/u-boot-spl.lds | 1 + arch/arm/cpu/u-boot-spl.lds | 1 + arch/arm/mach-zynq/u-boot-spl.lds | 4 ---- 3 files changed, 2 insertions(+), 4 deletions(-)

diff --git a/arch/arm/cpu/armv7/omap-common/u-boot-spl.lds b/arch/arm/cpu/armv7/omap-common/u-boot-spl.lds index ccd0c83..9dccdc0 100644 --- a/arch/arm/cpu/armv7/omap-common/u-boot-spl.lds +++ b/arch/arm/cpu/armv7/omap-common/u-boot-spl.lds @@ -35,6 +35,7 @@ SECTIONS

. = ALIGN(4); .u_boot_list : {

• KEEP(*(SORT(.u_boot_list_*_part_driver_*)));
  

  KEEP(*(SORT(.u_boot_list*_i2c_*))); } >.sram

diff --git a/arch/arm/cpu/u-boot-spl.lds b/arch/arm/cpu/u-boot-spl.lds index c5b4f7c..1805043 100644 --- a/arch/arm/cpu/u-boot-spl.lds +++ b/arch/arm/cpu/u-boot-spl.lds @@ -40,6 +40,7 @@ SECTIONS #endif . = .; .u_boot_list : {

• KEEP(*(SORT(.u_boot_list_*_part_driver_*)));
  

  KEEP(*(SORT(.u_boot_list*_i2c_*))); }

diff --git a/arch/arm/mach-zynq/u-boot-spl.lds b/arch/arm/mach-zynq/u-boot-spl.lds index ecdf6a0..a5c76a6 100644 --- a/arch/arm/mach-zynq/u-boot-spl.lds +++ b/arch/arm/mach-zynq/u-boot-spl.lds @@ -38,16 +38,12 @@ SECTIONS } > .sram

. = ALIGN(4); -#ifdef CONFIG_SPL_DM .u_boot_list : { KEEP(*(SORT(.u_boot_list_*_driver_*))); KEEP(*(SORT(.u_boot_list_*_uclass_*))); } > .sram

. = ALIGN(4); -#endif

• . = .;

This is getting more interested than I thought. We have for zynq added two lists for driver and uclass added by Simon but this is missing in linker script for others SoC. Does that mean that none else is using SPL with DM? Or that we have in linker script something what doesn't need to be there?

Zynq is ahead of the curve here, many SoCs are not doing DM in SPL yet. With my TI hat on, the series that lets us do FIT in SPL (and thus pick one of N DTBs for u-boot.img) is what unblocks moving over because we can retain having 'am335x_evm' support 4+ distinct HW platforms in a single binary.

ok.

...

Regarding this patch. Zynq is not enabling partition support but I still think that we should add also this list to zynq spl. Maybe we didn't understand each other. I wanted to say remove that SPL_DM and keep that part_driver lists even we are not enabling them by default.

Right, but keep in mind that u_boot_list_*_driver_* will match u_boot_list_2_part_driver_2_dos already so we don't need to glob it in.

All of that said, I'm going to do a v3 now. I played with the linker list stuff like I was thinking I should anyhow and I see things like: *(SORT(.u_boot_list_*)) .u_boot_list_2_cmd_1 0x0000000040310d50 0x0 common/built-in.o .u_boot_list_2_cmd_3 0x0000000040310d50 0x0 common/built-in.o .u_boot_list_2_i2c_1 0x0000000040310d50 0x0 drivers/built-in.o .u_boot_list_2_i2c_2_omap24_0 0x0000000040310d50 0x2c drivers/built-in.o

So when we include empty stuff into the file the linker still does the right thing.

You are right I missed that. But on the other hand maybe then make sense for others to just add this KEEP(*(SORT(.u_boot_list_*_driver_*))); instead of KEEP(*(SORT(.u_boot_list_*_part_driver_*)));

Thanks, Michal