On Tue, Feb 23, 2021 at 07:00:47AM +0000, Lim, Elly Siew Chin wrote:
Hi Tom,
-----Original Message----- From: Tom Rini trini@konsulko.com Sent: Saturday, February 20, 2021 5:02 AM To: Lim, Elly Siew Chin elly.siew.chin.lim@intel.com Cc: u-boot@lists.denx.de; Marek Vasut marex@denx.de; Tan, Ley Foon ley.foon.tan@intel.com; See, Chin Liang chin.liang.see@intel.com; Simon Goldschmidt simon.k.r.goldschmidt@gmail.com; Chee, Tien Fong tien.fong.chee@intel.com; Westergreen, Dalon dalon.westergreen@intel.com; Simon Glass sjg@chromium.org; Gan, Yau Wai yau.wai.gan@intel.com Subject: Re: [PATCH] Makefile: Add target to generate hex output for combined spl and dtb
On Fri, Feb 19, 2021 at 01:55:44PM +0800, Siew Chin Lim wrote:
From: Dalon Westergreen dalon.westergreen@intel.com
Some architectures, Stratix10/Agilex, require a hex formatted spl that combines the spl image and dtb. This adds a target to create said hex file with and offset of SPL_TEXT_BASE.
Signed-off-by: Dalon Westergreen dalon.westergreen@intel.com Signed-off-by: Siew Chin Lim elly.siew.chin.lim@intel.com
Makefile | 11 ++++++----- include/configs/socfpga_soc64_common.h | 2 +- scripts/Makefile.spl | 8 ++++++++ 3 files changed, 15 insertions(+), 6 deletions(-)
diff --git a/Makefile b/Makefile index 4da46dea39..f1adc9aa23 100644 --- a/Makefile +++ b/Makefile @@ -1263,11 +1263,6 @@ OBJCOPYFLAGS_u-boot-nodtb.bin := -O binary \ $(if $(CONFIG_X86_16BIT_INIT),-R .start16 -R .resetvec) \ $(if $(CONFIG_MPC85XX_HAVE_RESET_VECTOR),-R .bootpg -
R .resetvec)
-OBJCOPYFLAGS_u-boot-spl.hex = $(OBJCOPYFLAGS_u-boot.hex)
-spl/u-boot-spl.hex: spl/u-boot-spl FORCE
- $(call if_changed,objcopy)
binary_size_check: u-boot-nodtb.bin FORCE @file_size=$(shell wc -c u-boot-nodtb.bin | awk '{print $$1}') ; \ map_size=$(shell cat u-boot.map | \ @@ -1935,6 +1930,12 @@ spl/u-boot-spl.bin: spl/u-boot-spl @: $(SPL_SIZE_CHECK)
+spl/u-boot-spl-dtb.bin: spl/u-boot-spl
- @:
+spl/u-boot-spl-dtb.hex: spl/u-boot-spl
- @:
spl/u-boot-spl: tools prepare \ $(if
$(CONFIG_OF_SEPARATE)$(CONFIG_OF_EMBED)$(CONFIG_SPL_OF_PLATDATA ),dts/dt.dtb) \
$(if$(CONFIG_OF_SEPARATE)$(CONFIG_OF_EMBED)$(CONFIG_TPL_OF_PLATDATA ),dts/d
t.dtb) diff --git a/include/configs/socfpga_soc64_common.h b/include/configs/socfpga_soc64_common.h index fdcd7d3e9a..1af359466c 100644 --- a/include/configs/socfpga_soc64_common.h +++ b/include/configs/socfpga_soc64_common.h @@ -200,7 +200,7 @@ unsigned int cm_get_l4_sys_free_clk_hz(void);
- 0x8000_0000 ...... End of SDRAM_1 (assume 2GB)
*/ -#define CONFIG_SPL_TARGET "spl/u-boot-spl.hex" +#define CONFIG_SPL_TARGET "spl/u-boot-spl-dtb.hex" #define CONFIG_SPL_MAX_SIZE CONFIG_SYS_INIT_RAM_SIZE #define CONFIG_SPL_STACK CONFIG_SYS_INIT_SP_ADDR #define CONFIG_SPL_BSS_MAX_SIZE 0x100000 /* 1 MB */ diff --git a/scripts/Makefile.spl b/scripts/Makefile.spl index ea4e045769..625e06d0d9 100644 --- a/scripts/Makefile.spl +++ b/scripts/Makefile.spl @@ -229,6 +229,9 @@ ifneq
($(CONFIG_TARGET_SOCFPGA_GEN5)$(CONFIG_TARGET_SOCFPGA_ARRIA10),)
INPUTS-y += $(obj)/$(SPL_BIN).sfp endif
+INPUTS-$(CONFIG_TARGET_SOCFPGA_STRATIX10) += $(obj)/u-boot-spl-
dtb.hex
+INPUTS-$(CONFIG_TARGET_SOCFPGA_AGILEX) += $(obj)/u-
boot-spl-dtb.hex
ifdef CONFIG_ARCH_SUNXI INPUTS-y += $(obj)/sunxi-spl.bin
@@ -389,6 +392,11 @@ $(obj)/$(SPL_BIN).sfp: $(obj)/$(SPL_BIN).bin FORCE MKIMAGEFLAGS_sunxi-spl.bin = -T sunxi_egon \ -n $(CONFIG_DEFAULT_DEVICE_TREE)
+OBJCOPYFLAGS_u-boot-spl-dtb.hex := -I binary -O ihex +--change-address=$(CONFIG_SPL_TEXT_BASE)
+$(obj)/u-boot-spl-dtb.hex: $(obj)/u-boot-spl-dtb.bin FORCE
- $(call if_changed,objcopy)
$(obj)/sunxi-spl.bin: $(obj)/$(SPL_BIN).bin FORCE $(call if_changed,mkimage)
The commit message / subject doesn't make it clear that you're doing something for SoCFPGA here, and I seem to recall this change being proposed a while ago and it wasn't clear that this is really the best default path forward I thought?
Noted. I will resend the patch and enhance the message / subject .
I have checked again with the team about this, we still would like to generate the 'u-boot-spl-dtb.hex' for socfpga soc64 devices by default during u-boot compilation which is always needed by customer to generate the final configuration bitstream. Can you please advise if this is not the right way for the implementation?
I think the changes are fine then, please just update the commit message, thanks!