uboot-bcm4908: add package with BCM4908 U-Boot

New BCM4908 devices come with U-Boot instead of CFE. Firmwares for such
devices has to include U-Boot.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
(cherry picked from commit 0d45e1ea96)

package/boot/uboot-bcm4908/Makefile

@@ -0,0 +1,48 @@
+# SPDX-License-Identifier: GPL-2.0-or-later
+
+include $(TOPDIR)/rules.mk
+include $(INCLUDE_DIR)/kernel.mk
+
+PKG_RELEASE:=$(AUTORELEASE)
+
+PKG_SOURCE_PROTO:=git
+PKG_SOURCE_URL:=https://git.openwrt.org/project/bcm63xx/u-boot.git
+PKG_SOURCE_DATE:=2022-03-03
+PKG_SOURCE_VERSION:=92e9eca819c9c898d9d2010e1a217726c42c8c47
+PKG_MIRROR_HASH:=11c37fe4c18d55e799153600d1cfd8ee9ca7da8326d0024c1d825f4a327c5f0d
+
+include $(INCLUDE_DIR)/u-boot.mk
+include $(INCLUDE_DIR)/package.mk
+
+define U-Boot/Default
+  BUILD_TARGET:=bcm4908
+  UBOOT_IMAGE:=u-boot-nodtb.bin
+  DEFAULT:=y
+endef
+
+define U-Boot/bcm4908
+  NAME:=Broadcom's BCM4908
+  UBOOT_CONFIG:=bcm94908
+endef
+
+UBOOT_TARGETS := \
+	bcm4908
+
+define Build/Prepare
+	$(call Build/Prepare/Default)
+	mkdir -p $(PKG_BUILD_DIR)/include/generated/
+	( cd $(PKG_BUILD_DIR)/board/broadcom/bcmbca/httpd/html/ && \
+	  xxd -i index.html > ../../../../../include/generated/index.h && \
+	  xxd -i flashing.html > ../../../../../include/generated/flashing.h && \
+	  xxd -i fail.html > ../../../../../include/generated/fail.h && \
+	  xxd -i 404.html > ../../../../../include/generated/404.h )
+endef
+
+define Build/InstallDev
+	$(INSTALL_DIR) $(STAGING_DIR_IMAGE)/u-boot
+	$(INSTALL_BIN) $(PKG_BUILD_DIR)/$(UBOOT_IMAGE) $(STAGING_DIR_IMAGE)/u-boot/
+	$(INSTALL_BIN) $(PKG_BUILD_DIR)/u-boot.dtb $(STAGING_DIR_IMAGE)/u-boot/
+	$(INSTALL_BIN) $(PKG_BUILD_DIR)/arch/arm/dts/*.dtb $(STAGING_DIR_IMAGE)/u-boot/
+endef
+
+$(eval $(call BuildPackage/U-Boot))

package/boot/uboot-bcm4908/patches/100-check-config-allow-to-complete-build-even-with-ad-ho.patch

@@ -0,0 +1,40 @@
+From: Masahiro Yamada <yamada.masahiro@socionext.com>
+Date: Mon, 26 Sep 2016 13:05:02 +0900
+Subject: [PATCH] check-config: allow to complete build even with ad-hoc CONFIG
+ options
+
+Currently, the check-config.sh terminates the build when unknown
+ad-hoc options are detected.  I think it is too much because we may
+want to patch config headers locally in a build/deployment project.
+
+So, let's relax check-config.sh to just warn even if it detects
+options that are not in the whitelist.  Instead, this check can be
+done at the end of build, along with other checks.  It will catch
+more attention.
+
+Even with this change, the Buildman tool catches new warnings,
+so Tom can give NACK to new ad-hoc options.
+
+Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
+---
+ scripts/check-config.sh | 3 +--
+ 1 file changed, 1 insertion(+), 2 deletions(-)
+
+--- a/scripts/check-config.sh
++++ b/scripts/check-config.sh
+@@ -50,14 +50,13 @@ cat `find ${srctree} -name "Kconfig*"` |sed -n \
+ 	|sort |uniq > ${ok}
+ comm -23 ${suspects} ${ok} >${new_adhoc}
+ if [ -s ${new_adhoc} ]; then
+-	echo >&2 "Error: You must add new CONFIG options using Kconfig"
++	echo >&2 "Warning: You must add new CONFIG options using Kconfig"
+ 	echo >&2 "The following new ad-hoc CONFIG options were detected:"
+ 	cat >&2 ${new_adhoc}
+ 	echo >&2
+ 	echo >&2 "Please add these via Kconfig instead. Find a suitable Kconfig"
+ 	echo >&2 "file and add a 'config' or 'menuconfig' option."
+ 	# Don't delete the temporary files in case they are useful
+-	exit 1
+ else
+ 	rm ${suspects} ${ok} ${new_adhoc}
+ fi

package/boot/uboot-bcm4908/patches/101-tools-fix-mkimage-static-compilation.patch

@@ -0,0 +1,24 @@
+From 8efe417aa9de654425cc01d0fc93be355a3f648d Mon Sep 17 00:00:00 2001
+From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= <rafal@milecki.pl>
+Date: Wed, 13 Oct 2021 11:04:45 +0200
+Subject: [PATCH] tools: fix mkimage static compilation
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
+---
+ tools/Makefile | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+--- a/tools/Makefile
++++ b/tools/Makefile
+@@ -157,7 +157,7 @@ ifneq ($(CONFIG_MX23)$(CONFIG_MX28)$(CON
+ HOSTCFLAGS_kwbimage.o += \
+ 	$(shell pkg-config --cflags libssl libcrypto 2> /dev/null || echo "")
+ HOSTLOADLIBES_mkimage += \
+-	$(shell pkg-config --libs libssl libcrypto 2> /dev/null || echo "-lssl -lcrypto")
++	$(shell pkg-config --libs --static libssl libcrypto 2> /dev/null || echo "-lssl -lcrypto -lpthread")
+ 
+ # OS X deprecate openssl in favour of CommonCrypto, supress deprecation
+ # warnings on those systems

package/boot/uboot-bcm4908/patches/200-configs-bcm94908-unset-CONFIG_SPL.patch

@@ -0,0 +1,38 @@
+From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= <rafal@milecki.pl>
+Date: Fri, 4 Mar 2022 09:21:32 +0100
+Subject: [PATCH] configs: bcm94908: unset CONFIG_SPL
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+Compiling SPL is always tricky as it needs to fit limited resources.
+Fortunately in most cases there is no need to replace SPL or TPL while
+flashing a new firmware.
+
+Compiling SPL for BCM4908 seems to fail with non-Broadcom toolchain:
+aarch64-openwrt-linux-musl-ld.bfd: u-boot-spl section `.u_boot_list' will not fit in region `.sram'
+aarch64-openwrt-linux-musl-ld.bfd: section .bss VMA [00000000822b9000,00000000822b93ef] overlaps section .u_boot_list VMA [00000000822b8f60,00000000822b9a87]
+aarch64-openwrt-linux-musl-ld.bfd: region `.sram' overflowed by 2696 bytes
+
+It also requires hashtable.h which has to be generated using some
+Broadcom's custom perl script that isn't integrated as this point.
+
+For now just disable SPL and use only last-stage U-Boot that must be
+shipped with every firmware.
+
+Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
+---
+ configs/bcm94908_defconfig | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+--- a/configs/bcm94908_defconfig
++++ b/configs/bcm94908_defconfig
+@@ -21,7 +21,7 @@ CONFIG_SPL_SERIAL_SUPPORT=y
+ CONFIG_TPL_LIBCOMMON_SUPPORT=y
+ CONFIG_TPL_LIBGENERIC_SUPPORT=y
+ CONFIG_NR_DRAM_BANKS=1
+-CONFIG_SPL=y
++# CONFIG_SPL is not set
+ CONFIG_SPL_LIBDISK_SUPPORT=y
+ CONFIG_ENV_VARS_UBOOT_CONFIG=y
+ CONFIG_TPL_SYS_MALLOC_F_LEN=0x11000

package/boot/uboot-bcm4908/patches/201-Assume-TPL-support-for-ATF-when-compiling-U-Boot-wit.patch

@@ -0,0 +1,55 @@
+From: =?UTF-8?q?Rafa=C5=82=20Mi=C5=82ecki?= <rafal@milecki.pl>
+Date: Fri, 4 Mar 2022 09:23:34 +0100
+Subject: [PATCH] Assume TPL support for ATF when compiling U-Boot without TPL
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+Broadcom's U-Boot behaviour depends on compilation time check whether
+TPL was compiled with or without ATF support. There is no proper runtime
+check.
+
+When compiling just U-Boot (without SPL & TPL) there is no way to tell
+if it's going to work with TPL with or without ATF support.
+
+Modify code to blindly assume ATF support in TPL in such cases. It seems
+to be always true for Broadcom and we need some assumption as we don't
+deal with compiling SPL or TPL.
+
+Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
+---
+ arch/arm/mach-bcmbca/bcm4908/cpu.c | 2 +-
+ board/broadcom/bcmbca/board.c      | 4 ++--
+ 2 files changed, 3 insertions(+), 3 deletions(-)
+
+--- a/arch/arm/mach-bcmbca/bcm4908/cpu.c
++++ b/arch/arm/mach-bcmbca/bcm4908/cpu.c
+@@ -138,7 +138,7 @@ int get_nr_cpus()
+ 	return nr_cpus;
+ }
+ 
+-#if !defined(CONFIG_TPL_ATF)
++#if defined(CONFIG_TPL) && !defined(CONFIG_TPL_ATF)
+ void boot_secondary_cpu(unsigned long vector)
+ {
+ 	uint32_t cpu, nr_cpus = QUAD_CPUS;
+--- a/board/broadcom/bcmbca/board.c
++++ b/board/broadcom/bcmbca/board.c
+@@ -103,7 +103,7 @@ void board_spinor_init(void)
+ 
+ int board_init(void)
+ {
+-#if !defined(CONFIG_TPL_ATF)
++#if defined(CONFIG_TPL) && !defined(CONFIG_TPL_ATF)
+ 	unsigned long vector;
+ #endif
+ 	board_sdk_init_e();
+@@ -121,7 +121,7 @@ int board_init(void)
+ 	printf("$Uboot: "BUILD_TAG" $\n");
+ #endif
+ 
+-#if !defined(CONFIG_TPL_ATF)
++#if defined(CONFIG_TPL) && !defined(CONFIG_TPL_ATF)
+ #if defined(CONFIG_ARM64)
+ 	vector  = (unsigned long)&_start;
+ #else