[PATCH 0/8] Support 64-bit U-Boot for Slim Bootloader
From: Aiden Park aiden.park@intel.com
This patchset will build 64-bit U-Boot as a Slim Bootloader payload and introduce a new X86_RUN_64BIT_ONLY to build pure 64-bit U-Boot binary without SPL. In addition, this supports 64-bit kernel booting from 64-bit U-Boot.
Aiden Park (8): x86: Add a new X86_RUN_64BIT_ONLY to Kconfig x86: start64: Add a hook at 64-bit entry x86: start64: Support HOB in 64-bit U-Boot x86: slimbootloader: Support 64-bit operation configs: slimbootloader: Add x86_64 slimbootloader config x86: Fix 64-bit compile warning x86: enable 64-bit kernel boot from 64-bit U-Boot doc: slimbootloader: Update 64-bit build instruction
arch/x86/Kconfig | 8 ++++++ arch/x86/cpu/slimbootloader/Makefile | 9 ++++-- arch/x86/cpu/slimbootloader/entry64.S | 14 ++++++++++ arch/x86/cpu/slimbootloader/slimbootloader.c | 17 ++++++++++-- arch/x86/cpu/start64.S | 10 +++++++ arch/x86/cpu/x86_64/Makefile | 2 +- arch/x86/cpu/x86_64/call64.S | 21 ++++++++++++++ arch/x86/cpu/x86_64/cpu.c | 10 +++++++ arch/x86/include/asm/bootparam.h | 10 ++++++- arch/x86/include/asm/zimage.h | 2 +- arch/x86/lib/bootm.c | 10 +++++-- arch/x86/lib/zimage.c | 24 ++++++++++------ cmd/x86/hob.c | 8 +++--- configs/slimbootloader-x86_64_defconfig | 24 ++++++++++++++++ doc/board/intel/slimbootloader.rst | 29 ++++++++++++++++++++ 15 files changed, 176 insertions(+), 22 deletions(-) create mode 100644 arch/x86/cpu/slimbootloader/entry64.S create mode 100644 arch/x86/cpu/x86_64/call64.S create mode 100644 configs/slimbootloader-x86_64_defconfig
From: Aiden Park aiden.park@intel.com
This will build U-Boot as a pure 64-bit binary with no SPL. It can be used with a pre-stage boot firmware which has already done 16-bit, 32-bit and 64-bit init.
Signed-off-by: Aiden Park aiden.park@intel.com --- arch/x86/Kconfig | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index c8eae24c07..89add17e69 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -40,6 +40,14 @@ config X86_RUN_64BIT runs through the 16-bit and 32-bit init, then switches to 64-bit mode and jumps to U-Boot proper.
+config X86_RUN_64BIT_ONLY + bool "64-bit only" + select X86_64 + help + Build U-Boot as a pure 64-bit binary with no 32-bit SPL. This can + be used with a pre-stage boot firmware which has already done + 16-bit, 32-bit and 64-bit init. + endchoice
config X86_64
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This will build U-Boot as a pure 64-bit binary with no SPL. It can be used with a pre-stage boot firmware which has already done 16-bit, 32-bit and 64-bit init.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/Kconfig | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index c8eae24c07..89add17e69 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -40,6 +40,14 @@ config X86_RUN_64BIT runs through the 16-bit and 32-bit init, then switches to 64-bit mode and jumps to U-Boot proper.
+config X86_RUN_64BIT_ONLY
Perhaps X86_RUN_64BIT_NO_SPL? Bin might hav a better idea.
bool "64-bit only"select X86_64helpBuild U-Boot as a pure 64-bit binary with no 32-bit SPL. This canbe used with a pre-stage boot firmware which has already done16-bit, 32-bit and 64-bit init.endchoice
config X86_64
2.20.1
Regards, Simon
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 1/8] x86: Add a new X86_RUN_64BIT_ONLY to Kconfig
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This will build U-Boot as a pure 64-bit binary with no SPL. It can be used with a pre-stage boot firmware which has already done 16-bit, 32-bit and 64-bit init.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/Kconfig | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index c8eae24c07..89add17e69 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -40,6 +40,14 @@ config X86_RUN_64BIT runs through the 16-bit and 32-bit init, then switches to 64-bit mode and jumps to U-Boot proper.
+config X86_RUN_64BIT_ONLY
Perhaps X86_RUN_64BIT_NO_SPL? Bin might hav a better idea.
I considered X64_RUN_64BIT_NO_SPL, X64_RUN_64BIT_PURE and X64_RUN_64BIT_NATIVE. Hi Bin, can you recommend better naming?
bool "64-bit only"select X86_64helpBuild U-Boot as a pure 64-bit binary with no 32-bit SPL. This canbe used with a pre-stage boot firmware which has already done16-bit, 32-bit and 64-bit init.endchoice
config X86_64
2.20.1
Regards, Simon
Best Regards, Aiden
Hi Aiden,
On Wed, Apr 29, 2020 at 1:44 PM Park, Aiden aiden.park@intel.com wrote:
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 1/8] x86: Add a new X86_RUN_64BIT_ONLY to Kconfig
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This will build U-Boot as a pure 64-bit binary with no SPL. It can be used with a pre-stage boot firmware which has already done 16-bit, 32-bit and 64-bit init.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/Kconfig | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index c8eae24c07..89add17e69 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -40,6 +40,14 @@ config X86_RUN_64BIT runs through the 16-bit and 32-bit init, then switches to 64-bit mode and jumps to U-Boot proper.
+config X86_RUN_64BIT_ONLY
Perhaps X86_RUN_64BIT_NO_SPL? Bin might hav a better idea.
I considered X64_RUN_64BIT_NO_SPL, X64_RUN_64BIT_PURE and X64_RUN_64BIT_NATIVE. Hi Bin, can you recommend better naming?
Cannot figure out a better name :) Let's use it.
Reviewed-by: Bin Meng bmeng.cn@gmail.com
Regards, Bin
Hi Bin,
-----Original Message----- From: Bin Meng bmeng.cn@gmail.com Sent: Thursday, April 30, 2020 2:14 AM To: Park, Aiden aiden.park@intel.com Cc: Simon Glass sjg@chromium.org; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 1/8] x86: Add a new X86_RUN_64BIT_ONLY to Kconfig
Hi Aiden,
On Wed, Apr 29, 2020 at 1:44 PM Park, Aiden aiden.park@intel.com wrote:
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 1/8] x86: Add a new X86_RUN_64BIT_ONLY to Kconfig
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This will build U-Boot as a pure 64-bit binary with no SPL. It can be used with a pre-stage boot firmware which has already done 16-bit, 32-bit and 64-bit init.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/Kconfig | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index c8eae24c07..89add17e69 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -40,6 +40,14 @@ config X86_RUN_64BIT runs through the 16-bit and 32-bit init, then switches to 64-bit mode and jumps to U-Boot proper.
+config X86_RUN_64BIT_ONLY
Perhaps X86_RUN_64BIT_NO_SPL? Bin might hav a better idea.
I considered X64_RUN_64BIT_NO_SPL, X64_RUN_64BIT_PURE and
X64_RUN_64BIT_NATIVE.
Hi Bin, can you recommend better naming?
Cannot figure out a better name :) Let's use it.
Thanks Bin. Let me use X86_RUN_64BIT_ONLY for now.
Hi Simon, is this naming okay for you?
Reviewed-by: Bin Meng bmeng.cn@gmail.com
Regards, Bin
Best Regards, Aiden
Hi Aiden,
On Fri, 1 May 2020 at 12:36, Park, Aiden aiden.park@intel.com wrote:
Hi Bin,
-----Original Message----- From: Bin Meng bmeng.cn@gmail.com Sent: Thursday, April 30, 2020 2:14 AM To: Park, Aiden aiden.park@intel.com Cc: Simon Glass sjg@chromium.org; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 1/8] x86: Add a new X86_RUN_64BIT_ONLY to Kconfig
Hi Aiden,
On Wed, Apr 29, 2020 at 1:44 PM Park, Aiden aiden.park@intel.com wrote:
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 1/8] x86: Add a new X86_RUN_64BIT_ONLY to Kconfig
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This will build U-Boot as a pure 64-bit binary with no SPL. It can be used with a pre-stage boot firmware which has already done 16-bit, 32-bit and 64-bit init.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/Kconfig | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index c8eae24c07..89add17e69 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -40,6 +40,14 @@ config X86_RUN_64BIT runs through the 16-bit and 32-bit init, then switches to 64-bit mode and jumps to U-Boot proper.
+config X86_RUN_64BIT_ONLY
Perhaps X86_RUN_64BIT_NO_SPL? Bin might hav a better idea.
I considered X64_RUN_64BIT_NO_SPL, X64_RUN_64BIT_PURE and
X64_RUN_64BIT_NATIVE.
Hi Bin, can you recommend better naming?
Cannot figure out a better name :) Let's use it.
Thanks Bin. Let me use X86_RUN_64BIT_ONLY for now.
Hi Simon, is this naming okay for you?
Yes that's fine thanks.
Regards, Simon
From: Aiden Park aiden.park@intel.com
This will allow a board or cpu to do its specific initialization at 64-bit entry if U-Boot is a pure 64-bit binary.
Signed-off-by: Aiden Park aiden.park@intel.com --- arch/x86/cpu/start64.S | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/arch/x86/cpu/start64.S b/arch/x86/cpu/start64.S index 7be834788b..b8ac5aab57 100644 --- a/arch/x86/cpu/start64.S +++ b/arch/x86/cpu/start64.S @@ -13,6 +13,11 @@ .globl _start .type _start, @function _start: +#if defined(CONFIG_X86_RUN_64BIT_ONLY) + jmp init_64bit_entry +.globl init_64bit_entry_ret +init_64bit_entry_ret: +#endif /* Set up memory using the existing stack */ mov %rsp, %rdi call board_init_f_alloc_reserve
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This will allow a board or cpu to do its specific initialization at 64-bit entry if U-Boot is a pure 64-bit binary.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/cpu/start64.S | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/arch/x86/cpu/start64.S b/arch/x86/cpu/start64.S index 7be834788b..b8ac5aab57 100644 --- a/arch/x86/cpu/start64.S +++ b/arch/x86/cpu/start64.S @@ -13,6 +13,11 @@ .globl _start .type _start, @function _start: +#if defined(CONFIG_X86_RUN_64BIT_ONLY)
jmp init_64bit_entry+.globl init_64bit_entry_ret +init_64bit_entry_ret: +#endif
Would it be safe to do this always? Is that a standard register to have the HOB pointer be in? If so, we could always store it. We could also make sure that U-Boot passes it on too.
Note that U-Boot has a struct arch_spl_handoff which is how it passes the HOB through.
/* Set up memory using the existing stack */ mov %rsp, %rdi call board_init_f_alloc_reserve-- 2.20.1
Regards, Simon
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 2/8] x86: start64: Add a hook at 64-bit entry
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This will allow a board or cpu to do its specific initialization at 64-bit entry if U-Boot is a pure 64-bit binary.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/cpu/start64.S | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/arch/x86/cpu/start64.S b/arch/x86/cpu/start64.S index 7be834788b..b8ac5aab57 100644 --- a/arch/x86/cpu/start64.S +++ b/arch/x86/cpu/start64.S @@ -13,6 +13,11 @@ .globl _start .type _start, @function _start: +#if defined(CONFIG_X86_RUN_64BIT_ONLY)
jmp init_64bit_entry+.globl init_64bit_entry_ret +init_64bit_entry_ret: +#endif
Would it be safe to do this always? Is that a standard register to have the HOB pointer be in? If so, we could always store it. We could also make sure that U- Boot passes it on too.
Unfortunately, it's not. This is to cover pre-stage firmware's calling convention. The calling convention of 64-bit Slim Bootloader is that HOB address is in rcx.
Note that U-Boot has a struct arch_spl_handoff which is how it passes the HOB through.
Thanks for your information. Let me try to leverage arch_spl_handoff to pass HOB.
/* Set up memory using the existing stack */ mov %rsp, %rdi call board_init_f_alloc_reserve-- 2.20.1
Regards, Simon
Best Regards, Aiden
From: Aiden Park aiden.park@intel.com
This will allow 64-bit U-Boot to use HOB.
Signed-off-by: Aiden Park aiden.park@intel.com --- arch/x86/cpu/start64.S | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/arch/x86/cpu/start64.S b/arch/x86/cpu/start64.S index b8ac5aab57..7faf4f6f13 100644 --- a/arch/x86/cpu/start64.S +++ b/arch/x86/cpu/start64.S @@ -25,6 +25,11 @@ init_64bit_entry_ret:
call board_init_f_init_reserve
+#if defined(CONFIG_X86_RUN_64BIT_ONLY) && defined(CONFIG_USE_HOB) + mov %r10, %rdi + call set_hob_list +#endif + xor %rdi, %rdi call board_init_f call board_init_f_r
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This will allow 64-bit U-Boot to use HOB.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/cpu/start64.S | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/arch/x86/cpu/start64.S b/arch/x86/cpu/start64.S index b8ac5aab57..7faf4f6f13 100644 --- a/arch/x86/cpu/start64.S +++ b/arch/x86/cpu/start64.S @@ -25,6 +25,11 @@ init_64bit_entry_ret:
call board_init_f_init_reserve+#if defined(CONFIG_X86_RUN_64BIT_ONLY) && defined(CONFIG_USE_HOB)
mov %r10, %rdicall set_hob_list+#endif
Similar question - would it be safe to always do this?
xor %rdi, %rdi call board_init_f call board_init_f_r-- 2.20.1
Regards, Simon
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 3/8] x86: start64: Support HOB in 64-bit U-Boot
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This will allow 64-bit U-Boot to use HOB.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/cpu/start64.S | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/arch/x86/cpu/start64.S b/arch/x86/cpu/start64.S index b8ac5aab57..7faf4f6f13 100644 --- a/arch/x86/cpu/start64.S +++ b/arch/x86/cpu/start64.S @@ -25,6 +25,11 @@ init_64bit_entry_ret:
call board_init_f_init_reserve+#if defined(CONFIG_X86_RUN_64BIT_ONLY) && defined(CONFIG_USE_HOB)
mov %r10, %rdicall set_hob_list+#endif
Similar question - would it be safe to always do this?
You are right. The r10 is a scratch register. The r12-r15 would be correct choice. Instead of this, let me try to leverage arch_spl_handoff.
xor %rdi, %rdi call board_init_f call board_init_f_r-- 2.20.1
Regards, Simon
Best Regards, Aiden
From: Aiden Park aiden.park@intel.com
This supports 64-bit U-Boot as a Slim Bootloader payload.
Signed-off-by: Aiden Park aiden.park@intel.com --- arch/x86/cpu/slimbootloader/Makefile | 9 +++++++-- arch/x86/cpu/slimbootloader/entry64.S | 14 ++++++++++++++ arch/x86/cpu/slimbootloader/slimbootloader.c | 17 +++++++++++++++-- 3 files changed, 36 insertions(+), 4 deletions(-) create mode 100644 arch/x86/cpu/slimbootloader/entry64.S
diff --git a/arch/x86/cpu/slimbootloader/Makefile b/arch/x86/cpu/slimbootloader/Makefile index aac9fa3db8..79fa699501 100644 --- a/arch/x86/cpu/slimbootloader/Makefile +++ b/arch/x86/cpu/slimbootloader/Makefile @@ -1,5 +1,10 @@ # SPDX-License-Identifier: GPL-2.0+ # -# Copyright (C) 2019 Intel Corporation <www.intel.com> +# Copyright (C) 2019-2020 Intel Corporation <www.intel.com>
-obj-y += car.o slimbootloader.o sdram.o serial.o +ifeq ($(CONFIG_X86_64),y) +obj-y += entry64.o +else +obj-y += car.o +endif +obj-y += slimbootloader.o sdram.o serial.o diff --git a/arch/x86/cpu/slimbootloader/entry64.S b/arch/x86/cpu/slimbootloader/entry64.S new file mode 100644 index 0000000000..5e101e18a9 --- /dev/null +++ b/arch/x86/cpu/slimbootloader/entry64.S @@ -0,0 +1,14 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +/* + * Copyright (C) 2020 Intel Corporation <www.intel.com> + */ + +#include <generated/asm-offsets.h> + +.section .text + +.globl init_64bit_entry +init_64bit_entry: + /* Save hob pointer parameter */ + mov %rcx, %r10 + jmp init_64bit_entry_ret diff --git a/arch/x86/cpu/slimbootloader/slimbootloader.c b/arch/x86/cpu/slimbootloader/slimbootloader.c index 21dcfb2142..7857e4cd8b 100644 --- a/arch/x86/cpu/slimbootloader/slimbootloader.c +++ b/arch/x86/cpu/slimbootloader/slimbootloader.c @@ -1,6 +1,6 @@ // SPDX-License-Identifier: GPL-2.0+ /* - * Copyright (C) 2019 Intel Corporation <www.intel.com> + * Copyright (C) 2019-2020 Intel Corporation <www.intel.com> */
#include <common.h> @@ -43,11 +43,23 @@ static void tsc_init(void)
int arch_cpu_init(void) { + int ret = 0; + tsc_init();
- return x86_cpu_init_f(); +#if !CONFIG_IS_ENABLED(X86_64) + ret = x86_cpu_init_f(); +#endif + return ret; }
+#if CONFIG_IS_ENABLED(X86_64) +int set_hob_list(void *hob_list) +{ + gd->arch.hob_list = hob_list; + return 0; +} +#else int checkcpu(void) { return 0; @@ -57,3 +69,4 @@ int print_cpuinfo(void) { return default_print_cpuinfo(); } +#endif
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This supports 64-bit U-Boot as a Slim Bootloader payload.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/cpu/slimbootloader/Makefile | 9 +++++++-- arch/x86/cpu/slimbootloader/entry64.S | 14 ++++++++++++++ arch/x86/cpu/slimbootloader/slimbootloader.c | 17 +++++++++++++++-- 3 files changed, 36 insertions(+), 4 deletions(-) create mode 100644 arch/x86/cpu/slimbootloader/entry64.S
diff --git a/arch/x86/cpu/slimbootloader/Makefile b/arch/x86/cpu/slimbootloader/Makefile index aac9fa3db8..79fa699501 100644 --- a/arch/x86/cpu/slimbootloader/Makefile +++ b/arch/x86/cpu/slimbootloader/Makefile @@ -1,5 +1,10 @@ # SPDX-License-Identifier: GPL-2.0+ # -# Copyright (C) 2019 Intel Corporation <www.intel.com> +# Copyright (C) 2019-2020 Intel Corporation <www.intel.com>
-obj-y += car.o slimbootloader.o sdram.o serial.o +ifeq ($(CONFIG_X86_64),y) +obj-y += entry64.o +else +obj-y += car.o +endif +obj-y += slimbootloader.o sdram.o serial.o diff --git a/arch/x86/cpu/slimbootloader/entry64.S b/arch/x86/cpu/slimbootloader/entry64.S new file mode 100644 index 0000000000..5e101e18a9 --- /dev/null +++ b/arch/x86/cpu/slimbootloader/entry64.S @@ -0,0 +1,14 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +/*
- Copyright (C) 2020 Intel Corporation <www.intel.com>
- */
+#include <generated/asm-offsets.h>
+.section .text
+.globl init_64bit_entry +init_64bit_entry:
/* Save hob pointer parameter */mov %rcx, %r10jmp init_64bit_entry_retdiff --git a/arch/x86/cpu/slimbootloader/slimbootloader.c b/arch/x86/cpu/slimbootloader/slimbootloader.c index 21dcfb2142..7857e4cd8b 100644 --- a/arch/x86/cpu/slimbootloader/slimbootloader.c +++ b/arch/x86/cpu/slimbootloader/slimbootloader.c @@ -1,6 +1,6 @@ // SPDX-License-Identifier: GPL-2.0+ /*
- Copyright (C) 2019 Intel Corporation <www.intel.com>
*/
- Copyright (C) 2019-2020 Intel Corporation <www.intel.com>
#include <common.h> @@ -43,11 +43,23 @@ static void tsc_init(void)
int arch_cpu_init(void) {
int ret = 0;tsc_init();
return x86_cpu_init_f();+#if !CONFIG_IS_ENABLED(X86_64)
Can you use if() instead of #if ?
ret = x86_cpu_init_f();+#endif
return ret;}
+#if CONFIG_IS_ENABLED(X86_64)
It should be safe to define both of these functions so I don't think you need the #ifdef
+int set_hob_list(void *hob_list) +{
gd->arch.hob_list = hob_list;return 0;+} +#else int checkcpu(void) { return 0; @@ -57,3 +69,4 @@ int print_cpuinfo(void) { return default_print_cpuinfo(); }
+#endif
2.20.1
Regards, Simon
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 4/8] x86: slimbootloader: Support 64-bit operation
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This supports 64-bit U-Boot as a Slim Bootloader payload.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/cpu/slimbootloader/Makefile | 9 +++++++-- arch/x86/cpu/slimbootloader/entry64.S | 14 ++++++++++++++ arch/x86/cpu/slimbootloader/slimbootloader.c | 17 +++++++++++++++-- 3 files changed, 36 insertions(+), 4 deletions(-) create mode 100644 arch/x86/cpu/slimbootloader/entry64.S
diff --git a/arch/x86/cpu/slimbootloader/Makefile b/arch/x86/cpu/slimbootloader/Makefile index aac9fa3db8..79fa699501 100644 --- a/arch/x86/cpu/slimbootloader/Makefile +++ b/arch/x86/cpu/slimbootloader/Makefile @@ -1,5 +1,10 @@ # SPDX-License-Identifier: GPL-2.0+ # -# Copyright (C) 2019 Intel Corporation <www.intel.com> +# Copyright (C) 2019-2020 Intel Corporation <www.intel.com>
-obj-y += car.o slimbootloader.o sdram.o serial.o +ifeq ($(CONFIG_X86_64),y) +obj-y += entry64.o +else +obj-y += car.o +endif +obj-y += slimbootloader.o sdram.o serial.o diff --git a/arch/x86/cpu/slimbootloader/entry64.S b/arch/x86/cpu/slimbootloader/entry64.S new file mode 100644 index 0000000000..5e101e18a9 --- /dev/null +++ b/arch/x86/cpu/slimbootloader/entry64.S @@ -0,0 +1,14 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +/*
- Copyright (C) 2020 Intel Corporation <www.intel.com> */
+#include <generated/asm-offsets.h>
+.section .text
+.globl init_64bit_entry +init_64bit_entry:
/* Save hob pointer parameter */mov %rcx, %r10jmp init_64bit_entry_retdiff --git a/arch/x86/cpu/slimbootloader/slimbootloader.c b/arch/x86/cpu/slimbootloader/slimbootloader.c index 21dcfb2142..7857e4cd8b 100644 --- a/arch/x86/cpu/slimbootloader/slimbootloader.c +++ b/arch/x86/cpu/slimbootloader/slimbootloader.c @@ -1,6 +1,6 @@ // SPDX-License-Identifier: GPL-2.0+ /*
- Copyright (C) 2019 Intel Corporation <www.intel.com>
*/
- Copyright (C) 2019-2020 Intel Corporation <www.intel.com>
#include <common.h> @@ -43,11 +43,23 @@ static void tsc_init(void)
int arch_cpu_init(void) {
int ret = 0;tsc_init();
return x86_cpu_init_f();+#if !CONFIG_IS_ENABLED(X86_64)
Can you use if() instead of #if ?
I will do it.
ret = x86_cpu_init_f();+#endif
return ret;}
+#if CONFIG_IS_ENABLED(X86_64)
It should be safe to define both of these functions so I don't think you need the #ifdef
These are defined in arch/x86/cpu/x86_64/cpu.c already. Is it okay to make them weak reference or can I keep this as it is?
+int set_hob_list(void *hob_list) +{
gd->arch.hob_list = hob_list;return 0;+} +#else int checkcpu(void) { return 0; @@ -57,3 +69,4 @@ int print_cpuinfo(void) { return default_print_cpuinfo(); }
+#endif
2.20.1
Regards, Simon
Best Regards, Aiden
Hi Aiden,
On Wed, 29 Apr 2020 at 00:01, Park, Aiden aiden.park@intel.com wrote:
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 4/8] x86: slimbootloader: Support 64-bit operation
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This supports 64-bit U-Boot as a Slim Bootloader payload.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/cpu/slimbootloader/Makefile | 9 +++++++-- arch/x86/cpu/slimbootloader/entry64.S | 14 ++++++++++++++ arch/x86/cpu/slimbootloader/slimbootloader.c | 17 +++++++++++++++-- 3 files changed, 36 insertions(+), 4 deletions(-) create mode 100644 arch/x86/cpu/slimbootloader/entry64.S
diff --git a/arch/x86/cpu/slimbootloader/Makefile b/arch/x86/cpu/slimbootloader/Makefile index aac9fa3db8..79fa699501 100644 --- a/arch/x86/cpu/slimbootloader/Makefile +++ b/arch/x86/cpu/slimbootloader/Makefile @@ -1,5 +1,10 @@ # SPDX-License-Identifier: GPL-2.0+ # -# Copyright (C) 2019 Intel Corporation <www.intel.com> +# Copyright (C) 2019-2020 Intel Corporation <www.intel.com>
-obj-y += car.o slimbootloader.o sdram.o serial.o +ifeq ($(CONFIG_X86_64),y) +obj-y += entry64.o +else +obj-y += car.o +endif +obj-y += slimbootloader.o sdram.o serial.o diff --git a/arch/x86/cpu/slimbootloader/entry64.S b/arch/x86/cpu/slimbootloader/entry64.S new file mode 100644 index 0000000000..5e101e18a9 --- /dev/null +++ b/arch/x86/cpu/slimbootloader/entry64.S @@ -0,0 +1,14 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +/*
- Copyright (C) 2020 Intel Corporation <www.intel.com> */
+#include <generated/asm-offsets.h>
+.section .text
+.globl init_64bit_entry +init_64bit_entry:
/* Save hob pointer parameter */mov %rcx, %r10jmp init_64bit_entry_retdiff --git a/arch/x86/cpu/slimbootloader/slimbootloader.c b/arch/x86/cpu/slimbootloader/slimbootloader.c index 21dcfb2142..7857e4cd8b 100644 --- a/arch/x86/cpu/slimbootloader/slimbootloader.c +++ b/arch/x86/cpu/slimbootloader/slimbootloader.c @@ -1,6 +1,6 @@ // SPDX-License-Identifier: GPL-2.0+ /*
- Copyright (C) 2019 Intel Corporation <www.intel.com>
*/
- Copyright (C) 2019-2020 Intel Corporation <www.intel.com>
#include <common.h> @@ -43,11 +43,23 @@ static void tsc_init(void)
int arch_cpu_init(void) {
int ret = 0;tsc_init();
return x86_cpu_init_f();+#if !CONFIG_IS_ENABLED(X86_64)
Can you use if() instead of #if ?
I will do it.
ret = x86_cpu_init_f();+#endif
return ret;}
+#if CONFIG_IS_ENABLED(X86_64)
It should be safe to define both of these functions so I don't think you need the #ifdef
These are defined in arch/x86/cpu/x86_64/cpu.c already. Is it okay to make them weak reference or can I keep this as it is?
Oh I see. I am not a fan of weak functions so perhaps we should keep them as is.
+int set_hob_list(void *hob_list) +{
gd->arch.hob_list = hob_list;return 0;+} +#else int checkcpu(void) { return 0; @@ -57,3 +69,4 @@ int print_cpuinfo(void) { return default_print_cpuinfo(); }
+#endif
2.20.1
Regards, Simon
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Wednesday, April 29, 2020 11:04 AM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 4/8] x86: slimbootloader: Support 64-bit operation
Hi Aiden,
On Wed, 29 Apr 2020 at 00:01, Park, Aiden aiden.park@intel.com wrote:
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 4/8] x86: slimbootloader: Support 64-bit operation
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This supports 64-bit U-Boot as a Slim Bootloader payload.
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/cpu/slimbootloader/Makefile | 9 +++++++-- arch/x86/cpu/slimbootloader/entry64.S | 14 ++++++++++++++ arch/x86/cpu/slimbootloader/slimbootloader.c | 17 +++++++++++++++-- 3 files changed, 36 insertions(+), 4 deletions(-) create mode 100644 arch/x86/cpu/slimbootloader/entry64.S
diff --git a/arch/x86/cpu/slimbootloader/Makefile b/arch/x86/cpu/slimbootloader/Makefile index aac9fa3db8..79fa699501 100644 --- a/arch/x86/cpu/slimbootloader/Makefile +++ b/arch/x86/cpu/slimbootloader/Makefile @@ -1,5 +1,10 @@ # SPDX-License-Identifier: GPL-2.0+ # -# Copyright (C) 2019 Intel Corporation <www.intel.com> +# Copyright (C) 2019-2020 Intel Corporation <www.intel.com>
-obj-y += car.o slimbootloader.o sdram.o serial.o +ifeq ($(CONFIG_X86_64),y) +obj-y += entry64.o +else +obj-y += car.o +endif +obj-y += slimbootloader.o sdram.o serial.o diff --git a/arch/x86/cpu/slimbootloader/entry64.S b/arch/x86/cpu/slimbootloader/entry64.S new file mode 100644 index 0000000000..5e101e18a9 --- /dev/null +++ b/arch/x86/cpu/slimbootloader/entry64.S @@ -0,0 +1,14 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +/*
- Copyright (C) 2020 Intel Corporation <www.intel.com> */
+#include <generated/asm-offsets.h>
+.section .text
+.globl init_64bit_entry +init_64bit_entry:
/* Save hob pointer parameter */mov %rcx, %r10jmp init_64bit_entry_retdiff --git a/arch/x86/cpu/slimbootloader/slimbootloader.c b/arch/x86/cpu/slimbootloader/slimbootloader.c index 21dcfb2142..7857e4cd8b 100644 --- a/arch/x86/cpu/slimbootloader/slimbootloader.c +++ b/arch/x86/cpu/slimbootloader/slimbootloader.c @@ -1,6 +1,6 @@ // SPDX-License-Identifier: GPL-2.0+ /*
- Copyright (C) 2019 Intel Corporation <www.intel.com>
*/
- Copyright (C) 2019-2020 Intel Corporation <www.intel.com>
#include <common.h> @@ -43,11 +43,23 @@ static void tsc_init(void)
int arch_cpu_init(void) {
int ret = 0;tsc_init();
return x86_cpu_init_f();+#if !CONFIG_IS_ENABLED(X86_64)
Can you use if() instead of #if ?
I will do it.
ret = x86_cpu_init_f();+#endif
return ret;}
+#if CONFIG_IS_ENABLED(X86_64)
It should be safe to define both of these functions so I don't think you need the #ifdef
These are defined in arch/x86/cpu/x86_64/cpu.c already. Is it okay to make them weak reference or can I keep this as it is?
Oh I see. I am not a fan of weak functions so perhaps we should keep them as is.
I agree with you about the weak functions. Let me keep this as is. Thanks.
+int set_hob_list(void *hob_list) +{
gd->arch.hob_list = hob_list;return 0;+} +#else int checkcpu(void) { return 0; @@ -57,3 +69,4 @@ int print_cpuinfo(void) { return default_print_cpuinfo(); }
+#endif
2.20.1
Regards, Simon
Best Regards, Aiden
From: Aiden Park aiden.park@intel.com
Add slimbootloader-x86_64_defconfig for 64-bit slimbootloader board.
Signed-off-by: Aiden Park aiden.park@intel.com --- configs/slimbootloader-x86_64_defconfig | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 configs/slimbootloader-x86_64_defconfig
diff --git a/configs/slimbootloader-x86_64_defconfig b/configs/slimbootloader-x86_64_defconfig new file mode 100644 index 0000000000..f046c30179 --- /dev/null +++ b/configs/slimbootloader-x86_64_defconfig @@ -0,0 +1,24 @@ +CONFIG_X86=y +CONFIG_X86_RUN_64BIT_ONLY=y +CONFIG_ENV_SIZE=0x1000 +CONFIG_VENDOR_INTEL=y +CONFIG_TARGET_SLIMBOOTLOADER=y +# CONFIG_USE_CAR is not set +CONFIG_BOOTSTAGE=y +CONFIG_BOOTSTAGE_REPORT=y +CONFIG_BOOTDELAY=10 +CONFIG_SYS_CONSOLE_INFO_QUIET=y +CONFIG_BOARD_EARLY_INIT_R=y +CONFIG_LAST_STAGE_INIT=y +CONFIG_HUSH_PARSER=y +CONFIG_CMD_MMC=y +CONFIG_CMD_USB=y +CONFIG_CMD_EXT2=y +CONFIG_CMD_FAT=y +CONFIG_EFI_PARTITION=y +CONFIG_DEFAULT_DEVICE_TREE="slimbootloader" +CONFIG_SYS_RELOC_GD_ENV_ADDR=y +CONFIG_REGMAP=y +CONFIG_SYSCON=y +# CONFIG_PCI_PNP is not set +CONFIG_CONSOLE_SCROLL_LINES=5
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
Add slimbootloader-x86_64_defconfig for 64-bit slimbootloader board.
Signed-off-by: Aiden Park aiden.park@intel.com
configs/slimbootloader-x86_64_defconfig | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 configs/slimbootloader-x86_64_defconfig
How about slimbootloader64?
We always know it is x86.
Regards, Simon
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:32 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 5/8] configs: slimbootloader: Add x86_64 slimbootloader config
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
Add slimbootloader-x86_64_defconfig for 64-bit slimbootloader board.
Signed-off-by: Aiden Park aiden.park@intel.com
configs/slimbootloader-x86_64_defconfig | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 configs/slimbootloader-x86_64_defconfig
How about slimbootloader64?
We always know it is x86.
Thanks for your suggestion. I will rename it to slimbootloader64.
Regards, Simon
Best Regards, Aiden
On Mon, Apr 27, 2020 at 4:32 AM Simon Glass sjg@chromium.org wrote:
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
Add slimbootloader-x86_64_defconfig for 64-bit slimbootloader board.
Signed-off-by: Aiden Park aiden.park@intel.com
configs/slimbootloader-x86_64_defconfig | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 configs/slimbootloader-x86_64_defconfig
How about slimbootloader64?
We always know it is x86.
Agreed with Simon,
Aiden, please also make sure the defconfig file is written with all options in the correct order. You can do it like this:
$ make slimbootloader64_defconfig $ make savedefconfig $ cp defconfig configs/slimbootloader64_defconfig
Regards, Bin
From: Aiden Park aiden.park@intel.com
This is to fix pointer type cast warning in hob command.
Signed-off-by: Aiden Park aiden.park@intel.com --- cmd/x86/hob.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/cmd/x86/hob.c b/cmd/x86/hob.c index 3967a7ca5a..8c2c1ffc63 100644 --- a/cmd/x86/hob.c +++ b/cmd/x86/hob.c @@ -36,14 +36,14 @@ static int do_hob(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
hdr = gd->arch.hob_list;
- printf("HOB list address: 0x%08x\n\n", (unsigned int)hdr); + printf("HOB list address: 0x%p\n\n", hdr);
- printf("# | Address | Type | Len | "); + printf("# | Address | Type | Len | "); printf("%36s\n", "GUID"); - printf("---|----------|-----------|------|-"); + printf("---|------------------|-----------|------|-"); printf("------------------------------------\n"); while (!end_of_hob(hdr)) { - printf("%02x | %08x | ", i, (unsigned int)hdr); + printf("%02x | %p | ", i, hdr); type = hdr->type; if (type == HOB_TYPE_UNUSED) desc = "*Unused*";
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This is to fix pointer type cast warning in hob command.
Signed-off-by: Aiden Park aiden.park@intel.com
cmd/x86/hob.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/cmd/x86/hob.c b/cmd/x86/hob.c index 3967a7ca5a..8c2c1ffc63 100644 --- a/cmd/x86/hob.c +++ b/cmd/x86/hob.c @@ -36,14 +36,14 @@ static int do_hob(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
hdr = gd->arch.hob_list;
printf("HOB list address: 0x%08x\n\n", (unsigned int)hdr);
printf("HOB list address: 0x%p\n\n", hdr);
That 0x should not be there as hex is assumed in U-Boot.
Do we actually need a 64-bit pointer here, or will the HOB be below 4GB? I'm really not a fan of 16-character addresses in a user command.
printf("# | Address | Type | Len | ");
printf("# | Address | Type | Len | "); printf("%36s\n", "GUID");
printf("---|----------|-----------|------|-");
printf("---|------------------|-----------|------|-"); printf("------------------------------------\n"); while (!end_of_hob(hdr)) {
printf("%02x | %08x | ", i, (unsigned int)hdr);
printf("%02x | %p | ", i, hdr); type = hdr->type; if (type == HOB_TYPE_UNUSED) desc = "*Unused*";-- 2.20.1
Regards, Simon
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 12:46 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 6/8] x86: Fix 64-bit compile warning
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This is to fix pointer type cast warning in hob command.
Signed-off-by: Aiden Park aiden.park@intel.com
cmd/x86/hob.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/cmd/x86/hob.c b/cmd/x86/hob.c index 3967a7ca5a..8c2c1ffc63 100644 --- a/cmd/x86/hob.c +++ b/cmd/x86/hob.c @@ -36,14 +36,14 @@ static int do_hob(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
hdr = gd->arch.hob_list;
printf("HOB list address: 0x%08x\n\n", (unsigned int)hdr);
printf("HOB list address: 0x%p\n\n", hdr);That 0x should not be there as hex is assumed in U-Boot.
Thanks. Let me remove the 0x prefix.
Do we actually need a 64-bit pointer here, or will the HOB be below 4GB? I'm really not a fan of 16-character addresses in a user command.
Currently, HOB is always below 4GB. This was just in case for 64-bit address space. Let me assume HOB < 4GB for now.
printf("# | Address | Type | Len | ");
printf("# | Address | Type | Len | "); printf("%36s\n", "GUID");
printf("---|----------|-----------|------|-");
printf("---|------------------|-----------|------|-"); printf("------------------------------------\n"); while (!end_of_hob(hdr)) {
printf("%02x | %08x | ", i, (unsigned int)hdr);
printf("%02x | %p | ", i, hdr); type = hdr->type; if (type == HOB_TYPE_UNUSED) desc = "*Unused*";-- 2.20.1
Regards, Simon
Best Regards, Aiden
From: Aiden Park aiden.park@intel.com
Signed-off-by: Aiden Park aiden.park@intel.com --- arch/x86/cpu/x86_64/Makefile | 2 +- arch/x86/cpu/x86_64/call64.S | 21 +++++++++++++++++++++ arch/x86/cpu/x86_64/cpu.c | 10 ++++++++++ arch/x86/include/asm/bootparam.h | 10 +++++++++- arch/x86/include/asm/zimage.h | 2 +- arch/x86/lib/bootm.c | 10 +++++++--- arch/x86/lib/zimage.c | 24 ++++++++++++++++-------- 7 files changed, 65 insertions(+), 14 deletions(-) create mode 100644 arch/x86/cpu/x86_64/call64.S
diff --git a/arch/x86/cpu/x86_64/Makefile b/arch/x86/cpu/x86_64/Makefile index 400f0ffe39..951e503a1f 100644 --- a/arch/x86/cpu/x86_64/Makefile +++ b/arch/x86/cpu/x86_64/Makefile @@ -3,4 +3,4 @@ # Written by Simon Glass sjg@chromium.org #
-obj-y += cpu.o interrupts.o setjmp.o +obj-y += cpu.o interrupts.o setjmp.o call64.o diff --git a/arch/x86/cpu/x86_64/call64.S b/arch/x86/cpu/x86_64/call64.S new file mode 100644 index 0000000000..e2cfe15080 --- /dev/null +++ b/arch/x86/cpu/x86_64/call64.S @@ -0,0 +1,21 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +/* + * Copyright (C) 2020 Intel Corporation <www.intel.com> + */ + +.code64 +.globl cpu_call64 +cpu_call64: + /* + * cpu_call64(ulong pgtable, ulong setup_base, ulong target) + * + * rdi - pgtable (unused - already in 64-bit with paging) + * rsi - setup_base + * rdx - target + * + */ + cli + mov %rdx, %rcx + mov %rsi, %rdx + call *%rcx + ret diff --git a/arch/x86/cpu/x86_64/cpu.c b/arch/x86/cpu/x86_64/cpu.c index 90925e46ea..0730c43f1c 100644 --- a/arch/x86/cpu/x86_64/cpu.c +++ b/arch/x86/cpu/x86_64/cpu.c @@ -7,6 +7,7 @@ #include <common.h> #include <cpu_func.h> #include <debug_uart.h> +#include <asm/cpu.h>
/* * Global declaration of gd. @@ -67,3 +68,12 @@ int x86_cpu_reinit_f(void) { return 0; } + +int cpu_jump_to_64bit(ulong setup_base, ulong load_address) +{ + ulong target = load_address + 0x200; + + cpu_call64(0, setup_base, target); + + return -EFAULT; +} diff --git a/arch/x86/include/asm/bootparam.h b/arch/x86/include/asm/bootparam.h index d961dddc9e..73c94a33ec 100644 --- a/arch/x86/include/asm/bootparam.h +++ b/arch/x86/include/asm/bootparam.h @@ -59,7 +59,15 @@ struct setup_header { __u32 initrd_addr_max; __u32 kernel_alignment; __u8 relocatable_kernel; - __u8 _pad2[3]; + __u8 min_alignment; + __u16 xloadflags; +#define XLF_KERNEL_64 BIT(0) +#define XLF_CAN_BE_LOADED_ABOVE_4G BIT(1) +#define XLF_EFI_HANDOVER_32 BIT(2) +#define XLF_EFI_HANDOVER_64 BIT(3) +#define XLF_EFI_KEXEC BIT(4) +#define XLF_5LEVEL BIT(5) +#define XLF_5LEVEL_ENABLED BIT(6) __u32 cmdline_size; __u32 hardware_subarch; __u64 hardware_subarch_data; diff --git a/arch/x86/include/asm/zimage.h b/arch/x86/include/asm/zimage.h index 80e128ccf3..cadeb04168 100644 --- a/arch/x86/include/asm/zimage.h +++ b/arch/x86/include/asm/zimage.h @@ -31,7 +31,7 @@ #define ZIMAGE_LOAD_ADDR 0x10000
struct boot_params *load_zimage(char *image, unsigned long kernel_size, - ulong *load_addressp); + ulong *load_addressp, bool *image_64bit); int setup_zimage(struct boot_params *setup_base, char *cmd_line, int auto_boot, unsigned long initrd_addr, unsigned long initrd_size); void setup_video(struct screen_info *screen_info); diff --git a/arch/x86/lib/bootm.c b/arch/x86/lib/bootm.c index 07d8f1f279..ebed9e4605 100644 --- a/arch/x86/lib/bootm.c +++ b/arch/x86/lib/bootm.c @@ -74,6 +74,7 @@ static int boot_prep_linux(bootm_headers_t *images) void *data = NULL; size_t len; int ret; + bool image_64bit;
#ifdef CONFIG_OF_LIBFDT if (images->ft_len) { @@ -116,7 +117,8 @@ static int boot_prep_linux(bootm_headers_t *images) ulong load_address; char *base_ptr;
- base_ptr = (char *)load_zimage(data, len, &load_address); + base_ptr = (char *)load_zimage(data, len, &load_address, + &image_64bit); if (!base_ptr) { puts("## Kernel loading failed ...\n"); goto error; @@ -124,6 +126,10 @@ static int boot_prep_linux(bootm_headers_t *images) images->os.load = load_address; cmd_line_dest = base_ptr + COMMAND_LINE_OFFSET; images->ep = (ulong)base_ptr; +#if CONFIG_IS_ENABLED(X86_64) + if (image_64bit) + images->os.arch = IH_ARCH_X86_64; +#endif } else if (images->ep) { cmd_line_dest = (void *)images->ep + COMMAND_LINE_OFFSET; } else { @@ -164,9 +170,7 @@ int boot_linux_kernel(ulong setup_base, ulong load_address, bool image_64bit) * TODO(sjg@chromium.org): Support booting both 32-bit and * 64-bit kernels from 64-bit U-Boot. */ -#if !CONFIG_IS_ENABLED(X86_64) return cpu_jump_to_64bit(setup_base, load_address); -#endif } else { /* * Set %ebx, %ebp, and %edi to 0, %esi to point to the diff --git a/arch/x86/lib/zimage.c b/arch/x86/lib/zimage.c index ffc09630b7..c84e2d39b2 100644 --- a/arch/x86/lib/zimage.c +++ b/arch/x86/lib/zimage.c @@ -129,7 +129,7 @@ static int setup_device_tree(struct setup_header *hdr, const void *fdt_blob) }
struct boot_params *load_zimage(char *image, unsigned long kernel_size, - ulong *load_addressp) + ulong *load_addressp, bool *image_64bit) { struct boot_params *setup_base; int setup_size; @@ -179,6 +179,9 @@ struct boot_params *load_zimage(char *image, unsigned long kernel_size, big_image = (bootproto >= 0x0200) && (hdr->loadflags & BIG_KERNEL_FLAG);
+ /* Determine 64-bit kernel */ + *image_64bit = (hdr->xloadflags & XLF_KERNEL_64) ? true : false; + /* Determine load address */ if (big_image) *load_addressp = BZIMAGE_LOAD_ADDR; @@ -313,12 +316,13 @@ void __setup_pcat_compatibility(void) int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]) { struct boot_params *base_ptr; - void *bzImage_addr = NULL; + void *bzimage_addr = NULL; ulong load_address; char *s; - ulong bzImage_size = 0; + ulong bzimage_size = 0; ulong initrd_addr = 0; ulong initrd_size = 0; + bool image_64bit;
disable_interrupts();
@@ -333,11 +337,11 @@ int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]) }
if (s) - bzImage_addr = (void *)simple_strtoul(s, NULL, 16); + bzimage_addr = (void *)simple_strtoul(s, NULL, 16);
if (argc >= 3) { /* argv[2] holds the size of the bzImage */ - bzImage_size = simple_strtoul(argv[2], NULL, 16); + bzimage_size = simple_strtoul(argv[2], NULL, 16); }
if (argc >= 4) @@ -346,8 +350,13 @@ int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]) initrd_size = simple_strtoul(argv[4], NULL, 16);
/* Lets look for */ - base_ptr = load_zimage(bzImage_addr, bzImage_size, &load_address); + base_ptr = load_zimage(bzimage_addr, bzimage_size, &load_address, + &image_64bit); +#if !CONFIG_IS_ENABLED(X86_64) + image_64bit = false; +#endif
+ /* we assume that the kernel is in place */ if (!base_ptr) { puts("## Kernel loading failed ...\n"); return -1; @@ -358,8 +367,7 @@ int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]) return -1; }
- /* we assume that the kernel is in place */ - return boot_linux_kernel((ulong)base_ptr, load_address, false); + return boot_linux_kernel((ulong)base_ptr, load_address, image_64bit); }
U_BOOT_CMD(
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/cpu/x86_64/Makefile | 2 +- arch/x86/cpu/x86_64/call64.S | 21 +++++++++++++++++++++ arch/x86/cpu/x86_64/cpu.c | 10 ++++++++++ arch/x86/include/asm/bootparam.h | 10 +++++++++- arch/x86/include/asm/zimage.h | 2 +- arch/x86/lib/bootm.c | 10 +++++++--- arch/x86/lib/zimage.c | 24 ++++++++++++++++-------- 7 files changed, 65 insertions(+), 14 deletions(-) create mode 100644 arch/x86/cpu/x86_64/call64.S
diff --git a/arch/x86/cpu/x86_64/Makefile b/arch/x86/cpu/x86_64/Makefile index 400f0ffe39..951e503a1f 100644 --- a/arch/x86/cpu/x86_64/Makefile +++ b/arch/x86/cpu/x86_64/Makefile @@ -3,4 +3,4 @@ # Written by Simon Glass sjg@chromium.org #
-obj-y += cpu.o interrupts.o setjmp.o +obj-y += cpu.o interrupts.o setjmp.o call64.o diff --git a/arch/x86/cpu/x86_64/call64.S b/arch/x86/cpu/x86_64/call64.S new file mode 100644 index 0000000000..e2cfe15080 --- /dev/null +++ b/arch/x86/cpu/x86_64/call64.S @@ -0,0 +1,21 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +/*
- Copyright (C) 2020 Intel Corporation <www.intel.com>
- */
+.code64 +.globl cpu_call64 +cpu_call64:
/** cpu_call64(ulong pgtable, ulong setup_base, ulong target)** rdi - pgtable (unused - already in 64-bit with paging)* rsi - setup_base* rdx - target**/climov %rdx, %rcxmov %rsi, %rdxcall *%rcxretdiff --git a/arch/x86/cpu/x86_64/cpu.c b/arch/x86/cpu/x86_64/cpu.c index 90925e46ea..0730c43f1c 100644 --- a/arch/x86/cpu/x86_64/cpu.c +++ b/arch/x86/cpu/x86_64/cpu.c @@ -7,6 +7,7 @@ #include <common.h> #include <cpu_func.h> #include <debug_uart.h> +#include <asm/cpu.h>
/*
- Global declaration of gd.
@@ -67,3 +68,12 @@ int x86_cpu_reinit_f(void) { return 0; }
+int cpu_jump_to_64bit(ulong setup_base, ulong load_address) +{
ulong target = load_address + 0x200;cpu_call64(0, setup_base, target);return -EFAULT;+} diff --git a/arch/x86/include/asm/bootparam.h b/arch/x86/include/asm/bootparam.h index d961dddc9e..73c94a33ec 100644 --- a/arch/x86/include/asm/bootparam.h +++ b/arch/x86/include/asm/bootparam.h @@ -59,7 +59,15 @@ struct setup_header { __u32 initrd_addr_max; __u32 kernel_alignment; __u8 relocatable_kernel;
__u8 _pad2[3];
__u8 min_alignment;__u16 xloadflags;+#define XLF_KERNEL_64 BIT(0) +#define XLF_CAN_BE_LOADED_ABOVE_4G BIT(1) +#define XLF_EFI_HANDOVER_32 BIT(2) +#define XLF_EFI_HANDOVER_64 BIT(3) +#define XLF_EFI_KEXEC BIT(4) +#define XLF_5LEVEL BIT(5) +#define XLF_5LEVEL_ENABLED BIT(6) __u32 cmdline_size; __u32 hardware_subarch; __u64 hardware_subarch_data; diff --git a/arch/x86/include/asm/zimage.h b/arch/x86/include/asm/zimage.h index 80e128ccf3..cadeb04168 100644 --- a/arch/x86/include/asm/zimage.h +++ b/arch/x86/include/asm/zimage.h @@ -31,7 +31,7 @@ #define ZIMAGE_LOAD_ADDR 0x10000
struct boot_params *load_zimage(char *image, unsigned long kernel_size,
ulong *load_addressp);
ulong *load_addressp, bool *image_64bit);
Please put a 'p' on the end, so image_64bitp, since it a return value.
int setup_zimage(struct boot_params *setup_base, char *cmd_line, int auto_boot, unsigned long initrd_addr, unsigned long initrd_size); void setup_video(struct screen_info *screen_info); diff --git a/arch/x86/lib/bootm.c b/arch/x86/lib/bootm.c index 07d8f1f279..ebed9e4605 100644 --- a/arch/x86/lib/bootm.c +++ b/arch/x86/lib/bootm.c @@ -74,6 +74,7 @@ static int boot_prep_linux(bootm_headers_t *images) void *data = NULL; size_t len; int ret;
bool image_64bit;#ifdef CONFIG_OF_LIBFDT if (images->ft_len) { @@ -116,7 +117,8 @@ static int boot_prep_linux(bootm_headers_t *images) ulong load_address; char *base_ptr;
base_ptr = (char *)load_zimage(data, len, &load_address);
base_ptr = (char *)load_zimage(data, len, &load_address,&image_64bit); if (!base_ptr) { puts("## Kernel loading failed ...\n"); goto error;@@ -124,6 +126,10 @@ static int boot_prep_linux(bootm_headers_t *images) images->os.load = load_address; cmd_line_dest = base_ptr + COMMAND_LINE_OFFSET; images->ep = (ulong)base_ptr; +#if CONFIG_IS_ENABLED(X86_64)
if (image_64bit)
if (CONFIG_IS_ENABLED(X86_64) && image_64bit)
images->os.arch = IH_ARCH_X86_64;+#endif } else if (images->ep) { cmd_line_dest = (void *)images->ep + COMMAND_LINE_OFFSET; } else { @@ -164,9 +170,7 @@ int boot_linux_kernel(ulong setup_base, ulong load_address, bool image_64bit) * TODO(sjg@chromium.org): Support booting both 32-bit and * 64-bit kernels from 64-bit U-Boot. */ -#if !CONFIG_IS_ENABLED(X86_64)
if()
return cpu_jump_to_64bit(setup_base, load_address);-#endif } else { /* * Set %ebx, %ebp, and %edi to 0, %esi to point to the diff --git a/arch/x86/lib/zimage.c b/arch/x86/lib/zimage.c index ffc09630b7..c84e2d39b2 100644 --- a/arch/x86/lib/zimage.c +++ b/arch/x86/lib/zimage.c @@ -129,7 +129,7 @@ static int setup_device_tree(struct setup_header *hdr, const void *fdt_blob) }
struct boot_params *load_zimage(char *image, unsigned long kernel_size,
ulong *load_addressp)
ulong *load_addressp, bool *image_64bit)
image_64bitp
{ struct boot_params *setup_base; int setup_size; @@ -179,6 +179,9 @@ struct boot_params *load_zimage(char *image, unsigned long kernel_size, big_image = (bootproto >= 0x0200) && (hdr->loadflags & BIG_KERNEL_FLAG);
/* Determine 64-bit kernel */*image_64bit = (hdr->xloadflags & XLF_KERNEL_64) ? true : false;
Drop ()
In fact I suspect you can do:
*image_64bit = hdr->xloadflags & XLF_KERNEL_64)
/* Determine load address */ if (big_image) *load_addressp = BZIMAGE_LOAD_ADDR;@@ -313,12 +316,13 @@ void __setup_pcat_compatibility(void) int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]) { struct boot_params *base_ptr;
void *bzImage_addr = NULL;
void *bzimage_addr = NULL;
Good change, but really should be in a separate patch.
ulong load_address; char *s;
ulong bzImage_size = 0;
ulong bzimage_size = 0; ulong initrd_addr = 0; ulong initrd_size = 0;bool image_64bit; disable_interrupts();@@ -333,11 +337,11 @@ int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]) }
if (s)
bzImage_addr = (void *)simple_strtoul(s, NULL, 16);
bzimage_addr = (void *)simple_strtoul(s, NULL, 16); if (argc >= 3) { /* argv[2] holds the size of the bzImage */
bzImage_size = simple_strtoul(argv[2], NULL, 16);
bzimage_size = simple_strtoul(argv[2], NULL, 16); } if (argc >= 4)@@ -346,8 +350,13 @@ int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]) initrd_size = simple_strtoul(argv[4], NULL, 16);
/* Lets look for */
base_ptr = load_zimage(bzImage_addr, bzImage_size, &load_address);
base_ptr = load_zimage(bzimage_addr, bzimage_size, &load_address,&image_64bit);+#if !CONFIG_IS_ENABLED(X86_64)
if()
But why not just set it always?
image_64bit = false;+#endif
/* we assume that the kernel is in place */ if (!base_ptr) { puts("## Kernel loading failed ...\n"); return -1;@@ -358,8 +367,7 @@ int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]) return -1; }
/* we assume that the kernel is in place */return boot_linux_kernel((ulong)base_ptr, load_address, false);
return boot_linux_kernel((ulong)base_ptr, load_address, image_64bit);}
U_BOOT_CMD(
2.20.1
Regards, Simon
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 7/8] x86: enable 64-bit kernel boot from 64-bit U-Boot
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
Signed-off-by: Aiden Park aiden.park@intel.com
arch/x86/cpu/x86_64/Makefile | 2 +- arch/x86/cpu/x86_64/call64.S | 21 +++++++++++++++++++++ arch/x86/cpu/x86_64/cpu.c | 10 ++++++++++ arch/x86/include/asm/bootparam.h | 10 +++++++++- arch/x86/include/asm/zimage.h | 2 +- arch/x86/lib/bootm.c | 10 +++++++--- arch/x86/lib/zimage.c | 24 ++++++++++++++++-------- 7 files changed, 65 insertions(+), 14 deletions(-) create mode 100644 arch/x86/cpu/x86_64/call64.S
diff --git a/arch/x86/cpu/x86_64/Makefile b/arch/x86/cpu/x86_64/Makefile index 400f0ffe39..951e503a1f 100644 --- a/arch/x86/cpu/x86_64/Makefile +++ b/arch/x86/cpu/x86_64/Makefile @@ -3,4 +3,4 @@ # Written by Simon Glass sjg@chromium.org #
-obj-y += cpu.o interrupts.o setjmp.o +obj-y += cpu.o interrupts.o setjmp.o call64.o diff --git a/arch/x86/cpu/x86_64/call64.S b/arch/x86/cpu/x86_64/call64.S new file mode 100644 index 0000000000..e2cfe15080 --- /dev/null +++ b/arch/x86/cpu/x86_64/call64.S @@ -0,0 +1,21 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +/*
- Copyright (C) 2020 Intel Corporation <www.intel.com> */
+.code64 +.globl cpu_call64 +cpu_call64:
/** cpu_call64(ulong pgtable, ulong setup_base, ulong target)** rdi - pgtable (unused - already in 64-bit with paging)* rsi - setup_base* rdx - target**/climov %rdx, %rcxmov %rsi, %rdxcall *%rcxretdiff --git a/arch/x86/cpu/x86_64/cpu.c b/arch/x86/cpu/x86_64/cpu.c index 90925e46ea..0730c43f1c 100644 --- a/arch/x86/cpu/x86_64/cpu.c +++ b/arch/x86/cpu/x86_64/cpu.c @@ -7,6 +7,7 @@ #include <common.h> #include <cpu_func.h> #include <debug_uart.h> +#include <asm/cpu.h>
/*
- Global declaration of gd.
@@ -67,3 +68,12 @@ int x86_cpu_reinit_f(void) { return 0; }
+int cpu_jump_to_64bit(ulong setup_base, ulong load_address) {
ulong target = load_address + 0x200;cpu_call64(0, setup_base, target);return -EFAULT;+} diff --git a/arch/x86/include/asm/bootparam.h b/arch/x86/include/asm/bootparam.h index d961dddc9e..73c94a33ec 100644 --- a/arch/x86/include/asm/bootparam.h +++ b/arch/x86/include/asm/bootparam.h @@ -59,7 +59,15 @@ struct setup_header { __u32 initrd_addr_max; __u32 kernel_alignment; __u8 relocatable_kernel;
__u8 _pad2[3];
__u8 min_alignment;__u16 xloadflags;+#define XLF_KERNEL_64 BIT(0) +#define XLF_CAN_BE_LOADED_ABOVE_4G BIT(1) +#define XLF_EFI_HANDOVER_32 BIT(2) +#define XLF_EFI_HANDOVER_64 BIT(3) +#define XLF_EFI_KEXEC BIT(4) +#define XLF_5LEVEL BIT(5) +#define XLF_5LEVEL_ENABLED BIT(6) __u32 cmdline_size; __u32 hardware_subarch; __u64 hardware_subarch_data; diff --git a/arch/x86/include/asm/zimage.h b/arch/x86/include/asm/zimage.h index 80e128ccf3..cadeb04168 100644 --- a/arch/x86/include/asm/zimage.h +++ b/arch/x86/include/asm/zimage.h @@ -31,7 +31,7 @@ #define ZIMAGE_LOAD_ADDR 0x10000
struct boot_params *load_zimage(char *image, unsigned long kernel_size,
ulong *load_addressp);
ulong *load_addressp, bool- *image_64bit);
Please put a 'p' on the end, so image_64bitp, since it a return value.
Thanks. I will do that.
int setup_zimage(struct boot_params *setup_base, char *cmd_line, int
auto_boot,
unsigned long initrd_addr, unsigned longinitrd_size); void setup_video(struct screen_info *screen_info); diff --git a/arch/x86/lib/bootm.c b/arch/x86/lib/bootm.c index 07d8f1f279..ebed9e4605 100644 --- a/arch/x86/lib/bootm.c +++ b/arch/x86/lib/bootm.c @@ -74,6 +74,7 @@ static int boot_prep_linux(bootm_headers_t *images) void *data = NULL; size_t len; int ret;
bool image_64bit;#ifdef CONFIG_OF_LIBFDT if (images->ft_len) { @@ -116,7 +117,8 @@ static int boot_prep_linux(bootm_headers_t *images) ulong load_address; char *base_ptr;
base_ptr = (char *)load_zimage(data, len, &load_address);
base_ptr = (char *)load_zimage(data, len, &load_address,&image_64bit); if (!base_ptr) { puts("## Kernel loading failed ...\n"); goto error;@@ -124,6 +126,10 @@ static int boot_prep_linux(bootm_headers_t
*images)
images->os.load = load_address; cmd_line_dest = base_ptr + COMMAND_LINE_OFFSET; images->ep = (ulong)base_ptr;+#if CONFIG_IS_ENABLED(X86_64)
if (image_64bit)if (CONFIG_IS_ENABLED(X86_64) && image_64bit)
images->os.arch = IH_ARCH_X86_64; #endif } else if (images->ep) { cmd_line_dest = (void *)images->ep + COMMAND_LINE_OFFSET; } else {@@ -164,9 +170,7 @@ int boot_linux_kernel(ulong setup_base, ulong
load_address, bool image_64bit)
* TODO(sjg@chromium.org): Support booting both 32-bit and * 64-bit kernels from 64-bit U-Boot. */-#if !CONFIG_IS_ENABLED(X86_64)
if()
I will do that.
return cpu_jump_to_64bit(setup_base, load_address);-#endif } else { /* * Set %ebx, %ebp, and %edi to 0, %esi to point to the diff --git a/arch/x86/lib/zimage.c b/arch/x86/lib/zimage.c index ffc09630b7..c84e2d39b2 100644 --- a/arch/x86/lib/zimage.c +++ b/arch/x86/lib/zimage.c @@ -129,7 +129,7 @@ static int setup_device_tree(struct setup_header *hdr, const void *fdt_blob) }
struct boot_params *load_zimage(char *image, unsigned long kernel_size,
ulong *load_addressp)
ulong *load_addressp, bool- *image_64bit)
image_64bitp
I will do that.
{ struct boot_params *setup_base; int setup_size; @@ -179,6 +179,9 @@ struct boot_params *load_zimage(char *image,
unsigned long kernel_size,
big_image = (bootproto >= 0x0200) && (hdr->loadflags & BIG_KERNEL_FLAG);
/* Determine 64-bit kernel */*image_64bit = (hdr->xloadflags & XLF_KERNEL_64) ? true :- false;
Drop ()
In fact I suspect you can do:
*image_64bit = hdr->xloadflags & XLF_KERNEL_64)
I will do that.
/* Determine load address */ if (big_image) *load_addressp = BZIMAGE_LOAD_ADDR; @@ -313,12 +316,13@@ void __setup_pcat_compatibility(void) int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]) { struct boot_params *base_ptr;
void *bzImage_addr = NULL;
void *bzimage_addr = NULL;Good change, but really should be in a separate patch.
Let me make a separate patch for this.
ulong load_address; char *s;
ulong bzImage_size = 0;
ulong bzimage_size = 0; ulong initrd_addr = 0; ulong initrd_size = 0;bool image_64bit; disable_interrupts();@@ -333,11 +337,11 @@ int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc,
char *const argv[])
} if (s)
bzImage_addr = (void *)simple_strtoul(s, NULL, 16);
bzimage_addr = (void *)simple_strtoul(s, NULL, 16); if (argc >= 3) { /* argv[2] holds the size of the bzImage */
bzImage_size = simple_strtoul(argv[2], NULL, 16);
bzimage_size = simple_strtoul(argv[2], NULL, 16); } if (argc >= 4)@@ -346,8 +350,13 @@ int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc,
char *const argv[])
initrd_size = simple_strtoul(argv[4], NULL, 16); /* Lets look for */
base_ptr = load_zimage(bzImage_addr, bzImage_size, &load_address);
base_ptr = load_zimage(bzimage_addr, bzimage_size, &load_address,&image_64bit); #if+!CONFIG_IS_ENABLED(X86_64)
if()
But why not just set it always?
This was just to keep original boot flow in zboot to boot 32-bit kernel entry from 32-bit U-Boot. I also agree with you. 32-bit U-Boot is able to boot 64-bit kernel entry with 64-bit paging. Let me remove this.
image_64bit = false;+#endif
/* we assume that the kernel is in place */ if (!base_ptr) { puts("## Kernel loading failed ...\n"); return -1;@@ -358,8 +367,7 @@ int do_zboot(cmd_tbl_t *cmdtp, int flag, int argc,
char *const argv[])
return -1; }
/* we assume that the kernel is in place */return boot_linux_kernel((ulong)base_ptr, load_address, false);
return boot_linux_kernel((ulong)base_ptr, load_address,- image_64bit);
}
U_BOOT_CMD(
2.20.1
Regards, Simon
Best Regards, Aiden
From: Aiden Park aiden.park@intel.com
Add steps to build 64-bit Slim Bootloader and U-Boot.
Signed-off-by: Aiden Park aiden.park@intel.com --- doc/board/intel/slimbootloader.rst | 29 +++++++++++++++++++++++++++++ 1 file changed, 29 insertions(+)
diff --git a/doc/board/intel/slimbootloader.rst b/doc/board/intel/slimbootloader.rst index a8c41b1aa7..779ed78ef1 100644 --- a/doc/board/intel/slimbootloader.rst +++ b/doc/board/intel/slimbootloader.rst @@ -145,6 +145,35 @@ Also, the PayloadId needs to be set for APL board.
Use DediProg to flash IFWI. You should reach at U-Boot serial console.
+Build Instruction for 64-bit Slim Bootloader & U-Boot on QEMU target +-------------------------------------------------------------------- + +1. Build 64-bit U-Boot and obtain u-boot-dtb.bin:: + + $ make distclean + $ make slimbootloader-x86_64_defconfig + $ make all + +2. Copy u-boot-dtb.bin to Slim Bootloader:: + + $ mkdir -p <Slim Bootloader Dir>/PayloadPkg/PayloadBins/ + $ cp <U-Boot Dir>/u-boot-dtb.bin <Slim Bootloader Dir>/PayloadPkg/PayloadBins/u-boot-dtb.bin + +3. Update PayloadId with 'U-BT':: + + $ vi Platform/QemuBoardPkg/CfgData/CfgDataExt_Brd1.dlt + -GEN_CFG_DATA.PayloadId | 'AUTO' + +GEN_CFG_DATA.PayloadId | 'U-BT' + +4. Update payload text base:: + + $ vi Platform/QemuBoardPkg/BoardConfig.py + + self.PAYLOAD_LOAD_HIGH = 0 + + self.PAYLOAD_EXE_BASE = 0x00100000 + +5. Build QEMU target:: + + $ python BuildLoader.py build qemu -a x64 -p "OsLoader.efi:LLDR:Lz4;u-boot-dtb.bin:U-BT:Lzma"
Build Instruction to use ELF U-Boot -----------------------------------
On Wed, Apr 22, 2020 at 8:45 AM aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
Add steps to build 64-bit Slim Bootloader and U-Boot.
Signed-off-by: Aiden Park aiden.park@intel.com
doc/board/intel/slimbootloader.rst | 29 +++++++++++++++++++++++++++++ 1 file changed, 29 insertions(+)
Reviewed-by: Bin Meng bmeng.cn@gmail.com
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This patchset will build 64-bit U-Boot as a Slim Bootloader payload and introduce a new X86_RUN_64BIT_ONLY to build pure 64-bit U-Boot binary without SPL. In addition, this supports 64-bit kernel booting from 64-bit U-Boot.
Aiden Park (8): x86: Add a new X86_RUN_64BIT_ONLY to Kconfig x86: start64: Add a hook at 64-bit entry x86: start64: Support HOB in 64-bit U-Boot x86: slimbootloader: Support 64-bit operation configs: slimbootloader: Add x86_64 slimbootloader config x86: Fix 64-bit compile warning x86: enable 64-bit kernel boot from 64-bit U-Boot doc: slimbootloader: Update 64-bit build instruction
arch/x86/Kconfig | 8 ++++++ arch/x86/cpu/slimbootloader/Makefile | 9 ++++-- arch/x86/cpu/slimbootloader/entry64.S | 14 ++++++++++ arch/x86/cpu/slimbootloader/slimbootloader.c | 17 ++++++++++-- arch/x86/cpu/start64.S | 10 +++++++ arch/x86/cpu/x86_64/Makefile | 2 +- arch/x86/cpu/x86_64/call64.S | 21 ++++++++++++++ arch/x86/cpu/x86_64/cpu.c | 10 +++++++ arch/x86/include/asm/bootparam.h | 10 ++++++- arch/x86/include/asm/zimage.h | 2 +- arch/x86/lib/bootm.c | 10 +++++-- arch/x86/lib/zimage.c | 24 ++++++++++------ cmd/x86/hob.c | 8 +++--- configs/slimbootloader-x86_64_defconfig | 24 ++++++++++++++++ doc/board/intel/slimbootloader.rst | 29 ++++++++++++++++++++ 15 files changed, 176 insertions(+), 22 deletions(-) create mode 100644 arch/x86/cpu/slimbootloader/entry64.S create mode 100644 arch/x86/cpu/x86_64/call64.S create mode 100644 configs/slimbootloader-x86_64_defconfig
For the series:
Tested on qemu: Tested-by: Simon Glass sjg@chromium.org
One little thing that I noticed...your instructions show copying u-boot-dtb.bin into the slimbootloader directory. This is the same as u-boot.bin and I suggest you use that name instead. When I tried to, I got an error.
Regards, SImon
Hi Simon,
-----Original Message----- From: Simon Glass sjg@chromium.org Sent: Sunday, April 26, 2020 1:16 PM To: Park, Aiden aiden.park@intel.com Cc: Bin Meng bmeng.cn@gmail.com; U-Boot Mailing List <u- boot@lists.denx.de> Subject: Re: [PATCH 0/8] Support 64-bit U-Boot for Slim Bootloader
Hi Aiden,
On Tue, 21 Apr 2020 at 18:45, aiden.park@intel.com wrote:
From: Aiden Park aiden.park@intel.com
This patchset will build 64-bit U-Boot as a Slim Bootloader payload and introduce a new X86_RUN_64BIT_ONLY to build pure 64-bit U-Boot binary without SPL. In addition, this supports 64-bit kernel booting from 64-bit U-Boot.
Aiden Park (8): x86: Add a new X86_RUN_64BIT_ONLY to Kconfig x86: start64: Add a hook at 64-bit entry x86: start64: Support HOB in 64-bit U-Boot x86: slimbootloader: Support 64-bit operation configs: slimbootloader: Add x86_64 slimbootloader config x86: Fix 64-bit compile warning x86: enable 64-bit kernel boot from 64-bit U-Boot doc: slimbootloader: Update 64-bit build instruction
arch/x86/Kconfig | 8 ++++++ arch/x86/cpu/slimbootloader/Makefile | 9 ++++-- arch/x86/cpu/slimbootloader/entry64.S | 14 ++++++++++ arch/x86/cpu/slimbootloader/slimbootloader.c | 17 ++++++++++-- arch/x86/cpu/start64.S | 10 +++++++ arch/x86/cpu/x86_64/Makefile | 2 +- arch/x86/cpu/x86_64/call64.S | 21 ++++++++++++++ arch/x86/cpu/x86_64/cpu.c | 10 +++++++ arch/x86/include/asm/bootparam.h | 10 ++++++- arch/x86/include/asm/zimage.h | 2 +- arch/x86/lib/bootm.c | 10 +++++-- arch/x86/lib/zimage.c | 24 ++++++++++------ cmd/x86/hob.c | 8 +++--- configs/slimbootloader-x86_64_defconfig | 24 ++++++++++++++++ doc/board/intel/slimbootloader.rst | 29 ++++++++++++++++++++ 15 files changed, 176 insertions(+), 22 deletions(-) create mode 100644 arch/x86/cpu/slimbootloader/entry64.S create mode 100644 arch/x86/cpu/x86_64/call64.S create mode 100644 configs/slimbootloader-x86_64_defconfig
For the series:
Tested on qemu: Tested-by: Simon Glass sjg@chromium.org
One little thing that I noticed...your instructions show copying u-boot-dtb.bin into the slimbootloader directory. This is the same as u-boot.bin and I suggest you use that name instead. When I tried to, I got an error.
Thanks for your testing. Let me update the instruction with u-boot.bin.
Regards, Simon
Best Regards, Aiden
participants (4)
-
aiden.park＠intel.com -
Bin Meng -
Park, Aiden
-
Simon Glass