Skip to contentSkip to conversation

efi: Improvements to U-Boot running on top of UEFI

PATCH Superseded Simon Glass 6 participants 82 messages Sep 8, 2021 – Sep 24, 2021

Subsystems (4)

v1 of 9 Superseded Latest is v9 →

lore
Share & tools
lore
#0 lore raw

cover efi: Improvements to U-Boot running on top of UEFI

At present U-Boot can be built as an EFI app, but it is really just for testing, with very few features. Instead, the payload build is used for booting on top of UEFI, where U-Boot takes over the machine immediately and supplies its own drivers.

But the app could be made more useful.

This series provides access to EFI block devices and the video console within U-Boot. It also restructures the app to make it possible to boot a kernel, although further work is needed to implement this.

This can be thought of as making U-Boot perform a little like 'grub', in that it runs purely based on UEFI-provided services.

Of course a lot more work is required, but this is a start.

Note: It would be very useful to include qemu tests of the app and stub in CI.

This is available at u-boot-dm/efi-working

Simon Glass (35):
x86: Keep symbol information in u-boot ELF file
x86: Create a new header for EFI
x86: Show some EFI info with the bdinfo command
x86: Tidy up global_data pointer for 64-bit
efi: Add a script for building and testing U-Boot on UEFI
x86: Create a 32/64-bit selection for the app
efi: Create a 64-bit app
x86: Don't duplicate global_ptr in 64-bit EFI app
efi: Add a way to obtain boot services in the app
efi: Add video support to the app
RFC: efi: Drop code that doesn't work with driver model
efi: Add EFI uclass for media
efi: Add a media/block driver for EFI block devices
efi: Locate all block devices in the app
patman: Use a ValueError exception if tools.Run() fails
binman: Report an error if test files fail to compile
binman: Support reading the offset of an ELF-file symbol
binman: Allow timeout to occur in the image or its section
binman: Tidy up comments on _DoTestFile()
binman: Support updating the dtb in an ELF file
efi: serial: Support arrow keys
bloblist: Move to rST format
bloblist: Support allocating the bloblist
x86: Allow booting a kernel from the EFI app
x86: Don't process the kernel command line unless enabled
x86: efi: Add room for the binman definition in the dtb
efi: Add comments to struct efi_priv
efi: Fix ll_boot_init() operation with the app
efi: Add a few comments to the stub
efi: Share struct efi_priv between the app and stub code
efi: Move exit_boot_services into a function
efi: Check for failure when initing the app
efi: Mention that efi_info_get() is only used in the stub
efi: Show when allocated pages are used
efi: Allow easy selection of serial-only operation

 arch/sandbox/dts/test.dts                     |   4 +
 arch/x86/config.mk                            |   4 +-
 arch/x86/cpu/efi/payload.c                    |  13 +-
 arch/x86/cpu/intel_common/Makefile            |   2 +-
 arch/x86/cpu/u-boot-64.lds                    |   2 +
 arch/x86/cpu/x86_64/Makefile                  |   4 +
 arch/x86/cpu/x86_64/cpu.c                     |  32 +---
 arch/x86/cpu/x86_64/misc.c                    |  41 ++++++
 arch/x86/dts/Makefile                         |   2 +-
 arch/x86/dts/efi-x86_app.dts                  |   4 +
 arch/x86/include/asm/efi.h                    |  39 +++++
 arch/x86/include/asm/global_data.h            |   2 +
 arch/x86/include/asm/zimage.h                 |   3 -
 arch/x86/lib/Makefile                         |   1 +
 arch/x86/lib/bdinfo.c                         |  22 +++
 arch/x86/lib/bootm.c                          |  11 +-
 arch/x86/lib/zimage.c                         |  14 +-
 board/efi/Kconfig                             |  15 +-
 board/efi/efi-x86_app/Kconfig                 |   6 +-
 board/efi/efi-x86_app/MAINTAINERS             |  11 +-
 common/Kconfig                                |  15 +-
 common/bloblist.c                             |  16 +-
 common/board_f.c                              |   8 +-
 ..._app_defconfig => efi-x86_app32_defconfig} |   2 +-
 configs/efi-x86_app64_defconfig               |  39 +++++
 doc/{README.bloblist => develop/bloblist.rst} |  20 ++-
 doc/develop/index.rst                         |   1 +
 doc/develop/uefi/u-boot_on_efi.rst            |   8 +-
 drivers/block/Kconfig                         |  33 +++++
 drivers/block/Makefile                        |   4 +
 drivers/block/blk-uclass.c                    |   2 +-
 drivers/block/efi-media-uclass.c              |  15 ++
 drivers/block/efi_blk.c                       | 115 +++++++++++++++
 drivers/block/sb_efi_media.c                  |  21 +++
 drivers/serial/serial_efi.c                   |  11 +-
 drivers/video/Kconfig                         |   2 +-
 drivers/video/efi.c                           |  45 +++++-
 include/configs/efi-x86_app.h                 |  31 +++-
 include/dm/uclass-id.h                        |   1 +
 include/efi.h                                 | 117 ++++++++++++++-
 include/efi_api.h                             |  15 ++
 include/init.h                                |   2 +-
 lib/efi/Kconfig                               |  34 ++++-
 lib/efi/efi.c                                 | 108 ++++++++++++++
 lib/efi/efi_app.c                             | 139 ++++++++++++++++--
 lib/efi/efi_stub.c                            |  95 +++++-------
 lib/efi_driver/Makefile                       |   2 +-
 lib/efi_loader/Kconfig                        |   2 +
 lib/efi_loader/efi_device_path.c              |  96 +++---------
 lib/efi_loader/efi_disk.c                     |  48 ------
 scripts/build-efi.sh                          | 111 ++++++++++++++
 test/dm/Makefile                              |   1 +
 test/dm/efi_media.c                           |  24 +++
 tools/binman/binman.rst                       |  36 +++++
 tools/binman/cmdline.py                       |   2 +
 tools/binman/control.py                       |  11 ++
 tools/binman/elf.py                           |  74 +++++++++-
 tools/binman/elf_test.py                      |  45 +++++-
 tools/binman/ftest.py                         |  91 +++++++++++-
 tools/binman/test/Makefile                    |  13 +-
 tools/binman/test/bss_data.c                  |   2 +-
 tools/binman/test/embed_data.c                |  16 ++
 tools/binman/test/embed_data.lds              |  23 +++
 tools/binman/test/u_boot_binman_embed.c       |  13 ++
 tools/binman/test/u_boot_binman_embed.lds     |  29 ++++
 tools/binman/test/u_boot_binman_embed_sm.c    |  13 ++
 tools/patman/tools.py                         |   2 +-
 67 files changed, 1492 insertions(+), 293 deletions(-)
 create mode 100644 arch/x86/cpu/x86_64/misc.c
 create mode 100644 arch/x86/include/asm/efi.h
 create mode 100644 arch/x86/lib/bdinfo.c
 rename configs/{efi-x86_app_defconfig => efi-x86_app32_defconfig} (97%)
 create mode 100644 configs/efi-x86_app64_defconfig
 rename doc/{README.bloblist => develop/bloblist.rst} (77%)
 create mode 100644 drivers/block/efi-media-uclass.c
 create mode 100644 drivers/block/efi_blk.c
 create mode 100644 drivers/block/sb_efi_media.c
 create mode 100755 scripts/build-efi.sh
 create mode 100644 test/dm/efi_media.c
 create mode 100644 tools/binman/test/embed_data.c
 create mode 100644 tools/binman/test/embed_data.lds
 create mode 100644 tools/binman/test/u_boot_binman_embed.c
 create mode 100644 tools/binman/test/u_boot_binman_embed.lds
 create mode 100644 tools/binman/test/u_boot_binman_embed_sm.c
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#1 lore raw

PATCH 1/35 x86: Keep symbol information in u-boot ELF file

At present this information is stripped when linking. It is useful to keep it around. Strip it from the .efi files instead.

---

 arch/x86/config.mk | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
arch/x86/config.mk +2 -2
index 7a8242562db..589f2aed2bc 100644
@@ -37,7 +37,7 @@ KBUILD_LDFLAGS += -m $(if $(IS_32BIT),elf_i386,elf_x86_64)
 LDFLAGS_EFI_PAYLOAD := -Bsymbolic -Bsymbolic-functions -shared --no-undefined -s
 
 OBJCOPYFLAGS_EFI := -j .text -j .sdata -j .data -j .dynamic -j .dynsym \
-	-j .rel -j .rela -j .reloc
+	-j .rel -j .rela -j .reloc --strip-all
 
 # Compiler flags to be added when building UEFI applications
 CFLAGS_EFI := -fpic -fshort-wchar
@@ -65,7 +65,7 @@ CPPFLAGS_crt0-efi-$(EFIARCH).o += $(CFLAGS_EFI)
 ifeq ($(CONFIG_EFI_APP),y)
 
 PLATFORM_CPPFLAGS += $(CFLAGS_EFI)
-LDFLAGS_FINAL += -znocombreloc -shared -s
+LDFLAGS_FINAL += -znocombreloc -shared
 LDSCRIPT := $(LDSCRIPT_EFI)
 
 else
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#2 lore raw

PATCH 2/35 x86: Create a new header for EFI

In reply to Simon Glass

The setup routines are called from zimage but don't really belong in the zimage header. Add a new EFI header to house these. Add comments so it is clear what the functions do.

---

 arch/x86/include/asm/efi.h    | 32 ++++++++++++++++++++++++++++++++
 arch/x86/include/asm/zimage.h |  3 ---
 arch/x86/lib/zimage.c         |  1 +
 3 files changed, 33 insertions(+), 3 deletions(-)
 create mode 100644 arch/x86/include/asm/efi.h
arch/x86/include/asm/efi.h new file +32 -0
new file mode 100644
index 00000000000..c1735e4dc5f
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Copyright Google LLC
+ */
+
+#ifndef _ASM_EFI_H_
+#define _ASM_EFI_H_
+
+struct efi_info;
+struct screen_info;
+
+/**
+ * setup_video() - Set up the screen info in the x86 setup
+ *
+ * This is needed so Linux can use the display (when U-Boot is an EFI payload)
+ *
+ * @efi_info: Pointer to place to put the screen info in the x86 setup base
+ */
+void setup_video(struct screen_info *screen_info);
+
+/**
+ * setup_efi_info() - Set up the EFI info needed by Linux to boot
+ *
+ * This writes a suitable signature, table pointers, memory-map pointer, etc.
+ * These are needed for Linux to boot from U-Boot (when U-Boot is an EFI
+ * payload).
+ *
+ * @efi_info: Pointer to place to put the EFI info in the x86 setup base
+ */
+void setup_efi_info(struct efi_info *efi_info);
+
+#endif
arch/x86/include/asm/zimage.h +0 -3
index 6679767d16b..fa6e7f76e05 100644
@@ -72,7 +72,4 @@ int setup_zimage(struct boot_params *setup_base, char *cmd_line, int auto_boot,
  */
 void zimage_dump(struct boot_params *base_ptr);
 
-void setup_video(struct screen_info *screen_info);
-void setup_efi_info(struct efi_info *efi_info);
-
 #endif
arch/x86/lib/zimage.c +1 -0
index 9938c80a42b..7ce02226ef9 100644
@@ -29,6 +29,7 @@
 #include <asm/byteorder.h>
 #include <asm/bootm.h>
 #include <asm/bootparam.h>
+#include <asm/efi.h>
 #include <asm/global_data.h>
 #ifdef CONFIG_SYS_COREBOOT
 #include <asm/arch/timestamp.h>
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#3 lore raw

Re: [PATCH 02/35] x86: Create a new header for EFI

Quoted 4 lines · Simon Glass

The setup routines are called from zimage but don't really belong in the zimage header. Add a new EFI header to house these. Add comments so it is clear what the functions do.

Quoted 8 lines

---

  arch/x86/include/asm/efi.h    | 32 ++++++++++++++++++++++++++++++++
  arch/x86/include/asm/zimage.h |  3 ---
  arch/x86/lib/zimage.c         |  1 +
  3 files changed, 33 insertions(+), 3 deletions(-)

create mode 100644 arch/x86/include/asm/efi.h

In the future we should be able to run the EFI app on any UEFI architecture. Given this background are the function definitions really x86 specific?

Best regards

Heinrich

Quoted 64 lines
arch/x86/include/asm/efi.h new file +32 -0
new file mode 100644
index 00000000000..c1735e4dc5f
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Copyright Google LLC
+ */
+
+#ifndef _ASM_EFI_H_
+#define _ASM_EFI_H_
+
+struct efi_info;
+struct screen_info;
+
+/**
+ * setup_video() - Set up the screen info in the x86 setup
+ *
+ * This is needed so Linux can use the display (when U-Boot is an EFI payload)
+ *
+ * @efi_info: Pointer to place to put the screen info in the x86 setup base
+ */
+void setup_video(struct screen_info *screen_info);
+
+/**
+ * setup_efi_info() - Set up the EFI info needed by Linux to boot
+ *
+ * This writes a suitable signature, table pointers, memory-map pointer, etc.
+ * These are needed for Linux to boot from U-Boot (when U-Boot is an EFI
+ * payload).
+ *
+ * @efi_info: Pointer to place to put the EFI info in the x86 setup base
+ */
+void setup_efi_info(struct efi_info *efi_info);
+
+#endif
arch/x86/include/asm/zimage.h +0 -3
index 6679767d16b..fa6e7f76e05 100644
@@ -72,7 +72,4 @@ int setup_zimage(struct boot_params *setup_base, char *cmd_line, int auto_boot,
   */
  void zimage_dump(struct boot_params *base_ptr);
 
-void setup_video(struct screen_info *screen_info);
-void setup_efi_info(struct efi_info *efi_info);
-
  #endif
arch/x86/lib/zimage.c +1 -0
index 9938c80a42b..7ce02226ef9 100644
@@ -29,6 +29,7 @@
  #include <asm/byteorder.h>
  #include <asm/bootm.h>
  #include <asm/bootparam.h>
+#include <asm/efi.h>
  #include <asm/global_data.h>
  #ifdef CONFIG_SYS_COREBOOT
  #include <asm/arch/timestamp.h>
#4 lore raw

Re: [PATCH 02/35] x86: Create a new header for EFI

Hi Heinrich,

Quoted 17 lines · Heinrich Schuchardt
Quoted 4 lines · Simon Glass

The setup routines are called from zimage but don't really belong in the zimage header. Add a new EFI header to house these. Add comments so it is clear what the functions do.

Quoted 8 lines

---

  arch/x86/include/asm/efi.h    | 32 ++++++++++++++++++++++++++++++++
  arch/x86/include/asm/zimage.h |  3 ---
  arch/x86/lib/zimage.c         |  1 +
  3 files changed, 33 insertions(+), 3 deletions(-)

create mode 100644 arch/x86/include/asm/efi.h

Quoted 3 lines

In the future we should be able to run the EFI app on any UEFI architecture. Given this background are the function definitions really x86 specific?

Well firstly that seems like that would be future work for someone. Secondly, this relates to zimage which is really only used by x86 as far as I am aware.

Regards,
Simon

#5 lore raw

Re: [PATCH 02/35] x86: Create a new header for EFI

Quoted 20 lines · Simon Glass

Hi Heinrich,

Quoted 17 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

The setup routines are called from zimage but don't really belong in the zimage header. Add a new EFI header to house these. Add comments so it is clear what the functions do.

---

   arch/x86/include/asm/efi.h    | 32 ++++++++++++++++++++++++++++++++
   arch/x86/include/asm/zimage.h |  3 ---
   arch/x86/lib/zimage.c         |  1 +
   3 files changed, 33 insertions(+), 3 deletions(-)

create mode 100644 arch/x86/include/asm/efi.h

Quoted 5 lines
Quoted 3 lines

In the future we should be able to run the EFI app on any UEFI architecture. Given this background are the function definitions really x86 specific?

Well firstly that seems like that would be future work for someone.

I did not expect that you do that work. But we should not make that work more difficult than necessary.

Quoted 2 lines

Secondly, this relates to zimage which is really only used by x86 as far as I am aware.

setup_video() and setup_efi_info() do not sound like zImage nor like x86 specific.

Best regards

Heinrich

#6 lore raw

Re: [PATCH 02/35] x86: Create a new header for EFI

Hi Heinrich,

Quoted 30 lines · On Thu, 9 Sep 2021, 10:25 Heinrich Schuchardt, <[email protected]> wrote:
Quoted 20 lines · Simon Glass

Hi Heinrich,

Quoted 17 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

The setup routines are called from zimage but don't really belong in the zimage header. Add a new EFI header to house these. Add comments so it is clear what the functions do.

---

   arch/x86/include/asm/efi.h    | 32 ++++++++++++++++++++++++++++++++
   arch/x86/include/asm/zimage.h |  3 ---
   arch/x86/lib/zimage.c         |  1 +
   3 files changed, 33 insertions(+), 3 deletions(-)

create mode 100644 arch/x86/include/asm/efi.h

Quoted 5 lines
Quoted 3 lines

In the future we should be able to run the EFI app on any UEFI architecture. Given this background are the function definitions really x86 specific?

Well firstly that seems like that would be future work for someone.

Quoted 8 lines

I did not expect that you do that work. But we should not make that work more difficult than necessary.

Quoted 2 lines

Secondly, this relates to zimage which is really only used by x86 as far as I am aware.

setup_video() and setup_efi_info() do not sound like zImage nor like x86 specific.

But actually they are. Please can you take a closer look? This is just using existing code...

Regards,
Simon

#7 lore raw

PATCH 3/35 x86: Show some EFI info with the bdinfo command

In reply to Simon Glass

It is useful to see some basic EFI info with the command as it forms part of the information about a board.

Add a hook for this and show the table address as a start.

While here, fix an invalid cast in setup_efi_info().

---

 arch/x86/cpu/efi/payload.c | 13 +++++++++++--
 arch/x86/include/asm/efi.h |  7 +++++++
 arch/x86/lib/Makefile      |  1 +
 arch/x86/lib/bdinfo.c      | 22 ++++++++++++++++++++++
 4 files changed, 41 insertions(+), 2 deletions(-)
 create mode 100644 arch/x86/lib/bdinfo.c
arch/x86/cpu/efi/payload.c +11 -2
index 9a73b768e9b..3a9f7d72868 100644
@@ -280,15 +280,24 @@ void setup_efi_info(struct efi_info *efi_info)
 	}
 	efi_info->efi_memdesc_size = map->desc_size;
 	efi_info->efi_memdesc_version = map->version;
-	efi_info->efi_memmap = (u32)(map->desc);
+	efi_info->efi_memmap = (ulong)(map->desc);
 	efi_info->efi_memmap_size = size - sizeof(struct efi_entry_memmap);
 
 #ifdef CONFIG_EFI_STUB_64BIT
 	efi_info->efi_systab_hi = table->sys_table >> 32;
-	efi_info->efi_memmap_hi = (u64)(u32)(map->desc) >> 32;
+	efi_info->efi_memmap_hi = (u64)(ulong)map->desc >> 32;
 	signature = EFI64_LOADER_SIGNATURE;
 #else
 	signature = EFI32_LOADER_SIGNATURE;
 #endif
 	memcpy(&efi_info->efi_loader_signature, signature, 4);
 }
+
+void efi_show_bdinfo(void)
+{
+	struct efi_entry_systable *table = NULL;
+	int size, ret;
+
+	ret = efi_info_get(EFIET_SYS_TABLE, (void **)&table, &size);
+	bdinfo_print_num_l("efi_table", (ulong)table);
+}
arch/x86/include/asm/efi.h +7 -0
index c1735e4dc5f..dfd858b78ba 100644
@@ -29,4 +29,11 @@ void setup_video(struct screen_info *screen_info);
  */
 void setup_efi_info(struct efi_info *efi_info);
 
+/**
+ * efi_show_bdinfo() - Show information about EFI for the 'bdinfo' command
+ *
+ * This looks up the EFI table pointer and shows related info
+ */
+void efi_show_bdinfo(void);
+
 #endif
arch/x86/lib/Makefile +1 -0
index 65d9b3bd6a3..18757b29aa9 100644
@@ -3,6 +3,7 @@
 # (C) Copyright 2002-2006
 # Wolfgang Denk, DENX Software Engineering, [email protected].
 
+obj-y	+= bdinfo.o
 ifndef CONFIG_X86_64
 ifndef CONFIG_TPL_BUILD
 obj-y += bios.o
arch/x86/lib/bdinfo.c new file +22 -0
new file mode 100644
index 00000000000..0cb79b01bd3
@@ -0,0 +1,22 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * x86-specific information for the 'bd' command
+ *
+ * Copyright 2021 Google LLC
+ */
+
+#include <common.h>
+#include <efi.h>
+#include <init.h>
+#include <asm/efi.h>
+#include <asm/global_data.h>
+
+DECLARE_GLOBAL_DATA_PTR;
+
+void arch_print_bdinfo(void)
+{
+	bdinfo_print_num_l("prev table", gd->arch.table);
+
+	if (IS_ENABLED(CONFIG_EFI_STUB))
+		efi_show_bdinfo();
+}
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#8 lore raw

Re: [PATCH 03/35] x86: Show some EFI info with the bdinfo command

Quoted 22 lines · Simon Glass

It is useful to see some basic EFI info with the command as it forms part of the information about a board.

Add a hook for this and show the table address as a start.

While here, fix an invalid cast in setup_efi_info().

---

  arch/x86/cpu/efi/payload.c | 13 +++++++++++--
  arch/x86/include/asm/efi.h |  7 +++++++
  arch/x86/lib/Makefile      |  1 +
  arch/x86/lib/bdinfo.c      | 22 ++++++++++++++++++++++
  4 files changed, 41 insertions(+), 2 deletions(-)

create mode 100644 arch/x86/lib/bdinfo.c

arch/x86/cpu/efi/payload.c +0 -0
index 9a73b768e9b..3a9f7d72868 100644

arch/x86/cpu/efi/payload.c void setup_efi_info(struct efi_info *efi_info) · line 283

Quoted 5 lines
arch/x86/cpu/efi/payload.c +1 -1
@@ -280,15 +280,24 @@ void setup_efi_info(struct efi_info *efi_info)
  	}
  	efi_info->efi_memdesc_size = map->desc_size;
  	efi_info->efi_memdesc_version = map->version;
-	efi_info->efi_memmap = (u32)(map->desc);
+	efi_info->efi_memmap = (ulong)(map->desc);

When converting a pointer to integer we use (uintptr_t).

arch/x86/include/asm/bootparam.h defines efi_info->efi_memmap as u32. This type is too small to hold a pointer.

arch/x86/cpu/efi/payload.c void setup_efi_info(struct efi_info *efi_info) · line 288

Quoted 6 lines
arch/x86/cpu/efi/payload.c +1 -1
  	efi_info->efi_memmap_size = size - sizeof(struct efi_entry_memmap);
 
  #ifdef CONFIG_EFI_STUB_64BIT
  	efi_info->efi_systab_hi = table->sys_table >> 32;
-	efi_info->efi_memmap_hi = (u64)(u32)(map->desc) >> 32;
+	efi_info->efi_memmap_hi = (u64)(ulong)map->desc >> 32;

You should not use two fields to hold a single pointer.

Please, replace all of these duplicate fields.

Best regards

Heinrich

Quoted 72 lines
arch/x86/cpu/efi/payload.c +9 -0
  	signature = EFI64_LOADER_SIGNATURE;
  #else
  	signature = EFI32_LOADER_SIGNATURE;
  #endif
  	memcpy(&efi_info->efi_loader_signature, signature, 4);
  }
+
+void efi_show_bdinfo(void)
+{
+	struct efi_entry_systable *table = NULL;
+	int size, ret;
+
+	ret = efi_info_get(EFIET_SYS_TABLE, (void **)&table, &size);
+	bdinfo_print_num_l("efi_table", (ulong)table);
+}
arch/x86/include/asm/efi.h +7 -0
index c1735e4dc5f..dfd858b78ba 100644
@@ -29,4 +29,11 @@ void setup_video(struct screen_info *screen_info);
   */
  void setup_efi_info(struct efi_info *efi_info);
 
+/**
+ * efi_show_bdinfo() - Show information about EFI for the 'bdinfo' command
+ *
+ * This looks up the EFI table pointer and shows related info
+ */
+void efi_show_bdinfo(void);
+
  #endif
arch/x86/lib/Makefile +1 -0
index 65d9b3bd6a3..18757b29aa9 100644
@@ -3,6 +3,7 @@
  # (C) Copyright 2002-2006
  # Wolfgang Denk, DENX Software Engineering, [email protected].
 
+obj-y	+= bdinfo.o
  ifndef CONFIG_X86_64
  ifndef CONFIG_TPL_BUILD
  obj-y += bios.o
arch/x86/lib/bdinfo.c new file +22 -0
new file mode 100644
index 00000000000..0cb79b01bd3
@@ -0,0 +1,22 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * x86-specific information for the 'bd' command
+ *
+ * Copyright 2021 Google LLC
+ */
+
+#include <common.h>
+#include <efi.h>
+#include <init.h>
+#include <asm/efi.h>
+#include <asm/global_data.h>
+
+DECLARE_GLOBAL_DATA_PTR;
+
+void arch_print_bdinfo(void)
+{
+	bdinfo_print_num_l("prev table", gd->arch.table);
+
+	if (IS_ENABLED(CONFIG_EFI_STUB))
+		efi_show_bdinfo();
+}
#9 lore raw

Re: [PATCH 03/35] x86: Show some EFI info with the bdinfo command

Hi Heinrich,

Quoted 21 lines · Heinrich Schuchardt
Quoted 17 lines · Simon Glass

It is useful to see some basic EFI info with the command as it forms part of the information about a board.

Add a hook for this and show the table address as a start.

While here, fix an invalid cast in setup_efi_info().

---

  arch/x86/cpu/efi/payload.c | 13 +++++++++++--
  arch/x86/include/asm/efi.h |  7 +++++++
  arch/x86/lib/Makefile      |  1 +
  arch/x86/lib/bdinfo.c      | 22 ++++++++++++++++++++++
  4 files changed, 41 insertions(+), 2 deletions(-)

create mode 100644 arch/x86/lib/bdinfo.c

Quoted 12 lines

arch/x86/cpu/efi/payload.c void setup_efi_info(struct efi_info *efi_info) · line 283

Quoted 10 lines
arch/x86/cpu/efi/payload.c +1 -1
index 9a73b768e9b..3a9f7d72868 100644
@@ -280,15 +280,24 @@ void setup_efi_info(struct efi_info *efi_info)
      }
      efi_info->efi_memdesc_size = map->desc_size;
      efi_info->efi_memdesc_version = map->version;
-     efi_info->efi_memmap = (u32)(map->desc);
+     efi_info->efi_memmap = (ulong)(map->desc);

When converting a pointer to integer we use (uintptr_t).

Generally in U-Boot it is ulong so I try to be consistent here. Anyway, u32 is clearly not right for a 64-bit build as it gives warnings.

Quoted 6 lines

arch/x86/include/asm/bootparam.h defines efi_info->efi_memmap as u32. This type is too small to hold a pointer.

Quoted 2 lines
arch/x86/cpu/efi/payload.c +0 -0
      efi_info->efi_memmap_size = size - sizeof(struct efi_entry_memmap);
 
Quoted 8 lines

arch/x86/cpu/efi/payload.c void setup_efi_info(struct efi_info *efi_info) · line 288

Quoted 4 lines
arch/x86/cpu/efi/payload.c +1 -1
  #ifdef CONFIG_EFI_STUB_64BIT
      efi_info->efi_systab_hi = table->sys_table >> 32;
-     efi_info->efi_memmap_hi = (u64)(u32)(map->desc) >> 32;
+     efi_info->efi_memmap_hi = (u64)(ulong)map->desc >> 32;

You should not use two fields to hold a single pointer.

Please, replace all of these duplicate fields.

We do need to be compatible with what the kernel expects, otherwise there is no point in filling out this information.

Regards,
Simon

#10 lore raw

Re: [PATCH 03/35] x86: Show some EFI info with the bdinfo command

Quoted 37 lines · Simon Glass

Hi Heinrich,

Quoted 21 lines · Heinrich Schuchardt
Quoted 17 lines · Simon Glass

It is useful to see some basic EFI info with the command as it forms part of the information about a board.

Add a hook for this and show the table address as a start.

While here, fix an invalid cast in setup_efi_info().

---

   arch/x86/cpu/efi/payload.c | 13 +++++++++++--
   arch/x86/include/asm/efi.h |  7 +++++++
   arch/x86/lib/Makefile      |  1 +
   arch/x86/lib/bdinfo.c      | 22 ++++++++++++++++++++++
   4 files changed, 41 insertions(+), 2 deletions(-)

create mode 100644 arch/x86/lib/bdinfo.c

Quoted 12 lines

arch/x86/cpu/efi/payload.c void setup_efi_info(struct efi_info *efi_info) · line 283

Quoted 10 lines
arch/x86/cpu/efi/payload.c +1 -1
index 9a73b768e9b..3a9f7d72868 100644
@@ -280,15 +280,24 @@ void setup_efi_info(struct efi_info *efi_info)
       }
       efi_info->efi_memdesc_size = map->desc_size;
       efi_info->efi_memdesc_version = map->version;
-     efi_info->efi_memmap = (u32)(map->desc);
+     efi_info->efi_memmap = (ulong)(map->desc);

When converting a pointer to integer we use (uintptr_t).

Quoted 1 line

Generally in U-Boot it is ulong so I try to be consistent here.

Currently ulong and uintprt_t are the same technically. But for the sake of readability uintptr_t is much clearer.

We use uintptr_t in 935 code locations already. Just grep for it.

Quoted 14 lines

Anyway, u32 is clearly not right for a 64-bit build as it gives warnings.

Quoted 11 lines

arch/x86/include/asm/bootparam.h defines efi_info->efi_memmap as u32. This type is too small to hold a pointer.

Quoted 6 lines
arch/x86/cpu/efi/payload.c +1 -1
       efi_info->efi_memmap_size = size - sizeof(struct efi_entry_memmap);
 
   #ifdef CONFIG_EFI_STUB_64BIT
       efi_info->efi_systab_hi = table->sys_table >> 32;
-     efi_info->efi_memmap_hi = (u64)(u32)(map->desc) >> 32;
+     efi_info->efi_memmap_hi = (u64)(ulong)map->desc >> 32;
Quoted 6 lines
Quoted 3 lines

You should not use two fields to hold a single pointer.

Please, replace all of these duplicate fields.

We do need to be compatible with what the kernel expects, otherwise there is no point in filling out this information.

Why should we have separate fields for the low and high 32 bits of a pointer?

Best regards

Heinrich

#11 lore raw

Re: [PATCH 03/35] x86: Show some EFI info with the bdinfo command

Hi Heinrich,

Quoted 41 lines · Heinrich Schuchardt
Quoted 37 lines · Simon Glass

Hi Heinrich,

Quoted 33 lines · Heinrich Schuchardt
Quoted 22 lines · Simon Glass

It is useful to see some basic EFI info with the command as it forms part of the information about a board.

Add a hook for this and show the table address as a start.

While here, fix an invalid cast in setup_efi_info().

---

   arch/x86/cpu/efi/payload.c | 13 +++++++++++--
   arch/x86/include/asm/efi.h |  7 +++++++
   arch/x86/lib/Makefile      |  1 +
   arch/x86/lib/bdinfo.c      | 22 ++++++++++++++++++++++
   4 files changed, 41 insertions(+), 2 deletions(-)

create mode 100644 arch/x86/lib/bdinfo.c

arch/x86/cpu/efi/payload.c +0 -0
index 9a73b768e9b..3a9f7d72868 100644

arch/x86/cpu/efi/payload.c void setup_efi_info(struct efi_info *efi_info) · line 283

Quoted 5 lines
arch/x86/cpu/efi/payload.c +1 -1
@@ -280,15 +280,24 @@ void setup_efi_info(struct efi_info *efi_info)
       }
       efi_info->efi_memdesc_size = map->desc_size;
       efi_info->efi_memdesc_version = map->version;
-     efi_info->efi_memmap = (u32)(map->desc);
+     efi_info->efi_memmap = (ulong)(map->desc);

When converting a pointer to integer we use (uintptr_t).

Quoted 6 lines
Quoted 1 line

Generally in U-Boot it is ulong so I try to be consistent here.

Currently ulong and uintprt_t are the same technically. But for the sake of readability uintptr_t is much clearer.

We use uintptr_t in 935 code locations already. Just grep for it.

$ git grep uintptr_t  |wc -l
935
$ git grep ulong  |wc -l
6006
Quoted 17 lines
Quoted 17 lines

Anyway, u32 is clearly not right for a 64-bit build as it gives warnings.

Quoted 13 lines

arch/x86/include/asm/bootparam.h defines efi_info->efi_memmap as u32. This type is too small to hold a pointer.

arch/x86/cpu/efi/payload.c void setup_efi_info(struct efi_info *efi_info) · line 288

Quoted 6 lines
arch/x86/cpu/efi/payload.c +1 -1
       efi_info->efi_memmap_size = size - sizeof(struct efi_entry_memmap);
 
   #ifdef CONFIG_EFI_STUB_64BIT
       efi_info->efi_systab_hi = table->sys_table >> 32;
-     efi_info->efi_memmap_hi = (u64)(u32)(map->desc) >> 32;
+     efi_info->efi_memmap_hi = (u64)(ulong)map->desc >> 32;

You should not use two fields to hold a single pointer.

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

Please, replace all of these duplicate fields.

We do need to be compatible with what the kernel expects, otherwise there is no point in filling out this information.

Why should we have separate fields for the low and high 32 bits of a pointer?

This patch does not change that...you need to take it up with the Linux project, which created this requirement. We should not rely on endianness being the same between a u64 and two u32 values, if that is what you are asking.

But in any case, why do you insist on critiquing code that is already there? By all means send a patch if you want to change it.

But please just review what this patch does.

Regards,
Simon

#12 lore raw

PATCH 4/35 x86: Tidy up global_data pointer for 64-bit

In reply to Simon Glass

Add an extern declaration so that it is possible to use this macro in files other than the one that defines it.

---

 arch/x86/cpu/x86_64/cpu.c          | 3 +++
 arch/x86/include/asm/global_data.h | 2 ++
 2 files changed, 5 insertions(+)
arch/x86/cpu/x86_64/cpu.c +3 -0
index 90a766c3c57..e090b1b478a 100644
@@ -8,6 +8,9 @@
 #include <cpu_func.h>
 #include <debug_uart.h>
 #include <init.h>
+#include <asm/global_data.h>
+
+DECLARE_GLOBAL_DATA_PTR;
 
 /*
  * Global declaration of gd.
arch/x86/include/asm/global_data.h +2 -0
index 3e4044593c8..f95fb5a1931 100644
@@ -133,6 +133,8 @@ struct arch_global_data {
 #ifndef __ASSEMBLY__
 # if defined(CONFIG_EFI_APP) || CONFIG_IS_ENABLED(X86_64)
 
+extern struct global_data *global_data_ptr;
+
 /* TODO([email protected]): Consider using a fixed register for gd on x86_64 */
 #define gd global_data_ptr
 
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
Reviewed-by on v1 4/35
#13 lore raw

Re: [PATCH 04/35] x86: Tidy up global_data pointer for 64-bit

Quoted 4 lines · Simon Glass

Add an extern declaration so that it is possible to use this macro in files other than the one that defines it.

Quoted 34 lines

---

  arch/x86/cpu/x86_64/cpu.c          | 3 +++
  arch/x86/include/asm/global_data.h | 2 ++
  2 files changed, 5 insertions(+)
arch/x86/cpu/x86_64/cpu.c +3 -0
index 90a766c3c57..e090b1b478a 100644
@@ -8,6 +8,9 @@
  #include <cpu_func.h>
  #include <debug_uart.h>
  #include <init.h>
+#include <asm/global_data.h>
+
+DECLARE_GLOBAL_DATA_PTR;
 
  /*
   * Global declaration of gd.
arch/x86/include/asm/global_data.h +2 -0
index 3e4044593c8..f95fb5a1931 100644
@@ -133,6 +133,8 @@ struct arch_global_data {
  #ifndef __ASSEMBLY__
  # if defined(CONFIG_EFI_APP) || CONFIG_IS_ENABLED(X86_64)
 
+extern struct global_data *global_data_ptr;
+
  /* TODO([email protected]): Consider using a fixed register for gd on x86_64 */
  #define gd global_data_ptr
 
#14 lore raw

PATCH 5/35 efi: Add a script for building and testing U-Boot on UEFI

In reply to Simon Glass

It is quite complicating to run U-Boot on qemu since we have four different builds and they must use different versions of qemu and the UEFI binaries.

Add a script to help.

---

 doc/develop/uefi/u-boot_on_efi.rst |   4 ++
 scripts/build-efi.sh               | 111 +++++++++++++++++++++++++++++
 2 files changed, 115 insertions(+)
 create mode 100755 scripts/build-efi.sh
doc/develop/uefi/u-boot_on_efi.rst +4 -0
index c9a41bc919f..59ee3885295 100644
@@ -96,6 +96,10 @@ that EFI does not support booting a 64-bit application from a 32-bit
 EFI (or vice versa). Also it will often fail to print an error message if
 you get this wrong.
 
+You may find the script `scripts/build-efi.sh` helpful for building and testing
+U-Boot on UEFI on QEMU. It also includes links to UEFI binaries dating from
+2021.
+
 
 Inner workings
 --------------
scripts/build-efi.sh new file +111 -0

111 lines not on this page · raw message

Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#15 lore raw

PATCH 6/35 x86: Create a 32/64-bit selection for the app

In reply to Simon Glass

Most EFI implementations use 64-bit. In order to spupose a 64-bit app, update the Kconfig to add an option for 32/64 bit. Update the prompt for the existing option so it is clear it relates to the stub. Move both up to just under the choice that controls them, since this looks better and the menu.

Use CONFIG_EFI_APP in the Makefile instead of CONFIG_TARGET_EFI_APP, since the latter is specific to a single target and we will have two.

---

 arch/x86/cpu/intel_common/Makefile |  2 +-
 lib/efi/Kconfig                    | 34 +++++++++++++++++++++++-------
 2 files changed, 27 insertions(+), 9 deletions(-)
arch/x86/cpu/intel_common/Makefile +1 -1
index 8b9a810f66d..1dc17b45879 100644
@@ -27,7 +27,7 @@ obj-y += fast_spi.o
 obj-y += lpc.o
 obj-y += lpss.o
 obj-$(CONFIG_$(SPL_)INTEL_GENERIC_WIFI) += generic_wifi.o
-ifndef CONFIG_TARGET_EFI_APP
+ifndef CONFIG_EFI_APP
 obj-$(CONFIG_$(SPL_TPL_)X86_32BIT_INIT) += microcode.o
 ifndef CONFIG_$(SPL_)X86_64
 obj-y += microcode.o
lib/efi/Kconfig +26 -8
index 93b85644920..e818cef0d15 100644
@@ -26,18 +26,26 @@ config EFI_STUB
 
 endchoice
 
-config EFI_RAM_SIZE
-	hex "Amount of EFI RAM for U-Boot"
+choice
+	prompt "EFI app 32/64-bit selection"
 	depends on EFI_APP
-	default 0x2000000
 	help
-	  Set the amount of EFI RAM which is claimed by U-Boot for its own
-	  use. U-Boot allocates this from EFI on start-up (along with a few
-	  other smaller amounts) and it can never be increased after that.
-	  It is used as the RAM size in with U-Boot.
+	  EFI does not support mixing 32-bit and 64-bit modes. This is a
+	  significant problem because it means that you must build a stub with
+	  the correct type for EFI to load it correctly. If you are using
+	  32-bit EFI, select 32-bit here, else select 64-bit. Failure to do
+	  this may produce no error message - it just won't start!
+
+config EFI_APP_32BIT
+	bool "Produce an app for running with 32-bit EFI"
+
+config EFI_APP_64BIT
+	bool "Produce an app for running with 64-bit EFI"
+
+endchoice
 
 choice
-	prompt "EFI 32/64-bit selection"
+	prompt "EFI stub 32/64-bit selection"
 	depends on EFI_STUB
 	help
 	  EFI does not support mixing 32-bit and 64-bit modes. This is a
@@ -53,3 +61,13 @@ config EFI_STUB_64BIT
 	bool "Produce a stub for running with 64-bit EFI"
 
 endchoice
+
+config EFI_RAM_SIZE
+	hex "Amount of EFI RAM for U-Boot"
+	depends on EFI_APP
+	default 0x2000000
+	help
+	  Set the amount of EFI RAM which is claimed by U-Boot for its own
+	  use. U-Boot allocates this from EFI on start-up (along with a few
+	  other smaller amounts) and it can never be increased after that.
+	  It is used as the RAM size in with U-Boot.
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#16 lore raw

Re: [PATCH 06/35] x86: Create a 32/64-bit selection for the app

Quoted 1 line · Simon Glass

Most EFI implementations use 64-bit. In order to spupose a 64-bit app,

%s/spupose/support

Quoted 1 line

update the Kconfig to add an option for 32/64 bit. Update the prompt for

This commit message is unclear. We already can compile U-Boot as an 64bit EFI application. "In order to support" would imply that we don't.

Quoted 67 lines

the existing option so it is clear it relates to the stub. Move both up to just under the choice that controls them, since this looks better and the menu.

Use CONFIG_EFI_APP in the Makefile instead of CONFIG_TARGET_EFI_APP, since the latter is specific to a single target and we will have two.

---

  arch/x86/cpu/intel_common/Makefile |  2 +-
  lib/efi/Kconfig                    | 34 +++++++++++++++++++++++-------
  2 files changed, 27 insertions(+), 9 deletions(-)
arch/x86/cpu/intel_common/Makefile +1 -1
index 8b9a810f66d..1dc17b45879 100644
@@ -27,7 +27,7 @@ obj-y += fast_spi.o
  obj-y += lpc.o
  obj-y += lpss.o
  obj-$(CONFIG_$(SPL_)INTEL_GENERIC_WIFI) += generic_wifi.o
-ifndef CONFIG_TARGET_EFI_APP
+ifndef CONFIG_EFI_APP
  obj-$(CONFIG_$(SPL_TPL_)X86_32BIT_INIT) += microcode.o
  ifndef CONFIG_$(SPL_)X86_64
  obj-y += microcode.o
lib/efi/Kconfig +16 -8
index 93b85644920..e818cef0d15 100644
@@ -26,18 +26,26 @@ config EFI_STUB
 
  endchoice
 
-config EFI_RAM_SIZE
-	hex "Amount of EFI RAM for U-Boot"
+choice
+	prompt "EFI app 32/64-bit selection"
  	depends on EFI_APP
-	default 0x2000000
  	help
-	  Set the amount of EFI RAM which is claimed by U-Boot for its own
-	  use. U-Boot allocates this from EFI on start-up (along with a few
-	  other smaller amounts) and it can never be increased after that.
-	  It is used as the RAM size in with U-Boot.
+	  EFI does not support mixing 32-bit and 64-bit modes. This is a
+	  significant problem because it means that you must build a stub with
+	  the correct type for EFI to load it correctly. If you are using
+	  32-bit EFI, select 32-bit here, else select 64-bit. Failure to do
+	  this may produce no error message - it just won't start!
+
+config EFI_APP_32BIT
+	bool "Produce an app for running with 32-bit EFI"
+
+config EFI_APP_64BIT
+	bool "Produce an app for running with 64-bit EFI"
+
+endchoice
 
  choice
-	prompt "EFI 32/64-bit selection"
+	prompt "EFI stub 32/64-bit selection"
  	depends on EFI_STUB
  	help
  	  EFI does not support mixing 32-bit and 64-bit modes. This is a

lib/efi/Kconfig config EFI_STUB_64BIT · line 68

Quoted 8 lines
lib/efi/Kconfig +5 -0
@@ -53,3 +61,13 @@ config EFI_STUB_64BIT
  	bool "Produce a stub for running with 64-bit EFI"
 
  endchoice
+
+config EFI_RAM_SIZE
+	hex "Amount of EFI RAM for U-Boot"
+	depends on EFI_APP
+	default 0x2000000

32 MiB is quite small for loading a kernel and an initrd.

Best regards

Heinrich

Quoted 6 lines
lib/efi/Kconfig +5 -0
+	help
+	  Set the amount of EFI RAM which is claimed by U-Boot for its own
+	  use. U-Boot allocates this from EFI on start-up (along with a few
+	  other smaller amounts) and it can never be increased after that.
+	  It is used as the RAM size in with U-Boot.
#17 lore raw

Re: [PATCH 06/35] x86: Create a 32/64-bit selection for the app

Hi Heinrich,

Quoted 4 lines · Heinrich Schuchardt

On 9/8/21 3:33 PM, Simon Glass wrote:

Quoted 8 lines
Quoted 1 line

Most EFI implementations use 64-bit. In order to spupose a 64-bit app,

%s/spupose/support

Quoted 1 line

update the Kconfig to add an option for 32/64 bit. Update the prompt for

This commit message is unclear. We already can compile U-Boot as an 64bit EFI application. "In order to support" would imply that we don't.

Which board are you talking about? There is only the 64-bit payload at present, certainly no 64-bit app?

Quoted 73 lines
Quoted 73 lines

the existing option so it is clear it relates to the stub. Move both up to just under the choice that controls them, since this looks better and the menu.

Use CONFIG_EFI_APP in the Makefile instead of CONFIG_TARGET_EFI_APP, since the latter is specific to a single target and we will have two.

---

  arch/x86/cpu/intel_common/Makefile |  2 +-
  lib/efi/Kconfig                    | 34 +++++++++++++++++++++++-------
  2 files changed, 27 insertions(+), 9 deletions(-)
arch/x86/cpu/intel_common/Makefile +1 -1
index 8b9a810f66d..1dc17b45879 100644
@@ -27,7 +27,7 @@ obj-y += fast_spi.o
  obj-y += lpc.o
  obj-y += lpss.o
  obj-$(CONFIG_$(SPL_)INTEL_GENERIC_WIFI) += generic_wifi.o
-ifndef CONFIG_TARGET_EFI_APP
+ifndef CONFIG_EFI_APP
  obj-$(CONFIG_$(SPL_TPL_)X86_32BIT_INIT) += microcode.o
  ifndef CONFIG_$(SPL_)X86_64
  obj-y += microcode.o
lib/efi/Kconfig +16 -8
index 93b85644920..e818cef0d15 100644
@@ -26,18 +26,26 @@ config EFI_STUB
 
  endchoice
 
-config EFI_RAM_SIZE
-     hex "Amount of EFI RAM for U-Boot"
+choice
+     prompt "EFI app 32/64-bit selection"
      depends on EFI_APP
-     default 0x2000000
      help
-       Set the amount of EFI RAM which is claimed by U-Boot for its own
-       use. U-Boot allocates this from EFI on start-up (along with a few
-       other smaller amounts) and it can never be increased after that.
-       It is used as the RAM size in with U-Boot.
+       EFI does not support mixing 32-bit and 64-bit modes. This is a
+       significant problem because it means that you must build a stub with
+       the correct type for EFI to load it correctly. If you are using
+       32-bit EFI, select 32-bit here, else select 64-bit. Failure to do
+       this may produce no error message - it just won't start!
+
+config EFI_APP_32BIT
+     bool "Produce an app for running with 32-bit EFI"
+
+config EFI_APP_64BIT
+     bool "Produce an app for running with 64-bit EFI"
+
+endchoice
 
  choice
-     prompt "EFI 32/64-bit selection"
+     prompt "EFI stub 32/64-bit selection"
      depends on EFI_STUB
      help
        EFI does not support mixing 32-bit and 64-bit modes. This is a
@@ -53,3 +61,13 @@ config EFI_STUB_64BIT
      bool "Produce a stub for running with 64-bit EFI"
 
Quoted 8 lines

lib/efi/Kconfig config EFI_STUB_64BIT · line 68

Quoted 6 lines
lib/efi/Kconfig +5 -0
  endchoice
+
+config EFI_RAM_SIZE
+     hex "Amount of EFI RAM for U-Boot"
+     depends on EFI_APP
+     default 0x2000000

32 MiB is quite small for loading a kernel and an initrd.

It does run nicely in qemu though without needing a special -m value. What do you suggest?

Regards,
Simon

#18 lore raw

PATCH 7/35 efi: Create a 64-bit app

In reply to Simon Glass

Most modern platforms use 64-bit EFI so it is useful to have a U-Boot app that runs under that. Add a (non-functional) build for this.

Note that --whole-archive causes the gcc 9.2 linker to crash, so disable this for now. Once this is resolved, things should work.

For now, avoid meantioning the documentation for the 64-bit app, since it does not work.

---

 arch/x86/cpu/u-boot-64.lds                    |  2 +
 arch/x86/cpu/x86_64/Makefile                  |  4 ++
 arch/x86/cpu/x86_64/cpu.c                     | 17 --------
 arch/x86/cpu/x86_64/misc.c                    | 25 ++++++++++++
 board/efi/Kconfig                             | 15 ++++++-
 board/efi/efi-x86_app/Kconfig                 |  2 +-
 board/efi/efi-x86_app/MAINTAINERS             | 11 +++++-
 ..._app_defconfig => efi-x86_app32_defconfig} |  2 +-
 configs/efi-x86_app64_defconfig               | 39 +++++++++++++++++++
 doc/develop/uefi/u-boot_on_efi.rst            |  4 +-
 10 files changed, 96 insertions(+), 25 deletions(-)
 create mode 100644 arch/x86/cpu/x86_64/misc.c
 rename configs/{efi-x86_app_defconfig => efi-x86_app32_defconfig} (97%)
 create mode 100644 configs/efi-x86_app64_defconfig
arch/x86/cpu/u-boot-64.lds +2 -0
index ee0812aefbc..92a30c2a387 100644
@@ -15,7 +15,9 @@ SECTIONS
 	/DISCARD/ : { *(.u_boot_list_2_cmd_*) }
 #endif
 
+#ifdef CONFIG_SYS_TEXT_BASE
 	. = CONFIG_SYS_TEXT_BASE;	/* Location of bootcode in flash */
+#endif
 	__text_start = .;
 
 	.text.start : { *(.text.start); }
arch/x86/cpu/x86_64/Makefile +4 -0
index 400f0ffe397..e929563b2c1 100644
@@ -4,3 +4,7 @@
 #
 
 obj-y += cpu.o interrupts.o setjmp.o
+
+ifndef CONFIG_EFI
+obj-y += misc.o
+endif
arch/x86/cpu/x86_64/cpu.c +0 -17
index e090b1b478a..fb501218e25 100644
@@ -52,23 +52,6 @@ int x86_mp_init(void)
 	return 0;
 }
 
-int misc_init_r(void)
-{
-	return 0;
-}
-
-#ifndef CONFIG_SYS_COREBOOT
-int checkcpu(void)
-{
-	return 0;
-}
-
-int print_cpuinfo(void)
-{
-	return 0;
-}
-#endif
-
 int x86_cpu_reinit_f(void)
 {
 	return 0;
arch/x86/cpu/x86_64/misc.c new file +25 -0
new file mode 100644
index 00000000000..02587ff0c50
@@ -0,0 +1,25 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * (C) Copyright 2016 Google, Inc
+ * Written by Simon Glass <[email protected]>
+ */
+
+#include <common.h>
+#include <init.h>
+
+int misc_init_r(void)
+{
+	return 0;
+}
+
+#ifndef CONFIG_SYS_COREBOOT
+int checkcpu(void)
+{
+	return 0;
+}
+
+int print_cpuinfo(void)
+{
+	return 0;
+}
+#endif
board/efi/Kconfig +13 -2
index 291bd2ca154..3df6e31c8ba 100644
@@ -4,14 +4,25 @@ choice
 	prompt "Mainboard model"
 	optional
 
-config TARGET_EFI_APP
-	bool "efi application"
+config TARGET_EFI_APP32
+	bool "32-bit efi application"
+	select EFI_APP
 	help
 	  This target is used for running U-Boot on top of EFI. In
 	  this case EFI does the early initialisation, and U-Boot
 	  takes over once the RAM, video and CPU are fully running.
 	  U-Boot is loaded as an application from EFI.
 
+config TARGET_EFI_APP64
+	bool "64-bit efi application"
+	select EFI_APP
+	select X86_64
+	help
+	  This target is used for running U-Boot on top of EFI in 64-bit mode.
+	  In this case EFI does the early initialisation, and U-Boot
+	  takes over once the RAM, video and CPU are fully running.
+	  U-Boot is loaded as an application from EFI.
+
 config TARGET_EFI_PAYLOAD
 	bool "efi payload"
 	help
board/efi/efi-x86_app/Kconfig +1 -1
index ae87bf34d37..e412702eed7 100644
@@ -1,4 +1,4 @@
-if TARGET_EFI_APP
+if EFI_APP
 
 config SYS_BOARD
 	default "efi-x86_app"
board/efi/efi-x86_app/MAINTAINERS +9 -2
index fb8a6b1c2fa..b292811a8f0 100644
@@ -1,6 +1,13 @@
-EFI-X86_APP BOARD
+EFI-X86_APP32 BOARD
 M:	Simon Glass <[email protected]>
 S:	Maintained
 F:	board/efi/efi-x86_app/
 F:	include/configs/efi-x86_app.h
-F:	configs/efi-x86_app_defconfig
+F:	configs/efi-x86_app32_defconfig
+
+EFI-X86_APP64 BOARD
+M:	Simon Glass <[email protected]>
+S:	Maintained
+F:	board/efi/efi-x86_app/
+F:	include/configs/efi-x86_app.h
+F:	configs/efi-x86_app64_defconfig
configs/efi-x86_app32_defconfig renamed from configs/efi-x86_app_defconfig +1 -1
similarity index 97%
rename from configs/efi-x86_app_defconfig
rename to configs/efi-x86_app32_defconfig
index e9ee66250cf..43ab3fb99c7 100644
@@ -5,7 +5,7 @@ CONFIG_DEFAULT_DEVICE_TREE="efi-x86_app"
 CONFIG_DEBUG_UART_BASE=0
 CONFIG_DEBUG_UART_CLOCK=0
 CONFIG_VENDOR_EFI=y
-CONFIG_TARGET_EFI_APP=y
+CONFIG_TARGET_EFI_APP32=y
 CONFIG_DEBUG_UART=y
 CONFIG_FIT=y
 CONFIG_SHOW_BOOT_PROGRESS=y
configs/efi-x86_app64_defconfig new file +39 -0
new file mode 100644
index 00000000000..a395d3eb27b
@@ -0,0 +1,39 @@
+CONFIG_X86=y
+CONFIG_NR_DRAM_BANKS=8
+CONFIG_ENV_SIZE=0x1000
+CONFIG_DEFAULT_DEVICE_TREE="efi-x86_app"
+CONFIG_DEBUG_UART_BASE=0
+CONFIG_DEBUG_UART_CLOCK=0
+CONFIG_VENDOR_EFI=y
+CONFIG_TARGET_EFI_APP64=y
+CONFIG_DEBUG_UART=y
+CONFIG_FIT=y
+CONFIG_SHOW_BOOT_PROGRESS=y
+CONFIG_USE_BOOTARGS=y
+CONFIG_BOOTARGS="root=/dev/sdb3 init=/sbin/init rootwait ro"
+CONFIG_SYS_CONSOLE_INFO_QUIET=y
+CONFIG_DISPLAY_BOARDINFO_LATE=y
+CONFIG_LAST_STAGE_INIT=y
+CONFIG_HUSH_PARSER=y
+# CONFIG_CMD_BOOTM is not set
+CONFIG_CMD_PART=y
+# CONFIG_CMD_NET is not set
+CONFIG_CMD_TIME=y
+CONFIG_CMD_EXT2=y
+CONFIG_CMD_EXT4=y
+CONFIG_CMD_EXT4_WRITE=y
+CONFIG_CMD_FAT=y
+CONFIG_CMD_FS_GENERIC=y
+CONFIG_MAC_PARTITION=y
+CONFIG_ISO_PARTITION=y
+CONFIG_EFI_PARTITION=y
+CONFIG_OF_EMBED=y
+CONFIG_ENV_OVERWRITE=y
+CONFIG_SYS_RELOC_GD_ENV_ADDR=y
+CONFIG_REGMAP=y
+CONFIG_SYSCON=y
+# CONFIG_DM_ETH is not set
+# CONFIG_REGEX is not set
+# CONFIG_GZIP is not set
+CONFIG_EFI=y
+# CONFIG_EFI_LOADER is not set
doc/develop/uefi/u-boot_on_efi.rst +2 -2
index 59ee3885295..5de7215f6d1 100644
@@ -48,10 +48,10 @@ for that board. It will be either 32-bit or 64-bit. Alternatively, you can
 opt for using QEMU [1] and the OVMF [2], as detailed below.
 
 To build U-Boot as an EFI application (32-bit EFI required), enable CONFIG_EFI
-and CONFIG_EFI_APP. The efi-x86_app config (efi-x86_app_defconfig) is set up
+and CONFIG_EFI_APP. The efi-x86_app config (efi-x86_app32_defconfig) is set up
 for this. Just build U-Boot as normal, e.g.::
 
-   make efi-x86_app_defconfig
+   make efi-x86_app32_defconfig
    make
 
 To build U-Boot as an EFI payload (32-bit or 64-bit EFI can be used), enable
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#19 lore raw

Re: [PATCH 07/35] efi: Create a 64-bit app

Quoted 7 lines · Simon Glass

Most modern platforms use 64-bit EFI so it is useful to have a U-Boot app that runs under that. Add a (non-functional) build for this.

Note that --whole-archive causes the gcc 9.2 linker to crash, so disable this for now. Once this is resolved, things should work.

For now, avoid meantioning the documentation for the 64-bit app, since it

%s/meantioning/mentioning/

Quoted 246 lines

does not work.

---

  arch/x86/cpu/u-boot-64.lds                    |  2 +
  arch/x86/cpu/x86_64/Makefile                  |  4 ++
  arch/x86/cpu/x86_64/cpu.c                     | 17 --------
  arch/x86/cpu/x86_64/misc.c                    | 25 ++++++++++++
  board/efi/Kconfig                             | 15 ++++++-
  board/efi/efi-x86_app/Kconfig                 |  2 +-
  board/efi/efi-x86_app/MAINTAINERS             | 11 +++++-
  ..._app_defconfig => efi-x86_app32_defconfig} |  2 +-
  configs/efi-x86_app64_defconfig               | 39 +++++++++++++++++++
  doc/develop/uefi/u-boot_on_efi.rst            |  4 +-
  10 files changed, 96 insertions(+), 25 deletions(-)

create mode 100644 arch/x86/cpu/x86_64/misc.c
rename configs/{efi-x86_app_defconfig => efi-x86_app32_defconfig} (97%) create mode 100644 configs/efi-x86_app64_defconfig

arch/x86/cpu/u-boot-64.lds +2 -0
index ee0812aefbc..92a30c2a387 100644
@@ -15,7 +15,9 @@ SECTIONS
  	/DISCARD/ : { *(.u_boot_list_2_cmd_*) }
  #endif
 
+#ifdef CONFIG_SYS_TEXT_BASE
  	. = CONFIG_SYS_TEXT_BASE;	/* Location of bootcode in flash */
+#endif
  	__text_start = .;
 
  	.text.start : { *(.text.start); }
arch/x86/cpu/x86_64/Makefile +4 -0
index 400f0ffe397..e929563b2c1 100644
@@ -4,3 +4,7 @@
  #
 
  obj-y += cpu.o interrupts.o setjmp.o
+
+ifndef CONFIG_EFI
+obj-y += misc.o
+endif
arch/x86/cpu/x86_64/cpu.c +0 -17
index e090b1b478a..fb501218e25 100644
@@ -52,23 +52,6 @@ int x86_mp_init(void)
  	return 0;
  }
 
-int misc_init_r(void)
-{
-	return 0;
-}
-
-#ifndef CONFIG_SYS_COREBOOT
-int checkcpu(void)
-{
-	return 0;
-}
-
-int print_cpuinfo(void)
-{
-	return 0;
-}
-#endif
-
  int x86_cpu_reinit_f(void)
  {
  	return 0;
arch/x86/cpu/x86_64/misc.c new file +25 -0
new file mode 100644
index 00000000000..02587ff0c50
@@ -0,0 +1,25 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * (C) Copyright 2016 Google, Inc
+ * Written by Simon Glass <[email protected]>
+ */
+
+#include <common.h>
+#include <init.h>
+
+int misc_init_r(void)
+{
+	return 0;
+}
+
+#ifndef CONFIG_SYS_COREBOOT
+int checkcpu(void)
+{
+	return 0;
+}
+
+int print_cpuinfo(void)
+{
+	return 0;
+}
+#endif
board/efi/Kconfig +13 -2
index 291bd2ca154..3df6e31c8ba 100644
@@ -4,14 +4,25 @@ choice
  	prompt "Mainboard model"
  	optional
 
-config TARGET_EFI_APP
-	bool "efi application"
+config TARGET_EFI_APP32
+	bool "32-bit efi application"
+	select EFI_APP
  	help
  	  This target is used for running U-Boot on top of EFI. In
  	  this case EFI does the early initialisation, and U-Boot
  	  takes over once the RAM, video and CPU are fully running.
  	  U-Boot is loaded as an application from EFI.
 
+config TARGET_EFI_APP64
+	bool "64-bit efi application"
+	select EFI_APP
+	select X86_64
+	help
+	  This target is used for running U-Boot on top of EFI in 64-bit mode.
+	  In this case EFI does the early initialisation, and U-Boot
+	  takes over once the RAM, video and CPU are fully running.
+	  U-Boot is loaded as an application from EFI.
+
  config TARGET_EFI_PAYLOAD
  	bool "efi payload"
  	help
board/efi/efi-x86_app/Kconfig +1 -1
index ae87bf34d37..e412702eed7 100644
@@ -1,4 +1,4 @@
-if TARGET_EFI_APP
+if EFI_APP
 
  config SYS_BOARD
  	default "efi-x86_app"
board/efi/efi-x86_app/MAINTAINERS +9 -2
index fb8a6b1c2fa..b292811a8f0 100644
@@ -1,6 +1,13 @@
-EFI-X86_APP BOARD
+EFI-X86_APP32 BOARD
  M:	Simon Glass <[email protected]>
  S:	Maintained
  F:	board/efi/efi-x86_app/
  F:	include/configs/efi-x86_app.h
-F:	configs/efi-x86_app_defconfig
+F:	configs/efi-x86_app32_defconfig
+
+EFI-X86_APP64 BOARD
+M:	Simon Glass <[email protected]>
+S:	Maintained
+F:	board/efi/efi-x86_app/
+F:	include/configs/efi-x86_app.h
+F:	configs/efi-x86_app64_defconfig
configs/efi-x86_app32_defconfig renamed from configs/efi-x86_app_defconfig +1 -1
similarity index 97%
rename from configs/efi-x86_app_defconfig
rename to configs/efi-x86_app32_defconfig
index e9ee66250cf..43ab3fb99c7 100644
@@ -5,7 +5,7 @@ CONFIG_DEFAULT_DEVICE_TREE="efi-x86_app"
  CONFIG_DEBUG_UART_BASE=0
  CONFIG_DEBUG_UART_CLOCK=0
  CONFIG_VENDOR_EFI=y
-CONFIG_TARGET_EFI_APP=y
+CONFIG_TARGET_EFI_APP32=y
  CONFIG_DEBUG_UART=y
  CONFIG_FIT=y
  CONFIG_SHOW_BOOT_PROGRESS=y
configs/efi-x86_app64_defconfig new file +39 -0
new file mode 100644
index 00000000000..a395d3eb27b
@@ -0,0 +1,39 @@
+CONFIG_X86=y
+CONFIG_NR_DRAM_BANKS=8
+CONFIG_ENV_SIZE=0x1000
+CONFIG_DEFAULT_DEVICE_TREE="efi-x86_app"
+CONFIG_DEBUG_UART_BASE=0
+CONFIG_DEBUG_UART_CLOCK=0
+CONFIG_VENDOR_EFI=y
+CONFIG_TARGET_EFI_APP64=y
+CONFIG_DEBUG_UART=y
+CONFIG_FIT=y
+CONFIG_SHOW_BOOT_PROGRESS=y
+CONFIG_USE_BOOTARGS=y
+CONFIG_BOOTARGS="root=/dev/sdb3 init=/sbin/init rootwait ro"
+CONFIG_SYS_CONSOLE_INFO_QUIET=y
+CONFIG_DISPLAY_BOARDINFO_LATE=y
+CONFIG_LAST_STAGE_INIT=y
+CONFIG_HUSH_PARSER=y
+# CONFIG_CMD_BOOTM is not set
+CONFIG_CMD_PART=y
+# CONFIG_CMD_NET is not set
+CONFIG_CMD_TIME=y
+CONFIG_CMD_EXT2=y
+CONFIG_CMD_EXT4=y
+CONFIG_CMD_EXT4_WRITE=y
+CONFIG_CMD_FAT=y
+CONFIG_CMD_FS_GENERIC=y
+CONFIG_MAC_PARTITION=y
+CONFIG_ISO_PARTITION=y
+CONFIG_EFI_PARTITION=y
+CONFIG_OF_EMBED=y
+CONFIG_ENV_OVERWRITE=y
+CONFIG_SYS_RELOC_GD_ENV_ADDR=y
+CONFIG_REGMAP=y
+CONFIG_SYSCON=y
+# CONFIG_DM_ETH is not set
+# CONFIG_REGEX is not set
+# CONFIG_GZIP is not set
+CONFIG_EFI=y
+# CONFIG_EFI_LOADER is not set
doc/develop/uefi/u-boot_on_efi.rst +2 -2
index 59ee3885295..5de7215f6d1 100644
@@ -48,10 +48,10 @@ for that board. It will be either 32-bit or 64-bit. Alternatively, you can
  opt for using QEMU [1] and the OVMF [2], as detailed below.
 
  To build U-Boot as an EFI application (32-bit EFI required), enable CONFIG_EFI
-and CONFIG_EFI_APP. The efi-x86_app config (efi-x86_app_defconfig) is set up
+and CONFIG_EFI_APP. The efi-x86_app config (efi-x86_app32_defconfig) is set up
  for this. Just build U-Boot as normal, e.g.::
 
-   make efi-x86_app_defconfig
+   make efi-x86_app32_defconfig
     make
 
  To build U-Boot as an EFI payload (32-bit or 64-bit EFI can be used), enable
#20 lore raw

PATCH 8/35 x86: Don't duplicate global_ptr in 64-bit EFI app

In reply to Simon Glass

This variable is already defined by the EFI code. Drop the duplicate definition when building a 64-bit EFI app.

---

 arch/x86/cpu/x86_64/cpu.c  | 16 ----------------
 arch/x86/cpu/x86_64/misc.c | 16 ++++++++++++++++
 lib/efi/efi.c              |  9 +++++++++
 3 files changed, 25 insertions(+), 16 deletions(-)
arch/x86/cpu/x86_64/cpu.c +0 -16
index fb501218e25..a3674e8e29a 100644
@@ -10,22 +10,6 @@
 #include <init.h>
 #include <asm/global_data.h>
 
-DECLARE_GLOBAL_DATA_PTR;
-
-/*
- * Global declaration of gd.
- *
- * As we write to it before relocation we have to make sure it is not put into
- * a .bss section which may overlap a .rela section. Initialization forces it
- * into a .data section which cannot overlap any .rela section.
- */
-struct global_data *global_data_ptr = (struct global_data *)~0;
-
-void arch_setup_gd(gd_t *new_gd)
-{
-	global_data_ptr = new_gd;
-}
-
 int cpu_has_64bit(void)
 {
 	return true;
arch/x86/cpu/x86_64/misc.c +16 -0
index 02587ff0c50..691b67ff68a 100644
@@ -7,6 +7,22 @@
 #include <common.h>
 #include <init.h>
 
+DECLARE_GLOBAL_DATA_PTR;
+
+/*
+ * Global declaration of gd.
+ *
+ * As we write to it before relocation we have to make sure it is not put into
+ * a .bss section which may overlap a .rela section. Initialization forces it
+ * into a .data section which cannot overlap any .rela section.
+ */
+struct global_data *global_data_ptr = (struct global_data *)~0;
+
+void arch_setup_gd(gd_t *new_gd)
+{
+	global_data_ptr = new_gd;
+}
+
 int misc_init_r(void)
 {
 	return 0;
lib/efi/efi.c +9 -0
index 0c16a5fdd38..69e52e45748 100644
@@ -17,6 +17,15 @@
 #include <efi.h>
 #include <efi_api.h>
 
+/*
+ * Global declaration of gd.
+ *
+ * As we write to it before relocation we have to make sure it is not put into
+ * a .bss section which may overlap a .rela section. Initialization forces it
+ * into a .data section which cannot overlap any .rela section.
+ */
+struct global_data *global_data_ptr = (struct global_data *)~0;
+
 /*
  * Unfortunately we cannot access any code outside what is built especially
  * for the stub. lib/string.c is already being built for the U-Boot payload
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#21 lore raw

PATCH 9/35 efi: Add a way to obtain boot services in the app

In reply to Simon Glass

Add a function to return this information along with a stub for the efi_info_get() function, since calling it otherwise hangs U-Boot.

---

 include/efi.h     |  8 +++++++-
 lib/efi/efi_app.c | 10 ++++++++++
 2 files changed, 17 insertions(+), 1 deletion(-)
include/efi.h +7 -1
index 18c13e0370a..b5835422b95 100644
@@ -444,9 +444,15 @@ extern char _binary_u_boot_bin_start[], _binary_u_boot_bin_end[];
  *
  * @return pointer to EFI system table
  */
-
 struct efi_system_table *efi_get_sys_table(void);
 
+/**
+ * efi_get_boot() - Get access to the EFI boot services table
+ *
+ * @return pointer to EFI boot services table
+ */
+struct efi_boot_services *efi_get_boot(void);
+
 /**
  * efi_get_ram_base() - Find the base of RAM
  *
lib/efi/efi_app.c +10 -0
index 907bacd716a..f61665686c5 100644
@@ -31,11 +31,21 @@ struct efi_system_table *efi_get_sys_table(void)
 	return global_priv->sys_table;
 }
 
+struct efi_boot_services *efi_get_boot(void)
+{
+	return global_priv->boot;
+}
+
 unsigned long efi_get_ram_base(void)
 {
 	return global_priv->ram_base;
 }
 
+int efi_info_get(enum efi_entry_t type, void **datap, int *sizep)
+{
+	return -ENOSYS;
+}
+
 static efi_status_t setup_memory(struct efi_priv *priv)
 {
 	struct efi_boot_services *boot = priv->boot;
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#22 lore raw

PATCH 10/35 efi: Add video support to the app

In reply to Simon Glass

The current EFI video driver only works when running in the stub. In that case the stub calls boot services (before jumping to U-Boot proper) and copies the graphics info over to the efi table. This is necessary because the stub exits boot services before jumping to U-Boot.

The app maintains access to boot services throughout its life, so does not need to do this. Update the driver to support calling boot services directly.

Enable video output for the app.

A sample qemu command-line for this case is:

qemu-system-x86_64 -bios /usr/share/edk2.git/ovmf-ia32/OVMF-pure-efi.fd -drive id=disk,file=try.img,if=none,format=raw -nic none -device ahci,id=ahci -device ide-hd,drive=disk,bus=ahci.0

---

 arch/x86/dts/efi-x86_app.dts  |  4 ++++
 board/efi/efi-x86_app/Kconfig |  4 ++++
 drivers/video/Kconfig         |  2 +-
 drivers/video/efi.c           | 45 ++++++++++++++++++++++++++++-------
 include/configs/efi-x86_app.h |  6 ++---
 5 files changed, 49 insertions(+), 12 deletions(-)
arch/x86/dts/efi-x86_app.dts +4 -0
index 04e044a07a8..a5316e2a1a7 100644
@@ -25,4 +25,8 @@
 		compatible = "efi,reset";
 		u-boot,dm-pre-reloc;
 	};
+	efi-fb {
+		compatible = "efi-fb";
+	};
+
 };
board/efi/efi-x86_app/Kconfig +4 -0
index e412702eed7..ecd08d73146 100644
@@ -12,4 +12,8 @@ config SYS_SOC
 config SYS_CONFIG_NAME
 	default "efi-x86_app"
 
+config BOARD_SPECIFIC_OPTIONS # dummy
+	def_bool y
+	imply VIDEO_EFI
+
 endif
drivers/video/Kconfig +1 -1
index 8b940d70eb2..38a421690d9 100644
@@ -250,7 +250,7 @@ config VIDEO_COREBOOT
 
 config VIDEO_EFI
 	bool "Enable EFI framebuffer driver support"
-	depends on EFI_STUB
+	depends on EFI_STUB || EFI_APP
 	help
 	  Turn on this option to enable a framebuffeer driver when U-Boot is
 	  loaded as a payload (see README.u-boot_on_efi) by an EFI BIOS where
drivers/video/efi.c +37 -8

59 lines not on this page · raw message

include/configs/efi-x86_app.h +3 -3
index 33418cfbec4..6061a6db0a4 100644
@@ -10,8 +10,8 @@
 
 #undef CONFIG_TPM_TIS_BASE_ADDRESS
 
-#define CONFIG_STD_DEVICES_SETTINGS     "stdin=usbkbd,vga,serial\0" \
-					"stdout=vga,serial\0" \
-					"stderr=vga,serial\0"
+#define CONFIG_STD_DEVICES_SETTINGS	"stdin=serial\0" \
+					"stdout=vidconsole\0" \
+					"stderr=vidconsole\0"
 
 #endif
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#23 lore raw

Re: [PATCH 10/35] efi: Add video support to the app

Quoted 10 lines · Simon Glass

The current EFI video driver only works when running in the stub. In that case the stub calls boot services (before jumping to U-Boot proper) and copies the graphics info over to the efi table. This is necessary because the stub exits boot services before jumping to U-Boot.

The app maintains access to boot services throughout its life, so does not need to do this. Update the driver to support calling boot services directly.

Enable video output for the app.

Please, use the EFI_GRAPHICS_OUTPUT_PROTOCOL to implement video. Don't assume any VESA or VGA compatible card. This would not be portable to other architectures.

Best regards

Heinrich

Quoted 138 lines

A sample qemu command-line for this case is:

qemu-system-x86_64 -bios /usr/share/edk2.git/ovmf-ia32/OVMF-pure-efi.fd -drive id=disk,file=try.img,if=none,format=raw -nic none -device ahci,id=ahci -device ide-hd,drive=disk,bus=ahci.0

---

  arch/x86/dts/efi-x86_app.dts  |  4 ++++
  board/efi/efi-x86_app/Kconfig |  4 ++++
  drivers/video/Kconfig         |  2 +-
  drivers/video/efi.c           | 45 ++++++++++++++++++++++++++++-------
  include/configs/efi-x86_app.h |  6 ++---
  5 files changed, 49 insertions(+), 12 deletions(-)
arch/x86/dts/efi-x86_app.dts +4 -0
index 04e044a07a8..a5316e2a1a7 100644
@@ -25,4 +25,8 @@
  		compatible = "efi,reset";
  		u-boot,dm-pre-reloc;
  	};
+	efi-fb {
+		compatible = "efi-fb";
+	};
+
  };
board/efi/efi-x86_app/Kconfig +4 -0
index e412702eed7..ecd08d73146 100644
@@ -12,4 +12,8 @@ config SYS_SOC
  config SYS_CONFIG_NAME
  	default "efi-x86_app"
 
+config BOARD_SPECIFIC_OPTIONS # dummy
+	def_bool y
+	imply VIDEO_EFI
+
  endif
drivers/video/Kconfig +1 -1
index 8b940d70eb2..38a421690d9 100644
@@ -250,7 +250,7 @@ config VIDEO_COREBOOT
 
  config VIDEO_EFI
  	bool "Enable EFI framebuffer driver support"
-	depends on EFI_STUB
+	depends on EFI_STUB || EFI_APP
  	help
  	  Turn on this option to enable a framebuffeer driver when U-Boot is
  	  loaded as a payload (see README.u-boot_on_efi) by an EFI BIOS where
drivers/video/efi.c +37 -8
index c248bd352a7..4e13a3efcb1 100644
@@ -50,6 +50,28 @@ static void efi_find_pixel_bits(u32 mask, u8 *pos, u8 *size)
  	*size = len;
  }
 
+static int get_mode_info(struct vesa_mode_info *vesa)
+{
+	efi_guid_t efi_gop_guid = EFI_GRAPHICS_OUTPUT_PROTOCOL_GUID;
+	struct efi_boot_services *boot = efi_get_boot();
+	struct efi_gop_mode *mode;
+	struct efi_gop *gop;
+	int ret;
+
+	if (!boot)
+		return log_msg_ret("sys", -ENOSYS);
+	ret = boot->locate_protocol(&efi_gop_guid, NULL, (void **)&gop);
+	if (ret)
+		return log_msg_ret("prot", -ENOTSUPP);
+
+	mode = gop->mode;
+	vesa->phys_base_ptr = mode->fb_base;
+	vesa->x_resolution = mode->info->width;
+	vesa->y_resolution = mode->info->height;
+
+	return 0;
+}
+
  static int save_vesa_mode(struct vesa_mode_info *vesa)
  {
  	struct efi_entry_gopmode *mode;
@@ -57,16 +79,23 @@ static int save_vesa_mode(struct vesa_mode_info *vesa)
  	int size;
  	int ret;
 
-	ret = efi_info_get(EFIET_GOP_MODE, (void **)&mode, &size);
-	if (ret == -ENOENT) {
-		debug("efi graphics output protocol mode not found\n");
-		return -ENXIO;
+	if (IS_ENABLED(CONFIG_EFI_APP)) {
+		ret = get_mode_info(vesa);
+		if (ret) {
+			printf("efi graphics output protocol not found\n");
+			return -ENXIO;
+		}
+	} else {
+		ret = efi_info_get(EFIET_GOP_MODE, (void **)&mode, &size);
+		if (ret == -ENOENT) {
+			printf("efi graphics output protocol mode not found\n");
+			return -ENXIO;
+		}
+		vesa->phys_base_ptr = mode->fb_base;
+		vesa->x_resolution = mode->info->width;
+		vesa->y_resolution = mode->info->height;
  	}
 
-	vesa->phys_base_ptr = mode->fb_base;
-	vesa->x_resolution = mode->info->width;
-	vesa->y_resolution = mode->info->height;
-
  	if (mode->info->pixel_format < EFI_GOT_BITMASK) {
  		fbinfo = &efi_framebuffer_format_map[mode->info->pixel_format];
  		vesa->red_mask_size = fbinfo->red.size;
include/configs/efi-x86_app.h +3 -3
index 33418cfbec4..6061a6db0a4 100644
@@ -10,8 +10,8 @@
 
  #undef CONFIG_TPM_TIS_BASE_ADDRESS
 
-#define CONFIG_STD_DEVICES_SETTINGS     "stdin=usbkbd,vga,serial\0" \
-					"stdout=vga,serial\0" \
-					"stderr=vga,serial\0"
+#define CONFIG_STD_DEVICES_SETTINGS	"stdin=serial\0" \
+					"stdout=vidconsole\0" \
+					"stderr=vidconsole\0"
 
  #endif
#24 lore raw

Re: [PATCH 10/35] efi: Add video support to the app

Hi Heinrich,

Quoted 13 lines · Heinrich Schuchardt
Quoted 9 lines · Simon Glass

The current EFI video driver only works when running in the stub. In that case the stub calls boot services (before jumping to U-Boot proper) and copies the graphics info over to the efi table. This is necessary because the stub exits boot services before jumping to U-Boot.

The app maintains access to boot services throughout its life, so does not need to do this. Update the driver to support calling boot services directly.

Quoted 5 lines
Quoted 1 line

Enable video output for the app.

Please, use the EFI_GRAPHICS_OUTPUT_PROTOCOL to implement video. Don't assume any VESA or VGA compatible card. This would not be portable to other architectures.

Isn't that what this patch is doing? If not, can you please be more specific?

Regards,
Simon

#25 lore raw

PATCH 11/35 RFC: efi: Drop code that doesn't work with driver model

In reply to Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

---

 lib/efi_driver/Makefile          |  2 +-
 lib/efi_loader/Kconfig           |  2 +
 lib/efi_loader/efi_device_path.c | 96 +++++++-------------------------
 lib/efi_loader/efi_disk.c        | 48 ----------------
 4 files changed, 24 insertions(+), 124 deletions(-)
lib/efi_driver/Makefile +1 -1
index 83baa1c9a49..f2b6c05cc24 100644
@@ -6,6 +6,6 @@
 # object inclusion implicitly depends on it
 
 obj-y += efi_uclass.o
-ifeq ($(CONFIG_BLK)$(CONFIG_PARTITIONS),yy)
+ifeq ($(CONFIG_PARTITIONS),y)
 obj-y += efi_block_device.o
 endif
lib/efi_loader/Kconfig +2 -0
index dacc3b58810..799aa1a7512 100644
@@ -10,6 +10,8 @@ config EFI_LOADER
 	depends on !EFI_STUB || !X86_64 || EFI_STUB_64BIT
 	# We need EFI_STUB_32BIT to be set on x86_32 with EFI_STUB
 	depends on !EFI_STUB || !X86 || X86_64 || EFI_STUB_32BIT
+	depends on BLK
+	depends on DM_ETH || !NET
 	default y if !ARM || SYS_CPU = armv7 || SYS_CPU = armv8
 	select LIB_UUID
 	select PARTITION_UUIDS
lib/efi_loader/efi_device_path.c +21 -75

197 lines not on this page · raw message

lib/efi_loader/efi_disk.c +0 -48

61 lines not on this page · raw message

Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#26 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

Quoted 6 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Best regards

Heinrich

Quoted 315 lines

---

  lib/efi_driver/Makefile          |  2 +-
  lib/efi_loader/Kconfig           |  2 +
  lib/efi_loader/efi_device_path.c | 96 +++++++-------------------------
  lib/efi_loader/efi_disk.c        | 48 ----------------
  4 files changed, 24 insertions(+), 124 deletions(-)
lib/efi_driver/Makefile +1 -1
index 83baa1c9a49..f2b6c05cc24 100644
@@ -6,6 +6,6 @@
  # object inclusion implicitly depends on it
 
  obj-y += efi_uclass.o
-ifeq ($(CONFIG_BLK)$(CONFIG_PARTITIONS),yy)
+ifeq ($(CONFIG_PARTITIONS),y)
  obj-y += efi_block_device.o
  endif
lib/efi_loader/Kconfig +2 -0
index dacc3b58810..799aa1a7512 100644
@@ -10,6 +10,8 @@ config EFI_LOADER
  	depends on !EFI_STUB || !X86_64 || EFI_STUB_64BIT
  	# We need EFI_STUB_32BIT to be set on x86_32 with EFI_STUB
  	depends on !EFI_STUB || !X86 || X86_64 || EFI_STUB_32BIT
+	depends on BLK
+	depends on DM_ETH || !NET
  	default y if !ARM || SYS_CPU = armv7 || SYS_CPU = armv8
  	select LIB_UUID
  	select PARTITION_UUIDS
lib/efi_loader/efi_device_path.c +21 -75
index cbdb466da41..a09090a32e4 100644
@@ -46,7 +46,7 @@ static const struct efi_device_path_vendor ROOT = {
  	.guid = U_BOOT_GUID,
  };
 
-#if defined(CONFIG_DM_MMC) && defined(CONFIG_MMC)
+#if defined(CONFIG_MMC)
  /*
   * Determine if an MMC device is an SD card.
   *
@@ -486,7 +486,6 @@ bool efi_dp_is_multi_instance(const struct efi_device_path *dp)
  	return p->sub_type == DEVICE_PATH_SUB_TYPE_INSTANCE_END;
  }
 
-#ifdef CONFIG_DM
  /* size of device-path not including END node for device and all parents
   * up to the root device.
   */
@@ -503,7 +502,6 @@ __maybe_unused static unsigned int dp_size(struct udevice *dev)
  	case UCLASS_ETH:
  		return dp_size(dev->parent) +
  			sizeof(struct efi_device_path_mac_addr);
-#ifdef CONFIG_BLK
  	case UCLASS_BLK:
  		switch (dev->parent->uclass->uc_drv->id) {
  #ifdef CONFIG_IDE
@@ -511,12 +509,12 @@ __maybe_unused static unsigned int dp_size(struct udevice *dev)
  			return dp_size(dev->parent) +
  				sizeof(struct efi_device_path_atapi);
  #endif
-#if defined(CONFIG_SCSI) && defined(CONFIG_DM_SCSI)
+#if defined(CONFIG_SCSI)
  		case UCLASS_SCSI:
  			return dp_size(dev->parent) +
  				sizeof(struct efi_device_path_scsi);
  #endif
-#if defined(CONFIG_DM_MMC) && defined(CONFIG_MMC)
+#if defined(CONFIG_MMC)
  		case UCLASS_MMC:
  			return dp_size(dev->parent) +
  				sizeof(struct efi_device_path_sd_mmc_path);
@@ -554,8 +552,7 @@ __maybe_unused static unsigned int dp_size(struct udevice *dev)
  		default:
  			return dp_size(dev->parent);
  		}
-#endif
-#if defined(CONFIG_DM_MMC) && defined(CONFIG_MMC)
+#if defined(CONFIG_MMC)
  	case UCLASS_MMC:
  		return dp_size(dev->parent) +
  			sizeof(struct efi_device_path_sd_mmc_path);
@@ -590,7 +587,7 @@ __maybe_unused static void *dp_fill(void *buf, struct udevice *dev)
  		*vdp = ROOT;
  		return &vdp[1];
  	}
-#ifdef CONFIG_DM_ETH
+#ifdef CONFIG_NET
  	case UCLASS_ETH: {
  		struct efi_device_path_mac_addr *dp =
  			dp_fill(buf, dev->parent);
@@ -607,7 +604,6 @@ __maybe_unused static void *dp_fill(void *buf, struct udevice *dev)
  		return &dp[1];
  	}
  #endif
-#ifdef CONFIG_BLK
  	case UCLASS_BLK:
  		switch (dev->parent->uclass->uc_drv->id) {
  #ifdef CONFIG_SANDBOX
@@ -662,7 +658,7 @@ __maybe_unused static void *dp_fill(void *buf, struct udevice *dev)
  			return &dp[1];
  			}
  #endif
-#if defined(CONFIG_SCSI) && defined(CONFIG_DM_SCSI)
+#if defined(CONFIG_SCSI)
  		case UCLASS_SCSI: {
  			struct efi_device_path_scsi *dp =
  				dp_fill(buf, dev->parent);
@@ -676,7 +672,7 @@ __maybe_unused static void *dp_fill(void *buf, struct udevice *dev)
  			return &dp[1];
  			}
  #endif
-#if defined(CONFIG_DM_MMC) && defined(CONFIG_MMC)
+#if defined(CONFIG_MMC)
  		case UCLASS_MMC: {
  			struct efi_device_path_sd_mmc_path *sddp =
  				dp_fill(buf, dev->parent);
@@ -727,8 +723,7 @@ __maybe_unused static void *dp_fill(void *buf, struct udevice *dev)
  			      dev->name, dev->parent->uclass->uc_drv->id);
  			return dp_fill(buf, dev->parent);
  		}
-#endif
-#if defined(CONFIG_DM_MMC) && defined(CONFIG_MMC)
+#if defined(CONFIG_MMC)
  	case UCLASS_MMC: {
  		struct efi_device_path_sd_mmc_path *sddp =
  			dp_fill(buf, dev->parent);
@@ -770,24 +765,18 @@ __maybe_unused static void *dp_fill(void *buf, struct udevice *dev)
  		return dp_fill(buf, dev->parent);
  	}
  }
-#endif
 
  static unsigned dp_part_size(struct blk_desc *desc, int part)
  {
  	unsigned dpsize;
+	struct udevice *dev;
+	int ret;
 
-#ifdef CONFIG_BLK
-	{
-		struct udevice *dev;
-		int ret = blk_find_device(desc->if_type, desc->devnum, &dev);
+	ret = blk_find_device(desc->if_type, desc->devnum, &dev);
 
-		if (ret)
-			dev = desc->bdev->parent;
-		dpsize = dp_size(dev);
-	}
-#else
-	dpsize = sizeof(ROOT) + sizeof(struct efi_device_path_usb);
-#endif
+	if (ret)
+		dev = desc->bdev->parent;
+	dpsize = dp_size(dev);
 
  	if (part == 0) /* the actual disk, not a partition */
  		return dpsize;
@@ -877,36 +866,14 @@ static void *dp_part_node(void *buf, struct blk_desc *desc, int part)
   */
  static void *dp_part_fill(void *buf, struct blk_desc *desc, int part)
  {
-#ifdef CONFIG_BLK
-	{
-		struct udevice *dev;
-		int ret = blk_find_device(desc->if_type, desc->devnum, &dev);
+	struct udevice *dev;
+	int ret;
 
-		if (ret)
-			dev = desc->bdev->parent;
-		buf = dp_fill(buf, dev);
-	}
-#else
-	/*
-	 * We *could* make a more accurate path, by looking at if_type
-	 * and handling all the different cases like we do for non-
-	 * legacy (i.e. CONFIG_BLK=y) case. But most important thing
-	 * is just to have a unique device-path for if_type+devnum.
-	 * So map things to a fictitious USB device.
-	 */
-	struct efi_device_path_usb *udp;
-
-	memcpy(buf, &ROOT, sizeof(ROOT));
-	buf += sizeof(ROOT);
-
-	udp = buf;
-	udp->dp.type = DEVICE_PATH_TYPE_MESSAGING_DEVICE;
-	udp->dp.sub_type = DEVICE_PATH_SUB_TYPE_MSG_USB;
-	udp->dp.length = sizeof(*udp);
-	udp->parent_port_number = desc->if_type;
-	udp->usb_interface = desc->devnum;
-	buf = &udp[1];
-#endif
+	ret = blk_find_device(desc->if_type, desc->devnum, &dev);
+
+	if (ret)
+		dev = desc->bdev->parent;
+	buf = dp_fill(buf, dev);
 
  	if (part == 0) /* the actual disk, not a partition */
  		return buf;
@@ -1051,39 +1018,18 @@ struct efi_device_path *efi_dp_from_uart(void)
  #ifdef CONFIG_NET
  struct efi_device_path *efi_dp_from_eth(void)
  {
-#ifndef CONFIG_DM_ETH
-	struct efi_device_path_mac_addr *ndp;
-#endif
  	void *buf, *start;
  	unsigned dpsize = 0;
 
  	assert(eth_get_dev());
 
-#ifdef CONFIG_DM_ETH
  	dpsize += dp_size(eth_get_dev());
-#else
-	dpsize += sizeof(ROOT);
-	dpsize += sizeof(*ndp);
-#endif
 
  	start = buf = dp_alloc(dpsize + sizeof(END));
  	if (!buf)
  		return NULL;
 
-#ifdef CONFIG_DM_ETH
  	buf = dp_fill(buf, eth_get_dev());
-#else
-	memcpy(buf, &ROOT, sizeof(ROOT));
-	buf += sizeof(ROOT);
-
-	ndp = buf;
-	ndp->dp.type = DEVICE_PATH_TYPE_MESSAGING_DEVICE;
-	ndp->dp.sub_type = DEVICE_PATH_SUB_TYPE_MSG_MAC_ADDR;
-	ndp->dp.length = sizeof(*ndp);
-	ndp->if_type = 1; /* Ethernet */
-	memcpy(ndp->mac.addr, eth_get_ethaddr(), ARP_HLEN);
-	buf = &ndp[1];
-#endif
 
  	*((struct efi_device_path *)buf) = END;
 
lib/efi_loader/efi_disk.c +0 -48
index 988907ecb91..ef8b5c88ff9 100644
@@ -555,7 +555,6 @@ efi_status_t efi_disk_register(void)
  	struct efi_disk_obj *disk;
  	int disks = 0;
  	efi_status_t ret;
-#ifdef CONFIG_BLK
  	struct udevice *dev;
 
  	for (uclass_first_device_check(UCLASS_BLK, &dev); dev;
@@ -583,54 +582,7 @@ efi_status_t efi_disk_register(void)
  					&disk->header, desc, if_typename,
  					desc->devnum, dev->name);
  	}
-#else
-	int i, if_type;
 
-	/* Search for all available disk devices */
-	for (if_type = 0; if_type < IF_TYPE_COUNT; if_type++) {
-		const struct blk_driver *cur_drvr;
-		const char *if_typename;
-
-		cur_drvr = blk_driver_lookup_type(if_type);
-		if (!cur_drvr)
-			continue;
-
-		if_typename = cur_drvr->if_typename;
-		log_info("Scanning disks on %s...\n", if_typename);
-		for (i = 0; i < 4; i++) {
-			struct blk_desc *desc;
-			char devname[32] = { 0 }; /* dp->str is u16[32] long */
-
-			desc = blk_get_devnum_by_type(if_type, i);
-			if (!desc)
-				continue;
-			if (desc->type == DEV_TYPE_UNKNOWN)
-				continue;
-
-			snprintf(devname, sizeof(devname), "%s%d",
-				 if_typename, i);
-
-			/* Add block device for the full device */
-			ret = efi_disk_add_dev(NULL, NULL, if_typename, desc,
-					       i, NULL, 0, &disk);
-			if (ret == EFI_NOT_READY) {
-				log_notice("Disk %s not ready\n", devname);
-				continue;
-			}
-			if (ret) {
-				log_err("ERROR: failure to add disk device %s, r = %lu\n",
-					devname, ret & ~EFI_ERROR_MASK);
-				return ret;
-			}
-			disks++;
-
-			/* Partitions show up as block devices in EFI */
-			disks += efi_disk_create_partitions
-						(&disk->header, desc,
-						 if_typename, i, devname);
-		}
-	}
-#endif
  	log_info("Found %d disks\n", disks);
 
  	return EFI_SUCCESS;
#27 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

Hi Heinrich,

Quoted 7 lines · Heinrich Schuchardt
Quoted 3 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Quoted 8 lines
Quoted 3 lines

Boards which want EFI_LOADER should migrate to driver model first.

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

I need this patch for this series to work. You can still review things for other maintainers and in this case it is common for one maintainer to pick up the series once the others are happy.

Regards,
Simon

#28 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

Quoted 12 lines · Simon Glass

Hi Heinrich,

Quoted 9 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

Until 623b3a57976 ("efi_selftest: provide an EFI selftest application") there was no official maintainer for lib/efi/ but you were the main contributor.

But with that patch directory lib/efi/ was assigned to EFI PAYLOAD.

I am happy if you would continue to care about U-Boot on EFI.

Quoted 4 lines

I need this patch for this series to work. You can still review things for other maintainers and in this case it is common for one maintainer to pick up the series once the others are happy.

The direction of this patch is completely correct.

There are some things that will have to be changed, e.g we should not require CONFIG_DM_ETH=y. I will work on reviewing this patch in detail.

I already added CONFIG_BLK as a requirement for CONFIG_EFI_LOADER in a submitted patch.

Removing legacy code from lib/efi_loader/efi_disk.c and lib/efi_loader/efi_device_path.c could be done before all U-Boot on EFI patches.

Therefore I still think it makes sense to split the series in two:

1) Cleanup of the UEFI implementation
2) Rework of U-Boot on EFI

I hope merging in this sequence of patch series makes send to you.

Best regards

Heinrich

#29 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

Hi Heinrich,

Quoted 25 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

Hi Heinrich,

Quoted 9 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

Quoted 7 lines

Until 623b3a57976 ("efi_selftest: provide an EFI selftest application") there was no official maintainer for lib/efi/ but you were the main contributor.

But with that patch directory lib/efi/ was assigned to EFI PAYLOAD.

I am happy if you would continue to care about U-Boot on EFI.

OK.

Quoted 10 lines
Quoted 5 lines

I need this patch for this series to work. You can still review things for other maintainers and in this case it is common for one maintainer to pick up the series once the others are happy.

The direction of this patch is completely correct.

There are some things that will have to be changed, e.g we should not require CONFIG_DM_ETH=y. I will work on reviewing this patch in detail.

OK, but why not require DM_ETH? The deadline passed a year ago.

Quoted 3 lines

I already added CONFIG_BLK as a requirement for CONFIG_EFI_LOADER in a submitted patch.

OK good.

Quoted 11 lines

Removing legacy code from lib/efi_loader/efi_disk.c and lib/efi_loader/efi_device_path.c could be done before all U-Boot on EFI patches.

Therefore I still think it makes sense to split the series in two:

1) Cleanup of the UEFI implementation
2) Rework of U-Boot on EFI

I hope merging in this sequence of patch series makes send to you.

I am fine if you want to take on the clean-up stuff for BLK, etc. In that case we can just drop this patch when applying. But I do need some sort of patch here, for things to work.

Regards,
Simon

#30 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

Quoted 44 lines · Simon Glass

Hi Heinrich,

Quoted 25 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

Hi Heinrich,

Quoted 9 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

Quoted 7 lines

Until 623b3a57976 ("efi_selftest: provide an EFI selftest application") there was no official maintainer for lib/efi/ but you were the main contributor.

But with that patch directory lib/efi/ was assigned to EFI PAYLOAD.

I am happy if you would continue to care about U-Boot on EFI.

OK.

Quoted 6 lines
Quoted 5 lines

I need this patch for this series to work. You can still review things for other maintainers and in this case it is common for one maintainer to pick up the series once the others are happy.

Quoted 6 lines
Quoted 4 lines

The direction of this patch is completely correct.

There are some things that will have to be changed, e.g we should not require CONFIG_DM_ETH=y. I will work on reviewing this patch in detail.

OK, but why not require DM_ETH? The deadline passed a year ago.

Note that it's still a while (about a year) before I forcefully yank out unmigrated systems.

Signature · 1 line
Tom
#31 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

In reply to Simon Glass
Quoted 47 lines

From: Simon Glass <[email protected]>
Date: Thu, 9 Sep 2021 13:57:39 -0600

Hi Heinrich,

Quoted 25 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

Hi Heinrich,

Quoted 9 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

Quoted 7 lines

Until 623b3a57976 ("efi_selftest: provide an EFI selftest application") there was no official maintainer for lib/efi/ but you were the main contributor.

But with that patch directory lib/efi/ was assigned to EFI PAYLOAD.

I am happy if you would continue to care about U-Boot on EFI.

OK.

Quoted 6 lines
Quoted 5 lines

I need this patch for this series to work. You can still review things for other maintainers and in this case it is common for one maintainer to pick up the series once the others are happy.

Quoted 6 lines
Quoted 4 lines

The direction of this patch is completely correct.

There are some things that will have to be changed, e.g we should not require CONFIG_DM_ETH=y. I will work on reviewing this patch in detail.

OK, but why not require DM_ETH? The deadline passed a year ago.

Because we support boards without network ports?

#32 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

Quoted 47 lines · Mark Kettenis
Quoted 47 lines

From: Simon Glass <[email protected]>
Date: Thu, 9 Sep 2021 13:57:39 -0600

Hi Heinrich,

Quoted 25 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

Hi Heinrich,

Quoted 9 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

Quoted 7 lines

Until 623b3a57976 ("efi_selftest: provide an EFI selftest application") there was no official maintainer for lib/efi/ but you were the main contributor.

But with that patch directory lib/efi/ was assigned to EFI PAYLOAD.

I am happy if you would continue to care about U-Boot on EFI.

OK.

Quoted 6 lines
Quoted 5 lines

I need this patch for this series to work. You can still review things for other maintainers and in this case it is common for one maintainer to pick up the series once the others are happy.

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

The direction of this patch is completely correct.

There are some things that will have to be changed, e.g we should not require CONFIG_DM_ETH=y. I will work on reviewing this patch in detail.

OK, but why not require DM_ETH? The deadline passed a year ago.

Because we support boards without network ports?

Boards without networking should disable the relevant code, and as needed the EFI code return the proper error code?

Signature · 1 line
Tom
#33 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

Quoted 56 lines

Date: Thu, 9 Sep 2021 16:23:08 -0400
From: Tom Rini <[email protected]>

Quoted 52 lines · Mark Kettenis
Quoted 52 lines

From: Simon Glass <[email protected]>
Date: Thu, 9 Sep 2021 13:57:39 -0600

Hi Heinrich,

Quoted 25 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

Hi Heinrich,

Quoted 9 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

Quoted 7 lines

Until 623b3a57976 ("efi_selftest: provide an EFI selftest application") there was no official maintainer for lib/efi/ but you were the main contributor.

But with that patch directory lib/efi/ was assigned to EFI PAYLOAD.

I am happy if you would continue to care about U-Boot on EFI.

OK.

Quoted 10 lines
Quoted 5 lines

I need this patch for this series to work. You can still review things for other maintainers and in this case it is common for one maintainer to pick up the series once the others are happy.

The direction of this patch is completely correct.

There are some things that will have to be changed, e.g we should not require CONFIG_DM_ETH=y. I will work on reviewing this patch in detail.

Quoted 6 lines
Quoted 3 lines
Quoted 1 line

OK, but why not require DM_ETH? The deadline passed a year ago.

Because we support boards without network ports?

Boards without networking should disable the relevant code, and as needed the EFI code return the proper error code?

Yes, but it means you can't make DM_ETH a (hard) requirement for EFI_LOADER support. What I mean is that it should still be possible to build U-Boot with EFI_LOADER support even if DM_EFI isn't set for a board. It should just result in a UEFI implementation with no network support instead.

#34 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

Quoted 60 lines · Mark Kettenis
Quoted 60 lines

Date: Thu, 9 Sep 2021 16:23:08 -0400
From: Tom Rini <[email protected]>

Quoted 55 lines · Mark Kettenis
Quoted 47 lines

From: Simon Glass <[email protected]>
Date: Thu, 9 Sep 2021 13:57:39 -0600

Hi Heinrich,

Quoted 25 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

Hi Heinrich,

Quoted 9 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

Quoted 7 lines

Until 623b3a57976 ("efi_selftest: provide an EFI selftest application") there was no official maintainer for lib/efi/ but you were the main contributor.

But with that patch directory lib/efi/ was assigned to EFI PAYLOAD.

I am happy if you would continue to care about U-Boot on EFI.

OK.

Quoted 6 lines
Quoted 5 lines

I need this patch for this series to work. You can still review things for other maintainers and in this case it is common for one maintainer to pick up the series once the others are happy.

Quoted 6 lines
Quoted 4 lines

The direction of this patch is completely correct.

There are some things that will have to be changed, e.g we should not require CONFIG_DM_ETH=y. I will work on reviewing this patch in detail.

OK, but why not require DM_ETH? The deadline passed a year ago.

Because we support boards without network ports?

Quoted 8 lines
Quoted 2 lines

Boards without networking should disable the relevant code, and as needed the EFI code return the proper error code?

Yes, but it means you can't make DM_ETH a (hard) requirement for EFI_LOADER support. What I mean is that it should still be possible to build U-Boot with EFI_LOADER support even if DM_EFI isn't set for a board. It should just result in a UEFI implementation with no network support instead.

Yes, agreed. I was just trying to say that in the context of what DM code EFI_LOADER can demand, the deadline for BLK has passed and everything that didn't support it has been removed, so that's a good requirement and area of code to clean up as needed. But DM_ETH-or-bust isn't there, yet.

Signature · 1 line
Tom
#35 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

In reply to Mark Kettenis

Hi Mark,

Quoted 61 lines · Mark Kettenis
Quoted 61 lines

Date: Thu, 9 Sep 2021 16:23:08 -0400
From: Tom Rini <[email protected]>

Quoted 55 lines · Mark Kettenis
Quoted 47 lines

From: Simon Glass <[email protected]>
Date: Thu, 9 Sep 2021 13:57:39 -0600

Hi Heinrich,

Quoted 25 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

Hi Heinrich,

Quoted 9 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

Quoted 7 lines

Until 623b3a57976 ("efi_selftest: provide an EFI selftest application") there was no official maintainer for lib/efi/ but you were the main contributor.

But with that patch directory lib/efi/ was assigned to EFI PAYLOAD.

I am happy if you would continue to care about U-Boot on EFI.

OK.

Quoted 6 lines
Quoted 5 lines

I need this patch for this series to work. You can still review things for other maintainers and in this case it is common for one maintainer to pick up the series once the others are happy.

Quoted 6 lines
Quoted 4 lines

The direction of this patch is completely correct.

There are some things that will have to be changed, e.g we should not require CONFIG_DM_ETH=y. I will work on reviewing this patch in detail.

OK, but why not require DM_ETH? The deadline passed a year ago.

Because we support boards without network ports?

Quoted 8 lines
Quoted 2 lines

Boards without networking should disable the relevant code, and as needed the EFI code return the proper error code?

Yes, but it means you can't make DM_ETH a (hard) requirement for EFI_LOADER support. What I mean is that it should still be possible to build U-Boot with EFI_LOADER support even if DM_EFI isn't set for a board. It should just result in a UEFI implementation with no network support instead.

I think you are misunderstanding my patch. I have:

depends on DM_ETH || !NET

which means that if NET is used, DM_ETH must be. I think that is reasonable.

Regards,
Simon

#36 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

Quoted 76 lines

From: Simon Glass <[email protected]>
Date: Thu, 23 Sep 2021 20:48:42 -0600

Hi Mark,

Quoted 61 lines · Mark Kettenis
Quoted 61 lines

Date: Thu, 9 Sep 2021 16:23:08 -0400
From: Tom Rini <[email protected]>

Quoted 55 lines · Mark Kettenis
Quoted 47 lines

From: Simon Glass <[email protected]>
Date: Thu, 9 Sep 2021 13:57:39 -0600

Hi Heinrich,

Quoted 25 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

Hi Heinrich,

Quoted 9 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

Quoted 7 lines

Until 623b3a57976 ("efi_selftest: provide an EFI selftest application") there was no official maintainer for lib/efi/ but you were the main contributor.

But with that patch directory lib/efi/ was assigned to EFI PAYLOAD.

I am happy if you would continue to care about U-Boot on EFI.

OK.

Quoted 6 lines
Quoted 5 lines

I need this patch for this series to work. You can still review things for other maintainers and in this case it is common for one maintainer to pick up the series once the others are happy.

Quoted 6 lines
Quoted 4 lines

The direction of this patch is completely correct.

There are some things that will have to be changed, e.g we should not require CONFIG_DM_ETH=y. I will work on reviewing this patch in detail.

OK, but why not require DM_ETH? The deadline passed a year ago.

Because we support boards without network ports?

Quoted 8 lines
Quoted 2 lines

Boards without networking should disable the relevant code, and as needed the EFI code return the proper error code?

Yes, but it means you can't make DM_ETH a (hard) requirement for EFI_LOADER support. What I mean is that it should still be possible to build U-Boot with EFI_LOADER support even if DM_EFI isn't set for a board. It should just result in a UEFI implementation with no network support instead.

Quoted 5 lines

I think you are misunderstanding my patch. I have:

depends on DM_ETH || !NET

which means that if NET is used, DM_ETH must be. I think that is reasonable.

Yes, that should be fine (Kconfig stuff doesn't always make sense to me).

#37 lore raw

Re: [PATCH 11/35] RFC: efi: Drop code that doesn't work with driver model

In reply to Heinrich Schuchardt

Hi Heinrich,

Quoted 25 lines · Heinrich Schuchardt
Quoted 12 lines · Simon Glass

Hi Heinrich,

Quoted 9 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

This code should never have been added as it builds a new feature on top of legacy code. Drop it and add a dependency on BLK for this feature.

Boards which want EFI_LOADER should migrate to driver model first.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

This patch is not related to the rest of the series and the code has a different maintainer.

So, please, separate it from the series.

Who is the maintainer?

Quoted 7 lines

Until 623b3a57976 ("efi_selftest: provide an EFI selftest application") there was no official maintainer for lib/efi/ but you were the main contributor.

But with that patch directory lib/efi/ was assigned to EFI PAYLOAD.

I am happy if you would continue to care about U-Boot on EFI.

I can add myself as maintainer of that part of EFI if you like. I will include it in the v2 series. But it is a little fiddling because there are shared files, so I will just add a few for now.

Perhaps you could be co-maintainer for the app, given your EFI knowledge and that you review them? Also I expect they would still go through your tree.

Regards,
Simon

#38 lore raw

PATCH 12/35 efi: Add EFI uclass for media

In reply to Simon Glass

At present UCLASS_EFI is used to represent an EFI filesystem The description of this uclass is "EFI managed devices" which is pretty vague. The only driver that uses this uclass is in fact not a real U-Boot driver, since its operations do not include a struct udevice.

Rather than mess with this, create a new UCLASS_EFI_MEDIA uclass to handle EFI media such as disks. Make this the uclass to use for EFI media.

The existing implementation using UCLASS_EFI remains as is, for discussion.

---

 arch/sandbox/dts/test.dts        |  4 ++++
 drivers/block/Kconfig            | 23 +++++++++++++++++++++++
 drivers/block/Makefile           |  3 +++
 drivers/block/blk-uclass.c       |  2 +-
 drivers/block/efi-media-uclass.c | 15 +++++++++++++++
 drivers/block/sb_efi_media.c     | 21 +++++++++++++++++++++
 include/dm/uclass-id.h           |  1 +
 test/dm/Makefile                 |  1 +
 test/dm/efi_media.c              | 24 ++++++++++++++++++++++++
 9 files changed, 93 insertions(+), 1 deletion(-)
 create mode 100644 drivers/block/efi-media-uclass.c
 create mode 100644 drivers/block/sb_efi_media.c
 create mode 100644 test/dm/efi_media.c
arch/sandbox/dts/test.dts +4 -0
index 962bdbe5567..af88cf681ae 100644
@@ -486,6 +486,10 @@
 		compatible = "sandbox,clk-ccf";
 	};
 
+	efi-media {
+		compatible = "sandbox,efi-media";
+	};
+
 	eth@10002000 {
 		compatible = "sandbox,eth";
 		reg = <0x10002000 0x1000>;
drivers/block/Kconfig +23 -0
index 4023332dd98..058956ee8ee 100644
@@ -63,6 +63,29 @@ config TPL_BLOCK_CACHE
 	help
 	  This option enables the disk-block cache in TPL
 
+config EFI_MEDIA
+	bool "Support EFI media drivers"
+	default y if EFI || SANDBOX
+	help
+	  Enable this to support media devices on top of UEFI. This enables
+	  just the uclass so you also need a specific driver to make this do
+	  anything.
+
+	  For sandbox there is a test driver.
+
+if EFI_MEDIA
+
+config EFI_MEDIA_SANDBOX
+	bool "Sandbox EFI media driver"
+	depends on SANDBOX
+	default y
+	help
+	  Enables a simple sandbox media driver, used for testing just the
+	  EFI_MEDIA uclass. It does not do anything useful, since sandbox does
+	  not actually support running on top of UEFI.
+
+endif  # EFI_MEDIA
+
 config IDE
 	bool "Support IDE controllers"
 	select HAVE_BLOCK_DEVICE
drivers/block/Makefile +3 -0
index 94ab5c6f906..3778633da1d 100644
@@ -14,3 +14,6 @@ obj-$(CONFIG_IDE) += ide.o
 endif
 obj-$(CONFIG_SANDBOX) += sandbox.o
 obj-$(CONFIG_$(SPL_TPL_)BLOCK_CACHE) += blkcache.o
+
+obj-$(CONFIG_EFI_MEDIA) += efi-media-uclass.o
+obj-$(CONFIG_EFI_MEDIA_SANDBOX) += sb_efi_media.o
drivers/block/blk-uclass.c +1 -1
index 83682dcc181..2db7ce5de20 100644
@@ -44,7 +44,7 @@ static enum uclass_id if_type_uclass_id[IF_TYPE_COUNT] = {
 	[IF_TYPE_SATA]		= UCLASS_AHCI,
 	[IF_TYPE_HOST]		= UCLASS_ROOT,
 	[IF_TYPE_NVME]		= UCLASS_NVME,
-	[IF_TYPE_EFI]		= UCLASS_EFI,
+	[IF_TYPE_EFI]		= UCLASS_EFI_MEDIA,
 	[IF_TYPE_VIRTIO]	= UCLASS_VIRTIO,
 	[IF_TYPE_PVBLOCK]	= UCLASS_PVBLOCK,
 };
drivers/block/efi-media-uclass.c new file +15 -0
new file mode 100644
index 00000000000..e012f6f2f4c
@@ -0,0 +1,15 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Uclass for EFI media devices
+ *
+ * Copyright 2021 Google LLC
+ */
+
+#include <common.h>
+#include <dm.h>
+
+UCLASS_DRIVER(efi_media) = {
+	.id		= UCLASS_EFI_MEDIA,
+	.name		= "efi_media",
+	.flags		= DM_UC_FLAG_SEQ_ALIAS,
+};
drivers/block/sb_efi_media.c new file +21 -0
new file mode 100644
index 00000000000..4f242aa13a4
@@ -0,0 +1,21 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * EFI_MEDIA driver for sandbox
+ *
+ * Copyright 2021 Google LLC
+ */
+
+#include <common.h>
+#include <dm.h>
+
+static const struct udevice_id sandbox_efi_media_ids[] = {
+	{ .compatible = "sandbox,efi-media" },
+	{ }
+};
+
+
+U_BOOT_DRIVER(sandbox_efi_media) = {
+	.name		= "sandbox_efi_media",
+	.id		= UCLASS_EFI_MEDIA,
+	.of_match	= sandbox_efi_media_ids,
+};
include/dm/uclass-id.h +1 -0
index e7edd409f30..4906eb836c2 100644
@@ -49,6 +49,7 @@ enum uclass_id {
 	UCLASS_DSA,		/* Distributed (Ethernet) Switch Architecture */
 	UCLASS_ECDSA,		/* Elliptic curve cryptographic device */
 	UCLASS_EFI,		/* EFI managed devices */
+	UCLASS_EFI_MEDIA,	/* EFI media (e.g. a disk) */
 	UCLASS_ETH,		/* Ethernet device */
 	UCLASS_ETH_PHY,		/* Ethernet PHY device */
 	UCLASS_FIRMWARE,	/* Firmware */
test/dm/Makefile +1 -0
index 516f69d61cb..d9bc37d8313 100644
@@ -37,6 +37,7 @@ obj-$(CONFIG_DMA) += dma.o
 obj-$(CONFIG_VIDEO_MIPI_DSI) += dsi_host.o
 obj-$(CONFIG_DM_DSA) += dsa.o
 obj-$(CONFIG_ECDSA_VERIFY) += ecdsa.o
+obj-$(CONFIG_EFI_MEDIA_SANDBOX) += efi_media.o
 obj-$(CONFIG_DM_ETH) += eth.o
 ifneq ($(CONFIG_EFI_PARTITION),)
 obj-$(CONFIG_FASTBOOT_FLASH_MMC) += fastboot.o
test/dm/efi_media.c new file +24 -0
new file mode 100644
index 00000000000..e343a0e9c85
@@ -0,0 +1,24 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Test for EFI_MEDIA uclass
+ *
+ * Copyright 2021 Google LLC
+ */
+
+#include <common.h>
+#include <dm.h>
+#include <asm/test.h>
+#include <dm/test.h>
+#include <test/test.h>
+#include <test/ut.h>
+
+/* Test that we can use the EFI_MEDIA uclass */
+static int dm_test_efi_media(struct unit_test_state *uts)
+{
+	struct udevice *dev;
+
+	ut_assertok(uclass_first_device_err(UCLASS_EFI_MEDIA, &dev));
+
+	return 0;
+}
+DM_TEST(dm_test_efi_media, UT_TESTF_SCAN_PDATA | UT_TESTF_SCAN_FDT);
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#39 lore raw

Re: [PATCH 12/35] efi: Add EFI uclass for media

Quoted 10 lines · Simon Glass

At present UCLASS_EFI is used to represent an EFI filesystem The description of this uclass is "EFI managed devices" which is pretty vague. The only driver that uses this uclass is in fact not a real U-Boot driver, since its operations do not include a struct udevice.

Rather than mess with this, create a new UCLASS_EFI_MEDIA uclass to handle EFI media such as disks. Make this the uclass to use for EFI media.

The existing implementation using UCLASS_EFI remains as is, for discussion.

UCLASS_EFI is not related to block devices. It is for drivers implementing the EFI_DRIVER_BINDING_PROTOCOL. efi_uclass.c simplifies writing such drivers. So your patch does not make sense to me.

Please, separate the series in two: one for U-Boot on EFI, one for the UEFI implementation in U-Boot.

Best regards

Heinrich

Quoted 194 lines

---

  arch/sandbox/dts/test.dts        |  4 ++++
  drivers/block/Kconfig            | 23 +++++++++++++++++++++++
  drivers/block/Makefile           |  3 +++
  drivers/block/blk-uclass.c       |  2 +-
  drivers/block/efi-media-uclass.c | 15 +++++++++++++++
  drivers/block/sb_efi_media.c     | 21 +++++++++++++++++++++
  include/dm/uclass-id.h           |  1 +
  test/dm/Makefile                 |  1 +
  test/dm/efi_media.c              | 24 ++++++++++++++++++++++++
  9 files changed, 93 insertions(+), 1 deletion(-)

create mode 100644 drivers/block/efi-media-uclass.c create mode 100644 drivers/block/sb_efi_media.c create mode 100644 test/dm/efi_media.c

arch/sandbox/dts/test.dts +4 -0
index 962bdbe5567..af88cf681ae 100644
@@ -486,6 +486,10 @@
  		compatible = "sandbox,clk-ccf";
  	};
 
+	efi-media {
+		compatible = "sandbox,efi-media";
+	};
+
  	eth@10002000 {
  		compatible = "sandbox,eth";
  		reg = <0x10002000 0x1000>;
drivers/block/Kconfig +23 -0
index 4023332dd98..058956ee8ee 100644
@@ -63,6 +63,29 @@ config TPL_BLOCK_CACHE
  	help
  	  This option enables the disk-block cache in TPL
 
+config EFI_MEDIA
+	bool "Support EFI media drivers"
+	default y if EFI || SANDBOX
+	help
+	  Enable this to support media devices on top of UEFI. This enables
+	  just the uclass so you also need a specific driver to make this do
+	  anything.
+
+	  For sandbox there is a test driver.
+
+if EFI_MEDIA
+
+config EFI_MEDIA_SANDBOX
+	bool "Sandbox EFI media driver"
+	depends on SANDBOX
+	default y
+	help
+	  Enables a simple sandbox media driver, used for testing just the
+	  EFI_MEDIA uclass. It does not do anything useful, since sandbox does
+	  not actually support running on top of UEFI.
+
+endif  # EFI_MEDIA
+
  config IDE
  	bool "Support IDE controllers"
  	select HAVE_BLOCK_DEVICE
drivers/block/Makefile +3 -0
index 94ab5c6f906..3778633da1d 100644
@@ -14,3 +14,6 @@ obj-$(CONFIG_IDE) += ide.o
  endif
  obj-$(CONFIG_SANDBOX) += sandbox.o
  obj-$(CONFIG_$(SPL_TPL_)BLOCK_CACHE) += blkcache.o
+
+obj-$(CONFIG_EFI_MEDIA) += efi-media-uclass.o
+obj-$(CONFIG_EFI_MEDIA_SANDBOX) += sb_efi_media.o
drivers/block/blk-uclass.c +1 -1
index 83682dcc181..2db7ce5de20 100644
@@ -44,7 +44,7 @@ static enum uclass_id if_type_uclass_id[IF_TYPE_COUNT] = {
  	[IF_TYPE_SATA]		= UCLASS_AHCI,
  	[IF_TYPE_HOST]		= UCLASS_ROOT,
  	[IF_TYPE_NVME]		= UCLASS_NVME,
-	[IF_TYPE_EFI]		= UCLASS_EFI,
+	[IF_TYPE_EFI]		= UCLASS_EFI_MEDIA,
  	[IF_TYPE_VIRTIO]	= UCLASS_VIRTIO,
  	[IF_TYPE_PVBLOCK]	= UCLASS_PVBLOCK,
  };
drivers/block/efi-media-uclass.c new file +15 -0
new file mode 100644
index 00000000000..e012f6f2f4c
@@ -0,0 +1,15 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Uclass for EFI media devices
+ *
+ * Copyright 2021 Google LLC
+ */
+
+#include <common.h>
+#include <dm.h>
+
+UCLASS_DRIVER(efi_media) = {
+	.id		= UCLASS_EFI_MEDIA,
+	.name		= "efi_media",
+	.flags		= DM_UC_FLAG_SEQ_ALIAS,
+};
drivers/block/sb_efi_media.c new file +21 -0
new file mode 100644
index 00000000000..4f242aa13a4
@@ -0,0 +1,21 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * EFI_MEDIA driver for sandbox
+ *
+ * Copyright 2021 Google LLC
+ */
+
+#include <common.h>
+#include <dm.h>
+
+static const struct udevice_id sandbox_efi_media_ids[] = {
+	{ .compatible = "sandbox,efi-media" },
+	{ }
+};
+
+
+U_BOOT_DRIVER(sandbox_efi_media) = {
+	.name		= "sandbox_efi_media",
+	.id		= UCLASS_EFI_MEDIA,
+	.of_match	= sandbox_efi_media_ids,
+};
include/dm/uclass-id.h +1 -0
index e7edd409f30..4906eb836c2 100644
@@ -49,6 +49,7 @@ enum uclass_id {
  	UCLASS_DSA,		/* Distributed (Ethernet) Switch Architecture */
  	UCLASS_ECDSA,		/* Elliptic curve cryptographic device */
  	UCLASS_EFI,		/* EFI managed devices */
+	UCLASS_EFI_MEDIA,	/* EFI media (e.g. a disk) */
  	UCLASS_ETH,		/* Ethernet device */
  	UCLASS_ETH_PHY,		/* Ethernet PHY device */
  	UCLASS_FIRMWARE,	/* Firmware */
test/dm/Makefile +1 -0
index 516f69d61cb..d9bc37d8313 100644
@@ -37,6 +37,7 @@ obj-$(CONFIG_DMA) += dma.o
  obj-$(CONFIG_VIDEO_MIPI_DSI) += dsi_host.o
  obj-$(CONFIG_DM_DSA) += dsa.o
  obj-$(CONFIG_ECDSA_VERIFY) += ecdsa.o
+obj-$(CONFIG_EFI_MEDIA_SANDBOX) += efi_media.o
  obj-$(CONFIG_DM_ETH) += eth.o
  ifneq ($(CONFIG_EFI_PARTITION),)
  obj-$(CONFIG_FASTBOOT_FLASH_MMC) += fastboot.o
test/dm/efi_media.c new file +24 -0
new file mode 100644
index 00000000000..e343a0e9c85
@@ -0,0 +1,24 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Test for EFI_MEDIA uclass
+ *
+ * Copyright 2021 Google LLC
+ */
+
+#include <common.h>
+#include <dm.h>
+#include <asm/test.h>
+#include <dm/test.h>
+#include <test/test.h>
+#include <test/ut.h>
+
+/* Test that we can use the EFI_MEDIA uclass */
+static int dm_test_efi_media(struct unit_test_state *uts)
+{
+	struct udevice *dev;
+
+	ut_assertok(uclass_first_device_err(UCLASS_EFI_MEDIA, &dev));
+
+	return 0;
+}
+DM_TEST(dm_test_efi_media, UT_TESTF_SCAN_PDATA | UT_TESTF_SCAN_FDT);
#40 lore raw

Re: [PATCH 12/35] efi: Add EFI uclass for media

Hi Heinrich,

Quoted 12 lines · Heinrich Schuchardt
Quoted 8 lines · Simon Glass

At present UCLASS_EFI is used to represent an EFI filesystem The description of this uclass is "EFI managed devices" which is pretty vague. The only driver that uses this uclass is in fact not a real U-Boot driver, since its operations do not include a struct udevice.

Rather than mess with this, create a new UCLASS_EFI_MEDIA uclass to handle EFI media such as disks. Make this the uclass to use for EFI media.

Quoted 6 lines
Quoted 2 lines

The existing implementation using UCLASS_EFI remains as is, for discussion.

UCLASS_EFI is not related to block devices. It is for drivers implementing the EFI_DRIVER_BINDING_PROTOCOL. efi_uclass.c simplifies writing such drivers. So your patch does not make sense to me.

Do you understand the issue here? The existing EFI uclass is not a real U-Boot uclass. It does not use DM-style operations or even devices. It actually isn't clear to me what it is for. So for the purposes of this patch, I ignored it.

Quoted 3 lines

Please, separate the series in two: one for U-Boot on EFI, one for the UEFI implementation in U-Boot.

This series is all about the former. If there is anything related to the latter, it is likely just refactoring to avoid duplicating code. But I can't think of anything at present - can you please be more specific?

Quoted 22 lines

Best regards

Heinrich

Quoted 17 lines

---

  arch/sandbox/dts/test.dts        |  4 ++++
  drivers/block/Kconfig            | 23 +++++++++++++++++++++++
  drivers/block/Makefile           |  3 +++
  drivers/block/blk-uclass.c       |  2 +-
  drivers/block/efi-media-uclass.c | 15 +++++++++++++++
  drivers/block/sb_efi_media.c     | 21 +++++++++++++++++++++
  include/dm/uclass-id.h           |  1 +
  test/dm/Makefile                 |  1 +
  test/dm/efi_media.c              | 24 ++++++++++++++++++++++++
  9 files changed, 93 insertions(+), 1 deletion(-)

create mode 100644 drivers/block/efi-media-uclass.c create mode 100644 drivers/block/sb_efi_media.c create mode 100644 test/dm/efi_media.c

Regards,
Simon

#41 lore raw

PATCH 13/35 efi: Add a media/block driver for EFI block devices

In reply to Simon Glass

Add a block driver which handles read/write for EFI block devices. This driver actually already exists ('efi_block') but is not really suitable for use as a real U-Boot driver:

- The operations do not provide a udevice
- The code is designed for running as part of EFI loader, so uses

    EFI_PRINT() and EFI_CALL().

- It creates block devices for all the partitions too, which is not somthing we want to support in this way
- The bind method probes the device, which is not permitted
- It uses 'EFI' as its parent device

The new driver is more 'normal', just requiring its platform data be set up in advance.

---

 drivers/block/Kconfig   |  10 ++++
 drivers/block/Makefile  |   1 +
 drivers/block/efi_blk.c | 115 ++++++++++++++++++++++++++++++++++++++++
 include/efi.h           |  11 ++++
 4 files changed, 137 insertions(+)
 create mode 100644 drivers/block/efi_blk.c
drivers/block/Kconfig +10 -0
index 058956ee8ee..ab450a52e9e 100644
@@ -84,6 +84,16 @@ config EFI_MEDIA_SANDBOX
 	  EFI_MEDIA uclass. It does not do anything useful, since sandbox does
 	  not actually support running on top of UEFI.
 
+config EFI_MEDIA_BLK
+	bool "EFI media block driver"
+	depends on EFI_APP
+	default y
+	help
+	  Enables a block driver for providing access to UEFI devices. This
+	  allows use of block devices detected by the underlying UEFI
+	  implementation. With this it is possible to use filesystems on these
+	  devices, for example.
+
 endif  # EFI_MEDIA
 
 config IDE
drivers/block/Makefile +1 -0
index 3778633da1d..b221a7c6eea 100644
@@ -17,3 +17,4 @@ obj-$(CONFIG_$(SPL_TPL_)BLOCK_CACHE) += blkcache.o
 
 obj-$(CONFIG_EFI_MEDIA) += efi-media-uclass.o
 obj-$(CONFIG_EFI_MEDIA_SANDBOX) += sb_efi_media.o
+obj-$(CONFIG_EFI_MEDIA_BLK) += efi_blk.o
drivers/block/efi_blk.c new file +115 -0

115 lines not on this page · raw message

include/efi.h +11 -0
index b5835422b95..0ec5913ddd1 100644
@@ -414,6 +414,17 @@ struct efi_priv {
 	void *next_hdr;
 };
 
+/*
+ * EFI attributes of the udevice handled by efi_media driver
+ *
+ * @handle: handle of the controller on which this driver is installed
+ * @blkio: block io protocol proxied by this driver
+ */
+struct efi_media_plat {
+	efi_handle_t		handle;
+	struct efi_block_io	*blkio;
+};
+
 /* Base address of the EFI image */
 extern char image_base[];
 
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#42 lore raw

Re: [PATCH 13/35] efi: Add a media/block driver for EFI block devices

Quoted 5 lines · Simon Glass

Add a block driver which handles read/write for EFI block devices. This driver actually already exists ('efi_block') but is not really suitable for use as a real U-Boot driver:

- The operations do not provide a udevice

efi_bl_bind() creates a udevice by calling blk_create_device() when an EFI application calls ConnectController() for a handle with an EFI_BLOCK_IO_PROTOCOL.

Please, explain in some detail what you think is wrong with the existing code.

Quoted 11 lines

- The code is designed for running as part of EFI loader, so uses

     EFI_PRINT() and EFI_CALL().

- It creates block devices for all the partitions too, which is not somthing we want to support in this way
- The bind method probes the device, which is not permitted
- It uses 'EFI' as its parent device

The new driver is more 'normal', just requiring its platform data be set up in advance.

Please, separate this series in two. One for U-Boot on EFI and one for U-Boot's UEFI implementation.

Best regardss

Heinrich

Quoted 183 lines

---

  drivers/block/Kconfig   |  10 ++++
  drivers/block/Makefile  |   1 +
  drivers/block/efi_blk.c | 115 ++++++++++++++++++++++++++++++++++++++++
  include/efi.h           |  11 ++++
  4 files changed, 137 insertions(+)

create mode 100644 drivers/block/efi_blk.c

drivers/block/Kconfig +10 -0
index 058956ee8ee..ab450a52e9e 100644
@@ -84,6 +84,16 @@ config EFI_MEDIA_SANDBOX
  	  EFI_MEDIA uclass. It does not do anything useful, since sandbox does
  	  not actually support running on top of UEFI.
 
+config EFI_MEDIA_BLK
+	bool "EFI media block driver"
+	depends on EFI_APP
+	default y
+	help
+	  Enables a block driver for providing access to UEFI devices. This
+	  allows use of block devices detected by the underlying UEFI
+	  implementation. With this it is possible to use filesystems on these
+	  devices, for example.
+
  endif  # EFI_MEDIA
 
  config IDE
drivers/block/Makefile +1 -0
index 3778633da1d..b221a7c6eea 100644
@@ -17,3 +17,4 @@ obj-$(CONFIG_$(SPL_TPL_)BLOCK_CACHE) += blkcache.o
 
  obj-$(CONFIG_EFI_MEDIA) += efi-media-uclass.o
  obj-$(CONFIG_EFI_MEDIA_SANDBOX) += sb_efi_media.o
+obj-$(CONFIG_EFI_MEDIA_BLK) += efi_blk.o
drivers/block/efi_blk.c new file +115 -0
new file mode 100644
index 00000000000..c00b0cc0b1c
@@ -0,0 +1,115 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Block driver for EFI devices
+ * This supports a media driver of UCLASS_EFI with a child UCLASS_BLK
+ * It allows block-level access to EFI devices made available via EFI boot
+ * services
+ *
+ * Copyright 2021 Google LLC
+ */
+
+#include <common.h>
+#include <blk.h>
+#include <dm.h>
+#include <efi.h>
+#include <efi_api.h>
+
+struct efi_block_plat {
+	struct efi_block_io *blkio;
+};
+
+/**
+ * Read from block device
+ *
+ * @dev:	device
+ * @blknr:	first block to be read
+ * @blkcnt:	number of blocks to read
+ * @buffer:	output buffer
+ * Return:	number of blocks transferred
+ */
+static ulong efi_bl_read(struct udevice *dev, lbaint_t blknr, lbaint_t blkcnt,
+			 void *buffer)
+{
+	struct efi_block_plat *plat = dev_get_plat(dev);
+	struct efi_block_io *io = plat->blkio;
+	efi_status_t ret;
+
+	log_debug("read buf=%p, block=%lx, count=%lx: ", buffer, (ulong)blknr,
+		  (ulong)blkcnt);
+	ret = io->read_blocks(io, io->media->media_id, blknr,
+			      blkcnt * io->media->block_size, buffer);
+	log_debug("ret=%lx (dec %ld)\n", ret & ~EFI_ERROR_MASK,
+		  ret & ~EFI_ERROR_MASK);
+	if (ret)
+		return 0;
+
+	return blkcnt;
+}
+
+/**
+ * Write to block device
+ *
+ * @dev:	device
+ * @blknr:	first block to be write
+ * @blkcnt:	number of blocks to write
+ * @buffer:	input buffer
+ * Return:	number of blocks transferred
+ */
+static ulong efi_bl_write(struct udevice *dev, lbaint_t blknr, lbaint_t blkcnt,
+			  const void *buffer)
+{
+	struct efi_block_plat *plat = dev_get_plat(dev);
+	struct efi_block_io *io = plat->blkio;
+	efi_status_t ret;
+
+	log_debug("write buf=%p, block=%lx, count=%lx: ", buffer, (ulong)blknr,
+		  (ulong)blkcnt);
+	ret = io->write_blocks(io, io->media->media_id, blknr,
+			       blkcnt * io->media->block_size, (void *)buffer);
+	log_debug("ret=%lx (dec %ld)\n", ret & ~EFI_ERROR_MASK,
+		  ret & ~EFI_ERROR_MASK);
+	if (ret)
+		return 0;
+
+	return blkcnt;
+}
+
+/* Block device driver operators */
+static const struct blk_ops efi_blk_ops = {
+	.read	= efi_bl_read,
+	.write	= efi_bl_write,
+};
+
+U_BOOT_DRIVER(efi_block) = {
+	.name		= "efi_block",
+	.id		= UCLASS_BLK,
+	.ops		= &efi_blk_ops,
+	.plat_auto	= sizeof(struct efi_block_plat),
+};
+
+static int efi_media_bind(struct udevice *dev)
+{
+	struct efi_media_plat *plat = dev_get_plat(dev);
+	struct efi_block_plat *blk_plat;
+	struct udevice *blk;
+	int ret;
+
+	ret = blk_create_devicef(dev, "efi_block", "blk", IF_TYPE_EFI,
+				 dev_seq(dev), plat->blkio->media->block_size,
+				 plat->blkio->media->last_block, &blk);
+	if (ret) {
+		debug("Cannot create block device\n");
+		return ret;
+	}
+	blk_plat = dev_get_plat(blk);
+	blk_plat->blkio = plat->blkio;
+
+	return 0;
+}
+
+U_BOOT_DRIVER(efi_media) = {
+	.name		= "efi_media",
+	.id		= UCLASS_EFI_MEDIA,
+	.bind		= efi_media_bind,
+	.plat_auto	= sizeof(struct efi_media_plat),
+};
include/efi.h +11 -0
index b5835422b95..0ec5913ddd1 100644
@@ -414,6 +414,17 @@ struct efi_priv {
  	void *next_hdr;
  };
 
+/*
+ * EFI attributes of the udevice handled by efi_media driver
+ *
+ * @handle: handle of the controller on which this driver is installed
+ * @blkio: block io protocol proxied by this driver
+ */
+struct efi_media_plat {
+	efi_handle_t		handle;
+	struct efi_block_io	*blkio;
+};
+
  /* Base address of the EFI image */
  extern char image_base[];
 
#43 lore raw

Re: [PATCH 13/35] efi: Add a media/block driver for EFI block devices

Hi Heinrich,

Quoted 8 lines · Heinrich Schuchardt
Quoted 4 lines · Simon Glass

Add a block driver which handles read/write for EFI block devices. This driver actually already exists ('efi_block') but is not really suitable for use as a real U-Boot driver:

Quoted 8 lines
Quoted 1 line

- The operations do not provide a udevice

efi_bl_bind() creates a udevice by calling blk_create_device() when an EFI application calls ConnectController() for a handle with an EFI_BLOCK_IO_PROTOCOL.

Please, explain in some detail what you think is wrong with the existing code.

See below:

Quoted 8 lines
Quoted 8 lines

- The code is designed for running as part of EFI loader, so uses

     EFI_PRINT() and EFI_CALL().

- It creates block devices for all the partitions too, which is not somthing we want to support in this way
- The bind method probes the device, which is not permitted
- It uses 'EFI' as its parent device

Quoted 7 lines
Quoted 4 lines

The new driver is more 'normal', just requiring its platform data be set up in advance.

Please, separate this series in two. One for U-Boot on EFI and one for U-Boot's UEFI implementation.

Again I'm not sure what you mean here. Please point to something you don't want in this series, which is focussed on the app.

Quoted 14 lines

Best regardss

Heinrich

Quoted 9 lines

---

  drivers/block/Kconfig   |  10 ++++
  drivers/block/Makefile  |   1 +
  drivers/block/efi_blk.c | 115 ++++++++++++++++++++++++++++++++++++++++
  include/efi.h           |  11 ++++
  4 files changed, 137 insertions(+)

create mode 100644 drivers/block/efi_blk.c

Regards,
Simon

#44 lore raw

Re: [PATCH 13/35] efi: Add a media/block driver for EFI block devices

Quoted 13 lines · Simon Glass

Hi Heinrich,

Quoted 10 lines · Heinrich Schuchardt
Quoted 5 lines · Simon Glass

Add a block driver which handles read/write for EFI block devices. This driver actually already exists ('efi_block') but is not really suitable for use as a real U-Boot driver:

- The operations do not provide a udevice

Quoted 14 lines
Quoted 6 lines

efi_bl_bind() creates a udevice by calling blk_create_device() when an EFI application calls ConnectController() for a handle with an EFI_BLOCK_IO_PROTOCOL.

Please, explain in some detail what you think is wrong with the existing code.

See below:

Quoted 5 lines
Quoted 5 lines

- The code is designed for running as part of EFI loader, so uses

      EFI_PRINT() and EFI_CALL().

- It creates block devices for all the partitions too, which is not somthing we want to support in this way

No, it creates a block device for the handle that was passed to ConnectController(). The handles for the partitions are created by U-Boot and these are not backed by separate block devices.

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

- The bind method probes the device, which is not permitted
- It uses 'EFI' as its parent device

All devices that the UEFI sub-system uses must be probed.

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

The new driver is more 'normal', just requiring its platform data be set up in advance.

No clue what you mean by "in advance". It is a software like iPXE that produces a handle at runtime and then calls ConnectController(). When that call is done we must create a block device which is in status probed, create handles for the partitions and install protocol interfaces on these partitions.

Best regards

Heinrich

Quoted 27 lines
Quoted 5 lines
Quoted 2 lines

Please, separate this series in two. One for U-Boot on EFI and one for U-Boot's UEFI implementation.

Again I'm not sure what you mean here. Please point to something you don't want in this series, which is focussed on the app.

Quoted 14 lines

Best regardss

Heinrich

Quoted 9 lines

---

   drivers/block/Kconfig   |  10 ++++
   drivers/block/Makefile  |   1 +
   drivers/block/efi_blk.c | 115 ++++++++++++++++++++++++++++++++++++++++
   include/efi.h           |  11 ++++
   4 files changed, 137 insertions(+)

create mode 100644 drivers/block/efi_blk.c

Regards,
Simon

#45 lore raw

Re: [PATCH 13/35] efi: Add a media/block driver for EFI block devices

Hi Heinrich,

Quoted 27 lines · Heinrich Schuchardt
Quoted 23 lines · Simon Glass

Hi Heinrich,

Quoted 8 lines · Heinrich Schuchardt
Quoted 4 lines · Simon Glass

Add a block driver which handles read/write for EFI block devices. This driver actually already exists ('efi_block') but is not really suitable for use as a real U-Boot driver:

Quoted 8 lines
Quoted 1 line

- The operations do not provide a udevice

efi_bl_bind() creates a udevice by calling blk_create_device() when an EFI application calls ConnectController() for a handle with an EFI_BLOCK_IO_PROTOCOL.

Please, explain in some detail what you think is wrong with the existing code.

See below:

Quoted 1 line
Quoted 8 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

- The code is designed for running as part of EFI loader, so uses

      EFI_PRINT() and EFI_CALL().

- It creates block devices for all the partitions too, which is not somthing we want to support in this way

No, it creates a block device for the handle that was passed to ConnectController(). The handles for the partitions are created by U-Boot and these are not backed by separate block devices.

OK, that is hard to discover so thank you for explaining it. The other items stand, in any case.

Quoted 9 lines

Best regards

Heinrich

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

- The bind method probes the device, which is not permitted
- It uses 'EFI' as its parent device

All devices that the UEFI sub-system uses must be probed.,

It is fine to probe a device but this must not be done in the bind() method. Hopefully the reason for this is obvious?

Quoted 10 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

The new driver is more 'normal', just requiring its platform data be set up in advance.

No clue what you mean by "in advance". It is a software like iPXE that produces a handle at runtime and then calls ConnectController(). When that call is done we must create a block device which is in status probed, create handles for the partitions and install protocol interfaces on these partitions.

In advance of using the device. Basically we can pass the platform data into the device_bind() call, so it has it right from the start.

Please understand that, while the existing EFI uclass is wonky, for the reasons I have clearly explained in this patch, my intent in explaining this is to motivate adding a new uclass. I am not trying to change the existing uclass. It is just one of many wonky things in the original EFI implementation that predates your efforts.

- Simon

Quoted 32 lines

Best regards

Heinrich

Quoted 27 lines
Quoted 5 lines
Quoted 2 lines

Please, separate this series in two. One for U-Boot on EFI and one for U-Boot's UEFI implementation.

Again I'm not sure what you mean here. Please point to something you don't want in this series, which is focussed on the app.

Quoted 14 lines

Best regardss

Heinrich

Quoted 9 lines

---

   drivers/block/Kconfig   |  10 ++++
   drivers/block/Makefile  |   1 +
   drivers/block/efi_blk.c | 115 ++++++++++++++++++++++++++++++++++++++++
   include/efi.h           |  11 ++++
   4 files changed, 137 insertions(+)

create mode 100644 drivers/block/efi_blk.c

Regards,
Simon

#46 lore raw

PATCH 14/35 efi: Locate all block devices in the app

In reply to Simon Glass

When starting the app, locate all block devices and make them available to U-Boot. This allows listing partitions and accessing files in filesystems.

EFI also has the concept of 'disks', meaning boot media. For now, this is not obviously useful in U-Boot, but add code to at least locate these. This can be expanded later as needed.

---

 include/efi.h     |  15 ++++++
 include/efi_api.h |  15 ++++++
 lib/efi/efi_app.c | 119 ++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 149 insertions(+)
include/efi.h +15 -0
index 0ec5913ddd1..c0fddf7f6cd 100644
@@ -529,4 +529,19 @@ void efi_putc(struct efi_priv *priv, const char ch);
  */
 int efi_info_get(enum efi_entry_t type, void **datap, int *sizep);
 
+/**
+ * efi_bind_block() - bind a new block device to an EFI device
+ *
+ * Binds a new top-level EFI_MEDIA device as well as a child block device so
+ * that the block device can be accessed in U-Boot.
+ *
+ * The device can then be accessed using 'part list efi 0', 'fat ls efi 0:1',
+ * for example, just like any other interface type.
+ *
+ * @handle: handle of the controller on which this driver is installed
+ * @blkio: block io protocol proxied by this driver
+ * @return 0 if OK, -ve on error
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio);
+
 #endif /* _LINUX_EFI_H */
include/efi_api.h +15 -0
index c8f959bb720..0e88b3e5dbe 100644
@@ -1994,4 +1994,19 @@ struct efi_firmware_management_protocol {
 			const u16 *package_version_name);
 };
 
+#define EFI_DISK_IO_PROTOCOL_GUID	\
+	EFI_GUID(0xce345171, 0xba0b, 0x11d2, 0x8e, 0x4f, \
+		 0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b)
+
+struct efi_disk {
+	u64 revision;
+	efi_status_t (EFIAPI *read_disk)(struct efi_disk *this, u32 media_id,
+					 u64 offset, efi_uintn_t buffer_size,
+					 void *buffer);
+
+	efi_status_t (EFIAPI *write_disk)(struct efi_disk *this, u32 media_id,
+					  u64 offset, efi_uintn_t buffer_size,
+					  void *buffer);
+};
+
 #endif
lib/efi/efi_app.c +119 -0

137 lines not on this page · raw message

Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#47 lore raw

Re: [PATCH 14/35] efi: Locate all block devices in the app

Quoted 7 lines · Simon Glass

When starting the app, locate all block devices and make them available to U-Boot. This allows listing partitions and accessing files in filesystems.

EFI also has the concept of 'disks', meaning boot media. For now, this is not obviously useful in U-Boot, but add code to at least locate these. This can be expanded later as needed.

UEFI firmware handles with the EFI_BLOCK_IO_PROTOCOL for raw access to disks and partitions. It further provides the EFI_SIMPLE_FILE_SYSTEM_PROTOCOL on these handles to access files on formatted media.

Quoted 136 lines

---

  include/efi.h     |  15 ++++++
  include/efi_api.h |  15 ++++++
  lib/efi/efi_app.c | 119 ++++++++++++++++++++++++++++++++++++++++++++++
  3 files changed, 149 insertions(+)
include/efi.h +15 -0
index 0ec5913ddd1..c0fddf7f6cd 100644
@@ -529,4 +529,19 @@ void efi_putc(struct efi_priv *priv, const char ch);
   */
  int efi_info_get(enum efi_entry_t type, void **datap, int *sizep);
 
+/**
+ * efi_bind_block() - bind a new block device to an EFI device
+ *
+ * Binds a new top-level EFI_MEDIA device as well as a child block device so
+ * that the block device can be accessed in U-Boot.
+ *
+ * The device can then be accessed using 'part list efi 0', 'fat ls efi 0:1',
+ * for example, just like any other interface type.
+ *
+ * @handle: handle of the controller on which this driver is installed
+ * @blkio: block io protocol proxied by this driver
+ * @return 0 if OK, -ve on error
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio);
+
  #endif /* _LINUX_EFI_H */
include/efi_api.h +15 -0
index c8f959bb720..0e88b3e5dbe 100644
@@ -1994,4 +1994,19 @@ struct efi_firmware_management_protocol {
  			const u16 *package_version_name);
  };
 
+#define EFI_DISK_IO_PROTOCOL_GUID	\
+	EFI_GUID(0xce345171, 0xba0b, 0x11d2, 0x8e, 0x4f, \
+		 0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b)
+
+struct efi_disk {
+	u64 revision;
+	efi_status_t (EFIAPI *read_disk)(struct efi_disk *this, u32 media_id,
+					 u64 offset, efi_uintn_t buffer_size,
+					 void *buffer);
+
+	efi_status_t (EFIAPI *write_disk)(struct efi_disk *this, u32 media_id,
+					  u64 offset, efi_uintn_t buffer_size,
+					  void *buffer);
+};
+
  #endif
lib/efi/efi_app.c +57 -0
index f61665686c5..9ba48517422 100644
@@ -21,6 +21,9 @@
  #include <efi.h>
  #include <efi_api.h>
  #include <sysreset.h>
+#include <dm/device-internal.h>
+#include <dm/lists.h>
+#include <dm/root.h>
 
  DECLARE_GLOBAL_DATA_PTR;
 
@@ -46,6 +49,33 @@ int efi_info_get(enum efi_entry_t type, void **datap, int *sizep)
  	return -ENOSYS;
  }
 
+/**
+ * Create a block device so U-Boot can access an EFI device
+ *
+ * @handle:	EFI handle to bind
+ * @blkio:	block io protocol
+ * Return:	0 = success
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio)
+{
+	struct efi_media_plat plat;
+	struct udevice *dev;
+	char name[18];
+	int ret;
+
+	plat.handle = handle;
+	plat.blkio = blkio;
+	ret = device_bind(dm_root(), DM_DRIVER_GET(efi_media), "efi_media",
+			  &plat, ofnode_null(), &dev);
+	if (ret)
+		return log_msg_ret("bind", ret);
+
+	snprintf(name, sizeof(name), "efi_media_%x", dev_seq(dev));
+	device_set_name(dev, name);
+
+	return 0;
+}
+
  static efi_status_t setup_memory(struct efi_priv *priv)
  {
  	struct efi_boot_services *boot = priv->boot;
@@ -105,6 +135,95 @@ static void free_memory(struct efi_priv *priv)
  	global_data_ptr = NULL;
  }
 
+static int setup_disks(void)
+{
+	/* This is not fully implemented yet */
+	return 0;
+
+	efi_guid_t efi_disk_guid = EFI_DISK_IO_PROTOCOL_GUID;
+	struct efi_boot_services *boot = efi_get_boot();
+	struct efi_disk *disk;
+	int ret;
+
+	if (!boot)
+		return log_msg_ret("sys", -ENOSYS);
+	ret = boot->locate_protocol(&efi_disk_guid, NULL, (void **)&disk);
+	if (ret)
+		return log_msg_ret("prot", -ENOTSUPP);
+
+	return 0;
+}
+
+static int setup_block(void)
+{
+	efi_guid_t efi_blkio_guid = EFI_BLOCK_IO_PROTOCOL_GUID;
+	efi_guid_t efi_devpath_guid = EFI_DEVICE_PATH_PROTOCOL_GUID;
+	struct efi_boot_services *boot = efi_get_boot();
+	struct efi_block_io *blkio;
+	struct efi_device_path device_path;
+	efi_handle_t handle[100];

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 174

Quoted 10 lines
lib/efi/efi_app.c +10 -0
+	efi_uintn_t buf_size;
+	int num_handles;
+	int ret, i;
+
+	if (!boot)
+		return log_msg_ret("sys", -ENOSYS);
+
+	buf_size = sizeof(handle);
+	ret = boot->locate_handle(BY_PROTOCOL, &efi_blkio_guid, NULL,
+				  &buf_size, handle);

You could use LocateHandleBuffer() here which will allocate enough memory for all matching handles.

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 183

Quoted 9 lines
lib/efi/efi_app.c +9 -0
+	if (ret)
+		return log_msg_ret("loc", -ENOTSUPP);
+
+	num_handles = buf_size / sizeof(efi_handle_t);
+	log_info("Found %d EFI handles\n", num_handles);
+
+	for (i = 0; i < num_handles; i++) {
+		ret = boot->handle_protocol(handle[i], &efi_devpath_guid,
+					    (void **)&device_path);

Why do you read the devicepath if you don't use it?

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 187

Quoted 4 lines
lib/efi/efi_app.c +4 -0
+		if (ret) {
+			log_warning("- devpath %d failed (ret=%d)\n", i, ret);
+			continue;
+		}

Here some analysis of devicepaths and installed protocols is missing to find out which of the handles represents a block device and which represents a partition:

If the last devicepath node is type 4 , Media Device Path with SubType 1 Hard Drive and the partition number is non-zero it is a partition.

If the devicepath without the last node relates to a handle with the EFI_BLOCK_IO_PROTOCOL, this also indicates that the current handle is for a partition.

Best regards

Heinrich

Quoted 43 lines
lib/efi/efi_app.c +39 -0
+
+		ret = boot->handle_protocol(handle[i], &efi_blkio_guid,
+					    (void **)&blkio);
+		if (ret) {
+			log_warning("- blkio %d failed (ret=%d)\n", i, ret);
+			continue;
+		}
+
+		ret = efi_bind_block(handle[i], blkio);
+		if (ret) {
+			log_warning("- blkio bind %d failed (ret=%d)\n", i, ret);
+			continue;
+		}
+	}
+	if (ret)
+		return log_msg_ret("prot", -ENOTSUPP);
+
+	return 0;
+}
+
+int dm_scan_other(bool pre_reloc_only)
+{
+	if (gd->flags & GD_FLG_RELOC) {
+		int ret;
+
+		/* Find all block devices and setup EFI devices for them */
+		ret = setup_block();
+		if (ret)
+			return ret;
+
+		/* Not needed at present, but could be useful one day? */
+		ret = setup_disks();
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
  /**
   * efi_main() - Start an EFI image
   *
#48 lore raw

Re: [PATCH 14/35] efi: Locate all block devices in the app

Hi Simon,

Quoted 5 lines · Heinrich Schuchardt
Quoted 4 lines · Simon Glass

When starting the app, locate all block devices and make them available to U-Boot. This allows listing partitions and accessing files in filesystems.

Quoted 8 lines
Quoted 3 lines

EFI also has the concept of 'disks', meaning boot media. For now, this is not obviously useful in U-Boot, but add code to at least locate these. This can be expanded later as needed.

UEFI firmware handles with the EFI_BLOCK_IO_PROTOCOL for raw access to disks and partitions. It further provides the EFI_SIMPLE_FILE_SYSTEM_PROTOCOL on these handles to access files on formatted media.

Do you want to implement "efifs" as a U-Boot's file system on top of SIMPLE_FILE_SYSTEM_PROTO?
Just kidding.

I have one concern:

Quoted 183 lines
Quoted 136 lines

---

  include/efi.h     |  15 ++++++
  include/efi_api.h |  15 ++++++
  lib/efi/efi_app.c | 119 ++++++++++++++++++++++++++++++++++++++++++++++
  3 files changed, 149 insertions(+)
include/efi.h +15 -0
index 0ec5913ddd1..c0fddf7f6cd 100644
@@ -529,4 +529,19 @@ void efi_putc(struct efi_priv *priv, const char ch);
   */
  int efi_info_get(enum efi_entry_t type, void **datap, int *sizep);
 
+/**
+ * efi_bind_block() - bind a new block device to an EFI device
+ *
+ * Binds a new top-level EFI_MEDIA device as well as a child block device so
+ * that the block device can be accessed in U-Boot.
+ *
+ * The device can then be accessed using 'part list efi 0', 'fat ls efi 0:1',
+ * for example, just like any other interface type.
+ *
+ * @handle: handle of the controller on which this driver is installed
+ * @blkio: block io protocol proxied by this driver
+ * @return 0 if OK, -ve on error
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio);
+
  #endif /* _LINUX_EFI_H */
include/efi_api.h +15 -0
index c8f959bb720..0e88b3e5dbe 100644
@@ -1994,4 +1994,19 @@ struct efi_firmware_management_protocol {
  			const u16 *package_version_name);
  };
 
+#define EFI_DISK_IO_PROTOCOL_GUID	\
+	EFI_GUID(0xce345171, 0xba0b, 0x11d2, 0x8e, 0x4f, \
+		 0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b)
+
+struct efi_disk {
+	u64 revision;
+	efi_status_t (EFIAPI *read_disk)(struct efi_disk *this, u32 media_id,
+					 u64 offset, efi_uintn_t buffer_size,
+					 void *buffer);
+
+	efi_status_t (EFIAPI *write_disk)(struct efi_disk *this, u32 media_id,
+					  u64 offset, efi_uintn_t buffer_size,
+					  void *buffer);
+};
+
  #endif
lib/efi/efi_app.c +57 -0
index f61665686c5..9ba48517422 100644
@@ -21,6 +21,9 @@
  #include <efi.h>
  #include <efi_api.h>
  #include <sysreset.h>
+#include <dm/device-internal.h>
+#include <dm/lists.h>
+#include <dm/root.h>
 
  DECLARE_GLOBAL_DATA_PTR;
 
@@ -46,6 +49,33 @@ int efi_info_get(enum efi_entry_t type, void **datap, int *sizep)
  	return -ENOSYS;
  }
 
+/**
+ * Create a block device so U-Boot can access an EFI device
+ *
+ * @handle:	EFI handle to bind
+ * @blkio:	block io protocol
+ * Return:	0 = success
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio)
+{
+	struct efi_media_plat plat;
+	struct udevice *dev;
+	char name[18];
+	int ret;
+
+	plat.handle = handle;
+	plat.blkio = blkio;
+	ret = device_bind(dm_root(), DM_DRIVER_GET(efi_media), "efi_media",
+			  &plat, ofnode_null(), &dev);
+	if (ret)
+		return log_msg_ret("bind", ret);
+
+	snprintf(name, sizeof(name), "efi_media_%x", dev_seq(dev));
+	device_set_name(dev, name);
+
+	return 0;
+}
+
  static efi_status_t setup_memory(struct efi_priv *priv)
  {
  	struct efi_boot_services *boot = priv->boot;
@@ -105,6 +135,95 @@ static void free_memory(struct efi_priv *priv)
  	global_data_ptr = NULL;
  }
 
+static int setup_disks(void)
+{
+	/* This is not fully implemented yet */
+	return 0;
+
+	efi_guid_t efi_disk_guid = EFI_DISK_IO_PROTOCOL_GUID;
+	struct efi_boot_services *boot = efi_get_boot();
+	struct efi_disk *disk;
+	int ret;
+
+	if (!boot)
+		return log_msg_ret("sys", -ENOSYS);
+	ret = boot->locate_protocol(&efi_disk_guid, NULL, (void **)&disk);
+	if (ret)
+		return log_msg_ret("prot", -ENOTSUPP);
+
+	return 0;
+}
+
+static int setup_block(void)
+{
+	efi_guid_t efi_blkio_guid = EFI_BLOCK_IO_PROTOCOL_GUID;
+	efi_guid_t efi_devpath_guid = EFI_DEVICE_PATH_PROTOCOL_GUID;
+	struct efi_boot_services *boot = efi_get_boot();
+	struct efi_block_io *blkio;
+	struct efi_device_path device_path;
+	efi_handle_t handle[100];

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 174

Quoted 10 lines
lib/efi/efi_app.c +10 -0
+	efi_uintn_t buf_size;
+	int num_handles;
+	int ret, i;
+
+	if (!boot)
+		return log_msg_ret("sys", -ENOSYS);
+
+	buf_size = sizeof(handle);
+	ret = boot->locate_handle(BY_PROTOCOL, &efi_blkio_guid, NULL,
+				  &buf_size, handle);

You could use LocateHandleBuffer() here which will allocate enough memory for all matching handles.

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 183

Quoted 9 lines
lib/efi/efi_app.c +9 -0
+	if (ret)
+		return log_msg_ret("loc", -ENOTSUPP);
+
+	num_handles = buf_size / sizeof(efi_handle_t);
+	log_info("Found %d EFI handles\n", num_handles);
+
+	for (i = 0; i < num_handles; i++) {
+		ret = boot->handle_protocol(handle[i], &efi_devpath_guid,
+					    (void **)&device_path);

Why do you read the devicepath if you don't use it?

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 187

Quoted 4 lines
lib/efi/efi_app.c +4 -0
+		if (ret) {
+			log_warning("- devpath %d failed (ret=%d)\n", i, ret);
+			continue;
+		}

Here some analysis of devicepaths and installed protocols is missing to find out which of the handles represents a block device and which represents a partition:

If the last devicepath node is type 4 , Media Device Path with SubType 1 Hard Drive and the partition number is non-zero it is a partition.

If the devicepath without the last node relates to a handle with the EFI_BLOCK_IO_PROTOCOL, this also indicates that the current handle is for a partition.

Best regards

Heinrich

Quoted 9 lines

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 196

Quoted 9 lines
lib/efi/efi_app.c +9 -0
+
+		ret = boot->handle_protocol(handle[i], &efi_blkio_guid,
+					    (void **)&blkio);
+		if (ret) {
+			log_warning("- blkio %d failed (ret=%d)\n", i, ret);
+			continue;
+		}
+
+		ret = efi_bind_block(handle[i], blkio);

Here you are trying to create a U-Boot block device for every UEFI block device (i.e.BLOCK_IO_PROTOCOL interface), but please remember that U-Boot UEFI has created UEFI block devices for all the existing U-Boot block devices at the initialization time.

So any physical disk may end up having one (original) U-Boot block device and another U-Boot block device rooted in UEFI object (and yet another U-Boot block device rooted in the second one and so on?).

-Takahiro Akashi

Quoted 34 lines
Quoted 34 lines
lib/efi/efi_app.c +30 -0
+		if (ret) {
+			log_warning("- blkio bind %d failed (ret=%d)\n", i, ret);
+			continue;
+		}
+	}
+	if (ret)
+		return log_msg_ret("prot", -ENOTSUPP);
+
+	return 0;
+}
+
+int dm_scan_other(bool pre_reloc_only)
+{
+	if (gd->flags & GD_FLG_RELOC) {
+		int ret;
+
+		/* Find all block devices and setup EFI devices for them */
+		ret = setup_block();
+		if (ret)
+			return ret;
+
+		/* Not needed at present, but could be useful one day? */
+		ret = setup_disks();
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
  /**
   * efi_main() - Start an EFI image
   *
#49 lore raw

Re: [PATCH 14/35] efi: Locate all block devices in the app

Hi Takahiro,

Quoted 13 lines · AKASHI Takahiro

Hi Simon,

Quoted 9 lines · Heinrich Schuchardt
Quoted 7 lines · Simon Glass

When starting the app, locate all block devices and make them available to U-Boot. This allows listing partitions and accessing files in filesystems.

EFI also has the concept of 'disks', meaning boot media. For now, this is not obviously useful in U-Boot, but add code to at least locate these. This can be expanded later as needed.

Quoted 8 lines
Quoted 4 lines

UEFI firmware handles with the EFI_BLOCK_IO_PROTOCOL for raw access to disks and partitions. It further provides the EFI_SIMPLE_FILE_SYSTEM_PROTOCOL on these handles to access files on formatted media.

Do you want to implement "efifs" as a U-Boot's file system on top of SIMPLE_FILE_SYSTEM_PROTO?
Just kidding.

Eek.

Quoted 186 lines

I have one concern:

Quoted 183 lines
Quoted 136 lines

---

  include/efi.h     |  15 ++++++
  include/efi_api.h |  15 ++++++
  lib/efi/efi_app.c | 119 ++++++++++++++++++++++++++++++++++++++++++++++
  3 files changed, 149 insertions(+)
include/efi.h +15 -0
index 0ec5913ddd1..c0fddf7f6cd 100644
@@ -529,4 +529,19 @@ void efi_putc(struct efi_priv *priv, const char ch);
   */
  int efi_info_get(enum efi_entry_t type, void **datap, int *sizep);
 
+/**
+ * efi_bind_block() - bind a new block device to an EFI device
+ *
+ * Binds a new top-level EFI_MEDIA device as well as a child block device so
+ * that the block device can be accessed in U-Boot.
+ *
+ * The device can then be accessed using 'part list efi 0', 'fat ls efi 0:1',
+ * for example, just like any other interface type.
+ *
+ * @handle: handle of the controller on which this driver is installed
+ * @blkio: block io protocol proxied by this driver
+ * @return 0 if OK, -ve on error
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio);
+
  #endif /* _LINUX_EFI_H */
include/efi_api.h +15 -0
index c8f959bb720..0e88b3e5dbe 100644
@@ -1994,4 +1994,19 @@ struct efi_firmware_management_protocol {
                    const u16 *package_version_name);
  };
 
+#define EFI_DISK_IO_PROTOCOL_GUID  \
+   EFI_GUID(0xce345171, 0xba0b, 0x11d2, 0x8e, 0x4f, \
+            0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b)
+
+struct efi_disk {
+   u64 revision;
+   efi_status_t (EFIAPI *read_disk)(struct efi_disk *this, u32 media_id,
+                                    u64 offset, efi_uintn_t buffer_size,
+                                    void *buffer);
+
+   efi_status_t (EFIAPI *write_disk)(struct efi_disk *this, u32 media_id,
+                                     u64 offset, efi_uintn_t buffer_size,
+                                     void *buffer);
+};
+
  #endif
lib/efi/efi_app.c +57 -0
index f61665686c5..9ba48517422 100644
@@ -21,6 +21,9 @@
  #include <efi.h>
  #include <efi_api.h>
  #include <sysreset.h>
+#include <dm/device-internal.h>
+#include <dm/lists.h>
+#include <dm/root.h>
 
  DECLARE_GLOBAL_DATA_PTR;
 
@@ -46,6 +49,33 @@ int efi_info_get(enum efi_entry_t type, void **datap, int *sizep)
    return -ENOSYS;
  }
 
+/**
+ * Create a block device so U-Boot can access an EFI device
+ *
+ * @handle:        EFI handle to bind
+ * @blkio: block io protocol
+ * Return: 0 = success
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio)
+{
+   struct efi_media_plat plat;
+   struct udevice *dev;
+   char name[18];
+   int ret;
+
+   plat.handle = handle;
+   plat.blkio = blkio;
+   ret = device_bind(dm_root(), DM_DRIVER_GET(efi_media), "efi_media",
+                     &plat, ofnode_null(), &dev);
+   if (ret)
+           return log_msg_ret("bind", ret);
+
+   snprintf(name, sizeof(name), "efi_media_%x", dev_seq(dev));
+   device_set_name(dev, name);
+
+   return 0;
+}
+
  static efi_status_t setup_memory(struct efi_priv *priv)
  {
    struct efi_boot_services *boot = priv->boot;
@@ -105,6 +135,95 @@ static void free_memory(struct efi_priv *priv)
    global_data_ptr = NULL;
  }
 
+static int setup_disks(void)
+{
+   /* This is not fully implemented yet */
+   return 0;
+
+   efi_guid_t efi_disk_guid = EFI_DISK_IO_PROTOCOL_GUID;
+   struct efi_boot_services *boot = efi_get_boot();
+   struct efi_disk *disk;
+   int ret;
+
+   if (!boot)
+           return log_msg_ret("sys", -ENOSYS);
+   ret = boot->locate_protocol(&efi_disk_guid, NULL, (void **)&disk);
+   if (ret)
+           return log_msg_ret("prot", -ENOTSUPP);
+
+   return 0;
+}
+
+static int setup_block(void)
+{
+   efi_guid_t efi_blkio_guid = EFI_BLOCK_IO_PROTOCOL_GUID;
+   efi_guid_t efi_devpath_guid = EFI_DEVICE_PATH_PROTOCOL_GUID;
+   struct efi_boot_services *boot = efi_get_boot();
+   struct efi_block_io *blkio;
+   struct efi_device_path device_path;
+   efi_handle_t handle[100];

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 174

Quoted 10 lines
lib/efi/efi_app.c +10 -0
+   efi_uintn_t buf_size;
+   int num_handles;
+   int ret, i;
+
+   if (!boot)
+           return log_msg_ret("sys", -ENOSYS);
+
+   buf_size = sizeof(handle);
+   ret = boot->locate_handle(BY_PROTOCOL, &efi_blkio_guid, NULL,
+                             &buf_size, handle);

You could use LocateHandleBuffer() here which will allocate enough memory for all matching handles.

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 183

Quoted 9 lines
lib/efi/efi_app.c +9 -0
+   if (ret)
+           return log_msg_ret("loc", -ENOTSUPP);
+
+   num_handles = buf_size / sizeof(efi_handle_t);
+   log_info("Found %d EFI handles\n", num_handles);
+
+   for (i = 0; i < num_handles; i++) {
+           ret = boot->handle_protocol(handle[i], &efi_devpath_guid,
+                                       (void **)&device_path);

Why do you read the devicepath if you don't use it?

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 187

Quoted 4 lines
lib/efi/efi_app.c +4 -0
+           if (ret) {
+                   log_warning("- devpath %d failed (ret=%d)\n", i, ret);
+                   continue;
+           }

Here some analysis of devicepaths and installed protocols is missing to find out which of the handles represents a block device and which represents a partition:

If the last devicepath node is type 4 , Media Device Path with SubType 1 Hard Drive and the partition number is non-zero it is a partition.

If the devicepath without the last node relates to a handle with the EFI_BLOCK_IO_PROTOCOL, this also indicates that the current handle is for a partition.

Best regards

Heinrich

Quoted 14 lines
Quoted 9 lines

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 196

Quoted 9 lines
lib/efi/efi_app.c +9 -0
+
+           ret = boot->handle_protocol(handle[i], &efi_blkio_guid,
+                                       (void **)&blkio);
+           if (ret) {
+                   log_warning("- blkio %d failed (ret=%d)\n", i, ret);
+                   continue;
+           }
+
+           ret = efi_bind_block(handle[i], blkio);

Here you are trying to create a U-Boot block device for every UEFI block device (i.e.BLOCK_IO_PROTOCOL interface), but please remember that U-Boot UEFI has created UEFI block devices for all the existing U-Boot block devices at the initialization time.

Do you mean with EFI_LOADER? This code is for U-Boot as an app, where EFI_LOADER is disabled, at least at present.

Quoted 4 lines

So any physical disk may end up having one (original) U-Boot block device and another U-Boot block device rooted in UEFI object (and yet another U-Boot block device rooted in the second one and so on?).

Hopefully that is not an issue, for now. We will need to figure it out later, if we enable EFI_LOADER.

Regards,
Simon

[..]

#50 lore raw

Re: [PATCH 14/35] efi: Locate all block devices in the app

Quoted 229 lines · Simon Glass

Hi Takahiro,

Quoted 13 lines · AKASHI Takahiro

Hi Simon,

Quoted 9 lines · Heinrich Schuchardt
Quoted 7 lines · Simon Glass

When starting the app, locate all block devices and make them available to U-Boot. This allows listing partitions and accessing files in filesystems.

EFI also has the concept of 'disks', meaning boot media. For now, this is not obviously useful in U-Boot, but add code to at least locate these. This can be expanded later as needed.

Quoted 8 lines
Quoted 4 lines

UEFI firmware handles with the EFI_BLOCK_IO_PROTOCOL for raw access to disks and partitions. It further provides the EFI_SIMPLE_FILE_SYSTEM_PROTOCOL on these handles to access files on formatted media.

Do you want to implement "efifs" as a U-Boot's file system on top of SIMPLE_FILE_SYSTEM_PROTO?
Just kidding.

Eek.

Quoted 186 lines

I have one concern:

Quoted 183 lines
Quoted 136 lines

---

  include/efi.h     |  15 ++++++
  include/efi_api.h |  15 ++++++
  lib/efi/efi_app.c | 119 ++++++++++++++++++++++++++++++++++++++++++++++
  3 files changed, 149 insertions(+)
include/efi.h +15 -0
index 0ec5913ddd1..c0fddf7f6cd 100644
@@ -529,4 +529,19 @@ void efi_putc(struct efi_priv *priv, const char ch);
   */
  int efi_info_get(enum efi_entry_t type, void **datap, int *sizep);
 
+/**
+ * efi_bind_block() - bind a new block device to an EFI device
+ *
+ * Binds a new top-level EFI_MEDIA device as well as a child block device so
+ * that the block device can be accessed in U-Boot.
+ *
+ * The device can then be accessed using 'part list efi 0', 'fat ls efi 0:1',
+ * for example, just like any other interface type.
+ *
+ * @handle: handle of the controller on which this driver is installed
+ * @blkio: block io protocol proxied by this driver
+ * @return 0 if OK, -ve on error
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio);
+
  #endif /* _LINUX_EFI_H */
include/efi_api.h +15 -0
index c8f959bb720..0e88b3e5dbe 100644
@@ -1994,4 +1994,19 @@ struct efi_firmware_management_protocol {
                    const u16 *package_version_name);
  };
 
+#define EFI_DISK_IO_PROTOCOL_GUID  \
+   EFI_GUID(0xce345171, 0xba0b, 0x11d2, 0x8e, 0x4f, \
+            0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b)
+
+struct efi_disk {
+   u64 revision;
+   efi_status_t (EFIAPI *read_disk)(struct efi_disk *this, u32 media_id,
+                                    u64 offset, efi_uintn_t buffer_size,
+                                    void *buffer);
+
+   efi_status_t (EFIAPI *write_disk)(struct efi_disk *this, u32 media_id,
+                                     u64 offset, efi_uintn_t buffer_size,
+                                     void *buffer);
+};
+
  #endif
lib/efi/efi_app.c +57 -0
index f61665686c5..9ba48517422 100644
@@ -21,6 +21,9 @@
  #include <efi.h>
  #include <efi_api.h>
  #include <sysreset.h>
+#include <dm/device-internal.h>
+#include <dm/lists.h>
+#include <dm/root.h>
 
  DECLARE_GLOBAL_DATA_PTR;
 
@@ -46,6 +49,33 @@ int efi_info_get(enum efi_entry_t type, void **datap, int *sizep)
    return -ENOSYS;
  }
 
+/**
+ * Create a block device so U-Boot can access an EFI device
+ *
+ * @handle:        EFI handle to bind
+ * @blkio: block io protocol
+ * Return: 0 = success
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio)
+{
+   struct efi_media_plat plat;
+   struct udevice *dev;
+   char name[18];
+   int ret;
+
+   plat.handle = handle;
+   plat.blkio = blkio;
+   ret = device_bind(dm_root(), DM_DRIVER_GET(efi_media), "efi_media",
+                     &plat, ofnode_null(), &dev);
+   if (ret)
+           return log_msg_ret("bind", ret);
+
+   snprintf(name, sizeof(name), "efi_media_%x", dev_seq(dev));
+   device_set_name(dev, name);
+
+   return 0;
+}
+
  static efi_status_t setup_memory(struct efi_priv *priv)
  {
    struct efi_boot_services *boot = priv->boot;
@@ -105,6 +135,95 @@ static void free_memory(struct efi_priv *priv)
    global_data_ptr = NULL;
  }
 
+static int setup_disks(void)
+{
+   /* This is not fully implemented yet */
+   return 0;
+
+   efi_guid_t efi_disk_guid = EFI_DISK_IO_PROTOCOL_GUID;
+   struct efi_boot_services *boot = efi_get_boot();
+   struct efi_disk *disk;
+   int ret;
+
+   if (!boot)
+           return log_msg_ret("sys", -ENOSYS);
+   ret = boot->locate_protocol(&efi_disk_guid, NULL, (void **)&disk);
+   if (ret)
+           return log_msg_ret("prot", -ENOTSUPP);
+
+   return 0;
+}
+
+static int setup_block(void)
+{
+   efi_guid_t efi_blkio_guid = EFI_BLOCK_IO_PROTOCOL_GUID;
+   efi_guid_t efi_devpath_guid = EFI_DEVICE_PATH_PROTOCOL_GUID;
+   struct efi_boot_services *boot = efi_get_boot();
+   struct efi_block_io *blkio;
+   struct efi_device_path device_path;
+   efi_handle_t handle[100];

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 174

Quoted 10 lines
lib/efi/efi_app.c +10 -0
+   efi_uintn_t buf_size;
+   int num_handles;
+   int ret, i;
+
+   if (!boot)
+           return log_msg_ret("sys", -ENOSYS);
+
+   buf_size = sizeof(handle);
+   ret = boot->locate_handle(BY_PROTOCOL, &efi_blkio_guid, NULL,
+                             &buf_size, handle);

You could use LocateHandleBuffer() here which will allocate enough memory for all matching handles.

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 183

Quoted 9 lines
lib/efi/efi_app.c +9 -0
+   if (ret)
+           return log_msg_ret("loc", -ENOTSUPP);
+
+   num_handles = buf_size / sizeof(efi_handle_t);
+   log_info("Found %d EFI handles\n", num_handles);
+
+   for (i = 0; i < num_handles; i++) {
+           ret = boot->handle_protocol(handle[i], &efi_devpath_guid,
+                                       (void **)&device_path);

Why do you read the devicepath if you don't use it?

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 187

Quoted 4 lines
lib/efi/efi_app.c +4 -0
+           if (ret) {
+                   log_warning("- devpath %d failed (ret=%d)\n", i, ret);
+                   continue;
+           }

Here some analysis of devicepaths and installed protocols is missing to find out which of the handles represents a block device and which represents a partition:

If the last devicepath node is type 4 , Media Device Path with SubType 1 Hard Drive and the partition number is non-zero it is a partition.

If the devicepath without the last node relates to a handle with the EFI_BLOCK_IO_PROTOCOL, this also indicates that the current handle is for a partition.

Best regards

Heinrich

Quoted 14 lines
Quoted 9 lines

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 196

Quoted 9 lines
lib/efi/efi_app.c +9 -0
+
+           ret = boot->handle_protocol(handle[i], &efi_blkio_guid,
+                                       (void **)&blkio);
+           if (ret) {
+                   log_warning("- blkio %d failed (ret=%d)\n", i, ret);
+                   continue;
+           }
+
+           ret = efi_bind_block(handle[i], blkio);

Here you are trying to create a U-Boot block device for every UEFI block device (i.e.BLOCK_IO_PROTOCOL interface), but please remember that U-Boot UEFI has created UEFI block devices for all the existing U-Boot block devices at the initialization time.

Quoted 2 lines

Do you mean with EFI_LOADER? This code is for U-Boot as an app, where EFI_LOADER is disabled, at least at present.

Ah ok, but in this patch series, you touched some files under lib/efi_loader and lib/efi_driver and I thought you have assumed the option as well. So your patch#11 are fully independent from the rest.

-Takahiro Akashi

Quoted 12 lines
Quoted 4 lines

So any physical disk may end up having one (original) U-Boot block device and another U-Boot block device rooted in UEFI object (and yet another U-Boot block device rooted in the second one and so on?).

Hopefully that is not an issue, for now. We will need to figure it out later, if we enable EFI_LOADER.

Regards,
Simon

[..]

#51 lore raw

Re: [PATCH 14/35] efi: Locate all block devices in the app

In reply to Heinrich Schuchardt

Hi Heinrich,

Quoted 6 lines · Heinrich Schuchardt
Quoted 4 lines · Simon Glass

When starting the app, locate all block devices and make them available to U-Boot. This allows listing partitions and accessing files in filesystems.

Quoted 8 lines
Quoted 3 lines

EFI also has the concept of 'disks', meaning boot media. For now, this is not obviously useful in U-Boot, but add code to at least locate these. This can be expanded later as needed.

UEFI firmware handles with the EFI_BLOCK_IO_PROTOCOL for raw access to disks and partitions. It further provides the EFI_SIMPLE_FILE_SYSTEM_PROTOCOL on these handles to access files on formatted media.

OK ta.

Quoted 142 lines
Quoted 142 lines

---

  include/efi.h     |  15 ++++++
  include/efi_api.h |  15 ++++++
  lib/efi/efi_app.c | 119 ++++++++++++++++++++++++++++++++++++++++++++++
  3 files changed, 149 insertions(+)
include/efi.h +15 -0
index 0ec5913ddd1..c0fddf7f6cd 100644
@@ -529,4 +529,19 @@ void efi_putc(struct efi_priv *priv, const char ch);
   */
  int efi_info_get(enum efi_entry_t type, void **datap, int *sizep);
 
+/**
+ * efi_bind_block() - bind a new block device to an EFI device
+ *
+ * Binds a new top-level EFI_MEDIA device as well as a child block device so
+ * that the block device can be accessed in U-Boot.
+ *
+ * The device can then be accessed using 'part list efi 0', 'fat ls efi 0:1',
+ * for example, just like any other interface type.
+ *
+ * @handle: handle of the controller on which this driver is installed
+ * @blkio: block io protocol proxied by this driver
+ * @return 0 if OK, -ve on error
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio);
+
  #endif /* _LINUX_EFI_H */
include/efi_api.h +15 -0
index c8f959bb720..0e88b3e5dbe 100644
@@ -1994,4 +1994,19 @@ struct efi_firmware_management_protocol {
                      const u16 *package_version_name);
  };
 
+#define EFI_DISK_IO_PROTOCOL_GUID    \
+     EFI_GUID(0xce345171, 0xba0b, 0x11d2, 0x8e, 0x4f, \
+              0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b)
+
+struct efi_disk {
+     u64 revision;
+     efi_status_t (EFIAPI *read_disk)(struct efi_disk *this, u32 media_id,
+                                      u64 offset, efi_uintn_t buffer_size,
+                                      void *buffer);
+
+     efi_status_t (EFIAPI *write_disk)(struct efi_disk *this, u32 media_id,
+                                       u64 offset, efi_uintn_t buffer_size,
+                                       void *buffer);
+};
+
  #endif
lib/efi/efi_app.c +62 -0
index f61665686c5..9ba48517422 100644
@@ -21,6 +21,9 @@
  #include <efi.h>
  #include <efi_api.h>
  #include <sysreset.h>
+#include <dm/device-internal.h>
+#include <dm/lists.h>
+#include <dm/root.h>
 
  DECLARE_GLOBAL_DATA_PTR;
 
@@ -46,6 +49,33 @@ int efi_info_get(enum efi_entry_t type, void **datap, int *sizep)
      return -ENOSYS;
  }
 
+/**
+ * Create a block device so U-Boot can access an EFI device
+ *
+ * @handle:  EFI handle to bind
+ * @blkio:   block io protocol
+ * Return:   0 = success
+ */
+int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio)
+{
+     struct efi_media_plat plat;
+     struct udevice *dev;
+     char name[18];
+     int ret;
+
+     plat.handle = handle;
+     plat.blkio = blkio;
+     ret = device_bind(dm_root(), DM_DRIVER_GET(efi_media), "efi_media",
+                       &plat, ofnode_null(), &dev);
+     if (ret)
+             return log_msg_ret("bind", ret);
+
+     snprintf(name, sizeof(name), "efi_media_%x", dev_seq(dev));
+     device_set_name(dev, name);
+
+     return 0;
+}
+
  static efi_status_t setup_memory(struct efi_priv *priv)
  {
      struct efi_boot_services *boot = priv->boot;
@@ -105,6 +135,95 @@ static void free_memory(struct efi_priv *priv)
      global_data_ptr = NULL;
  }
 
+static int setup_disks(void)
+{
+     /* This is not fully implemented yet */
+     return 0;
+
+     efi_guid_t efi_disk_guid = EFI_DISK_IO_PROTOCOL_GUID;
+     struct efi_boot_services *boot = efi_get_boot();
+     struct efi_disk *disk;
+     int ret;
+
+     if (!boot)
+             return log_msg_ret("sys", -ENOSYS);
+     ret = boot->locate_protocol(&efi_disk_guid, NULL, (void **)&disk);
+     if (ret)
+             return log_msg_ret("prot", -ENOTSUPP);
+
+     return 0;
+}
+
+static int setup_block(void)
+{
+     efi_guid_t efi_blkio_guid = EFI_BLOCK_IO_PROTOCOL_GUID;
+     efi_guid_t efi_devpath_guid = EFI_DEVICE_PATH_PROTOCOL_GUID;
+     struct efi_boot_services *boot = efi_get_boot();
+     struct efi_block_io *blkio;
+     struct efi_device_path device_path;
+     efi_handle_t handle[100];
+     efi_uintn_t buf_size;
+     int num_handles;
+     int ret, i;
+
+     if (!boot)
Quoted 8 lines

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 174

Quoted 5 lines
lib/efi/efi_app.c +5 -0
+             return log_msg_ret("sys", -ENOSYS);
+
+     buf_size = sizeof(handle);
+     ret = boot->locate_handle(BY_PROTOCOL, &efi_blkio_guid, NULL,
+                               &buf_size, handle);

You could use LocateHandleBuffer() here which will allocate enough memory for all matching handles.

OK ta.

Quoted 13 lines

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 183

Quoted 10 lines
lib/efi/efi_app.c +9 -0
+     if (ret)
+             return log_msg_ret("loc", -ENOTSUPP);
+
+     num_handles = buf_size / sizeof(efi_handle_t);
+     log_info("Found %d EFI handles\n", num_handles);
+
+     for (i = 0; i < num_handles; i++) {
+             ret = boot->handle_protocol(handle[i], &efi_devpath_guid,
+                                         (void **)&device_path);

Why do you read the devicepath if you don't use it?

Well there is quite a bit of code here I am unsure about...e.g. will we need it later or is there something missing now? If not I can drop it.

Quoted 10 lines

lib/efi/efi_app.c static void free_memory(struct efi_priv *priv) · line 187

Quoted 5 lines
lib/efi/efi_app.c +4 -0
+             if (ret) {
+                     log_warning("- devpath %d failed (ret=%d)\n", i, ret);
+                     continue;
+             }

Here some analysis of devicepaths and installed protocols is missing to find out which of the handles represents a block device and which represents a partition:

Quoted 6 lines

If the last devicepath node is type 4 , Media Device Path with SubType 1 Hard Drive and the partition number is non-zero it is a partition.

If the devicepath without the last node relates to a handle with the EFI_BLOCK_IO_PROTOCOL, this also indicates that the current handle is for a partition.

OK thanks, I will take a look.

Regards,
Simon
[..]

#52 lore raw

PATCH 15/35 patman: Use a ValueError exception if tools.Run() fails

In reply to Simon Glass

The Exception base class is a very vague and could be confusing to the test system. Use the more specific ValueError exception instead.

---

 tools/patman/tools.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/patman/tools.py +1 -1
index 877e37cd8da..23ddb124804 100644
@@ -348,7 +348,7 @@ def Run(name, *args, **kwargs):
         result = command.RunPipe([all_args], capture=True, capture_stderr=True,
                                  env=env, raise_on_error=False, binary=binary)
         if result.return_code:
-            raise Exception("Error %d running '%s': %s" %
+            raise ValueError("Error %d running '%s': %s" %
                (result.return_code,' '.join(all_args),
                 result.stderr))
         return result.stdout
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#53 lore raw

PATCH 16/35 binman: Report an error if test files fail to compile

In reply to Simon Glass

At present any error from the 'make' command is silently swallowed by the test system. Fix this by showing it when detected.

---

 tools/binman/elf_test.py | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
tools/binman/elf_test.py +6 -2
index 7a128018d9f..bcccd78c0a1 100644
@@ -70,8 +70,12 @@ def BuildElfTestFiles(target_dir):
     # correctly. So drop any make flags here.
     if 'MAKEFLAGS' in os.environ:
         del os.environ['MAKEFLAGS']
-    tools.Run('make', '-C', target_dir, '-f',
-              os.path.join(testdir, 'Makefile'), 'SRC=%s/' % testdir)
+    try:
+        tools.Run('make', '-C', target_dir, '-f',
+                  os.path.join(testdir, 'Makefile'), 'SRC=%s/' % testdir)
+    except ValueError as e:
+        # The test system seems to suppress this in a strange way
+        print(e)
 
 
 class TestElf(unittest.TestCase):
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#54 lore raw

PATCH 17/35 binman: Support reading the offset of an ELF-file symbol

In reply to Simon Glass

Binman needs to be able to update the contents of an ELF file after it has been build. To support this, add a function to locate the position of a symbol's contents within the file.

Fix the comments on bss_data.c and Symbol while we are here.

---

 tools/binman/elf.py              | 53 ++++++++++++++++++++++++++++++--
 tools/binman/elf_test.py         | 37 ++++++++++++++++++++++
 tools/binman/test/Makefile       |  5 ++-
 tools/binman/test/bss_data.c     |  2 +-
 tools/binman/test/embed_data.c   | 16 ++++++++++
 tools/binman/test/embed_data.lds | 23 ++++++++++++++
 6 files changed, 131 insertions(+), 5 deletions(-)
 create mode 100644 tools/binman/test/embed_data.c
 create mode 100644 tools/binman/test/embed_data.lds
tools/binman/elf.py +50 -3
index 03b49d7163c..4aca4f847ce 100644
@@ -24,7 +24,14 @@ try:
 except:  # pragma: no cover
     ELF_TOOLS = False
 
-Symbol = namedtuple('Symbol', ['section', 'address', 'size', 'weak'])
+# Information about an EFL symbol:
+# section (str): Name of the section containing this symbol
+# address (int): Address of the symbol (its value)
+# size (int): Size of the symbol in bytes
+# weak (bool): True if the symbol is weak
+# offset (int or None): Offset of the symbol's data in the ELF file, or None if
+#   not known
+Symbol = namedtuple('Symbol', ['section', 'address', 'size', 'weak', 'offset'])
 
 # Information about an ELF file:
 #    data: Extracted program contents of ELF file (this would be loaded by an
@@ -71,8 +78,48 @@ def GetSymbols(fname, patterns):
         section, size =  parts[:2]
         if len(parts) > 2:
             name = parts[2] if parts[2] != '.hidden' else parts[3]
-            syms[name] = Symbol(section, int(value, 16), int(size,16),
-                                flags[1] == 'w')
+            syms[name] = Symbol(section, int(value, 16), int(size, 16),
+                                flags[1] == 'w', None)
+
+    # Sort dict by address
+    return OrderedDict(sorted(syms.items(), key=lambda x: x[1].address))
+
+def GetSymbolFileOffset(fname, patterns):
+    """Get the symbols from an ELF file
+
+    Args:
+        fname: Filename of the ELF file to read
+        patterns: List of regex patterns to search for, each a string
+
+    Returns:
+        None, if the file does not exist, or Dict:
+          key: Name of symbol
+          value: Hex value of symbol
+    """
+    def _GetFileOffset(elf, addr):
+        for seg in elf.iter_segments():
+            seg_end = seg['p_vaddr'] + seg['p_filesz']
+            if seg.header['p_type'] == 'PT_LOAD':
+                if addr >= seg['p_vaddr'] and addr < seg_end:
+                    return addr - seg['p_vaddr'] + seg['p_offset']
+
+    if not ELF_TOOLS:
+        raise ValueError('Python elftools package is not available')
+
+    syms = {}
+    with open(fname, 'rb') as fd:
+        elf = ELFFile(fd)
+
+        re_syms = re.compile('|'.join(patterns))
+        for section in elf.iter_sections():
+            if isinstance(section, SymbolTableSection):
+                for symbol in section.iter_symbols():
+                    if not re_syms or re_syms.search(symbol.name):
+                        addr = symbol.entry['st_value']
+                        syms[symbol.name] = Symbol(
+                            section.name, addr, symbol.entry['st_size'],
+                            symbol.entry['st_info']['bind'] == 'STB_WEAK',
+                            _GetFileOffset(elf, addr))
 
     # Sort dict by address
     return OrderedDict(sorted(syms.items(), key=lambda x: x[1].address))
tools/binman/elf_test.py +37 -0
index bcccd78c0a1..ac69a95b654 100644
@@ -6,6 +6,7 @@
 
 import os
 import shutil
+import struct
 import sys
 import tempfile
 import unittest
@@ -221,6 +222,42 @@ class TestElf(unittest.TestCase):
                          elf.DecodeElf(data, load + 2))
         shutil.rmtree(outdir)
 
+    def testEmbedData(self):
+        """Test for the GetSymbolFileOffset() function"""
+        if not elf.ELF_TOOLS:
+            self.skipTest('Python elftools not available')
+
+        fname = self.ElfTestFile('embed_data')
+        offset = elf.GetSymbolFileOffset(fname, ['embed_start', 'embed_end'])
+        start = offset['embed_start'].offset
+        end = offset['embed_end'].offset
+        data = tools.ReadFile(fname)
+        embed_data = data[start:end]
+        expect = struct.pack('<III', 0x1234, 0x5678, 0)
+        self.assertEqual(expect, embed_data)
+
+    def testEmbedFail(self):
+        """Test calling GetSymbolFileOffset() without elftools"""
+        try:
+            old_val = elf.ELF_TOOLS
+            elf.ELF_TOOLS = False
+            fname = self.ElfTestFile('embed_data')
+            with self.assertRaises(ValueError) as e:
+                elf.GetSymbolFileOffset(fname, ['embed_start', 'embed_end'])
+            self.assertIn('Python elftools package is not available',
+                      str(e.exception))
+        finally:
+            elf.ELF_TOOLS = old_val
+
+    def testEmbedDataNoSym(self):
+        """Test for GetSymbolFileOffset() getting no symbols"""
+        if not elf.ELF_TOOLS:
+            self.skipTest('Python elftools not available')
+
+        fname = self.ElfTestFile('embed_data')
+        offset = elf.GetSymbolFileOffset(fname, ['missing_sym'])
+        self.assertEqual({}, offset)
+
 
 if __name__ == '__main__':
     unittest.main()
tools/binman/test/Makefile +4 -1
index 0b19b7d9935..6e6cc6de491 100644
@@ -31,7 +31,7 @@ LDS_BINMAN_X86 := -T $(SRC)u_boot_binman_syms_x86.lds
 
 TARGETS = u_boot_ucode_ptr u_boot_no_ucode_ptr bss_data \
 	u_boot_binman_syms u_boot_binman_syms.bin u_boot_binman_syms_bad \
-	u_boot_binman_syms_size u_boot_binman_syms_x86
+	u_boot_binman_syms_size u_boot_binman_syms_x86 embed_data
 
 all: $(TARGETS)
 
@@ -44,6 +44,9 @@ u_boot_ucode_ptr: u_boot_ucode_ptr.c
 bss_data: CFLAGS += $(SRC)bss_data.lds
 bss_data: bss_data.c
 
+embed_data: CFLAGS += $(SRC)embed_data.lds
+embed_data: embed_data.c
+
 u_boot_binman_syms.bin: u_boot_binman_syms
 	$(OBJCOPY) -O binary $< -R .note.gnu.build-id $@
 
tools/binman/test/bss_data.c +1 -1
index 79537c31b65..4f9b64cef9e 100644
@@ -2,7 +2,7 @@
 /*
  * Copyright (c) 2016 Google, Inc
  *
- * Simple program to create a _dt_ucode_base_size symbol which can be read
+ * Simple program to create a bss_data region so the symbol can be read
  * by binutils. This is used by binman tests.
  */
 
tools/binman/test/embed_data.c new file +16 -0
new file mode 100644
index 00000000000..47d8c38248c
@@ -0,0 +1,16 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright 2021 Google LLC
+ *
+ * Simple program including some embedded data that can be accessed by binman.
+ * This is used by binman tests.
+ */
+
+int first[10] = {1};
+int embed[3] __attribute__((section(".embed"))) = {0x1234, 0x5678};
+int second[10] = {1};
+
+int main(void)
+{
+	return 0;
+}
tools/binman/test/embed_data.lds new file +23 -0
new file mode 100644
index 00000000000..908bf66c294
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Copyright 2021 Google LLC
+ */
+
+OUTPUT_FORMAT("elf32-i386", "elf32-i386", "elf32-i386")
+OUTPUT_ARCH(i386)
+ENTRY(_start)
+
+SECTIONS
+{
+	_start = .;
+	__data_start = .;
+	.data :
+	{
+		. = ALIGN(32);
+		embed_start = .;
+		*(.embed*)
+		embed_end = .;
+		. = ALIGN(32);
+		*(.data*)
+	}
+}
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#55 lore raw

PATCH 18/35 binman: Allow timeout to occur in the image or its section

In reply to Simon Glass

At present testThreadTimeout() assumes that the expected timeout happens first when building the section, but it can just as easily happen at the top-level image. Update the test to cope with both.

---

 tools/binman/ftest.py | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
tools/binman/ftest.py +1 -2
index cea3ebf2b9f..8199a4fc7e0 100644
@@ -4565,8 +4565,7 @@ class TestFunctional(unittest.TestCase):
         with self.assertRaises(ValueError) as e:
             self._DoTestFile('202_section_timeout.dts',
                              test_section_timeout=True)
-        self.assertIn("Node '/binman/section@0': Timed out obtaining contents",
-                      str(e.exception))
+        self.assertIn("Timed out obtaining contents", str(e.exception))
 
     def testTiming(self):
         """Test output of timing information"""
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#56 lore raw

PATCH 19/35 binman: Tidy up comments on _DoTestFile()

In reply to Simon Glass

The comment for this function is missing an argument and the return value. Fix it.

---

 tools/binman/ftest.py | 5 +++++
 1 file changed, 5 insertions(+)
tools/binman/ftest.py +5 -0
index 8199a4fc7e0..39a4b94cd0b 100644
@@ -334,6 +334,11 @@ class TestFunctional(unittest.TestCase):
             extra_indirs: Extra input directories to add using -I
             threads: Number of threads to use (None for default, 0 for
                 single-threaded)
+            test_section_timeout: True to force the first time to timeout, as
+                used in testThreadTimeout()
+
+        Returns:
+            int return code, 0 on success
         """
         args = []
         if debug:
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#57 lore raw

PATCH 20/35 binman: Support updating the dtb in an ELF file

In reply to Simon Glass

WIth EFI we must embed the devicetree in an ELF image so that it is loaded as part of the executable file. We want it to include the binman definition in there also, which in some cases cannot be created until the ELF (u-boot) is built. Add an option to binman to support writing the updated dtb to the ELF file u-boot.out

This is useful with the EFI app, which is always packaged as an ELF file.

---

 tools/binman/binman.rst                    | 36 ++++++++++
 tools/binman/cmdline.py                    |  2 +
 tools/binman/control.py                    | 11 +++
 tools/binman/elf.py                        | 21 ++++++
 tools/binman/ftest.py                      | 83 +++++++++++++++++++++-
 tools/binman/test/Makefile                 | 10 ++-
 tools/binman/test/u_boot_binman_embed.c    | 13 ++++
 tools/binman/test/u_boot_binman_embed.lds  | 29 ++++++++
 tools/binman/test/u_boot_binman_embed_sm.c | 13 ++++
 9 files changed, 216 insertions(+), 2 deletions(-)
 create mode 100644 tools/binman/test/u_boot_binman_embed.c
 create mode 100644 tools/binman/test/u_boot_binman_embed.lds
 create mode 100644 tools/binman/test/u_boot_binman_embed_sm.c
tools/binman/binman.rst +36 -0
index 09e7b571982..e2c7083722d 100644
@@ -822,6 +822,42 @@ the 'warning' line in scripts/Makefile.lib to see what it has found::
    # u_boot_dtsi_options_debug = $(u_boot_dtsi_options_raw)
 
 
+Updating an ELF file
+====================
+
+For the EFI app, where U-Boot is loaded from UEFI and runs as an app, there is
+no way to update the devicetree after U-Boot is built. Normally this works by
+creating a new u-boot.dtb.out with he updated devicetree, which is automatically
+built into the output image. With ELF this is not possible since the ELF is
+not part of an image, just a stand-along file. We must create an updated ELF
+file with the new devicetree.
+
+This is handled by the --update-fdt-in-elf option. It takes four arguments,
+separated by comma:
+
+   infile     - filename of input ELF file, e.g. 'u-boot's
+   outfile    - filename of output ELF file, e.g. 'u-boot.out'
+   begin_sym - symbol at the start of the embedded devicetree, e.g.
+   '__dtb_dt_begin'
+   end_sym   - symbol at the start of the embedded devicetree, e.g.
+   '__dtb_dt_end'
+
+When this flag is used, U-Boot does all the normal packaging, but as an
+additional step, it creates a new ELF file with the new devicetree embedded in
+it.
+
+If logging is enabled you will see a message like this::
+
+   Updating file 'u-boot' with data length 0x400a (16394) between symbols
+   '__dtb_dt_begin' and '__dtb_dt_end'
+
+There must be enough space for the updated devicetree. If not, an error like
+the following is produced::
+
+   ValueError: Not enough space in 'u-boot' for data length 0x400a (16394);
+   size is 0x1744 (5956)
+
+
 Entry Documentation
 ===================
 
tools/binman/cmdline.py +2 -0
index d6156df408b..23729f16dcc 100644
@@ -71,6 +71,8 @@ controlled by a description in the board device tree.'''
              'given')
     build_parser.add_argument('-u', '--update-fdt', action='store_true',
         default=False, help='Update the binman node with offset/size info')
+    build_parser.add_argument('--update-fdt-in-elf', type=str,
+        help='Update an ELF file with the output dtb: infile,outfile,begin_sym,end_sym')
 
     entry_parser = subparsers.add_parser('entry-docs',
         help='Write out entry documentation (see entries.rst)')
tools/binman/control.py +11 -0
index dcba02ff7f8..e73bc55252c 100644
@@ -598,6 +598,13 @@ def Binman(args):
             tools.FinaliseOutputDir()
         return 0
 
+    elf_params = None
+    if args.update_fdt_in_elf:
+        elf_params = args.update_fdt_in_elf.split(',')
+        if len(elf_params) != 4:
+            raise ValueError('Invalid args %s to --update-fdt-in-elf: expected infile,outfile,begin_sym,end_sym' %
+                             elf_params)
+
     # Try to figure out which device tree contains our image description
     if args.dt:
         dtb_fname = args.dt
@@ -644,6 +651,10 @@ def Binman(args):
             for dtb_item in state.GetAllFdts():
                 tools.WriteFile(dtb_item._fname, dtb_item.GetContents())
 
+            if elf_params:
+                data = state.GetFdtForEtype('u-boot-dtb').GetContents()
+                elf.UpdateFile(*elf_params, data)
+
             if missing:
                 tout.Warning("\nSome images are invalid")
 
tools/binman/elf.py +21 -0
index 4aca4f847ce..de2bb4651fa 100644
@@ -348,3 +348,24 @@ def DecodeElf(data, location):
                 segment.data()[offset:])
     return ElfInfo(output, data_start, elf.header['e_entry'] + virt_to_phys,
                    mem_end - data_start)
+
+def UpdateFile(infile, outfile, start_sym, end_sym, insert):
+    tout.Notice("Creating file '%s' with data length %#x (%d) between symbols '%s' and '%s'" %
+                (outfile, len(insert), len(insert), start_sym, end_sym))
+    syms = GetSymbolFileOffset(infile, [start_sym, end_sym])
+    if len(syms) != 2:
+        raise ValueError("Expected two symbols '%s' and '%s': got %d: %s" %
+                         (start_sym, end_sym, len(syms),
+                          ','.join(syms.keys())))
+
+    size = syms[end_sym].offset - syms[start_sym].offset
+    if len(insert) > size:
+        raise ValueError("Not enough space in '%s' for data length %#x (%d); size is %#x (%d)" %
+                         (infile, len(insert), len(insert), size, size))
+
+    data = tools.ReadFile(infile)
+    newdata = data[:syms[start_sym].offset]
+    newdata += insert + tools.GetBytes(0, size - len(insert))
+    newdata += data[syms[end_sym].offset:]
+    tools.WriteFile(outfile, newdata)
+    tout.Info('Written to offset %#x' % syms[start_sym].offset)
tools/binman/ftest.py +82 -1

107 lines not on this page · raw message

tools/binman/test/Makefile +9 -1
index 6e6cc6de491..387ba163353 100644
@@ -28,10 +28,12 @@ LDS_UCODE := -T $(SRC)u_boot_ucode_ptr.lds
 LDS_BINMAN := -T $(SRC)u_boot_binman_syms.lds
 LDS_BINMAN_BAD := -T $(SRC)u_boot_binman_syms_bad.lds
 LDS_BINMAN_X86 := -T $(SRC)u_boot_binman_syms_x86.lds
+LDS_BINMAN_EMBED := -T $(SRC)u_boot_binman_embed.lds
 
 TARGETS = u_boot_ucode_ptr u_boot_no_ucode_ptr bss_data \
 	u_boot_binman_syms u_boot_binman_syms.bin u_boot_binman_syms_bad \
-	u_boot_binman_syms_size u_boot_binman_syms_x86 embed_data
+	u_boot_binman_syms_size u_boot_binman_syms_x86 embed_data \
+	u_boot_binman_embed u_boot_binman_embed_sm
 
 all: $(TARGETS)
 
@@ -62,6 +64,12 @@ u_boot_binman_syms_bad: u_boot_binman_syms_bad.c
 u_boot_binman_syms_size: CFLAGS += $(LDS_BINMAN)
 u_boot_binman_syms_size: u_boot_binman_syms_size.c
 
+u_boot_binman_embed: CFLAGS += $(LDS_BINMAN_EMBED)
+u_boot_binman_embed: u_boot_binman_embed.c
+
+u_boot_binman_embed_sm: CFLAGS += $(LDS_BINMAN_EMBED)
+u_boot_binman_embed_sm: u_boot_binman_embed_sm.c
+
 clean:
 	rm -f $(TARGETS)
 
tools/binman/test/u_boot_binman_embed.c new file +13 -0
new file mode 100644
index 00000000000..75874bb6e23
@@ -0,0 +1,13 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright 2021 Google LLC
+ *
+ * Simple program to embed a devicetree. This is used by binman tests.
+ */
+
+int __attribute__((section(".mydtb"))) dtb_data[4096];
+
+int main(void)
+{
+	return 0;
+}
tools/binman/test/u_boot_binman_embed.lds new file +29 -0
new file mode 100644
index 00000000000..e213fa8a84c
@@ -0,0 +1,29 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Copyright (c) 2016 Google, Inc
+ */
+
+OUTPUT_FORMAT("elf32-i386", "elf32-i386", "elf32-i386")
+OUTPUT_ARCH(i386)
+ENTRY(_start)
+
+SECTIONS
+{
+	. = 0x00000000;
+	_start = .;
+
+	. = ALIGN(4);
+	.text :
+	{
+		*(.text*)
+	}
+
+	. = ALIGN(4);
+	.data : {
+		dtb_embed_begin = .;
+		KEEP(*(.mydtb));
+		dtb_embed_end = .;
+	}
+	.interp : { *(.interp*) }
+
+}
tools/binman/test/u_boot_binman_embed_sm.c new file +13 -0
new file mode 100644
index 00000000000..ae245d78a6a
@@ -0,0 +1,13 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright 2021 Google LLC
+ *
+ * Simple program to embed a devicetree. This is used by binman tests.
+ */
+
+int __attribute__((section(".mydtb"))) dtb_data[16];
+
+int main(void)
+{
+	return 0;
+}
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#58 lore raw

PATCH 21/35 efi: serial: Support arrow keys

In reply to Simon Glass

At present only the backspace key is supported in U-Boot, when running as an EFI app. Add support for arrows, home and end as well, to make the CLI more friendly.

---

 drivers/serial/serial_efi.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
drivers/serial/serial_efi.c +9 -2
index 33ddbd6080c..0067576389d 100644
@@ -24,6 +24,9 @@ struct serial_efi_priv {
 	bool have_key;
 };
 
+/* Convert a lower-case character to its ctrl-char equivalent */
+#define CTL_CH(c)		((c) - 'a' + 1)
+
 int serial_efi_setbrg(struct udevice *dev, int baudrate)
 {
 	return 0;
@@ -49,6 +52,7 @@ static int serial_efi_get_key(struct serial_efi_priv *priv)
 static int serial_efi_getc(struct udevice *dev)
 {
 	struct serial_efi_priv *priv = dev_get_priv(dev);
+	char conv_scan[10] = {0, 'p', 'n', 'f', 'b', 'a', 'e', 0, 8};
 	int ret, ch;
 
 	ret = serial_efi_get_key(priv);
@@ -63,8 +67,11 @@ static int serial_efi_getc(struct udevice *dev)
 	 * key scan code of 8. Handle this so that backspace works correctly
 	 * in the U-Boot command line.
 	 */
-	if (!ch && priv->key.scan_code == 8)
-		ch = 8;
+	if (!ch && priv->key.scan_code < sizeof(conv_scan)) {
+		ch = conv_scan[priv->key.scan_code];
+		if (ch >= 'a')
+			ch -= 'a' - 1;
+	}
 	debug(" [%x %x %x] ", ch, priv->key.unicode_char, priv->key.scan_code);
 
 	return ch;
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#59 lore raw

PATCH 22/35 bloblist: Move to rST format

In reply to Simon Glass

Move this documentation to the new format.

---

 doc/{README.bloblist => develop/bloblist.rst} | 4 ++--
 doc/develop/index.rst                         | 1 +
 2 files changed, 3 insertions(+), 2 deletions(-)
 rename doc/{README.bloblist => develop/bloblist.rst} (97%)
doc/develop/bloblist.rst renamed from doc/README.bloblist +2 -2
similarity index 97%
rename from doc/README.bloblist
rename to doc/develop/bloblist.rst
index 274c4605571..317ebc4919d 100644
@@ -1,4 +1,4 @@
-# SPDX-License-Identifier: GPL-2.0+
+.. SPDX-License-Identifier: GPL-2.0+
 
 Blob Lists - bloblist
 =====================
@@ -39,7 +39,7 @@ Blob tags
 
 Each blob has a tag which is a 32-bit number. This uniquely identifies the
 owner of the blob. Blob tags are listed in enum blob_tag_t and are named
-with a BLOBT_ prefix.
+with a `BLOBT_` prefix.
 
 
 Single structure
doc/develop/index.rst +1 -0
index 83c929babda..2a32645cfd2 100644
@@ -9,6 +9,7 @@ Implementation
 .. toctree::
    :maxdepth: 1
 
+   bloblist
    ci_testing
    commands
    devicetree/index
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
Acked-by on v1 22/35
#60 lore raw

Re: [PATCH 22/35] bloblist: Move to rST format

Quoted 3 lines · Simon Glass

Move this documentation to the new format.

Quoted 42 lines

---

  doc/{README.bloblist => develop/bloblist.rst} | 4 ++--
  doc/develop/index.rst                         | 1 +
  2 files changed, 3 insertions(+), 2 deletions(-)

rename doc/{README.bloblist => develop/bloblist.rst} (97%)

doc/develop/bloblist.rst renamed from doc/README.bloblist +2 -2
similarity index 97%
rename from doc/README.bloblist
rename to doc/develop/bloblist.rst
index 274c4605571..317ebc4919d 100644
@@ -1,4 +1,4 @@
-# SPDX-License-Identifier: GPL-2.0+
+.. SPDX-License-Identifier: GPL-2.0+
 
  Blob Lists - bloblist
  =====================
@@ -39,7 +39,7 @@ Blob tags
 
  Each blob has a tag which is a 32-bit number. This uniquely identifies the
  owner of the blob. Blob tags are listed in enum blob_tag_t and are named
-with a BLOBT_ prefix.
+with a `BLOBT_` prefix.
 
 
  Single structure
doc/develop/index.rst +1 -0
index 83c929babda..2a32645cfd2 100644
@@ -9,6 +9,7 @@ Implementation
  .. toctree::
     :maxdepth: 1
 
+   bloblist
     ci_testing
     commands
     devicetree/index
#61 lore raw

PATCH 23/35 bloblist: Support allocating the bloblist

In reply to Simon Glass

Typically the bloblist is positioned at a fixed address in memory until relocation. This is convenient when it is set up in SPL or before relocation.

But for EFI we want to set it up only when U-Boot proper is running. Add a way to allocate it using malloc() and update the documentation to cover this aspect of bloblist.

Note there are no tests of this feature at present, nor any direct testing of bloblist_init().

This can be added, e.g. by making this option controllable at runtime.

---

 common/Kconfig           | 15 +++++++++++++--
 common/bloblist.c        | 16 ++++++++++++++--
 common/board_f.c         |  8 +++++++-
 doc/develop/bloblist.rst | 16 ++++++++++++++++
 4 files changed, 50 insertions(+), 5 deletions(-)
common/Kconfig +13 -2
index 2ab20a6c85b..36a648596d1 100644
@@ -719,6 +719,8 @@ config TPL_BLOBLIST
 	  This enables a bloblist in TPL. The bloblist is set up in TPL and
 	  passed to SPL and U-Boot proper.
 
+if BLOBLIST
+
 config BLOBLIST_SIZE
 	hex "Size of bloblist"
 	depends on BLOBLIST
@@ -729,17 +731,24 @@ config BLOBLIST_SIZE
 	  is set up in the first part of U-Boot to run (TPL, SPL or U-Boot
 	  proper), and this sane bloblist is used for subsequent stages.
 
+config BLOBLIST_ALLOC
+	bool "Allocate bloblist"
+	help
+	  Allocate the bloblist using malloc(). This avoids the need to
+	  specify a fixed address on systems where this is unknown or can
+	  change at runtime.
+
 config BLOBLIST_ADDR
 	hex "Address of bloblist"
-	depends on BLOBLIST
 	default 0xc000 if SANDBOX
 	help
 	  Sets the address of the bloblist, set up by the first part of U-Boot
 	  which runs. Subsequent U-Boot stages typically use the same address.
 
+	  This is not used if BLOBLIST_ALLOC is selected.
+
 config BLOBLIST_SIZE_RELOC
 	hex "Size of bloblist after relocation"
-	depends on BLOBLIST
 	default BLOBLIST_SIZE
 	help
 	  Sets the size of the bloblist in bytes after relocation. Since U-Boot
@@ -747,6 +756,8 @@ config BLOBLIST_SIZE_RELOC
 	  size than the one set up by SPL. This bloblist is set up during the
 	  relocation process.
 
+endif # BLOBLIST
+
 endmenu
 
 source "common/spl/Kconfig"
common/bloblist.c +14 -2
index 1290fff8504..01b04103d91 100644
@@ -7,6 +7,7 @@
 #include <common.h>
 #include <bloblist.h>
 #include <log.h>
+#include <malloc.h>
 #include <mapmem.h>
 #include <spl.h>
 #include <asm/global_data.h>
@@ -416,10 +417,21 @@ int bloblist_init(void)
 		ret = bloblist_check(CONFIG_BLOBLIST_ADDR,
 				     CONFIG_BLOBLIST_SIZE);
 	if (ret) {
+		ulong addr;
+
 		log(LOGC_BLOBLIST, expected ? LOGL_WARNING : LOGL_DEBUG,
 		    "Existing bloblist not found: creating new bloblist\n");
-		ret = bloblist_new(CONFIG_BLOBLIST_ADDR, CONFIG_BLOBLIST_SIZE,
-				   0);
+		if (IS_ENABLED(CONFIG_BLOBLIST_ALLOC)) {
+			void *ptr = memalign(BLOBLIST_ALIGN,
+					     CONFIG_BLOBLIST_SIZE);
+
+			if (!ptr)
+				return log_msg_ret("alloc", -ENOMEM);
+			addr = map_to_sysmem(ptr);
+		} else {
+			addr = CONFIG_BLOBLIST_ADDR;
+		}
+		ret = bloblist_new(addr, CONFIG_BLOBLIST_SIZE, 0);
 	} else {
 		log(LOGC_BLOBLIST, LOGL_DEBUG, "Found existing bloblist\n");
 	}
common/board_f.c +7 -1
index f2746537c96..a5448c50666 100644
@@ -655,8 +655,14 @@ static int reloc_bootstage(void)
 static int reloc_bloblist(void)
 {
 #ifdef CONFIG_BLOBLIST
-	if (gd->flags & GD_FLG_SKIP_RELOC)
+	/*
+	 * Relocate only if we are supposed to send it
+	 */
+	if ((gd->flags & GD_FLG_SKIP_RELOC) &&
+	    CONFIG_BLOBLIST_SIZE == CONFIG_BLOBLIST_SIZE_RELOC) {
+		debug("Not relocating bloblist\n");
 		return 0;
+	}
 	if (gd->new_bloblist) {
 		int size = CONFIG_BLOBLIST_SIZE;
 
doc/develop/bloblist.rst +16 -0
index 317ebc4919d..47274cf8e26 100644
@@ -59,6 +59,22 @@ Bloblist provides a fairly simple API which allows blobs to be created and
 found. All access is via the blob's tag. Blob records are zeroed when added.
 
 
+Placing the bloblist
+--------------------
+
+The bloblist is typically positioned at a fixed address by TPL, or SPL. This
+is controlled by `CONFIG_BLOBLIST_ADDR`. But in some cases it is preferable to
+allocate the bloblist in the malloc() space. Use the `CONFIG_BLOBLIST_ALLOC`
+option to enable this.
+
+The bloblist is automatically relocated as part of U-Boot relocation. Sometimes
+it is useful to expand the bloblist in U-Boot proper, since it may want to add
+information for use by Linux. Note that this does not mean that Linux needs to
+know anything about the bloblist format, just that it is convenient to use
+bloblist to place things contiguously in memory. Set
+`CONFIG_BLOBLIST_SIZE_RELOC` to define the expanded size, if needed.
+
+
 Finishing the bloblist
 ----------------------
 
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#62 lore raw

PATCH 24/35 x86: Allow booting a kernel from the EFI app

In reply to Simon Glass

At present this is disabled, but it should work so long as the kernel does not need EFI services. Enable it and add a note about remaining work.

---

 arch/x86/lib/bootm.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
arch/x86/lib/bootm.c +7 -4
index 733dd712570..ff0e9f6a085 100644
@@ -179,10 +179,14 @@ int boot_linux_kernel(ulong setup_base, ulong load_address, bool image_64bit)
 		* U-Boot is setting them up that way for itself in
 		* arch/i386/cpu/cpu.c.
 		*
-		* Note that we cannot currently boot a kernel while running as
-		* an EFI application. Please use the payload option for that.
+		* Note: this is incomplete for EFI kernels!
+		*
+		* This can boot a kernel while running as an EFI application,
+		* but if the kernel requires EFI support then that support needs
+		* to be enabled first (see EFI_LOADER). Also the EFI information
+		* must enabled with setup_efi_info(). See setup_zimage() for
+		* how this is done with the stub.
 		*/
-#ifndef CONFIG_EFI_APP
 		__asm__ __volatile__ (
 		"movl $0, %%ebp\n"
 		"cli\n"
@@ -191,7 +195,6 @@ int boot_linux_kernel(ulong setup_base, ulong load_address, bool image_64bit)
 		[boot_params] "S"(setup_base),
 		"b"(0), "D"(0)
 		);
-#endif
 	}
 
 	/* We can't get to here */
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#63 lore raw

Re: [PATCH 24/35] x86: Allow booting a kernel from the EFI app

Quoted 16 lines · Simon Glass

At present this is disabled, but it should work so long as the kernel does not need EFI services. Enable it and add a note about remaining work.

---

  arch/x86/lib/bootm.c | 11 +++++++----
  1 file changed, 7 insertions(+), 4 deletions(-)
arch/x86/lib/bootm.c +0 -0
index 733dd712570..ff0e9f6a085 100644
@@ -179,10 +179,14 @@ int boot_linux_kernel(ulong setup_base, ulong load_address, bool image_64bit)
  		* U-Boot is setting them up that way for itself in
  		* arch/i386/cpu/cpu.c.

arch/x86/lib/bootm.c int boot_linux_kernel(ulong setup_base, ulong load_address, bool image_64bit) · line 188

Quoted 10 lines
arch/x86/lib/bootm.c +7 -2
  		*
-		* Note that we cannot currently boot a kernel while running as
-		* an EFI application. Please use the payload option for that.
+		* Note: this is incomplete for EFI kernels!
+		*
+		* This can boot a kernel while running as an EFI application,
+		* but if the kernel requires EFI support then that support needs
+		* to be enabled first (see EFI_LOADER). Also the EFI information
+		* must enabled with setup_efi_info(). See setup_zimage() for
+		* how this is done with the stub.

Did you test launching an EFI application with EFI_LOADER from the EFI app? If you get it working, please provide a defconfig and a testcase that runs helloworld.efi.

Best regards

Heinrich

Quoted 14 lines
arch/x86/lib/bootm.c +0 -2
  		*/
-#ifndef CONFIG_EFI_APP
  		__asm__ __volatile__ (
  		"movl $0, %%ebp\n"
  		"cli\n"
@@ -191,7 +195,6 @@ int boot_linux_kernel(ulong setup_base, ulong load_address, bool image_64bit)
  		[boot_params] "S"(setup_base),
  		"b"(0), "D"(0)
  		);
-#endif
  	}
 
  	/* We can't get to here */
#64 lore raw

Re: [PATCH 24/35] x86: Allow booting a kernel from the EFI app

Hi Heinrich,

Quoted 26 lines · Heinrich Schuchardt
Quoted 22 lines · Simon Glass

At present this is disabled, but it should work so long as the kernel does not need EFI services. Enable it and add a note about remaining work.

---

  arch/x86/lib/bootm.c | 11 +++++++----
  1 file changed, 7 insertions(+), 4 deletions(-)
arch/x86/lib/bootm.c +3 -2
index 733dd712570..ff0e9f6a085 100644
@@ -179,10 +179,14 @@ int boot_linux_kernel(ulong setup_base, ulong load_address, bool image_64bit)
              * U-Boot is setting them up that way for itself in
              * arch/i386/cpu/cpu.c.
              *
-             * Note that we cannot currently boot a kernel while running as
-             * an EFI application. Please use the payload option for that.
+             * Note: this is incomplete for EFI kernels!
+             *
+             * This can boot a kernel while running as an EFI application,
Quoted 8 lines

arch/x86/lib/bootm.c int boot_linux_kernel(ulong setup_base, ulong load_address, bool image_64bit) · line 188

Quoted 4 lines
arch/x86/lib/bootm.c +4 -0
+             * but if the kernel requires EFI support then that support needs
+             * to be enabled first (see EFI_LOADER). Also the EFI information
+             * must enabled with setup_efi_info(). See setup_zimage() for
+             * how this is done with the stub.

Did you test launching an EFI application with EFI_LOADER from the EFI app? If you get it working, please provide a defconfig and a testcase that runs helloworld.efi.

No unfortunately that cannot be built at present. This is just for kernels that don't use EFI_LOADER.

I am pretty sure it could be made to work, perhaps by passing the original UEFI system table through to the kernel.

Regards,
Simon

#65 lore raw

PATCH 25/35 x86: Don't process the kernel command line unless enabled

In reply to Simon Glass

If the 'bootm' command is not enabled then this code is not available and this causes a link error. Fix it.

Note that for the EFI app, there is no indication of missing code. It just hangs!

---

 arch/x86/lib/zimage.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)
arch/x86/lib/zimage.c +8 -5
index 7ce02226ef9..9cc04490307 100644
@@ -365,11 +365,14 @@ int setup_zimage(struct boot_params *setup_base, char *cmd_line, int auto_boot,
 			strcpy(cmd_line, (char *)cmdline_force);
 		else
 			build_command_line(cmd_line, auto_boot);
-		ret = bootm_process_cmdline(cmd_line, max_size, BOOTM_CL_ALL);
-		if (ret) {
-			printf("Cmdline setup failed (max_size=%x, bootproto=%x, err=%d)\n",
-			       max_size, bootproto, ret);
-			return ret;
+		if (IS_ENABLED(CONFIG_CMD_BOOTM)) {
+			ret = bootm_process_cmdline(cmd_line, max_size,
+						    BOOTM_CL_ALL);
+			if (ret) {
+				printf("Cmdline setup failed (max_size=%x, bootproto=%x, err=%d)\n",
+				       max_size, bootproto, ret);
+				return ret;
+			}
 		}
 		printf("Kernel command line: \"");
 		puts(cmd_line);
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#66 lore raw

PATCH 26/35 x86: efi: Add room for the binman definition in the dtb

In reply to Simon Glass

At present only 4KB of spare space is left in the DTB when building the EFI app. Increase this to 32KB so there is plenty of space to insert the binman definition. This cannot be expanded later (as with OF_SEPARATE) because the ELF image has already been built.

---

 arch/x86/dts/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/dts/Makefile +1 -1
index be209aaaf8f..5c8c05ec499 100644
@@ -24,7 +24,7 @@ dtb-y += bayleybay.dtb \
 
 targets += $(dtb-y)
 
-DTC_FLAGS += -R 4 -p 0x1000
+DTC_FLAGS += -R 4 -p $(if $(CONFIG_EFI_APP),0x8000,0x1000)
 
 PHONY += dtbs
 dtbs: $(addprefix $(obj)/, $(dtb-y))
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#67 lore raw

PATCH 27/35 efi: Add comments to struct efi_priv

In reply to Simon Glass

This structure is uncommented. Fix it.

---

 include/efi.h | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)
include/efi.h +25 -1
index c0fddf7f6cd..2db60a36f0b 100644
@@ -400,15 +400,39 @@ static inline struct efi_mem_desc *efi_get_next_mem_desc(
 	return (struct efi_mem_desc *)((ulong)desc + map->desc_size);
 }
 
+/**
+ * struct efi_priv - Information about the environment provided by EFI
+ *
+ * @parent_image: image passed into the EFI app or stub
+ * @sys_table: Pointer to system table
+ * @boot: Pointer to boot-services table
+ * @run: Pointer to runtime-services table
+ *
+ * Used by app only:
+ * @use_pool_for_malloc: true if all allocation should go through the EFI 'pool'
+ *	methods allocate_pool() and free_pool(); false to use 'pages' methods
+ *	allocate_pages() and free_pages()
+ * @ram_base: Base address of RAM (size CONFIG_EFI_RAM_SIZE)
+ * @image_data_type: Type of the loaded image (e.g. EFI_LOADER_CODE)
+ *
+ * Used by stub only:
+ * @info: Header of the info list, holding info collected by the stub and passed
+ *	to U-Boot
+ * @info_size: Size of the info list, in bytes from @info
+ * @next_hdr: Pointer to where to put the next header when adding to the list
+ */
 struct efi_priv {
 	efi_handle_t parent_image;
-	struct efi_device_path *device_path;
 	struct efi_system_table *sys_table;
 	struct efi_boot_services *boot;
 	struct efi_runtime_services *run;
+
+	/* app: */
 	bool use_pool_for_malloc;
 	unsigned long ram_base;
 	unsigned int image_data_type;
+
+	/* stub: */
 	struct efi_info_hdr *info;
 	unsigned int info_size;
 	void *next_hdr;
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#68 lore raw

PATCH 28/35 efi: Fix ll_boot_init() operation with the app

In reply to Simon Glass

This should return false when the EFI app is running, since UEFI has done the required low-level init. Fix it.

---

 include/init.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/init.h +1 -1
index c781789e367..6d682b02b05 100644
@@ -15,7 +15,7 @@
 #include <linux/types.h>
 
 /* Avoid using CONFIG_EFI_STUB directly as we may boot from other loaders */
-#ifdef CONFIG_EFI_STUB
+#ifdef CONFIG_EFI
 #define ll_boot_init()	false
 #else
 #include <asm/global_data.h>
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#69 lore raw

PATCH 29/35 efi: Add a few comments to the stub

In reply to Simon Glass

Comment some functions that need more information.

---

 lib/efi/efi_stub.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
lib/efi/efi_stub.c +22 -0
index b3393e47fae..156cbf0b928 100644
@@ -225,6 +225,22 @@ static int get_codeseg32(void)
 	return cs32;
 }
 
+/**
+ * setup_info_table() - sets up a table containing information from EFI
+ *
+ * We must call exit_boot_services() before jumping out of the stub into U-Boot
+ * proper, so that U-Boot has full control of peripherals, memory, etc.
+ *
+ * Once we do this, we cannot call any boot-services functions so we must find
+ * out everything we need to before doing that.
+ *
+ * Set up a struct efi_info_hdr table which can hold various records (e.g.
+ * struct efi_entry_memmap) with information obtained from EFI.
+ *
+ * @priv: Pointer to our private information which contains the list
+ * @size: Size of the table to allocate
+ * @return 0 if OK, non-zero on error
+ */
 static int setup_info_table(struct efi_priv *priv, int size)
 {
 	struct efi_info_hdr *info;
@@ -248,6 +264,12 @@ static int setup_info_table(struct efi_priv *priv, int size)
 	return 0;
 }
 
+/**
+ * add_entry_addr() - Add a new entry to the efi_info list
+ *
+ * @priv: Pointer to our private information which contains the list
+ *
+ */
 static void add_entry_addr(struct efi_priv *priv, enum efi_entry_t type,
 			   void *ptr1, int size1, void *ptr2, int size2)
 {
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#70 lore raw

PATCH 30/35 efi: Share struct efi_priv between the app and stub code

In reply to Simon Glass

At present each of these has its own static variable and helper functions. Move them into a shared file.

---

 include/efi.h      | 21 +++++++++++++++++++++
 lib/efi/efi.c      | 29 +++++++++++++++++++++++++++++
 lib/efi/efi_app.c  | 21 ++-------------------
 lib/efi/efi_stub.c |  7 ++++---
 4 files changed, 56 insertions(+), 22 deletions(-)
include/efi.h +21 -0
index 2db60a36f0b..fd43dec15a2 100644
@@ -474,6 +474,27 @@ extern char _binary_u_boot_bin_start[], _binary_u_boot_bin_end[];
 				EFI_VARIABLE_TIME_BASED_AUTHENTICATED_WRITE_ACCESS | \
 				EFI_VARIABLE_APPEND_WRITE)
 
+/**
+ * efi_get_priv() - Get access to the EFI-private information
+ *
+ * This struct it used by both the stub and the app to record things about the
+ * EFI environment. It is not available in U-Boot proper after the stub has
+ * jumped there. Use efi_info_get() to obtain info in that case.
+ *
+ * @return pointer to private info
+ */
+struct efi_priv *efi_get_priv(void);
+
+/**
+ * efi_set_priv() - Set up a pointer to the EFI-private information
+ *
+ * This is called in the stub and app to record the location of this
+ * information.
+ *
+ * @priv: New location of private data
+ */
+void efi_set_priv(struct efi_priv *priv);
+
 /**
  * efi_get_sys_table() - Get access to the main EFI system table
  *
lib/efi/efi.c +29 -0
index 69e52e45748..cd6bf47b180 100644
@@ -1,5 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0+
 /*
+ * Functions shared by the app and stub
+ *
  * Copyright (c) 2015 Google, Inc
  *
  * EFI information obtained here:
@@ -17,6 +19,33 @@
 #include <efi.h>
 #include <efi_api.h>
 
+static struct efi_priv *global_priv;
+
+struct efi_priv *efi_get_priv(void)
+{
+	return global_priv;
+}
+
+void efi_set_priv(struct efi_priv *priv)
+{
+	global_priv = priv;
+}
+
+struct efi_system_table *efi_get_sys_table(void)
+{
+	return global_priv->sys_table;
+}
+
+struct efi_boot_services *efi_get_boot(void)
+{
+	return global_priv->boot;
+}
+
+unsigned long efi_get_ram_base(void)
+{
+	return global_priv->ram_base;
+}
+
 /*
  * Global declaration of gd.
  *
lib/efi/efi_app.c +2 -19
index 9ba48517422..c02f0c90c45 100644
@@ -27,23 +27,6 @@
 
 DECLARE_GLOBAL_DATA_PTR;
 
-static struct efi_priv *global_priv;
-
-struct efi_system_table *efi_get_sys_table(void)
-{
-	return global_priv->sys_table;
-}
-
-struct efi_boot_services *efi_get_boot(void)
-{
-	return global_priv->boot;
-}
-
-unsigned long efi_get_ram_base(void)
-{
-	return global_priv->ram_base;
-}
-
 int efi_info_get(enum efi_entry_t type, void **datap, int *sizep)
 {
 	return -ENOSYS;
@@ -240,7 +223,7 @@ efi_status_t EFIAPI efi_main(efi_handle_t image,
 	/* Set up access to EFI data structures */
 	efi_init(priv, "App", image, sys_table);
 
-	global_priv = priv;
+	efi_set_priv(priv);
 
 	/*
 	 * Set up the EFI debug UART so that printf() works. This is
@@ -266,7 +249,7 @@ efi_status_t EFIAPI efi_main(efi_handle_t image,
 
 static void efi_exit(void)
 {
-	struct efi_priv *priv = global_priv;
+	struct efi_priv *priv = efi_get_priv();
 
 	free_memory(priv);
 	printf("U-Boot EFI exiting\n");
lib/efi/efi_stub.c +4 -3
index 156cbf0b928..bc4c3a48720 100644
@@ -31,7 +31,6 @@
 #error "This file needs to be ported for use on architectures"
 #endif
 
-static struct efi_priv *global_priv;
 static bool use_uart;
 
 struct __packed desctab_info {
@@ -63,6 +62,8 @@ void _debug_uart_init(void)
 
 void putc(const char ch)
 {
+	struct efi_priv *priv = efi_get_priv();
+
 	if (ch == '\n')
 		putc('\r');
 
@@ -73,7 +74,7 @@ void putc(const char ch)
 			;
 		outb(ch, (ulong)&com_port->thr);
 	} else {
-		efi_putc(global_priv, ch);
+		efi_putc(priv, ch);
 	}
 }
 
@@ -313,7 +314,7 @@ efi_status_t EFIAPI efi_main(efi_handle_t image,
 		puts(" efi_init() failed\n");
 		return ret;
 	}
-	global_priv = priv;
+	efi_set_priv(priv);
 
 	cs32 = get_codeseg32();
 	if (cs32 < 0)
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#71 lore raw

PATCH 31/35 efi: Move exit_boot_services into a function

In reply to Simon Glass

At present this code is inline in the app and stub. But they do the same thing. The difference is that the stub does it immediately and the app doesn't want to do it until the end (when it boots a kernel) or not at all, if returning to UEFI.

Also store the memory map so that it can be accessed within the app if needed.

---

 include/efi.h      | 32 +++++++++++++++++++++
 lib/efi/efi.c      | 70 ++++++++++++++++++++++++++++++++++++++++++++++
 lib/efi/efi_app.c  |  3 ++
 lib/efi/efi_stub.c | 66 ++++++++-----------------------------------
 4 files changed, 116 insertions(+), 55 deletions(-)
include/efi.h +32 -0
index fd43dec15a2..abc3fb2f94a 100644
@@ -407,6 +407,12 @@ static inline struct efi_mem_desc *efi_get_next_mem_desc(
  * @sys_table: Pointer to system table
  * @boot: Pointer to boot-services table
  * @run: Pointer to runtime-services table
+ * @memmap_key: Key returned from get_memory_map()
+ * @memmap_desc: List of memory-map records
+ * @memmap_alloc: Amount of memory allocated for memory map list
+ * @memmap_size Size of memory-map list in bytes
+ * @memmap_desc_size: Size of an individual memory-map record, in bytes
+ * @memmap_version: Memory-map version
  *
  * Used by app only:
  * @use_pool_for_malloc: true if all allocation should go through the EFI 'pool'
@@ -426,6 +432,12 @@ struct efi_priv {
 	struct efi_system_table *sys_table;
 	struct efi_boot_services *boot;
 	struct efi_runtime_services *run;
+	efi_uintn_t memmap_key;
+	struct efi_mem_desc *memmap_desc;
+	efi_uintn_t memmap_alloc;
+	efi_uintn_t memmap_size;
+	efi_uintn_t memmap_desc_size;
+	u32 memmap_version;
 
 	/* app: */
 	bool use_pool_for_malloc;
@@ -589,4 +601,24 @@ int efi_info_get(enum efi_entry_t type, void **datap, int *sizep);
  */
 int efi_bind_block(efi_handle_t handle, struct efi_block_io *blkio);
 
+/**
+ * efi_store_memory_map() - Collect the memory-map info from EFI
+ *
+ * Collect the memory info and store it for later use, e.g. in calling
+ * exit_boot_services()
+ *
+ * @priv:	Pointer to private EFI structure
+ * @return 0 if OK, non-zero on error
+ */
+int efi_store_memory_map(struct efi_priv *priv);
+
+/**
+ * efi_call_exit_boot_services() - Handlet eh exit-boot-service procedure
+ *
+ * Tell EFI we don't want their boot services anymore
+ *
+ * @return 0 if OK, non-zero on error
+ */
+int efi_call_exit_boot_services(void);
+
 #endif /* _LINUX_EFI_H */
lib/efi/efi.c +70 -0

73 lines not on this page · raw message

lib/efi/efi_app.c +3 -0
index c02f0c90c45..6da3fbd5104 100644
@@ -237,6 +237,9 @@ efi_status_t EFIAPI efi_main(efi_handle_t image,
 		printf("Failed to set up memory: ret=%lx\n", ret);
 		return ret;
 	}
+	ret = efi_store_memory_map(priv);
+	if (ret)
+		return ret;
 
 	printf("starting\n");
 
lib/efi/efi_stub.c +11 -55

98 lines not on this page · raw message

Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#72 lore raw

PATCH 32/35 efi: Check for failure when initing the app

In reply to Simon Glass

The stub checks for failure with efi_init(). Add this for the app as well. It is unlikely that anything can be done, but we may as well stop.

---

 lib/efi/efi_app.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
lib/efi/efi_app.c +5 -2
index 6da3fbd5104..e0666d1fae2 100644
@@ -221,8 +221,11 @@ efi_status_t EFIAPI efi_main(efi_handle_t image,
 	efi_status_t ret;
 
 	/* Set up access to EFI data structures */
-	efi_init(priv, "App", image, sys_table);
-
+	ret = efi_init(priv, "App", image, sys_table);
+	if (ret) {
+		printf("Failed to set up ARP: err=%lx\n", ret);
+		return ret;
+	}
 	efi_set_priv(priv);
 
 	/*
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#73 lore raw

PATCH 33/35 efi: Mention that efi_info_get() is only used in the stub

In reply to Simon Glass

This provides access to EFI tables after U-Boot as exited boot services. It is not needed in the app since boot services remain alive and we can just call them whenever needed.

Add a comment to explain this.

---

 include/efi.h | 4 ++++
 1 file changed, 4 insertions(+)
include/efi.h +4 -0
index abc3fb2f94a..f624243d596 100644
@@ -578,6 +578,10 @@ void efi_putc(struct efi_priv *priv, const char ch);
 /**
  * efi_info_get() - get an entry from an EFI table
  *
+ * This is called from U-Boot proper to read information set up by the EFI stub.
+ * It can only be used when running from the EFI stuff, not when U-Boot is
+ * running as an app.
+ *
  * @type:	Entry type to search for
  * @datap:	Returns pointer to entry data
  * @sizep:	Returns pointer to entry size
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#74 lore raw

PATCH 34/35 efi: Show when allocated pages are used

In reply to Simon Glass

Add a message here so that both paths of memory allocation are reported.

---

 lib/efi/efi_app.c | 1 +
 1 file changed, 1 insertion(+)
lib/efi/efi_app.c +1 -0
index e0666d1fae2..c8f4784eb7d 100644
@@ -97,6 +97,7 @@ static efi_status_t setup_memory(struct efi_priv *priv)
 			return ret;
 		priv->use_pool_for_malloc = true;
 	} else {
+		printf("(using allocated RAM address %lx) ", (ulong)addr);
 		priv->ram_base = addr;
 	}
 	gd->ram_size = pages << 12;
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#75 lore raw

Re: [PATCH 34/35] efi: Show when allocated pages are used

Quoted 13 lines · Simon Glass

Add a message here so that both paths of memory allocation are reported.

---

  lib/efi/efi_app.c | 1 +
  1 file changed, 1 insertion(+)
lib/efi/efi_app.c +0 -0
index e0666d1fae2..c8f4784eb7d 100644

lib/efi/efi_app.c static efi_status_t setup_memory(struct efi_priv *priv) · line 100

Quoted 4 lines
lib/efi/efi_app.c +1 -0
@@ -97,6 +97,7 @@ static efi_status_t setup_memory(struct efi_priv *priv)
  			return ret;
  		priv->use_pool_for_malloc = true;
  	} else {
+		printf("(using allocated RAM address %lx) ", (ulong)addr);

Shouldn't this be log_debug()?

Best regards

Heinrich

Quoted 4 lines
lib/efi/efi_app.c +0 -0
  		priv->ram_base = addr;
  	}
  	gd->ram_size = pages << 12;
#76 lore raw

Re: [PATCH 34/35] efi: Show when allocated pages are used

Hi Heinrich,

Quoted 12 lines · Heinrich Schuchardt
Quoted 8 lines · Simon Glass

Add a message here so that both paths of memory allocation are reported.

---

  lib/efi/efi_app.c | 1 +
  1 file changed, 1 insertion(+)
Quoted 11 lines

lib/efi/efi_app.c static efi_status_t setup_memory(struct efi_priv *priv) · line 100

Quoted 9 lines
lib/efi/efi_app.c +1 -0
index e0666d1fae2..c8f4784eb7d 100644
@@ -97,6 +97,7 @@ static efi_status_t setup_memory(struct efi_priv *priv)
                      return ret;
              priv->use_pool_for_malloc = true;
      } else {
+             printf("(using allocated RAM address %lx) ", (ulong)addr);

Shouldn't this be log_debug()?

It matches the one a few lines higher. I suppose it could be a debug statement, but allocation is an area that seems to different on different UEFIs, so I have it more prominent for the moment, until we figure it out.

Regards,
Simon

#77 lore raw

Re: [PATCH 34/35] efi: Show when allocated pages are used

Quoted 27 lines · Simon Glass

Hi Heinrich,

Quoted 12 lines · Heinrich Schuchardt
Quoted 8 lines · Simon Glass

Add a message here so that both paths of memory allocation are reported.

---

   lib/efi/efi_app.c | 1 +
   1 file changed, 1 insertion(+)
Quoted 11 lines

lib/efi/efi_app.c static efi_status_t setup_memory(struct efi_priv *priv) · line 100

Quoted 9 lines
lib/efi/efi_app.c +1 -0
index e0666d1fae2..c8f4784eb7d 100644
@@ -97,6 +97,7 @@ static efi_status_t setup_memory(struct efi_priv *priv)
                       return ret;
               priv->use_pool_for_malloc = true;
       } else {
+             printf("(using allocated RAM address %lx) ", (ulong)addr);

Shouldn't this be log_debug()?

Quoted 4 lines

It matches the one a few lines higher. I suppose it could be a debug statement, but allocation is an area that seems to different on different UEFIs, so I have it more prominent for the moment, until we figure it out.

If you want it more prominent, isn't it log_info() that you were promoting?

Best regards

Heinrich

#78 lore raw

Re: [PATCH 34/35] efi: Show when allocated pages are used

Hi Heinrich.

Quoted 31 lines · Heinrich Schuchardt
Quoted 27 lines · Simon Glass

Hi Heinrich,

Quoted 23 lines · Heinrich Schuchardt
Quoted 13 lines · Simon Glass

Add a message here so that both paths of memory allocation are reported.

---

   lib/efi/efi_app.c | 1 +
   1 file changed, 1 insertion(+)
lib/efi/efi_app.c +0 -0
index e0666d1fae2..c8f4784eb7d 100644

lib/efi/efi_app.c static efi_status_t setup_memory(struct efi_priv *priv) · line 100

Quoted 4 lines
lib/efi/efi_app.c +1 -0
@@ -97,6 +97,7 @@ static efi_status_t setup_memory(struct efi_priv *priv)
                       return ret;
               priv->use_pool_for_malloc = true;
       } else {
+             printf("(using allocated RAM address %lx) ", (ulong)addr);

Shouldn't this be log_debug()?

Quoted 6 lines
Quoted 4 lines

It matches the one a few lines higher. I suppose it could be a debug statement, but allocation is an area that seems to different on different UEFIs, so I have it more prominent for the moment, until we figure it out.

If you want it more prominent, isn't it log_info() that you were promoting?

OK I can change them both, so long as logging is enabled for EFI...will check.

Regards,
Simon

#79 lore raw

PATCH 35/35 efi: Allow easy selection of serial-only operation

In reply to Simon Glass

Add info about how to select vidconsole or serial.

---

 include/configs/efi-x86_app.h | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)
include/configs/efi-x86_app.h +25 -0
index 6061a6db0a4..cdd60be9f85 100644
@@ -10,8 +10,33 @@
 
 #undef CONFIG_TPM_TIS_BASE_ADDRESS
 
+/*
+ * Select the output device: Put an 'x' prefix before one of these to disable it
+ */
+
+/*
+ * Video output - can normally continue after exit_boot_services has been
+ * called, since output to the display does not require EFI services at that
+ * point. U-Boot sets up the console memory and does its own drawing.
+ */
 #define CONFIG_STD_DEVICES_SETTINGS	"stdin=serial\0" \
 					"stdout=vidconsole\0" \
 					"stderr=vidconsole\0"
 
+/*
+ * Serial output with no console. Run qemu with:
+ *
+ *    -display none -serial mon:stdio
+ *
+ * This will hang or fail to output on the console after exit_boot_services is
+ * called.
+ */
+#define xCONFIG_STD_DEVICES_SETTINGS	"stdin=serial\0" \
+					"stdout=serial\0" \
+					"stderr=serial\0"
+
+#undef CONFIG_BOOTCOMMAND
+
+#define CONFIG_BOOTCOMMAND "part list efi 0"
+
 #endif
Signature · 1 line
2.33.0.153.gba50c8fa24-goog
#80 lore raw
In reply to Simon Glass

Hi Simon,

Quoted 8 lines · Simon Glass

At present U-Boot can be built as an EFI app, but it is really just for testing, with very few features. Instead, the payload build is used for booting on top of UEFI, where U-Boot takes over the machine immediately and supplies its own drivers.

But the app could be made more useful.

Quoted 8 lines

This series provides access to EFI block devices and the video console within U-Boot. It also restructures the app to make it possible to boot a kernel, although further work is needed to implement this.

This can be thought of as making U-Boot perform a little like 'grub', in that it runs purely based on UEFI-provided services.

Of course a lot more work is required, but this is a start.

Thanks for working on this!

Quoted 6 lines

Note: It would be very useful to include qemu tests of the app and stub in CI.

This is available at u-boot-dm/efi-working

Do you have any idea of why this series did not land on patchwork?

Regards,
Bin

#81 lore raw
Quoted 23 lines · Bin Meng

Hi Simon,

Quoted 8 lines · Simon Glass

At present U-Boot can be built as an EFI app, but it is really just for testing, with very few features. Instead, the payload build is used for booting on top of UEFI, where U-Boot takes over the machine immediately and supplies its own drivers.

But the app could be made more useful.

Quoted 8 lines

This series provides access to EFI block devices and the video console within U-Boot. It also restructures the app to make it possible to boot a kernel, although further work is needed to implement this.

This can be thought of as making U-Boot perform a little like 'grub', in that it runs purely based on UEFI-provided services.

Of course a lot more work is required, but this is a start.

Thanks for working on this!

Quoted 1 line
Quoted 7 lines
Quoted 5 lines

Note: It would be very useful to include qemu tests of the app and stub in CI.

This is available at u-boot-dm/efi-working

Do you have any idea of why this series did not land on patchwork?

It did, I've just moved it to "Changes Requested" based on the feedback so far.

Signature · 1 line
Tom