- Concept - U-Boot List

[PATCH 00/22] efi: Improvements for the EFI app on ARM
by Simon Glass 18 Aug '25

18 Aug '25

From: Simon Glass <sjg(a)chromium.org> This series provides a number of minor improvements for the EFI app when running on ARM machines (on top of Tianocore, for example): - Tidy up various comments - Show the ARM exception level in bdinfo - Tidy up the output of 'meminfo' - Get the addr_find command running - Reduce verbosity when not debugging - Show the model at the top of the diplay when using vidconsole - Allow faking the boot right into the actual OS jump (for debugging) Simon Glass (22): arm: bootm: Add some debugging arm: Fix swtiching typo arm: Show the exception level with bdinfo arm: Drop kernel_entry for arm64 meminfo: Allow for up to 10 hex digits event: Add a dummy function for event_notify() boot: Improve debugging in bootm_load_os() boot: Pass flags to the bootm_final event fdt: Add debugging for fdt_simplefb efi: app: Show only a summary of disks and partitions efi: Fix up the addr_find command efi: app: Tidy up some stale comments in setup_memory() efi: app: Allocate pages in any region efi: app: Pick up the SMBIOS table efi: app: Use EFI_PAGE_SHIFT instead of 12 in free_memory() efi: app: Print the final message before freeing memory efi: app: Use the relocated global_data efi: app: Only show the memory map when debugging efi: app: Add a simplefb node to the devicetree efi: arm: Increase the cyclic timeout again efi: app: Show the model when the vidconsole starts efi: app: Enable the cat command arch/arm/lib/bdinfo.c | 7 ++++++ arch/arm/lib/bootm.c | 15 +++++++------ board/efi/efi-arm_app/board.c | 4 ++++ boot/bootm.c | 3 ++- boot/bootm_final.c | 13 ++++++----- boot/fdt_simplefb.c | 4 ++++ cmd/Kconfig | 3 ++- cmd/addr_find.c | 40 +++++++++++---------------------- cmd/meminfo.c | 9 ++++---- configs/efi-arm_app64_defconfig | 5 ++++- include/bootm.h | 13 +---------- include/event.h | 17 +++++++++++++- include/event_decl.h | 27 ++++++++++++++++++++++ lib/Kconfig | 3 ++- lib/efi_client/efi_app.c | 37 +++++++++++++++++++++--------- lib/efi_client/efi_app_init.c | 10 ++++++--- test/cmd/meminfo.c | 2 +- 17 files changed, 138 insertions(+), 74 deletions(-) create mode 100644 include/event_decl.h -- 2.43.0 base-commit: 6268c3c145ff27e9df34e790ec3b612194660733 branch: loadp

2 33

[PATCH v2 00/10] smbios: Refactor the smbios command and parser
by Simon Glass 18 Aug '25

18 Aug '25

From: Simon Glass <sjg(a)chromium.org> The 'smbios' command has its own code for parsing SMBIOS tables. There in also lib/smbios-parser.c which parses tables, although only version 2. This series moved parsing to lib/ and rationalises the code a little. It also adds support for a few more tables. Finally, a hook is added so operation from coreboot can be tested on ellesmere. Changes in v2: - Fix the return values on error Simon Glass (10): hooks: ellesmere: Add a coreboot board hooks: Add ellesmere riscv64_spl board coreboot: Increase the cyclic limit smbios: x86: Enable command by default smbios: Rename smbios_header() smbios: Move some parsing code to smbios-parser.c smbios: Create a function to locate the tables smbios: Pass smbios_info to smbios_next_table() smbios: coreboot: Update smbios_get_header() to use smbios_info smbios: Add support for showing table types 16 and 19 board/coreboot/coreboot/sysinfo.c | 12 +- cmd/Kconfig | 1 + cmd/smbios.c | 184 +++++++++++------- configs/chromebook_link_defconfig | 1 + configs/coreboot_defconfig | 1 + include/smbios.h | 82 +++++++- include/smbios_def.h | 34 ++++ lib/Kconfig | 1 + lib/efi_loader/efi_tcg2.c | 3 +- lib/smbios-parser.c | 143 ++++++++++---- test/hooks/bin/ellesmere/conf.coreboot_qemu | 28 +++ .../bin/ellesmere/conf.qemu-riscv64_spl_na | 1 + 12 files changed, 370 insertions(+), 121 deletions(-) create mode 100644 test/hooks/bin/ellesmere/conf.coreboot_qemu create mode 120000 test/hooks/bin/ellesmere/conf.qemu-riscv64_spl_na -- 2.43.0 base-commit: 711a77d57380f21ddd90ba4884ce0a7ae8f0e591 branch: loado2

1 10

[PATCH 00/10] smbios: Refactor the smbios command and parser
by Simon Glass 18 Aug '25

18 Aug '25

From: Simon Glass <sjg(a)chromium.org> The 'smbios' command has its own code for parsing SMBIOS tables. There in also lib/smbios-parser.c which parses tables, although only version 2. This series moved parsing to lib/ and rationalises the code a little. It also adds support for a few more tables. Finally, a hook is added so operation from coreboot can be tested on ellesmere. Simon Glass (10): hooks: ellesmere: Add a coreboot board hooks: Add ellesmere riscv64_spl board coreboot: Increase the cyclic limit smbios: x86: Enable command by default smbios: Rename smbios_header() smbios: Move some parsing code to smbios-parser.c smbios: Create a function to locate the tables smbios: Pass smbios_info to smbios_next_table() smbios: coreboot: Update smbios_get_header() to use smbios_info smbios: Add support for showing table types 16 and 19 board/coreboot/coreboot/sysinfo.c | 12 +- cmd/Kconfig | 1 + cmd/smbios.c | 184 +++++++++++------- configs/chromebook_link_defconfig | 1 + configs/coreboot_defconfig | 1 + include/smbios.h | 82 +++++++- include/smbios_def.h | 34 ++++ lib/Kconfig | 1 + lib/efi_loader/efi_tcg2.c | 3 +- lib/smbios-parser.c | 143 ++++++++++---- test/hooks/bin/ellesmere/conf.coreboot_qemu | 28 +++ .../bin/ellesmere/conf.qemu-riscv64_spl_na | 1 + 12 files changed, 370 insertions(+), 121 deletions(-) create mode 100644 test/hooks/bin/ellesmere/conf.coreboot_qemu create mode 120000 test/hooks/bin/ellesmere/conf.qemu-riscv64_spl_na -- 2.43.0 base-commit: a633c1d1d2a75ff85620747454847aa87976b897 branch: loado

2 11

[PATCH 00/10] test: Split up the image-creation code
by Simon Glass 17 Aug '25

17 Aug '25

From: Simon Glass <sjg(a)chromium.org> This series moves the image-creation code into separate files so it is easier to find and maintain. Simon Glass (10): test: Move image creation to test/py/img test: Create a common file for image utilities test: Move Fedora image-creation to its own file test: Move Ubuntu image-creation to its own file test: Move Armbian image-creation to its own file test: Move ChromeOS image-creation to its own file test: Move Android image-creation to its own file test: Move EFI image-creation to its own file test: Move the configuration-editor setup to its own file test: Move localboot image-creation to its own file MAINTAINERS | 5 + test/py/img/android.py | 146 ++++++++ test/py/img/armbian.py | 132 +++++++ test/py/img/cedit.py | 25 ++ test/py/img/chromeos.py | 154 +++++++++ test/py/img/common.py | 88 +++++ test/py/img/efi.py | 37 ++ test/py/img/fedora.py | 35 ++ test/py/img/localboot.py | 27 ++ test/py/img/ubuntu.py | 47 +++ test/py/{tests => }/img/vbe.py | 0 test/py/tests/test_ut.py | 616 +-------------------------------- 12 files changed, 705 insertions(+), 607 deletions(-) create mode 100644 test/py/img/android.py create mode 100644 test/py/img/armbian.py create mode 100644 test/py/img/cedit.py create mode 100644 test/py/img/chromeos.py create mode 100644 test/py/img/common.py create mode 100644 test/py/img/efi.py create mode 100644 test/py/img/fedora.py create mode 100644 test/py/img/localboot.py create mode 100644 test/py/img/ubuntu.py rename test/py/{tests => }/img/vbe.py (100%) -- 2.43.0 base-commit: 0d523d8d55dddbd1efd0c734845ceff9a975ff0b branch: makeb

1 10

[PATCH 00/19] test: Move towards using a fixture for test images
by Simon Glass 17 Aug '25

17 Aug '25

From: Simon Glass <sjg(a)chromium.org> There are quite a few test images created by the test system. Most are needed for bootstd tests. At present these are created by a test, called test_ut_dm_init_bootstd() which runs before the test_ut() tests. This means that it is possible to create the images once, then iterate on a test without the images being recreated each time. While this is not a terrible approach, it does have some drawbacks: - If an image is missing or corrupted for some reason, the test fails - The relationship between the images and the tests is not obvious - Images are often created even if the tests to be run do not need them Overall, it seems worthwhile to explore moving to using a fixture for the images. Another problem is that 'make qcheck' and 'make pcheck' are not tested in CI, so can break over time. The latter is a good way of seeing whether each test is independent, e.g. not reliant on files created by another test. Overall this area of U-Boot could use some attention. This series makes a very small start on this: - Update test_fdt_add_pubkey() to work without test_vboot_base() - Fix 'make qcheck' so that it passes - Minor fixes to a few tests - Pass separate fixtures to the image-setup functions instead of the overall ubman fixture Simon Glass (19): test: Disable LTO with make qcheck test: vbe: Free the bootflow after use test: vbe: Avoid using fdt_util to compile a devicetree test: Mark the sleep tests as slow test: Correct implementation of Spawn() in ConsoleSandbox test: Move create_rsa_pair() out of test_vboot_base() test: Update test_fdt_add_pubkey() to create its own files test/py: Provide a run_and_log() function that uses log test/py: Update setup_bootmenu_image() to avoid ubman test/py: Update setup_extlinux_image() to avoid ubman test/py: Update setup_fedora_image() to avoid ubman test/py: Update setup_ubuntu_image() to avoid ubman test/py: Update setup_cros_image() to avoid ubman test/py: Update setup_android_image() to avoid ubman test/py: Update setup_efi_image() to avoid ubman test/py: Update setup_cedit_file() to avoid ubman test/py: Update setup_localboot_image() to avoid ubman test/py: Update setup_vbe_image() to avoid ubman test/py: Update test_ut_dm_init_bootstd() to avoid ubman test/boot/vbe_abrec_os.c | 1 + test/py/console_sandbox.py | 2 +- test/py/tests/img/vbe.py | 57 ++--- test/py/tests/test_android/test_abootimg.py | 37 ++-- test/py/tests/test_sleep.py | 2 + test/py/tests/test_ut.py | 234 +++++++++++--------- test/py/tests/test_vboot.py | 70 +++--- test/py/utils.py | 42 +++- test/run | 3 + 9 files changed, 268 insertions(+), 180 deletions(-) -- 2.43.0 base-commit: a633c1d1d2a75ff85620747454847aa87976b897 branch: makea

1 19

[PATCH 0/5] sandbox: Fix memory corruption around 1MB
by Simon Glass 17 Aug '25

17 Aug '25

From: Simon Glass <sjg(a)chromium.org> After many hours of debugging, it turns out that the PCI EA driver is mapping itself into RAM at 1MB. This happens to be where the kernel ends up, with the vbe_abrec_os bootmeth. Since measurement is enabled, bootm_measure() calls map_sysmem() on the kernel in order to measure it. easurement takes place, although of course using the wrong data. Then, through a strange sequence of events, which I have found very hard to narrow down, the malloc() heap is corrupted. This series provides a fix. To repeat this problem: NO_LTO=1 ./test/py/test.py --bd sandbox --build -k "(ut or vbe) and not efi and not slow and not dm" which dies when running vbe_test_abrec_oem_norun: ... => echo $? 0 => host bind 0 [...]/build-sandbox/persistent-data/vbe1.img => ut -f bootstd vbe_test_abrec_oem_norun Test: bootstd_setup_for_tests: bootstd_common.c common/dlmalloc.c:793: do_check_free_chunk: Assertion `next->prev_size == sz' failed. resetting ... With this series, the above now passes. Simon Glass (5): sandbox: Swap the wanted and got messages in map_physmem() sandbox: Move PCI EA space out of RAM sandbox: Use the exact mapping size for PCI in pmc test: pci: Use the exact mapping size for PCI EA tests sandbox: Abort if a partition memory map is detected arch/sandbox/cpu/mem.c | 13 +++++++++++-- arch/sandbox/include/asm/test.h | 8 ++++---- doc/arch/sandbox/sandbox.rst | 1 + drivers/power/acpi_pmc/sandbox.c | 2 +- test/dm/pci.c | 14 +++++++++----- 5 files changed, 26 insertions(+), 12 deletions(-) -- 2.43.0 base-commit: f85258951dfb3f818e19daeaaddf543ddfc433bc branch: maked

1 5

[PATCH v2 00/19] efi: Support booting an OS with the EFI app on ARM
by Simon Glass 14 Aug '25

14 Aug '25

From: Simon Glass <sjg(a)chromium.org> The ARM EFI app does not successfully boot an OS at present. This series includes patches to allow it to boot an OS via extlinux: The main changes are: - Call exit-boot-services before jumping to the OS - Write the physical-memory info to the FDT - Tidy up some Kconfig options - Improve keyboard compatibility Note that QEMU 8.2 does not boot Ubuntu 24.04 with this series[1]. The kernel dies quite early with: ERROR:target/arm/internals.h:767:regime_is_user: code should not be reached Bail out! ERROR:target/arm/internals.h:767:regime_is_user: code should not be reached QEMU 9.2 resolves this problem. [1] https://lists.opensuse.org/archives/list/bugs@lists.opensuse.org/ message/XVO4M4LDEFOORP5M4ZFINZDX36DWO3G5/ Changes in v2: - Add a comment to serial_efi_get_key() about the 1ms delay Simon Glass (19): treewide: Correct paramters typo scripts: Add an option for the build directory efi: Move efi_store_memory_map() to the stub efi: Adjust serial to work better on real devices efi: Use the same format for all EFI_GUID() declarations efi: Add a few more GUIDs efi: Move memory-map dumping into a common file efi: Add a function to obtain the memory type as a string efi: Run happily without a video display efi: Rename efi_block_device driver efi: Exit EFI boot-services before starting the next app efi: Set the FDT address for QEMU efi: boot: Correct calculation of load address in app efi: Update the ARM app to boot from extlinux efi: arm: Expand app to use 512M of memory efi: arm: Add the memory map to the FDT efi: arm: Increase the cyclic timeout test: Provide a test for the ARM EFI app booting Ubuntu CI: Provide a QEMU instance for efi-arm_app64 .gitlab-ci.yml | 6 + arch/arm/include/asm/arch-mx6/mx6_plugin.S | 2 +- arch/arm/include/asm/arch-mx7/mx7_plugin.S | 2 +- board/efi/efi-arm_app/board.c | 30 ++++ board/efi/efi-arm_app/efi-arm_app.env | 1 + boot/bootm.c | 2 + cmd/efi.c | 119 +------------- configs/efi-arm_app64_defconfig | 4 + drivers/ddr/fsl/options.c | 2 +- drivers/ddr/marvell/axp/ddr3_spd.c | 2 +- drivers/serial/serial_efi.c | 37 ++++- drivers/video/efi.c | 8 +- include/efi.h | 46 +++++- include/efi_api.h | 178 +++++++++++---------- include/image.h | 2 +- lib/efi/Makefile | 1 + lib/efi/memory.c | 132 +++++++++++++++ lib/efi_client/efi.c | 48 ------ lib/efi_client/efi_app.c | 64 +++++++- lib/efi_client/stub.c | 57 +++++++ lib/efi_driver/efi_block_device.c | 2 +- lib/uuid.c | 4 + scripts/build-efi | 19 ++- scripts/build_helper.py | 1 + test/py/tests/test_distro.py | 22 ++- 25 files changed, 517 insertions(+), 274 deletions(-) create mode 100644 lib/efi/memory.c -- 2.43.0 base-commit: 97a083c86cfaab2268dc8a0f90aedbac6d419738 branch: loadn2

1 19

[PATCH 00/14] RFC: measure FIT signing keys with the TPM
by Ludwig Nussel 13 Aug '25

13 Aug '25

(The fork button on the gitlab instance is greyed out and I don't have privileged to push into a branch of u-boot so trying mail) This is a proof of concept for measured boot without EFI. The idea is to measure the RSA keys contained in the control FDT into PCR#7. The Linux system can then bind it's full disk encryption to that value. Assuming that everything up to and including u-boot is trusted, a devices' original disk content is only accessible then with the same set of trusted keys active. When adding more or different keys (not sure how yet) a device could still be booted with custom software. The most relevant patches are the three ones with exclamation mark in the list below. The others are random fixes Feedback welcome. Ludwig Nussel (14): arm: qemu: fix 64bit initrd_high and fdt_high bootm: treat absent load address as zero vsnprintf: add hex string dump format modifier iminfo: also verify signatures ! qemu: overlay signature nodes ! tpm: allocate tpm event log if missing ! boot: measure FIT signing keys with the TPM imx8mm: add ftpm to device tree if enabled tools: imx8image: implement printing header test: fix build with enabled CONFIG_MEASURED_BOOT sandbox: fix booting host images bootm: add measure subcommand mkimage: define log_err and log_info image-fit-sig: require signatures arch/arm/dts/imx8mm-u-boot.dtsi | 5 + board/emulation/qemu-arm/qemu-arm.c | 30 +++++- board/emulation/qemu-arm/qemu-arm.env | 5 + boot/Kconfig | 8 ++ boot/bootm.c | 93 +++++++++++++++--- boot/image-fdt.c | 65 +++++++++++++ boot/image-fit-sig.c | 15 ++- boot/image-fit.c | 31 ++++++ boot/image-pre-load.c | 3 - cmd/bootm.c | 8 ++ doc/develop/printf.rst | 4 + drivers/block/Makefile | 2 +- drivers/block/host_bootdev.c | 40 ++++++++ fs/sandbox/Makefile | 2 +- .../{host_bootdev.c => sandboxfs_bootdev.c} | 18 ++-- include/image.h | 1 + include/spbuf.h | 28 ++++++ include/tpm-common.h | 1 + include/tpm_tcg2.h | 4 + lib/Makefile | 2 + lib/spbuf.c | 64 ++++++++++++ lib/tpm_tcg2.c | 97 +++++++++++++++---- lib/vsprintf.c | 18 ++++ test/boot/measurement.c | 1 + tools/imx8image.c | 93 +++++++++++++++++- tools/mkimage.h | 2 + 26 files changed, 582 insertions(+), 58 deletions(-) create mode 100644 drivers/block/host_bootdev.c rename fs/sandbox/{host_bootdev.c => sandboxfs_bootdev.c} (65%) create mode 100644 include/spbuf.h create mode 100644 lib/spbuf.c -- 2.34.1

1 14

[PATCH 00/19] efi: Support booting an OS with the EFI app on ARM
by Simon Glass 12 Aug '25

12 Aug '25

From: Simon Glass <sjg(a)chromium.org> The ARM EFI app does not successfully boot an OS at present. This series includes patches to allow it to boot an OS via extlinux: The main changes are: - Call exit-boot-services before jumping to the OS - Write the physical-memory info to the FDT - Tidy up some Kconfig options - Improve keyboard compatibility Note that QEMU 8.2 does not boot Ubuntu 24.04 with this series[1]. The kernel dies quite early with: ERROR:target/arm/internals.h:767:regime_is_user: code should not be reached Bail out! ERROR:target/arm/internals.h:767:regime_is_user: code should not be reached QEMU 9.2 resolves this problem. [1] https://lists.opensuse.org/archives/list/bugs@lists.opensuse.org/ message/XVO4M4LDEFOORP5M4ZFINZDX36DWO3G5/ Simon Glass (19): treewide: Correct paramters typo scripts: Add an option for the build directory efi: Move efi_store_memory_map() to the stub efi: Adjust serial to work better on real devices efi: Use the same format for all EFI_GUID() declarations efi: Add a few more GUIDs efi: Move memory-map dumping into a common file efi: Add a function to obtain the memory type as a string efi: Run happily without a video display efi: Rename efi_block_device driver efi: Exit EFI boot-services before starting the next app efi: Set the FDT address for QEMU efi: boot: Correct calculation of load address in app efi: Update the ARM app to boot from extlinux efi: arm: Expand app to use 512M of memory efi: arm: Add the memory map to the FDT efi: arm: Increase the cyclic timeout test: Provide a test for the ARM EFI app booting Ubuntu CI: Provide a QEMU instance for efi-arm_app64 .gitlab-ci.yml | 6 + arch/arm/include/asm/arch-mx6/mx6_plugin.S | 2 +- arch/arm/include/asm/arch-mx7/mx7_plugin.S | 2 +- board/efi/efi-arm_app/board.c | 30 ++++ board/efi/efi-arm_app/efi-arm_app.env | 1 + boot/bootm.c | 2 + cmd/efi.c | 119 +------------- configs/efi-arm_app64_defconfig | 4 + drivers/ddr/fsl/options.c | 2 +- drivers/ddr/marvell/axp/ddr3_spd.c | 2 +- drivers/serial/serial_efi.c | 29 +++- drivers/video/efi.c | 8 +- include/efi.h | 46 +++++- include/efi_api.h | 178 +++++++++++---------- include/image.h | 2 +- lib/efi/Makefile | 1 + lib/efi/memory.c | 132 +++++++++++++++ lib/efi_client/efi.c | 48 ------ lib/efi_client/efi_app.c | 64 +++++++- lib/efi_client/stub.c | 57 +++++++ lib/efi_driver/efi_block_device.c | 2 +- lib/uuid.c | 4 + scripts/build-efi | 19 ++- scripts/build_helper.py | 1 + test/py/tests/test_distro.py | 22 ++- 25 files changed, 509 insertions(+), 274 deletions(-) create mode 100644 lib/efi/memory.c -- 2.43.0 base-commit: 97a083c86cfaab2268dc8a0f90aedbac6d419738 branch: loadn

1 19

[PATCH 00/11] boot: Allow lmb-allocation of compression buffer with booti
by Simon Glass 11 Aug '25

11 Aug '25

From: Simon Glass <sjg(a)chromium.org> ARM 'booti' images (as created by Linux 'make Image') typically need to be compressed. At present the decompression buffer is described by a pair of environment variables: kernel_comp_addr_r and kernel_comp_size This is quite inflexible, since the buffer size and location must be pre-set in the environment, before the OS size is known. For the EFI app, it is quite difficult to work with a fixed buffer. It is not in control of the memory map so it may be that the requested buffer address is not available. This series updates the boot implementation to allow the variables to be omitted, with lmb allocations done in this case. This resolves the problem in the EFI app, with kernels which come from extlinux.conf When the EFI app loads another app, a similar problem is present, but with the EFI bootmeth, also resolved in this series. Simon Glass (11): efi: Allocate the kernel buffer when running in the app boot: Add more debugging to bootm code boot: Add more debugging to image-baord and image-fdt boot: Read environment variables at start boot: Move compression-buffer handling to its own function boot: Add an OS-size field in bootm_info boot: Record the kernel size in pxe_utils boot: Check OS size when decompressing with booti boot: Support lmb reservation in resolve_os_comp_buf() boot: Correct a comment in image_setup_libfdt() boot: Emit an event just before starting the OS boot/bootm.c | 174 +++++++++++++++++++++++++++++++++++--------- boot/bootm_final.c | 10 +++ boot/bootmeth_efi.c | 25 ++++++- boot/image-board.c | 16 +++- boot/image-fdt.c | 5 +- boot/pxe_utils.c | 23 ++++-- cmd/booti.c | 1 + cmd/bootm.c | 1 + common/event.c | 1 + include/bootm.h | 18 +++++ include/event.h | 8 ++ include/pxe_utils.h | 2 + 12 files changed, 238 insertions(+), 46 deletions(-) -- 2.43.0 base-commit: c9dbf55855b3b0ca143e71e9204b74fc88117808 branch: loadm

1 11