November 2023 - U-Boot

[BUG] Video does not work as expected on efi-x86_app64.defconfig
by Heinrich Schuchardt 21 Nov '23

21 Nov '23

Hello Simon, I have built efi-x86_app64_defconfig on v2024.01-rc3. On my laptop the video output appears twice. See https://gist.githubusercontent.com/xypron/24ceedb749a68031790864c3a35c5f30/… Are we accessing the video buffer directly instead of using the EFI API? Best regards Heinrich

2 1

[PATCH v3] smbios: arm64: Ensure table is written at a good address
by Simon Glass 21 Nov '23

21 Nov '23

U-Boot typically sets up its malloc() pool near the top of memory. On ARM64 systems this can result in an SMBIOS table above 4GB which is not supported by SMBIOSv2. Work around this problem by automatically choosing an address below 4GB (but as high as possible), if needed. Tell efi_loader about the address in the same way as we do for ACPI tables. Signed-off-by: Simon Glass <sjg(a)chromium.org> --- Please see previous discussion here: https://patchwork.ozlabs.org/project/uboot/patch/ 20231025123117.v2.1.I054cad60e00f8cfde39256c7c7d9c960987fb9be@changeid/ Changes in v3: - Add RISC-V to the condition, too - Update the commit message Changes in v2: - Update to search for a suitable area automatically, if enabled lib/Kconfig | 13 ++++++++ lib/efi_loader/efi_smbios.c | 63 ++++++++++++++++++++++++++++++++++++- 2 files changed, 75 insertions(+), 1 deletion(-) diff --git a/lib/Kconfig b/lib/Kconfig index 19649517a39b..c70faeeebed5 100644 --- a/lib/Kconfig +++ b/lib/Kconfig @@ -998,6 +998,19 @@ config GENERATE_SMBIOS_TABLE See also SMBIOS_SYSINFO which allows SMBIOS values to be provided in the devicetree. +config SMBIOS_TABLE_FIXED + bool "Place the SMBIOS table at a special address" + depends on GENERATE_SMBIOS_TABLE && SMBIOS && EFI_LOADER + depends on ARM64 || RISCV + default y + help + Use this option to place the SMBIOS table at a special address. + + U-Boot typically sets up its malloc() pool near the top of memory. On + ARM64 systems this can result in an SMBIOS table above 4GB which is + not supported by SMBIOSv2. This option works around this problem by + chosing an address just below 4GB, if needed. + endmenu config LIB_RATIONAL diff --git a/lib/efi_loader/efi_smbios.c b/lib/efi_loader/efi_smbios.c index 48446f654d9b..566f206af5b6 100644 --- a/lib/efi_loader/efi_smbios.c +++ b/lib/efi_loader/efi_smbios.c @@ -47,6 +47,60 @@ efi_status_t efi_smbios_register(void) map_sysmem(addr, 0)); } +/** + * find_addr_below() - Find a usable region below the given max_addr + * + * Check if *addrp is suitable to define a memory region which finishes below + * @max_addr + @req_size. If so, do nothing and return 0 + * + * As a backup, if CONFIG_SMBIOS_TABLE_FIXED is enabled, search for a + * 4KB-aligned DRAM region which is large enough. Make sure it is below U-Boot's + * stack space, assumed to be 64KB. + * + * @max_addr: Maximum address that can be used (region must finish before here) + * @req:size: Required size of region + * @addrp: On entry: Current proposed address; on exit, holds the new address, + * on success + * Return 0 if OK (existing region was OK, or a new one was found), -ENOSPC if + * nothing suitable was found + */ +static int find_addr_below(ulong max_addr, ulong req_size, ulong *addrp) +{ + struct bd_info *bd = gd->bd; + ulong max_base; + int i; + + max_base = max_addr - req_size; + if (*addrp <= max_base) + return 0; + + if (!IS_ENABLED(CONFIG_SMBIOS_TABLE_FIXED)) + return -ENOSPC; + + /* Make sure that the base is at least 64KB below the stack */ + max_base = min(max_base, + ALIGN(gd->start_addr_sp - SZ_64K - req_size, SZ_4K)); + + for (i = CONFIG_NR_DRAM_BANKS - 1; i >= 0; i--) { + ulong start = bd->bi_dram[i].start; + ulong size = bd->bi_dram[i].size; + ulong addr; + + /* chose an address at most req_size bytes before the end */ + addr = min(max_base, start - req_size + size); + + /* check this is in the range */ + if (addr >= start && addr + req_size < start + size) { + *addrp = addr; + log_warning("Forcing SMBIOS table to address %lx\n", + addr); + return 0; + } + } + + return -ENOSPC; +} + static int install_smbios_table(void) { ulong addr; @@ -61,7 +115,14 @@ static int install_smbios_table(void) return log_msg_ret("mem", -ENOMEM); addr = map_to_sysmem(buf); - if (!write_smbios_table(addr)) { + + /* + * Deal with a fixed address if needed. For simplicity we assume that + * the SMBIOS-table size is <64KB. If a suitable address cannot be + * found, then write_smbios_table() returns an error. + */ + if (find_addr_below(SZ_4G - 1, SZ_64K, &addr) || + !write_smbios_table(addr)) { log_err("Failed to write SMBIOS table\n"); return log_msg_ret("smbios", -EINVAL); } -- 2.43.0.rc1.413.gea7ed67945-goog

2 1

[PATCH v2 0/5] video: Improve syncing performance with cyclic
by Simon Glass 21 Nov '23

21 Nov '23

Now that U-Boot has a background-processing feature, it is possible to reduce the amount of 'foreground' syncing of the display. At present this happens quite often. Foreground syncing blocks all other processing, sometimes for 10ms or more. When pasting commands into U-Boot over the UART, this typically result in characters being dropped. For example, on rpi_4 it isn't possible to paste in more than 35 characters before things fail. This makes updating the environment or entering long commands very painful over the console, since text must be pasted in chunks, or the vidconsole device must be dropped from stdout. This series introduces background syncing, enabled by default for boards which use video. The sync rates for foreground and background are configurable. With this series it is possible to paste in any amount of text to the command line. Some sandbox-specific workarounds can now be removed and sandbox video (./u-boot -Dl) is significantly more responsive. This obviously increases code size, since it enables a subsystem not normally used by default. However it only applies to boards which have VIDEO enabled, which are presumably less worried about memory space since the video code is fairly large. Also it is possible to disable CMD_CYCLIC and reduce the growth to: aarch64: (for 1/1 boards) all +1081.0 rodata +65.0 text +1016.0 arm: (for 1/1 boards) all +945.0 rodata +65.0 text +880.0 Without that, the increase doubles. It is of course possible to disable CYCLIC and still use VIDEO but this reverts to the current behaviour Changes in v2: - Add an SPL_CYCLIC symbol - Add a lot more explanation about the header files - Expand help for CONFIG_VIDEO - Fix 'groth' and 'work-around' typos in cover letter Simon Glass (5): cyclic: Add a symbol for SPL video: Move last_sync to private data video: Use cyclic to handle video sync sandbox: Increase cyclic CPU-time limit sandbox: Drop video-sync in serial driver common/Kconfig | 9 ++++++ common/Makefile | 2 +- drivers/serial/sandbox.c | 2 -- drivers/video/Kconfig | 35 +++++++++++++++++++++ drivers/video/video-uclass.c | 52 +++++++++++++++++++++++++------ drivers/watchdog/Kconfig | 1 + include/asm-generic/global_data.h | 2 +- include/cyclic.h | 6 ++-- include/video.h | 2 ++ 9 files changed, 96 insertions(+), 15 deletions(-) -- 2.43.0.rc1.413.gea7ed67945-goog

2 9

[PATCH 00/29] bootm: Refactoring to reduce reliance on CMDLINE (part A)
by Simon Glass 21 Nov '23

21 Nov '23

It would be useful to be able to boot an OS when CONFIG_CMDLINE is disabled. This could allow reduced code size. Standard boot provides a way to handle programmatic boot, without scripts, so such a feature is possible. The main impediment is the inability to use the booting features of U-Boot without a command line. So the solution is to avoid passing command arguments and the like to code in boot/ A similar process has taken place with filesystems, for example, where we have (somewhat) separate Kconfig options for the filesystem commands and the filesystems themselves. This series starts the process of refactoring the bootm logic so that it can be called from standard boot without using the command line. Mostly it removes the use of argc, argv and cmdtbl from the internal logic. Some limited tidy-up is included, but this is kept to smaller patches, rather than trying to remove all #ifdefs etc. Some function comments are added, however. A simple programmatic boot is provided as a starting point. This work will likely take many series, so this is just the start. Size growth with this series for firefly-rk3288 (Thumb2) is: arm: (for 1/1 boards) all +23.0 rodata -49.0 text +72.0 This should be removed by: https://source.denx.de/u-boot/custodians/u-boot-dm/-/issues/11 but it is not included in this series as it is already large enough. No functional change is intended in this series. Simon Glass (29): arm: x86: Drop discarding of command linker-lists mmc: env: Unify the U_BOOT_ENV_LOCATION conditions treewide: Tidy up semicolon after command macros bootstd: Add missing header file from bootdev.h bootstd: Introduce programmable boot bootm: Drop arguments from bootm_start() bootm: Simplify arguments for bootm_pre_load() bootm: Move boot_get_kernel() higher in the file image: Tidy up genimg_get_kernel_addr_fit() bootm: Reduce arguments to boot_get_kernel() image: Document error codes from fit_image_load() bootm: Adjust boot_get_kernel() to return an error bootm: Use the error return from boot_get_kernel() bootstage: Drop BOOTSTAGE_ID_FIT_KERNEL_INFO bootm: Move error printing out of boot_get_kernel() bootm: Reduce arguments to boot_find_os() bootm: Reduce arguments to boot_get_ramdisk() fdt: Allow use of fdt_support inside if() statements bootm: Drop #ifdef in bootm_find_images() bootm: Pass image buffer to boot_get_fdt() bootm: Reduce arguments to boot_get_fdt() bootm: Reduce arguments to boot_get_fpga() bootm: Reduce arguments to boot_get_loadables() bootm: Simplify Android ramdisk addr in bootm_find_images() bootm: efi: Drop special call to bootm_find_other() bootm: optee: Drop special call to bootm_find_other() bootm: Adjust the parameters of bootm_find_images() bootm: Add a function to check overlap bootm: Reduce arguments to bootm_find_other() arch/arm/cpu/u-boot.lds | 3 - arch/x86/cpu/u-boot-64.lds | 4 - arch/x86/cpu/u-boot-spl.lds | 4 - arch/x86/cpu/u-boot.lds | 4 - board/freescale/common/vid.c | 2 +- board/xilinx/common/fru.c | 2 +- board/xilinx/versal/cmds.c | 2 +- board/xilinx/zynqmp/cmds.c | 2 +- boot/Kconfig | 11 + boot/Makefile | 2 + boot/bootm.c | 576 +++++++++++++++++++---------------- boot/bootm_os.c | 16 - boot/image-board.c | 67 +--- boot/image-fdt.c | 39 +-- boot/prog_boot.c | 51 ++++ cmd/booti.c | 4 +- cmd/bootz.c | 4 +- cmd/btrfs.c | 2 +- cmd/eeprom.c | 2 +- cmd/ext2.c | 4 +- cmd/fs.c | 8 +- cmd/pinmux.c | 2 +- cmd/qfw.c | 2 +- common/main.c | 9 + env/mmc.c | 2 +- include/bootdev.h | 1 + include/bootm.h | 26 +- include/bootstage.h | 1 - include/bootstd.h | 9 + include/command.h | 2 +- include/fdt_support.h | 5 +- include/image.h | 127 ++++++-- 32 files changed, 550 insertions(+), 445 deletions(-) create mode 100644 boot/prog_boot.c -- 2.42.0.869.gea05f2083d-goog

3 59

[RFC PATCH 0/2] rpi5: initial support
by Dmitry Malkin 21 Nov '23

21 Nov '23

I was not able to boot rpi5 with the latest master. I run into dram_init() error: > DRAM: mbox: Timeout waiting for response > bcm2835: Could not query ARM memory size Apparently there is a new peripheral layout and ARMC MBOX has a different offset within. Turns out there is a compatible node "brcm,bcm2835-mbox" for all RPIs (bcm238x/bcm271x). After fixing this I got to the point where mmu_setup() was hanging (looks like it was due to incorrect mapping). So I've added a definition of first gigabyte of RAM and SOC range for bcm2712 into rpi.c. With those two patches and rpi_arm64_defconfig (with disabled CONFIG_VIDEO_BCM2835) I'm able to get a working u-boot prompt (over serial). FDT comes from firmware (EEPROM). These patches are based on v2024.01-rc2 and tested with RPI5 8GB. Known issues: * CONFIG_VIDEO_BCM2835 doesn't work. * MBOX get board revision (0x10002) doesn't work (the response status is 0x8000_0001). Dmitry Malkin (2): rpi5: add alternative way to get MBOX address via FDT node rpi5: add initial memory map for bcm2712 arch/arm/mach-bcm283x/include/mach/base.h | 1 + arch/arm/mach-bcm283x/include/mach/mbox.h | 4 ++-- arch/arm/mach-bcm283x/init.c | 29 ++++++++++++++++++++++- 3 files changed, 31 insertions(+), 3 deletions(-) -- 2.40.1

2 1

[PATCH] doc: sending_patches.rst: s/Superseeded/Superseded
by Mattijs Korpershoek 21 Nov '23

21 Nov '23

This is a common typo listed in scripts/spelling.txt. Fix it to match the patchwork status, which is superseded. Signed-off-by: Mattijs Korpershoek <mkorpershoek(a)baylibre.com> --- doc/develop/sending_patches.rst | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/doc/develop/sending_patches.rst b/doc/develop/sending_patches.rst index ba73d0d11b46..5a6962f1021c 100644 --- a/doc/develop/sending_patches.rst +++ b/doc/develop/sending_patches.rst @@ -363,7 +363,7 @@ A Custodian has additional privileges and can: * Awaiting Upstream - * Superseeded + * Superseded * Deferred @@ -399,7 +399,7 @@ today. Not all states are used by all custodians. and has not merged yet to master, or has queued the patch up to be submitted to be merged, but has not yet. -* Superseeded: Patches are marked as 'superseeded' when the poster submits a +* Superseded: Patches are marked as 'superseded' when the poster submits a new version of these patches. * Deferred: Deferred usually means the patch depends on something else that --- base-commit: 9e53e45292ee2f1d9d2ccc59914b161bef9b10d7 change-id: 20231121-doc-superseeded-d1ef005cda7e Best regards, -- Mattijs Korpershoek <mkorpershoek(a)baylibre.com>

2 1

[v2] doc: clang: Update and correct support notes
by Tom Rini 21 Nov '23

21 Nov '23

At this point, clang can be used on both 32bit and 64bit targets without issue. Make note of logic we have that will inform clang of the architecture to build for. Signed-off-by: Tom Rini <trini(a)konsulko.com> --- Changes in v2: - Reword the segment about CROSS_COMPILE and -target per Heinrich's suggestion. --- doc/build/clang.rst | 22 ++++++++++------------ 1 file changed, 10 insertions(+), 12 deletions(-) diff --git a/doc/build/clang.rst b/doc/build/clang.rst index cc265506c2fb..09bb988e9236 100644 --- a/doc/build/clang.rst +++ b/doc/build/clang.rst @@ -11,14 +11,6 @@ The ARM backend can be instructed not to use the r9 and x18 registers using supported inline assembly is needed to get and set the r9 or x18 value. This leads to larger code then strictly necessary, but at least works. -**NOTE:** target compilation only work for _some_ ARM boards at the moment. -Also AArch64 is not supported currently due to a lack of private libgcc -support. Boards which reassign gd in c will also fail to compile, but there is -in no strict reason to do so in the ARM world, since crt0.S takes care of this. -These assignments can be avoided by changing the init calls but this is not in -mainline yet. - - Debian based ------------ @@ -28,14 +20,20 @@ Required packages can be installed via apt, e.g. sudo apt-get install clang -Note that we still use binutils for some tools so we must continue to set -CROSS_COMPILE. To compile U-Boot with Clang on Linux without IAS use e.g. +We make use of the CROSS_COMPILE variable to derive the build target which is +passed as the --target parameter to clang. + +The CROSS_COMPILE variable further determines the paths to other build +tools. As assembler we use the binary pointed to by '$(CROSS_COMPILE)as' +instead of the LLVM integrated assembler (IAS). + +Here is an example demonstrating building U-Boot for the Raspberry Pi 2 +using clang: .. code-block:: bash make HOSTCC=clang rpi_2_defconfig - make HOSTCC=clang CROSS_COMPILE=arm-linux-gnueabi- \ - CC="clang -target arm-linux-gnueabi" -j8 + make HOSTCC=clang CROSS_COMPILE=arm-linux-gnueabi- CC=clang -j8 It can also be used to compile sandbox: -- 2.34.1

2 1

[PATCH 0/1] Fix DFU updates on the zynqmp platform
by Jared Baur 21 Nov '23

21 Nov '23

When using DFU alt info via EFI capsule updates on an Ultra96 V2 board, `dfu_alt_info` is automatically set to write a `boot.bin` file to the SD card when in SD boot mode and not using multiboot. After the update is applied, the board reboots and does not come back up. When writing an upper-case `BOOT.BIN` file to the SD card instead, the board updates and comes back up successfully. Jared Baur (1): Write upper-case BOOT.BIN to zynqmp MMC board/xilinx/zynqmp/zynqmp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) -- 2.42.0

2 2

Pull request efi-2024-01-rc4
by Heinrich Schuchardt 21 Nov '23

21 Nov '23

Dear Tom, The following changes since commit 24ca49b33af98d54d6cd2e845f071f6565345ffd: Prepare v2024.01-rc3 (2023-11-20 08:43:46 -0500) are available in the Git repository at: https://source.denx.de/u-boot/custodians/u-boot-efi.git tags/efi-2024-01-rc4 for you to fetch changes up to 64658007f3c61e32bf241f3048a23b1728774d57: cmd: eficonfig: create shortened boot options (2023-11-20 19:06:22 +0100) Gitlab CI showed no issues: https://source.denx.de/u-boot/custodians/u-boot-efi/-/pipelines/18655 ---------------------------------------------------------------- Pull request efi-2024-01-rc4 Documentation * Add HiSilicon board documentation to HTML docs * Fix building with Sphinx 6.0 UEFI * Increase default variable store size to 128K * Check return value of efi_append_scrtm version * Create shortened boot options in eficonfig command Other * Avoid incorrect error message in mkimage ---------------------------------------------------------------- Heinrich Schuchardt (5): mkimage: do not write incorrect error message doc: add HiSilicon board documentation to HTML docs doc: typo fdtaddr_addr_r efi_loader: improve efi_var_from_file() description cmd: eficonfig: create shortened boot options Ilias Apalodimas (2): efi_loader: Increase default variable store size to 128K efi_loader: Correctly account the SCRTM event creation Jonathan Corbet (1): docs: Fix the docs build with Sphinx 6.0 board/hisilicon/hikey/MAINTAINERS | 1 + board/hisilicon/hikey/README | 227 -------------------------- board/hisilicon/hikey960/MAINTAINERS | 1 + board/hisilicon/hikey960/README | 247 ---------------------------- board/hisilicon/poplar/MAINTAINERS | 1 + board/hisilicon/poplar/README | 288 --------------------------------- cmd/eficonfig.c | 5 +- doc/board/hisilicon/hikey.rst | 261 ++++++++++++++++++++++++++++++ doc/board/hisilicon/hikey960.rst | 284 ++++++++++++++++++++++++++++++++ doc/board/hisilicon/index.rst | 11 ++ doc/board/hisilicon/poplar.rst | 302 +++++++++++++++++++++++++++++++++++ doc/board/index.rst | 1 + doc/develop/bootstd.rst | 2 +- doc/sphinx/load_config.py | 6 +- lib/efi_loader/Kconfig | 14 +- lib/efi_loader/efi_tcg2.c | 5 +- lib/efi_loader/efi_var_file.c | 7 +- tools/renesas_spkgimage.c | 3 +- 18 files changed, 890 insertions(+), 776 deletions(-) delete mode 100644 board/hisilicon/hikey/README delete mode 100644 board/hisilicon/hikey960/README delete mode 100644 board/hisilicon/poplar/README create mode 100644 doc/board/hisilicon/hikey.rst create mode 100644 doc/board/hisilicon/hikey960.rst create mode 100644 doc/board/hisilicon/index.rst create mode 100644 doc/board/hisilicon/poplar.rst

2 1

[PATCH v3 0/3] Improve the experience with USB gadgets
by Miquel Raynal 21 Nov '23

21 Nov '23

Try to ease the use of USB gadgets for people not fully aware of all the terms and constraints. With more helpful error messages a little bit of guidance anyone should be able to adapt to the bind/unbind game that is now necessary with USB gadgets. The idea is to eventually get proper USB composite gadget support, but that is a longer term goal. Changes in v3: * First patch is new. * Collected the tags received. * Imply CMD_DM from CMD_BIND. Miquel Raynal (3): cmd: Change the dependencies between CMD_BIND and USB_GADGET cmd: bind: Try to improve the (un)bind help usb: udc: Try to clarify an error message cmd/Kconfig | 2 +- cmd/bind.c | 4 ++++ drivers/usb/gadget/Kconfig | 1 + drivers/usb/gadget/udc/udc-core.c | 13 ++++++++++++- 4 files changed, 18 insertions(+), 2 deletions(-) -- 2.34.1

2 5