November 2023 - U-Boot

[RFC PATCH 1/2] rpi5: initial support
by Dmitry Malkin 18 Nov '23

18 Nov '23

rpi5: add alternative way to get MBOX address via FDT node MBOX on RPI5/bcm2712 has a different offset. Find it via the "brcm,bcm2835-mbox" node. --- arch/arm/mach-bcm283x/include/mach/base.h | 1 + arch/arm/mach-bcm283x/include/mach/mbox.h | 4 ++-- arch/arm/mach-bcm283x/init.c | 8 +++++++- 3 files changed, 10 insertions(+), 3 deletions(-) diff --git a/arch/arm/mach-bcm283x/include/mach/base.h b/arch/arm/mach-bcm283x/include/mach/base.h index 4ccaf69693..8cd224921e 100644 --- a/arch/arm/mach-bcm283x/include/mach/base.h +++ b/arch/arm/mach-bcm283x/include/mach/base.h @@ -7,6 +7,7 @@ #define _BCM283x_BASE_H_ extern unsigned long rpi_bcm283x_base; +extern unsigned long rpi_bcm283x_mbox_addr; #ifdef CONFIG_ARMV7_LPAE #ifdef CONFIG_TARGET_RPI_4_32B diff --git a/arch/arm/mach-bcm283x/include/mach/mbox.h b/arch/arm/mach-bcm283x/include/mach/mbox.h index 490664f878..ad086be467 100644 --- a/arch/arm/mach-bcm283x/include/mach/mbox.h +++ b/arch/arm/mach-bcm283x/include/mach/mbox.h @@ -38,8 +38,8 @@ /* Raw mailbox HW */ -#define BCM2835_MBOX_PHYSADDR ({ BUG_ON(!rpi_bcm283x_base); \ - rpi_bcm283x_base + 0x0000b880; }) +#define BCM2835_MBOX_PHYSADDR ({ BUG_ON(!rpi_bcm283x_mbox_addr); \ + rpi_bcm283x_mbox_addr; }) struct bcm2835_mbox_regs { u32 read; diff --git a/arch/arm/mach-bcm283x/init.c b/arch/arm/mach-bcm283x/init.c index 7265faf6ce..d76c8e5eb2 100644 --- a/arch/arm/mach-bcm283x/init.c +++ b/arch/arm/mach-bcm283x/init.c @@ -116,6 +116,7 @@ static void rpi_update_mem_map(void) {} #endif unsigned long rpi_bcm283x_base = 0x3f000000; +unsigned long rpi_bcm283x_mbox_addr; int arch_cpu_init(void) { @@ -126,7 +127,7 @@ int arch_cpu_init(void) int mach_cpu_init(void) { - int ret, soc_offset; + int ret, soc_offset, mbox_offset; u64 io_base, size; rpi_update_mem_map(); @@ -142,6 +143,11 @@ int mach_cpu_init(void) return ret; rpi_bcm283x_base = io_base; + rpi_bcm283x_mbox_addr = rpi_bcm283x_base + 0xb880; + + mbox_offset = fdt_node_offset_by_compatible((void*)gd->fdt_blob, soc_offset, "brcm,bcm2835-mbox"); + if (mbox_offset > 0) + rpi_bcm283x_mbox_addr = fdt_get_base_address((void*)gd->fdt_blob, mbox_offset); return 0; } -- 2.40.1

1 0

[PATCH 1/1] cmd: eficonfig: create shortened boot options
by Heinrich Schuchardt 18 Nov '23

18 Nov '23

The boot options created by eficonfig should use shortened device-paths to avoid problems if drives are enumerated in a different sequence. Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt(a)canonical.com> --- cmd/eficonfig.c | 1 + 1 file changed, 1 insertion(+) diff --git a/cmd/eficonfig.c b/cmd/eficonfig.c index e6e8a0a488..3d95092109 100644 --- a/cmd/eficonfig.c +++ b/cmd/eficonfig.c @@ -528,6 +528,7 @@ struct efi_device_path *eficonfig_create_device_path(struct efi_device_path *dp_ p += fp_size; *((struct efi_device_path *)p) = END; + dp_volume = efi_dp_shorten(dp_volume); dp = efi_dp_append(dp_volume, (struct efi_device_path *)buf); free(buf); -- 2.40.1

3 3

[PATCH v2 00/32] bootm: Refactoring to reduce reliance on CMDLINE (part A)
by Simon Glass 18 Nov '23

18 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. Changes in v2: - Add new patch to adjust position of unmap_sysmem() in boot_get_kernel() - Add new patch to obtain command arguments - Fix 'boot_find_os' typo - Pass in the command name - Use the command table to provide the command name, instead of "bootm" Simon Glass (32): arm: x86: Drop discarding of command linker-lists README: Correct docs for CONFIG_SPL_BUILD 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: Adjust position of unmap_sysmem() in boot_get_kernel() 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 bootm_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() RFC: command: Introduce functions to obtain command arguments README | 26 +- 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 | 577 +++++++++++++++++++---------------- 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/disk.c | 4 +- cmd/eeprom.c | 2 +- cmd/ext2.c | 4 +- cmd/fs.c | 8 +- cmd/fuse.c | 2 +- cmd/mmc.c | 2 +- cmd/pinmux.c | 2 +- cmd/qfw.c | 2 +- common/main.c | 9 + drivers/misc/gsc.c | 6 +- env/mmc.c | 2 +- fs/fs.c | 4 +- include/bootdev.h | 1 + include/bootm.h | 26 +- include/bootstage.h | 1 - include/bootstd.h | 9 + include/command.h | 35 ++- include/fdt_support.h | 5 +- include/image.h | 127 ++++++-- test/cmd_ut.c | 2 +- 39 files changed, 612 insertions(+), 463 deletions(-) create mode 100644 boot/prog_boot.c -- 2.43.0.rc0.421.g78406f8d94-goog

3 42

[PATCH v5 0/3] BOOTP/DHCPv4 enhancements
by seanedmond＠linux.microsoft.com 18 Nov '23

18 Nov '23

From: Sean Edmond <seanedmond(a)microsoft.com> In our datacenter application, a single DHCP server is servicing 36000+ clients. Improvements are required to the DHCPv4 retransmission behavior to align with RFC and ensure less pressure is exerted on the server: - retransmission backoff interval maximum is configurable (environment variable bootpretransmitperiodmax) - initial retransmission backoff interval is configurable (environment variable bootpretransmitperiodinit) - transaction ID is kept the same for each BOOTP/DHCPv4 request (not recreated on each retry) For our application we'll use: - bootpretransmitperiodmax=16000 - bootpretransmitperiodinit=2000 A new configuration BOOTP_RANDOM_XID has been added to enable a randomized BOOTP/DHCPv4 transaction ID. Enhance DHCPv4 sending/parsing option 209 (PXE config file). A previous patch was accepted. A new patch fixes a possible double free() and addresses latest review comments. changes in v5: - fix depends for BOOTP_RANDOM_XID: "depends on CMD_BOOTP && (LIB_RAND || LIB_HW_RAND)" - add change log to individual patches changes in v4: - rebase master and resolve conflicts - change commit description for DHCP option 209 patch (this is now an enhancement patch) - default y for BOOTP_PXE_DHCP_OPTION (feedback from review) - Add "select LIB_RAND" for "config CMD_BOOTP" (retransmission improvements require rand()) changes in v3: - add define for option 209 and rfc5071 reference - Set RETRANSMIT_PERIOD_MAX_MS=60000 - Add randomization factor to retransmission timeout - Add depends for BOOTP_RANDOM_XID changes in v2: - use env_get_ulong() to get environment variables Sean Edmond (3): net: Additional fixes for dhcp option 209 net: bootp: BOOTP/DHCPv4 retransmission improvements net: bootp: add config option BOOTP_RANDOM_XID cmd/Kconfig | 10 ++++++- cmd/pxe.c | 2 ++ net/bootp.c | 76 +++++++++++++++++++++++++++++++++++++++++------------ net/bootp.h | 2 ++ 4 files changed, 72 insertions(+), 18 deletions(-) -- 2.42.0

2 6

[PATCH 0/4] rockchip: rk3588: Fix SPI flash bootsource id values
by Jonas Karlman 17 Nov '23

17 Nov '23

The commit fd6e425be243 ("rockchip: rk3588-rock-5b: Enable boot from SPI NOR flash") added a new BROM_BOOTSOURCE_SPINOR_RK3588 with value 6. At the time the reason for this new bootsource id value 6 was unknown. Since then the following bootsource id values have been observed on different RK3588 devices: - Radxa ROCK 5 Model A - fspim0_pins - rk3588s - Xunlong Orange Pi 5 - fspim0_pins - rk3588s brom_bootdevice_id: 3 - FriendlyElec NanoPC-T6 - fspim1_pins - rk3588 - Xunlong Orange Pi 5 Plus - fspim1_pins - rk3588 brom_bootdevice_id: 4 - Radxa ROCK 5 Model B - fspim2_pins - rk3588 brom_bootdevice_id: 6 It has now been confirmed that the BootRom on RK3588 use different bootsource id values depending on the iomux used by the flash spi controller, and not by the type of spi nor or spi nand flash used. This series adjusts the bootsource enum values, updates the boot_devices array to fix booting from SPI flash on RK3588 devices using fspim1_pins. It also enable building of a bootable SPI image for affected devices. John Clark (1): rockchip: rk3588-nanopc-t6: Build SPI image Jonas Karlman (2): rockchip: rk3588: Fix boot from SPI flash rockchip: rk3588s-orangepi-5: Build SPI image Slawomir Stepien (1): rockchip: rk3588-orangepi-5-plus: Build SPI image arch/arm/include/asm/arch-rockchip/bootrom.h | 4 +++- arch/arm/mach-rockchip/rk3588/rk3588.c | 5 +++-- configs/nanopc-t6-rk3588_defconfig | 1 + configs/orangepi-5-plus-rk3588_defconfig | 1 + configs/orangepi-5-rk3588s_defconfig | 1 + 5 files changed, 9 insertions(+), 3 deletions(-) -- 2.42.0

5 12

[PATCH] bootstd: Avoid freeing a non-allocated buffer
by Simon Glass 17 Nov '23

17 Nov '23

EFI applications can be very large and thus used to cause boot failures when malloc() space was exhausted. A recent changed fixed this by using the kernel_addr_r environment var as the address of the buffer. However, it still frees the buffer when the bootflow is discarded. Fix this by introducing a flag to indicate whether the buffer was allocated, or not. Note that kernel_addr_r is not the last word here. It might be better to use lmb to place images. But there is a lot of refactoring to do before we can remove the environment variables. The distro scripts rely on them so it is safe for bootstd to do so too. Fixes: 6a8c2f9781c bootstd: Avoid allocating memory for the EFI file Signed-off-by: Simon Glass <sjg(a)chromium.org> Reported by: Simon Glass <sjg(a)chromium.org> Reported by: Shantur Rathore <i(a)shantur.com> --- boot/bootflow.c | 3 ++- boot/bootmeth_efi.c | 1 + include/bootflow.h | 5 ++++- 3 files changed, 7 insertions(+), 2 deletions(-) diff --git a/boot/bootflow.c b/boot/bootflow.c index 6922e7e0c4e7..1ea2966ae9a5 100644 --- a/boot/bootflow.c +++ b/boot/bootflow.c @@ -467,7 +467,8 @@ void bootflow_free(struct bootflow *bflow) free(bflow->name); free(bflow->subdir); free(bflow->fname); - free(bflow->buf); + if (!(bflow->flags & BOOTFLOWF_STATIC_BUF)) + free(bflow->buf); free(bflow->os_name); free(bflow->fdt_fname); free(bflow->bootmeth_priv); diff --git a/boot/bootmeth_efi.c b/boot/bootmeth_efi.c index ae936c8daa18..9ba7734911e1 100644 --- a/boot/bootmeth_efi.c +++ b/boot/bootmeth_efi.c @@ -160,6 +160,7 @@ static int efiload_read_file(struct bootflow *bflow, ulong addr) if (ret) return log_msg_ret("read", ret); bflow->buf = map_sysmem(addr, bflow->size); + bflow->flags |= BOOTFLOWF_STATIC_BUF; set_efi_bootdev(desc, bflow); diff --git a/include/bootflow.h b/include/bootflow.h index 44d3741eacae..fede8f22a2b8 100644 --- a/include/bootflow.h +++ b/include/bootflow.h @@ -43,9 +43,12 @@ enum bootflow_state_t { * and it is using the prior-stage FDT, which is the U-Boot control FDT. * This is only possible with the EFI bootmeth (distro-efi) and only when * CONFIG_OF_HAS_PRIOR_STAGE is enabled + * @BOOTFLOWF_STATIC_BUF: Indicates that @bflow->buf is statically set, rather + * than being allocated by malloc(). */ enum bootflow_flags_t { BOOTFLOWF_USE_PRIOR_FDT = 1 << 0, + BOOTFLOWF_STATIC_BUF = 1 << 1, }; /** @@ -72,7 +75,7 @@ enum bootflow_flags_t { * @fname: Filename of bootflow file (allocated) * @logo: Logo to display for this bootflow (BMP format) * @logo_size: Size of the logo in bytes - * @buf: Bootflow file contents (allocated) + * @buf: Bootflow file contents (allocated unless @flags & BOOTFLOWF_STATIC_BUF) * @size: Size of bootflow file in bytes * @err: Error number received (0 if OK) * @os_name: Name of the OS / distro being booted, or NULL if not known -- 2.43.0.rc0.421.g78406f8d94-goog

4 9

[PATCH] spl: fix TPL_SYS_MALLOC_F description
by John Keeping 17 Nov '23

17 Nov '23

This config option enables the malloc() pool in TPL not the SPL. Fix the description to accurately reflect this. Fixes: fd8497dae54 (spl: Create proper symbols for enabling the malloc() pool) Signed-off-by: John Keeping <jkeeping(a)inmusicbrands.com> --- Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Kconfig b/Kconfig index 7df91d789f6..00ed1ecc173 100644 --- a/Kconfig +++ b/Kconfig @@ -327,7 +327,7 @@ config SPL_SYS_MALLOC_F_LEN malloc() region in SDRAM once it is inited. config TPL_SYS_MALLOC_F - bool "Enable malloc() pool in SPL" + bool "Enable malloc() pool in TPL" depends on SYS_MALLOC_F && TPL default y if SPL_SYS_MALLOC_F help -- 2.42.1

3 2

[PATCH] virtio: rng: gracefully handle 0 byte returns
by Andre Przywara 17 Nov '23

17 Nov '23

According to the virtio v1.x "entropy device" specification, a virtio-rng device is supposed to always return at least one byte of entropy. However the virtio v0.9 spec does not mention such a requirement. The Arm Fixed Virtual Platform (FVP) implementation of virtio-rng always returns 8 bytes less of entropy than requested. If 8 bytes or less are requested, it will return 0 bytes. This behaviour makes U-Boot's virtio_rng_read() implementation go into an endless loop, hanging the system. Work around this problem by always requesting 8 bytes more than needed, but only if a previous call to virtqueue_get_buf() returned 0 bytes. This should never trigger on a v1.x spec compliant implementation, but fixes the hang on the Arm FVP. Signed-off-by: Andre Przywara <andre.przywara(a)arm.com> Reported-by: Peter Hoyes <peter.hoyes(a)arm.com> --- drivers/virtio/virtio_rng.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/drivers/virtio/virtio_rng.c b/drivers/virtio/virtio_rng.c index b85545c2ee5..786359a6e36 100644 --- a/drivers/virtio/virtio_rng.c +++ b/drivers/virtio/virtio_rng.c @@ -20,7 +20,7 @@ struct virtio_rng_priv { static int virtio_rng_read(struct udevice *dev, void *data, size_t len) { int ret; - unsigned int rsize; + unsigned int rsize = 1; unsigned char buf[BUFFER_SIZE] __aligned(4); unsigned char *ptr = data; struct virtio_sg sg; @@ -29,7 +29,12 @@ static int virtio_rng_read(struct udevice *dev, void *data, size_t len) while (len) { sg.addr = buf; - sg.length = min(len, sizeof(buf)); + /* + * Work around implementations which always return 8 bytes + * less than requested, down to 0 bytes, which would + * cause an endless loop otherwise. + */ + sg.length = min(rsize ? len : len + 8, sizeof(buf)); sgs[0] = &sg; ret = virtqueue_add(priv->rng_vq, sgs, 0, 1); -- 2.25.1

4 8

[PATCH 0/3] Revert HAFDBS changes
by Chris Packham 17 Nov '23

17 Nov '23

As discussed this series reverts the HAFDBS changes that caused an issue on AC5/AC5X. I think there are some improvements that can be made to the initial memory map for the AC5/AC5X but so far nothing I've found makes it compatible with the HAFDBS changes. Chris Packham (3): Revert "armv8: enable HAFDBS for other ELx when FEAT_HAFDBS is present" Revert "arm64: Use level-2 for largest block mappings when FEAT_HAFDBS is present" Revert "arm64: Use FEAT_HAFDBS to track dirty pages when available" arch/arm/cpu/armv8/cache_v8.c | 30 +++--------------------------- arch/arm/include/asm/armv8/mmu.h | 20 ++++---------------- arch/arm/include/asm/global_data.h | 2 -- 3 files changed, 7 insertions(+), 45 deletions(-) -- 2.42.0

4 13

[PATCH 1/4] usb: gadget: g_dnl.c: Drop unused includes
by Tom Rini 17 Nov '23

17 Nov '23

We don't need <dfu.h> nor <thor.h> in this file, drop them. Signed-off-by: Tom Rini <trini(a)konsulko.com> --- drivers/usb/gadget/g_dnl.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/usb/gadget/g_dnl.c b/drivers/usb/gadget/g_dnl.c index afb7b74f3057..8423da377ca9 100644 --- a/drivers/usb/gadget/g_dnl.c +++ b/drivers/usb/gadget/g_dnl.c @@ -16,8 +16,6 @@ #include <g_dnl.h> #include <usb_mass_storage.h> -#include <dfu.h> -#include <thor.h> #include <env_callback.h> -- 2.34.1

3 6