u-boot.git/lib, branch main

efi_loader: Measure secure boot variables in the correct PCR if DeployedMode is set

2026-07-15T22:40:43Z

PK, KEK, db, dbx etc must always be measured in PCR7. DeployedMode and AuditMode should be measured in PCR1 if DeployedMode is set and PCR7 otherwise. Fix the u16_strcmp to only change the PCR value for those two variables. Signed-off-by: Ilias Apalodimas Acked-by: Heinrich Schuchardt

efi_loader: Change the return values if a TPM is missing during GetEventLog

2026-07-15T22:40:43Z

The function defined by the TCG spec looks like: typedef EFI_STATUS (EFIAPI *EFI_TCG2_GET_EVENT_LOG) ( IN EFI_TCG2_PROTOCOL *This, IN EFI_TCG2_EVENT_LOG_FORMAT EventLogFormat, OUT EFI_PHYSICAL_ADDRESS *EventLogLocation, OUT EFI_PHYSICAL_ADDRESS *EventLogLastEntry, OUT BOOLEAN *EventLogTruncated ); and the spec mandates that "If no TPM is present, the function SHALL set the following values and return EFI_SUCCESS: EventLogLocation = NULL EventLogLastEntry = NULL EventLogTruncated = FALSE" However, if we set it to NULL the local assignment is discarded when the function returns. Set it to 0, although on some platforms that's a valid address. Signed-off-by: Ilias Apalodimas Reviewed-by: Heinrich Schuchardt

efi_loader: Correctly count the size on DTB measurements

2026-07-15T22:40:42Z

When doing a sha256_update() for the measured DT, the size arguments for fdt_size_dt_struct() and fdt_size_dt_strings() are inversed. Signed-off-by: Ilias Apalodimas Acked-by: Heinrich Schuchardt

efi_loader: fix buffer overrun in efi_sigstore_parse_siglist

2026-07-15T22:40:42Z

In efi_sigstore_parse_siglist() sigdata is allocated. But instead of an allocation matching the size of sigdata, tainted external data was used to calculate the allocation size. This may lead to buffer overflows. * Correct the allocation size. * Follow the man-page. Use the structure size as second argument for calloc. Reviewed-by: Ilias Apalodimas Signed-off-by: Heinrich Schuchardt

efi_selftest: test block io revision and pointers

2026-07-15T22:40:42Z

Enhance the unit test to verify all Revision fields and all pointers of all the EFI_BLOCK_IO_PROTOCOL structures. As the unit test registers its own block io protocol for test purposes, make sure to initialize its revision properly, as it will be verified as well. This can run on the sandbox with the following command: ./u-boot -T -c 'setenv efi_selftest block device; bootefi selftest' Suggested-by: Heinrich Schuchardt Signed-off-by: Vincent Stehlé Cc: Ilias Apalodimas Cc: Tom Rini Reviewed-by: Heinrich Schuchardt

efi_selftest: free handles on teardown()

2026-07-15T22:40:42Z

In the block device selftest, make the handles pointer global and free it also in teardown(), to simplify error handling. We also need to nullify the pointer after freeing it on the normal path, to avoid freeing it a second time during teardown(). Signed-off-by: Vincent Stehlé Reviewed-by: Heinrich Schuchardt

efi_loader: fix building with CONFIG_EFI_VARIABLES_PRESEED=y

2026-07-15T22:40:42Z

Since commit bd3f9ee679b4 ("kbuild: Bump the build system to 6.1") out-of-tree builds with CONFIG_EFI_VARIABLES_PRESEED=y fail with errors like: ../lib/efi_loader/efi_var_seed.S:14: Error: file not found: ubootefi.var For out-of-tree build we cannot use CONFIG_EFI_VAR_SEED_FILE in the .incbin statement of file efi_var_seed.S. * We have to prepend $(srctree) if the path is relative. * We must not prepend $(srctree) if the path is absolute. Fixes: bd3f9ee679b4 ("kbuild: Bump the build system to 6.1") Reported-by: Jon Mason Closes: https://lore.kernel.org/u-boot/CAPoiz9zg4OXgHo5J3WtJHKOEuWOdCDrugWfAt6Z+d71j=+q8oA@mail.gmail.com/T/#mffaca10a9e812d03eceafad59999a02e57258b9a Tested-by: Nora Schiffer Reviewed-by: Ilias Apalodimas Tested-by: Ilias Apalodimas Signed-off-by: Heinrich Schuchardt

efi_loader: fix missing EFI_EXIT in efi_disconnect_controller

2026-07-15T22:40:42Z

We are missing a call to EFI_EXIT() when returning from efi_disconnect_controller(), which we need after having called EFI_ENTRY(). Fix this by jumping to the common error path, which does call EFI_EXIT(). Even though the common error path may try to free child_handle_buffer, this cannot harm in our case as it always NULL. This is inspired by a barebox fix. [1] Link: https://git.pengutronix.de/cgit/barebox/commit/?id=080db65e39a877b000baaf843c997a69821dfe69 [1] Fixes: 314bed6c854e ("efi_loader: fix DisconnectController() for sole child") Signed-off-by: Vincent Stehlé Cc: Heinrich Schuchardt Cc: Ilias Apalodimas Cc: Tom Rini Reviewed-by: Ilias Apalodimas Reviewed-by: Heinrich Schuchardt

power: regulator: add generic PMBus UCLASS_REGULATOR adapter

2026-07-13T01:16:57Z

PMBus regulators differ in numeric formats and quirks, not in how they are driven. Share that common behaviour as a regulator-uclass adapter so chip drivers and the pmbus CLI do not each reimplement the decode and transport, and add a catch-all driver on compatible = "pmbus" for compliant chips that have no dedicated driver yet. Gated by CONFIG_DM_REGULATOR_PMBUS_HELPER and CONFIG_DM_REGULATOR_PMBUS_GENERIC. Signed-off-by: Vincent Jardin Signed-off-by: Peng Fan

pmbus: add PMBus 1.x framework, CLI, and binding

2026-07-13T01:16:57Z

Add U-Boot's PMBus 1.x layer: the decoder/transport library, the pmbus CLI command and a generic DT binding. The subsequent commits provide the UCLASS_REGULATOR adapter and per-chip drivers. U-Boot's PMBus support is not a hwmon clone of Linux's drivers/hwmon/pmbus/. Linux owns the runtime side (polling, sysfs, alert IRQs, fan loops). U-Boot owns the boot-time side in order to, - identify the PMBus regulators a board carries: MFR_ID/ MFR_MODEL/MFR_REVISION + sanity checks. - print telemetry (VIN/VOUT/IIN/IOUT/POUT/TEMP) so an operator can confirm rail voltages and faults before the kernel - decode any chip alerts (STATUS_VOUT/STATUS_IOUT/STATUS_INPUT/ STATUS_TEMPERATURE/STATUS_CML) so a boot log shows why the previous boot failed or the board had been power cycled because of an outage (typically over temperature or under current). Out of scope by design: no periodic polling, no sysfs, no fan-speed control loop, no PMBUS_VIRT_* sensor virtualisation, no caching. If a use case needs any of those, the answer should be "wait until Linux comes up". It shall remain a thin layer. The constants and structural shape (command codes, status bit names, sensor-class enum, format enum, struct pmbus_driver_info) are mirrored from Linux drivers/hwmon/pmbus/pmbus.h verbatim. The decoders/encoders are reimplemented from the PMBus 1.3 specification because the surrounding hwmon context (struct pmbus_data, sysfs caching, hwmon publication) does not apply. The main benefits: - One framework + CLI for any board carrying PMBus regulators: no per-board PMBus implementation required anymore. - Boards call pmbus_print_telemetry() / pmbus_print_status_word() directly from boot init for a snapshot, sharing all decode + format-dispatch with the CLI. - Linux-compatible constants and DT binding so porting an existing drivers/hwmon/pmbus/ chip is mechanical. - Boot-time AVS/VID rail trim reuses the same decoders and encoders as the CLI and the regulator path: no duplicate math. Signed-off-by: Vincent Jardin Signed-off-by: Peng Fan