edk2-nvidia-202507.0

What’s Changed

• Migrated to edk2-stable202411.

Details

• fix: add missing guid in opteenvlib
• fix: By default, allow longer kernel cmdline in L4TLauncher
• feat: Allow VPR node to be zero size
• feat: add support for android init_boot
• chore(t264): enable Mgbe driver
• fix(PcieDWControllerDxe): re-enable threaded init
• chore: Remove ArmGicLib dependency for GicUtil
• fix: Include instance of RedfishPlatformWantedDeviceLib
• fix: correct the invalid i2c descriptor index for removal
• fix: address cache topology issues
• chore(EqosDeviceDxe): update nvethernetrm library
• chore: Upgrade to edk2-stable202411
• chore: Migrate to upstream AcpiSpmiLib
• fix(PcieDWControllerDxe): disable threaded init
• fix: methods to select rpmb partition and get emmc device
• fix: don't select the emmc device before the transaction
• fix: get the device cid from the diskinfo protocol
• fix: make egm bad pages memory default aligned
• fix: allocate egm bad pages memory

edk2-nvidia-202506.1

Details

• fix(DisplayDeviceTreeHelperLib): avoid using DeviceTreeHelperLib
• chore: report BMC event when host interface is missing
• fix: Increase the capsule size limit to 25MB
• fix: correct product name detection in SMBIOS Type1
• Fix: Cancel FW update on all FDs if an error occurs on any FD.
• fix: skip providing egm bad pages if none found
• fix: Vendor ID checking slowness in smmulib
• fix(BpmpIpc): protect blocking Communicate call
• fix(PcieDWControllerDxe): re-enable threaded init
• fix(ArmTimerDxe): use serial port debug lib
• fix(NvDisplayControllerDxe): introduce NvDisplayHwShutdown
• fix(t264): update boot chain register

edk2-nvidia-202405.3 (r36.4.4)

Details

• fix(stmm): add error handling for var store integrity

edk2-nvidia-202210.5.2 (r35.6.2)

Details

• fix(stmm): add error handling for var store integrity

edk2-nvidia-202506.0

Details

• fix: Change L4T/Android to black logo
• fix: Convert to SimpleTextOut for prints
• fix(PcieDWControllerDxe): disable threaded init
• chore(stmm): adjust the load address and image size
• chore(t264): add multi-socket support
• feat(t264): Include t264 GOP
• chore(t264): add t264 platform support
• chore: Introduce cppcheck suppression file
• refactor: Introduce BUILD_ANDROID to Kconfig
• chore: Remove legacy Kconfig SOC definitions
• fix: retry on FW update receive errors
• Revert "feat(t264): use blob FW update support"
• Revert "chore(t264): update capsule size"
• feat: Select product name dynamically using DTB and FRU data
• feat: support DTB cpu node numa-node-id field
• chore: trace cpus enabled bitmask
• chore(t264): enable host1x support
• chore: update firmware capsule size
• fix: only update inactive BR-BCT backup data
• chore(t264): update required capsule image names
• chore(t264): set inactive chain failed during capsule update
• chore(t264): support marker-based BR-BCT updates
• chore(t264): add UpdateBrBct support
• chore(t264): boot chain register updates
• fix: force timeout if SSIF takes too long on average
• fix: add ssdt only if present
• feat: Adds support for custom boot manager colors
• fix: Reduces duplicate prints to GOP consoles
• feat: Add scaled logo support
• fix: Allow for multiple top level caches
• fix: Correct Redfish Host Interface Device Type
• fix: Increase size of FPDT record
• fix: use TRM trim and tap values
• Fix: Update Chassis Manufacturer from DTB
• fix: Compiler warning in SBIOS UEFI stub library MemoryAllocationStubLib
• chore: ignore pcie io bar
• fix(stmm): Avoid installing the system memory HOB
• chore: Update function names to adhere to EDK2 coding conventions
• feat: Introduce support for T264 display
• feat: support more I2C controllers
• fix: count private resource correctly
• chore(smmuv3): update log levels
• fix: Initialize variables in ParseDefaultBootPriority
• chore(smmuv3): add a print telling busy
• fix: skip vid smmu override for presil
• fix: Intialize Status in RefreshNvBootOptions
• feat(smmuv3): enable bypass for vendor
• chore: Uncrustify GoogleTest cpp files
• feat: Add Timer, Gic, and PCIe UEFI shell apps
• fix(stmm): don't support multi socket mmio regions
• feat: add record index reading support for RAS log consumption
• fix: populate speed in type 17 per DRAM
• fix: change the way draminfo is indexed into
• chore: Fix compiler warning in TH500GpuDxe
• feat: Debug print memory locations early in boot
• fix(stmm): overwrite the status flag for a failed sfdp
• feat: Send system reset POST code when OS shutdown
• feat: add a SBIOS option to add some delay after PERST
• fix: Remove invalid i2c descriptor for address conversion
• fix(ufs): set initial adapt rate
• fix: Correct RasPayloadSize for log access overflow issue
• fix(stmm): add error handling for var store integrity
• fix(t264): Use HS series rate B

edk2-nvidia-202405.1.1 (r36.4.5)

Details

• fix: switch default value of PcdSocDisplayHandoffMode to Auto
• fix(varint): read records per erase block and fix leak
• fix: disable RCM boot of full UEFI for Jetson
• feat: add rcm support in launcher
• fix: Seperate L4T launcher efi from platform
• fix: change the logic to set physical presence pcd
• chore: add config to allow rcm boot
• fix: secure boot enablement in jetson minimal
• chore: use console debug lib for launcher

edk2-nvidia-202402.3

What’s Changed

• Upstream security fixes, including CVE-2022-36765, CVE-2022-36763, CVE-2022-36764.
• Various bug fixes.

Details

• fix: add error handling in fvbnorflashstandalonemm for var store integrity
• fix: disable RCM boot of full UEFI for Jetson
• feat: add rcm support in launcher
• fix: Seperate L4T launcher efi from platform
• fix: change the logic to set physical presence pcd
• chore: add config to allow rcm boot
• fix: secure boot enablement in jetson minimal
• chore: use console debug lib for launcher

edk2-nvidia-202504.1

What’s Changed

• SMMUv3 has been enabled by default on t24x (th500).

Details

• feat(smmuv3): add cache maintenance
• fix(smmuv3): reset strtab currently used
• feat(smmuv3): config stage 2 stes and create pts
• feat(smmuv3): configure stes
• feat(smmuv3): Add client driver

edk2-nvidia-202504.0

What’s Changed

Early code to support SMMUv3

Some code has been added to support SMMUv3, but SMMUv3 has not been completely enabled yet.

Early code to support T264

Some code has been added to support T264, but this code is not ready for use. This includes the T264 UEFI image as well as the StandaloneMmJetson image.

Details

• fix(smmuv3): do not free driver private data
• chore: add sid base for root bridge
• fix: add ecid to the hash that generates uuid
• chore: install smmu init complete protocol
• chore: add support for env matching in overlays
• fix: update smbios fields to read from CPUBL Params
• chore: add smmu v3 phandle for root bridge
• chore: update smmu protocol interfaces
• chore(t264): update capsule size
• fix: XusbControllerDxe Fix build with gcc-15
• feat(t264): use blob FW update support
• refactor: move FMP version support to FmpVersionLib
• chore: add PCDs for data flash and blob FW support
• feat: add bpmp ipc memory patching
• fix: install rt properties for presil
• feat(smmuv3): record event from invalid input sid
• feat(smmuv3): shutdown smmu on ReadyToBoot event
• feat(smmuv3): enable smmuv3 controller
• fix(smmuv3): use 64 bit write for STRTAB_BASE
• chore: Follow GCC5 toolchain to new name
• fix: don't try to get chip id in optee deployments
• fix: varstorecs method for new jetson soc
• refactor: Modify ethernet ASL to use pcds
• refactor: Remove SOC_GRACE
• fix(standalonemm_jetson): change memory address offsets
• fix: Fix memory leaks and improve memory management
• feat(smmuv3): add dummy functions
• feat(smmuv3): set up stream table
• fix: Last kernel command line variable update fix
• feat(smmuv3): set up event queue
• feat(smmuv3): set up command queue
• feat(smmuv3): configure smmu controller settings
• refactor: use per-SoC PlatformResourceLib
• feat(smmuv3): reset the smmu controllers
• fix: remove duplicate code for smmuv3
• fix: correct NCT serial number string overflow
• feat(smmuv3): add basic smmu protocol
• fix(smmuv3): use correct dtb library
• chore: add framework for smmu v3
• fix: Use emulated variables on android
• chore: increase max supported i2c controllers

edk2-nvidia-202503.1

Details

• fix: switch default value of PcdSocDisplayHandoffMode to Auto
• feat(th500): rename channel sparing menu option
• fix: Add version field in CPUBL Params
• fix: Update CPUBL Params V1