TF-A June 2026

tf-a@lists.trustedfirmware.org

6 participants
7 discussions

Proposal to unify tf.org vulnerability referencing in NVD

by Sandrine Bailleux

Hi everyone, I am sending this email to all tf.org project mailing lists to ensure all maintainers are aware and on board regarding this matter. If you have any concerns or questions, please reply on tf.org Discord #general channel, where I'll create a thread, as I think it will be much easier than dealing with cross-mailing lists emails. Background When a security vulnerability is discovered in one of the trustedfirmware.org projects, it is common to request a "Common Vulnerabilities and Exposures" (CVE) number. This number uniquely references the issue, which can then be searched in the vulnerability databases. One of these databases is NIST's "National Vulnerability Database" (NVD): https://nvd.nist.gov<https://nvd.nist.gov/vuln/detail/CVE-2023-51712> Entering a specific CVE number in NVD search engine will allow you to easily find the details of a specific issue, for example: https://nvd.nist.gov/vuln/detail/CVE-2023-51712 However, sometimes one is not looking for a specific CVE number but rather wants to list all known vulnerabilities affecting a particular project. For this, one can use the Common Platform Enumerations (CPE) search engine: https://nvd.nist.gov/products/cpe/search CPE is a structured naming scheme that includes information like the vendor name, the project name, the version / tag, and so on. See https://nvd.nist.gov/products/cpe for more details. So for example, https://nvd.nist.gov/vuln/detail/CVE-2023-51712 referenced above has the following CPE: cpe:2.3:o:arm:trusted_firmware-m:*:*:*:*:*:*:*:* This basically means * CPE version 2.3 is in use * 'o is the type of project, in this case it stands for Operating Systems (which is probably the closest match for low-level code like TF-M) * 'arm' is the vendor (that is wrong, see below) * 'trusted_firmware-m' is the project name, Problem statement It appears that CPEs used in NVD to reference vulnerabilities in tf.org projects differ a lot across projects. For some projects, there's even multiple of them. Sometimes the vendor is "arm", sometimes it's "linaro", or something else. Some of the TF-A and MbedTLS maintainers have initiated discussions with NVD to get this simplified and unified, but it would make sense to align other tf.org projects as well. Proposal CPE naming rules are that the vendor name should the parent organization of the project. Thus the proposal would be for all tf.org projects to use "trustedfirmware" as the vendor name in their CPE. For example: cpe:2.3:o:trustedfirmware:trusted_firmware-m:*:*:*:*:*:*:*:* cpe:2.3:a:trustedfirmware:mbed_tls:*:*:*:*:*:*:*:* We're only proposing to change the vendor name here ; each project is then free to choose how they want the project name or the type of software project they want to encode there. Thanks for reading, Best regards, Sandrine Afsa

5 days, 21 hours

Trusted Firmware-A LTS v2.14.3 Release Announcement

by svc_openci_enginfra＠arm.com

1 week, 1 day

Re: Guidance on direct Linux boot from TF-A (bypassing U-Boot) on ARM64

by Olivier Deprez

Hi, I would suggest looking into the platform ports: https://trustedfirmware-a.readthedocs.io/en/latest/plat/arm/fvp/fvp-specifi… https://trustedfirmware-a.readthedocs.io/en/latest/plat/rpi3.html#build-ins… RPI3_DIRECT_LINUX_BOOT https://trustedfirmware-a.readthedocs.io/en/latest/plat/qemu.html#qemu-virt… ARM_LINUX_KERNEL_AS_BL33 Regards, Olivier. ________________________________ From: Joita Mitra <joita.m(a)delopt.co.in> Sent: 11 June 2026 09:19 To: tf-a-owner(a)lists.trustedfirmware.org <tf-a-owner(a)lists.trustedfirmware.org> Subject: Guidance on direct Linux boot from TF-A (bypassing U-Boot) on ARM64 Hi, I am working on an RZ/G2L based platform and exploring the possibility of booting Linux directly from TF-A without using U-Boot as BL33. Current boot flow: BL2 -> BL31 -> U-Boot (BL33) -> Linux Target boot flow: BL2 -> BL31 -> Linux I have been studying the TF-A image loading framework, including: * desc_image_load.h * bl2_image_load_v2.c * bl2_plat_mem_params_desc.c * bl31_main.c * bl31_prepare_next_image_entry() My understanding is that BL2 loads images according to the image descriptors and passes entry point information to BL31 through the BL33 descriptor. I modified the BL33 descriptor to point to the Linux kernel entry address instead of U-Boot and attempted to transfer execution directly to Linux from BL31. However, Linux does not start successfully, which suggests that additional preparation normally performed by U-Boot may be missing. I would appreciate guidance on the following: 1. Is using the BL33 descriptor to describe the Linux kernel the correct architectural approach when bypassing U-Boot? 2. How should the Linux kernel Image and DTB be loaded from BL2? * Should Linux be treated as BL33? * Should the DTB be loaded as a separate image descriptor? * Is there an existing TF-A mechanism intended for this use case? 3. Which TF-A components are typically modified to support direct Linux boot? * BL2 image loading descriptors? * BL31 handoff logic? * Platform-specific image loading code? 4. Are there any reference implementations or platforms in TF-A that directly boot Linux without U-Boot? 5. Are there any ARM64 Linux boot protocol requirements that TF-A must satisfy before transferring control to Linux? Any guidance or examples would be greatly appreciated. Thank you. [cid:19da2837-2e52-4337-8b03-1cc393171495]<http://www.delopt.co.in/> Joita Mitra Embedded Software Engineer [cid:7d7fa5b9-a852-48ac-a797-7ed0723558b3] +91-7985313796 Fortune Serene, 3rd Floor, Plot No: 89/B, 90/A, West Avenue 9, Electronic City Phase 1,Bangalore 560100

1 week, 2 days

Trusted Firmware-A LTS v2.14.3 Release Announcement

by svc_openci_enginfra＠arm.com

1 week, 5 days

TF-A Tech Forum - Fuzzing TF-RMM Jun 11th 2026 4.00pm UK

by Olivier Deprez

Hi, On Jun 11th 2026 4.00pm UK, Rustam Ismayilov will present the topic of TF-RMM fuzzing with the following agenda: Fuzzed Interfaces fake_host AFL++ seed generation statefull fuzzing Results and performance Limitations and planned improvements Regards, Olivier. TF-A Tech Forum Thursday Jun 11, 2026 ⋅ 5pm – 6pm Central European Time - Paris Location https://linaro-org.zoom.us/j/93557863987?pwd=56a1l8cBnetDTZ6eazHGaE1Ctk4W34… https://www.google.com/url?q=https%3A%2F%2Flinaro-org.zoom.us%2Fj%2F9355786… Trusted Firmware is inviting you to a scheduled Zoom meeting.Topic: TF-A Tech ForumTime: May 15, 2025 02:00 PM London Every 2 weeks on Thu, 78 occurrence(s)Please download and import the following iCalendar (.ics) files to your calendar system.Weekly: https://linaro-org.zoom.us/meeting/tJcocu6gqDgjEtOkyBhSQauR1sUyFwIcNKLa/ics… Zoom Meetinghttps://linaro-org.zoom.us/j/93557863987?pwd=56a1l8cBnetDTZ6eazHGaE1Ctk4W34.1Meeting ID: 935 5786 3987Passcode: 939141---One tap mobile+12532158782,,93557863987# US (Tacoma)+13017158592,,93557863987# US (Washington DC)---Dial by your location• +1 253 215 8782 US (Tacoma)• +1 301 715 8592 US (Washington DC)• +1 305 224 1968 US• +1 309 205 3325 US• +1 312 626 6799 US (Chicago)• +1 346 248 7799 US (Houston)• +1 360 209 5623 US• +1 386 347 5053 US• +1 507 473 4847 US• +1 564 217 2000 US• +1 646 558 8656 US (New York)• +1 646 931 3860 US• +1 669 444 9171 US• +1 669 900 9128 US (San Jose)• +1 689 278 1000 US• +1 719 359 4580 US• +1 253 205 0468 US• 833 548 0276 US Toll-free• 833 548 0282 US Toll-free• 833 928 4608 US Toll-free• 833 928 4609 US Toll-free• 833 928 4610 US Toll-free• 877 853 5247 US Toll-free• 888 788 0099 US Toll-freeMeeting ID: 935 5786 3987Find your local number: https://linaro-org.zoom.us/u/adoz9mILli Guests d82620130(a)gmail.com namyoon(a)google.com shaikadnanafrid(a)gmail.com tf-a(a)lists.trustedfirmware.org

1 week, 5 days

EL3 SPMC shmem datastore sizing for FFA_MEM_RETRIEVE_REQ snapshot

by Daniel Boulby

Hi all, I am sending this note to get feedback from platform maintainers on an EL3 SPMC change in my current series (https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/50778/). The patch updates spmc_ffa_mem_retrieve_req() to take a snapshot of the caller's FFA_MEM_RETRIEVE_REQ descriptor into the EL3 SPMC shmem datastore before validation and response handling. The intent is to avoid repeatedly reading mutable fields from the caller-owned TX buffer while the request is being processed. In other words, once the request enters EL3 SPMC, we operate on a stable local copy rather than on memory that remains under the caller's control. The EL3 SPMC datastore is already used to cache shared memory transaction descriptors. With this change, it is also used to hold a temporary copy of the retrieve request. That means datastore sizing becomes more important, since a valid larger retrieve request may now fail if the backing datastore is too small. The datastore sizes are defined in platform_def.h (https://review.trustedfirmware.org/plugins/gitiles/TF-A/trusted-firmware-a/… for the fvp platform) and while looking through the existing platform definitions, I noticed that datastore sizes vary quite a lot. For example, some platforms provide a relatively large backing store, while others define much smaller regions. In particular, rdn2 currently defines only a 1024-byte datastore. My concern is that for such platforms, larger memory-share / retrieve flows could fail simply because the datastore is too small to hold the request snapshot. Before I proceed further, I wanted to ask platform maintainers: 1. Is there any platform-specific reason for these reduced datastore sizes? 2. If not, would maintainers be happy for these datastore sizes to be increased so they can accommodate larger valid transactions? In short, I want to confirm whether the smaller datastore definitions are deliberate platform constraints, or whether they should be treated as undersized defaults that now need revisiting. Feedback from maintainers would be very helpful before I post the next revision. Kind Regards, Daniel Boulby

2 weeks, 2 days

CI infrastructure scheduled maintenance: 4th June Sep 2026

by Saheer Babu

Hi all, We will be upgrading Cloudbees CI and clusters hosting review.trustedfirmware.org and ci.trustedfirmware.org on Wednesday, 3rd June 2025 at 16:00 GMT+1. During this maintenance window, both services will be unavailable for approximately 8 hours. A follow-up email will be sent once the services are fully restored. Best regards, Saheer [LOGO SMALL] Saheer Babu Principal Software Engineer CESW – Engineering Infrastructure

2 weeks, 3 days

2026

2025

2024

2023

2022

2021

2020

2019

2018

TF-A June 2026