November 2019 - TF-A - lists.trustedfirmware.org

Re: [TF-A] Design proposal: firmware debugfs interface

by Olivier Deprez

Hi, To complement the DebugFS design proposal document, here is a corresponding TF-A patchset: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/2422 https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/2431 https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/2432 https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/2423 https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/2421 An associated TFTF test case illustrates the usage: https://review.trustedfirmware.org/c/TF-A/tf-a-tests/+/2433 Regards, Olivier. ________________________________ From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org> on behalf of Olivier Deprez via TF-A <tf-a(a)lists.trustedfirmware.org> Sent: 25 October 2019 15:30 To: tf-a(a)lists.trustedfirmware.org <tf-a(a)lists.trustedfirmware.org> Subject: [TF-A] Design proposal: firmware debugfs interface Hi All, We are proposing a new firmware debug interface in the form of a "debug filesystem". The intent is to expose live firmware data or firmware driver HAL to the upper layers, in debug builds. For people interested, please review and comment the design proposal: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/2381 The design is not closed, the intent is to collect opinions, and have a discussion on options and implications. Thanks & Regards, Olivier. IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. -- TF-A mailing list TF-A(a)lists.trustedfirmware.org https://lists.trustedfirmware.org/mailman/listinfo/tf-a

4 years

1
1
0 0

Re: [TF-A] Design proposal: Property Access Layer

by Abhishek Pandit

Hi Louis, The term PAL is also used for platform abstraction layer so it "could" be confusing to have this being used for something different. Not necessarily suggesting to change anything but it might be worth considering this aspect. Thanks, Abhishek -----Original Message----- From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org> On Behalf Of Louis Mayencourt via TF-A Sent: 19 November 2019 13:40 To: tf-a(a)lists.trustedfirmware.org Subject: [TF-A] Design proposal: Property Access Layer Hi everyone, Please let me introduce the `Property Access Layer` prototype: The Property Access Layer (PAL) is an abstraction layer for platform specific data, allowing a "property" to be queried and a value retrieved without the requesting entity knowing what backing store is being used to hold the data. It is used to bridge new and old ways of providing platform-specific data: Today, information like the Chain of Trust is held within several, nested platform-defined tables. In the future, it may be provided as part of a device tree blob, along with the information about images to load. Introducing this abstraction layer will make migration easier and will preserve functionality for platforms that cannot / don't want to use device tree. Please have a look at the patches: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/2559/1 Regards, Louis IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. -- TF-A mailing list TF-A(a)lists.trustedfirmware.org https://lists.trustedfirmware.org/mailman/listinfo/tf-a

4 years, 4 months

2
2
0 0

RPi4 and CPU hotplugging

by Jan Kiszka

Hi Andre, I'm trying to get CPU offlining work for Linux on the RPi4. In fact, this is working already with current ATF master and the Raspberry kernel 4.19.85. Not it would be "nice" if onlining worked as well: [ 94.959294] CPU1: shutdown [ 94.959311] psci: CPU1 killed. [ 106.750960] CPU1: failed to come online [ 106.755425] CPU1: failed in unknown state : 0x0 Is this a known limitation? Or am I using the wrong kernel (obviously, it's a downstream one ;) )? Thanks, Jan PS: Current RPi firmware requires an explicit "armstub=armstub8.bin" in config.txt, no automatic pickup.

4 years, 4 months

3
6
0 0

Re: [TF-A] Insufficient ordering guarantees in bakery_lock_get() and bakery_lock_release() in bakery_lock_coherent.c

by Raghupathy Krishnamurthy

Hi Soby, Thanks for your response. >>it is needed to ensure the ordering of the succeeding sev(). Agree. Thanks for the clarification. >>Was this an issue that actually manifested on a hardware or is this something that you caught while reviewing the code? Noticed it while reviewing code and I have not observed it on hardware. Thanks -Raghu On November 26, 2019 at 8:55 AM, Soby Mathew <Soby.Mathew(a)arm.com> wrote: On 26/11/2019 16:30, Raghupathy Krishnamurthy via TF-A wrote: Hello! Reposting this from (https://developer.trustedfirmware.org/T589). bakery_lock_get() uses a dmbld() after lock acquisition which is insufficient in a lock acquire situation. With just dmbld(), stores in a critical section can be reordered before the dmbld() and indeed before the lock acquisition has taken place. similarly, bakery_lock_release() only uses dmbst(). A load in the critical section could be reordered after the dmbst() and write to the lock data structure releasing the lock. This is likely less of a problem but lock release needs to provide release semantics, and dmbst() is insufficient. For ex: A load in the critical section of CPU1 can be reordered after the store for the lock release, and it could read from a store that is executed on CPU2 in the same critical section, since CPU2 saw the store for the lock release first, and raced into the critical section. Hi Raghu, You are right on this. The dmbld() and dmbst() does not provide sufficient guarantees in the cases you mention. Was this an issue that actually manifested on a hardware or is this something that you caught while reviewing the code ? Also the dsb() after the write to the lock seems unnecessary. Am I missing something here ? It looks like the same issue is present even in bakery_lock_normal. If you are referring to the dsb() at this line : https://git.trustedfirmware.org/TF-A/trusted-firmware-a.git/tree/lib/locks/… it is needed to ensure the ordering of the succeeding sev(). Best Regards Soby Mathew Thanks Raghu

4 years, 4 months

2
2
0 0

Re: [TF-A] [RFC] New feature in TF-A to load encrypted FIP payloads

by Sandrine Bailleux

Hello Sumit, Sorry for not getting back to you earlier on this. I started looking at your patches and although I've not completely got my head around them yet, I've got some early comments and questions. I've tried to classify them by themes below. First of all, let me say that I'm no security expert and I did not know about the concept of authenticated encryption/decryption before looking at your patches. I spent a couple of hours reading about such algorithms in general and AES-GCM in particular. Most of what I've learned so far is based on my understanding of RFC 5116 [1]. What's the motivation for choosing GCM to start with? I've seen that it is free of patent [2], which I am guessing was a strong argument for it. I've also read that it is supposed to be quite lightweight and can take full advantage of parallel processing, although I've not looked into the details. Were these the reasons? Any other reasons? Key management -------------- fip_file_read() retrieves the key from the platform and stores it in a buffer on the stack. I don't see any code wiping it out of memory once we're done with it. Did I miss it? Unlike the root of trust public key, this is a (symmetric) secret key so it is sensitive data that we must not leave for grabs, even if the stack is in Trusted RAM and that it's likely to be overwritten by subsequent stack usage. Also, I am still trying to get my head around how this would integrate with a cryptographic engine where the key does not leave the chip. I can imagine that we could get the address of the encrypted firmware image from the FIP, pass that to a cryptographic engine, request it to decrypt it and store the result somewhere in Trusted RAM. In this case, we wouldn't call plat_get_fip_encryption_key(). Do you have any idea how we would pull this off? Like how the different modules (IO layer, crypto module, image parser module, ...) would integrate together? I have some concerns around the generation of the initialization vectors in the encrypt_fw tool. Right now, IVs are simply a random sequence of bytes (obtained through a call to OpenSSL's RAND_bytes() API). Now, I would imagine that RAND_bytes() is typically based on a good random number generator and thus will generate different sequences every time it is called. At least, as long as it is called from the same machine every time. But what if we encrypt a new FIP bundle from a different machine, say in the context of a firmware update? Is it not possible that it might choose the same IV out of bad luck? Perhaps that's an issue left to provisioning/manufacturing time and is out of the scope here. But it worries me because AFAIU, the security of AES-GCM is critically undermined if the same nonce is used multiple times with the same key (see section 5.1.1. "Nonce reuse" in RFC 5116). If the encryption key is the SSK (rather than the BSSK) then I guess the probability is even higher, as it is shared amongst a class of devices. Impact on memory footprint and performance ------------------------------------------ Do you know what the performance impact is when this feature is enabled in TF-A, to decrypt images at boot time? Obviously it depends on the platform and whether there is a dedicated cryptographic engine, and I suppose you cannot really get any relevant measurements out of QEMU but I would be interested if you've got any rough numbers. And what's the memory footprint impact? IIUC, AES-GCM almost does not inflate the size of the data it encrypts. The size of the ciphertext seems to be the same as the plaintext + the size of the authentication tag. So I guess there's no real impact on flash storage and Trusted RAM usage to hold decrypted firmware. But what about the mbedTLS primitives to decrypt the images? How much code and data does this add? encrypt_fw tool --------------- We have some floating ideas around re-implementing the tools (fiptool, certtool) in a scripting language (possibly python) in the future and also doing a better job at sharing a common description of the list of images to boot/authenticate between the firmware and the host tools. But we're not there yet, so I agree that implementing this new tool in C from the same "mold" as fiptool and certtool is what makes the most sense today. It's just another tool we will have to rework if and when we get there. I did not understand why this new tool needs to know what image it is encrypting. For example, one possible invocation could be: tools/encrypt_fw/encrypt_fw \ -k 1234567890abcdef1234567890abcdef1234567890abcdef1234567890abcdef \ --soc-fw bl31.bin \ --soc-fw-enc bl31_enc.bin \ --tos-fw bl32.bin \ --tos-fw-enc bl32_enc.bin Why not invoking the tool once per image instead? As in: encrypt_fw -k key -in ifile -out ofile for BL31, then for BL32? Does the tool do anything different based on the type of image it receives? Regards, Sandrine [1] https://tools.ietf.org/html/rfc5116 [2] https://csrc.nist.gov/csrc/media/projects/block-cipher-techniques/documents…

4 years, 4 months

2
1
0 0

Design proposal: new buildsystem

by Balint Dobszay

Hi All, The buildsystem of TF-A became complex and loaded with technical debt during the years, and it's time to do something about this. We made some plans and prototyping work to move to a CMake based solution and we would like to get feedback on the idea. Why CMake? In summary CMake is a mature tool having a wide acceptance in C and C++ projects. Also it has benefits of decreasing fragmentation in the developer community if we sync up with TF-M. How will it happen? This will be a slow process where the old build system will co-exist for a period with the new one. How long that period will be is an open question. For a more detailed summary please see https://developer.trustedfirmware.org/w/tf_a/cmake-buildsystem-proposal/ The design discussion will follow the design review proposal process of TF.org, as described on this page: https://ci.trustedfirmware.org/job/tf-m-build-test-nightly/lastSuccessfulBu… You can find the patch for capturing the design decisions and discussion here: https://review.trustedfirmware.org/#/c/TF-A/trusted-firmware-a/+/2662/ If you would like to contribute or have an opinion or any ideas please reply to this email or add a comment on Gerrit (link above). Regards, Balint IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

4 years, 4 months

1
0
0 0

Re: [TF-A] Insufficient ordering guarantees in bakery_lock_get() and bakery_lock_release() in bakery_lock_coherent.c

by Soby Mathew

On 26/11/2019 16:30, Raghupathy Krishnamurthy via TF-A wrote: > Hello! > > Reposting this from (https://developer.trustedfirmware.org/T589). > > bakery_lock_get() uses a dmbld() after lock acquisition which is insufficient in a lock acquire situation. With just dmbld(), stores in a critical section can be reordered before the dmbld() and indeed before the lock acquisition has taken place. similarly, bakery_lock_release() only uses dmbst(). A load in the critical section could be reordered after the dmbst() and write to the lock data structure releasing the lock. This is likely less of a problem but lock release needs to provide release semantics, and dmbst() is insufficient. For ex: A load in the critical section of CPU1 can be reordered after the store for the lock release, and it could read from a store that is executed on CPU2 in the same critical section, since CPU2 saw the store for the lock release first, and raced into the critical section. Hi Raghu, You are right on this. The dmbld() and dmbst() does not provide sufficient guarantees in the cases you mention. Was this an issue that actually manifested on a hardware or is this something that you caught while reviewing the code ? > Also the dsb() after the write to the lock seems unnecessary. Am I missing something here ? It looks like the same issue is present even in bakery_lock_normal. > If you are referring to the dsb() at this line : https://git.trustedfirmware.org/TF-A/trusted-firmware-a.git/tree/lib/locks/… it is needed to ensure the ordering of the succeeding sev(). Best Regards Soby Mathew > Thanks > Raghu >

4 years, 5 months

1
0
0 0

Insufficient ordering guarantees in bakery_lock_get() and bakery_lock_release() in bakery_lock_coherent.c

by Raghupathy Krishnamurthy

Hello! Reposting this from (https://developer.trustedfirmware.org/T589). bakery_lock_get() uses a dmbld() after lock acquisition which is insufficient in a lock acquire situation. With just dmbld(), stores in a critical section can be reordered before the dmbld() and indeed before the lock acquisition has taken place. similarly, bakery_lock_release() only uses dmbst(). A load in the critical section could be reordered after the dmbst() and write to the lock data structure releasing the lock. This is likely less of a problem but lock release needs to provide release semantics, and dmbst() is insufficient. For ex: A load in the critical section of CPU1 can be reordered after the store for the lock release, and it could read from a store that is executed on CPU2 in the same critical section, since CPU2 saw the store for the lock release first, and raced into the critical section. Also the dsb() after the write to the lock seems unnecessary. Am I missing something here ? It looks like the same issue is present even in bakery_lock_normal. Thanks Raghu

4 years, 5 months

1
0
0 0

Review hiccups

by Julius Werner

Hi Soby (and others), We've had a couple of MT8183 platform patches on Gerrit waiting for review/merge for a while (one over a week). Could you please help make sure they are in someone's queue at Arm? https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/2379 (passed CI) https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/2175 (passed CI) https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/2401 (need kick off CI) Let me know if there's anything else we can do to help increase turnaround time on Gerrit. Thanks, Julius

4 years, 5 months

2
1
0 0

How to verify image BL1

by Iñigo Vicente Waliño

Hi, I am using QEMu to simulate the ARMv8 platform. How can I verify that the BL1 image is in the ROM? And if not, can I not find it? Thanks, Iñigo.

4 years, 5 months

1
0
0 0

2024

2023

2022

2021

2020

2019

2018

TF-A November 2019