- mbed-tls - lists.trustedfirmware.org

MbedTLS - which protocols are advisable to support, and memory usage

by Peter

Hi All, I am new here. With a colleague, I am working on a product which previously used PolarSSL but was later changed to MbedTLS which, in the ST ARM 32F417 implementation, is believed to be less buggy. It is working ok as far as we have got, but we are finding that as well as needing some 100k more FLASH, it needs nearly 50k of RAM, which on the 32F417 (128k RAM) would normally be OK but due to other required functionality it leaves us just 10k. The architecture we are using for MbedTLS is a 48k (48k was found to be the smallest that works) static buffer within which TLS runs its own heap. I am aware that in the most generic case one has to have enough for one 16k buffer (plus a bit) for HTTPS, but what concerns me is that there appears to be no way to determine the worst case memory usage, across various usage scenarios. It would also be great to reduce this 48k to say 30k. Some of the protocols we know we don't need (e.g. PPP) but within any of them there is the question of which cipher suite needs supporting. What is the minimum cipher suite required to be able to use MbedTLS as an HTTPS client for use with typical current cloud-based file storage or data logging APIs such as Dropbox, Google Drive, Loggly etc? And what is the recommended minimum cipher suite required to implement an HTTPS server that would be able to negotiate a session with most current web-browsers? The goal is to ideally reduce both RAM and code size requirements for use in an embedded device that needs to work both as a general purpose server and client, without requiring support for multiple concurrent sessions. The other thing is that even the ST port of MbedTLS doesn't appear to make use of ST CPU hardware features such as AES256, DES, etc, which the 32F417 has in hardware. This speeds things up hugely but much more importantly in our case, saves a lot of RAM. For example AES256 can use about 10k if done in software. Thank you very much in advance for any input on what can be done. Peter

3 years, 11 months

1
0
0 0

Question about a research project: ESP32 mbed-tls

by Marco Portoni

Hello, I am Marco Portoni and I am working on a thesis research for the University of Study of Milan. It is focused on cryptographic libraries in the IoT world, and my question is: is it possible to run the benchmark on a ESP32? I have already runned the benchmark on 2 of my laptops to test and collected the results, but my ultimate goal is to make the benchmark run on the ESP. Is it possible to do it? Thanks for every help! Marco

3 years, 11 months

1
0
0 0

Build shared on macOS?

by Felipe Gasper

Hello, I’m trying to build shared from git on macOS and having no luck. I’m doing: > mkdir build; cd build > cmake DUSE_SHARED_MBEDTLS_LIBRARY=On .. … > make … > ls -la library total 2944 drwxr-xr-x 10 felipe staff 320 Dec 1 08:49 . drwxr-xr-x 15 felipe staff 480 Dec 1 08:49 .. drwxr-xr-x 8 felipe staff 256 Dec 1 08:49 CMakeFiles -rw-r--r-- 1 felipe staff 92755 Dec 1 08:49 Makefile -rw-r--r-- 1 felipe staff 2942 Dec 1 08:49 cmake_install.cmake -rw-r--r-- 1 felipe staff 34055 Dec 1 08:49 error.c -rw-r--r-- 1 felipe staff 916112 Dec 1 08:49 libmbedcrypto.a -rw-r--r-- 1 felipe staff 315856 Dec 1 08:49 libmbedtls.a -rw-r--r-- 1 felipe staff 104072 Dec 1 08:49 libmbedx509.a -rw-r--r-- 1 felipe staff 26265 Dec 1 08:49 version_features.c Am I just missing something? It seems like I’m following what the README says to do; I also did `export SHARED=1` just in case, but that made no difference. Thank you in advance! -FG

3 years, 11 months

3
3
0 0

Re: [mbed-tls] End of Life of LTS branch 2.16

by Dave Rodgman

Hi Radhika, We do not plan to extend support for Mbed TLS 2.16 beyond the three year support period. We do plan to release a 2.x LTS (probably 2.28) alongside the final 2.16 release, so that we always have an LTS available. Regards Dave Rodgman From: mbed-tls <mbed-tls-bounces(a)lists.trustedfirmware.org> on behalf of Radhika Jandhyala via mbed-tls <mbed-tls(a)lists.trustedfirmware.org> Reply to: Radhika Jandhyala <radhikaj(a)microsoft.com> Date: Monday, 22 November 2021 at 10:50 To: "mbed-tls(a)lists.trustedfirmware.org" <mbed-tls(a)lists.trustedfirmware.org> Subject: [mbed-tls] End of Life of LTS branch 2.16 Hi, I am a maintainer for https://github.com/openenclave/openenclave. We include mbedtls 2.16 as a submodule in our project. Reading the article below, it states that 2.16 support will last at least till Dec 2021. Will you continue to support 2.16 in 2022? https://tls.mbed.org/tech-updates/blog/announcing-lts-branch-mbedtls-2.16 We are specifically interested in CVE fixes being backported to 2.16 and the timeframe where that will continue. Is there an updated timeline for this? We do plan to upgrade to 3.0 in early 2022, but we anticipate that will take some effort due to breaking changes. Thank you very much! Radhika

3 years, 11 months

2
2
0 0

Mbed TLS long term support release plans

by Dave Rodgman

Hi, Mbed TLS 2.16 LTS is approaching the end of its support period – it was originally announced that it would be maintained for at least 3 years up until the end of 2021 – and currently there is no other LTS branch. To ensure that there is no period of time without a supported LTS branch, we plan to release 2.28 LTS in the near future (which will have the usual 3 year support period). We will continue supporting 2.16 LTS until this is available. The expectation is that as 2.28 will be API-compatible with 2.16, users of 2.16 LTS will be able to upgrade to 2.28 very easily. Progress towards 2.28 LTS can be followed here: https://github.com/orgs/ARMmbed/projects/18#column-15836286 Dave Rodgman

3 years, 11 months

1
0
0 0

End of Life of LTS branch 2.16

by Radhika Jandhyala

Hi, I am a maintainer for https://github.com/openenclave/openenclave. We include mbedtls 2.16 as a submodule in our project. Reading the article below, it states that 2.16 support will last at least till Dec 2021. Will you continue to support 2.16 in 2022? https://tls.mbed.org/tech-updates/blog/announcing-lts-branch-mbedtls-2.16 We are specifically interested in CVE fixes being backported to 2.16 and the timeframe where that will continue. Is there an updated timeline for this? We do plan to upgrade to 3.0 in early 2022, but we anticipate that will take some effort due to breaking changes. Thank you very much! Radhika

3 years, 11 months

1
0
0 0

PSA Crypto non-optional in Mbed TLS 3.1

by Manuel Pegourie-Gonnard

Hi, We'd like to announce a change we intend to make: starting with Mbed TLS 3.1, it will no longer be possible to build with TLS, X.509, or PK without support for PSA Crypto. Details: currently, use of PSA Crypto APIs from the TLS, X.509 and PK layers is controlled by the option MBEDTLS_USE_PSA_CRYPTO. When this option is disabled (which is the default), it's possible to build without MBEDTLS_PSA_CRYPTO_C. Starting with 3.1, we intend to start making TLS, X.509 and PK use PSA Crypto unconditionally, so MBEDTLS_PSA_CRYPTO_C will be automatically enabled in the build as soon as TLS, X.509 or PK is enabled. Impact: for users who already build with PSA Crypto enabled (the default), no impact. For users who currently disable MBEDTLS_PSA_CRYPTO_C in their configuration, starting with 3.1 there will be no functional changes, but the size of the built library will increase due to the additional features enabled. The increase depends on the configuration, application, platform, and toolchain, but the order of magnitude currently ranges from about 9 KB for a minimal TLS configuration with LTO (link-time optimisation) to about 30 KB for a full configuration without link-time garbage collection (though size-constrained devices are very unlikely to use such a configuration); we aim to reduce this overhead in the future. Rationale: maintaining two versions of every cryptographic operation in upper layers (one PSA, one non-PSA) imposes a significant burden on new developments. By removing that burden, we hope to progress faster on things that are more important in the long run, including better integration of PSA Crypto in TLS and X.509, and future code size optimisation of PSA Crypto. We realise the impact on users who were excluding PSA Crypto is significant. We're going to release Mbed TLS 2.28 in the coming months, which is not affected by this change and will receive bug fixes and security fixes for at least 3 years; we hope it provides an acceptable fall-back solution to affected users. In the long run, PSA Crypto will become the only Crypto API we offer; we want to make its footprint as small as possible considering its feature set and we hope this change will enable us to make faster progress towards that goal. Best regards, Manuel Pégourié-Gonnard for the Mbed TLS team.

3 years, 12 months

1
0
0 0

About `MBEDTLS_ALLOW_PRIVATE_ACCESS`

by Aditya Patwardhan

Hi, While working on updating mbedtls to v3.0, I saw that the internal fields in MBEDTLS have been made private. ( ref- here<https://github.com/ARMmbed/mbedtls/blob/development/docs/3.0-migration-guid…>). It says there and I quote "As a last resort, you can access the field foo of a structure bar by writing bar.MBEDTLS_PRIVATE(foo). Note that you do so at your own risk, since such code is likely to break in a future minor version of Mbed TLS.” I just wanted to know if there is any alternative solution to this, rather than using `MBEDTLS_PRIVATE` everywhere. I know the next release of mbedtls probably plans to fix this with appropriate solution. But we wanted to push out our feature branch As early as possible. I saw in the PR<https://github.com/zephyrproject-rtos/zephyr/pull/37753> in zephyr RTOS about updating to mbedtls-3.0. They have just added this line<https://github.com/ceolin/zephyr/blob/5bf3128a9703561e578651218f5bcdafb96f8…> #if !defined(MBEDTLS_ALLOW_PRIVATE_ACCESS) #define MBEDTLS_ALLOW_PRIVATE_ACCESS # endif Is this an alternative solution to using `MBEDTLS_PRIVATE` for accessing a private field. If yes, can somebody please point me to respective document as this would greatly reduce our code-changes. I understand there is some discussion going on in mbedtls about this change, and appropriate getter-setter functions shall be provided. But we wanted to push out out feature branch for early testing. Thanks and Regards, Aditya

3 years, 12 months

2
2
0 0

Mbed TLS Tech Forum

by Dave Rodgman

Hi, On Monday 8th we will have the next Mbed TLS Tech Forum. Please reply to the list if you have any agenda topics to raise. As a starter for the agenda, we will likely discuss: https://github.com/ARMmbed/mbedtls/pull/5067 Proposal for driver dispatch code gen Dave Rodgman Zoom details below. The call will be recorded and made available on the TF website : https://www.trustedfirmware.org/meetings/ Join Zoom Meeting https://linaro-org.zoom.us/j/99948462765?pwd=SGlHYlF1Z2owUDNFWWppaGlSRDh5UT… Meeting ID: 999 4846 2765 Passcode: 196117 One tap mobile +12532158782,,99948462765# US (Tacoma) +13462487799,,99948462765# US (Houston) Dial by your location +1 253 215 8782 US (Tacoma) +1 346 248 7799 US (Houston) +1 669 900 9128 US (San Jose) +1 301 715 8592 US (Washington DC) +1 312 626 6799 US (Chicago) +1 646 558 8656 US (New York) 888 788 0099 US Toll-free 877 853 5247 US Toll-free Meeting ID: 999 4846 2765 Find your local number: https://linaro-org.zoom.us/u/anpWWkRdt

4 years

1
0
0 0

mbedtls client: TLS 1.2 Client Hello triggers a -0x7780 fatal internal server error...?

by secretimap＠nickpelling.com

Hi everyone, The company I’m working for uses mbedtls to drive an embedded TLS server (for HTTPS): but now needs the same device to also act as a TLS client for logging in to remote servers (e.g. for LDAP etc). My embedded test code tries (but fails) to connect to a remote HTTPS server on port 443: I can connect to the same server via openssl on a desktop command line, so it seems to be working. Basically, the client hello message goes out fine (for brevity, I’ve skipped all the ciphersuites, but note that the list does contain the 0xC030 ciphersuite that desktop openssl is able to connect to OK): .ssl_cli.c:934: client hello, got 25 ciphersuites (excluding SCSVs) .ssl_cli.c:943: adding EMPTY_RENEGOTIATION_INFO_SCSV .ssl_cli.c:992: client hello, compress len.: 1 .ssl_cli.c:994: client hello, compress alg.: 0 .ssl_cli.c:186: client hello, adding signature_algorithms extension .ssl_cli.c:271: client hello, adding supported_elliptic_curves extension .ssl_cli.c:336: client hello, adding supported_point_formats extension .ssl_cli.c:518: client hello, adding encrypt_then_mac extension .ssl_cli.c:552: client hello, adding extended_master_secret extension .ssl_cli.c:585: client hello, adding session ticket extension .ssl_cli.c:1071: client hello, total extension length: 52 0000: 16 03 01 00 95 01 00 00 91 03 03 16 e9 7a ea aa 0010: 31 81 60 b6 a8 d3 32 93 a4 50 ca f3 e0 66 f8 47 0020: 56 95 0c cd a0 db b6 0e ae a7 d4 00 00 34 c0 30 0030: 00 9f c0 ad c0 9f c0 24 c0 28 00 6b c0 0a c0 14 0040: 00 39 c0 af c0 a3 c0 2b c0 2f 00 9e c0 ac c0 9e 0050: c0 23 c0 27 00 67 c0 09 c0 13 00 33 c0 ae c0 a2 0060: 00 ff 01 00 00 34 00 0d 00 16 00 14 06 03 06 01 0070: 05 03 05 01 04 03 04 01 03 03 03 01 02 03 02 01 0080: 00 0a 00 04 00 02 00 17 00 0b 00 02 01 00 00 16 0090: 00 00 00 17 00 00 00 23 00 00 However, the server hello that comes back is both unhappy and unhelpful: .ssl_tls.c:2721: in_left: 5, nb_want: 7 .ssl_tls.c:2722: ssl->f_recv(_timeout)() returned 2 (-0xfffffffe) .ssl_tls.c:2742: <= fetch input .ssl_tls.c:4233: dumping 'input record from network' (7 bytes) .ssl_tls.c:4233: 0000: 15 03 03 00 02 02 50 ......P .ssl_tls.c:5170: got an alert message, type: [2:80] .ssl_tls.c:5178: is a fatal alert message (msg 80) .ssl_tls.c:4369: mbedtls_ssl_handle_message_type() returned -30592 (-0x7780) .ssl_cli.c:1506: mbedtls_ssl_read_record() returned -30592 (-0x7780) Putting the ciphersuites to one side, I suspect that what is happening here is that the (remote) server is complaining about a TLS extension the mbedtls client is either including or omitting. (I’ve already disabled the max fragment extension). But because I can’t see inside the remote server, I can’t tell. ☹ Might the mbedtls TLS client need the supported_versions extension (43) to work with many remote TLS servers? Is this a known issue? I’ve gone through loads of mailing list archive messages here, but haven’t found anything that matches this. Thanks, Nick PS: when OpenSSL connects to the server, the client hello message (that works fine) looks like this in Wireshark: Transport Layer Security TLSv1.2 Record Layer: Handshake Protocol: Client Hello Content Type: Handshake (22) Version: TLS 1.2 (0x0303) Length: 358 Handshake Protocol: Client Hello Handshake Type: Client Hello (1) Length: 354 Version: TLS 1.2 (0x0303) Random: [SNIP] GMT Unix Time: Jan 21, 2095 20:50:07.000000000 GMT Standard Time Random Bytes: [SNIP] Session ID Length: 32 Session ID: [SNIP] Cipher Suites Length: 90 Cipher Suites (45 suites) [SNIP!] Cipher Suite: TLS_EMPTY_RENEGOTIATION_INFO_SCSV (0x00ff) Compression Methods Length: 1 Compression Methods (1 method) Compression Method: null (0) Extensions Length: 191 Extension: status_request (len=5) Type: status_request (5) Length: 5 Certificate Status Type: OCSP (1) Responder ID list Length: 0 Request Extensions Length: 0 Extension: supported_groups (len=22) Type: supported_groups (10) Length: 22 Supported Groups List Length: 20 Supported Groups (10 groups) Supported Group: x25519 (0x001d) Supported Group: secp256r1 (0x0017) Supported Group: secp384r1 (0x0018) Supported Group: secp521r1 (0x0019) Supported Group: x448 (0x001e) Supported Group: ffdhe2048 (0x0100) Supported Group: ffdhe3072 (0x0101) Supported Group: ffdhe4096 (0x0102) Supported Group: ffdhe6144 (0x0103) Supported Group: ffdhe8192 (0x0104) Extension: ec_point_formats (len=2) Type: ec_point_formats (11) Length: 2 EC point formats Length: 1 Elliptic curves point formats (1) EC point format: uncompressed (0) Extension: signature_algorithms (len=34) Type: signature_algorithms (13) Length: 34 Signature Hash Algorithms Length: 32 Signature Hash Algorithms (16 algorithms) [SNIP!] Extension: signature_algorithms_cert (len=34) Type: signature_algorithms_cert (50) Length: 34 Signature Hash Algorithms Length: 32 Signature Hash Algorithms (16 algorithms) [SNIP!] Extension: status_request_v2 (len=9) Type: status_request_v2 (17) Length: 9 Certificate Status List Length: 7 Certificate Status Type: OCSP Multi (2) Certificate Status Length: 4 Responder ID list Length: 0 Request Extensions Length: 0 Extension: extended_master_secret (len=0) Type: extended_master_secret (23) Length: 0 Extension: supported_versions (len=5) Type: supported_versions (43) Length: 5 Supported Versions length: 4 Supported Version: TLS 1.3 (0x0304) Supported Version: TLS 1.2 (0x0303) Extension: psk_key_exchange_modes (len=2) Type: psk_key_exchange_modes (45) Length: 2 PSK Key Exchange Modes Length: 1 PSK Key Exchange Mode: PSK with (EC)DHE key establishment (psk_dhe_ke) (1) Extension: key_share (len=38) Type: key_share (51) Length: 38 Key Share extension Client Key Share Length: 36 Key Share Entry: Group: x25519, Key Exchange length: 32 Group: x25519 (29) Key Exchange Length: 32 Key Exchange: 9f8e83a6859e6b4c6c3e43b524b81e6f63cc409c5757692a2343a2929c681c17

4 years

1
0
0 0

2025

2024

2023

2022

2021

2020

mbed-tls