Yutaka Niibe gniibe@identi.ca

Maebashi, Japan

GNU wannabee ['gnu:be], FSIJ chairman, Debian Developer

  • GnuPG 2.3.1 is released

    2021-04-21T00:23:02Z via Pumpa To: Public CC: Followers

    Bug fixing (mostly for build) release. It's a worth try.

    Scorpio shared this.

  • Libgcrypt 1.9.3 is released

    2021-04-19T23:46:45Z via Pumpa To: Public CC: Followers

    It would be good if this was the 1.9.0. I mean, there are more fixes for regressions (than we had expected).

    Good things are the build process has been improved.
  • Ploting X448

    2021-04-12T02:08:16Z via Pumpa To: Public CC: Followers

    Green plots are the first 200 points of X448.
    Green star is the generator.
  • GnuPG 2.3.0 released

    2021-04-07T23:54:08Z via Pumpa To: Public CC: Followers

    It comes with Ed448/X448 support.
  • Sway

    2021-04-06T05:45:29Z via Pumpa To: Public CC: Followers

    I start using Sway to test Wayland on this machine.
    It seems that input method support is not that mature in the protocol.

    Thus, I fixed my implementation (of Emacs Lisp) for Japanese input.
  • Ed448 for Gnuk

    2021-03-17T23:50:52Z via Pumpa To: Public CC: Followers

    Running 100 iterations, signing on FST-01 (Cortex-M3 @ 72MHz), it's like:

    real 0m15.929s
    user 0m0.973s
    sys 0m0.262s

    Yes, generated signature is good:
    ~$ gpg --verify /tmp/a1.asc /etc/motd
    gpg: NOTE: THIS IS A DEVELOPMENT VERSION!
    gpg: It is only intended for test purposes and should NOT be
    gpg: used in a production environment or with production keys!
    gpg: Signature made Thu Mar 18 08:47:45 2021 JST
    gpg: using EDDSA key 91B78334699E9155A9E0651CBE352A1DCF455BFF
    gpg: Good signature from "Kunisada Chu " [ultimate]
  • X448 for Gnuk

    2021-03-09T05:41:52Z via Pumpa To: Public CC: Followers

    Running 100 iterations, decrypting on FST-01G (Cortex-M3 @ 72MHz), it's like:

    gpg: encrypted with cv448 key, ID F6FA616F8E4378B9, created 2021-03-09
          "Kunisada Chu <chuji@gniibe.org>"
    
    real    0m34.189s
    user    0m2.512s
    sys 0m0.387s
    gniibe@iwagami:~/work/gnuk/gnuk/src$ 
    

    That is, 0.34189 second for a single decryption process.

    It's good enough for me.

  • Gnuk 1.2.17

    2021-02-26T05:17:30Z via Pumpa To: Public CC: Followers

    Mainly for GNU/Linux emulation, new Gnuk is out.

    https://lists.gnupg.org/pipermail/gnuk-users/2021-February/000293.html
  • sub and subs instruction (pre-UAL syntax, UAL syntax)

    2021-02-25T03:12:26Z via Pumpa To: Public CC: Followers

    I didn't know well about two different syntax in ARM assembler.

    For UAL syntax (newer, which is usually used), it requires an explicit "s" suffix to specify flags being set, while pre-UAL syntax (older, which is used for Cortex-M0) assumes that implicitly.

    In GCC, we have -masm-syntax-unified option.
  • libgcrypt 1.9.2 released

    2021-02-17T23:39:28Z via Pumpa To: Public CC: Followers

    Well, 1.9.0 had point-zero release problem (= no one tests well until its release).

    Now, it has support of Ed448/X448 curves.

    Enjoy,

    Alexandre Oliva, nukem likes this.

  • GD32VF103 and its possible use for Gnuk

    2020-09-30T23:59:47Z via Pumpa To: Public CC: Followers

    GD32VF103, which comes with RISC-V core, is cheap and fast.

    I ported Chopstx to GD32VF103, and am considering if it's good to port Gnuk.

    Last month, when I found riscv-openocd merged the support of GD32VF103, I tried to use the debugger interface.

    And... for my version (which I bought from Seeed), the Longan Nano, I can access its flash ROM content.
    If my analysys is correct, it's a kind of matter for CVE, I suppose.

    With help of my Chinese friends, I try to contact the vendor.

    Happy Chinese Holiday Season!

    Alexandre Oliva likes this.

    And then, I was informed for this having many CVEs:
    https://github.com/JohannesObermaier/f103-analysis/

    Well, mine is independent, and more easier to be attacked.

    Yutaka Niibe at 2020-10-02T01:46:05Z

    geekable likes this.

  • Supporting Ed448 (GnuPG)

    2020-06-24T01:21:24Z via Pumpa To: Public CC: Followers

    I pushed the change to support Ed448 to GnuPG master:
    https://dev.gnupg.org/rGa763bb2580b0d586a80b8ccd3654f41e49604f4f
  • Ed448 in libgcrypt

    2020-06-10T23:59:57Z via Pumpa To: Public CC: Followers

    X448 feature has been pushed to GnuPG master.

    Now, I'm adding the support of Ed448 to libgcrypt so that it will be also supported by GnuPG.
  • Supporting X448 (GnuPG)

    2020-06-10T23:58:44Z via Pumpa To: Public CC: Followers

    I pushed the change to support X448 to GnuPG master.
    UI (displaying the name of algo, selecting it from tty or command line) needs some more changes, though.
  • Bouncing GnuPG Logo

    2019-12-26T23:33:30Z via Identi.ca Web To: Public CC: Followers

    It's running on the Longan-Nano board with RISC-V MCU named GD32VF103 (Bumblebee core).

    I took a video, but I don't know how we can share a video with free software friendly way.
    So, it's an image only for now.

    salyavin likes this.

    Just for the record, you can upload videos here, to Pump.io, with a client capable of it, like Dianara.


    And there's always a Mediagoblin or Nextcloud instance 😃

    JanKusanagi at 2019-12-27T13:01:08Z

    salyavin likes this.

  • "FSIJ" logo on 160x80 screen by RISC-V MCU

    2019-12-01T23:49:16Z via Identi.ca Web To: Public CC: Followers

    Longan Nano comes with LCD display, which is connected through SPI.
    I wrote a SPI driver for Chopstx.
    This is an example.

    salyavin likes this.

  • 2019-11-26T00:23:14Z via Identi.ca Web To: Public CC: Followers

    Last week, I renewed IEEE membership, and got a sticker saying "KEEP CALM I'M IEEE MEMBER".

    Then, yesterday, I received FSF Bulletin.  #fsfbulletin
    It was sent on November 11th, the one of binary days.  :-)

    McClane likes this.

  • 2019-11-12T06:11:30Z via Identi.ca Web To: Public CC: Followers

    Chopstx is running on RISC-V MCU now.
    I got the board named "Longan-nano", which uses GD32VF103.

    Alexandre Oliva likes this.

  • Experiment with ICE40 for SHA256

    2019-09-04T23:41:02Z via Identi.ca Web To: Public CC: Followers

    With Migen, I wrote a SHA256 module.