On-going portability issue with 9.9p2

Darren Tucker dtucker at dtucker.net
Sat Mar 1 16:40:20 AEDT 2025


On Sat, 1 Mar 2025 at 14:23, Dennis Clarke via openssh-unix-dev <
openssh-unix-dev at mindrot.org> wrote:

>
> So, I applied the patch from :
>
>
> https://lists.mindrot.org/pipermail/openssh-unix-dev/2025-February/041816.html
>
> Then did the usual autoreconf dance and then configure fails :
>

Which version of autoconf did you use?


> configure: error: cannot run /opt/bw/bin/bash ../openssh-9.9p2/config.sub
>

Take a look in config.log to see what the error is, but my guess is
config.sub is missing or corrupt.  I suspect something went wrong in the
autoreconf step.

You can try this patch (
https://www.dtucker.net/tmp/openssh-mlkem-popcount.patch) against the V_9_9
branch from github or mindrot, which is the same thing except it includes
the autoconf changes.  It seems to work for me (on Linux anyway, while I do
have a sparc it is not easily accessible, and doesn't have the Sun
compiler anyway).

$ git clone https://github.com/openssh/openssh-portable/
$ cd openssh-portable
$ git checkout V_9_9
$ lynx -source https://www.dtucker.net/tmp/openssh-mlkem-popcount.patch |
patch -p1
patching file config.h.in
patching file configure
patching file configure.ac
patching file libcrux_mlkem768_sha3.h
patching file mlkem768.sh
$ ./configure | grep popc
checking whether compiler supports __builtin_popcount... yes


> Really some inline ASM would get this done.
>

I don't think anyone in the OpenSSH team is interested in maintaining
assembler for any architectures.

-- 
Darren Tucker (dtucker at dtucker.net)
GPG key 11EAA6FA / A86E 3E07 5B19 5880 E860  37F4 9357 ECEF 11EA A6FA
    Good judgement comes with experience. Unfortunately, the experience
usually comes from bad judgement.


More information about the openssh-unix-dev mailing list