[smartmontools-support] State of OS/2 port

Dave Yeo dyeo at gmx.com
Wed Aug 5 00:52:43 CEST 2026


Christian Franke wrote:
> Hi Dave,
>
> thanks for the positive feedback.
>
> Dave Yeo via Smartmontools-support wrote:
>> Hi, compiling the latest smartmontools git on ArcaOS (OEM version of
>> OS/2, arcanoae.com) configure output this at the tail,
>
> Is there any free (evaluation?) version of this OS available we could
> use locally for compile tests?

No. It is at heart OS/2 4.52, like eComStation (eCs), which did have an 
evaluation version available and there are 4.52 ISO's floating around 
the internet. Toolchain is installed as RPM's so can be installed on 
4.52 or updated on eCs

> Or alternatively, is there any free cross-toolchain available?
>

Yes, recently available. https://github.com/komh/cross-os2emx/ which 
should easily build smartmontools. Only available for Linux and Mac 
currently.

>
>> ```
>> configure: WARNING:
>> This version of smartmontools does not use C++14, C++17 or C++20
>> enhancements,
>> but future versions possibly will.
>> This script was unable to determine a compiler option to enable C++20.
>> Please send info about compiler and platform to
>> smartmontools-support at listi.jpberlin.de - Thanks!
>
> There are no plans to require C++20 very soon, but possibly C++14. So
> the above is good news.
>
>
>> configure: WARNING:
>> The OS/2 port of smartmontools is unmaintained.  Please report the test
>> results to smartmontools-support at listi.jpberlin.de - Thanks!
>
> Didn't expect such a quick feedback after I added the above message one
> week ago (88240355037d).

Good luck as I just happened to do a svn up and ended up with the git 
repository.

>
> I will keep this message for a while but remove it for OS/2 (and keep
> QNX) before next release.
>

Good. I'm now subscribed here so any support I can give.

>
>> ```
>> Smartmontools compiled fine and seems to work fine.
>
> Very good news, the recent "guessed" fix (20d527bc4a78) was apparently
> successful.
>
>
>> I've attached the results of make check from the system's GCC, 9.2.0
>> and the latest available, 15.2.0. Not sure of other tests that I can do.
>
> Did you test `smartctl -x -a ...`?

I tested -x and -a separately and now together. I'll attach the output 
from my HD. My SSD recently died, which is why I updated hoping a newer 
version would show some info. Unluckily it died with no SMART errors, 
seems to be the interface, sometimes detected during boot sometimes not.
Unluckily currently no smart support in the NVMe driver and most of my 
stuff on the NVMe drive.
I'll attach the output of smartctl -x -a ahci1

>
> The os_os2.cpp module still uses the (meantime very) old interface which
> limits ATA pass-through to traditional SMART functionality. So `smartctl
> -x` will likely show many messages like:
> "Non-SMART commands not implemented"
>

Surprisingly few, see log. Perhaps I should plug in an SSD to try?
Dave

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smartctl.log
Type: application/x-log
Size: 7617 bytes
Desc: not available
URL: <https://listi.jpberlin.de/pipermail/smartmontools-support/attachments/20260804/090b8d7e/attachment-0001.bin>


More information about the Smartmontools-support mailing list