[ofa-general] RE: [ofw] RE: porting IB management code to Windows

[Sean Hefty]

>I think you have to appreciate that the code written for Linux is
>written in C99 and for the POSIX API. WDK does not provide that kind
>of environment and trying to bake over that with #ifdef is a huge
>ongoing job. Just look at what Sean has done already, he had to
>provide an independent getopt implementation just for a tiny utility.

Note that the management code has an OS abstraction layer in the form on
complib.

Probably the biggest hurdle to porting the code is the use of the sys/class
filesystem.  We have some ideas on how we might work around this, but we haven't
studied the actual use of sys/class.

>MS already has a C99 POSIX environment they support and distribute, it
>is called SFU and the bundled Intrix library set.

I like to distinguish between build requirements, versus requirements placed on
end-users.  Requiring users to download and install SFU separate from WinOF is
something that I would like to avoid, if reasonable to do so.

So far, nothing that we've seen blocks the ability to share the majority of the
libibmad or diag code.  (Arlin can provide more details on changes.)  However,
if the management package is to support multiple OS's, then the code will
require some modifications to support that.  I'd like to think that the
companies involved here would like to support 

How flexible are people willing to be?  Are there objections on the Windows side
to taking IB management binaries and including them as part of WinOF?  Are there
objections on the Linux side from accepting patches that allow the code to be
built using the WDK?

- Sean