[Openib-windows] RE: MTHCA driver changes

[Fabian Tillier]

On 3/26/06, Leonid Keller <leonid at mellanox.co.il> wrote:
>
> > -----Original Message-----
> > From: ftillier.sst at gmail.com [mailto:ftillier.sst at gmail.com]
> > On Behalf Of Fabian Tillier
> > Sent: Saturday, March 25, 2006 1:38 AM
> > To: Leonid Keller
> > Cc: openib-windows at openib.org
> > Subject: MTHCA driver changes
> >
> > Before I start spending time on the MTHCA driver,
>
> It's up to you, how to use your precious time.

:)

> >I'd like to understand what kind of changes you will accept, and what
> > kind are just a waste of my time in putting them together.
>
> I'm open to any suggestions, but those, intended for increasing the
> driver's  effectiveness, stability and functionality are superior wrt.
> those, which just decrease the number of lines.

Ok, great.  I'll just send whatever I end up changing.

> > How much are you planning on keeping this code a straight
> > copy from the Linux project? Will you take and apply
> > Windows-specific patches to make the code more streamlined
> > and Windows oriented?
> > You have your own:
> > - time stamp abstraction (and ugly at that)
> > - spinlock abstraction
> > - wait_queue, semaphore, and mutex abstraction
> > - types
> > - memory abstraction
> > - pci bus abstraction
>
> These were/are not new abstractions, just a way to keep similarity to
> Linux code for the time being for facilitating bugfix merges.

I had never seen them before, so they were new to me!

> It was planned from the beginning to replace them gradually by Windows
> code.

Ok, that's what I wanted to know - it wasn't clear how much you
planned on keeping this source in sync with the Linux driver (i.e.
whether you would be open to changes or not).  Thanks for clearing
this up.

- Fab