Re: [MiNT] [PATCH] Export function XHNewCookie to kernel modules

[Alan Hourihane <alanh@fairlite.co.uk>]

On Mon, 2010-12-13 at 14:27 +0100, David Gálvez wrote:
> Hi Alan
> 
> 2010/12/13 Alan Hourihane <alanh@fairlite.co.uk>:
> > Sorry the the "self" conversation.
> >
> 
> No problem, just kidding :-)
> 
> > Getting back to your two points. I think a TSR before MiNT is loaded, is
> > a natural placement for this, just as HDDriver etc are loaded before
> > MiNT too. I guess we'd want to support just TOS, so a TSR makes sense.
> > Assuming this is the MASS storage USB stuff you are talking about.
> >
> 
> I'm very surprised that you and Vincent advise me to go for the TSR
> solution, I thought that MiNT philosophy was the opposite way, getting
> rid of all TSR and these kind of stuff. Integrating all those things
> into the kernel or as modules.

I guess my opinion for a TSR is trying to understand whether we just
support MiNT, or TOS as well. As I could see some people using just TOS
and no MiNT to access USB devices.

> > As for the alternative, if it were MiNT only, I'd suggest an XDD.
> >
> 
> Why a XDD? Wouldn't be better a KM?

Possibly. I've dealved into your code to understand the direction, so
I'll defer to you and say a KM is fine :-)

> > Can you explain again what you are currently doing that necessitates the
> > patch ?
> >
> 
> Yes of course, my final aim is to have a USB stack for MiNT, as this
> can take long, I try to go through steps that end with something
> useful so others can use. First step was an USB mass storage driver,
> now I'm implementing this driver as a kernel module, this is why i
> need the patch to be applied. In the future I would like to have a usb
> core, separate modules for each USB hardware (PCI, EtherNAt,
> NetUSBee,..) and also for each USB device driver (mass storage,
> keyboard, mouse, ...).

Sounds good. I'll apply the patch and we can see where things take us.

Alan.