Re: libusb+librump patch

[Antti Kantee]

On 16/10/15 16:56, Robert Millan wrote:
> El 15/10/15 a les 03:03, Bruno Félix Rezende Ribeiro ha escrit:
> > > OTOH I think this part of your patch:
> > >
> > > +  #define RUMP_SYS_OPEN
> > > +  #define RUMP_SYS_CLOSE
> > > +  #define RUMP_SYS_IOCTL
> > > +  #define RUMP_SYS_READWRITE
> > >
> > > is a bit dangerous. It would break any (current or future) usage of
> > > open() / close() / etc in that file which is not related to USB device
> > > nodes.
> >
> > I see.  What do you recommend?  #ifdefs for each occurrence?  Anyway,
> > I'm just playing around with it.
>
> I recommend explicit rump_sys_open(), e.g.
>
> int fd = rump_sys_open("/dev/ugenhc", RUMP_O_RDWR);
> if (fd == -1)
>       error(1, rump_errno2host(errno), "rump_sys_open");

The rump kernel ugenhc (= "ugen host controller") driver uses a 
/dev/ugen backend to emulate a USB host controller to the USB stack.  So 
the idea with ugenhc is that you can run the USB drivers in userspace, 
e.g. for development purposes, if your host supports /dev/ugen, and 
don't need to access the actual PCI devices directly.  So you can't open 
/dev/ugenhc, and not only because it doesn't exist.

If you don't have a USB stack on your host (which I assume is why we're 
having this discussion ;), ugenhc is completely useless.

If you want to do something like rump_sys_open("/dev/ugen"), you 
probably need the ugen driver as a rump kernel components (as was 
speculated in this thread a few days ago).