speedtch cxacru2.c, NONE, 1.1 cxacru2.h, NONE, 1.1 speedtch2.c, NONE, 1.1 speedtch2.h, NONE, 1.1 usbatm2.c, NONE, 1.1 Makefile, 1.8, 1.9 usbatm.c, 1.1, NONE

[matthieu castet]

Hi,

David Woodhouse wrote:
> On Thu, 2005-01-20 at 21:30 +0100, Duncan Sands wrote:
> 
>>>Sorry for asking again, but what does it buy you?
>>
>>I will answer this (and your previous mail) later.  For now, I would
>>prefer to just work on it.
> 
> 
> FWIW I'm with Roman on this one. We had a library which was available
> for the hardware-specific drivers to use as appropriate. Now AFAICT we
> have a single monolithic driver module which optionally contains support
> for various different types of hardware. That seems like a step
> backwards to me. 
> 

The new model is also problematic for driver alreading using driver info ...
Also some usb device comported like 2 usb devices. The first device is 
when there is no firmware, and when you give it the firmware, the device 
disconnect and appear as a new device (different id).
So in this case we need 2 driver (usbatm + the one for prefirmware 
device), instead of one before...

Now when you load all the driver code so you end with a big module, even 
if you want to use only one driver.


I think the library model was better and even can be split in a raw atm 
lib + usb lib, now it try to implement a driver model and is very 
restrictive.

I don't see any benefit of it, and I hope the future model will be 
clarified, because I won't spent time for finally having to use an 
something that won't fit for eagle-usb...


Matthieu