bewan : most uglier driver

matthieu castet castet.matthieu at
Mon Jul 18 14:44:22 EDT 2005


Gilles Espinasse wrote:
>>The pci driver seem to provide raw atm packet so I believe it could be
>>interresting to split usbatm into a raw atm driver and an usb part.
>>But the firmware seem to be excecuted on the host cpu, for that they
>>need to include in their driver a libm !!!
> it may be the case of the pci driver but this can't apply to the usb model
> ;-)

it seem it is also true for the usb chip :
$objdump -x arch/i386/modem_ant_USB_LINUX.o | grep -w sin
00000000         *UND*  00000000 sin
00035103 R_386_PC32        sin
0003512c R_386_PC32        sin
0003535d R_386_PC32        sin
00035381 R_386_PC32        sin

 > the pci flavor is a controller-less chip (meaning calculations are 
made on
 > the host computer)
So I think both usb and pci are controller-less chip.


More information about the Usbatm mailing list