Re: [Libusb-devel] Patch: libusb/os/linux_usbfs.c: Distingush between usbfs and other kernel mode drivers

[Michael P. <mic...@gm...>]

Uwe Bonnes wrote:
>> @@ -1242,7 +1242,8 @@ static int op_kernel_driver_active(struct
libusb_device_handle *handle,
>> 			"get driver failed error %d errno %d", r, errno);
>> 		return LIBUSB_ERROR_OTHER;
>> 	}
>> -
>> +	if (!(strcmp(getdrv.driver,"usbfs")))
>> +	      return 2;
>>  	return 1;
>>  }

Is a typical use then:

  int isactive = libusb_kernel_driver_active(...);
  // point A
  if(isactive==1) {
  	libusb_detach_kernel_driver(...);
  } else if(isactive==2) {
  	printf("in use\n");
  	exit(...);
  }
  libusb_claim_interface(...);
  // point B

?  If so....


Now what if two such programs are running?  Say program 1 checks and find
isactive==1, and stops at point A.  Program 2 checks and find isactive==1,
detaches ftdi_sio and claims the interface, and stops at point B (or later).
Now program 1 resumes, detaches the usbfs of program 2, and claims the
interface itself.  If I understand the original thread from Apr 2009
correctly, we're back where we started, if only rarely.

Is there also a similar situation when reattaching at the end?

Michael