With our LicenseChoice we need to ensure that GPL code is not mixed with double-licenced source file.

The files to review are for the ueagle-atm's driver ueagle.c , and to be compared to eagle-usb's following files : eu_main.c eu_utils.c

a diff may not be relevant...

Analysis / comparison of function by function :

• ueagle.c - static int uea_send_modem_cmd(struct usb_device *usb, u16 addr, u16 count, u8 * buff)
• eu_utils.c - int eu_cmd_to_modem (eu_instance_t *ins, uint32_t cmd, uint32_t idx, uint32_t count, uint8_t *data)
  • variable xfer_buff is the only common point (not same type...), not same content nor size allocated
  • structure of code is different (use of goto byebye in one case, not in the other)
  • use of queue in eu_utils.c, use of msg in ueagle.c=> code is different (hum normal => they do not do the same things...)
    
• eu_utils.c - static int eu_send_modem_cmd (eu_instance_t *ins, uint16_t addr, uint16_t count, uint8_t *buff)
  • use eu_instance_t *ins (instead of struct usb_device *usb)
  • call to usb_control_msg is quite the same, one difference being that CTRL_TIMEOUT is not hardcoded in ueagle.c)
  • ret is not managed in eu_utils.c whereas it is => structure of code is different, objects used are not the same, functions called are only similar

• ueagle.c - static int uea_load_firmware(struct usb_device *usb, unsigned int ver)
• eu_utils.c - int eu_load_firmware ( eu_instance_t *ins, uint32_t pid, unsigned int *ver )
  • load a firmware-in-a-file in ueagle.c
  • load a firmware-in-a-variable in eu_utils.c
  • De-assert reset : well the mechanism to manage usb is the same, the commands are nearly identical, though the code structure is different : print success in ueagle.c (not in eu_util.c)=> code is different

• ueagle.c - static int uea_getesi(struct uea_softc *sc, u_char * esi) - get mac address
• eu_utils.c - void eu_get_mac_addr ( eu_instance_t *ins )
  => not the same code

• ueagle.c - static int uea_read_proc(char *page, char **start, off_t off, int count, int *eof, void *data) - display information
• eu_main.c - static int eu_read_proc ( char *page, char **start, off_t off, int count, int *eof, void *data)
  • display is the same (hence same sprintf) - checksum of DSPcode could be added (see mail by ADI)
  • ueagle.c does not use the same structure
  • comments could be more meaningful in ueagle.c
  • less states of modem are managed in ueagle.c (why ?)
  • what is the overhead with 4K page in eu_main.c ? is it useful ?=> display functions are similar, structure of code is simpler in ueagle.c (rewritten)

• ueagle.c - static int __init uea_init(void)
• eu_main.c - static int __init eu_init (void)
  • error message says /proc/driver/eagle-usb/ instead of ueagle-atm (after name changed) => correction needed (minor)
  • module init code structure is standard (to be compared to other drivers)

• ueagle.c - static void __exit uea_exit(void)
• eu_main.c - static void __exit eu_exit (void)
  • error management is different
  • module exit code structure is standard

other functions to be compared ?!

---