By not always starting the polling loop from the same modulation, we avoid entering infinite loops where devices exporting 2 targets (on 2 different modulations) get the same target activated over and over. If this target is not readable (e.g. a wallet emulating a tag), we will stay in an error loop for ever. Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
It seems that some pn533 firmwares go belly up when being asked to send poll frames too frequently. Adding a 10ms delay between each of them calm the chip down and prevent it from crashing. Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
The extended information frame are sent by PN533 to exchange frames larger than 255 bytes. These extended frame are very close from the standard ones except for the header size length. On each incoming frame, we set the correct header length, and we do that only for the standard pn533 chipsets as the acr122 does not seem to support extended frames properly. Signed-off-by: Olivier Guiter <olivier.guiter@linux.intel.com> Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
On sending large frames (size > 262), we split it in multiple chunks and send them asynchronously with MI bit. Signed-off-by: Olivier Guiter <olivier.guiter@linux.intel.com> Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
Extended Information frames are slightly different from standard frames as they can (theorically) handle datas up tu 64kB. PN533 firmware only supports packet data up to 265 (incl. TFI byte) This kind of frame are used when the pn533 wants to exchange more than 255 bytes, and this patch handles the reception of such frames. Signed-off-by: Olivier Guiter <olivier.guiter@linux.intel.com> Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
The AUTO RFCA bit forbids the pn533 chipset to turn its radio on whenever an external field is present. Without this bit set, some devices seems to get over flood by the pn533 rf field and thus become hardly detectable. Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
p2p devices must be able to support 424 kbps, so we should always select that bitrate in initiator mode. Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
Some devices are getting confused when not being asked for their system code with type F. Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
By turning the radio off after each failed polling try, we dramatically improve the pn533 polling loop efficiency. Without this fix, all Android phones running the broadcom NFC stack are almost never detected. Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
By using the standard setting for the regular pn533 dongles, we no longer wait for ever for an ATR_RES. Without this, a failing ATR_REQ will put the hardware into a busy loop, constantly waiting for an ATR_RES. Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>