macraw infinite loop ws5500

[rFronteddu]

I was testing some code with a pico2 and the ws5500, I noticed that in MACRAW mode, if I send a lot of packets the board can get stuck in socket.c. It seems that tmp is always 0. I was able to avoid the problem by calling return when tmp is 0. I also added a print to see how often that would happen and it is not very often. I know that the ws is not bricked because I am able to query other flags from SPI such as getSn_TX_FSR(sn) and the WS also replies to ping.

Are you aware of anything that I could be doing wrong?

while(1)
{
	tmp = getSn_IR(sn);
        if(tmp & Sn_IR_SENDOK)
        {
            setSn_IR(sn, Sn_IR_SENDOK);
            printf("z9a %d\n", len);
            break;
        }
        //M:20131104
        //else if(tmp & Sn_IR_TIMEOUT) return SOCKERR_TIMEOUT;
        else if(tmp & Sn_IR_TIMEOUT)
        {
            setSn_IR(sn, Sn_IR_TIMEOUT);
            //M20150409 : Fixed the lost of sign bits by type casting.
            //len = (uint16_t)SOCKERR_TIMEOUT;
            //break;
            printf("z9b %d\n", len);
            #if _WIZCHIP_ < 5500    //M20150401 : for WIZCHIP Errata #4, #5 (ARP errata)
               if(taddr) setSUBR((uint8_t*)&taddr);
            #endif
            return SOCKERR_TIMEOUT;
        }
        print("Error"); // I added this
        return (int32_t)len;
}
    ...

[eadam-withings]

Not sure it's related, but on our side we had an infinite loop with W5500 + pico2 while waiting for DHCP, and the issue was coming from the commit 5b4b6bc20548d9292e78e48bcc7871ac2a38a55e in socket.c.

Changing back the line 202:
switch(protocol & 0x0F)
To
switch(protocol)
Fixed the issue

[rFronteddu]

I don' think it's related but I will update my fork with your fix too!