Tool/software:
Hi TI team,
I'm testing the Enet lwIP App on a on J784S4 using the Enet LLD and LWIP stack. The board runs SDK version: ti-processor-sdk-09_02_00. I enabled the interactive lwIP shell that listens on port 23 and accepts commands like:
lwipcfg.h
#define LWIP_SHELL_APP 1
$ telnet 192.168.10.80 23
> open 192.168.10.82 80
I'm trying to connect to an HTTP server running on my host PC (192.168.10.82:80), confirmed reachable from the board (ping successful, server running with python3 -m http.server 80)
==========================
Enet lwIP App - v 0.2
==========================
EnetBoard_setupPorts: 1 of 1 ports configurations found
CPU Load: 51%
Starting lwIP, local interface IP is dhcp-enabled
No valid MAC address found in EEPROM, falling back to static address
EnetMcm: CPSW_2G on MCU NAVSS
PHY 0 is alive
EnetPhy_bindDriver: PHY 0: OUI:00a0ef Model:03 Ver:00 <-> 'generic' : OK
Host MAC address: 70:ff:76:1d:92:c1
[LWIPIF_LWIP] Enet LLD netif initialized successfully
status_callback==UP, local interface IP is 0.0.0.0
Cpsw_handleLinkUp: Port 1: Link up: 1-Gbps Full-Duplex
MAC Port 1: link up
link_callback==UP
CPU Load: 1%
status_callback==UP, local interface IP is 192.168.10.80
Enet lwIP App: Added Network IP address I/F ti0: 192.168.10.80
Initializing apps
CPU Load: 1%
CPU Load: 1%
CPU Load: 1%
CPU Load: 1%
CPU Load: 1%
CPU Load: 13%
CPU Load: 100%
CPU Load: 100%
It seems the shell application hangs after netconn_connect() call.
shell.c
static s8_t
com_open(struct command *com)
{
... sendstr("Opening connection to ", com->conn);
netconn_write(com->conn, com->args[0], strlen(com->args[0]), NETCONN_COPY);
sendstr(":", com->conn);
netconn_write(com->conn, com->args[1], strlen(com->args[1]), NETCONN_COPY);
sendstr(NEWLINE, com->conn);
conns[i] = netconn_new(NETCONN_TCP);
if (conns[i] == NULL) {
sendstr("Could not create connection identifier (out of memory)."NEWLINE, com->conn);
return ESUCCESS;
}
err = netconn_connect(conns[i], &ipaddr, port); << HANGS HERE>>
Observations:
- The board receives the SYN-ACK from the server (confirmed via tcpdump).
- CPU load spikes to 100% immediately after open is issued.
- No debug logs are visible even after setting LWIP_DEBUG and enabling relevant debug macros.
Questions:
- Is this a known issue with netconn_connect() or netconn_write() in blocking shell applications?
- How can I get full LWIP debug logs from the SDK environment?
