Hallo all,
I am having trouble to securely switch between DHCP and static IP
assignment. The way I am doing this at the moment is sometimes leading to an
assert (please read further).
So far I am quite satisfied with the lwip-stack and its performance, but
this problem gives me a headache.
Description:
My environment needs to be able to switch between static IP and DHCP during
runtime. This should be possible at any point in time (although once
configured it will most likely stay the same.).
Info: I am using a Realtime OS.
So after initialization either with DHCP or static IP address assignment I
use two functions to switch between those 2 modes.
Both functions are called out of a user menu. The code looks like the
following:
int iEthCtrl_UseStaticIP(unsigned long ulIPAddress, unsigned long
ulIPGateway, unsigned long ulIPSubnetMask)
{
struct ip_addr xIpAddr, xNetMask, xGateway;
// Check, if the network connection is up and DHCP is activated
if ((sNetif.flags & NETIF_FLAG_UP) && (sNetif.flags &
NETIF_FLAG_DHCP)) {
// Release the DHCP lease (it already calls netif_set_down()
function)
dhcp_release(&sNetif);
// Stop the dhcp service
dhcp_stop(&sNetif);
xIpAddr.addr = ulIPAddress;
xGateway.addr = ulIPGateway;
xNetMask.addr = ulIPSubnetMask;
// Set the new ip address
netif_set_addr(&sNetif, &xIpAddr, &xNetMask, &xGateway);
// Bring the interface up again
netif_set_up(&sNetif);
}
...
}
int iEthCtrl_UseDHCP(void)
{
struct ip_addr xIpAddr, xNetMask, xGateway;
IP4_ADDR(&xIpAddr, 0, 0, 0, 0);
IP4_ADDR(&xNetMask, 0, 0, 0, 0);
IP4_ADDR(&xGateway, 0, 0, 0, 0);
// Check, if the network connection is up and DHCP is NOT activated
if ((sNetif.flags & NETIF_FLAG_UP) && !(sNetif.flags &
NETIF_FLAG_DHCP)) {
// Bring the current interface down
netif_set_down(&sNetif);
// Set the IP addresses to "0.0.0.0" so the DHCP server can send
a reply.
// Otherwise the previous static IP might be out of the network
class of the DHCP server
netif_set_addr(&sNetif, &xIpAddr, &xNetMask, &xGateway);
// Start DHCP negotiation
if (sNetif.dhcp == NULL) {
dhcp_start(&sNetif);
}
else {
dhcp_network_changed(&sNetif);
}
...
}
...
}
Furthermore I am using a callback function to read the (cable) link status
from the PHY every 500ms:
static void vPrvEthCtrl_CheckLinkStatusCallback(void* pvParams)
{
volatile avr32_macb_t *macb = &AVR32_MACB;
unsigned long ulStatusReg;
ulStatusReg = ulReadMDIO(macb, PHY_BMSR); // Can be blocking!
if (ulStatusReg & BMSR_LSTATUS) {
// If the cable was disconnected previously we set the link up
again
if (bLinkStatus == FALSE) {
netif_set_link_up(&sNetif);
}
bLinkStatus = TRUE;
}
else {
if (bLinkStatus == TRUE) {
netif_set_link_down(&sNetif);
netif_set_down(&sNetif);
}
bLinkStatus = FALSE;
}
}
So the problem occurs after calling the function iEthCtrl_UseDHCP().
After doing so it sometimes happens to stop executing on an assert in
dhcp_create_request().
Assert:
LWIP_ASSERT("dhcp_create_request: dhcp->p_out == NULL", dhcp->p_out ==
NULL);
Callstack:
dhcp_create_request
dhcp_select
dhcp_handle_offer
dhcp_recv
udp_input
ip_input
ethernet_input
tcpip_thread
So my questions are the following.
Why is a DHCP-request forwarded to the DHCP-handler, although dhcp_stop()
has been called before?
Is the way I switch from static IP to DHCP and vice versa correct?
Any comment or advice is appreciated.
Thanks in advance!
Best regards,
Benjamin Schelte
_______________________________________________
lwip-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/lwip-users
