Hi Zuki,
After several debugging test and as you told me yesterday, I came
to the conclusion that the ixp1200eb hangs within the Control_SystemStart().
Anyway it's still hard to tell where the bug is
hiding since, even with the help of led(), we have experienced different
behaviors according to different kernel used.
It seems that troubles come within this chunk of code:
Control_SystemStart()
{
....
// Initialize and Start interrupts
Int_Init(printPortData, NULL, ueInterruptHandler);
led(0xC1);
Int_addSlowHandler(timerCallbackFn);
led(0xC2);
control_init_ui(); // Initialize the user
interface
led(0xC3);
Int_addFastHandler(UI_checkUART);
led(0x25);
// Set the VRP to SA forwarding rate, used only for a test
scratchWrite32(VRP_TO_SA_RATE_ADDR, 0);
led(0xC4);
Fmt_print("--- Control_SystemStart done\n\n");
led(0xC5);
...
In some cases, once the sa hanged, the led shows the 0xC4 step, in
others, with different kernels, it hangs at 0xC2 step. In both cases
it seems that the board is not able to complete the Fmt_print() call, and
the last printed message is always truncated at 10th char sent to the
serial port.
Have you got any idea of how to get ride of this matter?
Any suggestions are appreciated.
Thanks,
Bonelli Nicola
On Fri, 28 Feb 2003, Bonelli Nicola wrote:
After several debugging test and as you told me yesterday, I came to the conclusion that the ixp1200eb hangs within the Control_SystemStart().
In both cases it seems that the board is not able to complete the Fmt_print() call, and the last printed message is always truncated at 10th char sent to the serial port.
Have you got any idea of how to get ride of this matter?
You might want to see if the UI_checkUART function is called after that tenth invocation. (You can put a putc or similar function in the checkUART function.) This seems strange to me; I'll give it some more thought and try to get back to you soon. Zuki -- Yitzchak Gottlieb zuki@CS.Princeton.EDU
What version of the IXP1200 chip are you using? The last nibble of the ID_CHIP register (shown at boot time) will tell you this: 0=A Stepping, 1=B Stepping, 3=C Stepping. The Vera firmware will not work for A Stepping parts. Scott
Intel provide peth to communication from core and uengine, but have some proble 1. the communication speed is low, 2. there is no priority in the communication queue 3. read lock will hang whole system who have some better method to communication from core to uengine, or where I can find ? best regards huang tao
participants (4)
-
Bonelli Nicola -
huangtao -
Scott Karlin -
Yitzchak M. Gottlieb