This days I came to a problem like this:

Environment:linux kernel 2.6.20, PPC 405EP bubinga demo board.

We have tested two PCI2IDE cards in this board,A)silicon Image's (cpu:cmd649);B)IT8212F's.The A card with a harddisk,which is well recognised by the demo board;but the same harddisk in B card, can't be recognised by the board,only recognised the PCI card.We trace the kernel,found something different in this funtion in the file linux-2.6.20/drivers/ide/ide-iops.c:

++++++++++++++++++++++++++++++++++++++++++

int ide_wait_not_busy(ide_hwif_t *hwif, unsigned long timeout)
{
u8 stat = 0;
u8 i = 0;
while(timeout--) {
  /*
   * Turn this into a schedule() sleep once I'm sure
   * about locking issues (2.5 work ?).
   * hwif->io_ports[IDE_STATUS_OFFSET] = 0xFFEF
   */
  mdelay(1);

  stat = hwif->INB(hwif->io_ports[IDE_STATUS_OFFSET]);
  printk(KERN_INFO "hwif->io_ports[IDE_STATUS_OFFSET] ==> %4x, 222222STAT ==> %4x\n",hwif->io_ports[IDE_STATUS_OFFSET],stat);
  if ((stat & BUSY_STAT) == 0){
   printk(KERN_INFO "BUSYING................\n");
   return 0;
  }
  /*
   * Assume a value of 0xff means nothing is connected to
   * the interface and it doesn't implement the pull-down
   * resistor on D7.n        isa_io_base = e7fee000
   */
  if (stat == 0xff){
   printk(KERN_INFO "NODEV................ \n");
   return -ENODEV;
  }
  touch_softlockup_watchdog();
  touch_nmi_watchdog();
}
return -EBUSY;
}

+++++++++++++++++++++++++++++++++++++++++++++++++++

stat = hwif->INB(hwif->io_ports[IDE_STATUS_OFFSET]);
printk(KERN_INFO "hwif->io_ports[IDE_STATUS_OFFSET] ==> %4x, 222222STAT ==> %4x\n",hwif->io_ports[IDE_STATUS_OFFSET],stat);

By printing this two parameters, we found that the former parameter is the same in two cards,but the value of "stat" varies between the two

cards.A card,stat equals to 0x50;B card,stat equals to 0xFF,and it keeps the same,won't change.

I'm puzzled about this.Is it about the driver of IT821x.c in the kernel?or something else?

Any advise is welcomed.Thank you.