MPC834x PCI problem

[Gary Thomas]

I have two [internal] boards with MPC8347.  Both have a PCI
bus, slightly different set of "wired" peripherals.

On one board, the PCI seems to be working fine.  I can talk
to all of my wired devices, plus one in a plugin slot.  The
[PCI portion] DTS for this board looks like this:
	pci0: pci at ff008500 {
		cell-index = <1>;
		interrupt-map-mask = <0xf800 0x0 0x0 0x7>;
		interrupt-map = <
				/* IDSEL 0x0A (External slot) */
				 0x5000 0x0 0x0 0x1 &fpga_ic 0
				 0x5000 0x0 0x0 0x2 &fpga_ic 1
				 0x5000 0x0 0x0 0x3 &fpga_ic 2
				 0x5000 0x0 0x0 0x4 &fpga_ic 3

				/* IDSEL 0x0B (Promise SATA) */
				 0x5800 0x0 0x0 0x1 &fpga_ic 5
				 0x5800 0x0 0x0 0x2 &fpga_ic 5
				 0x5800 0x0 0x0 0x3 &fpga_ic 5
				 0x5800 0x0 0x0 0x4 &fpga_ic 5

				/* IDSEL 0x0C (Fujitsu Coral-P) */
				 0x6000 0x0 0x0 0x1 &fpga_ic 4
				 0x6000 0x0 0x0 0x2 &fpga_ic 4
				 0x6000 0x0 0x0 0x3 &fpga_ic 4
				 0x6000 0x0 0x0 0x4 &fpga_ic 4

				/* IDSEL 0x0D (Philips USB) */
				 0x6800 0x0 0x0 0x1 &fpga_ic 12
				 0x6800 0x0 0x0 0x2 &fpga_ic 12
				 0x6800 0x0 0x0 0x3 &fpga_ic 12
				 0x6800 0x0 0x0 0x4 &fpga_ic 12

				/* IDSEL 0x1F (External slot) */
				 0xF800 0x0 0x0 0x1 &fpga_ic 0
				 0xF800 0x0 0x0 0x2 &fpga_ic 1
				 0xF800 0x0 0x0 0x3 &fpga_ic 2
				 0xF800 0x0 0x0 0x4 &fpga_ic 3
                                >;
		interrupt-parent = <&ipic>;
		interrupts = <0x13 0x8
                              0x14 0x8>;
		bus-range = <0 0>;
		ranges = <0x02000000 0x0 0xC0000000 0xC0000000 0x0 0x10000000
			  0x01000000 0x0 0x00000000 0xB8000000 0x0 0x00100000>;
		clock-frequency = <33333333>;
		#interrupt-cells = <1>;
		#size-cells = <2>;
		#address-cells = <3>;
		reg = <0xff008500 0x100		/* Internal registers */
		       0xff008300 0x8>;         /* Config Space registers */
		compatible = "fsl,mpc8349-pci";
		device_type = "pci";
	};
This board has a Promise SATA controller in slot 11 (drivers/ata/sata_promise.c)

The second board uses a DTS derived from the first.  In fact, the
*only* difference is in the PCI layout:
	pci0: pci at ff008500 {
		cell-index = <1>;
		interrupt-map-mask = <0xf800 0x0 0x0 0x7>;
		interrupt-map = <
				/* IDSEL 0x0B (Promise SATA) */
				 0x5800 0x0 0x0 0x1 &ipic 0x13 8
				 0x5800 0x0 0x0 0x2 &ipic 0x13 8
				 0x5800 0x0 0x0 0x3 &ipic 0x13 8
				 0x5800 0x0 0x0 0x4 &ipic 0x13 8
                                >;
		interrupt-parent = <&ipic>;
		interrupts = <0x13 0x8>;
		bus-range = <0 0>;
		ranges = <0x02000000 0x0 0xC0000000 0xC0000000 0x0 0x10000000
			  0x01000000 0x0 0x00000000 0xB8000000 0x0 0x00100000>;
		clock-frequency = <33333333>;
		#interrupt-cells = <1>;
		#size-cells = <2>;
		#address-cells = <3>;
		reg = <0xff008500 0x100		/* Internal registers */
		       0xff008300 0x8>;         /* Config Space registers */
		compatible = "fsl,mpc8349-pci";
		device_type = "pci";
	};
This board has a slightly different Promise controller (drivers/ata/pata_pdc2027x.c)
The basic PCI access code in both drivers is the same.

As mentioned, the first board works fine.  The second board
falls apart on PCI access.  It's obvious that the PDC2027x
driver is having access problems.  Here are some clues from
the boot log (PCI related items only):

  Found FSL PCI host bridge at 0x00000000ff008500. Firmware bus number: 0->0
  PCI host bridge /pci at ff008500 (primary) ranges:
   MEM 0x00000000c0000000..0x00000000cfffffff -> 0x00000000c0000000
    IO 0x00000000b8000000..0x00000000b80fffff -> 0x0000000000000000
  PCI: Probing PCI hardware
  PCI: Cannot allocate resource region 5 of device 0000:00:0b.0, will remap
  bus: 00 index 0 io port: [0x00-0xfffff]
  bus: 00 index 1 mmio: [0xc0000000-0xcfffffff]
  pdc_detect_pll_input_clock: scr[FFFFFFFF]
  pdc_read_counter: bccrh [7FFF] bccrl [7FFF]
  pdc_read_counter: bccrhv[7FFF] bccrlv[7FFF]
  pdc_read_counter: bccrh [7FFF] bccrl [7FFF]
  pdc_read_counter: bccrhv[7FFF] bccrlv[7FFF]
  pdc_detect_pll_input_clock: scr[FFFFFFFF]
  pdc_detect_pll_input_clock: start[1073741823] end[1073741823]
  pdc_detect_pll_input_clock: PLL input clock[0]Hz

It's pretty obvious that the PDC driver is having troubles.
The system seems happy otherwise:

  root at ppc_target:~ lspci -v
  00:00.0 Bridge: Unknown device 1957:0085 (rev 11)
        Flags: bus master, 66MHz, fast devsel, latency 248
        Memory at <unassigned> (64-bit, prefetchable)
        Memory at <unassigned> (64-bit, non-prefetchable)
        Capabilities: [48] #06 [0000]

  00:0b.0 Mass storage controller: Promise Technology, Inc. 20275 (rev 01) (prog-if 85)
        Subsystem: Promise Technology, Inc. 20275
        Flags: bus master, 66MHz, slow devsel, latency 0, IRQ 19
        I/O ports at 1000 [size=8]
        I/O ports at 1008 [size=4]
        I/O ports at 1010 [size=8]
        I/O ports at 1018 [size=4]
        I/O ports at 1020 [size=16]
        [virtual] Memory at c0000000 (32-bit, non-prefetchable) [size=16K]
        Capabilities: [60] Power Management version 1


Any ideas what might be going [wrong] here?  I did notice that
both RedBoot [my boot environment] and an older kernel (2.6.20)
on this board assigned a non-zero offset in the PCI space for
this device:
  Bus: 0, PCI Device: 11, PCI Func: 0
    Vendor Id: 0x105A, Device Id: 0x1275, Command: 0x0007, IRQ: 1
    BAR[0]    0x00001001 / probed size 0x00000000 / CPU addr 0xb8001000
    BAR[1]    0x00001009 / probed size 0x00000000 / CPU addr 0xb8001008
    BAR[2]    0x00001011 / probed size 0x00000000 / CPU addr 0xb8001010
    BAR[3]    0x00001019 / probed size 0x00000000 / CPU addr 0xb8001018
    BAR[4]    0x00001021 / probed size 0x00000000 / CPU addr 0xb8001020
    BAR[5]    0x00100000 / probed size 0x00000000 / CPU addr 0xc0100000

Could this be a possible problem (maybe the device doesn't like it)?

Thanks

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------