More information regarding this issue:
If I change the dts from:
pci2: pcie@ffe08000 {
cell-index = <3>;
compatible = "fsl,mpc8548-pcie";
device_type = "pci";
#interrupt-cells = <1>;
#size-cells = <2>;
#address-cells = <3>;
reg = <0 0xffe08000 0 0x1000>;
bus-range = <0 255>;
ranges = <0x2000000 0x0 0xc0000000 0 0xc0000000 0x0 0x20000000
0x1000000 0x0 0x00000000 0 0xffc40000 0x0 0x10000>;
clock-frequency = <33333333>;
interrupt-parent = <&mpic>;
interrupts = <27 2>;
interrupt-map-mask = <0xf800 0x0 0x0 0x7>;
interrupt-map = <
/* IDSEL 0x0 */
0000 0x0 0x0 0x1 &mpic 0x8 0x1
0000 0x0 0x0 0x2 &mpic 0x9 0x1
0000 0x0 0x0 0x3 &mpic 0xa 0x1
0000 0x0 0x0 0x4 &mpic 0xb 0x1
>;
pcie@0 {
reg = <0x0 0x0 0x0 0x0 0x0>;
#size-cells = <2>;
#address-cells = <3>;
device_type = "pci";
ranges = <0x2000000 0x0 0xc0000000
0x2000000 0x0 0xc0000000
0x0 0x20000000
0x1000000 0x0 0x0
0x1000000 0x0 0x0
0x0 0x100000>;
};
};
to:
pci2: pcie@ffe08000 {
cell-index = <3>;
compatible = "fsl,mpc8548-pcie";
device_type = "pci";
#interrupt-cells = <1>;
#size-cells = <2>;
#address-cells = <3>;
reg = <0 0xffe08000 0 0x1000>;
bus-range = <0 255>;
ranges = <0x2000000 0x0 0xc0000000 0 0xc0000000 0x0 0x20000000
0x1000000 0x0 0x00000000 0 0xffc40000 0x0 0x10000>;
clock-frequency = <33333333>;
interrupt-parent = <&mpic>;
interrupts = <27 2>;
interrupt-map-mask = <0xf800 0x0 0x0 0x7>;
interrupt-map = <
/* IDSEL 0x0 */
0000 0x0 0x0 0x1 &mpic 0x8 0x2
0000 0x0 0x0 0x2 &mpic 0x9 0x2
0000 0x0 0x0 0x3 &mpic 0xa 0x2
0000 0x0 0x0 0x4 &mpic 0xb 0x2
>;
pcie@0 {
reg = <0x0 0x0 0x0 0x0 0x0>;
#size-cells = <2>;
#address-cells = <3>;
device_type = "pci";
ranges = <0x2000000 0x0 0xc0000000
0x2000000 0x0 0xc0000000
0x0 0x20000000
0x1000000 0x0 0x0
0x1000000 0x0 0x0
0x0 0x100000>;
};
};
I find that the cards that failed begin to work and the one that
worked begins to fail.
Miguel Ángel
On Thu, Aug 2, 2012 at 11:28 AM, Miguel Ángel Álvarez
<caronteycerbero@xxxxxxxxx> wrote:
> Dear all,
>
> I have a 1 to 4 PCIe switch connected to a root port of a P2020.
>
> I am attaching a PCIe board to each one of the download ports of the
> switch (same kind of card with four serial ports in each one of them).
>
> When booting Linux, all cards are detected, so I can see them in
> /sys/bus/pci/devices. All 16 ttyS* are created. However, when trying
> to use them, I have no problems when using the serial ports of the
> card in the last downstream port of the switch, but if I try to use
> any of the others I obtain the following error:
>
> irq 16: nobody cared (try booting with the "irqpoll" option)
> Call Trace:
> [df83fca0] [c0007540] show_stack+0x84/0x1b0 (unreliable)
> [df83fcd8] [c0077324] __report_bad_irq+0x5c/0xe0
> [df83fcf0] [c0077534] note_interrupt+0x18c/0x238
> [df83fd10] [c0078520] handle_fasteoi_irq+0xe4/0x134
> [df83fd28] [c0004ea0] handle_one_irq+0x30/0x44
> [df83fd38] [c000c078] __ipipe_do_IRQ+0x84/0xc0
> [df83fd58] [c007bdc8] __ipipe_sync_stage+0x210/0x3b8
> [df83fdb0] [c007cf40] __ipipe_unstall_root+0x60/0x74
> [df83fdb8] [c0047798] __do_softirq+0xd8/0x200
> [df83fe08] [c0004d50] do_softirq+0x60/0x80
> [df83fe18] [c004751c] irq_exit+0x54/0xac
> [df83fe20] [c000c07c] __ipipe_do_IRQ+0x88/0xc0
> [df83fe40] [c007bdc8] __ipipe_sync_stage+0x210/0x3b8
> [df83fe98] [c000ba7c] __ipipe_handle_irq+0x1c4/0x1f8
> [df83fec8] [c000bd80] __ipipe_grab_irq+0x15c/0x1f0
> [df83fee8] [c001323c] __ipipe_ret_from_except+0x0/0xc
> [df83ffa8] [c0008684] cpu_idle+0x68/0xe0
> [df83ffc0] [c0544c2c] start_secondary+0x314/0x350
> [df83fff0] [c0001c9c] __secondary_start+0x30/0x84
> handlers:
> [<c0336e80>] (serial8250_interrupt+0x0/0x148)
> Disabling IRQ #16
>
> What I can see from /sys/bus/pci/devices/*/irq is that one irq is
> assigned to each card; 16 for the first, 17 for the second, 18 for the
> third and 19 for the last (and successfull) one.
>
> My question is... where are irqs mapped for the pci express cards? I
> have backtrace the code to pci_device_probe in pci-driver.c, but there
> it seems that the irqs are already assigned.
>
> I have searched the web, but have not found the clue that makes me
> understand the process.
>
> Any ideas?
>
> Thanks
>
> Miguel Ángel
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
