Francois,
Thanks for responding.
On 05/07/2016 04:35 AM, Francois Romieu wrote:
> Murali Karicheri <m-kariche...@ti.com> :
> [...]
>> I am trying to integrate the rtl8168 PCIe card to have Ethernet functional
>> on my Keystone EVM.
>
> Which (EVM) one ?
K2G EVM. This is a new EVM for which work has been started to add it to upstream
kernel. But it not there yet. The PCIe h/w in the K2G SoC is a re-use from
existing SoC such as K2E. The PCIe controller (Root complex a.k.a RC) driver
based on this PCIe h/w works fine on K2E EVM with a Marvel SATA controller
connected to the PCIe port. On K2G EVM, I have a PCIe slot to which I have
plugged the rtl8168 PCIe card and power on with the log I had sent you. The
serdes driver is responsible for link setup seems to work fine, link is up and
could detect the rtl8168 below.
>
>> I purchased the rtl8111c Gib card from Amazon. The Card is detected
>> by the RC and I can see it is enumerated and show up when doing lspci
>> command.
>
> What does "the RC" mean ? A different PC ?
RC - Root Complex (PCIe controller - drivers/pci/host/pci-keystone.c, a
designware
based PCIe controller driver)
The the PCIe controller driver bindings...
pcie0_phy: phy@2320000 {
#phy-cells = <0>;
compatible = "ti,keystone-serdes-pcie";
reg = <0x02320000 0x4000>;
link-rate-kbps = <5000000>;
num-lanes = <1>;
status = "disabled";
};
pcie0: pcie@21800000 {
compatible = "ti,keystone-pcie", "snps,dw-pcie";
power-domains = <&k2g_pds K2G_DEV_PCIE0>;
clocks = <&k2g_clks K2G_DEV_PCIE0
K2G_DEV_PCIE_VBUS_CLK>;
clock-names = "pcie";
#address-cells = <3>;
#size-cells = <2>;
reg = <0x21801000 0x2000>, <0x21800000 0x1000>,
<0x02620128 4>;
ranges = <0x81000000 0 0 0x23250000 0 0x4000
0x82000000 0 0x50000000 0x50000000 0
0x10000000>;
status = "disabled";
device_type = "pci";
num-lanes = <1>;
phys = <&pcie0_phy>;
#interrupt-cells = <1>;
interrupt-map-mask = <0 0 0 7>;
interrupt-map = <0 0 0 1 &pcie_intc0 0>, /* INT A */
<0 0 0 2 &pcie_intc0 1>, /* INT B */
<0 0 0 3 &pcie_intc0 2>, /* INT C */
<0 0 0 4 &pcie_intc0 3>; /* INT D */
pcie_msi_intc0: msi-interrupt-controller {
interrupt-controller;
#interrupt-cells = <1>;
interrupt-parent = <&gic>;
interrupts = <GIC_SPI 30 IRQ_TYPE_EDGE_RISING>,
<GIC_SPI 31 IRQ_TYPE_EDGE_RISING>,
<GIC_SPI 32 IRQ_TYPE_EDGE_RISING>,
<GIC_SPI 33 IRQ_TYPE_EDGE_RISING>,
<GIC_SPI 34 IRQ_TYPE_EDGE_RISING>,
<GIC_SPI 35 IRQ_TYPE_EDGE_RISING>,
<GIC_SPI 36 IRQ_TYPE_EDGE_RISING>,
<GIC_SPI 37 IRQ_TYPE_EDGE_RISING>;
};
pcie_intc0: legacy-interrupt-controller {
interrupt-controller;
#interrupt-cells = <1>;
interrupt-parent = <&gic>;
interrupts = <GIC_SPI 26 IRQ_TYPE_EDGE_RISING>,
<GIC_SPI 27 IRQ_TYPE_EDGE_RISING>,
<GIC_SPI 28 IRQ_TYPE_EDGE_RISING>,
<GIC_SPI 29 IRQ_TYPE_EDGE_RISING>;
};
};
>
>> However I can't get the Ethernet port functional.
>
>> Does this need MSI interrupt ?
>
> No.
Does it also work with MSI?? Is there anything to be set in the driver
to request MSI interrupt?
>
>> I can't see it has requested any.
>
> Yes, something went really, really wrong. See below.
>
> [...]
>> [ 2.303965] PCI host bridge /soc/pcie@21800000 ranges:
>> [ 2.309108] No bus range found for /soc/pcie@21800000, using [bus 00-ff]
>> [ 2.316269] IO 0x23250000..0x23253fff -> 0x00000000
>> [ 2.321499] MEM 0x50000000..0x5fffffff -> 0x50000000
>> [ 2.331666] keystone-pcie 21801000.pcie: PCI host bridge to bus 0000:00
>> [ 2.338283] pci_bus 0000:00: root bus resource [bus 00-ff]
>> [ 2.343937] pci_bus 0000:00: root bus resource [io 0x0000-0x3fff]
>> [ 2.350114] pci_bus 0000:00: root bus resource [mem 0x50000000-0x5fffffff]
>> [ 2.357095] pci 0000:00:00.0: [104c:b00b] type 01 class 0x060400
>> [ 2.357665] PCI: bus0: Fast back to back transfers disabled
>> [ 2.363717] pci 0000:01:00.0: [10ec:8168] type 00 class 0x020000
>> [ 2.363809] pci 0000:01:00.0: reg 0x10: [io 0x0000-0x00ff]
>> [ 2.363867] pci 0000:01:00.0: reg 0x18: [mem 0x00000000-0x00000fff 64bit]
>> [ 2.363909] pci 0000:01:00.0: reg 0x20: [mem 0x00000000-0x0000ffff 64bit
>> pref]
>> [ 2.363939] pci 0000:01:00.0: reg 0x30: [mem 0x00000000-0x0001ffff pref]
>> [ 2.364099] pci 0000:01:00.0: supports D1 D2
>> [ 2.364116] pci 0000:01:00.0: PME# supported from D0 D1 D2 D3hot D3cold
>> [ 2.381251] PCI: bus1: Fast back to back transfers disabled
>> [ 2.386989] pci 0000:00:00.0: BAR 8: assigned [mem 0x50000000-0x500fffff]
>> [ 2.393937] pci 0000:00:00.0: BAR 9: assigned [mem 0x50100000-0x501fffff
>> pref]
>> [ 2.401221] pci 0000:00:00.0: BAR 7: assigned [io 0x1000-0x1fff]
>> [ 2.407320] pci 0000:01:00.0: BAR 6: assigned [mem 0x50100000-0x5011ffff
>> pref]
>> [ 2.414597] pci 0000:01:00.0: BAR 4: assigned [mem 0x50120000-0x5012ffff
>> 64bit pref]
>> [ 2.422380] pci 0000:01:00.0: BAR 2: assigned [mem 0x50000000-0x50000fff
>> 64bit]
>> [ 2.429702] pci 0000:01:00.0: BAR 0: assigned [io 0x1000-0x10ff]
>> [ 2.435821] pci 0000:00:00.0: PCI bridge to [bus 01]
>> [ 2.440783] pci 0000:00:00.0: bridge window [io 0x1000-0x1fff]
>> [ 2.446896] pci 0000:00:00.0: bridge window [mem 0x50000000-0x500fffff]
>> [ 2.453699] pci 0000:00:00.0: bridge window [mem 0x50100000-0x501fffff
>> pref]
>> [ 2.461453] pcieport 0000:00:00.0: Signaling PME through PCIe PME
>> interrupt
>> [ 2.468411] pci 0000:01:00.0: Signaling PME through PCIe PME interrupt
>> [ 2.475075] pcie_pme 0000:00:00.0:pcie01: service driver pcie_pme loaded
>> [ 2.475392] aer 0000:00:00.0:pcie02: service driver aer loaded
>> [ 2.475652] r8169 Gigabit Ethernet driver 2.3LK-NAPI loaded
>> [ 2.481419] r8169 0000:01:00.0: enabling device (0140 -> 0143)
>> [ 2.488865] r8169 0000:01:00.0 eth0: RTL8169 at 0xf0d6a000,
>> 00:00:00:00:00:00, XID 00000000 IRQ 286
>
> No need to go further, there is a serious problem here.
>
> Most of the XID bits are read from a mapped register. They're definitely not
> expected to be null as the driver requires it to identify a proper chipset
> (the common PCI configuration registers only tell that you aren't dealing
> with a coffee machine).
Looks like during enumeration, the PCIe controller has correctly read the
Device ID a and Vendor ID over the PCIe bus. So the bus access is happening
at least to read the common PCI configuration space. But then what you are
confirming is the mapped XID register is not accessed correctly and is reading
zeors. Right? I will try to debug this first before proceeding any further.
Thanks.
Murali
> Any read returning zeroes would not surprize me.
>
> [...]
>> Can someone help me figure out what is missing ?
>
> Hardly at this point. I can only suggest to switch power off, plug
> the 8168 device in a x86 (32 or 64 bits) system and check how it behaves.
>
I need to continue work with my EVM as I have to make it functional.
--
Murali Karicheri
Linux Kernel, Keystone
