LKML Archive on lore.kernel.org
 help / color / Atom feed
* Re: Dts for eth network based on marvell's mv88e6390x crashes Xilinx's linux-kernel v5.4
       [not found] <CANBsoPmgct2UTq=Cuf1rXJRitiF1mWhWwdtH2=73yyZiJbT0rg@mail.gmail.com>
@ 2020-10-16  3:31 ` Andrew Lunn
       [not found]   ` <CANBsoPkCEZadmBaeZ=8EAOP6Ctw5deLen7yKQk__1-ZVoJE6yA@mail.gmail.com>
  0 siblings, 1 reply; 3+ messages in thread
From: Andrew Lunn @ 2020-10-16  3:31 UTC (permalink / raw)
  To: michael alayev; +Cc: vivien.didelot, netdev, linux-kernel

On Sun, Oct 11, 2020 at 03:31:52PM +0300, michael alayev wrote:
> Hello,
> 
> This is my system's network topology. It relies on Xilinx's Zynq-7000 soc.
> 
> zynq-7000      switch1       switch2
>  (cpu)       mv88e6390x    mv88e6390x
> -------      ---------     ---------
> |     |      |       |     |       |
> | gem0|------|p0  p10|-----|p10    |
> |     |      |       |     |       |
> |     |      ---------     ---------
> |     |         |mdio_addr=2   | mdio_addr=3
> |     |         |              |
> |     |-------------------------
> |     |    |
> |     |    |              mv88e1510(phy)
> |     |    |  mdio_addr=1 --------
> |     |    ---------------|      |
> |     |                   |      |
> | gem1|-------------------|      |
> |     |                   |      |
> |     |                   --------
> -------
> 
> I have built a v5.4 linux kernel for it and its ethernet network defined in the device-tree as follows
> 
> (arch/arm/boot/dts/zynq-zed.dts gem0 and gem1)   :
> 
> &gem0 {
> status = "okay";
> phy-mode = "rgmii-id";
> phy-handle = <&phy0>;
> 
> 
> mdio {
> #address-cells = <1>;
> #size-cells = <0>;
> 
> phy0: ethernet-phy@0 {
> compatible = "marvell";
> reg = <0>;
> device_type = "ethernet-phy";
> fixed-link {
> speed = <1000>;
> full-duplex;
> };
> };
> 
> debug_phy: ethernet-phy@1 {
> compatible = "marvell";
> reg = <1>;
> device_type = "ethernet-phy";
> label = "debug-phy";
> };

Hi Michael

This is pretty unreadable with all the white space removed. Please
could you post again with the white space.

> 
> Starting kernel ...
> ...
> [   40.953876] [<c05ab9cc>] (dsa_unregister_switch) from [<c042c980>] (mv88e6xxx_remove+0x1c/0x6c)
> [   40.962602] [<c042c980>] (mv88e6xxx_remove) from [<c042572c>] (mdio_remove+0x1c/0x38)
> [   40.970457] [<c042572c>] (mdio_remove) from [<c03cb4a4>] (device_release_driver_internal+0xf0/0x194)
> [   40.979612] [<c03cb4a4>] (device_release_driver_internal) from [<c03ca1e8>] (bus_remove_device+0xcc/0xdc)
> [   40.989201] [<c03ca1e8>] (bus_remove_device) from [<c03c7108>] (device_del+0x170/0x288)
> [   40.997227] [<c03c7108>] (device_del) from [<c0425674>] (mdio_device_remove+0xc/0x18)
> [   41.005073] [<c0425674>] (mdio_device_remove) from [<c0425024>] (mdiobus_unregister+0x50/0x84)
> [   41.013714] [<c0425024>] (mdiobus_unregister) from [<c043ac2c>] (macb_probe+0x88c/0xa68)
> [   41.021834] [<c043ac2c>] (macb_probe) from [<c03cc6d0>] (platform_drv_probe+0x48/0x98)
> [   41.029771] [<c03cc6d0>] (platform_drv_probe) from [<c03cad04>] (really_probe+0x140/0x2f8)
> [   41.038060] [<c03cad04>] (really_probe) from [<c03cb110>] (driver_probe_device+0x10c/0x154)
> [   41.046427] [<c03cb110>] (driver_probe_device) from [<c03cb2e4>] (device_driver_attach+0x44/0x5c)
> [   41.055316] [<c03cb2e4>] (device_driver_attach) from [<c03cb3a8>] (__driver_attach+0xac/0xb8)
> [   41.063856] [<c03cb3a8>] (__driver_attach) from [<c03c93b8>] (bus_for_each_dev+0x64/0xa0)
> [   41.072049] [<c03c93b8>] (bus_for_each_dev) from [<c03ca2d4>] (bus_add_driver+0xdc/0x1bc)
> [   41.080244] [<c03ca2d4>] (bus_add_driver) from [<c03cba60>] (driver_register+0xb0/0xf8)
> [   41.088266] [<c03cba60>] (driver_register) from [<c01026e0>] (do_one_initcall+0x74/0x164)
> [   41.096469] [<c01026e0>] (do_one_initcall) from [<c0800e38>] (kernel_init_freeable+0x108/0x1d8)
> [   41.105192] [<c0800e38>] (kernel_init_freeable) from [<c05e829c>] (kernel_init+0x8/0x110)
> [   41.113391] [<c05e829c>] (kernel_init) from [<c01010e8>] (ret_from_fork+0x14/0x2c)

This looks like the classic first time probe -EPROBE_DEFER. You expect
this path to happen, but not the invalid access. macb_probe will
register the mdio bus which causes the switches to probe. But because
macb_probe has not got so far as to call register_netdev, the gem
interface is not available, so DSA ends up returning
EPROBE_DEFER. What normally happens is that macb_probe continues, and
registers the network interface. The switches get probed again later
and it all works.

The real question is, why does the unroll go wrong when -EPROBE_DEFER
is returned?

   Andrew

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Dts for eth network based on marvell's mv88e6390x crashes Xilinx's linux-kernel v5.4
       [not found]   ` <CANBsoPkCEZadmBaeZ=8EAOP6Ctw5deLen7yKQk__1-ZVoJE6yA@mail.gmail.com>
@ 2020-10-18 15:58     ` Andrew Lunn
       [not found]       ` <CANBsoPm1Ln=59cGKbaA5OKdjA5dwEFA0pcg2tPUa5i2Db747Fw@mail.gmail.com>
  0 siblings, 1 reply; 3+ messages in thread
From: Andrew Lunn @ 2020-10-18 15:58 UTC (permalink / raw)
  To: michael alayev; +Cc: vivien.didelot, netdev, linux-kernel, adror

On Sun, Oct 18, 2020 at 10:18:53AM +0300, michael alayev wrote:
> Hello andrew,
> > This is pretty unreadable with all the white space removed.
> > Please could you post again with the white space.
> Its formatted better here:
> https://stackoverflow.com/q/64301750/8926995?sem=2

Better, but the indentation is still wrong.

Please fix your email client and post the DT file for review. I will
then point out some of the errors.

    Andrew

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Dts for eth network based on marvell's mv88e6390x crashes Xilinx's linux-kernel v5.4
       [not found]       ` <CANBsoPm1Ln=59cGKbaA5OKdjA5dwEFA0pcg2tPUa5i2Db747Fw@mail.gmail.com>
@ 2020-10-19 14:26         ` Andrew Lunn
  0 siblings, 0 replies; 3+ messages in thread
From: Andrew Lunn @ 2020-10-19 14:26 UTC (permalink / raw)
  To: michael alayev; +Cc: vivien.didelot, netdev, linux-kernel, adror

On Mon, Oct 19, 2020 at 05:05:08PM +0300, michael alayev wrote:
> Hello Andrew,
> 
> 
>     > Please fix your email client and
> 
> 
>     > post the DT file for review. I will
> 
>     > then point out some of the errors.
> 
> 
> 
>     &gem0 {
>         status = "okay";
>         phy-mode = "rgmii-id";
>         phy-handle = <&phy0>;

The diagram you showed had gem0 connected directly to the switch. So
this phy-handle is wrong. Or the diagram is wrong.

> 
>         mdio {
>             #address-cells = <1>;
>             #size-cells = <0>;
> 
>             phy0: ethernet-phy@0 {
>                 compatible = "marvell";
>                 reg = <0>;
>                 device_type = "ethernet-phy";
>                 fixed-link {
>                     speed = <1000>;
>                     full-duplex;
>                 };
>             };
> 
>             debug_phy: ethernet-phy@1 {
>             compatible = "marvell";
>             reg = <1>;
>             device_type = "ethernet-phy";
>             label = "debug-phy";
>         };

indentation is all wrong here.

>         switch0: switch@2 {
>             compatible = "marvell,mv88e6190";
>             #address-cells = <1>;
>             #size-cells = <0>;
>             reg = <2>;
> 
>             dsa,member = <0 0>;
> 
>             ports {
>                 #address-cells = <1>;
>                 #size-cells = <0>;
> 
>                 switch0phy1: port@0 {
>                     reg = <0>;
>                     label = "uid208-cpu";
>                     ethernet = <&gem0>;
>                     phy-mode = "rgmii-id";

You have gem0 using gphy-mode = "rgmii-id" as well. Both doing delays
will not work. You should drop the one in gem0.

>                     fixed-link {
>                         speed = <1000>;
>                         full-duplex;
>                     };
>                 };
> 
> 
>                 port@1 {
>                     reg = <1>;
>                     label = "uid201-1A";
>                 };
> 
>                 port@2 {
>                     reg = <2>;
>                     label = "uid202-2A-p9-1A";
>                     phy-mode = "1000base-x";
>                     fixed-link {
>                         speed = <1000>;
>                         full-duplex;
>                     };

Why both 1000base-x and fixed link? Do you have an SFP connected? If
so, describe the SFP in DT.

>                 };
> 
>                 switch0port10: port@10 {
>                     reg = <10>;
>                     label = "dsa";
>                     link = <&switch1port10>;
>                     phy-mode = "1000base-x";
>                     fixed-link {
>                         speed = <1000>;
>                         full-duplex;
>                     };
>                 };

This is a 6390X right? Why limit it to 1000base-X when it could be
doing 10G?

      Andrew

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, back to index

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <CANBsoPmgct2UTq=Cuf1rXJRitiF1mWhWwdtH2=73yyZiJbT0rg@mail.gmail.com>
2020-10-16  3:31 ` Dts for eth network based on marvell's mv88e6390x crashes Xilinx's linux-kernel v5.4 Andrew Lunn
     [not found]   ` <CANBsoPkCEZadmBaeZ=8EAOP6Ctw5deLen7yKQk__1-ZVoJE6yA@mail.gmail.com>
2020-10-18 15:58     ` Andrew Lunn
     [not found]       ` <CANBsoPm1Ln=59cGKbaA5OKdjA5dwEFA0pcg2tPUa5i2Db747Fw@mail.gmail.com>
2020-10-19 14:26         ` Andrew Lunn

LKML Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/lkml/0 lkml/git/0.git
	git clone --mirror https://lore.kernel.org/lkml/1 lkml/git/1.git
	git clone --mirror https://lore.kernel.org/lkml/2 lkml/git/2.git
	git clone --mirror https://lore.kernel.org/lkml/3 lkml/git/3.git
	git clone --mirror https://lore.kernel.org/lkml/4 lkml/git/4.git
	git clone --mirror https://lore.kernel.org/lkml/5 lkml/git/5.git
	git clone --mirror https://lore.kernel.org/lkml/6 lkml/git/6.git
	git clone --mirror https://lore.kernel.org/lkml/7 lkml/git/7.git
	git clone --mirror https://lore.kernel.org/lkml/8 lkml/git/8.git
	git clone --mirror https://lore.kernel.org/lkml/9 lkml/git/9.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 lkml lkml/ https://lore.kernel.org/lkml \
		linux-kernel@vger.kernel.org
	public-inbox-index lkml

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-kernel


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git