Petalinux boot hangs after "mali: loading out-of-tree module taints kernel."

[wtneo@leica]

Software:

  Petalinux 2018.1

  xilinx-zcu104-v2018.1-final.bsp

  https://github.com/Xilinx/hdf-examples/tree/rel-v2018.1/zcu104-zynqmp/system.hdf

Hardware: ZCU104

  I am trying to follow tutorial UG1209 (2018.1) for ZCU104.  UG1209 is for ZCU102 but I do not own it.  ZCU102 has 4GB of PS ram whereas ZCU102 has only 2GB.

  I followed UG1209 to configure makemenuconfig, inlcuding  Subsystem AUTO Hardware Settings->Memory Settings->System Memory Size = 0x6D5FFFFF.  I am not sure if this setting is applicable to ZCU104

  Compile was successful.  Seem like pmufw is auto packaged into boot files. But Petalinux hangs during boot and the last few messages were:

...

...

INIT: version 2.88 booting
Starting udev
[ 6.658973] udevd[1750]: starting version 3.2.2
[ 6.668610] udevd[1751]: starting eudev-3.2.2
[ 6.715919] mali: loading out-of-tree module taints kernel.

....HANGS....

  I found from this forum, in a successful boot, the next message should be

[ 4.696398] zynqmp_gpd_attach_dev error -13, node 22
[ 4.701374] dwc3-of-simple ff9d0000.usb0: failed to add to PM domain pd-usb0: -13
[ 4.709227] zynqmp_gpd_attach_dev error -13, node 59
[ 4.714199] nwl-pcie fd0e0000.pcie: failed to add to PM domain pd-pcie: -13
Populating dev cache
Starting internet superserver: inetd.

  Pls advise how to configure Petalinux correctly for ZCU104

Tks,

Neo

[sandeepg]

Hi wtneo@leica,

ZCU102 board has 4GB RAM https://www.xilinx.com/products/boards-and-kits/ek-u1-zcu102-g.html so you don't need to change the memory configuration.

We have a bug in tuning system memory size in petalinux which should be fixed in 2018.2 I would suggest to use the default configurations generated by petalinux.

[wtneo@leica]

Hi Sandeep,

    Do you mean that:

   1. I DO NOT follow UG1209 to configure makemenuconfig, inlcuding Subsystem AUTO Hardware Settings->Memory Settings->System Memory Size = 0x6D5FFFFF.

   2. Instead leave System Memory Size setting at 0x80000000 (default)

Neo

[sandeepg]

Yes correct.

[wtneo@leica]

Hi Sandeep,

  I still get the same error.

Step I did:

1. petalinux-config

2. Subsystem AUTO Hardware Settings->Memory Settings->System Memory Size = 0x80000000

3. petalinux-build

Could it have something to do with pmufw?  Because in a normal boot, the next dmesg line would be "zynqmp_gpd_attach_dev" ,something related to pmu

Regards,

Neo

[tmonaghan84]

Did you manage to get this working?

I am facing the same fault using a ZCU104 , I have been working through the Xilinx Guide UG1209, and creating the boot file with 4 partitions as described on page 64 on wards.

This create partitions  board boots the sequence below.

FSBL -> PMU -> ATF ->  UBOOT -> Petalinux (hangs output as below)

This is PetaLinux  2018.2, although UBoot (1) and ATF (2) are reporting 2018.1

1) U-Boot 2018.01 (Jul 27 2018 - 09:38:11 +0100) Xilinx ZynqMP ZCU104 revC

2) NOTICE:  BL31: v1.4(release):xilinx-v2018.1-4-g93a69a5a

PetaLinux output before hanging.

--------------------------------------------

[ 6.666718] ALSA device list:
[ 6.669627] #0: DisplayPort monitor
[ 6.754039] ata2: SATA link down (SStatus 0 SControl 330)
[ 6.759379] ata1: SATA link down (SStatus 0 SControl 330)
[ 6.764872] Freeing unused kernel memory: 512K
INIT: version 2.88 booting
Starting udev
[ 6.870912] udevd[1766]: starting version 3.2.2
[ 6.880688] udevd[1767]: starting eudev-3.2.2
[ 6.926780] mali: loading out-of-tree module taints kernel.                     <<<<Hangs here

Any ideas on how to proceed?

[swordwest]

I have the same question!

[tmonaghan84]

Swordwest could you elaborate on your process/setup just to understand the common component?

I am now failing to get any PetaLinux build working (previously this was working)

I have created a new fresh PetaLinux project and it builds but on boot hangs at the same "mali" point, commands used below;

petaLinux -t project -s Xilinx-zcu104-v2018.2-final.bsp"

cd <project dir>

petaLinux-build

petaLinux-package --boot --fsbl zynqmp_fsbl.elf --uboot

I then copied boot.bin & image.ub to the SD card,This boots on the ZCU104 then hang at during PetaLinux booting at the "Mali" line mentioned.

~I am wondering if my previous working build was on 2018.1?

[swordwest]

I use petalinux in the vmware ubuntu16.04. When I use preImage ,it workwell, if I use self building version,I have the same error.

[tmonaghan84]

Swordwest some questions;

1. Are using ZCU104 development board?
2. Are you using the 2018.2 BSP for the ZCU104 (Xilinx-zcu104-v2018.2-final.bsp)?
3. Are you using PetaLinux 2018.2 ?

Thanks

Tim

[swordwest]

I have ZCU104 and 2018.2 BSP for the ZCU104 (Xilinx-zcu104-v2018.2-final.bsp) and PetaLinux 2018.2

[swordwest]

petalinux-package --boot --format BIN --fsbl zynqmp_fsbl.elf --fpga system.bit --u-boot

It's OK now!

[tmonaghan84]

Managed to get it working thanks for the pointer @swordwest

I was not embedding the FPGA bit stream in the PetaLinux-package command, I guess the MALI requires some configuration of the hardware?

I ran the command below to get it working, the red section I omitted previously when it was hanging.

$petalinux-package --boot --fsbl zynqmp_fsbl.elf --fpga system.bit --u-boot

Ps. the "--format" bin isn't required as this is the default output format.

[leonliu]

Did you fix this nwl-pcie problem?

I am using 2018.3, it seems coming back

Print the same messages, but system keeps booting. Eventually, sysgtem is on with login prompt, but eth0 was not find 

[ 2.048806] nwl-pcie fd0e0000.pcie: failed to add to PM domain pd-pcie: -13
[ 2.056172] zynqmp_gpd_attach_dev error -13, node 41
[ 2.060654] xilinx-dpdma fd4c0000.dma: failed to add to PM domain pd-dp: -13

INIT: Entering runlevel: 5
Configuring network interfaces... Cannot find device "eth0"

I have also many failed PM-domain messages

[ 2.203990] Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled
[ 2.207640] zynqmp_gpd_attach_dev error -13, node 41
[ 2.209651] zynqmp-display fd4a0000.zynqmp-display: failed to add to PM domain pd-3
[ 2.217893] cacheinfo: Unable to detect cache hierarchy for CPU 0
[ 2.217926] PLL: shutdown
[ 2.217940] zynqmp_pll_disable() clock disable failed for dpll_int, ret = -13
[ 2.237888] brd: module loaded
[ 2.241397] loop: module loaded
[ 2.242298] zynqmp_gpd_attach_dev error -13, node 28
[ 2.244582] ahci-ceva fd0c0000.ahci: failed to add to PM domain pd-sata: -13
[ 2.251714] mtdoops: mtd device (mtddev=name/number) must be supplied
[ 2.258886] m25p80 spi0.0: found n25q00, expected n25q512a
[ 2.263886] m25p80 spi0.0: n25q00 (262144 Kbytes)
[ 2.268131] 4 ofpart partitions found on MTD device spi0.0
[ 2.273561] Creating 4 MTD partitions on "spi0.0":
[ 2.278320] 0x000000000000-0x000001e00000 : "boot"
[ 2.283563] 0x000001e00000-0x000001e40000 : "bootenv"
[ 2.288475] 0x000001e40000-0x000004240000 : "kernel"
[ 2.293396] 0x000004240000-0x000010000000 : "spare"
[ 2.299158] libphy: Fixed MDIO Bus: probed
[ 2.303004] tun: Universal TUN/TAP device driver, 1.6
[ 2.307078] CAN device driver interface
[ 2.311041] zynqmp-pinctrl ff180000.pinctrl: set mux failed for pin 24
[ 2.317267] xilinx_can ff070000.can: Error applying setting, reverse things back
[ 2.324647] zynqmp_gpd_attach_dev error -13, node 48
[ 2.329529] xilinx_can ff070000.can: failed to add to PM domain pd-can1: -13
[ 2.337873] macb ff0e0000.ethernet: Not enabling partial store and forward
[ 2.343841] libphy: MACB_mii_bus: probed
[ 2.377867] usbcore: registered new interface driver asix
[ 2.377950] usbcore: registered new interface driver ax88179_178a
[ 2.383702] usbcore: registered new interface driver cdc_ether
[ 2.389497] usbcore: registered new interface driver net1080
[ 2.395116] usbcore: registered new interface driver cdc_subset
[ 2.401004] usbcore: registered new interface driver zaurus
[ 2.406543] usbcore: registered new interface driver cdc_ncm
[ 2.412417] xilinx-axipmon ffa00000.perf-monitor: Probed Xilinx APM
[ 2.418970] zynqmp-pinctrl ff180000.pinctrl: set mux failed for pin 52