i.MX6DL boot hang on 4.9.88 or 4.14 kernel with latest galcore.ko

[yenchinlee]

Hi:

I test the customized board which has  i.MX6DL and i.MX6Q CPU on imx_4.9.88_2.0.0_ga and imx_4.14.98_2.0.0_ga.

The i.MX6DL kernel hang when insmod the galcore.ko (6.2.4.p1.8 with 4.9 kernel, 6.2.4.p4.0 with 4.14 kernel), but i.MX6Q can use the GPU driver correctly without any issue. (rootfs is builded from official's NXP yocto-bsp)

Here's the log of i.MX6DL:

```

<BOOT on i.MX6DL with custome board>

root@imx6dlsabresd:~# insmod /galcore.ko  
galcore: loading out-of-tree module taints kernel. 
galcore: clk_get vg clock failed, disable vg! 
Galcore version 6.2.4.150331 
alloc_contig_range: [3ca00, 3cc00) PFNs busy 
alloc_contig_range: [3cb00, 3cd00) PFNs busy

<KERNEL HANG WITHOUT FUTHER INFO>

```

And what I tested on i.MX6Q

```

<BOOT on i.MX6Q with custome board>

root@imx6dlsabresd:~# insmod /galcore.ko  
Galcore version 6.2.4.150331
root@imx6dlsabresd:~#
<SYSTEM NOT HANG AND GPU WORK CORRECTLY>

```

Does anyone make i.MX6DL board worked on latest NXP's BSP with GPU-VIV supported ? Or if i.MX6DL can only worked on imx_4.1.15_2.0.0_ga and below ?

[satyajit_bhuyan]

I am having a kernel oops on same processor. Any resolution?

[igorpadykov]

Hi YenChin

one can try with Linux Binary Demo Files - i.MX 6QuadPlus, i.MX 6Quad, i.MX 6DualPlus, i.MX 6Dual, i.MX 6DualLite, i....

gpu driver is described in sect.5.1 Graphics Processing Unit (GPU)

attached Linux Manual

Best regards
igor
-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------

[yenchinlee]

Hi:

The sect.5.1 Graphics Processing Unit (GPU) didn't give  enough information on this problem.

I can use the same u-boot to boot from following version of kernel with related galcore.ko:

- 3.0.35

- 3.10.17_1.0.0_ga

- 3.10.53_1.0.0_ga (gpu driver version 4.x)

- 3.14.52_1.1.0_ga

- 4.1.15_1.0.0_ga  (gpu driver version: 5.x)

All above version of kernel can worked on our customize board with i.MX6DL/D/Q SoC, but after upgrade to 4.9.88 or 4.14.98, only i.MX6D/Q can work with gpu driver.

In my view, the problem on i.MX6DL is the gpu driver (galcore.ko) with version 6.x , since the system hang without any info after doing

insmod galcore.ko

So it is not related to userspace libs. 

[igorpadykov]

one can try to rebuld image from scratch using "imx6dlsabresd"

build configuration as described in sect.5.1 Build configurations

attached Yocto Guide.

Best regards
igor

[yenchinlee]

Yes, I know Yocto Project very well.

The problem is, when I use galcore.ko ver 6.x, the i.MX6DL will hang after insert it, but i.MX6D/Q won't.

So there's might something wrong on device tree or u-boot or galcore.ko it self. 

I just need a clear info if it's my own problem or just a bug in galcore.ko. The galcore.ko ver 4.x or 5.x  was worked very well on my i.MX6DL/D/Q devices, but galcore.ko ver 6.x only worked on my i.MX6D/Q devices, not i.MX6DL.

[liyu1]

I have the same problem. How can I solve it?   Thanks ....    linux 4.14.98 

[yenchinlee]

No, I don't have time to find what's wrong :smileysad:

The 4.1 kernel or below version of BSPs works fine on my i.MX6DL devices, just 4.9.88 and 4.14.98 have such problem.

I still think there's something wrong on the galcore.ko, or maybe I need to update my u-boot to the latest version.

[satyajit_bhuyan]

I am using CPU type:CPU identified as i.MX6DL, silicon rev 1.3

kernel :4.14.174 with kernel-module-imx-gpu-viv 6.2.4.p1.2-r0

But after booting up from SD card, when I am doing login to root:

fusion login:
Poky (Yocto Project Reference Distro) 2.5.1 fusion ttymxc4

fusion login: rooalloc_contig_range: [44700, 44ae8) PFNs busy
alloc_contig_range: [44800, 44be8) PFNs busy
Unable to handle kernel paging request at virtual address 56225ba4
pgd = a94b8000
[56225ba4] *pgd=00000000
Internal error: Oops: 5 [#1] PREEMPT SMP ARM
Modules linked in:
CPU: 0 PID: 299 Comm: HMI Not tainted 4.14.174-ga-fusion+ga1a0dd213d72 #1
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
task: a8ceec00 task.stack: a911c000
PC is at gckDEVICE_Dispatch+0x30/0x124
LR is at drv_ioctl+0x144/0x1f8
pc : [<80705274>] lr : [<806fe068>] psr: 800d0013
sp : a911dd60 ip : 7ef3bca8 fp : 75479d34
r10: 00000036 r9 : a911c000 r8 : 7ef3ba20
r7 : 00000051 r6 : a90b21c0 r5 : a85d2800 r4 : 00000051
r3 : 56225b00 r2 : 0000000b r1 : a911dd88 r0 : a8634e00
Flags: Nzcv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none
Control: 10c53c7d Table: 394b8059 DAC: 00000051
Process HMI (pid: 299, stack limit = 0xa911c210)
Stack: (0xa911dd60 to 0xa911e000)
dd60: 00000051 806fe068 7ef3ba70 00000000 00000190 00000000 7ef3ba70 00000000
dd80: 00000190 00000000 00000028 7ef3bca8 76f6a808 7ef3ce14 75807b8c 00000009
dda0: 754774c0 754a4b20 00000001 0d696914 76f80a80 00000002 76f82ad8 7ef3bb18
ddc0: 000002fa 7ef3bb1c 753998d4 76f60870 00000001 7539c874 000002fa 7539d714
dde0: 76f82ad8 7ef3bb1c 7ef3bb18 7ef3bb48 0000013b 00000000 bb797b34 7539d714
de00: 753998d4 05dbcbd9 7539dc17 7539d714 7539c874 7ef3bb8c 76f87908 76f87fa4
de20: aaaaaaab 76f7f1c8 00000018 76f87908 00000000 00000000 7539dd82 bb797b34
de40: 00000000 76f82c90 7ef3bb8c 7539dc17 7ef3bbec 76f82ad8 7ef3bb94 76f612f8
de60: 7ef3bb94 76f88a64 00000016 00000000 00000001 00000000 00000001 76f82ad8
de80: 7ef3bc1c 76f82c90 00000000 76f612f8 7ef3bbc4 76f88a64 00000016 00000000
dea0: 00000001 037fd108 00000001 7539c874 76f82ad8 76f82c90 00000000 00000000
dec0: 00000001 020743a0 00000001 75398000 000f3518 76f7d4d0 00000000 00000001
dee0: 000f4240 75479d24 75479d34 76f66570 00000000 00000001 00000001 00000000
df00: 000f3214 7539c874 02075e68 754a35a4 00000000 7ef3bc54 ab71d600 7ef3ba20
df20: a86052f0 a8efb540 00007530 8021ae98 75479d34 80226df8 00000006 80226c9c
df40: a8c8a000 00000006 a8efb540 a8c8a000 00000006 a8efb540 a8c8a000 00000020
df60: a8efb548 80206ac4 00000000 00000000 a8efb540 a8efb540 00000006 00007530
df80: 7ef3ba20 a911c000 00000036 8021b6f4 00002710 00007530 754a35a4 00000036
dfa0: 80107ac4 801078c0 00002710 00007530 00000006 00007530 7ef3ba20 020761f8
dfc0: 00002710 00007530 754a35a4 00000036 00000000 7ef3ba70 00000190 75479d34
dfe0: 757909e0 7ef3ba0c 754732a0 757909ec 600d0010 00000006 00000000 00000000
[<80705274>] (gckDEVICE_Dispatch) from [<806fe068>] (drv_ioctl+0x144/0x1f8)
[<806fe068>] (drv_ioctl) from [<8021ae98>] (do_vfs_ioctl+0x9c/0x8c4)
[<8021ae98>] (do_vfs_ioctl) from [<8021b6f4>] (SyS_ioctl+0x34/0x58)
[<8021b6f4>] (SyS_ioctl) from [<801078c0>] (ret_fast_syscall+0x0/0x54)
Code: e5913008 e3a0200b e0233c92 e0803103 (e593c0a4)
---[ end trace 3694221408134be0 ]---
watchdog: watchdog0: watchdog did not stop!

[igorpadykov]

one can check advices from below link and try to debug it:

galcore.ko hangs while loading on 3.14.30 linux 

Best regards
igor

[liyu1]

I have the same problem. How can I solve it?   Thanks ....    linux 4.14.98