Re: [REGRESSION] boot failure on imx8mp due to e0203ddf9af7

["Ming Qian(OSS)" <ming.qian@oss.nxp.com>]

Hi,

On 3/5/2026 6:43 AM, Marco Felsch wrote:
> Hi Franz,
> 
> On 26-03-04, Franz Schnyder wrote:
>> Hello,
>>
>> Our iMX8MP SoMs (Verdin and Toradex SMARC) fail to boot after
>> commit e0203ddf9af7 ("media: verisilicon: Avoid G2 bus error while decoding H.264 and HEVC").
>> On rare occasions, it's still able to boot.
>> It fails after having registered g1 VPU, probably when it tries to
>> register g2 VPU.
>>
>> Ming: Any ideas on how to proceed here?
>>
>> Success::
>> [    9.040517] hantro-vpu 38300000.video-codec: registered nxp,imx8mm-vpu-g1-dec as /dev/video0
>> [    9.052133] hantro-vpu 38310000.video-codec: registered nxp,imx8mq-vpu-g2-dec as /dev/video1
>>
>>
>> Failure with decoded stacktrace::
>>
>> [    9.004477] hantro-vpu 38300000.video-codec: registered nxp,imx8mm-vpu-g1-dec as /dev/video0
>> [    9.027097] Unable to handle kernel paging request at virtual address ffffd9c18eb05000
> 
> This is quite a nasty bug you triggered here! I think the issue is that
> the newly added imx8mq_vpu_shared_resources[] is of type __initconst
> where as it should be __initconst_or_module. You can test this
> assumption if you switch from a module [M] build for the hantro_vpu
> driver to a builtin [x] build.

I will correct __initconst to __initconst_or_module.
Thank you for pointing out the issue.

Regards,
Ming

> 
> Regards,
>    Marco
> 
> 
>> [    9.027132] Mem abort info:
>> [    9.027139]   ESR = 0x0000000096000007
>> [    9.027147]   EC = 0x25: DABT (current EL), IL = 32 bits
>> [    9.027156]   SET = 0, FnV = 0
>> [    9.027164]   EA = 0, S1PTW = 0
>> [    9.027171]   FSC = 0x07: level 3 translation fault
>> [    9.027180] Data abort info:
>> [    9.027187]   ISV = 0, ISS = 0x00000007, ISS2 = 0x00000000
>> [    9.027196]   CM = 0, WnR = 0, TnD = 0, TagAccess = 0
>> [    9.027205]   GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
>> [    9.027215] swapper pgtable: 4k pages, 48-bit VAs, pgdp=000000004924f000
>> [    9.027226] [ffffd9c18eb05000] pgd=0000000000000000, p4d=10000000496dd003, pud=1000000102e42403, pmd=100000010c33d403, pte=0000000000000000
>> [    9.027258] Internal error: Oops: 0000000096000007 [#1]  SMP
>> [    9.101669] Modules linked in: snd_soc_fsl_sai crypto_engine rng_core imx_pcm_dma authenc snd_soc_fsl_aud2htx snd_soc_fsl_utils imx8mp_interconnect libdes i
>> mx8mp_hdmi_pai evdev imx8mp_hdmi_tx aes_ce_blk dw_hdmi ghash_ce drm_display_helper snd_soc_simple_card hantro_vpu phy_fsl_imx8mq_usb libaes phy_fsl_samsung_hdm
>> i snd_soc_simple_card_utils cec phy_fsl_imx8m_pcie v4l2_vp9 imx8mp_hdmi_pvi typec fsl_imx8_ddr_perf snd_soc_imx_hdmi gf128mul samsung_dsim v4l2_h264 etnaviv ti
>> _ads1015 industrialio_triggered_buffer snd_soc_nau8822 ina2xx lm75 gpu_sched i3c lontium_lt8912b kfifo_buf governor_userspace imx_sdma flexcan imx_aipstz dwc3_
>> imx8mp imx_bus can_dev spi_nxp_fspi caam spi_imx error pwm_imx27 imx8mm_thermal nvmem_snvs_lpgpr snvs_pwrkey usb_conn_gpio gpio_keys display_connector roles lo
>> op fuse ipv6 libsha1 autofs4
>> [    9.173690] CPU: 3 UID: 0 PID: 45 Comm: kworker/u16:2 Not tainted 6.19.0+ #47 PREEMPT
>> [    9.181623] Hardware name: Toradex Verdin iMX8M Plus on Verdin Development Board (DT)
>> [    9.189462] Workqueue: events_unbound deferred_probe_work_func
>> [    9.195314] pstate: 600000c5 (nZCv daIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
>> [    9.202284] pc : of_find_matching_node_and_match (drivers/of/base.c:1129 drivers/of/base.c:1120 drivers/of/base.c:1187)
>> [    9.207869] lr : of_find_matching_node_and_match (drivers/of/base.c:1185)
>> [    9.213452] sp : ffff8000805dbad0
>> [    9.216771] x29: ffff8000805dbad0 x28: 0000000000000000 x27: 0000000000000000
>> [    9.223927] x26: 0000000000000000 x25: 0000000000000000 x24: ffffd9c2060d48e0
>> [    9.231078] x23: 0000000000000000 x22: ffffd9c18eb05000 x21: ffff0000ff79d7f8
>> [    9.238231] x20: 0000000000000000 x19: ffffd9c18eb05000 x18: 00000000ffffffff
>> [    9.245385] x17: ffff263ef9a67000 x16: ffffd9c2055d3d54 x15: ffff8000805db870
>> [    9.252536] x14: 0000000000000000 x13: 006365646f632d6f x12: 0000000000000001
>> [    9.259689] x11: 0000000000000003 x10: 0000000000000020 x9 : 0000000000000000
>> [    9.266842] x8 : 0000000000000020 x7 : 00000000ffffffff x6 : ffff0000c4a9d8c8
>> [    9.273995] x5 : 0000000000000014 x4 : 0000000000000000 x3 : 0000000000000001
>> [    9.281145] x2 : 0000000000000000 x1 : 0000000000000000 x0 : 0000000000000000
>> [    9.288301] Call trace:
>> [    9.290753]  of_find_matching_node_and_match (drivers/of/base.c:1129 drivers/of/base.c:1120 drivers/of/base.c:1187) (P)
>> [    9.296336] hantro_probe (./include/linux/of.h:983 drivers/media/platform/verisilicon/hantro_drv.c:1047 drivers/media/platform/verisilicon/hantro_drv.c:1225) hantro_vpu
>> [    9.301144]  platform_probe (drivers/base/platform.c:1449)
>> [    9.304815]  really_probe (drivers/base/dd.c:583 drivers/base/dd.c:661)
>> [    9.308399]  __driver_probe_device (drivers/base/dd.c:803)
>> [    9.312765]  driver_probe_device (drivers/base/dd.c:833)
>> [    9.316956]  __device_attach_driver (drivers/base/dd.c:962)
>> [    9.321415]  bus_for_each_drv (drivers/base/bus.c:500)
>> [    9.325262]  __device_attach (drivers/base/dd.c:1035)
>> [    9.329109]  device_initial_probe (drivers/base/dd.c:1088)
>> [    9.333303]  bus_probe_device (drivers/base/bus.c:576)
>> [    9.337152]  deferred_probe_work_func (drivers/base/dd.c:125 (discriminator 14))
>> [    9.341697]  process_one_work (kernel/workqueue.c:3268)
>> [    9.345719]  worker_thread (kernel/workqueue.c:3334 kernel/workqueue.c:3421)
>> [    9.349477]  kthread (kernel/kthread.c:467)
>> [    9.352716]  ret_from_fork (arch/arm64/kernel/entry.S:861)
>> [    9.356303] Code: 6b14001f 9a93d35a 1a80d294 91032273 (39400264)
>> All code
>> ========
>>     0:   6b14001f        cmp     w0, w20
>>     4:   9a93d35a        csel    x26, x26, x19, le
>>     8:   1a80d294        csel    w20, w20, w0, le
>>     c:   91032273        add     x19, x19, #0xc8
>>    10:*  39400264        ldrb    w4, [x19]               <-- trapping instruction
>>
>> Code starting with the faulting instruction
>> ===========================================
>>     0:   39400264        ldrb    w4, [x19]
>> [    9.362413] ---[ end trace 0000000000000000 ]---
>> [    9.367037] note: kworker/u16:2[45] exited with irqs disabled
>> [    9.367134] note: kworker/u16:2[45] exited with preempt_count 1
>>
>>
>> Thanks,
>> Franz
>>
>>
>>
>