Re: drm/msm/gpu: Cancel idle/boost work on suspend

[Anders Roxell]

Hi

While booting Linux mainline on arm64 qcom db410c device the following
kernel NULL pointer dereference noticed due to this commit [1]
6aa89ae1fb04 ("drm/msm/gpu: Cancel idle/boost work on suspend")

 [   17.207382] Unable to handle kernel NULL pointer dereference at
 virtual address 0000000000000010
 [   17.207755] Mem abort info:
 [   17.215915]   ESR = 0x96000004
 [   17.217903]   EC = 0x25: DABT (current EL), IL = 32 bits
 [   17.225033]   SET = 0, FnV = 0
 [   17.226501]   EA = 0, S1PTW = 0
 [   17.233147]   FSC = 0x04: level 0 translation fault
 [   17.233332] Data abort info:
 [   17.238135]   ISV = 0, ISS = 0x00000004
 [   17.240720]   CM = 0, WnR = 0
 [   17.244061] user pgtable: 4k pages, 48-bit VAs, pgdp=000000008e7b9000
 [   17.247323] [0000000000000010] pgd=080000008e790003,
 p4d=080000008e790003, pud=080000008c2aa003, pmd=0000000000000000
 [   17.254369] Internal error: Oops: 96000004 [#1] PREEMPT SMP
 [   17.264042] Modules linked in: crct10dif_ce qcom_wcnss_pil
 adv7511(+) cec qcom_pon rtc_pm8xxx qcom_spmi_temp_alarm qcom_spmi_vadc
 qcom_vadc_common snd_soc_msm8916_digital snd_soc_msm8916_analog
 qcom_camss snd_soc_apq8016_sbc snd_soc_lpass_apq8016 snd_soc_lpass_cpu
 qcom_q6v5_mss qcom_pil_info snd_soc_lpass_platform qcom_q6v5
 snd_soc_qcom_common msm videobuf2_dma_sg qcom_sysmon venus_core
 qcom_common v4l2_fwnode qcom_rng qcom_glink_smem v4l2_async
 v4l2_mem2mem qmi_helpers qnoc_msm8916 gpu_sched mdt_loader
 videobuf2_memops qcom_stats videobuf2_v4l2 videobuf2_common
 i2c_qcom_cci display_connector icc_smd_rpm drm_kms_helper rpmsg_char
 socinfo rfkill rmtfs_mem drm qrtr fuse
 [   17.306825] CPU: 2 PID: 67 Comm: kworker/2:2 Not tainted 5.17.0-rc1 #1
 [   17.328796] Hardware name: Qualcomm Technologies, Inc. APQ 8016 SBC (DT)
 [   17.335137] Workqueue: pm pm_runtime_work
 [   17.341979] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
 [   17.345896] pc : hrtimer_active+0x14/0x80
 [   17.352652] lr : hrtimer_cancel+0x28/0x70

 Reported-by: Linux Kernel Functional Testing <lkft@xxxxxxxxxx>
 Reported-by: Anders Roxell <anders.roxell@xxxxxxxxxx>

 kernel crash log links [2][3].

Cheers,
Anders
[1] https://lore.kernel.org/all/20220108180913.814448-3-robdclark@xxxxxxxxx/
[2] https://lkft.validation.linaro.org/scheduler/job/4422191#L2360
[3] https://qa-reports.linaro.org/lkft/linux-mainline-master/build/v5.17-rc1-230-g145d9b498fc8/testrun/7713584/suite/linux-log-parser/test/check-kernel-oops-4422047/log