Kernel panic
Closed this issue · 10 comments
after compiling latest version, using the example config file, loading the module and running the following:
sudo ffmpeg -i /dev/video0 -f v4l2 /dev/video7
/dev/video2 output was weird, like this:
Then a few seconds later I got kernel panic:
Jan 23 15:56:40 laptop kernel: [ 4311.764840] ------------[ cut here ]------------
Jan 23 15:56:40 laptop kernel: [ 4311.764843] refcount_t: underflow; use-after-free.
Jan 23 15:56:40 laptop kernel: [ 4311.764879] WARNING: CPU: 6 PID: 23877 at lib/refcount.c:28 refcount_warn_saturate+0xae/0xf0
Jan 23 15:56:40 laptop kernel: [ 4311.764880] Modules linked in: akvcam(OE) thunderbolt rfcomm nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_masq xt_conntrack nft_ct xt_MASQUERADE nf_conntr
ack_netlink xfrm_user xfrm_algo xt_addrtype iptable_filter iptable_nat bpfilter br_netfilter nft_chain_nat bridge nf_nat stp nf_conntrack llc nf_defrag_ipv6 nf_defrag_ipv4 libcrc32c nf_tables nfnetlink vb
oxnetadp(OE) vboxnetflt(OE) vboxdrv(OE) ccm aufs cmac algif_hash algif_skcipher af_alg overlay bnep nls_iso8859_1 intel_rapl_msr nvidia_uvm(O) mei_hdcp intel_rapl_common x86_pkg_temp_thermal nvidia_drm(PO
) intel_powerclamp nvidia_modeset(PO) coretemp kvm_intel kvm nvidia(PO) intel_cstate snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_hda_codec_hdmi uvcvideo snd_hda_intel videobuf2_vmalloc s
nd_intel_dspcfg videobuf2_memops videobuf2_v4l2 snd_hda_codec videobuf2_common snd_hda_core videodev btusb snd_hwdep btrtl btbcm mc btintel snd_pcm bluetooth snd_seq_midi snd_seq_midi_event ecdh_generic e
cc snd_rawmidi
Jan 23 15:56:40 laptop kernel: [ 4311.764965] iwlmvm snd_seq mac80211 input_leds joydev libarc4 snd_seq_device snd_timer serio_raw intel_wmi_thunderbolt intel_rapl_perf iwlwifi wmi_bmof snd cfg80211 ipmi
_devintf ipmi_msghandler mei_me soundcore mei intel_pch_thermal ideapad_laptop sparse_keymap mac_hid acpi_pad sch_fq_codel parport_pc ppdev lp parport ip_tables x_tables autofs4 dm_crypt hid_logitech_hidp
p hid_logitech_dj usbhid hid_generic i915 crct10dif_pclmul crc32_pclmul ghash_clmulni_intel i2c_algo_bit aesni_intel drm_kms_helper crypto_simd cryptd glue_helper syscopyarea sysfillrect sysimgblt fb_sys_
fops psmouse cec rc_core nvme intel_lpss_pci drm i2c_i801 intel_lpss nvme_core r8169 ahci idma64 realtek libahci virt_dma i2c_hid hid wmi video pinctrl_sunrisepoint pinctrl_intel
Jan 23 15:56:40 laptop kernel: [ 4311.765047] CPU: 6 PID: 23877 Comm: akvcam-thread-3 Tainted: P OE 5.6.0-1026-oem #26-Ubuntu
Jan 23 15:56:40 laptop kernel: [ 4311.765050] Hardware name: LENOVO INVA/LNVNB161216, BIOS 4GCN38WW 07/18/2018
Jan 23 15:56:40 laptop kernel: [ 4311.765058] RIP: 0010:refcount_warn_saturate+0xae/0xf0
Jan 23 15:56:40 laptop kernel: [ 4311.765064] Code: 7b 31 2e 01 01 e8 77 22 b5 ff 0f 0b 5d c3 80 3d 68 31 2e 01 00 75 91 48 c7 c7 a0 8e 9e bc c6 05 58 31 2e 01 01 e8 57 22 b5 ff <0f> 0b 5d c3 80 3d 46 31
2e 01 00 0f 85 6d ff ff ff 48 c7 c7 f8 8e
Jan 23 15:56:40 laptop kernel: [ 4311.765067] RSP: 0018:ffffb40b46fcfe48 EFLAGS: 00010286
Jan 23 15:56:40 laptop kernel: [ 4311.765071] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000007
Jan 23 15:56:40 laptop kernel: [ 4311.765074] RDX: 0000000000000007 RSI: 0000000000000082 RDI: ffff8e7f9ed99800
Jan 23 15:56:40 laptop kernel: [ 4311.765077] RBP: ffffb40b46fcfe48 R08: 00000000000006fc R09: 0000000000000004
Jan 23 15:56:40 laptop kernel: [ 4311.765080] R10: 0000000000000000 R11: 0000000000000001 R12: ffff8e7e337a2bc0
Jan 23 15:56:40 laptop kernel: [ 4311.765082] R13: 0000000000000000 R14: ffff8e7daf811600 R15: ffffb40b46e7fc40
Jan 23 15:56:40 laptop kernel: [ 4311.765086] FS: 0000000000000000(0000) GS:ffff8e7f9ed80000(0000) knlGS:0000000000000000
Jan 23 15:56:40 laptop kernel: [ 4311.765089] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Jan 23 15:56:40 laptop kernel: [ 4311.765092] CR2: 00007f0408100188 CR3: 00000001f660a002 CR4: 00000000003606e0
Jan 23 15:56:40 laptop kernel: [ 4311.765095] Call Trace:
Jan 23 15:56:40 laptop kernel: [ 4311.765130] akvcam_frame_delete.part.0+0x52/0x60 [akvcam]
Jan 23 15:56:40 laptop kernel: [ 4311.765148] akvcam_frame_delete+0x13/0x20 [akvcam]
Jan 23 15:56:40 laptop kernel: [ 4311.765162] akvcam_default_frame.part.0+0x3e/0xd0 [akvcam]
Jan 23 15:56:40 laptop kernel: [ 4311.765177] akvcam_device_clock_run_once+0x1e3/0x260 [akvcam]
Jan 23 15:56:40 laptop kernel: [ 4311.765191] akvcam_device_clock_timeout+0x3d/0x60 [akvcam]
Jan 23 15:56:40 laptop kernel: [ 4311.765199] kthread+0x104/0x140
Jan 23 15:56:40 laptop kernel: [ 4311.765212] ? akvcam_device_clock_run_once+0x260/0x260 [akvcam]
Jan 23 15:56:40 laptop kernel: [ 4311.765218] ? kthread_park+0x90/0x90
Jan 23 15:56:40 laptop kernel: [ 4311.765227] ret_from_fork+0x35/0x40
Jan 23 15:56:40 laptop kernel: [ 4311.765234] ---[ end trace 24815edb4fa9ce14 ]---
Linux laptop 5.6.0-1026-oem #26-Ubuntu SMP Sun Sep 6 11:06:50 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
Ubuntu 20.04.1 LTS x86_64
Try, with:
ffmpeg -i /dev/video0 -f v4l2 -vcodec rawvideo -pix_fmt rgb24 /dev/videoX
and don't use sudo, sudo is not needed here.
Try, with:
ffmpeg -i /dev/video0 -f v4l2 -vcodec rawvideo -pix_fmt rgb24 /dev/videoXand don't use sudo, sudo is not needed here.
still the same issue. after about 20 seconds whole system just freezes. any ideas on further debugging?
any ideas on further debugging?
I also get crashes after a few seconds. I tried both RT and non RT kernel. Same issue.
Here are the logs:
https://gist.github.com/francescortiz/1fc016416f0db2738b0f4986d08519e9
I am using kernel 5.9.16 and 5.9.1-rt19
Additional information: for me the image looks good, though, I have no rendering issues.
@francescortiz It's crashing here, very similar to @sh4dowb bug, probably due to a race condition.
Hi @hipersayanX , thanks for feedback. Do you know if there is there a workaround? Thanks!
@francescortiz try replacing the following line here with:
akvcam_frame_t default_frame = NULL;Make the change in /usr/src/akvcam-1.2.0, and rebuild it with dkms autoinstall.
Good news! Given your input, I tried now but with default_frame option properly set and it is stable. I created a .bmp with the frame size matching the virtual device, 680x480. So I think that the direction you were pointing is the source of the problem.
This bug should be fixed withing the latest commits, I will make a new release in a few days.