marian34

Ok, I just didn't know where to turn with the problem. Don't take it as remorse. I certainly don't expect you to work on it, it's not worth it. I just thought, when it works on the older kernel, if the solution would be to turn off some new functionality in the newer kernels. We can safely let it float, thanks.

I have a 3A adapter there and I also tested it on a PC power supply. It certainly won't be an AC adapter. There will be a bug in the kernel. Although those pc sat cards of older data on a linux PC (ubuntu, kubuntu latest) work reliably.

Armbianmonitor: http://ix.io/2rfI This is an extract from dmesg when it starts to freeze: [ 268.944228] rcu: INFO: rcu_sched self-detected stall on CPU [ 268.944762] rcu: 1-....: (5385 ticks this GP) idle=ee2/1/0x40000004 softirq=10708/10708 fqs=2058 [ 268.944949] (t=5250 jiffies g=17313 q=123) [ 268.945122] NMI backtrace for cpu 1 [ 268.945429] CPU: 1 PID: 89 Comm: kworker/1:1 Tainted: G C 5.4.50-sunxi #trunk [ 268.945555] Hardware name: Allwinner sun8i Family [ 268.945958] Workqueue: events dbs_work_handler [ 268.946615] [<c010dc49>] (unwind_backtrace) from [<c010a245>] (show_stack+0x11/0x14) [ 268.947066] [<c010a245>] (show_stack) from [<c0861f5b>] (dump_stack+0x6f/0x7c) [ 268.947518] [<c0861f5b>] (dump_stack) from [<c08662bf>] (nmi_cpu_backtrace+0x6b/0x94) [ 268.947940] [<c08662bf>] (nmi_cpu_backtrace) from [<c08663b9>] (nmi_trigger_cpumask_backtrace+0xd1/0xdc) [ 268.948402] [<c08663b9>] (nmi_trigger_cpumask_backtrace) from [<c016f0e1>] (rcu_dump_cpu_stacks+0x7f/0x9c) [ 268.948851] [<c016f0e1>] (rcu_dump_cpu_stacks) from [<c016e48d>] (rcu_sched_clock_irq+0x5a5/0x71c) [ 268.949292] [<c016e48d>] (rcu_sched_clock_irq) from [<c0173f19>] (update_process_times+0x29/0x50) [ 268.949689] [<c0173f19>] (update_process_times) from [<c0181ac7>] (tick_sched_timer+0x37/0x74) [ 268.950081] [<c0181ac7>] (tick_sched_timer) from [<c01748eb>] (__hrtimer_run_queues+0xef/0x230) [ 268.950509] [<c01748eb>] (__hrtimer_run_queues) from [<c017528d>] (hrtimer_interrupt+0xd1/0x1fc) [ 268.950975] [<c017528d>] (hrtimer_interrupt) from [<c07015e3>] (arch_timer_handler_phys+0x27/0x2c) [ 268.951471] [<c07015e3>] (arch_timer_handler_phys) from [<c0165193>] (handle_percpu_devid_irq+0x57/0x1a0) [ 268.951955] [<c0165193>] (handle_percpu_devid_irq) from [<c01610e5>] (generic_handle_irq+0x1d/0x28) [ 268.952378] [<c01610e5>] (generic_handle_irq) from [<c016159f>] (__handle_domain_irq+0x43/0x84) [ 268.952754] [<c016159f>] (__handle_domain_irq) from [<c0517ffd>] (gic_handle_irq+0x39/0x6c) [ 268.953128] [<c0517ffd>] (gic_handle_irq) from [<c0101ae5>] (__irq_svc+0x65/0x94) [ 268.953303] Exception stack(0xef343b70 to 0xef343bb8) [ 268.953585] 3b60: ed3caf8c 80080113 80000000 00004052 [ 268.953950] 3b80: ed3cadb0 00001000 ed3caf8c 80080113 f08f3000 00000fdf 00000021 00000001 [ 268.954258] 3ba0: 00000000 ef343bc0 bf9cbd3d c0875a9c 00080133 ffffffff [ 268.954716] [<c0101ae5>] (__irq_svc) from [<c0875a9c>] (_raw_spin_unlock_irqrestore+0x1c/0x20) [ 268.956005] [<c0875a9c>] (_raw_spin_unlock_irqrestore) from [<bf9cbd3d>] (dvb_dmx_swfilter+0xcd/0x108 [dvb_core]) [ 268.957563] [<bf9cbd3d>] (dvb_dmx_swfilter [dvb_core]) from [<bf9958ab>] (dvb_usb_data_complete+0x27/0x28 [dvb_usb]) [ 268.958095] [<bf9958ab>] (dvb_usb_data_complete [dvb_usb]) from [<ef2e6114>] (0xef2e6114)

Ok, but I also had the latest version from compile.sh there and it worked the same way. The board is powered enough, I tested it.

Is there anything that can be done about it? Something will start(e.g. minisatip, mumudvb) to disproportionately use the cpu and after a short time it will crash completely. A similar configuration works well on an OPI PC(Linux orangepipc 4.10.1-sun8i #1 SMP Sun Mar 5 18:47:02 CET 2017 armv7l GNU/Linux), but with newer versions it behaves just as incorrectly. I've tested other images (ubuntu, debian), but it's the same everywhere.

The motivation is obvious, transcoding the stream (which produces OPi itself) to lower data streams. But it seems that it's too much for this great machine.

I have advanced, but the results are not good. A desktop is needed (/dev/cedar_dev). If you are doing ssh, an export display is needed "export DISPLAY=:0". I have a problem uploading the video itself, after a short time (10s) ends with an error "Stream ends prematurely..." marian@orangepiplus:~$ ffmpeg -hwaccel vdpau -i "http://192.168.1.125:4242/bysid/4905" -c:v rawvideo -an -f rawvideo raw.yuv . . . Stream mapping: Stream #0:0 -> #0:0 (h264 (native) -> rawvideo (native)) Press [q] to stop, [?] for help [h264 @ 0x7f6afe00] co located POCs unavailable [h264 @ 0x7f672bd0] co located POCs unavailable [http @ 0x7f60b920] Stream ends prematurely at 11811392, should be 18446744073709551615=11 drop=0 speed=0.555x http://192.168.1.125:4242/bysid/4905: Input/output error.60 bitrate=622080.0kbits/s dup=11 drop=0 speed=0.557x [h264 @ 0x7fd11560] error while decoding MB 42 10, bytestream -10 [h264 @ 0x7fd11560] concealing 6967 DC, 6967 AC, 6967 MV errors in P frame http://192.168.1.125:4242/bysid/4905: Input/output error Last message repeated 3 times http://192.168.1.125:4242/bysid/4905: Input/output error.84 bitrate=622080.0kbits/s dup=11 drop=0 speed=0.539x Last message repeated 1 times frame= 275 fps= 14 q=-0.0 Lsize= 835312kB time=00:00:11.00 bitrate=622080.0kbits/s dup=14 drop=0 speed=0.543x video:835312kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.000000% On this .yuv file processed it is possible to apply a cedrus encoder...ok. It is not possible for the decoder / encoder to work at the same time! (???). Something similar is also written elsewhere, it's a shame I did not find it earlier.

I wanted to test HW encoding, and I found only the Cedrus HW Encoder. It is possible to transcode(stream to stream) video(h264) using x264 libray? Probably not.

I have opi2 + (debian server) marian@orangepiplus:~/FFmpeg$ cat /etc/*release # PLEASE DO NOT EDIT THIS FILE BOARD=orangepiplus BOARD_NAME="Orange Pi+" VERSION=5.24 LINUXFAMILY=sun8i BRANCH=dev ARCH=arm IMAGE_TYPE=stable PRETTY_NAME="Debian GNU/Linux 8 (jessie)" NAME="Debian GNU/Linux" VERSION_ID="8" VERSION="8 (jessie)" ID=debian HOME_URL="http://www.debian.org/" SUPPORT_URL="http://www.debian.org/support" BUG_REPORT_URL="https://bugs.debian.org/" I have gone through a lot of instructions, but without success. I always end up on: marian@orangepiplus:~$ sudo ffmpeg -i vlc.ts -pix_fmt nv12 -qp 20 -c:v cedrus264 -f null /dev/null ... Stream #0:6[0x140f]: Unknown: none ([5][0][0][0] / 0x0005) Stream #0:7[0x14eb](slo): Subtitle: dvb_teletext ([6][0][0][0] / 0x0006) [cedrus264 @ 0xf2ca40] VE Open error. Output #0, null, to '/dev/null': Stream #0:0: Video: h264, q=2-31, 128 kb/s, SAR 1:1 DAR 0:0, 25 fps Metadata: encoder : Lavc56.0.101 cedrus264 Stream #0:1(slo): Audio: pcm_s16le, 0 channels Metadata: encoder : Lavc56.0.101 pcm_s16le Stream mapping: Stream #0:0 -> #0:0 (h264 (native) -> h264 (cedrus264)) Stream #0:1 -> #0:1 (mp2 (native) -> pcm_s16le (native)) Error while opening encoder for output stream #0:0 - maybe incorrect parameters such as bit_rate, rate, width or height ...VE open error. what am I doing wrong? I went after this ( https://github.com/avafinger/ffmpeg_cedrus264_H3_Armbian_5.23 ) , even with custom compilation. Somewhere is talk about "modprobe sunxi_cedar_mod", but I did not know what to do about the module. Or, the module is not needed and support is directly in ffmpeg_cedrus? ??? or is the module only in desktop?

Hi, Linux orangepiplus 4.9.4-sun8i #3 SMP Wed Feb 1 12:29:48 CET 2017 armv7l GNU/Linux it's ok(skystar usb_2_hd_ci). thanks. Except for a small bug-after connecting TT connect 3650 not work ssh. After entering the password, nothing happens, console freezes. But this is a detail of.

Does not work well. Unicable and also no_unicable system. Tools like mumudvb, dvblast, minisatip reported some time after launch errors and the picture falls apart. For example mumudvb: ERRO: SCAM_SEND: thread starved, channel STV1_Vypiname 3059 0 ERRO: SCAM_SEND: thread starved, channel STV1_Vypiname 3145 0 ERRO: SCAM_SEND: thread starved, channel STV1_Vypiname 3214 0 ERRO: SCAM_SEND: thread starved, channel STV1_Vypiname 3262 0

Hi, unicable(dvbs2) works: marian@orangepiplus2e:~$ uname -a Linux orangepiplus2e 4.9.0-sun8i #5 SMP Sat Dec 10 13:00:16 CET 2016 armv7l GNU/Linux tested with opi_pc(jessie desktop). Thanks.

http://sprunge.us/JfDK No unicable system work, including dvb-s2. Unicable convertor It is properly tuned, tested on another system trough sat_hub.

Hello, I have OPI_PC(or OPI_+2): marian@orangepipc:~$ cat /etc/*-release # PLEASE DO NOT EDIT THIS FILE BOARD=orangepipc BOARD_NAME="Orange Pi PC" VERSION=5.20 LINUXFAMILY=sun8i BRANCH=default ARCH=arm PRETTY_NAME="Debian GNU/Linux 8 (jessie)" NAME="Debian GNU/Linux" VERSION_ID="8" VERSION="8 (jessie)" ID=debian HOME_URL="http://www.debian.org/" SUPPORT_URL="http://www.debian.org/support" BUG_REPORT_URL="https://bugs.debian.org/" and Skystar usb_2_hd_ci(or TT connect 3650). marian@orangepipc:~$ Module Size Used by mali_drm 2732 1 drm 178255 2 mali_drm mali 123208 0 ump 29379 3 mali sunxi_cir 1601 0 dvb_usb_az6027 12997 10 stb0899 29431 1 dvb_usb_az6027 stb6100 5436 1 dvb_usb_az6027 dvb_usb 15734 1 dvb_usb_az6027 dvb_core 68509 2 dvb_usb,dvb_usb_az6027 8189es 1076034 0 Unicable dvb-s2 don't work(no lock signal).DVB-S transponders work.HW error excluded. Do you have any ideas?