| Summary: | Failure of recent DVB-T2 devices to be recognized | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | Len Lawrence <tarazed25> |
| Component: | RPM Packages | Assignee: | Kernel and Drivers maintainers <kernel> |
| Status: | NEW --- | QA Contact: | |
| Severity: | normal | ||
| Priority: | Normal | CC: | marja11, thierry.vignaud, tmb |
| Version: | Cauldron | Keywords: | NEEDINFO |
| Target Milestone: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | MGA5TOO | ||
| Source RPM: | kernel, kernel-firmware | CVE: | |
| Status comment: | |||
| Attachments: |
Section of journal around the connection time for 292e
Tail end of dmesg coveringdisconnection of 290e and connection of 292e. Verbose lsusb output for soloHD Verbose lspcidrake output Retry using downloaded firmware Some results using Nova triplestick Latest attempt to use the Nova PCTV stick dmesg after plugging in the PCTV 292e : 2018-08-29 Attempt to build channel scan file for PCtv 292e |
||
|
Description
Len Lawrence
2016-07-19 22:46:48 CEST
The front-end capabilities of the tuners are detected provided the firmware is in place.
Marja Van Waes
2016-07-20 18:42:20 CEST
CC:
(none) =>
marja11, thierry.vignaud, tmb The PCTV 290e worked out of the box but has been superseded by the 292e. If you order a Nova Triplestick expecting 290e what you get is the 292e. (In reply to Len Lawrence from comment #0) > > How reproducible: > Every time on different machines with kernels 4.4 - 4.7. So valid in cauldron, too. Which kernels did you try? (desktop, tmb, linus...?) CC:
thierry.vignaud =>
(none) As far as I can remember it was the desktop kernels but tmb and linus have visited at least one of the machines but been superseded by the latest desktop kernels in time. Is there a good reason to try tmb or linus with respect to this bug? I had assumed that Mageia had simply not kept up with the latest hardware because there may be few or no people in the dev teams who use these devices. (In reply to Len Lawrence from comment #4) > Is there a good reason to try tmb or linus with > respect to this bug? Not that I'm aware of, only asked to get this report as complete as possible. I may have been wrong to assign this bug to the kernel, maybe the problem isn't there, but with proper detection of newer devices? AFAIK we don't remove support from kernels we ship ;-) Do you mind attaching output.txt that is the result of running, as root, journalctl -af 2>&1 | tee output.txt from just before before attaching the devices until after output related to attaching them stops? And, could you also attach usb.txt that is the result of then running lsusb 2>&1 | tee usb.txt (install usbutils first if the lsusb command isn't recognized) Severity:
enhancement =>
normal Attaching journal output for before and after connection of PCTV 292e and subsequent module listing. End of dmesg also attached. $ lsusb Bus 002 Device 002: ID 8087:8001 Intel Corp. Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 001 Device 002: ID 8087:8009 Intel Corp. Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 004 Device 003: ID 1058:0820 Western Digital Technologies, Inc. Bus 004 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub Bus 003 Device 007: ID 2040:0264 Hauppauge Bus 003 Device 004: ID 046d:c534 Logitech, Inc. Unifying Receiver Bus 003 Device 002: ID 0a5c:21e8 Broadcom Corp. BCM20702A0 Bluetooth 4.0 Bus 003 Device 005: ID 1b1c:0c04 Corsair Bus 003 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub # lsmod | grep em28xx em28xx_rc 20480 0 rc_core 28672 2 em28xx_rc,rc_pinnacle_pctv_hd em28xx_dvb 32768 0 dvb_core 126976 2 em28xx_dvb,cxd2820r em28xx 90112 2 em28xx_dvb,em28xx_rc tveeprom 24576 1 em28xx v4l2_common 16384 1 em28xx videodev 172032 2 em28xx,v4l2_common usbcore 237568 10 uas,btusb,em28xx_dvb,usb_storage,em28xx,ehci_hcd,ehci_pci,usbhid,xhci_hcd,xhci_pci The rc_pinnacle_pctv_hd module belongs to the 290e I think, disconnected earlier. The 292e powers up and is detected but the system fails to recognize it as a DVB device as far as I can make out, in spite of the firmware in /usr/lib/firmware. I shall try to track down the reference to the patch needed for em28xx. Created attachment 8314 [details]
Section of journal around the connection time for 292e
Created attachment 8315 [details]
Tail end of dmesg coveringdisconnection of 290e and connection of 292e.
These links appear to be relevant to the em28xx and em28xx-dvb modules/drivers. https://patchwork.linuxtv.org/patch/23618/ https://patchwork.linuxtv.org/patch/23763/ https://patchwork.kernel.org/patch/9048201/ ( http://www.spinics.net/lists/linux-media/msg76076.html ) ( https://lwn.net/Articles/594862/ ) ( https://tvheadend.org/boards/4/topics/15050?r=15073 ) ( https://www.google.co.uk/#q=292e+em28xx+patch&start=10 ) ( http://linuxtv.fi/viewtopic.php?f=15&t=5099 ) Created attachment 8316 [details]
Verbose lsusb output for soloHD
em28xx is enabled as checked by running the following command rpm -ql kernel-desktop-4.7.0-2.mga6 |grep em28xx What's the output of "lspcidrake -v" on the systems where you see "unknown" as driver? (to be attached, not pasted) Keywords:
(none) =>
NEEDINFO commit 6db0e4d62071223b58f28f2718105a8ad51f3c18
Author: Thierry Vignaud <thierry.vignaud@...>
Date: Sat Aug 6 20:04:36 2016 +0200
em28xx drives TV cards, not webcam (mga#18995)
---
Commit Link:
http://gitweb.mageia.org/software/drakx/commit/?id=6db0e4d62071223b58f28f2718105a8ad51f3c18
Ran the command: $ rpm -ql 4.4.16-desktop-1.mga5 |grep em28xx No output. Attaching lspcidrake output. Created attachment 8317 [details]
Verbose lspcidrake output
This is the firmware installed manually: $ ls -l /lib/firmware | grep dvb -rw-r--r-- 1 root root 73960 Aug 6 17:25 dvb-demod-si2168-02.fw -rw-r--r-- 1 root root 74148 Aug 6 17:25 dvb-demod-si2168-a20-01.fw -rw-r--r-- 1 root root 74945 Aug 6 17:25 dvb-demod-si2168-a30-01.fw * -rw-r--r-- 1 root root 74945 Aug 6 17:25 dvb-demod-si2168-b40-01.fw * -rw-r--r-- 1 root root 74944 Aug 6 17:25 dvb-tuner-si2158-a20-01.fw * Three of these look suspiciously similar in size. Something very strange going on here. A diff indicates that two of the files contain mainly html type code. So clicking on the links at the URL given in the opening comment does not in fact download the firmware. (In reply to Len Lawrence from comment #13) > Ran the command: > > $ rpm -ql 4.4.16-desktop-1.mga5 |grep em28xx This cannot work... This is not a valid query as this is not a real package name. See: $ rpm -qpl kernel-desktop-4.4.16-1.mga5-1-1.mga5.x86_64.rpm |grep em28xx|wc -l 6 Sorry. Too much going on around here. That was a cut-and-paste error from uname -r. $ rpm -ql kernel-desktop-4.4.16-1.mga5 | grep em28xx/lib/modules/4.4.16-desktop-1.mga5/kernel/drivers/media/usb/em28xx /lib/modules/4.4.16-desktop-1.mga5/kernel/drivers/media/usb/em28xx/em28xx-alsa.ko.xz /lib/modules/4.4.16-desktop-1.mga5/kernel/drivers/media/usb/em28xx/em28xx-dvb.ko.xz /lib/modules/4.4.16-desktop-1.mga5/kernel/drivers/media/usb/em28xx/em28xx-rc.ko.xz /lib/modules/4.4.16-desktop-1.mga5/kernel/drivers/media/usb/em28xx/em28xx-v4l.ko.xz /lib/modules/4.4.16-desktop-1.mga5/kernel/drivers/media/usb/em28xx/em28xx.ko.xz Urrk! $ rpm -ql kernel-desktop-4.4.16-1.mga5 | grep em28xx /lib/modules/4.4.16-desktop-1.mga5/kernel/drivers/media/usb/em28xx Downloaded and installed the correct firmware and rebooted. See brief report attached. Created attachment 8318 [details]
Retry using downloaded firmware
Created attachment 8319 [details]
Some results using Nova triplestick
Mass-reassigning all bugs with "kernel" in the Source RPM field that are assigned to tmb, to the kernel packagers group, because tmb is currently MIA. Assignee:
tmb =>
kernel commit 47a242eb87e4f48a38a4f68d2b584c689e7cb517
Author: Thierry Vignaud <thierry.vignaud@...>
Date: Sat Aug 6 20:04:36 2016 +0200
em28xx drives TV cards, not webcam (mga#18995)
(cherry picked from commit 6db0e4d62071223b58f28f2718105a8ad51f3c18)
---
Commit Link:
http://gitweb.mageia.org/software/drakx/commit/?id=47a242eb87e4f48a38a4f68d2b584c689e7cb517
I have just had a look at this again and find that the latest PCTV stick still fails to be recognized. It shows a blue light when first plugged in but goes dark immediately. Although the Si2168 firmware is missing w_scan is able to identify the device and its capabilities but always reports "ERROR: Sorry - i couldn't get any working frequency/transponder Nothing to scan!!" after some six minutes. When scanning is successful it usually takes at least 12 minutes. Attaching the latest notes. Created attachment 9574 [details]
Latest attempt to use the Nova PCTV stick
Having another go at using the latest PCTV triplestick 292e. $ lsusb Bus 003 Device 002: ID 2013:025f PCTV Systems Both kaffeine and vlc were able to display SD channels without any retuning, but no HD. Attempted to use w_scan to rescan the channels and it failed after recognizing the Silicon Labs Si2168. ERROR: Sorry - i couldn't get any working frequency/transponder Nothing to scan!! kaffeine does not see any transponders although it knows about the Silicon Labs tuner. That means it cannot scan for SD channels either. Yet it can see the SD channels using an existing channels.dvb file. A brief tour of the internet shows that most of the users of this device have been unsuccessful with HD as well. Maybe we should give up on this and mark it WONTFIX. In spite of what the following two extracts show, neither kaffeine nor w_scan succeeds in using the channel tuner to find free-to-air terrestrial channels. $ sudo lsmod | grep em28 em28xx_rc 20480 0 rc_core 28672 3 rc_pinnacle_pctv_hd,em28xx_rc em28xx_dvb 32768 0 dvb_core 122880 1 em28xx_dvb em28xx 90112 2 em28xx_rc,em28xx_dvb tveeprom 24576 1 em28xx v4l2_common 16384 1 em28xx videodev 167936 2 v4l2_common,em28xx media 40960 3 videodev,em28xx,si2157 usbcore 245760 10 usbhid,usb_storage,ehci_hcd,xhci_pci,em28xx,em28xx_dvb,btusb,uas,xhci_hcd,ehci_pci $ dmesg | grep em28 [ 1.866177] usbcore: registered new interface driver em28xx [ 2.080788] em28xx: New device PCTV PCTV 292e @ 480 Mbps (2013:025f, interface 0, class 0) [ 2.080789] em28xx: DVB interface 0 found: isoc [ 2.080829] em28xx: chip ID is em28178 [ 4.061413] em28178 #0: EEPROM ID = 26 00 01 00, EEPROM hash = 0x7f10fc04 [ 4.061414] em28178 #0: EEPROM info: [ 4.061414] em28178 #0: microcode start address = 0x0004, boot configuration = 0x01 [ 4.067947] em28178 #0: AC97 audio (5 sample rates) [ 4.067948] em28178 #0: 500mA max power [ 4.067950] em28178 #0: Table at offset 0x27, strings=0x146a, 0x1888, 0x0a7e [ 4.067987] em28178 #0: Identified as PCTV tripleStick (292e) (card=94) [ 4.067988] em28178 #0: dvb set to isoc mode. [ 4.073317] em28178 #0: Binding DVB extension [ 4.085197] DVB: registering new adapter (em28178 #0) [ 4.085294] em28178 #0: DVB extension successfully initialized [ 4.085294] em28xx: Registered (Em28xx dvb Extension) extension [ 4.087769] em28178 #0: Registering input extension [ 4.117049] input: em28xx IR (em28178 #0) as /devices/pci0000:00/0000:00:14.0 Re comment 27 kernel-desktop-4.9.50-1.mga6-1-1.mga6 $ rpm -ql kernel-desktop-4.9.50-1.mga6-1-1.mga6 | grep em28xx /lib/modules/4.9.50-desktop-1.mga6/kernel/drivers/media/usb/em28xx /lib/modules/4.9.50-desktop-1.mga6/kernel/drivers/media/usb/em28xx/em28xx-alsa.ko.xz /lib/modules/4.9.50-desktop-1.mga6/kernel/drivers/media/usb/em28xx/em28xx-dvb.ko.xz /lib/modules/4.9.50-desktop-1.mga6/kernel/drivers/media/usb/em28xx/em28xx-rc.ko.xz /lib/modules/4.9.50-desktop-1.mga6/kernel/drivers/media/usb/em28xx/em28xx-v4l.ko.xz /lib/modules/4.9.50-desktop-1.mga6/kernel/drivers/media/usb/em28xx/em28xx.ko.xz I don't know if the following is relevant to this bug, or whether I should file a new one. Running Mageia 6 Official with the current packages for Kodi and tvheadend, generally, they work together well. Kodi is inclined to crash. In the log I see: ERROR: Could not find suitable input format: x-directory/normal At https://forum.odroid.com/viewtopic.php?f=144&t=26591, this was solved by upgrading tvheadend to version 4.2.1 CC:
(none) =>
laidlaws Created attachment 10338 [details]
dmesg after plugging in the PCTV 292e : 2018-08-29
Created attachment 10339 [details]
Attempt to build channel scan file for PCtv 292e
Time to reopen the bug on this affair. More and more people are complaining that Si2168 devices are not working in Mageia. There is a website which states that these are supported for kernels later than 4.17. Cauldron now stands at 4.18 so it seemed worth a test. $ uname -r 4.18.5-desktop-1.mga7 $ rpm -ql kernel-desktop-4.18.5-1.mga7-1-1.mga7 | grep em28xx /lib/modules/4.18.5-desktop-1.mga7/kernel/drivers/media/usb/em28xx /lib/modules/4.18.5-desktop-1.mga7/kernel/drivers/media/usb/em28xx/em28xx-alsa.ko.xz /lib/modules/4.18.5-desktop-1.mga7/kernel/drivers/media/usb/em28xx/em28xx-dvb.ko.xz /lib/modules/4.18.5-desktop-1.mga7/kernel/drivers/media/usb/em28xx/em28xx-rc.ko.xz /lib/modules/4.18.5-desktop-1.mga7/kernel/drivers/media/usb/em28xx/em28xx-v4l.ko.xz /lib/modules/4.18.5-desktop-1.mga7/kernel/drivers/media/usb/em28xx/em28xx.ko.xz The first hurdle remains - it seems that none of the relevant Silicon Labs firmware is packaged for Mageia so users have to find it and install it themselves. These are the files downloaded from https://github.com/OpenELEC/dvb-firmware/blob/master/firmware/ dvb-demod-si2168-02.fw dvb-demod-si2168-a20-01.fw dvb-demod-si2168-a30-01.fw dvb-tuner-si2157-a30-01.fw dvb-tuner-si2158-a20-01.fw dvb-demod-si2168-b40-01.fw $ lsusb Bus 002 Device 004: ID 2013:025f PCTV Systems See attachment re 292e. Scanned for channels using w_scan - see attachment. This failed. Borrowed the 290e from another workstation and tried it out here on mga7 to confirm that it would work in another environment. w_scan took about 18 minutes to scan the full spectrum and ran the tuning section without any problem and returned 130 channels. $ vlc channels.xspf showed the whole lot in the playlist and worked perfectly. So that is the old PCtv Nanostick 290e which uses the Sony CXD2820R chipset. # lsmod | grep em28 em28xx_rc 20480 0 rc_core 53248 3 em28xx_rc,rc_pinnacle_pctv_hd em28xx_dvb 36864 0 dvb_core 139264 2 cxd2820r,em28xx_dvb em28xx 98304 2 em28xx_rc,em28xx_dvb tveeprom 24576 1 em28xx v4l2_common 16384 1 em28xx media 45056 2 videodev,em28xx usbcore 294912 9 xhci_hcd,ehci_pci,usbhid,em28xx_dvb,usb_storage,ehci_hcd,xhci_pci,uas,em28xx kaffeine did not even get started with the 292e. Unsubscribing. My Comment 29 clearly has nothing to do with driver firmware, and I don't run Si devices. I have two dongles: RTL2832U Driver is in kernel. AF9015 No longer seems to be in use. It causes Windows to crash. Linux driver is available from LinuxTV site, and works well. Kaffeine is no measuring-stick. It won't recognize a dongle, even if dmesg shows it up. This seems to have something to do with the choice of USB port. If I plug the dongle into a different slot, both Kaffeine and VLC come good.
Doug Laidlaw
2018-08-30 03:59:19 CEST
CC:
laidlaws =>
(none) The following link demonstrates that the 292e was working in 2015 in Arch Linux. https://www.linuxquestions.org/questions/linux-hardware-18/pctv-triplestick-model-292e-%7C-no-channels-found-in-linux-but-in-windows-4175553960/ Three years later and it still does not work in Mageia. ?? The em28xx driver registers the device but w_scan seems unable to run the tuner after the demodulator section. Many transponders are found but at the tuning stage w_scan complains that it cannot find any. If I use scan it fails at the initial tuning stage. Some progress. While experimenting, although the scanning returned no useful data I discovered that both the newer devices work to some extent if given the channel scan returned from the tuning runs of the working device, the PCtv 290e. Both vlc and me-tv showed SD channels but could not manage the HD channels. During the attempts to tune the Hauppage WinTV and PCtv 292e devices the DVB-T2 capability of both was acknowledged. As far as I can tell, the two later devices do not see a strong enough signal to tune to on free-to-air. The good tuner comes back with values 2000dB stronger in some cases. No explanation for that yet. Since then there has been some success using w_scan data for a WinTV HD tuner, which also contains Silicon Labs chips. The firmware for several Silcon Labs devices is still missing from mga6 and Cauldron even though there is supposed to have been kernel support for these from way back in the 4.13 kernels. Users must download the appropriate packages and install them to /lib/firmware themselves. Once that is done there is a good chance that vlc can use the tuning data and that kaffeine can initiate a tuning scan. The implication of that, I think, is that the em28xx driver is up to date - it is only the firmware that is missing. |