Initial release of the satip-axe firmware is available (Minisatip for Digibit R1,DSI 400, IDL400s)

Zitat von fnu

Das ist Quatsch, niemand profitiert davon wenn der Maintainer hier nicht an Board ist und mitdiskutieren kann, vorallem spricht er offensichtlich kein deutsch ...

...ansichtssache. denn den Fortschritt der firmware, wenn dieser z.b. "ausserhalb" des boards von statten geht, der kõnnte dann...gerade für die rein deutschsprachigen Interessenten, eben hier AUCH im board mitverfolgt werden...
es geht ja nicht nur um den maintainer in dem thread....sondern um die fw in verbindung mit vdr und da sind wir hier doch richtig, oder


greets arel
Mod.: Was soll das mit den Fullquotes? Bitte dringend lesen: Richtig zitieren und URLs posten ...

Zitat von arel

...ansichtssache. denn den Fortschritt der firmware,

Ganz sicher nicht, sondern Erfahrung aus vielen Jahren. Was ihr dabei macht ist rumraten ohne das was dabei rauskommt.

Ihr könnt, sollt, dürft gerne hier diskutieren, aber dann bitte den Maintainer involvieren und das ganze am Besten in einer Sprache die er auch spricht.

Regards
fnu

arel

Ist schon gut so! Die es nicht interessiert müssen ja nicht mitlesen.

Grüße

Hat ja wenig bis nichts mit der Stock FW der Inverto OEM Boxen zu tun, daher in einen spezifischen Thread ausgelagert.

Zitat von zwick der agent

Die es nicht interessiert müssen ja nicht mitlesen.

Als Moderator werde ich jetzt jeden Beitrag mitlesen, ganz besonders Deine.

Regards
fnu

Zitat

Als Moderator werde ich jetzt jeden Beitrag mitlesen, ganz besonders Deine.

Kannst Du gerne machen, ich habe nichts zu verbergen

Grüße

satip-axe-201504082030-4 : Wed Apr 8 2015
- upgraded busybox to v1.23.2
- added ntpd configuration (network time daemon)
- changed dropbear default PATH to "/sbin:/usr/sbin:/bin:/usr/bin:/usr/local/bin"
- added /dev/sd* devices hotplug (mounted to /media tree using labels)
- added kernel modules for USB serial converters (most Phoenix readers)
- added /dev/ttyUSB* devices hotplug (persistent names like /dev/ttyUSB2-1:1.0)
- added o*c*m r10619
- allow customization of hostname (/etc/sysconfig/config)
- minisatip fixes (signal levels, cleanups)

satip-axe-201504061852-3 : Mon Apr 6 2015
- minisatip fixes (unicable-jess, corrected input selection - coax inputs)

satip-axe-201504061852-2 : Mon Apr 6 2015
- minisatip fixes (unicable, multiple tuners, h/v and lo/hi setup)
- load axe modules in rcSBB

satip-axe-201503311856-1 : Tue Mar 31 2015
- initial firmware release

Ich kann ja mal beschreiben, wie ich minisatip auf Digibit R1 zum laufen gebracht habe.

Ausgangssituation:
- Digibit R1 mit FW 1.16.0.120
- Raspi mit MLD 4.01
- alles über einen 100 MBit-Switch verbunden
- git clone https://github.com/perexg/satip-axe (gestern gemacht)
- Raspi läuft mit Digibit Stock-FW sehr gut

erster Versuch:
- die Versionen aus satip-axe/dist/ vom 31.3. und 6.4. gemäss README per USB ausprobiert
Ergebnis:
- ssh ok
- minisatip läuft
- kein Server im satip-Plugin sichtbar
- kein Stream wird abgespielt

zweiter Versuch:
- Firmware 1.17.0.120 aus satip-axe/firmware/ per Stock-Webfrontend eingespielt
- Version aus satip-axe/dist/ vom 6.4. auf USB
Ergebnis:
- ssh ok
- minisatip läuft
- erstmal kein Server im satip-Plugin sichtbar
- aber Stream läuft trotzdem
- Server nach einiger Zeit (automatischer Restart von minisatip) im satip-Plugin sichtbar
- Streams auch mit Zapping stabil (teste aber erst seit einer Stunde)

Bemerkenswert:
- Die angezeigten STR/SNR Werte sind mit satip-axe vom 6.4. wesentlich besser als mit Stock-Firmware

Gruss

plego

Thank you for the report. If I break forum rules with the english reactions, just give me a note and I'll stop. But using the google translator to write German sentences looks to me worse than do nothing.

The signal strength should be corrected in today's firmware #4. But my code uses only a basic linear conversion from values given by the inverto driver, so there is a room for improvements.

The problem that SAT>IP clients do not see the minisatip server might be that the multicasts are sent before the network is initialized, so clients probably react on the second announce. I created issue https://github.com/perexg/satip-axe/issues/3 for this problem. I'll fix it in the next build.

Also, if you have any idea what to add to this firmware, just create a new issue or give a note here.

NFS/Samba server: I don't like this functionality so much, but I can add it if there is a big demand. The CPU is not powerful, my tests show that two full muxes (transponders) can trigger 100% CPU usage with minisatip.

perex
You are doing a great job!
Shame on me, I should write in english.

I tried the basic functions of todays release and they are looking good.
I had an issue with Mondays release: After approx. 30 minutes of doing nothing (nothing in terms of just watching tv, no zapping), I wasn`t able to switch to another channel. I was stucked to the actual tuned channel and vdr did not throw the "channel not available". kill -9 to minisatip fixed it.
I will test this for todays release and open an issue, if it is still there.

Regards

plego

Zitat von perex

If I break forum rules with the english reactions, just give me a note and I'll stop.

Please go ahead, I strongly hope all interessted people do answer in common english, that you don't need to struggle with any translation tool.

Despite the fact we're vdr-portal.de, we're not fixed to german language ...

Regards
fnu

Update:

satip-axe-201504082030-4 running for 8 hours now and no stucked channel issue, I did some hardcore zapping after breakfast.

Just got a "[pll_wait_for_lock] pll locking timeout" in dmesg for the very first switch this morning. I will try to reproduce that.

Regards

plego

satip-axe-201504092035-5 - Thu Apr 9 2015
- fixed minisatip & dhcp issue - minisatip is started with good IP
- changed load order of AXE modules and AXE hw initialization

BTW: It seems that hw has a hardware input selector which is separated from the voltage/diseqc . I can select any coax input to any tuner, so I plan to add quattro possibility like the original firmware (fast tuning without extra diseqc when only one position is used) and even the master/slave configuration which will allow to share same diseqc position/polarization/lo-hi settings by multiple tuners and limited coax inputs (TVHeadend can be configured to use this restricted hardware).

perex :
You are doing an good Job !!!

I have testet the whole day Build 4.
My Tests with Analysing records with TS-Doctor Programm shows me, that the recorded Streams was without errors,

This Circumcances is not given with the original FW.
With original FW i have sometimes heavy artefacts.
There was gone by Build 2 to Build 4.

But now, with Build 5, i have a lot of artefacts in my live-streams.

Excuse me for my bad english

Greetings,
Darkover

Thanks. I'm struggling with the kernel modules from inverto (tuners, demodulators, demuxers). It seems that the initialization fails badly in some cases. Could you try build 6? Also, you may try to power-off the box for 5-30 minutes. It helps sometimes.

satip-axe-201504101512-6 - Fri Apr 10 2015
- extended minisatip (quattro LNB, master/slave, unicable on other input)
- another shuffling in the AXE modules load procedure

Sorry, the Artefacts are still in the Streams with Build 6

I tried to power off the GSS-Box one Hour.
But the Artefacts dont gone.

All the Tests with the Builds 2-6 i look with "netstat -anus" in the Section "Udp" on my VDR-Machine.
Till Build4 i have no Errors in udp packet receive errors.

With Build 5 and 6 i have now 40000 packet receive errors within 15 Minutes.

Zitat von darkover

Sorry, the Artefacts are still in the Streams with Build 6
All the Tests with the Builds 2-6 i look with "netstat -anus" in the Section "Udp" on my VDR-Machine.
Till Build4 i have no Errors in udp packet receive errors.

It seems like another error than I'm talking about. You may try to increase kernel UDP buffer sizes - https://www.linux-tips.org/art…-receive-error-difference . The build 4+ has bigger buffers for demux read operations, so the larger chunks of data can be sent to the network.

EDIT: I'm using this /etc/sysctl.conf settings on my workstation:

net.core.rmem_max=2621440
net.core.wmem_max=2621440
net.ipv4.tcp_rmem=4096 65536 65536
net.ipv4.tcp_wmem=4096 65536 65536

perex:

I have changed the values in the sysctl.conf and the Artifacts was gone.
Now i can use Build 6 without Errors in the Streams.

I have played around with the values and now my setup in the sysctl.conf is :

net.ipv4.ip_forward = 0
net.ipv4.conf.default.rp_filter = 1
net.ipv4.conf.all.rp_filter = 1
net.core.rmem_max=33554432
net.core.wmem_max=33554432
net.core.netdev_max_backlog=2000
net.ipv4.tcp_rmem=4096 65536 65536
net.ipv4.tcp_wmem=4096 65536 65536

This works for me without errors. Many Thanks for this Hint !
Build 6 seems to be good, i have test it for the last 2 Hours.

Greetings,
Darkover

Hi !

can anybody tell me whatfor the o***m is inside ? is it possible to decode on the fly before streaming ?

ciao gerd

Hi,

Zitat von gerdh

can anybody tell me whatfor the o***m is inside ? is it possible to decode on the fly before streaming ?

no CPU is too slow for that, as I can see. You could decode *one* SD stream, but a HD stream is already over the boundary. I think perex included it, to have a server, running anyway, to connect the USB cardadaper to.

Best

Zitat von tehlers

Hi,

no CPU is too slow for that, as I can see. You could decode *one* SD stream, but a HD stream is already over the boundary. I think perex included it, to have a server, running anyway, to connect the USB cardadaper to.

Best

Alles anzeigen

Yes, the reason to have this app included is to have an independent card server only. The ST7018 CPU is really slow for any software descrambling. This chip has the hardware decoder, but I think that it's only for few services (for the standard usage like standard set-top-box) and all is closed source.