Beiträge von perex

Build 15 of satip-axe firmware was released: Link to build 15 . Latest minisatip (minisatip8), some diseqc fixes, multistream support (tested with tvheadend and enigma2), VDR should work with minisatip8 now, too.

Build 14 of satip-axe firmware was released:

satip-axe-201705251044-14 - Wed Jun 7 2017


    kernel modules from idl4k-1.25.0.157 (frontend + demux)
    updated minisatip5 (0.5.54)
    fixed minisatip issue - multiple frontend device open - #72
    added minisatip7 (0.7.4) / NOTE: Some options are changed!
        this is the PREFERRED Sat>IP server now!
        added -0 (--diseqc-multi) option (it's just a workaround)
    upgraded busybox to 1.26.2
    upgraded dropbear to 2016.74
    upgraded rpcbind to 0.2.3
    upgraded nfsutils to 1.3.4
    upgraded o*s*c*a*m to rev.11384

Download

Portisch : axe_fe.ko - this module has only I2C driver (compatible with Linux DVB-API) for the used demodulators and tuners. No STAPI dependency here (data transfers are handled in the demux kernel modules). The list of I2C hardware is here:

https://github.com/perexg/sati…master/debug/hardware.txt

I added to my kernel a hook to "mangle" the i2c messages on the fly - so it's possible to modify the baseband amplifier settings for in STV6210 chip for example and perhaps to fix "init" issues when we know what to modify.

Portisch: wget error should be fixed in my satip-axe git repo . There is busybox's httpd server in the satip-axe firmware also with ftpd and nfs for data exports from the /media tree (mounted USB drives).

motze: Adding USB DVB-T tuners might be possible, if the USB subsystem is ok in the used kernel (I have some reports about card-readers which does not work - the kernel hangs). The good thing is that the DVB drivers from Inverto can co-exist with standard linuxdvb drivers. Unfortunately, I fight with a free time now.

Darkstar: I cannot find firmware files for gss.mux . Do you have any ?

Unfortunately, the original firmware for web has a little header which I've not analyzed, so you cannot flash the .fw file through the standard firmware, but if you flash the satip-axe firmware-v9 - there is upgrade-fw script which can do next upgrades without the USB flash.

Also, note that satip-axe modifies the uboot config in this way (look to last line):

https://github.com/perexg/sati…atches/uboot-flash.script

The fw2-nand0 mtd partition is used (original fw uses fw1-nand0).

Unfortunately, I got a reply from Portisch (which is also the author of these libraries) that another "part" is missing in the inverto modules:

"Actualy there is no possibility to use the hardware descrambling.
The ST sttkd_ioctl module is missing.
This modul is controlling the secure co processor."

The kernel modules are there (stapi_core/stapi_ioctl). These modules offer stapi devices - including /dev/stapi/stpti5_ioctl for communication with the PTI block in CPU which also does descrambling. Then we have static libraries (without sources) for o*c*m for stapi v5 which can allow o*c*m to set the descrambling keys through /dev/stapi/stpti5_ioctl device to hardware (in o*c*m - the dvba*i module does this job). So now, it's just about to put everything together (minisatip / o*c*m with stapi v5).

http://www.streamboard.tv/wbb2…?postid=541107#post541107

EDIT: It seems that multiple PMTs are passed to one PTI device - it seems that hw is capable to descramble multiple services from one mux (transponder)..

Little comments (if I understand things correctly):

- the CPU can descramble six streams in hardware (functionality through stapi v5 - closed API)
- o***m support stapi up to v4 using a closed source static library (so there's a chance that v5 will be supported in future)
- the Inverto drivers use only 4 descramblers (PTI) - one per demodulator, so only one service can be descrambled from a mux (transponder)

But maybe, I'm wrong with something above and it's out of scope of my interest. But anyone other might dig to it...

The final build 9 is released https://github.com/perexg/satip-axe/tree/master/dist . Added /etc/sysconfig/localtime file handling and /etc/sysconfig/profile support.

I tried to resolve all reported issues in build 9 - https://github.com/perexg/satip-axe/tree/beta/dist . Please, report any issues before I move it to the standard master branch.

- added [url]http://boxip:8080/axe-status.txt[/url] status page (shell parseable)
    - fixed ntpd start
    - added crond start / configuration
    - added customd start (custom daemon/service/setup - script) from sysinit
    - added nano text editor
    - updated o***m to 10663
    - added update-fw script for the internal flash upgrade without an USB stick

Zitat von hummel99

Hi,

I'm runing your Firmware since 3 Days and its very stable compared to the Stock Firmware, many thanks for that.
One Question, is it possible to get the State of the Tuner LEDs from the shell ?

No, it's not possible. The LEDs are controlled using a special char device /dev/axe/fp-0 . I don't think that this information can be read back.

But it's easy to write the "tuner busy" information from minisatip to a file or maybe, the http protocol can be used to serve this information.

Zitat von tehlers

Maybe this is easier, than passing the parameters via modification of uboot parameters? That's at least how the original kernel was configured.

Yes, sure. I was lazy to look for this option. Added to my repo - commit 26f0dff40a7e9567994d4c06bc46aee4e072cbd0 .

Zitat von CvH

Maybe this was already answered (can't find anything), does the box supports WOL ?
In combination with the autoshutdown script we could save a lot energy.

ethtool says:

Supports Wake-on: ug
Wake-on: g

But the question is if the kernel supports suspend for STi7108 .

Zitat von tehlers

Hi perexg,

I have a problem compiling the kernel. I have compiled the kernel without errors and injected it my own initrd, which is basically the original Image with my changes. The uImage.gz resulting boots well until it wants to use the DVB-devices, then there is a kernel oops. Same Image with original Kernel boots and works well. I have no idea where the problem could be, since you compile and use your own kernel, too. Maybe you have an idea by looking into the dmesg:

Thanks

Tim

Alles anzeigen

Check the uboot parameters. https://github.com/perexg/sati…ster/patches/uboot.script

The parameter 'bigphysarea=20000' is required for the kernel - it seems that the original firmware has this parameter somewhat build-in.

I put more information about the I2C bus - to talk with LNB drivers, tuners and demulators to the https://github.com/perexg/satip-axe/tree/master/debug directory. Also, I added datasheets for the LNB driver and tuner there. I cannot find datasheet (with register description) for STV0900 demodulator - but we have code in the mainline kernel - stv0900_core.c etc.

I think that the axe_fe.ko kernel module (which does the whole setup of these chips) can be replaced with own code. The TS packet data flow which uses proprietary STAPI is completely separated from the other settings.

Zitat von Fux

perex:
Do you think there is some tuning possibility to reduce idle power consumption?

I ask this, cause the satip devices have a quite big idle consumption higher than 10W.

I saw in a thread that the switch from 1000M to 100M in the ethernet card can reduce power usage by 2W. If you have configurable network switch, could you force 100M on ethernet port? I can add ethtool to axe-firmware.

Also, the I2C chips might be probably powered down (not sure - have to read specs deeply and linux kernel drivers for them).

You may also try to unload the axe_fe module (which is the driver for tuners). Use

touch /tmp/nosatip
killall -9 minisatip
rmmod axe_fe

Could you measure wattage now ?