<!-- Some styling for better description lists --><style type='text/css'>dt { font-weight: bold;float: left;display:inline;margin-right: 1em} dd { display:block; margin-left: 2em}</style>
dongs: o shit
<br> https://en.wikipedia.org/wiki/Socionext
<br> thats waht they got renamed to
***: [LOGGER] has quit IRC (Ping timeout: 240 seconds)
<br> kaspter has quit IRC (Ping timeout: 264 seconds)
<br> Sashmo has quit IRC (Ping timeout: 260 seconds)
<br> croppa_ has quit IRC ()
<br> sgtpepper has quit IRC (Ping timeout: 244 seconds)
<br> abruanese has quit IRC (Ping timeout: 252 seconds)
lja: <u>crope</u>: mchehab: around?
<br> any plans to support more than one tuner for a single frontend?
<br> done a driver for the si21832 demod (http://www.silabs.com/Support%20Documents/TechnicalDocs/Si21832-B60-short.pdf)
<br> but had to split the frontend in 2 because of the 2 tuners
<br> (total 2 adapter, 2 frontend each)
mchehab: <u>lja</u>: right now, satellite should be on a different frontend, because frequencies are in kHz
<br> I remember we've discussed this issue in the past, but didn't have time to work on a complete patch
<br> (I sent a RFC one - half baked)
lja: it's a single demod
mchehab: btw, we just merged support for an universal frontend with looks very similar from that one
lja: what's the name?
mchehab: from NetUp
lja: oik
mchehab: it is a sony frontend
<br> <u>Subject</u>: [GIT PULL] NetUP Universal DVB (revision 1.4)
<br> tuner HELENE (CXD2858ER) and new Sony demodulator ARTEMIS (CXD2854ER).
lja: how many tuners does it use?
mchehab: it's a very nice card
<br> 2 tuners
<br> 2 CI slots
lja: the card I worked is the tbs6522
<br> it's working good but I was hoping to get everything on the same frontend
<br> driver is very similar to s12168
<br> https://github.com/ljalves/linux_media/blob/latest/drivers/media/dvb-frontends/si2183.c
mchehab: well, feel free to work on a patch fixing it
<br> btw, it is really cool to see the configuration pages on Kaffeine for the NetUp device
lja: and it was tricky because the tuner i2c gate for the 2nd demod is on the 1st demod
<br> (both demods share the same i2c gate)
***: JackWinter has quit IRC (Ping timeout: 250 seconds)
lja: I will send a few patches to support the new tbs cards, all based on an fpga chip
mchehab: btw, using the Kaffeine's developer's hat, I guess having satellite on a separate demod actually is better
<br> DVB-S/S2 require several additional stuff, specially on setups with rotors
lja: I've used the set_property to set frequencies: https://github.com/ljalves/linux_media/blob/latest/drivers/media/dvb-frontends/si2183.c#L834
mchehab: (but, using my kernel developer's hat, just one frontend makes more sense)
lja: on a previous multi-standard demod I've put everything on the same frontend, but it had 1 tuner only
<br> the problem here is 2 tuner working for 1 demod
mchehab: not sure if that would work
<br> because userspace apps may do the wrong thing
lja: well, it works...
mchehab: as they don't expect the parameters to change
lja: (only tested tvheadend)
<br> and it did the right thing
mchehab: I guess Kaffeine 2.x won't do the right thing...
<br> but DVB-S is currently broken
-: mchehab is fixing it ATM
mchehab: I could fix it there too, but I would need either a card to test or a patch for the NetUp card to expose its frontends as just one
lja: how would you attach 2 tuners to a single frontend?
<br> dvb_frontend should have a list of tuner_ops
mchehab: netup has actually 2 adapters
<br> each adapter with 1 tuner, and 1 demod... with is exposed as 2, due to the Hz x kHz issue
lja: wait
<br> 1 tuner for 1 demod, right?
mchehab: yes
lja: ok, the si21683 has 4 tuners
<br> 2 tuners for 1 demod
mchehab: (afaikt)
lja: total 2 demods - 4 tuners
mchehab: ah, ok
<br> so, satellite is on a separate tuner then, I guess?
lja: yes
<br> http://www.silabs.com/Support%20Documents/TechnicalDocs/Si21832-B60-short.pdf
<br> check first picture
mchehab: ah, I see
lja: to properly support that, dvb_frontend would need a list of (dvb_tuner_ops + tuner_priv)
mchehab: on Netup, it is just one tuner, that can be ISDB-S/DVB-S/S2 + DVB/C/S/S2 + ISDB-T
<br> yeah, your setup is more complex
lja: I've split it in 2 adapters, each with 2 frontends
mchehab: you'll likely need to use 2 frontends anyway, because, on DVB, a frontend is demod + tuner
<br> well, I guess it is possible to map it also as one frontend with 2 tuners too
<br> with the kind of hack you did
<br> (in any case, it will be a hack)
<br> the only way to actually represent its internal connections is via the media controller
lja: yes
mchehab: nice device!
lja: well... it's working for dvb-t/c/s (tested)
mchehab: nice
<br> did you test with dvb-c2?
-: mchehab wanders if we have enough stuff for C2 at the Kernel
lja: no... no access to a signal
<br> what is the delsys?
<br> btw
mchehab: we didn't add yet
<br> no signal :)
<br> I have a generator that supports C2
<br> I need some signal and some board to test it
lja: i need a signal :)
mchehab: <u>lja</u>: with regards to Kaffeine setup, see the screenshots at: https://mchehab.fedorapeople.org/kaffeine_screenshots/tmp/
<br> a device with both DVB-S/S2 and a terrestrial or cable standard would require too many horizontal lines that won't likely fit on wide screen monitors
<br> btw, that's the screen Kaffeine 2.x produces with the NetUp device
lja: you mean the UI size
<br> for the setup
mchehab: yes
<br> fixable, but painful
lja: yes. I mostly use tvheadend and it kind solve it by making a tree and a setup panel for the selected device
<br> this card shows up like this: http://xsys.dyndns.org/pub/tvh_si2183.png
<br> (the tree)
mchehab: I see
<br> on Kaffeine, each device is a tab
<br> by "device", you should actually read frontend
lja: yes
mchehab: all parameters for that device are set at the window...
<br> well, for DVB-S, there is a popup window for LNBf settings
<br> but a device like yours will very likely produce a crap config setup
lja: anyway, its two different things (a dvb backend vs a "media player")
mchehab: yes
<br> and kaffeine is meant to be as simple as possible for the end user
lja: btw, should I send the pcie card patch as 1 file only?
<br> it's the full driver
<br> all inside here: https://github.com/ljalves/linux_media/tree/latest/drivers/media/pci/tbsecp3
mchehab: you should send the tuners/demods as separate drivers
<br> on separate patches
<br> the PCI device driver can be on a single patch
lja: yes, but as a start will only send the pcie card driver (there is 1 card that has already the tuners/demods in kernel)
<br> ok
<br> the tbs6502 uses the si2168+si2157
<br> so I can add that card as a start
<br> then I'll send the missing demod+tuners for other cards
mchehab: Ok.
<br> please send the demods/tuners before adding the cards that require those other tuners/demods
lja: yes
<br> got it
mchehab: good
***: ao2 has quit IRC (Quit: Leaving)
dongs: lja, what documentation is si2183.c based on
<br> anyway looks clean enough, bookmarking
<br> silabs vendor provided driver is fucking AWFUL
<br> the guy who wrote the 'silabs generic driver" they provide actually boasts this on his linked-in page
<br> honestly its a fucking embarassement
***: brfransen has quit IRC (Ping timeout: 272 seconds)
SupDudes: anyone here?
dongs: nope
SupDudes: it seems like LinuxTV is dead
<br> do people still use channels.conf for configuring TV?
<br> <u>dongs</u>: are you a Linux kernel developer?
dongs: <u>SupDudes</u>: i was until they made it so difficult to actually develop anything by changing APIs every 2 weeks
SupDudes: What did you develop?
dongs: lunixdvb drivers
<br> obviosuly
SupDudes: Which one(s)?
dongs: for hardware we make
<br> as a free non-support option for customerrs
SupDudes: okay. What hardware do you make?
dongs: but since UAPI happened and since they keep changing shit, and over 9000 lunix distros and differnet kernel versions
<br> we said fuckit
<br> since its impossible to support
SupDudes: What's UAPI?
dongs: when they made it impossible to compile kernel drivers out of tree
<br> by removeing all the useful includes
SupDudes: Oh
dongs: so in order for customers to compile our shit they'd need to download unknown amount of lunix kernel sores and match it with wahtever kernel they're running etc
SupDudes: Was that because of GPL?
<br> Does anyone even use Linux for DVB?
dongs: no
SupDudes: If not GPL, then why?
dongs: no, the "no" was for "nobody uses lunix for dvb"
SupDudes: So why are you here?
dongs: im not sur why uapi happened
<br> https://lwn.net/Articles/507794/
SupDudes: this? https://lwn.net/Articles/507794/
dongs: ya
<br> While such work does not improve the lives of users, it certainly improves the lives of developers, by easing later work that does add new features.
<br> they'er completely right.
<br> it just fucks users over 100%
<br> basically, unless you're a kernel developer, fuck off and dont build anything.
SupDudes: How does it un-improve lives of users?
dongs: because out of tree driver building is impossible
SupDudes: How does UAPI prevent that?
dongs: the headers don't exist
<br> distros ship post-uapi split stuff
SupDudes: You seem quite knowledgeable about Linux Kernel Development
dongs: which doesnt include anything needed to build drivers out of tree
SupDudes: I figured, by your description, that UAPI was to stop 3rd party non GPL drivers
<br> but that article says otherwise
dongs: so now if we were to support lunixtv, we'd need to ship entire 'media tree' or wahtever the fuck
<br> and somehow make users patch wahtever current kernel they have
<br> but they keep moving apis/files/headers/shit around
<br> it would be a fulltime job
<br> to supprot that shti
SupDudes: does this affect only DVB or drivers at large?
dongs: dunno, only looked at it in context of DVB
SupDudes: it seems you can do out-of-tree drivers still
<br> maybe DVB doesn't expose enough?
dongs: most likely.
<br> and if it hasnt been fixed in 4 year,s it isnt going to be. but its ok, decision to stop supporting this shit was super easy
SupDudes: What's the name of your driver?
dongs: as you said yourself
<br> < SupDudes> Does anyone even use Linux for DVB?
<br> <u>SupDudes</u>: doesnt matter, it was not included in kernel (and wasnt offered to be anyway), just distributed to users of our hardware who asked.
<br> under GPL too, not that it matters
<br> there was nothignto hide in it
SupDudes: okay
<br> You sound quite bitter
dongs: not at all
SupDudes: If you say so
dongs: Lunix sucks, end of story
<br> but thats been known for years.
SupDudes: Is that a troll
dongs: naw
SupDudes: <u>dongs</u>: What is your favourite Linux boxen?
dongs: I dont run lunix
<br> i use win10
<br> exclusively, on all machiens
<br> I have some VMs with lunix
<br> thats where it belongs
SupDudes: You enjoy spyware I guess
<br> I tried Windows 10 once, it was sending way too much telemetry bullshit
dongs: ? I havent had any spyware on any computers i own and I run windows since 3.1
SupDudes: Windows 10 is the spyware
dongs: about as much spyware as 'lunix experience project' that collects your information.
<br> or ubuntu's amazon tie-in
<br> etc.
<br> sorry dude, try harder trolling.
SupDudes: I'm not trolling
<br> http://betanews.com/2016/02/06/windows-10-phones-home-a-lot-even-with-all-reporting-and-telemetry-disabled/
dongs: who cares
<br> android/google collect way more info on what you do
<br> do you have something to hide?
<br> I dont
SupDudes: Windows 8.1 works fine, so I have no reason to upgrade
<br> Luckily for me, I'm neither paranoid or a troll, unlike some people I knew
<br> What phone do you use? iPhone? Windows Mobile?
dongs: lol.
<br> > not running windows 10
<br> > no paranoid
<br> k
<br> and a shitty troll at that
SupDudes: nice troll + blog
dongs: how's REcon? shitty troll
SupDudes: Well you fell for it
dongs: you wish
SupDudes: i didn't know UAPI was until you mentioned it
<br> only an expert Lunix Kernel Developer would know this
dongs: riight
SupDudes: the fact you're still here proves it
dongs: the fact that youre still a shitty troll remains
SupDudes: a non troll will always be shitty at trolling
<br> unlike an expert troll master as yourself
dongs: lunch time
<br> maybe you should hack some more gibsons.
SupDudes: enjoy fat american food
dongs: enjoy fat canadian food
SupDudes: i don't do anything illegal, unlike you
<br> you guys are retarded because of this https://git.linuxtv.org//dtv-scan-tables.git/
<br> way too much info that is totally useless and unnecessary
<br> is this for real: https://git.linuxtv.org/dtv-scan-tables.git/tree/channels-conf/dvb-t?id=4f192bdfa3cca852d0fb9b9a3b2df5c6e2cf2fe7
***: rjkm_ is now known as rjkm
<br> rjkm has quit IRC (Quit: leaving)
xdarklight: <u>dongs</u>: evening - any updates from pana yet?
<br> it would also be nice if someone could confirm that I'm right about BER (https://linuxtv.org/wiki/index.php/Panasonic_MN88472#BER_.28Bit_Error_Rate.29): I'm storing "error" in post_bit_error and "sum" in post_bit_count (or would this go into pre_bit_* instead?)
<br> also I'm taking the PER values and storing them in block_error=(error * 100) and block_count=per_len
***: iive has quit IRC (Quit: They came for me...)