[mythtv-users] Fix found for IR receiver with mceusb in newer kernels

Sun Jul 5 13:51:13 UTC 2015

References: <CA+SOsrQcDO5UzArzDOHhH0NEu1VADpRsHoTOnhqTKjJy9=74ng at mail.gmail.com> <682836388.20150621111006 at gmail.com> <CA+SOsrRazBDeuJH8_5B_f=ERRk1gYs=VwV5C0+xB85ubXVzz1w at mail.gmail.com> <CA+SOsrRSRB+sGGa4UvCK7-rA-uYuTCz5oZ_4vD2Z3c5w6XwG3w at mail.gmail.com> <i4ptoada182d4pgdsqohorlppib0tpifp7 at 4ax.com> <CA+SOsrRZ8gDrB7uVg_OdfFwjoJHEJw+KyQJTOfKPxwMCBiDk5g at mail.gmail.com> <67ttoapr74r7f2eofb6gvg2sp34b5snsle at 4ax.com> <CA+SOsrR8GAjjW6OAgC7CP3kC8Sz4Jg41VMdBA4MfSZR1e2pmpg at mail.gmail.com> <CAOQWjw2rygovr8Qcz3oax2CK=BPWqp=+1PXrbazFdGOQreck=g at mail.gmail.com> <CA+SOsrRWM5u7LhUwnCSf45DQ1Z=zYPcOGQG+zi5iS6v_sX-z9w at mail.gmail.com> <CAOQWjw0i2AbNMUr1trgzH=nRuhmvAovYofeRJ2ZOAL2BUwwNfw at mail.gmail.com> <CA+SOsrSyKri43+q4T=26LaTdRpi0G89BYUgFS2umzf=2pTocaQ at mail.gmail.com> <55903072.6070609 at gmail.com> <CA+SOsrS=1VMTaMjzcXy3VKu_O_gyDPx=Npg3=haOOWmGxL9whA at mail.gmail.com> <CA+SOsrR74CjtMynz4nCTTV01T43WHrY1WRHDsig2kUJxm0=3GA at mail.gmail.com> <55917230.5020006 at gmail.com> <20150629170452.49bdb339 at cs3.al.itld> <CA+
 SOsrRxUDuEgQaWo7-U9Z43JCjs+R0EAczNThNWb0UiGikbBA at mail.gmail.com> <CAAMCDefWm5nSWtq97nw0Xr+MAZsKvw5=6ZvwG17MFwx6=Q7fhg at mail.gmail.com> <CA+SOsrQ2dPeQE8MKgSj2FUuMb7wSGx3h12t-DOqmbXvgPEFYnA at mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Return-Path: hikavdh at gmail.com
X-OriginalArrivalTime: 05 Jul 2015 13:51:27.0954 (UTC) FILETIME=[B0AEEB20:01D0B729]

Hoi Jerry,

Friday, July 3, 2015, 4:06:44 PM, you wrote:

>>
>>
>> On Mon, Jun 29, 2015 at 12:27 PM, Jerry Rubinow <jerrymr at gmail.com> wrote:
>> > On Mon, Jun 29, 2015 at 1:04 PM, Mike Thomas <mt3 at pfw.demon.co.uk>
>> wrote:
>> >> On Mon, 29 Jun 2015 18:28:32 +0200
>> >> Alec Leamas <leamas.alec at gmail.com> wrote:
>> >>> On 29/06/15 17:31, Jerry Rubinow wrote:
>> >>> > On Sun, Jun 28, 2015 at 2:50 PM, Jerry Rubinow <jerrymr at gmail.com>
>> >>> > wrote:
>> >>> >>
>> >>> >> On Jun 28, 2015 13:36, "Alec Leamas" <leamas.alec at gmail.com> wrote:
>> >>> >>>
>> >>> >>> On 28/06/15 15:54, Jerry Rubinow wrote:
>> >>> >>>
>> >>> >>>>
>> >>> >>>> Nick, it looks like 3.16, what I've been running, has that patch
>> >>> >>>> integrated into the kernel driver.  I tried a clean Ubuntu 15.04
>> >>> >>>> install and the problem is there as well.  Unfortunately the
>> >>> >>>> bios doesn't let me disable xHCI.    I think it may be time to
>> >>> >>>> visit newegg for a new ir receiver.
>> >>> >>>
>> >>> >>>
>> >>> >>> TBH, I havn't read the complete thread. That said, I understand
>> >>> >>> your situation as if the kernel decoding route is at a dead end
>> >>> >>> for you (?).
>> >>> >>>
>> >>> >>> Note, the alternative is to let lirc decode the raw timing data.
>> >>> >>> I know this has been mentioned before, but perhaps you missed
>> >>> >>> this opportunity?
>> >>> >>>
>> >>> >>> Obviously, the existing config files doesn't work for you. One
>> >>> >>> option would be to use irrecord to create a new config file after
>> >>> >>> setting the protocol to 'lirc' and re-checking that mode2 still
>> >>> >>> works as expected. Might work, dunno, but given that mode2 is OK
>> >>> >>> for you I think you have a fair chance.
>> >>> >>>
>> >>> >>> Cheers!
>> >>> >>>
>> >>> >>> --alec
>> >>> >>
>> >>> >> Alec, that's probably worth trying, I'll try irrecord.  However, my
>> >>> >> expectation is that the bits coming out of the driver will be
>> >>> >> faulty due to something in the USB 3/xHCI implementation.
>> >>> >>
>> >>> >> I'll try that tonight.
>> >>> >>
>> >>> >> -Jerry
>> >>> >
>> >>> > Nope, I tried the config generated by irrecord and maybe one out of
>> >>> > 10 to 20 presses would generate something it could recognize in
>> >>> > irw, but usually not the key I pressed.
>> >>>
>> >>> But you did run irrecord successfully enough to create a config
>> >>> file?! This does indeed looks strange.
>> >>>
>> >>> Just a double check: you don't have physical (i. e., IR) disturbances
>> >>> around? Low energy/fluorescent lamps or so?
>> >>
>> >> Dear OP,
>> >>
>> >> You might like to look into the archives for my earlier comments
>> >> regarding IR disturbances. Some monitors and TV sets produce enough IR
>> >> on their own that you might as well give up using an infra red remote
>> >> control. All my monitors and TVs fall into that category. It just so
>> >> happens that the TV's own remote control protocol is transmitted in the
>> >> blanks between the IR flashes given off by the back light. I couldn't
>> >> find another remote control which would transmit in these bursts. I
>> >> even bought a Sharp IR receiver chip to go with my Sharp TV and that
>> >> didn't work. It took me weeks of trials before I eventually gave up.
>> >>
>> >> I settled on a Bluetooth remote control which I ordered from Maplin.
>> >> This has been flawless. They are coded so it is possible to have
>> >> several in one building without them conflicting. You just plug it in
>> >> and it comes up as a HID input device (a keyboard and mouse). No
>> >> drivers, no nothing. I have the key mappings if you want them.
>> >>
>> >> Cheers,
>> >>
>> >> Mike.
>> >>
>> >
>> > Mike, I've been running this setup (tv, a/v receiver, etc), in this
>> > room for years.  The only difference between totally working and
>> > totally not working is swapping in a different frontend computer with
>> > a newer version of mythbuntu.  No other moved/added/removed devices.
>>
>> On Mon, Jun 29, 2015 at 1:50 PM Roger Heflin <rogerheflin at gmail.com>
> wrote:

>> I had to modify the a kernel driver for xhci to get mine to work.
>> 3.16.6 did not work without the change.
>>
>> I have not been able to update kernels last time I tried as the newer
>> kernels don't yet have the modification included last time I checked.
>>
>> I think I posted the fix someplace, if not yell and I can probably
>> figure out the change that I made.  Without the change lirc did "see"
>> keypresses but none of it made any sense (garbled stuff).
>>
>>
> I can verify that making Roger's change to the usb core fixes the issue.

> To summarize, an older USB IR receiver will not work when plugged into a
> USB2 or USB3 port that is using xHCI.  Bits will be dropped and scancodes
> will not be received properly.  This is below the level of lirc or
> ir-keytables.  I know it happens in 14.04 and later versions.

> Patching config.c in drivers/usb/core fixes the issue (changing a line that
> says n = 32 to n = 10) and compiling a new kernel.

> Now my only problem is that after recompiling I no longer have any files in
> /sys/class/rc.  So I can use lirc, but not ir-keytable.  A wrong kernel
> config option?  Dunno.

> -Jerry

If you recompile your kernel, you also have to recompile any external
modules. Like from lirc or nvidia.

Tot mails,
  Hika                            mailto:hikavdh at gmail.com

"Zonder hoop kun je niet leven
Zonder leven is er geen hoop
Het eeuwige dilemma
Zeker als je hoop moet vernietigen om te kunnen overleven!"

De lerende Mens