Skip to contentSkip to conversation

cx88 totally fried in 2.6.15-rcX -was- Re: HD3000 - no NTSC via tuner

DISCUSSION Michael Krufky 6 participants 45 messages Nov 29, 2005 – Dec 2, 2005 15 messages not in the archive Media (V4L/DVB)

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#4 lore raw
Quoted 84 lines · Gene Heskett
Quoted 61 lines · Michael Krufky
Quoted 31 lines · Perry Gilfillan
Quoted 21 lines
Quoted 18 lines · Don Koch

(Followup to posting on pcHDTV forum.)

I recentely purchased a pcHDTV HD3000 card. Initially I was having problems with NTSC reception in general (via tuner, composite or svideo input), part of which was a problem with the version of linux I was using. Currently running 2.6.15-rc2-g458af543.

Symptoms:
- ATSC OTA, svideo and composite work. NTSC tuner does not (tried both OTA and cable).
- Xawtv segfaults unless given a specific device (/dev/video0) and then doesn't show anything in any mode including composite and svideo. This may just be an xawtv problem.
- Using tvtime works for composite and svideo, but the tuner doesn't work in either broadcast or cable settings. Disabling signal detection shows the same "image" for any channel.

Quoted 3 lines
Quoted 3 lines

All the appropriate modules seem to be in place.

I can report some similar troubles. It has worked with vanilla kernel up to 2.6.13 on x86 up untill two weeks ago, when I moved the card to an amd64 system.

I'm currently running gentoo-sources-2.6.13 on amd64. I have not tried the newly merged tree with this kernel.

Quoted 9 lines

However with gentoo-sources-2.6.14-r2 I tried the newly merged tree, and do not get NTSC reception. I have not tried the drivers as released with 2.6.14 yet, so I'll follow up later on that.

I'm also having some issues with the VIA Unichrome drivers with 2.6.14, so I may not be able to get far enough to really see whats happening.

I know that this board uses a Thomson DDT 7610 tuner. I have a FusionHDTV3 Gold-T card that uses Thomson DDT 7611 tuner. I've seen a copy of the datasheet -- both tuners are the same.

Now, here's my story: My card was working fine for a good long time, but ever since about 2 months ago, the analog tuner no longer works. ONLY the ATSC digital tuner is working. The tuner does NOT include a tda9887, AFAIK, and I have also tested in windows. Regardless of whether I am using Linux or Windows, I can only view an ATSC digital stream, NTSC analog no longer works at all. Both used to work in both OS's.

AFAIK, this is a physical hardware problem, and has nothing to do with any bad coding. For the life of me, I cannot figure out what went wrong.

-Mike

Quoted 8 lines

Maybe Herman, or someone else, might be able to tell us what factors can burn out a hybrid tuner's analog capabilities, and leave digital tuning capabilities intact.

Maybe the Thomson DDT 761x tuners are faulty?

Can anybody enlighten us?

Well, based on the note about -rc2-git6 having the v4l stuff reverted, I just built it. Same song, same verse. The tuner acts as if the antenna has been disconnected and held a fraction of an inch away. The video is all digital noise (in VSB), and the audio has a fraction of a word now and then, buried in white noise. As my dish receiver has about a 65,000 u-volt output, I find that sort of signal loss almost unbelievable.

It still takes a cold reboot back to 2.6.14.3 or I think any earlier kernel (that worked ok) in order to restore the cards normal operation, I'm watching cnn on it right now after having done so.

Quoted 9 lines

So to me, there is little difference between git3 and git6. Neither works, failing in the same manner exactly, and matching the performance of 15-rc2 as issued. I didn't build -rc1, so I can't say exactly where the hose got cut.

I'd suggest just blowing the v4l directory in the -rc2 kernel away, and
replacing it with that from 2.6.14.3 just to get back to a working
sitiuation that you can then start from scratch on. But whatdoIknow? :)Gene Heskett <[email protected]>

EEK! This is not the problem that I'm having at all..... Then again, I have a different board.

Gene, telling someone to revert all the v4l changes in 2.6.15 doesn't help us to fix the actual problem at all, nor will it help them use their hardware.

Gene, I believe that I asked you to install the cvs modules against 2.6.14, and you told me that doing that works. The code in cvs contains all the patches that we have sent to 2.6.15 and THEN some. Can you please confirm that installing the cvs modules from v4l-dvb cvs (v4l and dvb have merged cvs repo's) against 2.6.14 is NOT broken?? This would rule out any possibility of the v4l changes in 2.6.15 being the cause of your problems.

It is WELL established that there are memory errors in 2.6.15 (although I thought they were all fixed -- guess not). I understand that you are using some MakeIt script to build your kernel -- have you tried the standard method?

[...a few minutes go by...]

OKAY, I concur -- When I did my testing for 2.6.15-rc2-git6, that was with my saa7134-based card, using nxt200x ... success!

However, when I tried using my FusionHDTV3 Gold-T (cx88 card using lgdt3302), neither analog nor digital works in 2.6.15-rc2-git6.... Under 2.6.13 (and 2.6.14, i think), however, digital DOES work. Analog still doesnt work, but i believe that my hardware is damaged.

SO, looks like we have a regression somewhere in the kernel that breaks the cx88 driver :-(

I don't even know where to begin.

...One idea... We also know that upstream changes created some compile warnings in tuner-core... Hans has fixed that in cvs -- maybe Hans' patch in v4l-dvb cvs could fix it? Gene, try installing v4l-dvb cvs against 2.6.15-rc2-git6 (or later) and see if that might fix NTSC. Somehow, I doubt it -- but it is certainly worth a try.

Regards,

Michael Krufky

#5 lore raw
Quoted 90 lines · Michael Krufky
Quoted 81 lines · Gene Heskett
Quoted 55 lines · Michael Krufky
Quoted 28 lines · Perry Gilfillan
Quoted 19 lines
Quoted 18 lines · Don Koch

(Followup to posting on pcHDTV forum.)

I recentely purchased a pcHDTV HD3000 card. Initially I was having problems with NTSC reception in general (via tuner, composite or svideo input), part of which was a problem with the version of linux I was using. Currently running 2.6.15-rc2-g458af543.

Symptoms:
- ATSC OTA, svideo and composite work. NTSC tuner does not (tried both OTA and cable).
- Xawtv segfaults unless given a specific device (/dev/video0) and then doesn't show anything in any mode including composite and svideo. This may just be an xawtv problem.
- Using tvtime works for composite and svideo, but the tuner doesn't work in either broadcast or cable settings. Disabling signal detection shows the same "image" for any channel.

Quoted 1 line
Quoted 1 line

All the appropriate modules seem to be in place.

I can report some similar troubles. It has worked with vanilla kernel up to 2.6.13 on x86 up untill two weeks ago, when I moved the card to an amd64 system.

I'm currently running gentoo-sources-2.6.13 on amd64. I have not tried the newly merged tree with this kernel.

Quoted 7 lines

However with gentoo-sources-2.6.14-r2 I tried the newly merged tree, and do not get NTSC reception. I have not tried the drivers as released with 2.6.14 yet, so I'll follow up later on that.

I'm also having some issues with the VIA Unichrome drivers with 2.6.14, so I may not be able to get far enough to really see whats happening.

I know that this board uses a Thomson DDT 7610 tuner. I have a FusionHDTV3 Gold-T card that uses Thomson DDT 7611 tuner. I've seen a copy of the datasheet -- both tuners are the same.

Now, here's my story: My card was working fine for a good long time, but ever since about 2 months ago, the analog tuner no longer works. ONLY the ATSC digital tuner is working. The tuner does NOT include a tda9887, AFAIK, and I have also tested in windows. Regardless of whether I am using Linux or Windows, I can only view an ATSC digital stream, NTSC analog no longer works at all. Both used to work in both OS's.

AFAIK, this is a physical hardware problem, and has nothing to do with any bad coding. For the life of me, I cannot figure out what went wrong.

-Mike

Quoted 7 lines

Maybe Herman, or someone else, might be able to tell us what factors can burn out a hybrid tuner's analog capabilities, and leave digital tuning capabilities intact.

Maybe the Thomson DDT 761x tuners are faulty?

Can anybody enlighten us?

Well, based on the note about -rc2-git6 having the v4l stuff reverted, I just built it. Same song, same verse. The tuner acts as if the antenna has been disconnected and held a fraction of an inch away. The video is all digital noise (in VSB), and the audio has a fraction of a word now and then, buried in white noise. As my dish receiver has about a 65,000 u-volt output, I find that sort of signal loss almost unbelievable.

It still takes a cold reboot back to 2.6.14.3 or I think any earlier kernel (that worked ok) in order to restore the cards normal operation, I'm watching cnn on it right now after having done so.

So to me, there is little difference between git3 and git6. Neither works, failing in the same manner exactly, and matching the performance of 15-rc2 as issued. I didn't build -rc1, so I can't say exactly where the hose got cut.

Quoted 4 lines

I'd suggest just blowing the v4l directory in the -rc2 kernel away, and replacing it with that from 2.6.14.3 just to get back to a working sitiuation that you can then start from scratch on. But whatdoIknow? :)Gene Heskett <[email protected]>

EEK! This is not the problem that I'm having at all..... Then again, I have a different board.

Quoted 6 lines

Gene, telling someone to revert all the v4l changes in 2.6.15 doesn't help us to fix the actual problem at all, nor will it help them use their hardware.

Gene, I believe that I asked you to install the cvs modules against 2.6.14, and you told me that doing that works.

No, I never did get it to work Michael, and I believe I said so, mainly because it wouldn't even compile. ISTR I sent a message with the compiler exit messages at the time.

Now I'd be willing to try it again, but I'd need exactly the proceedure as it would apply to a working 2.6.14.x kernel. I don't think I did it right the last time. My script, fwiw, renames one generation back so that a bad kernel can be reverted easily by renameing the vmlinuz and /lib/modules/version-number trees. Its kind of a swiss army knife in that I comment/uncomment stuff in the buildit (thats another script I use to apply then patches etc), but the makeit script only needs the version number updated to match the Makefile and from there its a 'time ./makeit' till I'm done editing grub.conf & ready to reboot. I've taken note that the recent makefiles apparently does its own depmod operation when doing the modules_install but haven't taken my command out of it yet so that gets done twice..

Quoted 6 lines

The code in cvs
contains all the patches that we have sent to 2.6.15 and THEN some. Can you please confirm that installing the cvs modules from v4l-dvb cvs (v4l and dvb have merged cvs repo's) against 2.6.14 is NOT broken?? This would rule out any possibility of the v4l changes in 2.6.15 being the cause of your problems.

See above.

Quoted 4 lines

It is WELL established that there are memory errors in 2.6.15 (although I thought they were all fixed -- guess not). I understand that you are using some MakeIt script to build your kernel -- have you tried the standard method?

My script does it essentially the same as one would do it by hand. The exception being that I understand the Makefile now has an install target for the kernel, but my script does the copying rather than calling that, and *I* do the grub.conf editing.

Quoted 9 lines

[...a few minutes go by...]

OKAY, I concur -- When I did my testing for 2.6.15-rc2-git6, that was with my saa7134-based card, using nxt200x ... success!

However, when I tried using my FusionHDTV3 Gold-T (cx88 card using lgdt3302), neither analog nor digital works in 2.6.15-rc2-git6.... Under 2.6.13 (and 2.6.14, i think), however, digital DOES work. Analog still doesnt work, but i believe that my hardware is damaged.

He should do a full shutdown and cold boot to a kernel he knows works, and I expect the analog will work again. Warm reboots DO NOT DO IT!

Quoted 11 lines

SO, looks like we have a regression somewhere in the kernel that breaks the cx88 driver :-(

I don't even know where to begin.

...One idea... We also know that upstream changes created some compile warnings in tuner-core... Hans has fixed that in cvs -- maybe Hans' patch in v4l-dvb cvs could fix it? Gene, try installing v4l-dvb cvs against 2.6.15-rc2-git6 (or later) and see if that might fix NTSC. Somehow, I doubt it -- but it is certainly worth a try.

Like I said, complete instructions please so that we are on the same page. I still have the rc2-git6 tree that didn't work, so as my script does a make clean, it should be easy enough to do with the right instructions. Like what dir in the kernel tree am I supposed to be in when I issue the cvs checkout command etc.

Quoted 3 lines

Regards,

Michael Krufky

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#6 lore raw
Quoted 101 lines · Gene Heskett
Quoted 28 lines · Michael Krufky
Quoted 16 lines · Gene Heskett

Well, based on the note about -rc2-git6 having the v4l stuff reverted,
I just built it. Same song, same verse. The tuner acts as if the
antenna has been disconnected and held a fraction of an inch away. The video is all digital noise (in VSB), and the audio has a fraction of a word now and then, buried in white noise. As my dish receiver
has about a 65,000 u-volt output, I find that sort of signal loss
almost unbelievable.

It still takes a cold reboot back to 2.6.14.3 or I think any earlier kernel (that worked ok) in order to restore the cards normal operation, I'm watching cnn on it right now after having done so.

So to me, there is little difference between git3 and git6. Neither works, failing in the same manner exactly, and matching the performance of 15-rc2 as issued. I didn't build -rc1, so I can't say exactly where the hose got cut.

Quoted 6 lines

I'd suggest just blowing the v4l directory in the -rc2 kernel away, and replacing it with that from 2.6.14.3 just to get back to a working sitiuation that you can then start from scratch on. But whatdoIknow? :)Gene Heskett <[email protected]>

EEK! This is not the problem that I'm having at all..... Then again, I have a different board.

Quoted 8 lines

Gene, telling someone to revert all the v4l changes in 2.6.15 doesn't help us to fix the actual problem at all, nor will it help them use their hardware.

Gene, I believe that I asked you to install the cvs modules against 2.6.14, and you told me that doing that works.

No, I never did get it to work Michael, and I believe I said so, mainly because it wouldn't even compile. ISTR I sent a message with the compiler exit messages at the time.

Now I'd be willing to try it again, but I'd need exactly the proceedure as it would apply to a working 2.6.14.x kernel. I don't think I did it right the last time. My script, fwiw, renames one generation back so that a bad kernel can be reverted easily by renameing the vmlinuz and /lib/modules/version-number trees. Its kind of a swiss army knife in that I comment/uncomment stuff in the buildit (thats another script I use to apply then patches etc), but the makeit script only needs the version number updated to match the Makefile and from there its a 'time ./makeit' till I'm done editing grub.conf & ready to reboot. I've taken note that the recent makefiles apparently does its own depmod operation when doing the modules_install but haven't taken my command out of it yet so that gets done twice..

Quoted 7 lines

The code in cvs
contains all the patches that we have sent to 2.6.15 and THEN some. Can you please confirm that installing the cvs modules from v4l-dvb cvs (v4l and dvb have merged cvs repo's) against 2.6.14 is NOT
broken?? This would rule out any possibility of the v4l changes in
2.6.15 being the cause of your problems.

See above.

Quoted 5 lines

It is WELL established that there are memory errors in 2.6.15 (although I thought they were all fixed -- guess not). I understand that you are using some MakeIt script to build your kernel -- have you
tried the standard method?

My script does it essentially the same as one would do it by hand. The exception being that I understand the Makefile now has an install target for the kernel, but my script does the copying rather than calling that, and *I* do the grub.conf editing.

Quoted 11 lines

[...a few minutes go by...]

OKAY, I concur -- When I did my testing for 2.6.15-rc2-git6, that was with my saa7134-based card, using nxt200x ... success!

However, when I tried using my FusionHDTV3 Gold-T (cx88 card using lgdt3302), neither analog nor digital works in 2.6.15-rc2-git6.... Under 2.6.13 (and 2.6.14, i think), however, digital DOES work. Analog still doesnt work, but i believe that my hardware is damaged.

He should do a full shutdown and cold boot to a kernel he knows works, and I expect the analog will work again. Warm reboots DO NOT DO IT!

Quoted 12 lines

SO, looks like we have a regression somewhere in the kernel that breaks the cx88 driver :-(

I don't even know where to begin.

...One idea... We also know that upstream changes created some compile warnings in tuner-core... Hans has fixed that in cvs -- maybe Hans' patch in v4l-dvb cvs could fix it? Gene, try installing v4l-dvb cvs against 2.6.15-rc2-git6 (or later) and see if that might fix NTSC. Somehow, I doubt it -- but it is certainly worth a try.

Quoted 6 lines

Like I said, complete instructions please so that we are on the same page. I still have the rc2-git6 tree that didn't work, so as my script does a make clean, it should be easy enough to do with the right instructions. Like what dir in the kernel tree am I supposed to be in when I issue the cvs checkout command etc.

Gene-

Let's go a slightly different route... First off, I must say:

I think that I over-reacted in my previous email. To test, I have installed v4l-dvb cvs over kernels 2.6.13, 2.6.14, and 2.6.15-rc2-git6, and all three worked fine with my FusionHDTV3 Gold-T in ATSC mode. It was the vanilla 2.6.15-rc2-git6 that didnt work for me. With the cvs modules installed, however, it is working well.

Right now, I am in the process of recompiling my vanilla 2.6.15-rc2-git6 kernel, to confirm that this actually is a new regression. If it is in fact a regression in 2.6.15, then it is likely to be a v4l regression. Luckily, Mauro and I have prepared some patchsets to fix some 2.6.15 bugs and compile warnings. It seems that this might be one of the bugs that is now fixed in cvs.

Anyhow, would you like to give cvs a whirl under your 2.6.15-rc2-git6 kernel?

There is a wiki-howto, located at:
http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS

... but I will include instructions in this email for your convenience.

Here's how:

1) Please start with vanilla 2.6.15-rc2-git6 ... Have the kernel already installed and running.

2) Check-out the newly merged v4l-dvb cvs repository:

cvs -d :pserver:[email protected]:/cvs/video4linux login cvs -d :pserver:[email protected]:/cvs/video4linux co v4l-dvb

3) Change into the v4l-dvb directory:

cd v4l-dvb

4) (optional) If you are recompiling the cvs modules against a different kernel, clean the tree and kernel version info:

make distclean

5) Compile the modules:

make

6) Install them: (as root)

make install

7) Reboot the machine

Hopefully, this will fix your problem. Please let me know.

-Michael Krufky

#7 lore raw
Quoted 8 lines · Michael Krufky
Quoted 5 lines · Gene Heskett

Like I said, complete instructions please so that we are on the same page. I still have the rc2-git6 tree that didn't work, so as my script does a make clean, it should be easy enough to do with the right instructions. Like what dir in the kernel tree am I supposed to be in when I issue the cvs checkout command etc.

Oops.... I forgot to answer this question....

It doesnt matter in what directory you are issuing the commands below... Although you certainly should NOT issue these within your kernel source, and you should be inside the newly-downloaded v4l-dvb tree after you "cd" into it. I recommend either doing this in your ~home directory, or in /usr/src

Quoted 21 lines

Here's how:

1) Please start with vanilla 2.6.15-rc2-git6 ... Have the kernel already installed and running.

2) Check-out the newly merged v4l-dvb cvs repository:

cvs -d :pserver:[email protected]:/cvs/video4linux login cvs -d :pserver:[email protected]:/cvs/video4linux co v4l-dvb

3) Change into the v4l-dvb directory:

cd v4l-dvb

4) (optional) If you are recompiling the cvs modules against a different kernel, clean the tree and kernel version info:

make distclean

5) Compile the modules:

make

Quoted 7 lines

6) Install them: (as root)

make install

7) Reboot the machine

Hopefully, this will fix your problem. Please let me know.

Cheers,

Michael

#8 lore raw
Quoted 8 lines · Michael Krufky
Quoted 6 lines · Michael Krufky
Quoted 5 lines · Gene Heskett

Like I said, complete instructions please so that we are on the same page. I still have the rc2-git6 tree that didn't work, so as my script does a make clean, it should be easy enough to do with the right instructions. Like what dir in the kernel tree am I supposed to be in when I issue the cvs checkout command etc.

Quoted 7 lines

Oops.... I forgot to answer this question....

It doesnt matter in what directory you are issuing the commands below... Although you certainly should NOT issue these within your kernel source, and you should be inside the newly-downloaded v4l-dvb tree after you "cd" into it. I recommend either doing this in your ~home directory, or in /usr/src

I built it in /usr/src, seemed to be ok ANAICT. It just didn't work. Bear in mind I have no ATSC signals out here in the West Virginia hills yet, all ntsc. So I can't as yet test the ATSC side.

[...]

Quoted 1 line

Michael

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#9 lore raw
Quoted 26 lines · Gene Heskett
Quoted 13 lines · Michael Krufky
Quoted 10 lines · Michael Krufky
Quoted 7 lines · Gene Heskett

Like I said, complete instructions please so that we are on the same page. I still have the rc2-git6 tree that didn't work, so as my script does a make clean, it should be easy enough to do with the right instructions. Like what dir in the kernel tree am I supposed to be in when I issue the cvs checkout command etc.

Quoted 9 lines

Oops.... I forgot to answer this question....

It doesnt matter in what directory you are issuing the commands below... Although you certainly should NOT issue these within your kernel source, and you should be inside the newly-downloaded v4l-dvb tree after you "cd" into it. I recommend either doing this in your ~home directory, or in /usr/src

Quoted 3 lines

I built it in /usr/src, seemed to be ok ANAICT. It just didn't work. Bear in mind I have no ATSC signals out here in the West Virginia hills yet, all ntsc. So I can't as yet test the ATSC side.

Very interesting... This whole time I thought that your testing was using a VSB stream.

So, you're only talking about problems with analog NTSC?

It seems that I might be having a similar (but NOT the same) problem with my FusionHDTV3 Gold-T, but it has nothing to do with the kernel version. Let me remind you that this card uses Thomson DDT 7611, and your card uses Thomson DDT 7610. It simply doesnt work anymore, regardless of whether I am in Linux or Windows, no matter which kernel version. Only digital ATSC (using QAM256) is working for me.

I will have to test NTSC using my FusionHDTV5 Gold, installed in my other box. It doesnt have 2.6.15 installed right now. I will wait for 2.6.15-rc3 to propogate to the kernel.org mirrors before I install it, so this may take some time. The other card, however, has a TUA6034 tuner, inside an LG TDVS-H062F ... not the same tuner as you this time, but still a cx88 board.

Anyway, after 2.6.15-rc3 propogates and I get it installed, I'll let you know how my testing goes. Would you do the same?

Cheers,

Mike

#10 lore raw
Quoted 1 line · Michael Krufky

Anyway, after 2.6.15-rc3 propogates and I get it installed, I'll let you know how my testing goes. Would you do the same?

My FusionHDTV5 Gold card works fine under 2.6.15-rc3 ... Hopefully yours does too.

/me crosses fingers

-Mike

#11 lore raw
In reply to Michael Krufky
Quoted 18 lines · Michael Krufky
Quoted 17 lines · Gene Heskett
Quoted 15 lines · Michael Krufky
Quoted 6 lines · Michael Krufky
Quoted 5 lines · Gene Heskett

Like I said, complete instructions please so that we are on the same page. I still have the rc2-git6 tree that didn't work, so as my script does a make clean, it should be easy enough to do with the right instructions. Like what dir in the kernel tree am I supposed to be in when I issue the cvs checkout command etc.

Oops.... I forgot to answer this question....

It doesnt matter in what directory you are issuing the commands below... Although you certainly should NOT issue these within your kernel source, and you should be inside the newly-downloaded v4l-dvb tree after you "cd" into it. I recommend either doing this in your ~home directory, or in /usr/src

Quoted 9 lines
Quoted 3 lines

I built it in /usr/src, seemed to be ok ANAICT. It just didn't work. Bear in mind I have no ATSC signals out here in the West Virginia hills yet, all ntsc. So I can't as yet test the ATSC side.

Very interesting... This whole time I thought that your testing was using a VSB stream.

So, you're only talking about problems with analog NTSC?

Correct.

And there is one difference between the modprobe cx88-dvb that works, and the ones that don't. The ones that do report:
kernel: DVB: registering new adapter (cx88[0]).
kernel: DVB: registering frontend 0 (Oren OR51132 VSB/QAM Frontend)...

echoed on the cli, those that don't work are silent.

The complete section from the messages file is:

Nov 28 23:25:55 coyote kernel: Linux video capture interface: v1.00
Nov 28 23:25:55 coyote kernel: cx2388x v4l2 driver version 0.0.5 loaded
Nov 28 23:25:55 coyote kernel: ACPI: PCI Interrupt Link [LNK2] enabled
at IRQ 5
Nov 28 23:25:55 coyote kernel: ACPI: PCI Interrupt 0000:01:07.0[A] ->
Link [LNK2] -> GSI 5 (level, low) -> IRQ 5
Nov 28 23:25:55 coyote kernel: CORE cx88[0]: subsystem: 7063:3000,
board: pcHDTV HD3000 HDTV [card=22,autodetected]
Nov 28 23:25:55 coyote kernel: TV tuner 52 at 0x1fe, Radio tuner -1 at
0x1fe
Nov 28 23:25:55 coyote kernel: i2c-algo-bit.o: cx88[0] passed test.
Nov 28 23:25:55 coyote kernel: tuner 2-0061: chip found @ 0xc2 (cx88[0])
Nov 28 23:25:55 coyote kernel: tuner 2-0061: type set to 52 (Thomson DDT
7610 (ATSC/NTSC))
Nov 28 23:25:55 coyote kernel: cx88[0]/0: found at 0000:01:07.0, rev: 5,
irq: 5, latency: 32, mmio: 0xea000000
Nov 28 23:25:55 coyote kernel: cx88[0]/0: registered device video0
[v4l2]
Nov 28 23:25:55 coyote kernel: cx88[0]/0: registered device vbi0
Nov 28 23:25:55 coyote kernel: cx88[0]/0: registered device radio0
Nov 28 23:25:55 coyote kernel: cx2388x dvb driver version 0.0.5 loaded
Nov 28 23:25:55 coyote kernel: ACPI: PCI Interrupt 0000:01:07.2[A] ->
Link [LNK2] -> GSI 5 (level, low) -> IRQ 5
Nov 28 23:25:55 coyote kernel: cx88[0]/2: found at 0000:01:07.2, rev: 5,
irq: 5, latency: 32, mmio: 0xeb000000
Nov 28 23:25:55 coyote kernel: cx88[0]/2: cx2388x based dvb card

Now, from a 2.6.15-rc2-git6 + cvs boot that doesn't work:

Nov 28 23:15:24 coyote kernel: Linux video capture interface: v1.00
Nov 28 23:15:24 coyote kernel: cx2388x v4l2 driver version 0.0.5 loaded
Nov 28 23:15:24 coyote kernel: ACPI: PCI Interrupt Link [LNK2] enabled
at IRQ 5
Nov 28 23:15:24 coyote kernel: ACPI: PCI Interrupt 0000:01:07.0[A] ->
Link [LNK2] -> GSI 5 (level, low) -> IRQ 5
Nov 28 23:15:24 coyote kernel: CORE cx88[0]: subsystem: 7063:3000,
board: pcHDTV HD3000 HDTV [card=22,autodetected]
Nov 28 23:15:24 coyote kernel: TV tuner 52 at 0x1fe, Radio tuner -1 at
0x1fe
Nov 28 23:15:24 coyote kernel: i2c-algo-bit.o: cx88[0] passed test.
Nov 28 23:15:24 coyote kernel: tuner 2-0061: chip found @ 0xc2 (cx88[0])
Nov 28 23:15:24 coyote kernel: tuner 2-0061: type set to 52 (Thomson DDT
7610 (ATSC/NTSC))
Nov 28 23:15:24 coyote kernel: cx88[0]/0: found at 0000:01:07.0, rev: 5,
irq: 5, latency: 32, mmio: 0xea000000
Nov 28 23:15:24 coyote kernel: cx88[0]/0: registered device video0
[v4l2]
Nov 28 23:15:24 coyote kernel: cx88[0]/0: registered device vbi0
Nov 28 23:15:24 coyote kernel: cx88[0]/0: registered device radio0
Nov 28 23:15:25 coyote kernel: cx2388x dvb driver version 0.0.5 loaded
Nov 28 23:15:25 coyote kernel: ACPI: PCI Interrupt 0000:01:07.2[A] ->
Link [LNK2] -> GSI 5 (level, low) -> IRQ 5
Nov 28 23:15:25 coyote kernel: cx88[0]/2: found at 0000:01:07.2, rev: 5,
irq: 5, latency: 32, mmio: 0xeb000000
Nov 28 23:15:25 coyote kernel: cx88[0]/2: cx2388x based dvb card
Nov 28 23:15:25 coyote kernel: DVB: registering new adapter (cx88[0]).
Nov 28 23:15:25 coyote kernel: DVB: registering frontend 0 (Oren OR51132

VSB/QAM Frontend)...

Which *looks* the same to me, but it doesn't work. Those last 2 lines that are normally echoed to the cli, are not with the newer code although they are in the messages log.

Quoted 14 lines

It seems that I might be having a similar (but NOT the same) problem with my FusionHDTV3 Gold-T, but it has nothing to do with the kernel version. Let me remind you that this card uses Thomson DDT 7611, and your card uses Thomson DDT 7610. It simply doesnt work anymore, regardless of whether I am in Linux or Windows, no matter which kernel version. Only digital ATSC (using QAM256) is working for me.

I will have to test NTSC using my FusionHDTV5 Gold, installed in my other box. It doesnt have 2.6.15 installed right now. I will wait for 2.6.15-rc3 to propogate to the kernel.org mirrors before I install it, so this may take some time. The other card, however, has a TUA6034 tuner, inside an LG TDVS-H062F ... not the same tuner as you this time, but still a cx88 board.

Quoted 2 lines

Anyway, after 2.6.15-rc3 propogates and I get it installed, I'll let you know how my testing goes. Would you do the same?

Yes of course. I need to do some housekeeping in my grub.conf though, I'm currently up to 24 entries. :) So stuff before 2.6.14 is about to go away.

Quoted 3 lines

Cheers,

Mike

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#12 lore raw
Quoted 6 lines · Gene Heskett
Quoted 3 lines

I should also add this is with 2.6.14.2 and todays v4l-dvb CVS.

And I should add that 2.6.14, any version without the cvs update, works perfectly in ntsc here. I have to atsc signals available.

Gene,

Would you mind trying to install the cvs modules on top of 2.6.14, using the instructions that I gave you last night? This will confirm once and for all whether your problems are due to a v4l regression, or an upstream regression in 2.6.15.

Thanks,

Signature · 1 line
Michael Krufky
#13 lore raw
Quoted 6 lines · Michael Krufky
Quoted 4 lines · Gene Heskett
Quoted 1 line

I should also add this is with 2.6.14.2 and todays v4l-dvb CVS.

And I should add that 2.6.14, any version without the cvs update, works perfectly in ntsc here. I have to atsc signals available.

Quoted 8 lines

Gene,

Would you mind trying to install the cvs modules on top of 2.6.14, using the instructions that I gave you last night? This will confirm once and for all whether your problems are due to a v4l regression, or an upstream regression in 2.6.15.

Thanks,

Yes, I can do that. I can also confirm that its still borked in -rc3.

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#14 lore raw
Quoted 9 lines · Gene Heskett
Quoted 8 lines · Michael Krufky
Quoted 4 lines · Gene Heskett
Quoted 1 line

I should also add this is with 2.6.14.2 and todays v4l-dvb CVS.

And I should add that 2.6.14, any version without the cvs update, works perfectly in ntsc here. I have to atsc signals available.

Gene,

Quoted 8 lines
Quoted 6 lines

Would you mind trying to install the cvs modules on top of 2.6.14, using the instructions that I gave you last night? This will confirm once and for all whether your problems are due to a v4l regression, or an upstream regression in 2.6.15.

Thanks,

Yes, I can do that. I can also confirm that its still borked in -rc3.

Turns out I'd borked the linkages in /lib/modules/2.6.14 with a previous attempt, so I'm rebuilding 2.6.14 first in order to restore the linakges. The cvs build is slow, is the -j=1 in the Makefile? Or do I need to spec that on the cli? I've forgotten. Anyway, its done, so a reboot is coming up.

And its broken in the same manner for 2.6.14 + cvs. I tried to do a screen snapshot but the tvtime window was blank in the snapshot. Didn't we used to have a snapshotter that grabbed only the window with ficus way back when?

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#15 lore raw
In reply to Michael Krufky
Quoted 134 lines · Michael Krufky
Quoted 95 lines · Gene Heskett
Quoted 26 lines · Michael Krufky
Quoted 17 lines · Gene Heskett

Well, based on the note about -rc2-git6 having the v4l stuff reverted, I just built it. Same song, same verse. The tuner acts as if the antenna has been disconnected and held a fraction of an inch away. The video is all digital noise (in VSB), and the audio has a fraction of a word now and then, buried in white noise. As my dish receiver has about a 65,000 u-volt output, I find that sort of signal loss almost unbelievable.

It still takes a cold reboot back to 2.6.14.3 or I think any earlier kernel (that worked ok) in order to restore the cards normal operation, I'm watching cnn on it right now after having done so.

So to me, there is little difference between git3 and git6. Neither works, failing in the same manner exactly, and matching the performance of 15-rc2 as issued. I didn't build -rc1, so I can't say exactly where the hose got cut.

Quoted 4 lines

I'd suggest just blowing the v4l directory in the -rc2 kernel away, and replacing it with that from 2.6.14.3 just to get back to a working sitiuation that you can then start from scratch on. But whatdoIknow? :)Gene Heskett <[email protected]>

EEK! This is not the problem that I'm having at all..... Then again, I have a different board.

Quoted 6 lines

Gene, telling someone to revert all the v4l changes in 2.6.15 doesn't help us to fix the actual problem at all, nor will it help them use their hardware.

Gene, I believe that I asked you to install the cvs modules against 2.6.14, and you told me that doing that works.

No, I never did get it to work Michael, and I believe I said so, mainly because it wouldn't even compile. ISTR I sent a message with the compiler exit messages at the time.

Now I'd be willing to try it again, but I'd need exactly the proceedure as it would apply to a working 2.6.14.x kernel. I don't think I did it right the last time. My script, fwiw, renames one generation back so that a bad kernel can be reverted easily by renameing the vmlinuz and /lib/modules/version-number trees. Its kind of a swiss army knife in that I comment/uncomment stuff in the buildit (thats another script I use to apply then patches etc), but the makeit script only needs the version number updated to match the Makefile and from there its a 'time ./makeit' till I'm done editing grub.conf & ready to reboot. I've taken note that the recent makefiles apparently does its own depmod operation when doing the modules_install but haven't taken my command out of it yet so that gets done twice..

Quoted 6 lines

The code in cvs
contains all the patches that we have sent to 2.6.15 and THEN some. Can you please confirm that installing the cvs modules from v4l-dvb cvs (v4l and dvb have merged cvs repo's) against 2.6.14 is NOT broken?? This would rule out any possibility of the v4l changes in 2.6.15 being the cause of your problems.

See above.

Quoted 4 lines

It is WELL established that there are memory errors in 2.6.15 (although I thought they were all fixed -- guess not). I understand that you are using some MakeIt script to build your kernel -- have you tried the standard method?

My script does it essentially the same as one would do it by hand. The exception being that I understand the Makefile now has an install target for the kernel, but my script does the copying rather than calling that, and *I* do the grub.conf editing.

Quoted 9 lines

[...a few minutes go by...]

OKAY, I concur -- When I did my testing for 2.6.15-rc2-git6, that was with my saa7134-based card, using nxt200x ... success!

However, when I tried using my FusionHDTV3 Gold-T (cx88 card using lgdt3302), neither analog nor digital works in 2.6.15-rc2-git6.... Under 2.6.13 (and 2.6.14, i think), however, digital DOES work. Analog still doesnt work, but i believe that my hardware is damaged.

He should do a full shutdown and cold boot to a kernel he knows works, and I expect the analog will work again. Warm reboots DO NOT DO IT!

Quoted 10 lines

SO, looks like we have a regression somewhere in the kernel that breaks the cx88 driver :-(

I don't even know where to begin.

...One idea... We also know that upstream changes created some compile warnings in tuner-core... Hans has fixed that in cvs -- maybe Hans' patch in v4l-dvb cvs could fix it? Gene, try installing v4l-dvb cvs against 2.6.15-rc2-git6 (or later) and see if that might fix NTSC. Somehow, I doubt it -- but it is certainly worth a try.

Quoted 5 lines

Like I said, complete instructions please so that we are on the same page. I still have the rc2-git6 tree that didn't work, so as my script does a make clean, it should be easy enough to do with the right instructions. Like what dir in the kernel tree am I supposed to be in when I issue the cvs checkout command etc.

Gene-

Let's go a slightly different route... First off, I must say:

I think that I over-reacted in my previous email. To test, I have installed v4l-dvb cvs over kernels 2.6.13, 2.6.14, and 2.6.15-rc2-git6, and all three worked fine with my FusionHDTV3 Gold-T in ATSC mode. It was the vanilla 2.6.15-rc2-git6 that didnt work for me. With the cvs modules installed, however, it is working well.

Right now, I am in the process of recompiling my vanilla 2.6.15-rc2-git6 kernel, to confirm that this actually is a new regression. If it is in fact a regression in 2.6.15, then it is likely to be a v4l regression. Luckily, Mauro and I have prepared some patchsets to fix some 2.6.15 bugs and compile warnings. It seems that this might be one of the bugs that is now fixed in cvs.

Anyhow, would you like to give cvs a whirl under your 2.6.15-rc2-git6 kernel?

There is a wiki-howto, located at:
http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS

... but I will include instructions in this email for your convenience.

Here's how:

1) Please start with vanilla 2.6.15-rc2-git6 ... Have the kernel already installed and running.

2) Check-out the newly merged v4l-dvb cvs repository:

Quoted 7 lines

cvs -d :pserver:[email protected]:/cvs/video4linux login cvs -d :pserver:[email protected]:/cvs/video4linux co
v4l-dvb

3) Change into the v4l-dvb directory:

cd v4l-dvb

Got it to that point, but I need to reboot AFTER I disable the modprobe statement in my r.clocal file. That loads about 25 modules it seems & I don't want the old code to touch it before I've rebuilt the cvs.

Ok, rebooted to 2.6.15-rc2-git6

Quoted 5 lines

4) (optional) If you are recompiling the cvs modules against a different kernel, clean the tree and kernel version info:

make distclean

Skipped

Quoted 3 lines

5) Compile the modules:

make

Working now. Done. The only odd thing was a couple of lines as it started stateing 'no version now' but several lines later it had found the kernel version ok.

Quoted 4 lines

6) Install them: (as root)

make install

is this supposed to be make modules_install? I did the plain one. That seemed to work, a repeat shows this:

[root@coyote v4l-dvb]# make install
make -C /usr/src/v4l-dvb/v4l install make[1]: Entering directory `/usr/src/v4l-dvb/v4l'

Eliminating old V4L modules (errors on this step is not a problem).. make[1]: [v4l-rminstall] Error 1 (ignored)

Installing new V4L modules at corresponding Kernel dir... install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/common install -m 644 -c ir-common.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/common install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/video install -m 644 -c btcx-risc.ko bttv.ko tda9887.ko tuner.ko tvaudio.ko tveeprom.ko saa6588.ko tvmixer.ko v4l1-compat.ko v4l2-common.ko wm8775.ko cs53l32a.ko video-buf.ko video-buf-dvb.ko ir-kbd-gpio.ko ir-kbd-i2c.ko msp3400.ko tvp5150.ko saa711x.ko saa7134-alsa.ko saa7134-oss.ko saa7115.ko cx25840.ko saa7127.ko compat_ioctl32.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/video install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/video/cx88 install -m 644 -c cx8800.ko cx8802.ko cx88-alsa.ko cx88-blackbird.ko cx88xx.ko cx88-dvb.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/video/cx88
install: cannot stat `cx88-alsa.ko': No such file or directory make[1]: [v4l-install] Error 1 (ignored)
install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/video/saa7134 install -m 644 -c saa6752hs.ko saa7134.ko saa7134-empress.ko saa7134-dvb.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/video/saa7134 install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/video/em28xx install -m 644 -c em28xx.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/video/em28xx /sbin/depmod -a

Eliminating old DVB modules (errors on this step is not a problem).. make[1]: [dvb-rminstall] Error 1 (ignored)

Installing new DVB modules at corresponding Kernel dir... install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/common install -m 644 -c saa7146.ko saa7146_vv.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/common install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/frontends install -m 644 -c cx22702.ko dvb-pll.ko lgdt330x.ko or51132.ko tda1004x.ko mt352.ko sp887x.ko nxt6000.ko cx24110.ko or51211.ko nxt200x.ko cx24123.ko mt312.ko stv0299.ko nxt2002.ko bcm3510.ko dib3000mb.ko dib3000mc.ko ves1820.ko cx22700.ko tda8083.ko ves1x93.ko stv0297.ko sp8870.ko l64781.ko s5h1420.ko tda10021.ko at76c651.ko tda80xx.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/frontends install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/bt8xx install -m 644 -c bt878.ko dvb-bt8xx.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/bt8xx install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/dvb-core install -m 644 -c dvb-core.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/dvb-core install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/b2c2 install -m 644 -c b2c2-flexcop.ko b2c2-flexcop-pci.ko b2c2-flexcop-usb.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/b2c2 install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/cinergyT2 install -m 644 -c cinergyT2.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/cinergyT2 install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/dvb-usb install -m 644 -c dvb-usb.ko dvb-usb-vp7045.ko dvb-usb-vp702x.ko dvb-usb-dtt200u.ko dvb-usb-a800.ko dvb-usb-dibusb-mb.ko dvb-usb-dibusb-mc.ko dvb-usb-nova-t-usb2.ko dvb-usb-umt-010.ko dvb-usb-digitv.ko dvb-usb-cxusb.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/dvb-usb install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/pluto2 install -m 644 -c pluto2.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/pluto2 install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/ttpci install -m 644 -c budget-core.ko budget.ko ttpci-eeprom.ko budget-av.ko budget-ci.ko budget-patch.ko dvb-ttpci.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/ttpci install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/ttusb-budget install -m 644 -c dvb-ttusb-budget.o /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/ttusb-budget install -d /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/ttusb-dec install -m 644 -c ttusb_dec.ko ttusbdecfe.ko /lib/modules/2.6.15-rc2-git6/kernel/drivers/media/dvb/ttusb-dec /sbin/depmod -a
make[1]: Leaving directory `/usr/src/v4l-dvb/v4l'
------------------
Now the acid test, reboot... Humm, rebooted, but the line in rc.local, that I ran from a vt after logging in as root, did not find the cx88-dvb module, so nothing is loaded. Odd, maybe a typu the first time, the second time after starting x seems to have worked.

So lets see if it works. Nope, same symptoms as before, huge amount of digital noise on-screen, and the audio is thoroughly buried in the noise. Just like the antenna is unplugged, but there is 60,000-65,000 u-v at the tuners input.

Quoted 3 lines

7) Reboot the machine

Hopefully, this will fix your problem. Please let me know.

And now you know. And I'm gonna cold reboot to 2.6.14.3, which will restore it, or has several times before.

Quoted 1 line

-Michael Krufky

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#19 lore raw

Re: Gene's pcHDTV 3000 analog problem

Kirk, Perry and Don, please see below...

Quoted 77 lines · Gene Heskett
Quoted 24 lines · Michael Krufky
Quoted 15 lines

me too. And I just rebuilt 2.6.14 w/o the cvs, did a cold boot, and it works just fine except the modprobe stage was silent at the cli. But my logs are being spammed with:

Nov 29 13:24:49 coyote kernel: CORE IOCTL: 0xc02c5638
Nov 29 13:24:49 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 13:24:49 coyote last message repeated 2 times
Nov 29 13:24:49 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 13:24:49 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 13:24:49 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 13:24:49 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 13:24:53 coyote last message repeated 131 times
Nov 29 13:24:53 coyote kernel: CORE IOCTL: 0xc008561c
Quoted 9 lines

My hard drives have been power cycled more times in the last few days playing with this than at any time in their history since I have a BIG ups.

So I think I'm back to 2.6.14.3 for a while. Lemme know if and when there is something else to test. Right now I'm vacuuming up water in the basement for the first time in years, it rained HARD here earlier.

Quoted 9 lines

Gene-

There IS something more to test, unless I missed an email from you....

...and it's going to be very hard for people to follow your problem... Please stick with your original thread to discuss your particular problem.

Gah, I marked the thread as important, but then it jumped to the v4l list from lkml and I got lost in the shuffle.

Quoted 3 lines

What happens when you install v4l-dvb cvs on top of 2.6.14.y ?

Already did that, blew up as usual, cold boot required.

Quoted 4 lines

(Gene's card uses Thomson DDT 7610 - we need to figure out what analog demod is in play, here.)

Ok, here is an lsmod after doing the modprobe cx88-dvb:

[root@coyote linux-2.6.14]# lsmod |grep cx88

cx88_dvb                7428  0
cx8800                 27276  0
v4l1_compat            13188  1 cx8800
v4l2_common             4864  1 cx8800
cx8802                  9092  1 cx88_dvb
cx88xx                 53024  3 cx88_dvb,cx8800,cx8802
i2c_algo_bit            8456  1 cx88xx
ir_common               7556  1 cx88xx
btcx_risc               3976  3 cx8800,cx8802,cx88xx
tveeprom               12304  1 cx88xx
videodev                7296  2 cx8800,cx88xx
mt352                   6148  1 cx88_dvb
or51132                 9220  1 cx88_dvb
video_buf_dvb           4612  1 cx88_dvb
video_buf              17540  5
cx88_dvb,cx8800,cx8802,cx88xx,video_buf_dvb
lgdt330x                7580  1 cx88_dvb
cx22702                 5892  1 cx88_dvb
dvb_pll                 7812  3 cx88_dvb,or51132,cx22702
i2c_core               17808  12

cx88_dvb,cx88xx,i2c_algo_bit,tuner,tveeprom,mt352,or51132,lgdt330x,cx227
02,w83627hf,i2c_isa,i2c_nforce2
---------
And the log spamming originates and stops according to whether or not
I'm watching tv on it, and continues to do so when booted to 2.6.14.3. Once up and running, its not very often, but hundreds of lines of it during the startup & a few during the shutdown (of tvtime)

Quoted 5 lines

Interesting to me is that I was under the impression that the nxt200x module was the tuner driver, but its not anyplace in the list whether I smunch it with the grep or not.

nxt200x is a dvb frontend driver -- it is NOT a tuner driver, although it does work with the tuner, but not on your board (pchdtv 3000). nxt200x is used by cx88-dvb for the ATi HDTV Wonder, and it's used by saa7134-dvb for the AVerTVHD A180. The nxt200x was not present in 2.6.14, although it IS present in 2.6.15 and v4l-dvb cvs. We know that this is not causing your trouble because Kirk Lapray has both the ATI HDTV Wonder AND pcHDTV 3000 installed on his system, and he says both are working properly.

However, I don't know if Kirk has tried 2.6.15-rc3 yet...

Kirk, do both cards work properly in both analog AND digital video? ...or were you only talking about digital? Kirk, I could really use your help in diagnosing the problem that Gene, Perry and Don are talking about. It looks like something in the 2.6.15 patchsets broke analog video on the pcHDTV 3000.... Would you kindly be able to test 2.6.15-rc3 on your hardware?

Quoted 6 lines

The logs do say its a Thompson 7610 tuner though.
-----------

Nov 29 13:38:31 coyote kernel: TV tuner 52 at 0x1fe, Radio tuner -1 at
0x1fe
Nov 29 13:38:31 coyote kernel: tuner 2-0061: chip found @ 0xc2 (cx88[0])
Nov 29 13:38:31 coyote kernel: tuner 2-0061: type set to 52 (Thomson DDT
Quoted 10 lines

7610 (ATSC/NTSC))
----------

Now, I'm out of pocket till late this evening. I'll check in of course when I get back. I'd like to get this resolved before I get stuck in an airplane & shipped to upstate MI for a while. I'm gettin too old
for this back to work crap. But a $1k/week & expenses comes in handy
too. :)

Gene, Perry and Don -

I wish that we could resolve this with a -git bisection test... Unfortunately, we cannot do it this time, as the VM_RESERVED memory problems existed throughout 2.6.15 development.

All I can think of doing is telling you to start reverting cx88 / tuner related v4l patches, but that is very tedious....

Another idea is to try installing cvs again on top of your kernel, but with each unsuccessful checkout, back out cvs a few days, using the -D option.....

If you need better details on how to do this, let me know... I'll try to write something up.

-Mike

A message that is not in the archive (sent off-list or not yet received): [email protected]

#21 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 6 lines · Kirk Lapray

I only run the cards in digital mode. I have no need to tune to any analog channels with them. When I was working on nxt200x I was able to tune both analog and digital channels on the HDTV Wonder, but I have never tried to tune to any analog channels using the HD3000. This was using kernel 2.6.13 and the cvs v4l and dvb trees.

I am not sure when I will have some time to test 2.6.15-rc3, but if I get some time I will try analog support on this and my current setup.

Kirk

Kirk-

Please test your cards in analog mode, or I fear that we may have to cause NXT200X to depend on BROKEN.

Here's why....

A few months ago, as we added the nxt200x module to cvs, I told you that with nxt200x loaded into memory, I had some extra devices showing up on my i2c bus. At the time, I was using another cx88 card, and it did not use the nxt200x module itself, although it loads up into memory automatically by the cx88-dvb module.

You told me that there was some code in nxt200x module that somehow opens up a channel to hidden i2c devices. Why would this code affect my system if my device is not using the nxt200x module?

Is there code being run at nxt200x module load that is causing this BEFORE cx88-dvb calls nxt200x_attach() ?

It seems that Gene, Perry and Don are having problems with their analog tuners (they each have pcHDTV 3000) ever since nxt200x got added.

Gene, Perry and Don - What happens if you have the cx88 module loaded, but you do NOT load up cx88-dvb (nxt200x will not be loaded) ... Does the problem persist?

You do not need cx88-dvb to view analog television.

Kirk, we need a control group! Please test analog on both boards.

Kirk, there is a thread on the v4l/dvb mailing lists right now about an i2c gate dealing with Hauppauge cards and cx22702 frontend. What Steve Toth has described about this 'i2c gate' is starting to sound similar to what you mentioned about making hidden i2c devices visible.

I'm getting the feeling that nxt200x is indeed the problem.

Gene, Perry and Don .... Another thing you can try -- Once again, install v4l-dvb cvs on top of your running kernel, but this time, before compiling, edit v4l-dvb/v4l/Makefile , and remove the line:

EXTRA_CFLAGS += -DHAVE_NXT200X=1

... This line appears twice, you only need to remove the top one, as it pertains to the cx88 card, although it is safe to remove both for the purposes of this test.

If this fixes your problem, then we know that nxt200x is the cause.

-Mike

#22 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 8 lines · On Tue, 29 Nov 2005 18:01:16 -0500 Michael Krufky wrote:

It seems that Gene, Perry and Don are having problems with their analog tuners (they each have pcHDTV 3000) ever since nxt200x got added.

Gene, Perry and Don - What happens if you have the cx88 module loaded, but you do NOT load up cx88-dvb (nxt200x will not be loaded) ... Does the problem persist?

You do not need cx88-dvb to view analog television.

Same thing.

Quoted 13 lines

Kirk, we need a control group! Please test analog on both boards.

Kirk, there is a thread on the v4l/dvb mailing lists right now about an i2c gate dealing with Hauppauge cards and cx22702 frontend. What Steve Toth has described about this 'i2c gate' is starting to sound similar to what you mentioned about making hidden i2c devices visible.

I'm getting the feeling that nxt200x is indeed the problem.

Gene, Perry and Don .... Another thing you can try -- Once again, install v4l-dvb cvs on top of your running kernel, but this time, before compiling, edit v4l-dvb/v4l/Makefile , and remove the line:

Quoted 7 lines

EXTRA_CFLAGS += -DHAVE_NXT200X=1

... This line appears twice, you only need to remove the top one, as it pertains to the cx88 card, although it is safe to remove both for the purposes of this test.

If this fixes your problem, then we know that nxt200x is the cause.

Working on this build, now.

Quoted 2 lines

-Mike

-d

#23 lore raw

Re: Gene's pcHDTV 3000 analog problem

In reply to Michael Krufky
Quoted 7 lines · On Tue, 29 Nov 2005 18:01:16 -0500 Michael Krufky wrote:

Gene, Perry and Don .... Another thing you can try -- Once again, install v4l-dvb cvs on top of your running kernel, but this time, before compiling, edit v4l-dvb/v4l/Makefile , and remove the line:

EXTRA_CFLAGS += -DHAVE_NXT200X=1

Quoted 8 lines

... This line appears twice, you only need to remove the top one, as it pertains to the cx88 card, although it is safe to remove both for the purposes of this test.

If this fixes your problem, then we know that nxt200x is the cause.

-Mike

No difference.

Back to looking at the code...

-d

A message that is not in the archive (sent off-list or not yet received): [email protected]

#25 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 5 lines · Kirk Lapray

There are some i2c commands sent in cx88-cards.c, but that should only happen with the HDTV Wonder.

The nxt200x should only be sending i2c commands if it detects a NXT2002 or NXT2004 chip. You can set the nxt200x module to load with debug=1 to see if it is doing something that it shouldn't. The only thing that I can think of is that the auto detecting of NXT2002 and NXT2004 is not working like it is supposed to. If you do not have a NXT2002 or NXT2004 based card the nxt200x module should not do anything. The only i2c command in the attach function is to read the card id.

From what I have seen the only device that is made visible is the tuner. This allows the tuner modules to connect to it. It basically turns on the tuner, and this should only happen on NXT2004 based cards.

Quoted 8 lines · Don Koch
Quoted 3 lines

If this fixes your problem, then we know that nxt200x is the cause.

No difference.

Back to looking at the code...

Okay, it looks like nxt200x is a red herring, sorry for the false alarm. Even still, Kirk, it would be very helpful if you could confirm analog video functionality of your pcHDTV 3000 board, using either 2.6.15-rc3, or v4l-dvb cvs on top of any kernel version.

...But before that, please first try it in your current configuration.

I do not have this board, and there *IS* a problem with it... Based on the tests done by Gene and Don, it looks like the regression is inside the v4l code, although this doesnt make any sense, as other cards with similar hardware are not affected.

I don't know any of the specifics about this board, besides the fact that is has OR51132 (irrelevant - the problem exists even without cx88-dvb loaded) Thomson DDT 7610, and a cx2388x chip.

I have two cx88 boards of my own - FusionHDTV5 Gold, everything works fine, both analog and digital. FusionHDTV3 Gold-T, using Thomson DDT 7611 (same as 7610) , and digital works, but the analog tuner is completely broken, regardless of what kernel version, regardless of whether I'm using Linux or windows, and it used to work a few months ago.

I think my hardware is fried, but since Gene is able to restore functionality by a cold boot into 2.6.14.2, it makes me think his problem is because of bad code, where my problem is due to bad hardware.

All I can think of doing next is to have Gene, Don or Perry do a bisection test on our cvs repo.... checking out different cvs revisions until we can narrow it down to the day the problem patch was applied.

::sigh::

Who wants to do it? I'll give you detailed instructions if you're willing.

Regards,
Mike

#26 lore raw

Re: Gene's pcHDTV 3000 analog problem

On Tuesday 29 November 2005 20:26, Michael Krufky wrote:

[...]

Quoted 5 lines

All I can think of doing next is to have Gene, Don or Perry do a bisection test on our cvs repo.... checking out different cvs revisions until we can narrow it down to the day the problem patch was applied.

::sigh::

A sigh? More like an 'oh fudge' or whatever your fav expletive deleted is...

Quoted 5 lines

Who wants to do it? I'll give you detailed instructions if you're willing.

Regards,
Mike

Can you farm it out, one set of patches to each of us? Or do you want to setup a seperate cvs tree based on the v4l code in 2.6.14.3, and incrementally patch it as we each report its still ok, until it breaks again? I think I'd prefer the latter so we are all near the same page even if it takes 3x longer to arrive at the answer. How many actual patches in terms of dependency groups are there? I know, I coulda went all week without asking that :(

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#27 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 18 lines · Gene Heskett

On Tuesday 29 November 2005 20:26, Michael Krufky wrote:

[...]

Quoted 7 lines

ll I can think of doing next is to have Gene, Don or Perry do a bisection test on our cvs repo.... checking out different cvs revisions until we can narrow it down to the day the problem patch was applied.

::sigh::

A sigh? More like an 'oh fudge' or whatever your fav expletive deleted is...

Quoted 4 lines

Who wants to do it? I'll give you detailed instructions if you're willing.

Quoted 8 lines

Can you farm it out, one set of patches to each of us? Or do you want to setup a seperate cvs tree based on the v4l code in 2.6.14.3, and incrementally patch it as we each report its still ok, until it breaks again? I think I'd prefer the latter so we are all near the same page even if it takes 3x longer to arrive at the answer. How many actual patches in terms of dependency groups are there? I know, I coulda went all week without asking that :(

Actually, cvs has a parameter that lets you specify cutoff dates...

This is what I am suggesting that you do... Base this on my previous cvs instructions....

reminder: http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS

so....

1st:

cvs -d :pserver:[email protected]:/cvs/video4linux login cvs -d :pserver:[email protected]:/cvs/video4linux co v4l-dvb cd v4l-dvb
make clean
make
make install

test

(you already did this - you said doesnt work)

Now, try again, using last month's code:

cvs up -D 2005-11-01
make clean
make
make install

did it work? yes? ok, so try two weeks later:

cvs up -D 2005-11-15
make clean
make
make install

... or no it didnt? then another month earlier:

cvs up -D 2005-10-01
make clean
make
make install

works now? try two weeks later:

cvs up -D 2005-10-15
make clean
make
make install

doesnt work? 1 week earlier:

cvs up -D 2005-10-07
make clean
make
make install

doesnt work? 1 day earlier:

cvs up -D 2005-10-06
make clean
make
make install

it works? hmm... problem patch must have been committed between 10-06 and 10-07

At that point, just give me the date and it will probably be easy to narrow it down, or then again, you can use time of day in the -D parameter too... (see man cvs)

anyway, you get the picture... This is an adaptation for cvs based on Linus' description of the git bisection testing, and I'm sure that there are more efficient ways of doing it, but this should work.

Cheers,

Mike

#28 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 10 lines · On Wed, 30 Nov 2005 00:29:23 -0500 Michael Krufky wrote:
Quoted 8 lines · Gene Heskett

On Tuesday 29 November 2005 20:26, Michael Krufky wrote:

[...]

Quoted 4 lines

ll I can think of doing next is to have Gene, Don or Perry do a bisection test on our cvs repo.... checking out different cvs revisions until we can narrow it down to the day the problem patch was applied.

Do we know of a date where the code is known to work. First thing I'd like to do is verify that the card works at all. Remember, I've never seen NTSC tuner mode work and don't want to chase a red herring if the card is busted.

Thanks,
-d

#29 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 6 lines · Don Koch

On Wed, 30 Nov 2005 00:29:23 -0500

Quoted 3 lines · Michael Krufky
Quoted 2 lines · Gene Heskett

On Tuesday 29 November 2005 20:26, Michael Krufky wrote:

Quoted 8 lines
Quoted 6 lines
Quoted 6 lines

[...]

Quoted 4 lines

ll I can think of doing next is to have Gene, Don or Perry do a bisection test on our cvs repo.... checking out different cvs revisions until we can narrow it down to the day the problem patch was applied.

Do we know of a date where the code is known to work.

I assume this is actually a question. Its one I'm not privy to other than whats in 2.6.14.3 and earlier works. As to when that was merged into the kernel tarballs, I'll let Michael see if he can date it. And then we'ed want to look at anything post that merge date, using the bisect methods suggested.

However, its not going to be started here tonight, I need some sleep, drank way too much coffee yesterday & didn't sleep last night at all.

Quoted 4 lines

First thing I'd
like to do is verify that the card works at all. Remember, I've never seen NTSC tuner mode work and don't want to chase a red herring if the card is busted.

It should work with a stock 2.6.14.3 build if its going to work I think, although there may be other factors for cards other than my pcHDTV-3000. Michael?

Quoted 7 lines

Thanks,

(quoted diff) +0 -2
-d
-

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#30 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 37 lines · Gene Heskett
Quoted 12 lines · Don Koch

On Wed, 30 Nov 2005 00:29:23 -0500

Quoted 7 lines · Michael Krufky
Quoted 4 lines · Gene Heskett

On Tuesday 29 November 2005 20:26, Michael Krufky wrote:

[...]

Quoted 9 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

All I can think of doing next is to have Gene, Don or Perry do a bisection test on our cvs repo.... checking out different cvs revisions until we can narrow it down to the day the problem patch was applied.

Do we know of a date where the code is known to work.

I assume this is actually a question. Its one I'm not privy to other than whats in 2.6.14.3 and earlier works. As to when that was merged into the kernel tarballs, I'll let Michael see if he can date it. And then we'ed want to look at anything post that merge date, using the bisect methods suggested.

Quoted 6 lines

First thing I'd
like to do is verify that the card works at all. Remember, I've never seen NTSC tuner mode work and don't want to chase a red herring if the card is busted.

Quoted 5 lines

It should work with a stock 2.6.14.3 build if its going to work I think, although there may be other factors for cards other than my pcHDTV-3000. Michael?

2.6.14 was released about a month ago... so, to be safe, I'd say about 2 months ago in cvs...

The idea of a bisection test is to always make your next test based on half the amount of changesets... It's not as easy to do this with cvs as it is with git, but we can make the best of it......

so, ideally, your tests can be like this:

1- 2months ago.
2- 1month ago
3- 2 weeks ago
4- 3 weeks ago
5- 2 weeks, 3 days ago
6- 2 weeks, 1 day ago

.... at that point, we can look at the cvs commit logs and either

a) start checking out individual patches

or

b) start making the checkout based on time of day in addition of date, in the form, YYYY-MM-DD HH:NN

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#36 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 9 lines · Perry Gilfillan

I've been able to pinpoint the time when the v4l-dvb cvs failes for the pcHDTV 3000 tuner!!

This test is with gentoo-sources-2.6.13-r5, with v4l-dvb cvs. As Michael pointed out, 2.6.15 has bugs. For my hardware, 2.6.13, and 14 have major issues as well...

My procedure is to do a distclean in the v4l-dvb cvs directory, checkout the repository for the given date, make, and install. Then I inspect the /lib/modules/.../media/video dirtree to be sure no old ko files are lingering. I do a cold boot to be sure the hardware is fully initialized from a known state. After many iterations, I've narrowed it down to this change set:

cvs as of 2005-10-17 16:01:
Analog TV works, as well as composite and s-video.

Quoted 7 lines

Changes made at 2005-10-17 16:02: bROKEN!!

cvs as of 2005-10-17 16:10:
Analog failes, s-video and composite work.

I'll leave it to more capable persons to figure how it failed.

Hmm... These times of day don't coincide with any actual commits....

Anyhow, this was the day Hartmut added tda8275a support. It broke many things, which we thought that we had eventually fixed. Maybe there are still a few bugs as a result of this?

This is enough for me to put together a test patch for you, but I cant do it right now..... I'll try to email you an experimental patch tonight or tomorrow that will remove tda8275a support, just to confirm that this is in fact the problem.

If it turns out that I'm right, then we'll have to figure out a way to fix it, since reverting Hartmut's patch is NOT an option.

Thanks for the testing!

-Mike

#37 lore raw

Re: Gene's pcHDTV 3000 analog problem

After checking the datasheets of Thompson tuner, and I have one guess:

At board description, tda9887 is not there. This tuner needs to work
properly.

This small patch does enable it for your board.

You should notice that you may need to use some parameters for tda0887
to work properly, like using port1=0 port2=0 qss=0 as insmod options for this module. (these are some on/off bits at the chip, to enable some special functions - if 0/0/0 doesn't work you may need to test 0/0/1, .. 1/1/1).

Cheers,
Mauro.

Index: linux/drivers/media/video/cx88/cx88-cards.c
===================================================================
RCS file: /cvs/video4linux/v4l-dvb/linux/drivers/media/video/cx88/cx88-cards.c,v retrieving revision 1.108

drivers/media/video/cx88/cx88-cards.c +1 -0
diff -u -p -r1.108 cx88-cards.c
@@ -569,6 +569,7 @@ struct cx88_board cx88_boards[] = {
 		.radio_type     = UNSET,
 		.tuner_addr	= ADDR_UNSET,
 		.radio_addr	= ADDR_UNSET,
+		.tda9887_conf   = TDA9887_PRESENT,
 		.input          = {{
 			.type   = CX88_VMUX_TELEVISION,
 			.vmux   = 0,
#38 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 5 lines · Mauro Carvalho Chehab

After checking the datasheets of Thompson tuner, and I have one guess:

At board description, tda9887 is not there. This tuner needs to work
properly.

Quoted 9 lines

This small patch does enable it for your board.

You should notice that you may need to use some parameters for tda0887
to work properly, like using port1=0 port2=0 qss=0 as insmod options for this module. (these are some on/off bits at the chip, to enable some special functions - if 0/0/0 doesn't work you may need to test 0/0/1, .. 1/1/1).

This has fixed it for me!! I compiled todays cvs, with out this patch, to watch it fail, then added the line as noted, and reloaded the modules without rebooting, and it worked. I did a cold start to see that it is repeatable, and it continues to work. I used no extra parameters, so the defaults are working fine.

Cheers,

Perry

Quoted 22 lines

Cheers, Mauro.

------------------------------------------------------------------------

Index: linux/drivers/media/video/cx88/cx88-cards.c
===================================================================
RCS file: /cvs/video4linux/v4l-dvb/linux/drivers/media/video/cx88/cx88-cards.c,v retrieving revision 1.108
diff -u -p -r1.108 cx88-cards.c

drivers/media/video/cx88/cx88-cards.c +1 -0
@@ -569,6 +569,7 @@ struct cx88_board cx88_boards[] = {
 		.radio_type     = UNSET,
 		.tuner_addr	= ADDR_UNSET,
 		.radio_addr	= ADDR_UNSET,
+		.tda9887_conf   = TDA9887_PRESENT,
 		.input          = {{
 			.type   = CX88_VMUX_TELEVISION,
 			.vmux   = 0,
#39 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 36 lines · Perry Gilfillan
Quoted 6 lines · Mauro Carvalho Chehab

After checking the datasheets of Thompson tuner, and I have one
guess:

At board description, tda9887 is not there. This tuner needs to work
properly.

Quoted 8 lines

This small patch does enable it for your board.

You should notice that you may need to use some parameters for
tda0887
to work properly, like using port1=0 port2=0 qss=0 as insmod options for this module. (these are some on/off bits at the chip, to enable some special functions - if 0/0/0 doesn't work you may need to test 0/0/1, .. 1/1/1).

This has fixed it for me!! I compiled todays cvs, with out this patch, to watch it fail, then added the line as noted, and reloaded the modules without rebooting, and it worked. I did a cold start to see that it is repeatable, and it continues to work. I used no extra parameters, so the defaults are working fine.

Cheers,

Perry

Quoted 13 lines

Cheers, Mauro.

------------------------------------------------------------------------

Index: linux/drivers/media/video/cx88/cx88-cards.c
===================================================================
RCS file: /cvs/video4linux/v4l-dvb/linux/drivers/media/video/cx88/cx88-cards.c,v retrieving revision 1.108
diff -u -p -r1.108 cx88-cards.c
--- linux/drivers/media/video/cx88/cx88-cards.c 25 Nov 2005
10:24:13 -0000 1.108
+++ linux/drivers/media/video/cx88/cx88-cards.c 1 Dec 2005
20:56:43 -0000

Quoted 9 lines

drivers/media/video/cx88/cx88-cards.c struct cx88_board cx88_boards[] = { · line 575

Quoted 8 lines
drivers/media/video/cx88/cx88-cards.c +1 -0
@@ -569,6 +569,7 @@ struct cx88_board cx88_boards[] = {
        .radio_type     = UNSET,
        .tuner_addr    = ADDR_UNSET,
        .radio_addr    = ADDR_UNSET,
+        .tda9887_conf   = TDA9887_PRESENT,
        .input          = {{
            .type   = CX88_VMUX_TELEVISION,
            .vmux   = 0, 

HURRAY!!!!!

... I'll go add this to cvs now.

#40 lore raw

Re: Gene's pcHDTV 3000 analog problem

In reply to Perry Gilfillan
Quoted 15 lines · Perry Gilfillan
Quoted 6 lines · Mauro Carvalho Chehab

After checking the datasheets of Thompson tuner, and I have one guess:

At board description, tda9887 is not there. This tuner needs to work properly.

Quoted 7 lines

This small patch does enable it for your board.

You should notice that you may need to use some parameters for tda0887 to work properly, like using port1=0 port2=0 qss=0 as insmod options for this module. (these are some on/off bits at the chip, to enable some special functions - if 0/0/0 doesn't work you may need to test 0/0/1, .. 1/1/1).

Quoted 5 lines

This has fixed it for me!! I compiled todays cvs, with out this patch, to watch it fail, then added the line as noted, and reloaded the modules without rebooting, and it worked. I did a cold start to see that it is repeatable, and it continues to work. I used no extra parameters, so the defaults are working fine.

I haven't built it yet, had to apply the patch by hand for some reason here, after doing a cvs up -D today.

2.6.15-rc4 under construction to test it.

Quoted 29 lines

Cheers,

Perry

Quoted 24 lines

Cheers,
Mauro.

----------------------------------------------------------------------
--

Index: linux/drivers/media/video/cx88/cx88-cards.c
===================================================================
RCS file:
/cvs/video4linux/v4l-dvb/linux/drivers/media/video/cx88/cx88-cards.c, v retrieving revision 1.108
diff -u -p -r1.108 cx88-cards.c
--- linux/drivers/media/video/cx88/cx88-cards.c 25 Nov 2005 10:24:13
-0000 1.108 +++ linux/drivers/media/video/cx88/cx88-cards.c 1 Dec
2005 20:56:43 -0000 @@ -569,6 +569,7 @@ struct cx88_board cx88_boards[] = {
.radio_type = UNSET,
.tuner_addr = ADDR_UNSET,
.radio_addr = ADDR_UNSET,

drivers/media/video/cx88/cx88-cards.c +1 -0
+		.tda9887_conf   = TDA9887_PRESENT,
		.input          = {{
			.type   = CX88_VMUX_TELEVISION,
			.vmux   = 0,
Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#41 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 29 lines · Gene Heskett
Quoted 18 lines · Perry Gilfillan
Quoted 6 lines · Mauro Carvalho Chehab

After checking the datasheets of Thompson tuner, and I have one guess:

At board description, tda9887 is not there. This tuner needs to work properly.

Quoted 9 lines

This small patch does enable it for your board.

You should notice that you may need to use some parameters for tda0887 to work properly, like using port1=0 port2=0 qss=0 as insmod options for this module. (these are some on/off bits at the chip, to enable some special functions - if 0/0/0 doesn't work you may need to test 0/0/1, .. 1/1/1).

Quoted 7 lines

This has fixed it for me!! I compiled todays cvs, with out this patch, to watch it fail, then added the line as noted, and reloaded the modules without rebooting, and it worked. I did a cold start to see that it is repeatable, and it continues to work. I used no extra parameters, so the defaults are working fine.

Quoted 5 lines

I haven't built it yet, had to apply the patch by hand for some reason here, after doing a cvs up -D today.

2.6.15-rc4 under construction to test it.

Gene, dont bother applying the patch to the cvs code -- It is obviously correct, so I have merged it into v4l-dvb cvs.

A simple cvs checkout will get you everything you need.

If you like, try to apply it against the kernel, without using cvs.

Anyhow, surely it will work. Thank you Gene, Don and Perry for helping us to solve this bug......... Hopefully we'll be able to get the fix into 2.6.15.

Cheers,

Michael Krufky

#42 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 23 lines · Michael Krufky
Quoted 22 lines · Gene Heskett
Quoted 20 lines · Perry Gilfillan
Quoted 6 lines · Mauro Carvalho Chehab

After checking the datasheets of Thompson tuner, and I have one guess:

At board description, tda9887 is not there. This tuner needs to work properly.

Quoted 7 lines

This small patch does enable it for your board.

You should notice that you may need to use some parameters for tda0887 to work properly, like using port1=0 port2=0 qss=0 as insmod options for this module. (these are some on/off bits at the chip, to enable some special functions - if 0/0/0 doesn't work you may need to test 0/0/1, .. 1/1/1).

This has fixed it for me!! I compiled todays cvs, with out this patch, to watch it fail, then added the line as noted, and reloaded the modules without rebooting, and it worked. I did a cold start to see that it is repeatable, and it continues to work. I used no extra parameters, so the defaults are working fine.

Quoted 7 lines
Quoted 4 lines

I haven't built it yet, had to apply the patch by hand for some reason here, after doing a cvs up -D today.

2.6.15-rc4 under construction to test it.

Gene, dont bother applying the patch to the cvs code -- It is obviously correct, so I have merged it into v4l-dvb cvs.

I did a cvs up -D 2005-12-1, about 1.5 hours ago and it hadn't been applied yet.

Quoted 1 line

A simple cvs checkout will get you everything you need.

I'll resuck the whole thing.

Quoted 1 line

If you like, try to apply it against the kernel, without using cvs.

Ok, that I can probably do. Unforch, there is a miss-cue in that file, rc4 says:
struct cx88_board cx88_boards[] = {
[CX88_BOARD_UNKNOWN] = {
.name = "UNKNOWN/GENERIC",
.tuner_type = UNSET,
.radio_type = UNSET,
.tuner_addr = ADDR_UNSET,
.radio_addr = ADDR_UNSET,
.input = {{
.type = CX88_VMUX_COMPOSITE1,

Note also that the .type = label has been changed. I'm going to change that line back to CX88_VMUX_TELEVISION, just for grins.

While the patch says:
@@ -569,6 +569,7 @@ struct cx88_board cx88_boards[] = {

                .radio_type     = UNSET,
                .tuner_addr     = ADDR_UNSET,
                .radio_addr     = ADDR_UNSET,
                .tda9887_conf   = TDA9887_PRESENT,
                .input          = {{
                        .type   = CX88_VMUX_TELEVISION,
                        .vmux   = 0,

And I've added that line from the patch to just above the ".input" line, its rebuilding now. And a cold reboot will test it, so I'll send this first just to let you know not all is well in tv land.

Quoted 7 lines

Anyhow, surely it will work. Thank you Gene, Don and Perry for helping us to solve this bug......... Hopefully we'll be able to get the fix into 2.6.15.

Cheers,

Michael Krufky

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.

A message that is not in the archive (sent off-list or not yet received): 1133486311.2969.11.camel@frolic

#44 lore raw

Re: Gene's pcHDTV 3000 analog problem

On Thursday 01 December 2005 20:18, R C wrote:
Gene adds: And damn, I've lost track of the original message that contained this patch being discussed. So some have been inadvertantly left out of this thread, my apologies.

Quoted 21 lines

Gene;

On Thu, 2005-12-01 at 19:03 -0500, Gene Heskett wrote:

[snip]

Quoted 15 lines

Ok, that I can probably do. Unforch, there is a miss-cue in that file, rc4 says:
struct cx88_board cx88_boards[] = {
[CX88_BOARD_UNKNOWN] = {
.name = "UNKNOWN/GENERIC",
.tuner_type = UNSET,
.radio_type = UNSET,
.tuner_addr = ADDR_UNSET,
.radio_addr = ADDR_UNSET,
.input = {{
.type = CX88_VMUX_COMPOSITE1,

Note also that the .type = label has been changed. I'm going to change that line back to CX88_VMUX_TELEVISION, just for grins.

Quoted 15 lines

drivers/media/video/cx88/cx88-cards.c struct cx88_board cx88_boards[] = { · line 575

Quoted 9 lines

While the patch says:

drivers/media/video/cx88/cx88-cards.c +0 -0
@@ -569,6 +569,7 @@ struct cx88_board cx88_boards[] = {
                .radio_type     = UNSET,
                .tuner_addr     = ADDR_UNSET,
                .radio_addr     = ADDR_UNSET,
                .tda9887_conf   = TDA9887_PRESENT,
                .input          = {{
                        .type   = CX88_VMUX_TELEVISION,
                        .vmux   = 0,

This means the patch applies _somewhere in_ that struct, not at the top of said struct. You're changing the entry for the default (unrecognized/generic) board (which won't get you anywhere), and the patch edits the pchdtv3000 entry (around line 569, or 530 lines below where you're editing)

After properly applying the patch, I'm now watching tv while booted to 2.6.15-rc4. And it didn't take a cold boot to do it. Hurrah!

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#45 lore raw

Re: Gene's pcHDTV 3000 analog problem

In reply to Mauro Carvalho Chehab

On Thursday 01 December 2005 16:00, Mauro Carvalho Chehab wrote:
I also took Hartmut out of the Cc: list, he is mailbombing me with a sender id scheme of some kind for every reply I make to All:

Its very poor form to do that to a mailing list although I suspect its because he is in the CC list and getting it directly if I don't remove him from it. He doesn't need 2 copies anyway.

Quoted 5 lines

After checking the datasheets of Thompson tuner, and I have one guess:

At board description, tda9887 is not there. This tuner needs to work properly.

Quoted 7 lines

This small patch does enable it for your board.

You should notice that you may need to use some parameters for tda0887 to work properly, like using port1=0 port2=0 qss=0 as insmod options for this module. (these are some on/off bits at the chip, to enable some special functions - if 0/0/0 doesn't work you may need to test 0/0/1, .. 1/1/1).

I'm puzzled. In 2.6.14.3, that tuner is not loaded, and its working fine other than spamming the logs by the megabyte about

Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc02c5638
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 22:56:18 coyote last message repeated 2 times
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 22:56:49 coyote last message repeated 907 times
Nov 29 22:57:50 coyote last message repeated 1828 times
Nov 29 22:58:46 coyote last message repeated 1703 times
Nov 29 22:58:46 coyote kernel: CORE IOCTL: 0xc008561c

Other than that, its fine in ntsc.

I'm currently doing a 'modprobe cx88-dvb' to load everything, but you mention using insmod with arguments above, so please clarify as to how I should proceed after applying that patch to, oh lets live dangerously and do it to 2.6.15-rc4.

Quoted 2 lines

Cheers,
Mauro.

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#46 lore raw

Re: Gene's pcHDTV 3000 analog problem

In reply to Mauro Carvalho Chehab
Quoted 7 lines · On Thu, 01 Dec 2005 19:00:59 -0200 Mauro Carvalho Chehab wrote:

After checking the datasheets of Thompson tuner, and I have one guess:

At board description, tda9887 is not there. This tuner needs to work
properly.

This small patch does enable it for your board.

Quoted 9 lines

You should notice that you may need to use some parameters for tda0887
to work properly, like using port1=0 port2=0 qss=0 as insmod options for this module. (these are some on/off bits at the chip, to enable some special functions - if 0/0/0 doesn't work you may need to test 0/0/1, .. 1/1/1).

Cheers,
Mauro.

Works fine here! Thanks, Mauro!!

The default settings worked.

-d

#47 lore raw

Re: Gene's pcHDTV 3000 analog problem

In reply to Michael Krufky
Quoted 49 lines · On Wed, 30 Nov 2005 00:29:23 -0500 Michael Krufky wrote:
Quoted 18 lines · Gene Heskett

On Tuesday 29 November 2005 20:26, Michael Krufky wrote:

[...]

Quoted 7 lines

ll I can think of doing next is to have Gene, Don or Perry do a bisection test on our cvs repo.... checking out different cvs revisions until we can narrow it down to the day the problem patch was applied.

::sigh::

A sigh? More like an 'oh fudge' or whatever your fav expletive deleted is...

Quoted 4 lines

Who wants to do it? I'll give you detailed instructions if you're willing.

Quoted 8 lines

Can you farm it out, one set of patches to each of us? Or do you want to setup a seperate cvs tree based on the v4l code in 2.6.14.3, and incrementally patch it as we each report its still ok, until it breaks again? I think I'd prefer the latter so we are all near the same page even if it takes 3x longer to arrive at the answer. How many actual patches in terms of dependency groups are there? I know, I coulda went all week without asking that :(

Actually, cvs has a parameter that lets you specify cutoff dates...

This is what I am suggesting that you do... Base this on my previous cvs instructions....

reminder: http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS

so....

1st:

cvs -d :pserver:[email protected]:/cvs/video4linux login cvs -d :pserver:[email protected]:/cvs/video4linux co v4l-dvb cd v4l-dvb
make clean
make
make install

test

(you already did this - you said doesnt work)

[...]

Quoted 11 lines

cvs up -D 2005-10-15
make clean
make
make install

doesnt work? 1 week earlier:

cvs up -D 2005-10-07
make clean
make
make install

Let's put it this way: for me, 2005-10-10 doesn't work and anything earlier doesn't build. I've tried building against 2.6.15-rc3, 2.6.14 and 2.6.13. The card doesn't work against the built-in 2.6.13 code, but the tuner is sending bizarre stuff
to it (channel 2 is *not* at 97.25 MHz). 2.6.14 kept spewing:

CORE IOCTL: 0xc054561d
cx88[0]: ioctl 0xc054561d (v4l2, rw, VIDIOC_G_TUNER)

all over the syslog.

Build issues include broken Makefiles (around 10-08) and missing header files.

Quoted 4 lines

Cheers,

Mike

-d

#48 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 88 lines · Don Koch
Quoted 62 lines · On Wed, 30 Nov 2005 00:29:23 -0500 Michael Krufky wrote:
Quoted 26 lines · Gene Heskett

On Tuesday 29 November 2005 20:26, Michael Krufky wrote:

[...]

Quoted 9 lines

ll I can think of doing next is to have Gene, Don or Perry do a bisection test on our cvs repo.... checking out different cvs revisions until we can narrow it down to the day the problem patch was applied.

::sigh::

A sigh? More like an 'oh fudge' or whatever your fav expletive deleted is...

Quoted 6 lines

Who wants to do it? I'll give you detailed instructions if you're willing.

Quoted 10 lines

Can you farm it out, one set of patches to each of us? Or do you want to setup a seperate cvs tree based on the v4l code in 2.6.14.3, and incrementally patch it as we each report its still ok, until it breaks again? I think I'd prefer the latter so we are all near the same page even if it takes 3x longer to arrive at the answer. How many actual patches in terms of dependency groups are there? I know, I coulda went all week without asking that :(

Actually, cvs has a parameter that lets you specify cutoff dates...

This is what I am suggesting that you do... Base this on my previous cvs instructions....

reminder: http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS

so....

1st:

cvs -d :pserver:[email protected]:/cvs/video4linux login cvs -d :pserver:[email protected]:/cvs/video4linux co v4l-dvb cd v4l-dvb
make clean
make
make install

test

(you already did this - you said doesnt work)

[...]

Quoted 5 lines

cvs up -D 2005-10-15
make clean
make
make install

Quoted 8 lines

doesnt work? 1 week earlier:

cvs up -D 2005-10-07
make clean
make
make install

Let's put it this way: for me, 2005-10-10 doesn't work and anything earlier doesn't build. I've tried building against 2.6.15-rc3, 2.6.14 and 2.6.13. The card doesn't work against the built-in 2.6.13 code, but the tuner is sending bizarre stuff
to it (channel 2 is *not* at 97.25 MHz). 2.6.14 kept spewing:

Quoted 14 lines

CORE IOCTL: 0xc054561d
cx88[0]: ioctl 0xc054561d (v4l2, rw, VIDIOC_G_TUNER)

all over the syslog.

Build issues include broken Makefiles (around 10-08) and missing header files.

Quoted 5 lines

Cheers,

Mike

hmm... would those be unresolved symbols in relation to dvb frontends?

This would be because we only recently merged the cvs trees, and the older builds dont include the dvb frontends in the v4l build. You can fix it by doing the following:

Be sure to "make clean" before and after each checkout -- it's necessary because of what the scripts below do to the build configuration.

Before running make (EACH TIME) do this:

1- edit v4l-dvb/v4l/scripts/merge-trees.sh ... delete everything BETWEEN these two lines... being sure to leave both of them intact.:

#!/bin/sh

...

patch -p2 <<'DIFF'

...do the same thing to v4l-dvb/v4l/scripts/unmerge-trees.sh

You will have to do it for each checkout, unfortunately. If you are using a more recent checkout (past few days) then you do not have to edit these files, as they have been deleted when the build was updated for the merger of v4l+dvb cvs.

2) make merge-trees

3) make

4) make install

...because of the "merge-trees" command, you'll have to run "make clean" each time before you cvs check out (cvs co) again.

I hope this works........

-Mike

#49 lore raw

Re: Gene's pcHDTV 3000 analog problem

On Wednesday 30 November 2005 23:00, Michael Krufky wrote:
cvs up -D 2005-10-15
was done inside the checkouts v4l-dvb dir
make (in the v4l subdir) clean, ok
make then blows up with:
root@coyote v4l]# make
make -C /lib/modules/2.6.14.3/build SUBDIRS=/usr/src/v4l-dvb/v4l modules make[1]: Entering directory /usr/src/linux-2.6.14.3' make[2]: *** No rule to make target /usr/src/v4l-dvb/v4l/video-buf.c', needed by `/usr/src/v4l-dvb/v4l/video-buf.o'. Stop. make[1]: *** [_module_/usr/src/v4l-dvb/v4l] Error 2
make[1]: Leaving directory `/usr/src/linux-2.6.14.3'
make: *** [default] Error 2

So what I figured to be a good starting point isn't.

So I stepped back to the parent v4l-dvb dir, and did a
cvs up -D 2005-11-1, which updated a whole bunch of stuff, and then I get this:
[root@coyote v4l-dvb]# cd v4l
[root@coyote v4l]# make clean
.version:1: *** missing separator. Stop.
[root@coyote v4l]# make
.version:1: *** missing separator. Stop.

So it looks as if I don't know what I'm doing (and I obviously don't) Where am I losing it here?

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#50 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 24 lines · Gene Heskett

On Wednesday 30 November 2005 23:00, Michael Krufky wrote:
cvs up -D 2005-10-15
was done inside the checkouts v4l-dvb dir
make (in the v4l subdir) clean, ok
make then blows up with:
root@coyote v4l]# make
make -C /lib/modules/2.6.14.3/build SUBDIRS=/usr/src/v4l-dvb/v4l modules make[1]: Entering directory /usr/src/linux-2.6.14.3' make[2]: *** No rule to make target /usr/src/v4l-dvb/v4l/video-buf.c', needed by `/usr/src/v4l-dvb/v4l/video-buf.o'. Stop. make[1]: *** [_module_/usr/src/v4l-dvb/v4l] Error 2
make[1]: Leaving directory `/usr/src/linux-2.6.14.3'
make: *** [default] Error 2

So what I figured to be a good starting point isn't.

So I stepped back to the parent v4l-dvb dir, and did a cvs up -D 2005-11-1, which updated a whole bunch of stuff, and then I get this:
[root@coyote v4l-dvb]# cd v4l
[root@coyote v4l]# make clean
.version:1: *** missing separator. Stop.
[root@coyote v4l]# make
.version:1: *** missing separator. Stop.

Quoted 5 lines

So it looks as if I don't know what I'm doing (and I obviously don't) Where am I losing it here?

I've been able to pin point the time of the fault to changes commited to v4l cvs at "2005-10-17 16:02."

In summary:

cvs as of 2005-10-17 16:01:
Analog TV works, as well as composite and s-video.

Changes made at 2005-10-17 16:02: bROKEN!!

cvs as of 2005-10-17 16:10:
Analog failes, s-video and composite work.

To test this, fron v4l-dvb/ ( not v4l-dvb/v4l/ ):

cvs up -D "2005-10-17 16:02"
make distclean
make
make install

Do a cold boot, and test.

cvs up -D "2005-10-17 16:01"
make distclean
make
make install

Do a cold boot and test.

Cheers,

Perry

#51 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 2 lines · Perry Gilfillan
Quoted 1 line · Gene Heskett

On Wednesday 30 November 2005 23:00, Michael Krufky wrote:

[...]

Quoted 29 lines

I've been able to pin point the time of the fault to changes commited to v4l cvs at "2005-10-17 16:02."

In summary:

cvs as of 2005-10-17 16:01:
Analog TV works, as well as composite and s-video.

Changes made at 2005-10-17 16:02: bROKEN!!

cvs as of 2005-10-17 16:10:
Analog failes, s-video and composite work.

To test this, fron v4l-dvb/ ( not v4l-dvb/v4l/ ):

cvs up -D "2005-10-17 16:02"
make distclean
make
make install

Do a cold boot, and test.

cvs up -D "2005-10-17 16:01"
make distclean
make
make install

Quoted 5 lines

Do a cold boot and test.

Cheers,

Perry

Unforch, it appears that Makefile is broken as its expecting to find a link named linux in the /usr/src dir that points at the present kernels tree. Per Linus's instructions that link was removed what, 2 years ago?

Anywho, its exiting rather quickly with this message:

[root@coyote v4l-dvb]# make distclean
make -C /usr/src/v4l-dvb/v4l distclean make[1]: Entering directory `/usr/src/v4l-dvb/v4l' find . -name '*.c' -type l -exec rm '{}' \; find . -name '*.h' -type l -exec rm '{}' \; rm -f *~ *.o *.ko *.mod.c
rm -f .version .*.o.flags .*.o.d .*.o.cmd .*.ko.cmd rm -rf .tmp_versions
make[1]: Leaving directory `/usr/src/v4l-dvb/v4l'

Which looks ok, so...

[root@coyote v4l-dvb]# make
make -C /usr/src/v4l-dvb/v4l
make[1]: Entering directory `/usr/src/v4l-dvb/v4l'
echo "No version yet."
No version yet.
uname -r|perl -ne 'if (/^([0-9]*)\.([0-9])*\.([0-9]*)(.*)$/) { printf ("VERSION=%s\nPATCHLEVEL:=%s\nSUBLEVEL:=%s\nKERNELRELEASE:=%s.%s.%s%s\n" ,$1,$2,$3,$1,$2,$3,$4); };' > ./.version
make[1]: Leaving directory `/usr/src/v4l-dvb/v4l'
make[1]: Entering directory `/usr/src/v4l-dvb/v4l'
creating symbolic links...
find: ../linux/drivers/usb/media: No such file or directory make[1]: *** [links] Error 1
make[1]: Leaving directory `/usr/src/v4l-dvb/v4l'
make: *** [all] Error 2

The ./v4l/.version file:
VERSION=2
PATCHLEVEL:=6
SUBLEVEL:=14
KERNELRELEASE:=2.6.14.3

which appears correct to me.

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#52 lore raw

Re: Gene's pcHDTV 3000 analog problem

In reply to Michael Krufky
Quoted 118 lines · Michael Krufky
Quoted 71 lines · Don Koch

On Wed, 30 Nov 2005 00:29:23 -0500

Quoted 47 lines · Michael Krufky
Quoted 17 lines · Gene Heskett

On Tuesday 29 November 2005 20:26, Michael Krufky wrote:

[...]

Quoted 6 lines

ll I can think of doing next is to have Gene, Don or Perry do a bisection test on our cvs repo.... checking out different cvs revisions until we can narrow it down to the day the problem patch was applied.

::sigh::

A sigh? More like an 'oh fudge' or whatever your fav expletive deleted is...

Quoted 2 lines

Who wants to do it? I'll give you detailed instructions if you're willing.

Quoted 7 lines

Can you farm it out, one set of patches to each of us? Or do you want to setup a seperate cvs tree based on the v4l code in 2.6.14.3, and incrementally patch it as we each report its still ok, until it breaks again? I think I'd prefer the latter so we are all near the same page even if it takes 3x longer to arrive at the answer. How many actual patches in terms of dependency groups are there? I know, I coulda went all week without asking that :(

Actually, cvs has a parameter that lets you specify cutoff dates...

This is what I am suggesting that you do... Base this on my previous cvs instructions....

reminder: http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS

so....

1st:

cvs -d :pserver:[email protected]:/cvs/video4linux login cvs -d :pserver:[email protected]:/cvs/video4linux co v4l-dvb cd v4l-dvb
make clean
make
make install

test

(you already did this - you said doesnt work)

[...]

Quoted 11 lines

cvs up -D 2005-10-15
make clean
make
make install

doesnt work? 1 week earlier:

cvs up -D 2005-10-07
make clean
make
make install

Let's put it this way: for me, 2005-10-10 doesn't work and anything earlier doesn't build. I've tried building against 2.6.15-rc3, 2.6.14 and 2.6.13. The card doesn't work against the built-in 2.6.13 code, but the tuner is sending bizarre stuff to it (channel 2 is *not* at 97.25 MHz). 2.6.14 kept spewing:

Quoted 11 lines

CORE IOCTL: 0xc054561d
cx88[0]: ioctl 0xc054561d (v4l2, rw, VIDIOC_G_TUNER)

all over the syslog.

Build issues include broken Makefiles (around 10-08) and missing header files.

Quoted 3 lines

Cheers,

Mike

hmm... would those be unresolved symbols in relation to dvb frontends?

This would be because we only recently merged the cvs trees, and the older builds dont include the dvb frontends in the v4l build. You can fix it by doing the following:

Be sure to "make clean" before and after each checkout -- it's necessary because of what the scripts below do to the build configuration.

Before running make (EACH TIME) do this:

1- edit v4l-dvb/v4l/scripts/merge-trees.sh ... delete everything
BETWEEN these two lines... being sure to leave both of them intact.:

#!/bin/sh

...

patch -p2 <<'DIFF'

...do the same thing to v4l-dvb/v4l/scripts/unmerge-trees.sh

You will have to do it for each checkout, unfortunately. If you are using a more recent checkout (past few days) then you do not have to edit these files, as they have been deleted when the build was updated for the merger of v4l+dvb cvs.

2) make merge-trees

Quoted 8 lines

3) make

4) make install

...because of the "merge-trees" command, you'll have to run "make clean" each time before you cvs check out (cvs co) again.

I hope this works........

Nope, see previous post, I can't get past this for say 2005-11-1:

Quoted 1 line

[root@coyote scripts]# vim merge-trees.sh

[root@coyote scripts]# vim unmerge-trees.sh
[root@coyote scripts]# cd ..
[root@coyote v4l]# make clean
.version:1: *** missing separator. Stop.
[root@coyote v4l]# make
.version:1: *** missing separator. Stop.
[root@coyote v4l]# make merge-trees
.version:1: *** missing separator. Stop.

So let me go back to 2005-10-1 & try that. But that ends the cvs up -D 2005-10-1 with:
Merging differences between 1.47 and 1.25 into merge-trees.sh
rcsmerge: warning: conflicts during merge
cvs update: conflicts found in v4l/scripts/merge-trees.sh
C v4l/scripts/merge-trees.sh
P v4l/scripts/prepare-ChangeLog.pl
P v4l/scripts/strip-trailing-whitespaces.sh
RCS file: /cvs/video4linux/v4l-dvb/v4l/scripts/Attic/unmerge-trees.sh,v retrieving revision 1.44
retrieving revision 1.23
Merging differences between 1.44 and 1.23 into unmerge-trees.sh
rcsmerge: warning: conflicts during merge
cvs update: conflicts found in v4l/scripts/unmerge-trees.sh C v4l/scripts/unmerge-trees.sh

& that bails out with the message:
[root@coyote v4l-dvb]# cd v4l/scripts
[root@coyote scripts]# vim unmerge-trees.sh
[root@coyote scripts]# vim merge-trees.sh
[root@coyote scripts]# cd ..
[root@coyote v4l]# make merge-trees
make: *** No rule to make target `merge-trees'. Stop. [root@coyote v4l]# make clean
rm -f *~ *.o *.ko *.mod.c
rm -f .version .*.o.flags .*.o.d .*.o.cmd .*.ko.cmd [root@coyote v4l]# make merge-trees
make: *** No rule to make target `merge-trees'. Stop. [root@coyote v4l]# make
make -C /lib/modules/2.6.14.3/build SUBDIRS=/usr/src/v4l-dvb/v4l modules make[1]: Entering directory /usr/src/linux-2.6.14.3' make[2]: *** No rule to make target /usr/src/v4l-dvb/v4l/video-buf.c', needed by `/usr/src/v4l-dvb/v4l/video-buf.o'. Stop. make[1]: *** [_module_/usr/src/v4l-dvb/v4l] Error 2
make[1]: Leaving directory `/usr/src/linux-2.6.14.3'
make: *** [default] Error 2

So this doesn't seem to be a productive effort at all, Michael. Or I'm doing things out of order as usual & need to start from scratch with a fresh checkout each time.

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#53 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 4 lines · Gene Heskett

So this doesn't seem to be a productive effort at all, Michael. Or I'm doing things out of order as usual & need to start from scratch with a fresh checkout each time.

Wipe out the cvs tree and do a fresh checkout each time.

-Mike

#54 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 8 lines · Michael Krufky
Quoted 3 lines · Gene Heskett

So this doesn't seem to be a productive effort at all, Michael. Or I'm doing things out of order as usual & need to start from scratch with a fresh checkout each time.

Wipe out the cvs tree and do a fresh checkout each time.

-Mike

Thats what I just did and got the broken makefile.

I just posted another msg that gives a blow by blow.

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#55 lore raw

Re: Gene's pcHDTV 3000 analog problem

In reply to Michael Krufky

On Wednesday 30 November 2005 23:00, Michael Krufky wrote:

And somebody in the Cc: list of this thread is using an SPI from spam.com and the confirm message contains no indication of who you are. I do NOT do sender verifications as it just seems to confirm I'm a good address to spam. I get several hundred a day the way it is. So if you aren't seeing my messages, only the replies to them, well... Tough.

[...]

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#56 lore raw

Re: Gene's pcHDTV 3000 analog problem

In reply to Michael Krufky
Quoted 17 lines · Michael Krufky
Quoted 10 lines · Kirk Lapray

I only run the cards in digital mode. I have no need to tune to any analog channels with them. When I was working on nxt200x I was able to tune both analog and digital channels on the HDTV Wonder, but I have never tried to tune to any analog channels using the HD3000. This was using kernel 2.6.13 and the cvs v4l and dvb trees.

I am not sure when I will have some time to test 2.6.15-rc3, but if I get some time I will try analog support on this and my current setup.

Kirk

Kirk-

Please test your cards in analog mode, or I fear that we may have to cause NXT200X to depend on BROKEN.

Quoted 7 lines

Here's why....

A few months ago, as we added the nxt200x module to cvs, I told you that with nxt200x loaded into memory, I had some extra devices showing up on my i2c bus. At the time, I was using another cx88 card, and it did not use the nxt200x module itself, although it loads up into memory automatically by the cx88-dvb module.

It was not being loaded by a modprobe cx88-dvb here, see my lsmod listing from earlier today.

Quoted 7 lines

You told me that there was some code in nxt200x module that somehow opens up a channel to hidden i2c devices. Why would this code affect my system if my device is not using the nxt200x module?

Is there code being run at nxt200x module load that is causing this BEFORE cx88-dvb calls nxt200x_attach() ?

Quoted 6 lines

It seems that Gene, Perry and Don are having problems with their analog tuners (they each have pcHDTV 3000) ever since nxt200x got added.

Gene, Perry and Don - What happens if you have the cx88 module loaded, but you do NOT load up cx88-dvb (nxt200x will not be loaded) ... Does the problem persist?

Having done an rmmod of everything related, installling the cvs but no reboot, an attempt to modprobe cx88 back into the system gets a module not found message. So maybe I have to reboot anyway, brb.

Quoted 10 lines

You do not need cx88-dvb to view analog television.

Kirk, we need a control group! Please test analog on both boards.

Kirk, there is a thread on the v4l/dvb mailing lists right now about an i2c gate dealing with Hauppauge cards and cx22702 frontend. What Steve Toth has described about this 'i2c gate' is starting to sound similar to what you mentioned about making hidden i2c devices visible.

Quoted 7 lines

I'm getting the feeling that nxt200x is indeed the problem.

Gene, Perry and Don .... Another thing you can try -- Once again, install v4l-dvb cvs on top of your running kernel, but this time, before compiling, edit v4l-dvb/v4l/Makefile , and remove the line:

EXTRA_CFLAGS += -DHAVE_NXT200X=1

I took these out before I installed the cvs to 2.6.14.3. So here goes a reboot.

Quoted 6 lines

... This line appears twice, you only need to remove the top one, as it pertains to the cx88 card, although it is safe to remove both for the purposes of this test.

If this fixes your problem, then we know that nxt200x is the cause.

As I said above, nxt200x wasn't being loaded by the modprobe cx88-dvb statement. I'm now rebooted after having done a cvs install after modifying the v4l/Makefile, and its now broken again. So I'm not convinced that the nxt200x code has anything to do with my problem, at least not yet. Now I have to see if I can recover by blowing away the /lib/modules dir and doing another make modules_install in the 2.6.14.3 tree, plus a cold reboot. Nah, to heck with it, just rerun my makeit script again, only takes about 9 minutes for a clean rebuild. But this cold reboot will take about 20 minutes, I have umpty gigabytes of partitions that will e2fsck, /dev/hdd3 (180GB for amandas virtual tapes) taking at least 15 minutes, much of it with the progress bar stuck at about 13%. And another 80 GB partition is set for this round too. It sticks for 5 minutes at about 30%, e2fsck, version 1.35, needs help IMNSHO. But thats what we have I guess...

Anyway, rebooted, and it works again after modprobe cx88-dvb from a cli. I let it run about 3 minutes, and got this in the messages log:

Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc04c561a
Nov 29 22:56:18 coyote last message repeated 3 times
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x800476c6
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0045627
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x80045626
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc04c561a
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x80085617
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x40085618
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x4054561e
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0045627
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x80045626
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc04c561a
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x80085617
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x40085618
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x4054561e
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x402c5639
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc02c5638
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 22:56:18 coyote last message repeated 2 times
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 22:56:49 coyote last message repeated 907 times
Nov 29 22:57:50 coyote last message repeated 1828 times
Nov 29 22:58:46 coyote last message repeated 1703 times
Nov 29 22:58:46 coyote kernel: CORE IOCTL: 0xc008561c

Is this related to the memory problems you commented on a couple of times? The system seems stable though. And I've watched tv by the hour many times, not knowing I should be tailing the log.

The log size so far for this week is ~500k.

Quoted 7 lines
(quoted diff) +0 -2
-Mike
 
-

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#57 lore raw

Re: Gene's pcHDTV 3000 analog problem

Gene,

Some points:

1) Is it a 64bit machine?

2) The problem also occurs at vanilla 2.6.15-rc3?

Cheers,
Mauro.

_______________________________________________________
Yahoo! doce lar. Fa�a do Yahoo! sua homepage.
http://br.yahoo.com/homepageset.html

#58 lore raw

Re: Gene's pcHDTV 3000 analog problem

Quoted 5 lines · Mauro Carvalho Chehab

Gene,

Some points:

1) Is it a 64bit machine?

No, x86-32, T.B.E. an Xp2800 Athlon, gig of ram.

Quoted 1 line

2) The problem also occurs at vanilla 2.6.15-rc3?

It works ok as late as 2.6.14.3, fails on 2.6.15-rc2 & rc3. I never build rc1's.

Quoted 17 lines

Cheers,
Mauro.

_______________________________________________________
Yahoo! doce lar. Faça do Yahoo! sua homepage.
http://br.yahoo.com/homepageset.html

--
video4linux-list mailing list
Unsubscribe mailto:[email protected]?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Signature · 8 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
#59 lore raw

Re: Gene's pcHDTV 3000 analog problem

In reply to Gene Heskett

Gene,

Two questions:

1) Is it a 64bit machine?

2) please try to modprobe tuner with offset=100, then with offset=-100. offset option will force tuner to make a freq offset. It might be the case.

Cheers,
Mauro.

Em Ter, 2005-11-29 às 23:06 -0500, Gene Heskett escreveu:

Quoted 159 lines · Michael Krufky
Quoted 17 lines
Quoted 10 lines · Kirk Lapray

I only run the cards in digital mode. I have no need to tune to any analog channels with them. When I was working on nxt200x I was able to tune both analog and digital channels on the HDTV Wonder, but I have never tried to tune to any analog channels using the HD3000. This was using kernel 2.6.13 and the cvs v4l and dvb trees.

I am not sure when I will have some time to test 2.6.15-rc3, but if I get some time I will try analog support on this and my current setup.

Kirk

Kirk-

Please test your cards in analog mode, or I fear that we may have to cause NXT200X to depend on BROKEN.

Quoted 7 lines

Here's why....

A few months ago, as we added the nxt200x module to cvs, I told you that with nxt200x loaded into memory, I had some extra devices showing up on my i2c bus. At the time, I was using another cx88 card, and it did not use the nxt200x module itself, although it loads up into memory automatically by the cx88-dvb module.

It was not being loaded by a modprobe cx88-dvb here, see my lsmod listing from earlier today.

Quoted 7 lines

You told me that there was some code in nxt200x module that somehow opens up a channel to hidden i2c devices. Why would this code affect my system if my device is not using the nxt200x module?

Is there code being run at nxt200x module load that is causing this BEFORE cx88-dvb calls nxt200x_attach() ?

Quoted 6 lines

It seems that Gene, Perry and Don are having problems with their analog tuners (they each have pcHDTV 3000) ever since nxt200x got added.

Gene, Perry and Don - What happens if you have the cx88 module loaded, but you do NOT load up cx88-dvb (nxt200x will not be loaded) ... Does the problem persist?

Having done an rmmod of everything related, installling the cvs but no reboot, an attempt to modprobe cx88 back into the system gets a module not found message. So maybe I have to reboot anyway, brb.

Quoted 10 lines

You do not need cx88-dvb to view analog television.

Kirk, we need a control group! Please test analog on both boards.

Kirk, there is a thread on the v4l/dvb mailing lists right now about an i2c gate dealing with Hauppauge cards and cx22702 frontend. What Steve Toth has described about this 'i2c gate' is starting to sound similar to what you mentioned about making hidden i2c devices visible.

Quoted 7 lines

I'm getting the feeling that nxt200x is indeed the problem.

Gene, Perry and Don .... Another thing you can try -- Once again, install v4l-dvb cvs on top of your running kernel, but this time, before compiling, edit v4l-dvb/v4l/Makefile , and remove the line:

EXTRA_CFLAGS += -DHAVE_NXT200X=1

I took these out before I installed the cvs to 2.6.14.3. So here goes a reboot.

Quoted 6 lines

... This line appears twice, you only need to remove the top one, as it pertains to the cx88 card, although it is safe to remove both for the purposes of this test.

If this fixes your problem, then we know that nxt200x is the cause.

As I said above, nxt200x wasn't being loaded by the modprobe cx88-dvb statement. I'm now rebooted after having done a cvs install after modifying the v4l/Makefile, and its now broken again. So I'm not convinced that the nxt200x code has anything to do with my problem, at least not yet. Now I have to see if I can recover by blowing away the /lib/modules dir and doing another make modules_install in the 2.6.14.3 tree, plus a cold reboot. Nah, to heck with it, just rerun my makeit script again, only takes about 9 minutes for a clean rebuild. But this cold reboot will take about 20 minutes, I have umpty gigabytes of partitions that will e2fsck, /dev/hdd3 (180GB for amandas virtual tapes) taking at least 15 minutes, much of it with the progress bar stuck at about 13%. And another 80 GB partition is set for this round too. It sticks for 5 minutes at about 30%, e2fsck, version 1.35, needs help IMNSHO. But thats what we have I guess...

Anyway, rebooted, and it works again after modprobe cx88-dvb from a cli. I let it run about 3 minutes, and got this in the messages log:

Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc04c561a
Nov 29 22:56:18 coyote last message repeated 3 times
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x800476c6
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0045627
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x80045626
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc04c561a
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x80085617
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x40085618
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x4054561e
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0045627
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x80045626
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc04c561a
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x80085617
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x40085618
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x4054561e
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0x402c5639
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561b
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc02c5638
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 22:56:18 coyote last message repeated 2 times
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc0445624
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc008561c
Nov 29 22:56:18 coyote kernel: CORE IOCTL: 0xc054561d
Nov 29 22:56:49 coyote last message repeated 907 times
Nov 29 22:57:50 coyote last message repeated 1828 times
Nov 29 22:58:46 coyote last message repeated 1703 times
Nov 29 22:58:46 coyote kernel: CORE IOCTL: 0xc008561c

Is this related to the memory problems you commented on a couple of times? The system seems stable though. And I've watched tv by the hour many times, not knowing I should be tailing the log.

The log size so far for this week is ~500k.

Quoted 7 lines
(quoted diff) +0 -2
-Mike
 
-

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Signature · 10 lines
Mauro Carvalho Chehab <[email protected]>


	

	
		
_______________________________________________________ 
Yahoo! doce lar. Faça do Yahoo! sua homepage. 
http://br.yahoo.com/homepageset.html