bttv problems in 2.4.0/2.4.1

bttv problems in 2.4.0/2.4.1

[Matthew Gabeler-Lee]

In 2.4.0 and 2.4.1, when I try to load the bttv driver, one of two
things happens: the system hangs (even alt-sysrq doesn't work!), or the
system powers off by itself (ATX mobo).  Instant power-off usually
happens after a soft reboot (init 6), while it usually hangs up after a
hard reboot (power cycling).

When it hangs, I noticed a very strange thing.  If I push the power
on/off button briefly, it un-hangs and seems to proceed as normal.  The
kernel does report an APIC error on each cpu (dual p3 700 system) when
this happens.

These errors all occur in the same way (as near as I can tell) in
kernels 2.4.0 and 2.4.1, using bttv drivers 0.7.50 (incl. w/ kernel),
0.7.53, and 0.7.55.

I am currently using 2.4.0-test10 with bttv 0.7.47, which works fine.

I have sent all this info to Gerd Knorr but, as far as I know, he hasn't
been able to track down the bug yet.  I thought that by posting here,
more eyes might at least make more reports of similar situations that
might help track down the problem.

PS: I'm not on the linux-kernel list, so please CC replies to me.

-- 
	-Matt

Today's weirdness is tomorrow's reason why.
                -- Hunter S. Thompson


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

Re: bttv problems in 2.4.0/2.4.1

[John Jasen]

On Tue, 30 Jan 2001, Matthew Gabeler-Lee wrote:

> These errors all occur in the same way (as near as I can tell) in
> kernels 2.4.0 and 2.4.1, using bttv drivers 0.7.50 (incl. w/ kernel),
> 0.7.53, and 0.7.55.
>
> I am currently using 2.4.0-test10 with bttv 0.7.47, which works fine.
>
> I have sent all this info to Gerd Knorr but, as far as I know, he hasn't
> been able to track down the bug yet.  I thought that by posting here,
> more eyes might at least make more reports of similar situations that
> might help track down the problem.

Try flipping the card into a different slot. A lot of the cards
exceptionally do not like IRQ/DMA sharing, and a lot of the motherboards
share them between different slots.

--
-- John E. Jasen (jjasen1@umbc.edu)
-- In theory, theory and practise are the same. In practise, they aren't.

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

Re: bttv problems in 2.4.0/2.4.1

[Matthew Gabeler-Lee]

On Tue, 30 Jan 2001, John Jasen wrote:

> On Tue, 30 Jan 2001, Matthew Gabeler-Lee wrote:
>
> > These errors all occur in the same way (as near as I can tell) in
> > kernels 2.4.0 and 2.4.1, using bttv drivers 0.7.50 (incl. w/ kernel),
> > 0.7.53, and 0.7.55.
> >
> > I am currently using 2.4.0-test10 with bttv 0.7.47, which works fine.
> >
> > I have sent all this info to Gerd Knorr but, as far as I know, he hasn't
> > been able to track down the bug yet.  I thought that by posting here,
> > more eyes might at least make more reports of similar situations that
> > might help track down the problem.
>
> Try flipping the card into a different slot. A lot of the cards
> exceptionally do not like IRQ/DMA sharing, and a lot of the motherboards
> share them between different slots.

I will try this, but my card has (and does) worked with irq sharing for
a long time.  Its entry in /proc/interrupts:
  9:     164935     165896   IO-APIC-level  acpi, bttv
I find it strange that a driver that had worked with shared interrupts
for a long time would suddenly cease to function with shared interrupts,
and would consider this a bug.  I will try changing the slot, but
getting it to not share interrupts will be difficult considering the
number of pci devices I have.

-- 
	-Matt

Today's weirdness is tomorrow's reason why.
		-- Hunter S. Thompson

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

Re: bttv problems in 2.4.0/2.4.1

[Gerd Knorr]

> > > I have sent all this info to Gerd Knorr but, as far as I know, he hasn't
> > > been able to track down the bug yet.  I thought that by posting here,
> > > more eyes might at least make more reports of similar situations that
> > > might help track down the problem.
> >
> > Try flipping the card into a different slot. A lot of the cards
> > exceptionally do not like IRQ/DMA sharing, and a lot of the motherboards
> > share them between different slots.
> 
> I will try this, but my card has (and does) worked with irq sharing for
> a long time.  Its entry in /proc/interrupts:
>   9:     164935     165896   IO-APIC-level  acpi, bttv
                                              ^^^^
What happens with acpi disabled?  The power-down at boot could be caused by
the acpi power management maybe ...

  Gerd

-- 
Get back there in front of the computer NOW. Christmas can wait.
	-- Linus "the Grinch" Torvalds,  24 Dec 2000 on linux-kernel
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: bttv problems in 2.4.0/2.4.1

[Prasanna P Subash]

[-- Attachment #1: Type: text/plain, Size: 1820 bytes --]


My bttv is at IRQ 3 and it still hangs the machine :(
I dont even have acpi built in.

btw I am testing with 2.4.1-pre9
-- 
Prasanna Subash   ---   psubash@turbolinux.com   ---     TurboLinux, INC
------------------------------------------------------------------------
Linux, the choice          | When the sun shineth, make hay.   -- John
of a GNU generation   -o)  | Heywood 
Kernel 2.4.0-ac4      /\\  | 
on a i686            _\\_v | 
                           | 
------------------------------------------------------------------------

On Wed, Jan 31, 2001 at 08:41:45AM +0100, Gerd Knorr wrote:
> > > > I have sent all this info to Gerd Knorr but, as far as I know, he hasn't
> > > > been able to track down the bug yet.  I thought that by posting here,
> > > > more eyes might at least make more reports of similar situations that
> > > > might help track down the problem.
> > >
> > > Try flipping the card into a different slot. A lot of the cards
> > > exceptionally do not like IRQ/DMA sharing, and a lot of the motherboards
> > > share them between different slots.
> > 
> > I will try this, but my card has (and does) worked with irq sharing for
> > a long time.  Its entry in /proc/interrupts:
> >   9:     164935     165896   IO-APIC-level  acpi, bttv
>                                               ^^^^
> What happens with acpi disabled?  The power-down at boot could be caused by
> the acpi power management maybe ...
> 
>   Gerd
> 
> -- 
> Get back there in front of the computer NOW. Christmas can wait.
> 	-- Linus "the Grinch" Torvalds,  24 Dec 2000 on linux-kernel
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> Please read the FAQ at http://www.tux.org/lkml/



[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: bttv problems in 2.4.0/2.4.1

[Prasanna P Subash]

[-- Attachment #1: Type: text/plain, Size: 1883 bytes --]

I have experienced similar issues with 2.4.0 and its test. I have a bttv848 chipset.
I even tried compiling in kdb as a part of the kernel to see if it oopses, but no luck.

I will try trying 0.7.47 today.

this works on 2.2.16, last time i tried.

-- 
Prasanna Subash   ---   psubash@turbolinux.com   ---     TurboLinux, INC
------------------------------------------------------------------------
Linux, the choice          | "You've got to think about tomorrow!" 
of a GNU generation   -o)  | "TOMORROW!  I haven't even prepared for 
Kernel 2.4.0-ac4      /\\  | yesterday* yet!" 
on a i686            _\\_v | 
                           | 
------------------------------------------------------------------------

On Tue, Jan 30, 2001 at 07:53:11PM -0500, John Jasen wrote:
> On Tue, 30 Jan 2001, Matthew Gabeler-Lee wrote:
> 
> > These errors all occur in the same way (as near as I can tell) in
> > kernels 2.4.0 and 2.4.1, using bttv drivers 0.7.50 (incl. w/ kernel),
> > 0.7.53, and 0.7.55.
> >
> > I am currently using 2.4.0-test10 with bttv 0.7.47, which works fine.
> >
> > I have sent all this info to Gerd Knorr but, as far as I know, he hasn't
> > been able to track down the bug yet.  I thought that by posting here,
> > more eyes might at least make more reports of similar situations that
> > might help track down the problem.
> 
> Try flipping the card into a different slot. A lot of the cards
> exceptionally do not like IRQ/DMA sharing, and a lot of the motherboards
> share them between different slots.
> 
> --
> -- John E. Jasen (jjasen1@umbc.edu)
> -- In theory, theory and practise are the same. In practise, they aren't.
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> Please read the FAQ at http://www.tux.org/lkml/



[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: bttv problems in 2.4.0/2.4.1

[adrian]

Hmmm,

   I have a bt848 based video capture card, and get near the same results:
2.4.0-test10 through 2.4.1 all lock when i2c registers the device.  The
card has its own interrupt.  With 2.2.18, the card initialized and the
kernel continued to boot.  Interesting.

Regards,
Adrian


On Tue, 30 Jan 2001, Prasanna P Subash wrote:

> I have experienced similar issues with 2.4.0 and its test. I have a bttv848 chipset.
> I even tried compiling in kdb as a part of the kernel to see if it oopses, but no luck.
>
> I will try trying 0.7.47 today.
>
> this works on 2.2.16, last time i tried.
>
> --
> Prasanna Subash   ---   psubash@turbolinux.com   ---     TurboLinux, INC
> ------------------------------------------------------------------------
> Linux, the choice          | "You've got to think about tomorrow!"
> of a GNU generation   -o)  | "TOMORROW!  I haven't even prepared for
> Kernel 2.4.0-ac4      /\\  | yesterday* yet!"
> on a i686            _\\_v |
>                            |
> ------------------------------------------------------------------------
>
> On Tue, Jan 30, 2001 at 07:53:11PM -0500, John Jasen wrote:
> > On Tue, 30 Jan 2001, Matthew Gabeler-Lee wrote:
> >
> > > These errors all occur in the same way (as near as I can tell) in
> > > kernels 2.4.0 and 2.4.1, using bttv drivers 0.7.50 (incl. w/ kernel),
> > > 0.7.53, and 0.7.55.
> > >
> > > I am currently using 2.4.0-test10 with bttv 0.7.47, which works fine.
> > >
> > > I have sent all this info to Gerd Knorr but, as far as I know, he hasn't
> > > been able to track down the bug yet.  I thought that by posting here,
> > > more eyes might at least make more reports of similar situations that
> > > might help track down the problem.
> >
> > Try flipping the card into a different slot. A lot of the cards
> > exceptionally do not like IRQ/DMA sharing, and a lot of the motherboards
> > share them between different slots.
> >
> > --
> > -- John E. Jasen (jjasen1@umbc.edu)
> > -- In theory, theory and practise are the same. In practise, they aren't.
> >
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@vger.kernel.org
> > Please read the FAQ at http://www.tux.org/lkml/
>
>
>

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

Re: bttv problems in 2.4.0/2.4.1

[Matthew Gabeler-Lee]

On Tue, 30 Jan 2001, adrian wrote:

>    I have a bt848 based video capture card, and get near the same results:
> 2.4.0-test10 through 2.4.1 all lock when i2c registers the device.  The
> card has its own interrupt.  With 2.2.18, the card initialized and the
> kernel continued to boot.  Interesting.

2 questions:
What card in particular do you have?
What version of the bttv drivers were you using in 2.4.0-test10?
It comes with 0.7.38; did you patch it to a higher version?

-- 
	-Matt

Today's weirdness is tomorrow's reason why.
		-- Hunter S. Thompson

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

Re: bttv problems in 2.4.0/2.4.1

[adrian]

On Tue, 30 Jan 2001, Matthew Gabeler-Lee wrote:

> On Tue, 30 Jan 2001, adrian wrote:
>
> >    I have a bt848 based video capture card, and get near the same results:
> > 2.4.0-test10 through 2.4.1 all lock when i2c registers the device.  The
> > card has its own interrupt.  With 2.2.18, the card initialized and the
> > kernel continued to boot.  Interesting.
>
> 2 questions:
> What card in particular do you have?
> What version of the bttv drivers were you using in 2.4.0-test10?
> It comes with 0.7.38; did you patch it to a higher version?
>
> --
> 	-Matt
>

The card is a video only capture that came with a camera (and has a
connector to power that camera next to the video connector).  Hauppauge is
silkscreened on the PCB, along with Axiom Design.  I didn't buy it, so all
I have to go off of is what's on the card.  So here's all of it.

Sticker on back: 58051 REV A 231383
On Bt chip: Bt848kpf video decoder 255 9637

As far as the bttv driver version, I've always used the ones that came
originally with that kernel version.

This is what lspci says:
00:0f.0 Multimedia video controller: Brooktree Corporation Bt848 (rev 11)
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr- Stepping- SERR- FastB2B-
        Status: Cap- 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
        Latency: 16 min, 40 max, 64 set
        Interrupt: pin A routed to IRQ 9
        Region 0: Memory at eddfe000 (32-bit, prefetchable)

A cat of /proc/pci reports REV 17 instead of 11.

Regards,
Adrian


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

Re: bttv problems in 2.4.0/2.4.1

[Gerd Knorr]

> The card is a video only capture that came with a camera (and has a
> connector to power that camera next to the video connector).

Sure the box is really dead?  These very cheap cards with just the bt848
and nothing else often have a non-working i2c bus (because they have no
chips connected to it, maybe even the i2c pins unconnected).  The
i2c initialization takes forever (minutes) on these boards due to timeouts
and retries of the i2c code unless you tell the i2c layer it should make
some sanity checks on the i2c bus first (options i2c-algo-bit bit_test=1).

  Gerd

-- 
Get back there in front of the computer NOW. Christmas can wait.
	-- Linus "the Grinch" Torvalds,  24 Dec 2000 on linux-kernel
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: bttv problems in 2.4.0/2.4.1

[adrian]

On 31 Jan 2001, Gerd Knorr wrote:

> > The card is a video only capture that came with a camera (and has a
> > connector to power that camera next to the video connector).
>
> Sure the box is really dead?  These very cheap cards with just the bt848
[snip]
> some sanity checks on the i2c bus first (options i2c-algo-bit bit_test=1).
>
>   Gerd
>

Yup, the box was just waiting for the timeout.  Setting that option allows
booting to continue regularly (i2c reports the device as being "busy"
during bootup).

Adrian


PS.  I think that card has gone bad or the drivers don't support it
correctly, since in both Windows and Linux machines it now gives scrambled
video.





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

Re: bttv problems in 2.4.0/2.4.1 PAL_BG

[archan]

I am using "Pixel View TV tuner card" based on "bttv". It works perfect 
in Windows with default TV application, and also responding well in 
Linux 2.2.17 and 2.4.0-test10 kernel. The device is getting detected 
perfectly by 2.4 kernel but I could not be able to check whether the 
card in 2.4 kernel is responding on PAL-BG signal (here, my frequency 
table is PAL-BG, country India) as none of the Linux apps (xawtv, 
cabletv) are responding positively.

archan
devrootp@bigfoot.com
archanp@bigfoot.com

Prasanna P Subash wrote:

> I have experienced similar issues with 2.4.0 and its test. I have a bttv848 chipset.
> I even tried compiling in kdb as a part of the kernel to see if it oopses, but no luck.
> 
> I will try trying 0.7.47 today.
> 
> this works on 2.2.16, last time i tried.
> 


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

Re: bttv problems in 2.4.0/2.4.1 PAL_BG

[John Jasen]

On Wed, 31 Jan 2001, archan wrote:

> I am using "Pixel View TV tuner card" based on "bttv". It works perfect
> in Windows with default TV application, and also responding well in
> Linux 2.2.17 and 2.4.0-test10 kernel. The device is getting detected
> perfectly by 2.4 kernel but I could not be able to check whether the
> card in 2.4 kernel is responding on PAL-BG signal (here, my frequency
> table is PAL-BG, country India) as none of the Linux apps (xawtv,
> cabletv) are responding positively.

I think I ended up trying the bttv 0.8 drivers, and maybe video4linux2,
nowe that I think about it.

I'll have to doublecheck and get back to you on that.

--
-- John E. Jasen (jjasen1@umbc.edu)
-- In theory, theory and practise are the same. In practise, they aren't.

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

Re: bttv problems in 2.4.0/2.4.1 PAL_BG

[archan]

ok!!

archan
devrootp@bigfoot.com
archanp@bigfoot.com

John Jasen wrote:

> On Wed, 31 Jan 2001, archan wrote:
> 
>> I am using "Pixel View TV tuner card" based on "bttv". It works perfect
>> in Windows with default TV application, and also responding well in
>> Linux 2.2.17 and 2.4.0-test10 kernel. The device is getting detected
>> perfectly by 2.4 kernel but I could not be able to check whether the
>> card in 2.4 kernel is responding on PAL-BG signal (here, my frequency
>> table is PAL-BG, country India) as none of the Linux apps (xawtv,
>> cabletv) are responding positively.
> 
> 
> I think I ended up trying the bttv 0.8 drivers, and maybe video4linux2,
> nowe that I think about it.
> 
> I'll have to doublecheck and get back to you on that.

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

Re: bttv problems in 2.4.0/2.4.1

[Dr. Kelsey Hudson]

Do you have framebuffer console compiled into your kernel? I noticed
similar behavior on my system when I had framebuffer console compiled in,
ACPI or APM (cant remember which, probably ACPI) compiled in, and bttv as
modules. System would power off when ACPI was loaded. Other times it would
do other stupid things like hang abruptly for no apparent reason.

On Tue, 30 Jan 2001, Matthew Gabeler-Lee wrote:

> In 2.4.0 and 2.4.1, when I try to load the bttv driver, one of two
> things happens: the system hangs (even alt-sysrq doesn't work!), or the
> system powers off by itself (ATX mobo).  Instant power-off usually
> happens after a soft reboot (init 6), while it usually hangs up after a
> hard reboot (power cycling).
> 
> When it hangs, I noticed a very strange thing.  If I push the power
> on/off button briefly, it un-hangs and seems to proceed as normal.  The
> kernel does report an APIC error on each cpu (dual p3 700 system) when
> this happens.
> 
> These errors all occur in the same way (as near as I can tell) in
> kernels 2.4.0 and 2.4.1, using bttv drivers 0.7.50 (incl. w/ kernel),
> 0.7.53, and 0.7.55.
> 
> I am currently using 2.4.0-test10 with bttv 0.7.47, which works fine.
> 
> I have sent all this info to Gerd Knorr but, as far as I know, he hasn't
> been able to track down the bug yet.  I thought that by posting here,
> more eyes might at least make more reports of similar situations that
> might help track down the problem.
> 
> PS: I'm not on the linux-kernel list, so please CC replies to me.
> 
> 

-- 
 Kelsey Hudson                                           khudson@ctica.com 
 Software Engineer
 Compendium Technologies, Inc                               (619) 725-0771
---------------------------------------------------------------------------     

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

Re: bttv problems in 2.4.0/2.4.1

[Matthew Gabeler-Lee]

On Fri, 9 Feb 2001, Dr. Kelsey Hudson wrote:

> Do you have framebuffer console compiled into your kernel? I noticed
> similar behavior on my system when I had framebuffer console compiled in,
> ACPI or APM (cant remember which, probably ACPI) compiled in, and bttv as
> modules. System would power off when ACPI was loaded. Other times it would
> do other stupid things like hang abruptly for no apparent reason.

My video card isn't really supported by the framebuffer stuff, at least
not for accelleration, so I don't have framebuffer support.  I have been
experiencing abrupt hangs recently too ...

-- 
	-Matt

If you suspect a man, don't employ him.

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

Re: bttv problems in 2.4.0/2.4.1

[Matthew Gabeler-Lee]

Disabling ACPI (all of power management, really.  SMP so no APM) seems
to have made it work in 2.4.1 for me.

On Fri, 9 Feb 2001, Dr. Kelsey Hudson wrote:

> Do you have framebuffer console compiled into your kernel? I noticed
> similar behavior on my system when I had framebuffer console compiled in,
> ACPI or APM (cant remember which, probably ACPI) compiled in, and bttv as
> modules. System would power off when ACPI was loaded. Other times it would
> do other stupid things like hang abruptly for no apparent reason.

-- 
	-Matt

If you suspect a man, don't employ him.

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