Kernel 2.6.29 breaks DVB-T ASUSTeK Tiger LNA Hybrid Capture Device

DISCUSSION Ralph 4 participants 10 messages Apr 3, 2009 – Apr 23, 2009 1 message not in the archive Media (V4L/DVB)

Share & tools

Ralph

author

2009-04-03 21:15 UTC #0 lore raw

ASUSTeK Tiger LNA Hybrid Capture Device PCI - Analog/DVB-T card Multimedia controller: Philips Semiconductors SAA7131/SAA7133/SAA7135 Video Broadcast Decoder (rev d1)

Works perfectly with kernel 2.6.28.4 (or older).
Recently, I have switched to 2.6.29 (same .config as 2.6.28.4) and now, at boot
time, I get the message:

IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs

Signal strength is very low and Kaffeine is unable to tune in any channel. Same problem with kernel 2.6.29.1

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

Messages from /var/log/dmesg

saa7134 0000:03:0a.0: PCI INT A -> Link[APC3] -> GSI 18 (level, low) -> \
IRQ 18
saa7133[0]: found at 0000:03:0a.0, rev: 209, irq: 18, latency: 32, mmio: \
0xfdefe000
saa7133[0]: subsystem: 1043:4871, board: ASUS P7131 4871 \
[card=111,autodetected]
saa7133[0]: board init: gpio is 0
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]: i2c eeprom 00: 43 10 71 48 54 20 1c 00 43 43 a9 1c 55 d2 b2 92 saa7133[0]: i2c eeprom 10: ff ff ff 0f ff 20 ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 20: 01 40 01 02 03 00 01 03 08 ff 00 cf ff ff ff ff saa7133[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 40: ff 21 00 c2 96 10 03 22 15 50 ff ff ff ff ff ff saa7133[0]: i2c eeprom 50: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 60: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 70: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff tuner' 2-004b: chip found @ 0x96 (saa7133[0])
tda829x 2-004b: setting tuner address to 61
tda829x 2-004b: type set to tda8290+75a
saa7133[0]: registered device video0 [v4l2]
saa7133[0]: registered device vbi0
dvb_init() allocating 1 frontend
DVB: registering new adapter (saa7133[0])
DVB: registering adapter 0 frontend -32769 (Philips TDA10046H DVB-T)...
tda1004x: setting up plls for 48MHz sampling clock
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: trying to boot from eeprom
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: waiting for firmware upload...
saa7134 0000:03:0a.0: firmware: requesting dvb-fe-tda10046.fw
tda1004x: found firmware revision 29 -- ok
saa7134 ALSA driver for DMA sound loaded
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]/alsa: saa7133[0] at 0xfdefe000 irq 18 registered as card -1

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

hermann pitton

2009-04-04 00:50 UTC #1 lore raw

Hi Ralph,

Am Freitag, den 03.04.2009, 20:49 +0000 schrieb Ralph:

Quoted 54 lines

ASUSTeK Tiger LNA Hybrid Capture Device PCI - Analog/DVB-T card Multimedia controller: Philips Semiconductors SAA7131/SAA7133/SAA7135 Video Broadcast Decoder (rev d1)
Works perfectly with kernel 2.6.28.4 (or older).
Recently, I have switched to 2.6.29 (same .config as 2.6.28.4) and now, at boot
time, I get the message:
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs
Signal strength is very low and Kaffeine is unable to tune in any channel. Same problem with kernel 2.6.29.1
-------------------------------------
Messages from /var/log/dmesg
saa7134 0000:03:0a.0: PCI INT A -> Link[APC3] -> GSI 18 (level, low) -> \
IRQ 18
saa7133[0]: found at 0000:03:0a.0, rev: 209, irq: 18, latency: 32, mmio: \
0xfdefe000
saa7133[0]: subsystem: 1043:4871, board: ASUS P7131 4871 \
[card=111,autodetected]
saa7133[0]: board init: gpio is 0
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]: i2c eeprom 00: 43 10 71 48 54 20 1c 00 43 43 a9 1c 55 d2 b2 92 saa7133[0]: i2c eeprom 10: ff ff ff 0f ff 20 ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 20: 01 40 01 02 03 00 01 03 08 ff 00 cf ff ff ff ff saa7133[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 40: ff 21 00 c2 96 10 03 22 15 50 ff ff ff ff ff ff saa7133[0]: i2c eeprom 50: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 60: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 70: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff tuner' 2-004b: chip found @ 0x96 (saa7133[0])
tda829x 2-004b: setting tuner address to 61
tda829x 2-004b: type set to tda8290+75a
saa7133[0]: registered device video0 [v4l2]
saa7133[0]: registered device vbi0
dvb_init() allocating 1 frontend
DVB: registering new adapter (saa7133[0])
DVB: registering adapter 0 frontend -32769 (Philips TDA10046H DVB-T)...
tda1004x: setting up plls for 48MHz sampling clock
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: trying to boot from eeprom

Quoted 9 lines

tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: waiting for firmware upload...
saa7134 0000:03:0a.0: firmware: requesting dvb-fe-tda10046.fw
tda1004x: found firmware revision 29 -- ok
saa7134 ALSA driver for DMA sound loaded
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]/alsa: saa7133[0] at 0xfdefe000 irq 18 registered as card -1

thanks for your report, as announced previously, I unfortunately did not have time to run with latest always ... (guess why ...)

The driver always worked with shared IRQs, if not, it was always a limitation of certain hardware or mostly in some combination with binary only drivers.

If the above should be the case in general now, and not only caused by some blacklist, no print out in that direction, the driver is pretty broken again.

I for sure don't have all for last months, but that "IRQF_DISABLED is not guaranteed on shared IRQs" for sure does not come from us here.

Cheers,
Hermann

hermann pitton

2009-04-04 01:07 UTC #2 lore raw

Am Samstag, den 04.04.2009, 02:45 +0200 schrieb hermann pitton:

Quoted 74 lines

Hi Ralph,
Am Freitag, den 03.04.2009, 20:49 +0000 schrieb Ralph:
Quoted 54 lines
ASUSTeK Tiger LNA Hybrid Capture Device PCI - Analog/DVB-T card Multimedia controller: Philips Semiconductors SAA7131/SAA7133/SAA7135 Video Broadcast Decoder (rev d1)
Works perfectly with kernel 2.6.28.4 (or older).
Recently, I have switched to 2.6.29 (same .config as 2.6.28.4) and now, at boot
time, I get the message:
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs
Signal strength is very low and Kaffeine is unable to tune in any channel. Same problem with kernel 2.6.29.1
-------------------------------------
Messages from /var/log/dmesg
saa7134 0000:03:0a.0: PCI INT A -> Link[APC3] -> GSI 18 (level, low) -> \
IRQ 18
saa7133[0]: found at 0000:03:0a.0, rev: 209, irq: 18, latency: 32, mmio: \
0xfdefe000
saa7133[0]: subsystem: 1043:4871, board: ASUS P7131 4871 \
[card=111,autodetected]
saa7133[0]: board init: gpio is 0
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]: i2c eeprom 00: 43 10 71 48 54 20 1c 00 43 43 a9 1c 55 d2 b2 92 saa7133[0]: i2c eeprom 10: ff ff ff 0f ff 20 ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 20: 01 40 01 02 03 00 01 03 08 ff 00 cf ff ff ff ff saa7133[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 40: ff 21 00 c2 96 10 03 22 15 50 ff ff ff ff ff ff saa7133[0]: i2c eeprom 50: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 60: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 70: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff tuner' 2-004b: chip found @ 0x96 (saa7133[0])
tda829x 2-004b: setting tuner address to 61
tda829x 2-004b: type set to tda8290+75a
saa7133[0]: registered device video0 [v4l2]
saa7133[0]: registered device vbi0
dvb_init() allocating 1 frontend
DVB: registering new adapter (saa7133[0])
DVB: registering adapter 0 frontend -32769 (Philips TDA10046H DVB-T)...
tda1004x: setting up plls for 48MHz sampling clock
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: trying to boot from eeprom
Quoted 9 lines
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: waiting for firmware upload...
saa7134 0000:03:0a.0: firmware: requesting dvb-fe-tda10046.fw
tda1004x: found firmware revision 29 -- ok
saa7134 ALSA driver for DMA sound loaded
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]/alsa: saa7133[0] at 0xfdefe000 irq 18 registered as card -1
thanks for your report, as announced previously, I unfortunately did not have time to run with latest always ... (guess why ...)
The driver always worked with shared IRQs, if not, it was always a limitation of certain hardware or mostly in some combination with binary only drivers.

Quoted 7 lines

If the above should be the case in general now, and not only caused by some blacklist, no print out in that direction, the driver is pretty broken again.
I for sure don't have all for last months, but that "IRQF_DISABLED is not guaranteed on shared IRQs" for sure does not come from us here.

Do use something unusual like pollirq or something?

We only have in saa7134-core.c

	/* initialize hardware #1 */
	saa7134_board_init1(dev);
	saa7134_hwinit1(dev);

	/* get irq */
	err = request_irq(pci_dev->irq, saa7134_irq,
			  IRQF_SHARED | IRQF_DISABLED, dev->name, dev);
	if (err < 0) {
		printk(KERN_ERR "%s: can't get IRQ %d\n",
		       dev->name,pci_dev->irq);
		goto fail3;
	}

and in saa7134-alsa.c

	err = request_irq(dev->pci->irq, saa7134_alsa_irq,
				IRQF_SHARED | IRQF_DISABLED, dev->name,
				(void*) &dev->dmasound);

	if (err < 0) {
		printk(KERN_ERR "%s: can't get IRQ %d for ALSA\n",
			dev->name, dev->pci->irq);
		goto __nodev;
	}

Have fun ;)
Hermann

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

hermann pitton

2009-04-05 18:57 UTC #4 lore raw

Hi,

Am Samstag, den 04.04.2009, 17:20 +0200 schrieb Ra.M.:

Quoted 133 lines

hermann pitton ha scritto:
Quoted 115 lines
Am Samstag, den 04.04.2009, 02:45 +0200 schrieb hermann pitton:
Quoted 75 lines
Hi Ralph,
Am Freitag, den 03.04.2009, 20:49 +0000 schrieb Ralph:
Quoted 54 lines
ASUSTeK Tiger LNA Hybrid Capture Device PCI - Analog/DVB-T card Multimedia controller: Philips Semiconductors SAA7131/SAA7133/SAA7135 Video Broadcast Decoder (rev d1)
Works perfectly with kernel 2.6.28.4 (or older).
Recently, I have switched to 2.6.29 (same .config as 2.6.28.4) and now, at boot
time, I get the message:
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs
Signal strength is very low and Kaffeine is unable to tune in any channel. Same problem with kernel 2.6.29.1
-------------------------------------
Messages from /var/log/dmesg
saa7134 0000:03:0a.0: PCI INT A -> Link[APC3] -> GSI 18 (level, low) -> \
IRQ 18
saa7133[0]: found at 0000:03:0a.0, rev: 209, irq: 18, latency: 32, mmio: \
0xfdefe000
saa7133[0]: subsystem: 1043:4871, board: ASUS P7131 4871 \
[card=111,autodetected]
saa7133[0]: board init: gpio is 0
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]: i2c eeprom 00: 43 10 71 48 54 20 1c 00 43 43 a9 1c 55 d2 b2 92 saa7133[0]: i2c eeprom 10: ff ff ff 0f ff 20 ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 20: 01 40 01 02 03 00 01 03 08 ff 00 cf ff ff ff ff saa7133[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 40: ff 21 00 c2 96 10 03 22 15 50 ff ff ff ff ff ff saa7133[0]: i2c eeprom 50: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 60: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 70: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff tuner' 2-004b: chip found @ 0x96 (saa7133[0])
tda829x 2-004b: setting tuner address to 61
tda829x 2-004b: type set to tda8290+75a
saa7133[0]: registered device video0 [v4l2]
saa7133[0]: registered device vbi0
dvb_init() allocating 1 frontend
DVB: registering new adapter (saa7133[0])
DVB: registering adapter 0 frontend -32769 (Philips TDA10046H DVB-T)...
tda1004x: setting up plls for 48MHz sampling clock
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: trying to boot from eeprom
Quoted 10 lines
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: waiting for firmware upload...
saa7134 0000:03:0a.0: firmware: requesting dvb-fe-tda10046.fw
tda1004x: found firmware revision 29 -- ok
saa7134 ALSA driver for DMA sound loaded
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]/alsa: saa7133[0] at 0xfdefe000 irq 18 registered as card -1
thanks for your report, as announced previously, I unfortunately did not have time to run with latest always ... (guess why ...)
The driver always worked with shared IRQs, if not, it was always a limitation of certain hardware or mostly in some combination with binary only drivers.
Quoted 8 lines
If the above should be the case in general now, and not only caused by some blacklist, no print out in that direction, the driver is pretty broken again.
I for sure don't have all for last months, but that "IRQF_DISABLED is not guaranteed on shared IRQs" for sure does not come from us here.
Do use something unusual like pollirq or something?
We only have in saa7134-core.c
	/* initialize hardware #1 */
	saa7134_board_init1(dev);
	saa7134_hwinit1(dev);

	/* get irq */
	err = request_irq(pci_dev->irq, saa7134_irq,
			  IRQF_SHARED | IRQF_DISABLED, dev->name, dev);
	if (err < 0) {
		printk(KERN_ERR "%s: can't get IRQ %d\n",
		       dev->name,pci_dev->irq);
		goto fail3;
	}
and in saa7134-alsa.c
	err = request_irq(dev->pci->irq, saa7134_alsa_irq,
				IRQF_SHARED | IRQF_DISABLED, dev->name,
				(void*) &dev->dmasound);

	if (err < 0) {
		printk(KERN_ERR "%s: can't get IRQ %d for ALSA\n",
			dev->name, dev->pci->irq);
		goto __nodev;
	}
Quoted 6 lines
Have fun ;)
Hermann
No, I do not use pollirq.
I have read that many users have had problems with 2.6.29 and IRQs. Those problems affect WiFi cards, Ethernet cards, DVB-T cards, etc.
For example:
http://article.gmane.org/gmane.linux.uml.devel/12098
http://www.gossamer-threads.com/lists/linux/kernel/1044282
http://zen-sources.org/content/irqfshared-irqfdisabled-fix-2629-rc

Quoted 7 lines

In all cases, at boot time appears the message:
IRQ XY: IRQF_DISABLED is not guaranteed on shared IRQs
So, probably, there is a kernel bug in the IRQs management of the 2.6.29 and 2.6.29.1

did build a 2.6.29.1 now and your report is correct!

DVB-T on saa7134 is broken at least for all tda10046 and tda8275 stuff and it is not restricted to devices with LNA.

For what I can see so far, it is not related to the IRQF_DISABLED print out, since only a warning for now and removing it from the driver doesn't change anything.

saa7134 DVB-S, analog TV and saa7134-alsa are not affected.

Installing the current mercurial v4l-dvb on 2.6.29.1 does fix it.

If on that saa7134-dvb.ko and saa7134.ko are replaced with the ones from 2.6.29.1 the breakage is back again. The related dvb and tuner modules tolerate such exchange on a first rough test.

As you reported, symptoms are tumbling signal and SNR between very low and 100%, as if tuning and AGC would never stabilize.

I suspect failing i2c stuff is involved. Did not notice anything like that on various mercurial versions during the last months.

Cheers,
Hermann

Mauro Carvalho Chehab

2009-04-14 03:23 UTC #5 lore raw

Quoted 161 lines · On Sun, 05 Apr 2009 20:22:33 +0200 hermann pitton <[email protected]>…

Hi,
Am Samstag, den 04.04.2009, 17:20 +0200 schrieb Ra.M.:
Quoted 133 lines
hermann pitton ha scritto:
Quoted 115 lines
Am Samstag, den 04.04.2009, 02:45 +0200 schrieb hermann pitton:
Quoted 75 lines
Hi Ralph,
Am Freitag, den 03.04.2009, 20:49 +0000 schrieb Ralph:
Quoted 54 lines
ASUSTeK Tiger LNA Hybrid Capture Device PCI - Analog/DVB-T card Multimedia controller: Philips Semiconductors SAA7131/SAA7133/SAA7135 Video Broadcast Decoder (rev d1)
Works perfectly with kernel 2.6.28.4 (or older).
Recently, I have switched to 2.6.29 (same .config as 2.6.28.4) and now, at boot
time, I get the message:
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs
Signal strength is very low and Kaffeine is unable to tune in any channel. Same problem with kernel 2.6.29.1
-------------------------------------
Messages from /var/log/dmesg
saa7134 0000:03:0a.0: PCI INT A -> Link[APC3] -> GSI 18 (level, low) -> \
IRQ 18
saa7133[0]: found at 0000:03:0a.0, rev: 209, irq: 18, latency: 32, mmio: \
0xfdefe000
saa7133[0]: subsystem: 1043:4871, board: ASUS P7131 4871 \
[card=111,autodetected]
saa7133[0]: board init: gpio is 0
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]: i2c eeprom 00: 43 10 71 48 54 20 1c 00 43 43 a9 1c 55 d2 b2 92 saa7133[0]: i2c eeprom 10: ff ff ff 0f ff 20 ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 20: 01 40 01 02 03 00 01 03 08 ff 00 cf ff ff ff ff saa7133[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 40: ff 21 00 c2 96 10 03 22 15 50 ff ff ff ff ff ff saa7133[0]: i2c eeprom 50: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 60: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 70: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff tuner' 2-004b: chip found @ 0x96 (saa7133[0])
tda829x 2-004b: setting tuner address to 61
tda829x 2-004b: type set to tda8290+75a
saa7133[0]: registered device video0 [v4l2]
saa7133[0]: registered device vbi0
dvb_init() allocating 1 frontend
DVB: registering new adapter (saa7133[0])
DVB: registering adapter 0 frontend -32769 (Philips TDA10046H DVB-T)...
tda1004x: setting up plls for 48MHz sampling clock
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: trying to boot from eeprom
Quoted 10 lines
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: waiting for firmware upload...
saa7134 0000:03:0a.0: firmware: requesting dvb-fe-tda10046.fw
tda1004x: found firmware revision 29 -- ok
saa7134 ALSA driver for DMA sound loaded
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]/alsa: saa7133[0] at 0xfdefe000 irq 18 registered as card -1
thanks for your report, as announced previously, I unfortunately did not have time to run with latest always ... (guess why ...)
The driver always worked with shared IRQs, if not, it was always a limitation of certain hardware or mostly in some combination with binary only drivers.
Quoted 8 lines
If the above should be the case in general now, and not only caused by some blacklist, no print out in that direction, the driver is pretty broken again.
I for sure don't have all for last months, but that "IRQF_DISABLED is not guaranteed on shared IRQs" for sure does not come from us here.
Do use something unusual like pollirq or something?
We only have in saa7134-core.c
	/* initialize hardware #1 */
	saa7134_board_init1(dev);
	saa7134_hwinit1(dev);

	/* get irq */
	err = request_irq(pci_dev->irq, saa7134_irq,
			  IRQF_SHARED | IRQF_DISABLED, dev->name, dev);
	if (err < 0) {
		printk(KERN_ERR "%s: can't get IRQ %d\n",
		       dev->name,pci_dev->irq);
		goto fail3;
	}
and in saa7134-alsa.c
	err = request_irq(dev->pci->irq, saa7134_alsa_irq,
				IRQF_SHARED | IRQF_DISABLED, dev->name,
				(void*) &dev->dmasound);

	if (err < 0) {
		printk(KERN_ERR "%s: can't get IRQ %d for ALSA\n",
			dev->name, dev->pci->irq);
		goto __nodev;
	}
Quoted 6 lines
Have fun ;)
Hermann
No, I do not use pollirq.
I have read that many users have had problems with 2.6.29 and IRQs. Those problems affect WiFi cards, Ethernet cards, DVB-T cards, etc.
For example:
http://article.gmane.org/gmane.linux.uml.devel/12098
http://www.gossamer-threads.com/lists/linux/kernel/1044282
http://zen-sources.org/content/irqfshared-irqfdisabled-fix-2629-rc
Quoted 7 lines
In all cases, at boot time appears the message:
IRQ XY: IRQF_DISABLED is not guaranteed on shared IRQs
So, probably, there is a kernel bug in the IRQs management of the 2.6.29 and 2.6.29.1
did build a 2.6.29.1 now and your report is correct!
DVB-T on saa7134 is broken at least for all tda10046 and tda8275 stuff and it is not restricted to devices with LNA.
For what I can see so far, it is not related to the IRQF_DISABLED print out, since only a warning for now and removing it from the driver doesn't change anything.
saa7134 DVB-S, analog TV and saa7134-alsa are not affected.
Installing the current mercurial v4l-dvb on 2.6.29.1 does fix it.
If on that saa7134-dvb.ko and saa7134.ko are replaced with the ones from 2.6.29.1 the breakage is back again. The related dvb and tuner modules tolerate such exchange on a first rough test.

Quoted 5 lines

As you reported, symptoms are tumbling signal and SNR between very low and 100%, as if tuning and AGC would never stabilize.
I suspect failing i2c stuff is involved. Did not notice anything like that on various mercurial versions during the last months.

Hermann,

Could you please try to bisect the patch that broke it? The instructions for bisecting with mercurial are available at README.patches file.

Cheers,
Mauro

David Wong

2009-04-14 09:30 UTC #6 lore raw

Quoted 175 lines · Mauro Carvalho Chehab

Quoted 161 lines · On Sun, 05 Apr 2009 20:22:33 +0200 hermann pitton <[email protected]>…
Hi,
Am Samstag, den 04.04.2009, 17:20 +0200 schrieb Ra.M.:
Quoted 133 lines
hermann pitton ha scritto:
Quoted 109 lines
Am Samstag, den 04.04.2009, 02:45 +0200 schrieb hermann pitton:
Quoted 75 lines
Hi Ralph,
Am Freitag, den 03.04.2009, 20:49 +0000 schrieb Ralph:
Quoted 54 lines
ASUSTeK Tiger LNA Hybrid Capture Device PCI - Analog/DVB-T card Multimedia controller: Philips Semiconductors SAA7131/SAA7133/SAA7135 Video Broadcast Decoder (rev d1)
Works perfectly with kernel 2.6.28.4 (or older).
Recently, I have switched to 2.6.29 (same .config as 2.6.28.4) and now, at boot
time, I get the message:
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs
Signal strength is very low and Kaffeine is unable to tune in any channel. Same problem with kernel 2.6.29.1
-------------------------------------
Messages from /var/log/dmesg
saa7134 0000:03:0a.0: PCI INT A -> Link[APC3] -> GSI 18 (level, low) -> \
IRQ 18
saa7133[0]: found at 0000:03:0a.0, rev: 209, irq: 18, latency: 32, mmio: \
0xfdefe000
saa7133[0]: subsystem: 1043:4871, board: ASUS P7131 4871 \
[card=111,autodetected]
saa7133[0]: board init: gpio is 0
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]: i2c eeprom 00: 43 10 71 48 54 20 1c 00 43 43 a9 1c 55 d2 b2 92 saa7133[0]: i2c eeprom 10: ff ff ff 0f ff 20 ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 20: 01 40 01 02 03 00 01 03 08 ff 00 cf ff ff ff ff saa7133[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 40: ff 21 00 c2 96 10 03 22 15 50 ff ff ff ff ff ff saa7133[0]: i2c eeprom 50: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 60: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 70: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff tuner' 2-004b: chip found @ 0x96 (saa7133[0])
tda829x 2-004b: setting tuner address to 61
tda829x 2-004b: type set to tda8290+75a
saa7133[0]: registered device video0 [v4l2]
saa7133[0]: registered device vbi0
dvb_init() allocating 1 frontend
DVB: registering new adapter (saa7133[0])
DVB: registering adapter 0 frontend -32769 (Philips TDA10046H DVB-T)...
tda1004x: setting up plls for 48MHz sampling clock
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: trying to boot from eeprom
Quoted 10 lines
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: waiting for firmware upload...
saa7134 0000:03:0a.0: firmware: requesting dvb-fe-tda10046.fw
tda1004x: found firmware revision 29 -- ok
saa7134 ALSA driver for DMA sound loaded
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]/alsa: saa7133[0] at 0xfdefe000 irq 18 registered as card -1
thanks for your report, as announced previously, I unfortunately did not have time to run with latest always ... (guess why ...)
The driver always worked with shared IRQs, if not, it was always a limitation of certain hardware or mostly in some combination with binary only drivers.
Quoted 8 lines
If the above should be the case in general now, and not only caused by some blacklist, no print out in that direction, the driver is pretty broken again.
I for sure don't have all for last months, but that "IRQF_DISABLED is not guaranteed on shared IRQs" for sure does not come from us here.
Do use something unusual like pollirq or something?
We only have in saa7134-core.c
/* initialize hardware #1 */
saa7134_board_init1(dev);
saa7134_hwinit1(dev);
/* get irq */
err = request_irq(pci_dev->irq, saa7134_irq, IRQF_SHARED | IRQF_DISABLED, dev->name, dev); if (err < 0) {
printk(KERN_ERR "%s: can't get IRQ %d\n", dev->name,pci_dev->irq);
goto fail3;
}
and in saa7134-alsa.c
err = request_irq(dev->pci->irq, saa7134_alsa_irq, IRQF_SHARED | IRQF_DISABLED, dev->name, (void*) &dev->dmasound);
Quoted 12 lines
if (err < 0) {
printk(KERN_ERR "%s: can't get IRQ %d for ALSA\n", dev->name, dev->pci->irq);
goto __nodev;
}
Have fun ;)
Hermann
No, I do not use pollirq.
I have read that many users have had problems with 2.6.29 and IRQs. Those problems affect WiFi cards, Ethernet cards, DVB-T cards, etc.
For example:
http://article.gmane.org/gmane.linux.uml.devel/12098
http://www.gossamer-threads.com/lists/linux/kernel/1044282
http://zen-sources.org/content/irqfshared-irqfdisabled-fix-2629-rc
Quoted 7 lines
In all cases, at boot time appears the message:
IRQ XY: IRQF_DISABLED is not guaranteed on shared IRQs
So, probably, there is a kernel bug in the IRQs management of the 2.6.29 and 2.6.29.1
did build a 2.6.29.1 now and your report is correct!
DVB-T on saa7134 is broken at least for all tda10046 and tda8275 stuff and it is not restricted to devices with LNA.
For what I can see so far, it is not related to the IRQF_DISABLED print out, since only a warning for now and removing it from the driver doesn't change anything.
saa7134 DVB-S, analog TV and saa7134-alsa are not affected.
Installing the current mercurial v4l-dvb on 2.6.29.1 does fix it.
If on that saa7134-dvb.ko and saa7134.ko are replaced with the ones from 2.6.29.1 the breakage is back again. The related dvb and tuner modules tolerate such exchange on a first rough test.
Quoted 5 lines
As you reported, symptoms are tumbling signal and SNR between very low and 100%, as if tuning and AGC would never stabilize.
I suspect failing i2c stuff is involved. Did not notice anything like that on various mercurial versions during the last months.
Hermann,
Could you please try to bisect the patch that broke it? The instructions for bisecting with mercurial are available at README.patches file.

Quoted 7 lines

Cheers,
Mauro
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Sorry for interrupt.
Would your saa7134 i2c problem is due to the i2c quirk?
I have problem on the saa7134 i2c quirk that I have to totally disable it on my work-in-progress card.
Just a little suggestion that trying disable the i2c quirk like this change set:
http://linuxtv.org/hg/~mkrufky/dmbth/rev/781ffa6c43d3

David.

hermann pitton

2009-04-14 21:42 UTC #7 lore raw

Hi,

Am Dienstag, den 14.04.2009, 17:30 +0800 schrieb David Wong:

Quoted 169 lines

Quoted 167 lines · Mauro Carvalho Chehab
Quoted 164 lines · On Sun, 05 Apr 2009 20:22:33 +0200 hermann pitton <[email protected]>…
Hi,
Am Samstag, den 04.04.2009, 17:20 +0200 schrieb Ra.M.:
Quoted 133 lines
hermann pitton ha scritto:
Quoted 109 lines
Am Samstag, den 04.04.2009, 02:45 +0200 schrieb hermann pitton:
Quoted 75 lines
Hi Ralph,
Am Freitag, den 03.04.2009, 20:49 +0000 schrieb Ralph:
Quoted 54 lines
ASUSTeK Tiger LNA Hybrid Capture Device PCI - Analog/DVB-T card Multimedia controller: Philips Semiconductors SAA7131/SAA7133/SAA7135 Video Broadcast Decoder (rev d1)
Works perfectly with kernel 2.6.28.4 (or older).
Recently, I have switched to 2.6.29 (same .config as 2.6.28.4) and now, at boot
time, I get the message:
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs
Signal strength is very low and Kaffeine is unable to tune in any channel. Same problem with kernel 2.6.29.1
-------------------------------------
Messages from /var/log/dmesg
saa7134 0000:03:0a.0: PCI INT A -> Link[APC3] -> GSI 18 (level, low) -> \
IRQ 18
saa7133[0]: found at 0000:03:0a.0, rev: 209, irq: 18, latency: 32, mmio: \
0xfdefe000
saa7133[0]: subsystem: 1043:4871, board: ASUS P7131 4871 \
[card=111,autodetected]
saa7133[0]: board init: gpio is 0
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]: i2c eeprom 00: 43 10 71 48 54 20 1c 00 43 43 a9 1c 55 d2 b2 92 saa7133[0]: i2c eeprom 10: ff ff ff 0f ff 20 ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 20: 01 40 01 02 03 00 01 03 08 ff 00 cf ff ff ff ff saa7133[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 40: ff 21 00 c2 96 10 03 22 15 50 ff ff ff ff ff ff saa7133[0]: i2c eeprom 50: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 60: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 70: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff tuner' 2-004b: chip found @ 0x96 (saa7133[0])
tda829x 2-004b: setting tuner address to 61
tda829x 2-004b: type set to tda8290+75a
saa7133[0]: registered device video0 [v4l2]
saa7133[0]: registered device vbi0
dvb_init() allocating 1 frontend
DVB: registering new adapter (saa7133[0])
DVB: registering adapter 0 frontend -32769 (Philips TDA10046H DVB-T)...
tda1004x: setting up plls for 48MHz sampling clock
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: trying to boot from eeprom
Quoted 10 lines
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: waiting for firmware upload...
saa7134 0000:03:0a.0: firmware: requesting dvb-fe-tda10046.fw
tda1004x: found firmware revision 29 -- ok
saa7134 ALSA driver for DMA sound loaded
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]/alsa: saa7133[0] at 0xfdefe000 irq 18 registered as card -1
thanks for your report, as announced previously, I unfortunately did not have time to run with latest always ... (guess why ...)
The driver always worked with shared IRQs, if not, it was always a limitation of certain hardware or mostly in some combination with binary only drivers.
Quoted 8 lines
If the above should be the case in general now, and not only caused by some blacklist, no print out in that direction, the driver is pretty broken again.
I for sure don't have all for last months, but that "IRQF_DISABLED is not guaranteed on shared IRQs" for sure does not come from us here.
Do use something unusual like pollirq or something?
We only have in saa7134-core.c
  /* initialize hardware #1 */
  saa7134_board_init1(dev);
  saa7134_hwinit1(dev);
/* get irq */
err = request_irq(pci_dev->irq, saa7134_irq,
                    IRQF_SHARED | IRQF_DISABLED, dev->name, dev);
  if (err < 0) {
          printk(KERN_ERR "%s: can't get IRQ %d\n",
                 dev->name,pci_dev->irq);
          goto fail3;
}
and in saa7134-alsa.c
err = request_irq(dev->pci->irq, saa7134_alsa_irq,
                          IRQF_SHARED | IRQF_DISABLED, dev->name,
                          (void*) &dev->dmasound);
Quoted 12 lines
  if (err < 0) {
          printk(KERN_ERR "%s: can't get IRQ %d for ALSA\n",
                  dev->name, dev->pci->irq);
          goto __nodev;
  }
Have fun ;)
Hermann
No, I do not use pollirq.
I have read that many users have had problems with 2.6.29 and IRQs. Those problems affect WiFi cards, Ethernet cards, DVB-T cards, etc.
For example:
http://article.gmane.org/gmane.linux.uml.devel/12098
http://www.gossamer-threads.com/lists/linux/kernel/1044282
http://zen-sources.org/content/irqfshared-irqfdisabled-fix-2629-rc
Quoted 7 lines
In all cases, at boot time appears the message:
IRQ XY: IRQF_DISABLED is not guaranteed on shared IRQs
So, probably, there is a kernel bug in the IRQs management of the 2.6.29 and 2.6.29.1
did build a 2.6.29.1 now and your report is correct!
DVB-T on saa7134 is broken at least for all tda10046 and tda8275 stuff and it is not restricted to devices with LNA.
For what I can see so far, it is not related to the IRQF_DISABLED print out, since only a warning for now and removing it from the driver doesn't change anything.
saa7134 DVB-S, analog TV and saa7134-alsa are not affected.
Installing the current mercurial v4l-dvb on 2.6.29.1 does fix it.
If on that saa7134-dvb.ko and saa7134.ko are replaced with the ones from 2.6.29.1 the breakage is back again. The related dvb and tuner modules tolerate such exchange on a first rough test.
As you reported, symptoms are tumbling signal and SNR between very low and 100%, as if tuning and AGC would never stabilize.

Quoted 7 lines

Quoted 7 lines
Quoted 2 lines
I suspect failing i2c stuff is involved. Did not notice anything like that on various mercurial versions during the last months.
Hermann,
Could you please try to bisect the patch that broke it? The instructions for bisecting with mercurial are available at README.patches file.

hmm, current mercurial is not broken for DVB-T on the 2.6.29.1. So we would have to find out if mercurial was ever bad.

Your on the 2.6.29 missing i2c gate control fixes made no difference too porting them to 2.6.29.1. I was thinking about to test next if it is already broken on the rc1 to come closer to it.

Quoted 8 lines

Sorry for interrupt.
Would your saa7134 i2c problem is due to the i2c quirk?
I have problem on the saa7134 i2c quirk that I have to totally disable it on my work-in-progress card.
Just a little suggestion that trying disable the i2c quirk like this change set:
http://linuxtv.org/hg/~mkrufky/dmbth/rev/781ffa6c43d3
David.

I have quite some cards and never had problems with that. The tuner status read out is also the same on the broken 2.6.29 and on working mercurial v4l-dvb and quirks are still enabled.

Unfortunately I can't promise any time soon for excessive bisecting, but I'll try to check if it is already on 29-rc1.

BTW, someone must disable the IR on the VIDEOMATE_T750 or add the missing. People can't even boot up it seems. Must go to stable too.

Cheers,
Hermann

hermann pitton

2009-04-15 23:00 UTC #8 lore raw

Hi,

just a short update.

Am Dienstag, den 14.04.2009, 23:33 +0200 schrieb hermann pitton:

Quoted 92 lines

Hi,
Am Dienstag, den 14.04.2009, 17:30 +0800 schrieb David Wong:
Quoted 89 lines
Quoted 87 lines · Mauro Carvalho Chehab
Quoted 84 lines · On Sun, 05 Apr 2009 20:22:33 +0200 hermann pitton <[email protected]>…
Hi,
Am Samstag, den 04.04.2009, 17:20 +0200 schrieb Ra.M.:
Quoted 81 lines
hermann pitton ha scritto:
Quoted 80 lines
Am Samstag, den 04.04.2009, 02:45 +0200 schrieb hermann pitton:
Quoted 78 lines
Hi Ralph,
Am Freitag, den 03.04.2009, 20:49 +0000 schrieb Ralph:
Quoted 54 lines
ASUSTeK Tiger LNA Hybrid Capture Device PCI - Analog/DVB-T card Multimedia controller: Philips Semiconductors SAA7131/SAA7133/SAA7135 Video Broadcast Decoder (rev d1)
Works perfectly with kernel 2.6.28.4 (or older).
Recently, I have switched to 2.6.29 (same .config as 2.6.28.4) and now, at boot
time, I get the message:
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs
Signal strength is very low and Kaffeine is unable to tune in any channel. Same problem with kernel 2.6.29.1
-------------------------------------
Messages from /var/log/dmesg
saa7134 0000:03:0a.0: PCI INT A -> Link[APC3] -> GSI 18 (level, low) -> \
IRQ 18
saa7133[0]: found at 0000:03:0a.0, rev: 209, irq: 18, latency: 32, mmio: \
0xfdefe000
saa7133[0]: subsystem: 1043:4871, board: ASUS P7131 4871 \
[card=111,autodetected]
saa7133[0]: board init: gpio is 0
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]: i2c eeprom 00: 43 10 71 48 54 20 1c 00 43 43 a9 1c 55 d2 b2 92 saa7133[0]: i2c eeprom 10: ff ff ff 0f ff 20 ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 20: 01 40 01 02 03 00 01 03 08 ff 00 cf ff ff ff ff saa7133[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 40: ff 21 00 c2 96 10 03 22 15 50 ff ff ff ff ff ff saa7133[0]: i2c eeprom 50: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 60: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 70: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff tuner' 2-004b: chip found @ 0x96 (saa7133[0])
tda829x 2-004b: setting tuner address to 61
tda829x 2-004b: type set to tda8290+75a
saa7133[0]: registered device video0 [v4l2]
saa7133[0]: registered device vbi0
dvb_init() allocating 1 frontend
DVB: registering new adapter (saa7133[0])
DVB: registering adapter 0 frontend -32769 (Philips TDA10046H DVB-T)...
tda1004x: setting up plls for 48MHz sampling clock
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: trying to boot from eeprom
Quoted 10 lines
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: waiting for firmware upload...
saa7134 0000:03:0a.0: firmware: requesting dvb-fe-tda10046.fw
tda1004x: found firmware revision 29 -- ok
saa7134 ALSA driver for DMA sound loaded
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]/alsa: saa7133[0] at 0xfdefe000 irq 18 registered as card -1
thanks for your report, as announced previously, I unfortunately did not have time to run with latest always ... (guess why ...)
The driver always worked with shared IRQs, if not, it was always a limitation of certain hardware or mostly in some combination with binary only drivers.
If the above should be the case in general now, and not only caused by some blacklist, no print out in that direction, the driver is pretty broken again.

The problem seems to be limited to only some cards.

Quoted 55 lines

Quoted 55 lines
Quoted 55 lines
Quoted 55 lines
Quoted 55 lines
Quoted 28 lines
Quoted 4 lines
I for sure don't have all for last months, but that "IRQF_DISABLED is not guaranteed on shared IRQs" for sure does not come from us here.
Do use something unusual like pollirq or something?
We only have in saa7134-core.c
  /* initialize hardware #1 */
  saa7134_board_init1(dev);
  saa7134_hwinit1(dev);
/* get irq */
err = request_irq(pci_dev->irq, saa7134_irq,
                    IRQF_SHARED | IRQF_DISABLED, dev->name, dev);
  if (err < 0) {
          printk(KERN_ERR "%s: can't get IRQ %d\n",
                 dev->name,pci_dev->irq);
          goto fail3;
}
and in saa7134-alsa.c
err = request_irq(dev->pci->irq, saa7134_alsa_irq,
                          IRQF_SHARED | IRQF_DISABLED, dev->name,
                          (void*) &dev->dmasound);
Quoted 12 lines
  if (err < 0) {
          printk(KERN_ERR "%s: can't get IRQ %d for ALSA\n",
                  dev->name, dev->pci->irq);
          goto __nodev;
  }
Have fun ;)
Hermann
No, I do not use pollirq.
I have read that many users have had problems with 2.6.29 and IRQs. Those problems affect WiFi cards, Ethernet cards, DVB-T cards, etc.
For example:
http://article.gmane.org/gmane.linux.uml.devel/12098
http://www.gossamer-threads.com/lists/linux/kernel/1044282
http://zen-sources.org/content/irqfshared-irqfdisabled-fix-2629-rc
In all cases, at boot time appears the message:
IRQ XY: IRQF_DISABLED is not guaranteed on shared IRQs

Quoted 8 lines

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 3 lines
So, probably, there is a kernel bug in the IRQs management of the 2.6.29 and 2.6.29.1
did build a 2.6.29.1 now and your report is correct!
DVB-T on saa7134 is broken at least for all tda10046 and tda8275 stuff and it is not restricted to devices with LNA.

That was too soon, but devices with and without LNA are reported.

Quoted 7 lines

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
For what I can see so far, it is not related to the IRQF_DISABLED print out, since only a warning for now and removing it from the driver doesn't change anything.
saa7134 DVB-S, analog TV and saa7134-alsa are not affected.
Installing the current mercurial v4l-dvb on 2.6.29.1 does fix it.

That is confirmed on the Tiger LNA Hybrid too.

Quoted 35 lines

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
If on that saa7134-dvb.ko and saa7134.ko are replaced with the ones from 2.6.29.1 the breakage is back again. The related dvb and tuner modules tolerate such exchange on a first rough test.
As you reported, symptoms are tumbling signal and SNR between very low and 100%, as if tuning and AGC would never stabilize.
Quoted 7 lines
Quoted 7 lines
Quoted 2 lines
I suspect failing i2c stuff is involved. Did not notice anything like that on various mercurial versions during the last months.
Hermann,
Could you please try to bisect the patch that broke it? The instructions for bisecting with mercurial are available at README.patches file.
hmm, current mercurial is not broken for DVB-T on the 2.6.29.1. So we would have to find out if mercurial was ever bad.
Your on the 2.6.29 missing i2c gate control fixes made no difference too porting them to 2.6.29.1. I was thinking about to test next if it is already broken on the rc1 to come closer to it.
Quoted 8 lines
Sorry for interrupt.
Would your saa7134 i2c problem is due to the i2c quirk?
I have problem on the saa7134 i2c quirk that I have to totally disable it on my work-in-progress card.
Just a little suggestion that trying disable the i2c quirk like this change set:
http://linuxtv.org/hg/~mkrufky/dmbth/rev/781ffa6c43d3
David.
I have quite some cards and never had problems with that. The tuner status read out is also the same on the broken 2.6.29 and on working mercurial v4l-dvb and quirks are still enabled.

Quoted 5 lines

Unfortunately I can't promise any time soon for excessive bisecting, but I'll try to check if it is already on 29-rc1.
BTW, someone must disable the IR on the VIDEOMATE_T750 or add the missing. People can't even boot up it seems. Must go to stable too.

I saw the trouble so far only on the Medion Quad CTX944 in the blue dual MSI PCI slot and on the CTX948 in the same machine in a standard slot. It is treated like the 0x007 subdevice of the Quad.

I can't easily change cards in that AMD quad core mini ATX machine.

So trying some other direction first.

After 1 Gigabyte of Fedora 10 updates on a x86_64 K8 machine, not used for a few weeks, and having a 2.6.29.1 there with the same config, the Asus Tiger 3in1 with still working LNA loading external firmware and the first Asus Tiger ever, that one with firmware eeprom and antenna input switch not suitable for the mounted connectors, are fine.

A second Medion Quad in this machine, only the 0x0007 subdevice is functional in an orange MSI PCI slot, doesn't matter and just to have mentioned it, shows the same problems for DVB-T.

Hopefully only a few users are affected then and a pattern should become visible. Thanks to all reporting and giving hints so far.

Cheers,
Hermann

Please report any other cards with new DVB-T problems on 2.6.29.

hermann pitton

2009-04-16 23:45 UTC #9 lore raw

Am Donnerstag, den 16.04.2009, 00:49 +0200 schrieb hermann pitton:

Quoted 3 lines

Hi,
just a short update.

Just again, don't read it if not affected, but I don't cut anything yet, in case it should be for longer.

Quoted 231 lines

Am Dienstag, den 14.04.2009, 23:33 +0200 schrieb hermann pitton:
Quoted 92 lines
Hi,
Am Dienstag, den 14.04.2009, 17:30 +0800 schrieb David Wong:
Quoted 89 lines
Quoted 87 lines · Mauro Carvalho Chehab
Quoted 84 lines · On Sun, 05 Apr 2009 20:22:33 +0200 hermann pitton <[email protected]>…
Hi,
Am Samstag, den 04.04.2009, 17:20 +0200 schrieb Ra.M.:
Quoted 81 lines
hermann pitton ha scritto:
Quoted 80 lines
Am Samstag, den 04.04.2009, 02:45 +0200 schrieb hermann pitton:
Quoted 78 lines
Hi Ralph,
Am Freitag, den 03.04.2009, 20:49 +0000 schrieb Ralph:
Quoted 54 lines
ASUSTeK Tiger LNA Hybrid Capture Device PCI - Analog/DVB-T card Multimedia controller: Philips Semiconductors SAA7131/SAA7133/SAA7135 Video Broadcast Decoder (rev d1)
Works perfectly with kernel 2.6.28.4 (or older).
Recently, I have switched to 2.6.29 (same .config as 2.6.28.4) and now, at boot
time, I get the message:
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs
Signal strength is very low and Kaffeine is unable to tune in any channel. Same problem with kernel 2.6.29.1
-------------------------------------
Messages from /var/log/dmesg
saa7134 0000:03:0a.0: PCI INT A -> Link[APC3] -> GSI 18 (level, low) -> \
IRQ 18
saa7133[0]: found at 0000:03:0a.0, rev: 209, irq: 18, latency: 32, mmio: \
0xfdefe000
saa7133[0]: subsystem: 1043:4871, board: ASUS P7131 4871 \
[card=111,autodetected]
saa7133[0]: board init: gpio is 0
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]: i2c eeprom 00: 43 10 71 48 54 20 1c 00 43 43 a9 1c 55 d2 b2 92 saa7133[0]: i2c eeprom 10: ff ff ff 0f ff 20 ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 20: 01 40 01 02 03 00 01 03 08 ff 00 cf ff ff ff ff saa7133[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 40: ff 21 00 c2 96 10 03 22 15 50 ff ff ff ff ff ff saa7133[0]: i2c eeprom 50: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 60: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 70: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom 90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom a0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff tuner' 2-004b: chip found @ 0x96 (saa7133[0])
tda829x 2-004b: setting tuner address to 61
tda829x 2-004b: type set to tda8290+75a
saa7133[0]: registered device video0 [v4l2]
saa7133[0]: registered device vbi0
dvb_init() allocating 1 frontend
DVB: registering new adapter (saa7133[0])
DVB: registering adapter 0 frontend -32769 (Philips TDA10046H DVB-T)...
tda1004x: setting up plls for 48MHz sampling clock
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: trying to boot from eeprom
Quoted 10 lines
tda1004x: timeout waiting for DSP ready
tda1004x: found firmware revision 0 -- invalid
tda1004x: waiting for firmware upload...
saa7134 0000:03:0a.0: firmware: requesting dvb-fe-tda10046.fw
tda1004x: found firmware revision 29 -- ok
saa7134 ALSA driver for DMA sound loaded
IRQ 18/saa7133[0]: IRQF_DISABLED is not guaranteed on shared IRQs saa7133[0]/alsa: saa7133[0] at 0xfdefe000 irq 18 registered as card -1
thanks for your report, as announced previously, I unfortunately did not have time to run with latest always ... (guess why ...)
The driver always worked with shared IRQs, if not, it was always a limitation of certain hardware or mostly in some combination with binary only drivers.
If the above should be the case in general now, and not only caused by some blacklist, no print out in that direction, the driver is pretty broken again.
The problem seems to be limited to only some cards.
Quoted 55 lines
Quoted 55 lines
Quoted 55 lines
Quoted 55 lines
Quoted 55 lines
Quoted 28 lines
Quoted 4 lines
I for sure don't have all for last months, but that "IRQF_DISABLED is not guaranteed on shared IRQs" for sure does not come from us here.
Do use something unusual like pollirq or something?
We only have in saa7134-core.c
  /* initialize hardware #1 */
  saa7134_board_init1(dev);
  saa7134_hwinit1(dev);
/* get irq */
err = request_irq(pci_dev->irq, saa7134_irq,
                    IRQF_SHARED | IRQF_DISABLED, dev->name, dev);
  if (err < 0) {
          printk(KERN_ERR "%s: can't get IRQ %d\n",
                 dev->name,pci_dev->irq);
          goto fail3;
}
and in saa7134-alsa.c
err = request_irq(dev->pci->irq, saa7134_alsa_irq,
                          IRQF_SHARED | IRQF_DISABLED, dev->name,
                          (void*) &dev->dmasound);
Quoted 12 lines
  if (err < 0) {
          printk(KERN_ERR "%s: can't get IRQ %d for ALSA\n",
                  dev->name, dev->pci->irq);
          goto __nodev;
  }
Have fun ;)
Hermann
No, I do not use pollirq.
I have read that many users have had problems with 2.6.29 and IRQs. Those problems affect WiFi cards, Ethernet cards, DVB-T cards, etc.
For example:
http://article.gmane.org/gmane.linux.uml.devel/12098
http://www.gossamer-threads.com/lists/linux/kernel/1044282
http://zen-sources.org/content/irqfshared-irqfdisabled-fix-2629-rc
In all cases, at boot time appears the message:
IRQ XY: IRQF_DISABLED is not guaranteed on shared IRQs
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 3 lines
So, probably, there is a kernel bug in the IRQs management of the 2.6.29 and 2.6.29.1
did build a 2.6.29.1 now and your report is correct!
DVB-T on saa7134 is broken at least for all tda10046 and tda8275 stuff and it is not restricted to devices with LNA.
That was too soon, but devices with and without LNA are reported.
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
For what I can see so far, it is not related to the IRQF_DISABLED print out, since only a warning for now and removing it from the driver doesn't change anything.
saa7134 DVB-S, analog TV and saa7134-alsa are not affected.
Installing the current mercurial v4l-dvb on 2.6.29.1 does fix it.
That is confirmed on the Tiger LNA Hybrid too.
Quoted 35 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
If on that saa7134-dvb.ko and saa7134.ko are replaced with the ones from 2.6.29.1 the breakage is back again. The related dvb and tuner modules tolerate such exchange on a first rough test.
As you reported, symptoms are tumbling signal and SNR between very low and 100%, as if tuning and AGC would never stabilize.
Quoted 7 lines
Quoted 7 lines
Quoted 2 lines
I suspect failing i2c stuff is involved. Did not notice anything like that on various mercurial versions during the last months.
Hermann,
Could you please try to bisect the patch that broke it? The instructions for bisecting with mercurial are available at README.patches file.
hmm, current mercurial is not broken for DVB-T on the 2.6.29.1. So we would have to find out if mercurial was ever bad.
Your on the 2.6.29 missing i2c gate control fixes made no difference too porting them to 2.6.29.1. I was thinking about to test next if it is already broken on the rc1 to come closer to it.
Quoted 8 lines
Sorry for interrupt.
Would your saa7134 i2c problem is due to the i2c quirk?
I have problem on the saa7134 i2c quirk that I have to totally disable it on my work-in-progress card.
Just a little suggestion that trying disable the i2c quirk like this change set:
http://linuxtv.org/hg/~mkrufky/dmbth/rev/781ffa6c43d3
David.
I have quite some cards and never had problems with that. The tuner status read out is also the same on the broken 2.6.29 and on working mercurial v4l-dvb and quirks are still enabled.
Quoted 5 lines
Unfortunately I can't promise any time soon for excessive bisecting, but I'll try to check if it is already on 29-rc1.
BTW, someone must disable the IR on the VIDEOMATE_T750 or add the missing. People can't even boot up it seems. Must go to stable too.
I saw the trouble so far only on the Medion Quad CTX944 in the blue dual MSI PCI slot and on the CTX948 in the same machine in a standard slot. It is treated like the 0x007 subdevice of the Quad.
I can't easily change cards in that AMD quad core mini ATX machine.
So trying some other direction first.
After 1 Gigabyte of Fedora 10 updates on a x86_64 K8 machine, not used for a few weeks, and having a 2.6.29.1 there with the same config, the Asus Tiger 3in1 with still working LNA loading external firmware and the first Asus Tiger ever, that one with firmware eeprom and antenna input switch not suitable for the mounted connectors, are fine.
A second Medion Quad in this machine, only the 0x0007 subdevice is functional in an orange MSI PCI slot, doesn't matter and just to have mentioned it, shows the same problems for DVB-T.

Quoted 8 lines

Hopefully only a few users are affected then and a pattern should become visible. Thanks to all reporting and giving hints so far.
Cheers,
Hermann
Please report any other cards with new DVB-T problems on 2.6.29.

I did revive another little bit noisy old machine, this time with a K7 i686 and compiled and installed a 2.6.29.1, sigh.

Two different CTX cards with FMD1216ME MK3 hybrid tuners and tda10046 do work, an old Asus P7131 Dual does work, but the CTX953 card=134 fails.

If there would not be that report for the DVB-T ASUSTeK Tiger LNA Hybrid, I would say cards with tuners at 0x60 fail ;)

So it goes for all other card specific stuff, even shared IRQs, binary drivers and firmware versions don't matter.

Hopefully I did overlook something on a first review, or we are tricked into something really dirty.

Cheers,
Hermann

hermann pitton

2009-04-23 23:37 UTC #10 lore raw

In reply to David Wong

Hi David and all interested,

[snip]

Quoted 9 lines

Sorry for interrupt.
Would your saa7134 i2c problem is due to the i2c quirk?
I have problem on the saa7134 i2c quirk that I have to totally disable it on my work-in-progress card.
Just a little suggestion that trying disable the i2c quirk like this change set:
http://linuxtv.org/hg/~mkrufky/dmbth/rev/781ffa6c43d3
David.

I cross post this in for the record.

Commenting the i2c quirk does not help at all on these card, as already assumed.

And I also can't see any pattern yet, which should cause this from within v4l-dvb.

At least, but who has hardware to test on all cases, the i2c quirk Gerd needed for the first saa7134 DVB card ever, the Pinnacle 300i, seems not to be needed for a few other cards I could test on only for now with saa7134 and saa7131e.

So, Jean likely is right, that we should have it the other way round and this might help you and Mike with a decision still to make.

Since 2.6.28 did work for those cards and 2.6.30-rc2-git4 seems to work within limited usability in my user environment, it seems to be something in between and might not even be the same.

Cheers,
Hermann