Skip to contentSkip to conversation

Current saa7134 driver breaks KNC One Tv-Station DVR (card=24)

DISCUSSION Bernhard Rosenkraenzer 3 participants 5 messages Dec 28, 2004 – Dec 31, 2004 Media (V4L/DVB)

lore
Share & tools
lore
#0 lore raw

Trying to modprobe saa7134 with this card results in a hanging modprobe process.

dmesg gets:
saa7130/34: v4l2 driver version 0.2.12 loaded
ACPI: PCI interrupt 0000:00:06.0[A] -> GSI 17 (level, low) -> IRQ 177 saa7134[0]: found at 0000:00:06.0, rev: 1, irq: 177, latency: 64, mmio:
0xdfff9c00
saa7134[0]: subsystem: 1894:a006, board: KNC One TV-Station DVR [card=24,autodetected]
saa7134[0]: board init: gpio is 830000
saa7134[0]: i2c eeprom 00: 94 18 06 a0 06 80 00 01 00 00 00 00 00 00 01 00 saa7134[0]: i2c eeprom 10: 00 ff 86 0e ff 20 ff ff ff ff ff ff ff ff ff ff saa7134[0]: i2c eeprom 20: 01 40 01 02 02 03 01 03 06 ff 01 df ff ff ff ff saa7134[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
tuner: chip found at addr 0xc0 i2c-bus saa7134[0]
tuner: type set to 38 (Philips PAL/SECAM multi (FM1216ME MK3)) by saa7134[0] tda9885/6/7: chip found @ 0x86
saa7134[0]/irq[10,-133081]: r=0x20 s=0x00 PE
saa7134[0]/irq: looping -- clearing PE (parity error!) enable bit

At that point, modprobe hangs; I can still use the box in other ttys, but [of course] the saa7134 card doesn't work.

This is reproducable at least with with 2.6.10-rc3-mm1, 2.6.10 and 2.6.10-ac1; the card used to work with earlier releases (I can't remember the last version that worked though, haven't used it for a while).

#1 lore raw
Quoted 2 lines · Bernhard Rosenkraenzer

Trying to modprobe saa7134 with this card results in a hanging modprobe process.

Some more debug info (with 2.6.10-ac1):

saa7130/34: v4l2 driver version 0.2.12 loaded
ACPI: PCI interrupt 0000:00:06.0[A] -> GSI 17 (level, low) -> IRQ 177 saa7134[0]: found at 0000:00:06.0, rev: 1, irq: 177,
latency: 64, mmio: 0xdfff9c00
saa7134[0]: subsystem: 1894:a006, board: KNC One TV-Station DVR [card=24,autodetected]
saa7134[0]: board init: gpio is 830000
saa7134[0]/oss: mixer input = TV
saa7134[0]: i2c xfer: < a0 00 >
saa7134[0]: i2c xfer: < a1 =94 =18 =06 =a0 =06 =80 =00 =01 =00 =00 =00 =00 =00 =00 =01 =00 =00 =ff =86 =0e =ff =20 =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =01 =40 =01 =02 =02 =03 =01 =03 =06 =ff =01 =df =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff >
saa7134[0]: i2c eeprom 00: 94 18 06 a0 06 80 00 01 00 00 00 00 00 00 01 00 saa7134[0]: i2c eeprom 10: 00 ff 86 0e ff 20 ff ff ff ff ff ff ff ff ff ff saa7134[0]: i2c eeprom 20: 01 40 01 02 02 03 01 03 06 ff 01 df ff ff ff ff saa7134[0]: i2c eeprom 30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff saa7134[0]/core: hwinit2
saa7134[0]/video: set tv norm = PAL
saa7134[0]/video: video input = 0 [Television]
saa7134[0]/video: set tv norm = PAL
saa7134: Loading i2c helpers<7>saa7134[0]: i2c xfer: < c0 >
tuner: chip found at addr 0xc0 i2c-bus saa7134[0]
tuner: type set to 38 (Philips PAL/SECAM multi (FM1216ME MK3)) by saa7134[0] saa7134[0]: i2c xfer: < c2 ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < c4 ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < c6 ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < c8 ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < ca ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < cc ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < ce ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < d0 ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < d2 ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < d4 ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < d6 ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < d8 ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < da ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < dc ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < de ERROR: NO_DEVICE
saa7134[0]: i2c xfer: < 86 >
tda9885/6/7: chip found @ 0x86
saa7134[0]: i2c xfer: < 86 00 c2 10 00 >
saa7134[0]: i2c xfer: < 96 ERROR: NO_DEVICE saa7134[0]/irq[0,-227141]: r=0x20 s=0x00 PE saa7134[0]/irq[1,-227141]: r=0x20 s=0x00 PE saa7134[0]/irq[2,-227141]: r=0x20 s=0x00 PE saa7134[0]/irq[3,-227141]: r=0x20 s=0x00 PE saa7134[0]/irq[4,-227141]: r=0x20 s=0x00 PE saa7134[0]/irq[5,-227141]: r=0x20 s=0x00 PE saa7134[0]/irq[6,-227141]: r=0x20 s=0x00 PE saa7134[0]/irq[7,-227141]: r=0x20 s=0x00 PE saa7134[0]/irq[8,-227141]: r=0x20 s=0x00 PE saa7134[0]/irq[9,-227141]: r=0x20 s=0x00 PE saa7134[0]/irq[10,-227141]: r=0x20 s=0x00 PE
saa7134[0]/irq: looping -- clearing PE (parity error!) enable bit saa7134[0]/irq[0,-227141]: r=0x20 s=0x00 PE
[modprobe hangs]

Looks like it's hanging at request_module("saa7134-empress");

#2 lore raw
Quoted 10 lines · Bernhard Rosenkraenzer
Quoted 2 lines · Bernhard Rosenkraenzer

Trying to modprobe saa7134 with this card results in a hanging modprobe process.

Some more debug info (with 2.6.10-ac1):

saa7130/34: v4l2 driver version 0.2.12 loaded
ACPI: PCI interrupt 0000:00:06.0[A] -> GSI 17 (level, low) -> IRQ 177 saa7134[0]: found at 0000:00:06.0, rev: 1, irq: 177, latency: 64, mmio: 0xdfff9c00 saa7134[0]: subsystem: 1894:a006, board: KNC One TV-Station DVR [card=24,autodetected]

...

Quoted 3 lines

saa7134: Loading i2c helpers<7>saa7134[0]: i2c xfer: < c0 >
tuner: chip found at addr 0xc0 i2c-bus saa7134[0]
tuner: type set to 38 (Philips PAL/SECAM multi (FM1216ME MK3)) by saa7134[0]

...

Quoted 1 line

tda9885/6/7: chip found @ 0x86

...

Quoted 5 lines

saa7134[0]/irq: looping -- clearing PE (parity error!) enable bit saa7134[0]/irq[0,-227141]: r=0x20 s=0x00 PE
[modprobe hangs]

Looks like it's hanging at request_module("saa7134-empress");

I've reported this before:
http://lkml.org/lkml/2004/11/17/275

The patch that Rusty posted doesn't seem to have it made into module-init-tools-3.1. Bummer :-(

The Bug in saa7134-core.c doesn't seem to have been addressed yet, so I suggest the following patch as a temporary measure:

drivers/media/video/saa7134/saa7134-core.c renamed from drivers/media/video/saa7134/saa7134-core.c.orig +4 -4
@@ -940,12 +940,12 @@ static int __devinit saa7134_initdev(str
 		request_module("tuner");
 	if (dev->tda9887_conf)
 		request_module("tda9887");
-  	if (card_is_empress(dev)) {
-		request_module("saa7134-empress");
+	if (card_is_empress(dev)) {
+//		request_module("saa7134-empress");
 		request_module("saa6752hs");
 	}
-  	if (card_is_dvb(dev))
-		request_module("saa7134-dvb");
+//	if (card_is_dvb(dev))
+//		request_module("saa7134-dvb");
 
 	v4l2_prio_init(&dev->prio);
 

Johannes

#3 lore raw
Quoted 5 lines · Johannes Stezenbach

I've reported this before:
http://lkml.org/lkml/2004/11/17/275

The patch that Rusty posted doesn't seem to have it made into module-init-tools-3.1. Bummer :-(

Because it was a dumb idea, as your reply showed. You can't load any dependent modules because this module is not initialized yet (what if you loaded a module that depended on you, then failed your initialization?). Gerd has the real fix, IIRC.

Rusty.

A bad analogy is like a leaky screwdriver -- Richard Braakman

#4 lore raw
Quoted 11 lines · Rusty Russell
Quoted 5 lines · Johannes Stezenbach

I've reported this before:
http://lkml.org/lkml/2004/11/17/275

The patch that Rusty posted doesn't seem to have it made into module-init-tools-3.1. Bummer :-(

Because it was a dumb idea, as your reply showed. You can't load any dependent modules because this module is not initialized yet (what if you loaded a module that depended on you, then failed your initialization?). Gerd has the real fix, IIRC.

Yeah, but IMHO it would be better if modprobe would bail out with an error instead of just hanging. For me hotplug loads the drivers at boot time and it's a bad good thing if the whole boot process hangs.

Johannes