Re: CX8800 driver and 2.6.15-RC2

[Michael Krufky]

Lukas Hejtmanek wrote:

> Hello,
>
> I tried cx8800 driver with Leadtek Winfast 2000XP Expert card.
> This is from dmesg:
> cx2388x v4l2 driver version 0.0.5 loaded
> ACPI: PCI Interrupt 0000:01:05.0[A] -> GSI 18 (level, low) -> IRQ 19
> CORE cx88[0]: subsystem: 107d:6611, board: Leadtek Winfast 2000XP Expert
> [card=5,autodetected]
> TV tuner 44 at 0x1fe, Radio tuner -1 at 0x1fe
> cx88[0]: Leadtek Winfast 2000XP Expert config: tuner=38, eeprom[0]=0x01
> input: cx88 IR (Leadtek Winfast 2000XP as /class/input/input2
> cx88[0]/0: found at 0000:01:05.0, rev: 5, irq: 19, latency: 64, mmio: 0xfd000000
> tuner 0-0060: All bytes are equal. It is not a TEA5767
> tuner 0-0060: chip found @ 0xc0 (cx88[0])
> tuner 0-0060: type set to 38 (Philips PAL/SECAM multi (FM1216ME MK3))
> tda9887 0-0043: chip found @ 0x86 (cx88[0])
> cx88[0]/0: registered device video0 [v4l2]
> cx88[0]/0: registered device vbi0
> cx88[0]/0: registered device radio0
>
> seems to be ok.
>
> However, using xawtv I got:
> This is xawtv-3.94, running on Linux/i686 (2.6.15-rc2-git1)
> /dev/video0 [v4l2]: ioctl VIDIOC_G_FBUF: Invalid argument
> v4l-conf had some trouble, trying to continue anyway
> ioctl: VIDIOC_G_FBUF(capability=0x0 [];flags=0x0
> [];base=(nil);fmt.width=0;fmt.height=0;fmt.pixelformat=0x00000000
> [....];fmt.field=ANY;fmt.bytesperline=0;fmt.sizeimage=0;fmt.colorspace=unknown;fmt.priv=0):
> Invalid argument
> ioctl:
> VIDIOC_TRY_FMT(type=VIDEO_OVERLAY;fmt.win.w.left=852;fmt.win.w.top=45;fmt.win.w.width=384;fmt.win.w.height=288;fmt.win.field=ANY;fmt.win.chromakey=0;fmt.win.clips=(nil);fmt.win.clipcount=0;fmt.win.bitmap=(nil)):
> Invalid argument
> ioctl:
> VIDIOC_S_FMT(type=VIDEO_OVERLAY;fmt.win.w.left=852;fmt.win.w.top=45;fmt.win.w.width=384;fmt.win.w.height=288;fmt.win.field=ANY;fmt.win.chromakey=0;fmt.win.clips=(nil);fmt.win.clipcount=0;fmt.win.bitmap=(nil)):
> Invalid argument
> ioctl: VIDIOC_OVERLAY(int=0): Invalid argument
> ioctl: VIDIOC_QBUF(index=0;type=VIDEO_CAPTURE;bytesused=0;flags=0x0
> [];field=ANY;;timecode.type=0;timecode.flags=0;timecode.frames=0;timecode.seconds=0;timecode.minutes=0;timecode.hours=0;timecode.userbits="";sequence=0;memory=MMAP):
> Bad address
> ioctl: VIDIOC_QBUF(index=0;type=VIDEO_CAPTURE;bytesused=0;flags=0x0
> [];field=ANY;;timecode.type=0;timecode.flags=0;timecode.frames=0;timecode.seconds=0;timecode.minutes=0;timecode.hours=0;timecode.userbits="";sequence=0;memory=MMAP):
> Bad address
> ioctl: VIDIOC_DQBUF(index=0;type=VIDEO_CAPTURE;bytesused=0;flags=0x0
> [];field=ANY;;timecode.type=0;timecode.flags=0;timecode.frames=0;timecode.seconds=0;timecode.minutes=0;timecode.hours=0;timecode.userbits="";sequence=0;memory=unknown):
> Invalid argument
> ioctl: VIDIOC_QBUF(index=0;type=VIDEO_CAPTURE;bytesused=0;flags=0x0
> [];field=ANY;;timecode.type=0;timecode.flags=0;timecode.frames=0;timecode.seconds=0;timecode.minutes=0;timecode.hours=0;timecode.userbits="";sequence=0;memory=MMAP):
> Bad address
> ioctl: VIDIOC_QBUF(index=0;type=VIDEO_CAPTURE;bytesused=0;flags=0x0
> [];field=ANY;;timecode.type=0;timecode.flags=0;timecode.frames=0;timecode.seconds=0;timecode.minutes=0;timecode.hours=0;timecode.userbits="";sequence=0;memory=MMAP):
> Bad address
>
> I inserted video_buf module with debug option and I got:
> kernel: vbuf: reqbufs: bufs=2, size=0x36000 [108 pages total]
> kernel: vbuf: mmap setup: 2 buffers, 221184 bytes each
> kernel: vbuf: mmap ca74c420: q=ca8fe848 ae94c000-ae982000 pgoff 00000000 bufs
> 0-0
> kernel: vbuf: mmap ca74c3e0: q=ca8fe848 ae916000-ae94c000 pgoff 00000036 bufs
> 1-1
> kernel: vbuf: init user [0xae94c000+0x36000 => 54 pages]
> kernel: vbuf: get_user_pages: err=-14 [0]
> kernel: vbuf: init user [0xae94c000+0x36000 => 54 pages]
> kernel: vbuf: get_user_pages: err=-14 [0]
> kernel: vbuf: munmap ca74c420 q=ca8fe848
> kernel: vbuf: munmap ca74c3e0 q=ca8fe848
> kernel: vbuf: init user [0x8476fc8+0x6c000 => 109 pages]
> kernel: vbuf: reqbufs: bufs=2, size=0xe1000 [450 pages total]
> kernel: vbuf: mmap setup: 2 buffers, 921600 bytes each
> kernel: vbuf: mmap cfdd4f80: q=cec35848 b7304000-b73e5000 pgoff 00000000 bufs
> 0-0
> kernel: vbuf: init user [0xb7304000+0xe1000 => 225 pages]
> kernel: vbuf: get_user_pages: err=-14 [0]
> kernel: vbuf: munmap cfdd4f80 q=cec35848
>
>
> Is it a bug in kernel modules?
Close, but no cigar ;-)

It was a memory management bug.... Already fixed in -rc3 (where new bugs 
were introduced) ...

-rc4 isn't bad, but a whole slew of v4l / dvb bugfixes went in JUST 
after -rc4 release...

Can you try 2.6.15-rc4-git1 and let us know how things are?

Cheers,

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