Am Dienstag, den 24.03.2009, 13:42 -0600 schrieb Gordon Smith: > On Tue, Mar 24, 2009 at 1:15 PM, Hans Verkuil <hverkuil@xxxxxxxxx> wrote: > > On Tuesday 24 March 2009 19:47:13 hermann pitton wrote: > >> Am Dienstag, den 24.03.2009, 03:03 +0100 schrieb hermann pitton: > >> > Hi, > >> > > >> > Am Montag, den 23.03.2009, 11:33 -0600 schrieb Gordon Smith: > >> > > On Fri, Mar 20, 2009 at 10:02 PM, hermann pitton > >> > > > >> > > <hermann-pitton@xxxxxxxx> wrote: > >> > > > Hi, > >> > > > > >> > > > Am Freitag, den 20.03.2009, 14:24 -0600 schrieb Gordon Smith: > >> > > > > Hello - > >> > > > > > >> > > > > I'm unable to read or stream compressed data from > >> > > > > saa7134/saa6752hs. > >> > > > > > >> > > > > I have a RTD Technologies VFG7350 (saa7134 based, two channel, > >> > > > > hardware encoder per channel, no tuner) running current v4l-dvb > >> > > > > in debian 2.6.26-1. > >> > > > > > >> > > > > MPEG2-TS data is normally available on /dev/video2 and > >> > > > > /dev/video3. > >> > > > > > >> > > > > Previously there were parameters for the saa6752hs module named > >> > > > > "force" and "ignore" to modify i2c addresses. The current module > >> > > > > appears to lack those parameters and may be using incorrect i2c > >> > > > > addresses. > >> > > > > > >> > > > > Current dmesg: > >> > > > > > >> > > > > [ 13.388944] saa6752hs 3-0020: chip found @ 0x40 (saa7133[0]) > >> > > > > [ 13.588458] saa6752hs 4-0020: chip found @ 0x40 (saa7133[1]) > >> > > > > > >> > > > > Prior dmesg (~2.6.25-gentoo-r7 + v4l-dvb ???): > >> > > > > > >> > > > > saa6752hs 1-0021: saa6752hs: chip found @ 0x42 > >> > > > > saa6752hs 1-0021: saa6752hs: chip found @ 0x42 > >> > > > > > >> > > > > Prior modprobe.conf entry: > >> > > > > options saa6752hs force=0x1,0x21,0x2,0x21 > >> > > > > ignore=0x0,0x20,0x1,0x20,0x2,0x20 > >> > >> It was disabled by Hans when converting to v4l2_subdev here. > >> http://linuxtv.org/hg/v4l-dvb/rev/c41af551e79f > >> > >> It is only valid for kernels < 2.6.22 now in saa6752hs. > >> > >> +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 22) > >> /* Addresses to scan */ > >> static unsigned short normal_i2c[] = {0x20, I2C_CLIENT_END}; > >> > >> I2C_CLIENT_INSMOD; > >> +#endif > >> > >> And we only have that single 0x20 address in saa7134-core.c etc. > >> That should be the problem. > >> > >> > > > > $ v4l2-dbg --device /dev/video2 --info > >> > > > > Driver info: > >> > > > > Driver name : saa7134 > >> > > > > Card type : RTD Embedded Technologies VFG73 > >> > > > > Bus info : PCI:0000:02:08.0 > >> > > > > Driver version: 526 > >> > > > > Capabilities : 0x05000001 > >> > > > > Video Capture > >> > > > > Read/Write > >> > > > > Streaming > >> > > > > > >> > > > > Streaming is a listed capability but the capture example at > >> > > > > http://v4l2spec.bytesex.org/spec/capture-example.html fails > >> > > > > during request for buffers. > >> > > > > > >> > > > > $ v4l2-capture --device /dev/video2 --mmap > >> > > > > /dev/video2 does not support memory mapping > >> > > > > > >> > > > > v4l2-capture.c: > >> > > > > req.count = 4; > >> > > > > req.type = V4L2_BUF_TYPE_VIDEO_CAPTURE; > >> > > > > req.memory = V4L2_MEMORY_MMAP; > >> > > > > > >> > > > > if (-1 == xioctl (fd, VIDIOC_REQBUFS, &req)) { > >> > > > > if (EINVAL == errno) { > >> > > > > fprintf (stderr, "%s does not support " > >> > > > > "memory mapping\n", dev_name); > >> > > > > > >> > > > > > >> > > > > A read() results in EIO error: > >> > > > > > >> > > > > $ v4l2-capture --device /dev/video0 --read > >> > > > > read error 5, Input/output error > >> > > > > > >> > > > > v4l2-capture.c: > >> > > > > if (-1 == read (fd, buffers[0].start, > >> > > > > buffers[0].length)) { switch (errno) { > >> > > > > ... > >> > > > > default: > >> > > > > errno_exit ("read"); > >> > > > > > >> > > > > > >> > > > > This behavior does not change if the saa6752hs module is not > >> > > > > loaded. > >> > > > > > >> > > > > Is there still a way to modify the i2c address(es) for the > >> > > > > saa6752hs module? > >> > > > > > >> > > > > Any pointers are appreciated. > >> > > > > > >> > > > > Gordon > >> > > > > >> > > > thanks for the report. > >> > > > > >> > > > It was probably forgotten that the prior insmod option had a > >> > > > reason. > >> > > > > >> > > > Try to change it to 0x21 in saa7134-i2c.c > >> > > > > >> > > > static char *i2c_devs[128] = { > >> > > > [ 0x20 ] = "mpeg encoder (saa6752hs)", > >> > > > [ 0xa0 >> 1 ] = "eeprom", > >> > > > [ 0xc0 >> 1 ] = "tuner (analog)", > >> > > > [ 0x86 >> 1 ] = "tda9887", > >> > > > [ 0x5a >> 1 ] = "remote control", > >> > > > }; > >> > > > > >> > > > and report if your cards a usable again. > >> > > > > >> > > > Seems we need the chip address per card without that insmod option > >> > > > and reliable probing. > >> > > > > >> > > > Cheers, > >> > > > Hermann > >> > > > >> > > Hermann - > >> > > > >> > > I made the change to saa7134-i2c.c but the i2c address did not > >> > > change. I added my initials (gms) to dmesg, so I know I'm loading the > >> > > new module. > >> > > > >> > > I set i2c_debug=1 for saa7134. Here is one device: > >> > > > >> > > [ 13.369175] saa7130/34: v4l2 driver version 0.2.14-gms loaded > >> > > [ 13.369294] saa7133[0]: found at 0000:02:08.0, rev: 17, irq: 11, > >> > > latency: 32, mmio: 0x80000000 > >> > > [ 13.369310] saa7133[0]: subsystem: 1435:7350, board: RTD Embedded > >> > > Technologies VFG7350 [card=72,autodetected] > >> > > [ 13.369335] saa7133[0]: board init: gpio is 10000 > >> > > [ 13.472509] saa7133[0]: i2c xfer: < a0 00 > > >> > > [ 13.480139] saa7133[0]: i2c xfer: < a1 =35 =14 =50 =73 =ff =ff =ff > >> > > =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff > >> > > ....snip... > >> > > =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff > >> > > =ff =ff > [ 13.520135] saa7133[0]: i2c eeprom 00: 35 14 50 73 ff ff > >> > > ff ff ff ff ff ff ff ff ff ff > >> > > ....snip... > >> > > [ 13.520467] saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff > >> > > ff ff ff ff ff ff ff > >> > > [ 13.608743] saa6752hs 1-0020: chip found (gms) @ 0x40 (saa7133[0]) > >> > > [ 13.608762] saa7133[0]: i2c xfer: < 40 13 > > >> > > [ 13.616164] saa7133[0]: i2c xfer: < 41 =13 =13 =13 =13 =13 =13 =13 > >> > > =13 =13 =13 =13 =13 > > >> > > [ 13.624161] saa6752hs i2c attach [addr=0x20,client=saa6752hs] > >> > > [ 13.624337] saa7133[0]: registered device video0 [v4l2] > >> > > [ 13.624390] saa7133[0]: registered device vbi0 > >> > > >> > yes, sorry, there is more to change. > >> > > >> > The 0x20 address is also hard coded in saa7134-core.c, > >> > saa7134-empress.c and saa6752hs.c. > >> > > >> > I also don't find your old insmod option for multiple addresses back to > >> > 2.6.18 for now. I'll try to find some time to look up the history next > >> > days. > >> > >> See above, static unsigned short normal_i2c[] seems to have allowed to > >> set even 0x21, but only 0x20 was visible there. > >> > >> > With the wrong address and i2c_debug=1 you should see a bunch of ERROR: > >> > NO DEVICE stuff from 0x40 and 0x41 if you try "cat /dev/video2" mpeg. > >> > > >> > Is it really detected at 0x42 with i2c_scan=1 ? > >> > Except on i2c_debug level the code seems to tolerate all wrong > >> > addresses without warnings. > >> > >> Please provide the i2c_scan with the device detected at 0x42. I seem to > >> find all card related mails from Robert W. Boone from RTD, but not that > >> different address information. > >> > >> We should wait until Hans is back from the Linux embedded conference, > >> since he is working on it and I don't even have a working empress > >> device, but it seems we need a solution for multiple addresses here too. > >> > >> If I did not miss anything greping around after lunch, the attached > >> patch might work as an interim for you. > > > > It's good to know that this device can also appear on address 0x42. I'll fix > > this properly this weekend. I need to do some work to add saa6588 support > > to saa7134 as well, so I can do this at the same time. > > > > BTW, please post to the new linux-media list rather than the obsolete > > video4linux list. I'm no longer subscribed there which is why I didn't see > > this earlier. > > > > Regards, > > > > Hans > > > > -- > > Hans Verkuil - video4linux developer - sponsored by TANDBERG > > > > I have Hermann's data request to verify 0x42 address. > > >From RTD Linux Application Note, there is a pcf8574 for digital I/O at 0x40.. > The saa6752hs is at 0x21. > > Here is edited dmesg with i2_scan for first channel: > > [ 13.268713] Linux video capture interface: v2.00 > [ 13.418744] saa7130/34: v4l2 driver version 0.2.14-gms1 loaded > [ 13.418875] saa7133[0]: found at 0000:02:08.0, rev: 17, irq: 11, > latency: 32, mmio: 0x80000000 > [ 13.418891] saa7133[0]: subsystem: 1435:7350, board: RTD Embedded > Technologies VFG7350 [card=72,autodetected] > [ 13.418918] saa7133[0]: board init: gpio is 10000 > [ 13.520522] saa7133[0]: i2c xfer: < a0 00 > > [ 13.528156] saa7133[0]: i2c xfer: < a1 =35 =14 =50 =73 =ff =ff =ff > =ff =ff =ff =ff =ff =ff =ff =ff =ff > ....snip... > =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff > =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff =ff > > [ 13.568135] saa7133[0]: i2c eeprom 00: 35 14 50 73 ff ff ff ff ff > ff ff ff ff ff ff ff > ....snip... > [ 13.568467] saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff > ff ff ff ff ff ff ff > [ 13.812492] saa7133[1]: i2c xfer: < 01 ERROR: NO_DEVICE > [ 13.812640] saa7133[1]: i2c xfer: < 03 ERROR: NO_DEVICE > ....snip... > [ 13.816919] saa7133[1]: i2c xfer: < 3d ERROR: NO_DEVICE > [ 13.817066] saa7133[1]: i2c xfer: < 3f ERROR: NO_DEVICE > [ 13.817214] saa7133[1]: i2c xfer: < 41 > > [ 13.824136] saa7133[1]: i2c xfer: < 43 > > [ 13.832136] saa7133[1]: i2c xfer: < 45 ERROR: NO_DEVICE > [ 13.832285] saa7133[1]: i2c xfer: < 47 ERROR: NO_DEVICE > ...snip... > [ 13.838640] saa7133[1]: i2c xfer: < 9d ERROR: NO_DEVICE > [ 13.838788] saa7133[1]: i2c xfer: < 9f ERROR: NO_DEVICE > [ 13.838935] saa7133[1]: i2c xfer: < a1 > > [ 13.844136] saa7133[1]: i2c xfer: < a3 ERROR: NO_DEVICE > [ 13.844285] saa7133[1]: i2c xfer: < a5 ERROR: NO_DEVICE > ....snip... > [ 13.850787] saa7133[1]: i2c xfer: < fd ERROR: NO_DEVICE > [ 13.850935] saa7133[1]: i2c xfer: < ff ERROR: NO_DEVICE > [ 13.678908] saa6752hs 1-0021: chip found (gms1) @ 0x42 (saa7133[0]) > [ 13.678926] saa7133[0]: i2c xfer: < 42 13 > > [ 13.684221] saa7133[0]: i2c xfer: < 43 =05 =07 =00 =00 =56 =31 =42 > =34 =02 =06 =00 =00 > > [ 13.692156] saa6752hs 1-0021: support AC-3 > [ 13.692177] saa6752hs i2c attach [addr=0x21,client=saa6752hs] > [ 13.692380] saa7133[0]: registered device video0 [v4l2] > [ 13.692433] saa7133[0]: registered device vbi0 Ah, that pcf8574 was most confusing. Gordon and Hans, Thanks. Cheers, Hermann -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
