cpia-0.3.1 SMP results

Peter Pregler Peter.Pregler@risc.uni-linz.ac.at
Wed, 10 Nov 1999 14:02:07 +0100 (CET)


On 10-Nov-99 Bas Rijniersce wrote:
> Hi,
> 
> The last version doesn't lock my machine anymore, thanx Peter :)

That is simply because it does not use polled streaming mode anymore for a
simple read. It will still crash if you use an application that does use
mmap. I have now a smp-machine here at work that I can use to fix the
smp-stuff (until the box is used for production). So I guess I will fix
the smp-stuff sooner or later.

There is one thing about the new driver that might be of interest for the
non-v4l driver users. The new v4l_read uses frame capture instead of stream
capture mode (see the source and the developers guide for more info). With
this method which seems to be more suitable to polling I got frame rates of
about 30 for CIF and up to 60 for QCIF (both with auto compression). For
application writers using the non-v4l module this means that the
streaming-based implementation of the test-client (main.cc) is sub-optimal.
I guess streaming does only make sense for a true interrupt driven capture
though this is not my primary target anymore.

-Peter


> But it doesn't work correctly yet:
>>From dmesg
> ---------
> Linux video capture interface: v1.00
> parport0: PC-style at 0x378, dma 3 [SPP,ECP,ECPEPP,ECPPS2]
> parport0: detected irq 7; use procfs to enable interrupt-driven
> operation.
> V4L Parport CPiA Webcam driver (C) 1999 P.Pregler v0.3
>   using DMA mode (irq 7, DMA 3)
> ppcpia(38576):v4l.c:v4l_initialize(1398):v4l_initialize
> ppcpia(38577):v4l.c:goto_low_power(759):camera now in LOW power state
> ppcpia(38585):v4l.c:goto_high_power(748):camera now in HIGH power state
>   CPIA Version: 1.20 (2.0)
>   CPIA PnP-ID: 0553:0002:0100
>   VP-Version: 1.0 0100
> ppcpia(38597):v4l.c:save_camera_state(797):32/6/92
> ppcpia(38598):v4l.c:save_camera_state(815):0/0/185/0/220/214/214/230
> ppcpia(38599):v4l.c:goto_low_power(759):camera now in LOW power state
> ppcpia(46742):v4l.c:v4l_open(895):v4l_open
> ppcpia(46750):v4l.c:goto_high_power(748):camera now in HIGH power state
> ppcpia(46763):v4l.c:dispatch_commands(512):0 1 0 0 1 1 1 0 
> ppcpia(46781):v4l.c:v4l_ioctl(1020):VIDIOCGCAP
> ppcpia(46781):v4l.c:v4l_ioctl(1109):VIDIOCGWIN
> ppcpia(46781):v4l.c:v4l_ioctl(1074):VIDIOCGPICT
> ppcpia(46781):v4l.c:v4l_ioctl(1084):VIDIOCSPICT
> ppcpia(46781):v4l.c:v4l_ioctl(1090):palette: 4
> ppcpia(46781):v4l.c:v4l_ioctl(1093):depth: 24
> ppcpia(46781):v4l.c:v4l_ioctl(1102):VIDIOCSPICT: 24 / 4 // 32768 / 32768
> / 32768 / 32768
> ppcpia(46781):v4l.c:v4l_ioctl(1119):VIDIOCSWIN
> ppcpia(46781):v4l.c:v4l_ioctl(1199):1/352:288
> ---------
> 
> The last lines (i don't know from which one) is when I start ./gqcam.
> GQcam prints one line:
> flags: 49
> Start drawing the GTK window, but never finishes. Then gqcam is frozen,
> no way to kill it (kill -9 doesn't work!). I have left the program for
> 15-20 minutes, no change. The rest of the machines is working OK. The
> only way to kill gqcam is to reboot.
> 
> Bas
> ----
> Bas Rijniersce                       Phone +31 341 550545
> Oude Telgterweg 81                   Fax +31 341 562940
> 3851 EA  Ermelo                      http://www.brijn.nu
> The Netherlands                      bas@brijn.nu
> -------------------------------------------------------------------------
> ----
> To unsubscribe from this mailinglist, send the line "unsubscribe
> vision-webcam" in the
> body of a message to "majordomo@errors.no".
> 

-------------------------------
Email: Peter.Pregler@risc.uni-linz.ac.at
WWW:   http://www.risc.uni-linz.ac.at/people/ppregler
-----------------------------------------------------------------------------
To unsubscribe from this mailinglist, send the line "unsubscribe vision-webcam" in the
body of a message to "majordomo@errors.no".