Re: comedi_mark_buffer_read() not working ?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tuesday 19 November 2002 03:09 pm, Daniel Koch wrote:
> Hi,
>
> I still got that nasty buffer-overflow  problem I already mentioned some
> mails before. If I try to use mmap with the following options, the program
> exits after some time, depending on the buffersize I set with
> comedi_set_buffer_size():
>
> ./mmap -n16 -F100000 -N1000000.
>
> Dmesg reports a buffer-overflow, and when I do

Well, if you don't redirect the mmap demo's output to a file or something, 
then printing the results to your screen is going to slow down the mmap demo 
so much it wont be able to keep up with the data coming in from the board.  
At least at the speed you are taking data.  Also, I mmap.c is currently hard 
coded to use TRIG_NONE as a stop_src, so it isn't going to stop after your 
million scans.

>
> ./mmap -n16 -F100000 -N1000000 -v
>
> I get also get something like
>
> front = 0 back = 4678
>
> Why does the program exit there? Is it possible that
>

That's actually a bug in demo/mmap.c, it was fixed in the cvs version a few 
months ago.

- -- 
Frank

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE92t4l5vihyNWuA4URAqnhAJ97YhgZN08kLZc6tG2HgtDuc4iKFACcDC/o
+SrMrHmYw6VVSNsqrogL2Ug=
=8SWh
-----END PGP SIGNATURE-----

Received on 2002-11-20Z00:58:11