Yep, me again. Sorry.
I am capturing uncompressed with two LML33 cards (using an older zoran
driver, as I seem unable to get the new one to work). It seems there is
memory strangeness going on at the driver level, unfortunately.
As the cards are initialized, they get their own mmap:ed memory area,
and the buffers all have their own space and memory (and more than
sufficient).
When running with both, however, I get bits of the input from one card
into the input of the other. It is a bit non-deterministic as they run
in separate threads, but it is very clear it is happening - exactly how
it looks depends on which card is triggered to capture when.
Looking at the messages with dmesg, I notice something which may or may
not indicate a problem:
LML33[0]: mmap at 0x4091f000, size 4194304
V4L remap page range 0 0x1f400000 2097152 to 0x4091f000
V4L remap page range 1 0x1f600000 2097152 to 0x40b1f000
...
LML33[1]: mmap at 0x40d1f000, size 4194304
V4L remap page range 0 0x1f400000 2097152 to 0x40d1f000
V4L remap page range 1 0x1f600000 2097152 to 0x40f1f000
In other words, the two buffer areas are mapping to the same physical
address. Could this be the source of my trouble?
I would be grateful for any kind of response.
--
Trust the Computer. The Computer is your friend.
Tel. +46-046 222 8588 Dr. Janne Mor�n (mr)
Home: +46-046 211 4973 Dept. of Cognitive Science
Fax: +46-046 222 9758 Kungshuset, Lund
S-222 22 Lund, Sweden
-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
_______________________________________________
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users
