Skip site navigation (1)Skip section navigation (2)
Date:      29 Jun 1999 13:30:10 +0200
From:      Frode Vatvedt Fjeld <frodef@acm.org>
To:        freebsd-multimedia@freebsd.org
Subject:   Re: bt848 channel frequencies
Message-ID:  <2h3dzbs0ql.fsf@dslab7.cs.uit.no>
In-Reply-To: Roger Hardiman's message of "Tue, 29 Jun 1999 11:42:34 %2B0100"
References:  <19990627074316.A1600@ipass.net> <199906272257.QAA07116@orthanc.ab.ca> <19990627194019.A1726@ipass.net> <2hzp1ksk05.fsf@dslab7.cs.uit.no> <37783907.94141BC7@cs.strath.ac.uk> <2h9093s4q3.fsf@dslab7.cs.uit.no> <3778A31A.794B@cs.strath.ac.uk>

next in thread | previous in thread | raw e-mail | index | archive | help
Roger Hardiman <roger@cs.strath.ac.uk> writes:
> Well, you can use read().  It blocks on a frame (or a field).

Hum.. I've always assumed that read() is a lot slower than mmap(), but
I'm not really sure why. Do you think the performance of read() is
comparable to mmap() etc.?

> What is unstable in -stable?

Right now, I can crash my machine with something like "cat file
>/dev/fd0" (although this worked fine earlier this year).

There's something wrong with the SYSVSHM code that causes a crash
(very consistently) under certain conditions when there are no more
SHM buffers.

The meteor driver crashes consistently. I've submitted a pr with a fix
many months ago, but nobody with CVS access seems to care (hence I'm
not very eager to make more fixes). I'd guess this fix applies to
meteor in -current too.

> What [bktr] problems. Can you give some examples?  I've not had any
> problems myself with the API.

Right now I observe this behavior in my app:

This works (I get a working 768x576 YUV output):

  * Set size 384x288 and RGB format
  * Change size to 768x576
  * Change format to YUV

This doesn't work (the 768x576 YUV output is corrupted, mostly black
with a green vertical bar to the right):

  * Set size 384x288 and YUV format (works)
  * Change size to 768x576

However, if i continue to do this:

  * Set format to RGB
  * Set format to YUV

..the 768x576 YUV output is fine again.

I'm not 100% positive this is a fault in bktr, but, like I said, it is
too hard to tell from the bktr source. And (almost) the same
application code works fine with meteor.

-- 
Frode Vatvedt Fjeld


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-multimedia" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?2h3dzbs0ql.fsf>