Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 14 Aug 1998 08:46:21 -0400
From:      Randall Hopper <rhh@ct.picker.com>
To:        Gregory Hosler <greg@camelot.eno.ericsson.se>, Roger Hardiman <roger@cs.strath.ac.uk>
Cc:        multimedia@FreeBSD.ORG
Subject:   Re: Putting the bt848 driver into the GENERIC kernel
Message-ID:  <19980814084621.A6084@ct.picker.com>
In-Reply-To: <XFMail.980814084226.greg@camelot.eno.ericsson.se>; from Gregory Hosler on Fri, Aug 14, 1998 at 08:42:26AM %2B0730
References:  <35D2C9DC.47D0@cs.strath.ac.uk> <XFMail.980814084226.greg@camelot.eno.ericsson.se>

next in thread | previous in thread | raw e-mail | index | archive | help
Gregory Hosler:
 |On 13-Aug-98 Roger Hardiman wrote:
 |>Soren suggested a BOOTKERNEL and BIGKERNEL.
 |
 |why not just compile the driver as a loadable module ?
 |
 |just curious,

Roger recounted through that in a recent thread (attached).  The issue is
the vm_contig_alloc call the driver uses to alloc the DMA phys memory
doesn't work after the system has booted.  It doesn't have the smarts to
bump/swap/rearrange phys page mappings to get suffic contig phys mem.

Know any VM hackers that could take a crack at it?  Having a bt848 driver
LKM would be really cool. :-)  As would not having the driver alloc the
uffer on probe/attach, but rather on open where it makes the most sense.

Randall



Date: Thu, 13 Aug 1998 17:16:48 +0100
From: Roger Hardiman <roger@cs.strath.ac.uk>
Organization: University of Strathclyde
Subject: Re: FXTV and DGA

Scott
> 
> This may be a dumb suggestion (I don't use fxtv so I have no idea if
> this will work or not), but couldn't the driver be made into an LKM?

Funny you should mention it, but Amancio, Randall and later myself
talked about this around 8 months ago.

The problem was basically this.
You need to allocate configuous memory for the frame buffer.

At boot time, this is easy. Memory is empty and you can allocate
a 864k block of memory with no gaps in it.
However, I was told that by the time the system has booted,
and the LKM executed, there is quite possible you cannot get a
contiguous block of memory as the memory is fragmented.

Basically, you need a memeory manager that can shift memory blocks
around making a large space.

I was told the current memory management code cannot do this.


But if we could do it, then the LKM would be great.

Bye
Roger

Roger Hardiman
Strathclyde Uni Telepresence Group

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

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?19980814084621.A6084>