[linux-dvb] Re: Suggestion for channel tuning

[Klaus Schmidinger <Klaus.Schmidinger@cadsoft.de>]

Matthias Weingart wrote:
> 
> On 24 Apr 2001 ma.hoff@bigfoot.com wrote:
> 
> > > I think (k6 2-450 on MSI mobo) that it in my case is related to the fact
> > > that Linux disables DMA by default on this mobo chipset (forcing
> > > DMA srewed op a hard disk, so I am not going to try again).
> > > So maybe the trick here is to use DMA and maybe a mobo that allows it in
> > > Linux, that should allow some more data output.
> > > On the old mobo I had DMA disabled, and multi sector set with hpparm.
> > > And the multi-sector setting made a WHOLE lot of difference (for the better).
> > > This is all I understand about it at the moment, just a matter of
> > > data throughput.
> > > This is only for the 'hickups' in the picture though.
> > > Regards
> > > Jan
> > Is there a way to check this ? I thought if you disable dma - or if linux doesn't support the motherbord chipset it will disable dma - harddisk will run in PIO Mode - right ?
> 
> hdparm /dev/hda
> tells you what is on or off
> 
> It was already discussed here: VDR does not buffer disc-io yet. If vdr
> will use a ringbuffer in memory that disc-io problem will gone :-)

VDR 0.72 _does_ use a ring buffer for replay. However, it is handled by a single
thread, which may still cause timing problems.

> But first the whole thing must run with the new API ;-/

Yes, and once that's accomplished I'll use a double threaded ring buffer for
replay (as is already the case for recording).

Klaus
-- 
_______________________________________________________________

Klaus Schmidinger                       Phone: +49-8635-6989-10
CadSoft Computer GmbH                   Fax:   +49-8635-6989-40
Hofmark 2                               Email:   kls@cadsoft.de
D-84568 Pleiskirchen, Germany           URL:     www.cadsoft.de
_______________________________________________________________


-- 
Info:
To unsubscribe send a mail to listar@linuxtv.org with "unsubscribe linux-dvb" as subject.