[mythtv] Suggestion for improving SLOW channel changing inLiveTV.....

Kenneth Aafløy lists at kenneth.aafloy.net
Sun Sep 11 04:13:55 UTC 2005


Søndag 11 september 2005 05:33, skrev Isaac Richards:
> On Saturday 10 September 2005 10:49 pm, Kenneth Aafløy wrote:
> > I'm wondering though, if the truncate of the ringbuffer actually does
> > free the data of a file or not? If so, it could be more efficient to
> > rename the current file, create a new and defer the removal to another
> > thread.
>
> Naw, it just seeks to the beginning of the file.  No actual truncation
> going on, until live-tv is exited.

Only located in the Linux domain currently:

Right, so if the filesystem driver is implemented correctly userspace should 
see a seek, and nothing else. Could it be possible that some filesystems 
schedules the workload of removing the inodes? If not so, it could have an 
impact on the running program. If the workload was deferred to a thread with 
minimal scheduling priority, it could get a boost. Please don't state that 
filesystem x does this or that standard y specifies that, as it's merrily 
words, which seldomly work out when used in real world appliances.

I'm not about to argue this single point to death, I'm just curious as
to how stuff work in practise as opposed to real life.

Btw, Thanks for your support for the other items in the mail,
I'll be expecting a patch real soon ;)

Kenneth


More information about the mythtv-dev mailing list