MVallevand wrote:
> On Jan 23, 2008 1:29 PM, Roger Heflin <[EMAIL PROTECTED]> wrote:
>
>> The basic startup of mvpmc uses about 8-8.5M, I am not sure where
>> all of that comes from, the code and shared libraries come to
>> about 4M (2.7 exe, 1.2M shared libs).
>
> You're not seeing the first 4M bank with the wrapper filesystem.
>
> Martin
/proc/meminfo says there is 14M of ram that the kernel sees, so how
much does the device actually have?
What exactly is the wrapper filesystem?
At bootup I see about 8-8.5MB show by ps as being used by mvpmc,
with 400-500kb free, and with about 3M tied up in buffers.
About 1M is tied up in the tmpfs for /live/memory.
It looks like one could either just load the top level categories
in the menu, and discard everything else, and then when one goes
down a level again load just the entries that are needed there,
and discard the rest, and this would reduce the amount of ram
required to hold large numbers of recordings, but would increase
the network traffic ( and probably time) needed to produce
the next menu, but would work for larger number of recordings.
If I understand what is happening, one thread has a full listing
of the recordings while a delete is happening, and then a update
comes in and another copy starts being made, and before that finishes
another refresh happens and a 3rd copy comes in, and it just does
not have the ram to deal with all of them, but would eventually clean
everything up given enough time.
Maybe the solution is to either in the delete to save the info to
delete the show, and then remove the program info before doing the
delete so that there is enough space for the new info that will show
up on the refresh (after the delete), or maybe the solution could also be
to temporarily delay/stop the callback event for the update when doing
a delete and call that callback after the delete is done, it very
much looks like the update callback and the delete are interacting
in a very bad way. Mine only fails if I do several deletes quickly,
someone else (with around 800+ recordings) reports that delete always
messes things up.
Comments, Ideas? I have a full git here, and can build my own
dongles, and can duplicate the problem, does the problem sound
right? And what about the possible solution?
Roger
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Mvpmc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mvpmc-users
mvpmc wiki: http://mvpmc.wikispaces.com/
