[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gpsd-dev] [gpsd-users] Building the master branch and the split24 o
|
From: |
Gary E. Miller |
|
Subject: |
Re: [gpsd-dev] [gpsd-users] Building the master branch and the split24 option |
|
Date: |
Mon, 21 Oct 2013 12:19:44 -0700 |
Yo Eric!
On Mon, 21 Oct 2013 15:01:12 -0400
"Eric S. Raymond" <address@hidden> wrote:
> Gary E. Miller <address@hidden>:
> > On Mon, 21 Oct 2013 17:13:46 +0200
> > "address@hidden" <address@hidden> wrote:
> >
> > > i have seen this error from time to time during the last weeks, it
> > > comes and goes.
> >
> > I think this is related to the dropping of root at random times
> > during initialization. Eric has been working on this.
> >
> > The problem is root is needed at initialization in the PPS thread
> > but root is droppedin the main thread. Depending on the relative
> > progress of the threads rooot may be dropped to soon.
>
> I don't think this can be the explanation. Remember, Ray says he
> doesn't get the error with 3.9 but does with git head. If the
> thread-initialization race were at fault it shhould be the other way
> around - head has a mutex wait for the initialization to finish, 3.9
> does not.
The mutex ends too early. So the effect is for the main loop to wait
for the mutex, then drop root at one of the worser times to do so.
It is on my TODO list, but the doc discussion has gotten the list humming.
RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97701
address@hidden Tel:+1(541)382-8588
signature.asc
Description: PGP signature