malloc deadlocks & ioslaves not exiting

[Roger Larsson]

On Monday 10 February 2003 20:27, Thiago Macieira wrote:
> 
> Here's something interesting too: while Helio and I were trying to do an 
> initial search on the ioslaves not exiting and consuming too much CPU, we 
> ended up with this backtrace:
> 
> #0  0x40a81a57 in malloc_consolidate (av=0x40ab8640) at malloc.c:2934
> #1  0x40a80754 in fREe (mem=0x82244f8) at malloc.c:3105
> #2  0x40a7f9bc in free (m=0x82244f8) at malloc.c:5525
> #3  0x41197b4a in _XFreeDisplayStructure () from /usr/X11R6/lib/libX11.so.6
> [snip]
> 
> If I remember correctly from what he told me, it's BRANCH of Feb 4th, 
> --enable-fast-malloc=full.

Very interesting. I have the ioslave consuming to much CPU problem.
And I do have --enable-fast-malloc=yes

> 
> Perhaps the two problems are connected?
> 

Not unlikely. Because my conclusion is that some kind of deadlock / livelock
is the core of the ioslave problem.

Are there people that does not have the ioslave problem?
Do you use fast-malloc?

/RogerL

-- 
Roger Larsson
Skellefteå
Sweden