On 2013/7/24 17:07, Willy Tarreau wrote:
On Wed, Jul 24, 2013 at 04:58:31PM +0800, Godbach wrote:
The mechanism of poisoning/memset is used for memory pool. Memory pool
uses MALLOC in dev7 and CALLOC in master, most chunks of memory are
allocated from memory pool while processing sessions. But what we have
talked about is memory usage which is allocated directly by calloc()
during startup with almost no session.
I agree, but I thought that the difference could lie in *some* of
the pools being used. That said if this is just upon startup, I
agree that the pools should be empty and thus cannot explain the
difference.
So it seems that the memory usage we talk about here should not be
related to this mechanism in my opinion.
Indeed. I have no idea why we're observing these differences, and I
don't know if the libc uses heuristics to decide to memset() the
area or not.
I think we'd better define a zalloc() function aimed at replacing
calloc() and which would always clear the area, than rely on some
random behaviour we have no control over.
Willy
Yes, it is a good choice to clear the memory area explicitly. I will go
on reporting if anything new I get about this issue.
--
Best Regards,
Godbach
