< Previous by Date Date Index Next by Date >
< Previous in Thread Thread Index  

Re: [reSIProcate] Use of reSiprocate


We have expended considerable effort searching for and plugging memory leaks in the stack. We currently do not know of any leaks in the stack. I am not an authority on DUM, but Scott has done a lot of leak-hunting, and he does know DUM, so I imagine it is pretty tight there too. I have recently done some work on repro to tighten it up.

As for allocation sizes, there may be some cases where small things are heap-allocated, but I have experimented with bulk allocation of some of the commonly used small(ish) objects, and I saw no performance gain. The modern malloc libraries seem to cope with/ prevent memory fragmentation pretty well (at least on linux/OS X).

Best regards,
Byron Campen

Can anyone tell me if the reSiprocate stack has memory leaks? I just want
to know this before I decide on using it.

Also, what is the smallest size memory block it will allocate? A lot of small allocations/deallocations can fragment memory and slow performance.

Thanks...

_______________________________________________
resiprocate-devel mailing list
resiprocate-devel@xxxxxxxxxxxxxxxxxxx
https://list.sipfoundry.org/mailman/listinfo/resiprocate-devel

Attachment: smime.p7s
Description: S/MIME cryptographic signature