RFR[9u-dev] 8130425: libjvm crash due to stack overflow in executables with 32k tbss/tdata

Andrew Haley aph at redhat.com
Wed Jan 20 15:33:10 UTC 2016

On 01/20/2016 03:19 PM, Peter Levart wrote:
> So if we add 32k (or whatever the max. possible 
> overhead of TLS is) to the existing 32k, that should be the default for 
> reaper thread. 

I'm not sure how many thread-local variables are possible in theory,
but I suspect it's unlimited.  In machines with growable memory
regions for stacks allocating the default stack has very little
overhead, so it doesn't matter.  Systems which pre-allocate their
thread stacks have their own problems.  These tend to be special kinds
of cloud containers and some turnkey systems which cannot tolerate an
out-of-memory error.


More information about the core-libs-dev mailing list