the __morestack in ecall when using the multithreads

#0 raise (sig=) at ../sysdeps/unix/sysv/linux/raise.c:51
intel/confidential-computing.sgx#1 0x00007f4b80653902 in sig_handler(int, siginfo_t*, void*) () from /usr/lib/x86_64-linux-gnu/libsgx_urts.so
intel/confidential-computing.sgx#2
intel/confidential-computing.sgx#3 0x00007f4b8064c6f8 in __morestack () from /usr/lib/x86_64-linux-gnu/libsgx_urts.so
intel/confidential-computing.sgx#4 0x00007f4b8064f61f in CEnclave::ecall(int, void const*, void*, bool) () from /usr/lib/x86_64-linux-gnu/libsgx_urts.so
intel/confidential-computing.sgx#5 0x00007f4b8065113f in _sgx_ecall(unsigned long, int, void const*, void*, bool) [clone .part.0] () from /usr/lib/x86_64-linux-gnu/libsgx_urts.so

=====
1. i have confronted the ecall coredump in multithreads ecall, but when i changed the 6 threads to 2 threads, the program can continue but core again.
2.it seems not the interface parameters problem or multithreads question.
about other caused?

Start with the provided backtrace at __morestack, CEnclave::ecall, and _sgx_ecall, then reproduce the crash with six and two threads. Compare the runs to determine whether the failure is tied to multithreaded ecall behavior, and document a specific cause plus a validated fix or workaround.

Written by the indexing model from the issue text.