Hi, I am Purna from India. I got this mailing list from libmicrohttpd info page and would like to discuss with you about a crash that we are getting from libmicrohttpd application during our security testing.
Version we are using: libmicrohttpd-0.9.50-1.el7.centos.x86_64.rpm <https://acos.alcatel-lucent.com/frs/download.php/36551/libmicrohttpd-0.9.50-1.el7.centos.x86_64.rpm> Our application is using libmicrohttpd interface to receive http request. When we are doing security testing on our application, we got a crash with the following stack trace. Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fff7279c700 (LWP 31873)] 0x00007ffff6add570 in MHD_http_unescape () from /lib64/libmicrohttpd.so.12 (gdb) bt #0 0x00007ffff6add570 in MHD_http_unescape () from /lib64/libmicrohttpd.so.12 #1 0x00007ffff6ad6172 in MHD_connection_handle_idle () from /lib64/libmicrohttpd.so.12 #2 0x00007ffff6ad6c25 in call_handlers () from /lib64/libmicrohttpd.so.12 #3 0x00007ffff6ad8d2e in MHD_handle_connection () from /lib64/libmicrohttpd.so.12 #4 0x00007ffff609ae25 in start_thread () from /lib64/libpthread.so.0 #5 0x00007ffff5dc834d in clone () from /lib64/libc.so.6 (gdb) Looking at the below statements, the crash is happening in method *http_unescape*, possibly the webserver is trying to decode the received messages which is not encoded already and its crashing I guess. Any thoughts around this? from google, i have seen that MHD_http_unescape () method is having issues earlier, which is fixed in version 0.9.32. But we are still seeing this issue, even if we are using version 0.9.50. Do you have any other suggestions for us to try out? Are we missing any other dependant library or configuration for *libmicrohttpd*? Any help around this will be deeply appreciated. -- Regards Purna
