Monday, 4 April 2022

[gccsdk] OpenSSL issues

Hi

I'm trying to track down an issue with ssl in python, which used OpenSSL
1.1. To try to reduce the number of moving parts, I've managed to narrow
it down to a smaller test.

The setup is .. the C code below (which runs fine on linux) has been
compiled on RISC OS (native) with GCC10 (buult from the auto builder).
LibSSL (1.1.0) is the one I built with the auto-builder, but I had to
tweak it to use the 1.1.0n (rather than m) sources and I built it for
armv6z (IIRC) instaad of armv3.


#include <openssl/ssl.h>

int main(void)
{
SSL_CTX *ctx = SSL_CTX_new( TLS_client_method() );
printf("ctx: %x\n", ctx);
printf("load_verify_locations: %d\n",
SSL_CTX_load_verify_locations(ctx, "cacert.pem", NULL) );
SSL_CTX_free(ctx);
printf("Bye!\n");
}


*gcc -c ssltest.c -ILibSSL11:
*gcc ssltest.o -o ssltest -LLibSSLl11: -lssl -lcrypto -ldl
*ssltest
ctx: 71a88

Fatal signal received: Segmentation fault

Stack backtrace:

Running thread 0x58834 (Main Thread)
(50df5f40) lr: 48fa9dbc : 7ADBC :
SharedLibs:lib.armeabihf.libunixlib/so/5/0/0
(50df5fb0) lr: 48fab450 : 7C450 :
SharedLibs:lib.armeabihf.libunixlib/so/5/0/0
(50df5fb4) lr: 50ef6d3c
(48fab450) lr: ef020014
Stack frame has gone out of bounds with address ebfff8ec

*

If I rename the pem file, it reports SSL_CTX_load_verify_locations: 0
(which one would expect) then has a similar error (wihtout pritning Bye!)


Anyone any ideas?

Cheers

Chris
--
Chris Johns

This email is measured by weight not volume - contents may settle during
transit.

_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK

No comments:

Post a Comment