-
Notifications
You must be signed in to change notification settings - Fork 2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
luajit segmentation fault #2
Comments
Can you provide the version of nginx, and the stack backtrace when nginx crashed? |
And did all lua-nginx-module tests passed on your build? |
I tried it with 0.8.49 and 0.7.67. The test code is from you example:
The stack trace only had two entries: #0 0x0000000100085e48 in lj_alloc_malloc () I have done some fairly complicated with the non-jit version, including loading modules, so the non-jit version seems fine. When I start nginx, it fails immediately with the segment fault, before I issue any requests are handled. |
I've tried nginx 0.7.67 and 0.8.49 with luajit 2.0 git HEAD (almost identical to the beta5 release kicked out these days), the whole tests suite is passing successfully on my x86_64 slackware linux :) It may be an issue for LuaJIT 2.0's immature x86_64 support on Mac? I'll try to find a macbook to verify this. Thanks for reporting. Cheers, |
Could be a mac thing. My production machines are linux, but I do all my development on the mac. If you need more information, let me know. The module looks fantastic so far. |
I have someone tried nginx-0.8.49 and LuaJIT-2.0beta5 on a 32-bit Mac, and all tests passed. Could you run some Lua codes with your LuaJIT build? Does it work well when running as a standalone interpreter? |
Yes, I can run code on the standalone interpreter: /usr/local/bin/luajit-2.0.0-beta5 |
Could you download this example code and try it out on your Mac? Download both con_run1.cc and con_run1.lua. Compile and link con_run1.cc by: And run: Could you tell me if it crashes? |
Yes, it crashes. Stack trace is: #0 0x000000010002b048 in lj_alloc_malloc () |
OK...I've created a more minified example. Would you please try this again? Compile and link min_con.c by: Run: If this also crashes, I will submit a bug to LuaJIT. |
Yes, it crashed also. Backtrace is: #0 0x000000010002b048 in lj_alloc_malloc () |
It seems that LuaJIT requires some special linking options on 64-bit MacOS...Please try this cmdline and try min_con.c example again: $ gcc -I/usr/local/include/luajit-2.0 min_con.c -o min_con -Wl,-pagezero_size,10000 -Wl,-image_base,100000000 -L/usr/local/lib -lluajit-5.1 |
This works. I'll add these link options and try on nginx. |
Is lua-nginx-module working on your server after added these linking options? |
OK, consider it solved. |
I got same problem on my SnowLeopard. I think you should add this note to README. Thanks. |
Thanks for reminding me. The notes have been added to README, and I'll see if it can be auto done in config file later. |
This issue appeared in our EC2 test cluster, which compiles Nginx with `-DNGX_LUA_USE_ASSERT` and `-DNGX_LUA_ABORT_AT_PANIC`. The lua-resty-redis test: === TEST 1: github issue openresty#108: ngx.locaiton.capture + redis.set_keepalive in t/bugs.t [1] would produce core dumps in the check leak testing mode. The backtrace for these core dumps was: #0 0x00007fd417bee277 in raise () from /lib64/libc.so.6 openresty#1 0x00007fd417bef968 in abort () from /lib64/libc.so.6 openresty#2 0x00007fd417be7096 in __assert_fail_base () from /lib64/libc.so.6 openresty#3 0x00007fd417be7142 in __assert_fail () from /lib64/libc.so.6 openresty#4 0x000000000050d227 in ngx_http_lua_socket_tcp_resume_conn_op (spool=c/ngx_http_lua_socket_tcp.c:3963 openresty#5 0x000000000050e51a in ngx_http_lua_socket_tcp_finalize (r=r@entry=0x5628) at ../../src/ngx_http_lua_socket_tcp.c:4195 openresty#6 0x000000000050e570 in ngx_http_lua_socket_tcp_cleanup (data=0x7fd419p_lua_socket_tcp.c:3755 openresty#7 0x0000000000463aa5 in ngx_http_free_request (r=r@entry=0xbfaec0, rc=http_request.c:3508 ... Which was caused by the following assertion in ngx_http_lua_socket_tcp.c with `NGX_DEBUG`: #if (NGX_DEBUG) ngx_http_lua_assert(spool->connections >= 0); #else Thanks to Mozilla's rr, a recorded session showed that `spool->connections` was `-1`. Unfortunately, reproducing this case does not seem possible, since the failure is due to the request cleanup (`ngx_http_free_request`). Here is an explanation: -- thread 1 local sock = ngx.socket.tcp() sock:connect() sock:setkeepalive() -- pool created, connections: 1 -- thread 2 local sock = ngx.socket.tcp() sock:connect() -- from pool, connections: 1 -- thread 1 -- sock from thread 1 idle timeout, closes, and calls -- ngx_http_lua_socket_tcp_finalize, connections: 0 -- thread 2 sock:setkeepalive() -- connections: -1 -- ngx_http_lua_socket_tcp_resume_conn_op gets called, assertion fails In order to avoid this race condition, we must determine whether the socket pool exists or not, not from the `ngx_http_lua_socket_tcp_upstream` struct, but from the Lua Registry. This way, when thread 2's socket enters the keepalive state, it will respect the previous call to `ngx_http_lua_socket_free_pool` (which unset the pool from the registry). [1]: https://github.com/openresty/lua-resty-redis/blob/master/t/bugs.t
This issue appeared in our EC2 test cluster, which compiles Nginx with `-DNGX_LUA_USE_ASSERT` and `-DNGX_LUA_ABORT_AT_PANIC`. The lua-resty-redis test: === TEST 1: github issue openresty#108: ngx.locaiton.capture + redis.set_keepalive in t/bugs.t [1] would produce core dumps in the check leak testing mode. The backtrace for these core dumps was: #0 0x00007fd417bee277 in raise () from /lib64/libc.so.6 openresty#1 0x00007fd417bef968 in abort () from /lib64/libc.so.6 openresty#2 0x00007fd417be7096 in __assert_fail_base () from /lib64/libc.so.6 openresty#3 0x00007fd417be7142 in __assert_fail () from /lib64/libc.so.6 openresty#4 0x000000000050d227 in ngx_http_lua_socket_tcp_resume_conn_op (spool=c/ngx_http_lua_socket_tcp.c:3963 openresty#5 0x000000000050e51a in ngx_http_lua_socket_tcp_finalize (r=r@entry=0x5628) at ../../src/ngx_http_lua_socket_tcp.c:4195 openresty#6 0x000000000050e570 in ngx_http_lua_socket_tcp_cleanup (data=0x7fd419p_lua_socket_tcp.c:3755 openresty#7 0x0000000000463aa5 in ngx_http_free_request (r=r@entry=0xbfaec0, rc=http_request.c:3508 ... Which was caused by the following assertion in ngx_http_lua_socket_tcp.c with `NGX_DEBUG`: #if (NGX_DEBUG) ngx_http_lua_assert(spool->connections >= 0); #else Thanks to Mozilla's rr, a recorded session showed that `spool->connections` was `-1`. Here is a reproducible test case: local sock1 = ngx.socket.tcp() local sock2 = ngx.socket.tcp() sock1:connect() sock2:connect() sock1:setkeepalive() -- pool created, connections: 1 sock2:setkeepalive() -- connections: 1 sock1:connect() -- connections: 1 sock2:connect() -- connections: 1 sock1:close() -- connections: 0 sock2:close() -- connections: -1 -- ngx_http_lua_socket_tcp_resume_conn_op gets called, assertion fails In order to avoid this race condition, we must determine whether the socket pool exists or not, not from the `ngx_http_lua_socket_tcp_upstream` struct, but from the Lua Registry. This way, when thread 2's socket enters the keepalive state, it will respect the previous call to `ngx_http_lua_socket_free_pool` (which unset the pool from the registry). [1]: https://github.com/openresty/lua-resty-redis/blob/master/t/bugs.t
==openresty==70603==ERROR: AddressSanitizer: memcpy-param-overlap: memory ranges [0x621000001500,0x621000002181) and [0x62100000187f, 0x621000002500) overlap #0 0x7f3db1899ffe (/lib64/libasan.so.5+0x99ffe) #1 0x9da926 (/usr/local/openresty-debug/nginx/sbin/nginx+0x9da926) #2 0x9dd1a1 (/usr/local/openresty-debug/nginx/sbin/nginx+0x9dd1a1) #3 0x4c89c6 (/usr/local/openresty-debug/nginx/sbin/nginx+0x4c89c6) #4 0x5d1e4e (/usr/local/openresty-debug/nginx/sbin/nginx+0x5d1e4e) #5 0x4c89c6 (/usr/local/openresty-debug/nginx/sbin/nginx+0x4c89c6) #6 0x5b8583 (/usr/local/openresty-debug/nginx/sbin/nginx+0x5b8583) #7 0x4c89c6 (/usr/local/openresty-debug/nginx/sbin/nginx+0x4c89c6) #8 0x4b4419 (/usr/local/openresty-debug/nginx/sbin/nginx+0x4b4419) #9 0x427f16 (/usr/local/openresty-debug/nginx/sbin/nginx+0x427f16) #10 0x7f3daff27554 (/lib64/libc.so.6+0x22554) #11 0x42d537 (/usr/local/openresty-debug/nginx/sbin/nginx+0x42d537)
…t_peer at function end. Avoid inserting new parameters in the middle of the function to prevent core dumps when using old lua-resty-core with new lua-nginx-module. Example stack trace: ``` Message: Process 1414245 (nginx) of user 1000 dumped core. Stack trace of thread 1414245: #0 0x00007ff596938285 __strlen_avx2 (libc.so.6 + 0x162285) openresty#1 0x00007ff596f623d2 lj_cf_ffi_string (libluajit-5.1.so.2 + 0x523d2) openresty#2 0x00007ff596f1bb4b lj_BC_FUNCC (libluajit-5.1.so.2 + 0xbb4b) openresty#3 0x00007ff596f74223 lua_pcall (libluajit-5.1.so.2 + 0x64223) openresty#4 0x00000000005044b7 n/a (/home/jiahao/work/org/lua-resty-core/work/nginx/sbin/nginx + 0x1044b7) ```
…t_peer at function end. Avoid inserting new parameters in the middle of the function to prevent core dumps when using old lua-resty-core with new lua-nginx-module. Example stack trace: ``` Message: Process 1414245 (nginx) of user 1000 dumped core. Stack trace of thread 1414245: #0 0x00007ff596938285 __strlen_avx2 (libc.so.6 + 0x162285) openresty#1 0x00007ff596f623d2 lj_cf_ffi_string (libluajit-5.1.so.2 + 0x523d2) openresty#2 0x00007ff596f1bb4b lj_BC_FUNCC (libluajit-5.1.so.2 + 0xbb4b) openresty#3 0x00007ff596f74223 lua_pcall (libluajit-5.1.so.2 + 0x64223) openresty#4 0x00000000005044b7 n/a (/home/jiahao/work/org/lua-resty-core/work/nginx/sbin/nginx + 0x1044b7) ```
…t function end. Avoid inserting new parameters in the middle of the function to prevent core dumps when using old lua-resty-core with new lua-nginx-module. Example stack trace: ``` Message: Process 2199905 (nginx) of user 1000 dumped core. Stack trace of thread 2199905: #0 0x00007ffaf1e4b385 in OPENSSL_sk_num (st=st@entry=0xffffffff) at crypto/stack/stack.c:382 openresty#1 0x0000000000510aba in sk_X509_num (sk=0xffffffff) at /opt/ssl/include/openssl/x509.h:99 openresty#2 ngx_http_lua_ffi_ssl_verify_client (r=<optimized out>, client_certs=<optimized out>, trusted_certs=0xffffffff, depth=<optimized out>, err=0x0) at /home/jiahao/work/org/lua-nginx-module/src/ngx_http_lua_ssl_certby.c:1588 ```
…t function end. Avoid inserting new parameters in the middle of the function to prevent core dumps when using old lua-resty-core with new lua-nginx-module. Example stack trace: ``` Message: Process 2199905 (nginx) of user 1000 dumped core. Stack trace of thread 2199905: #0 0x00007ffaf1e4b385 in OPENSSL_sk_num (st=st@entry=0xffffffff) at crypto/stack/stack.c:382 openresty#1 0x0000000000510aba in sk_X509_num (sk=0xffffffff) at /opt/ssl/include/openssl/x509.h:99 openresty#2 ngx_http_lua_ffi_ssl_verify_client (r=<optimized out>, client_certs=<optimized out>, trusted_certs=0xffffffff, depth=<optimized out>, err=0x0) at /home/jiahao/work/org/lua-nginx-module/src/ngx_http_lua_ssl_certby.c:1588 ```
…t_peer at function end. Avoid inserting new parameters in the middle of the function to prevent core dumps when using old lua-resty-core with new lua-nginx-module. Example stack trace: ``` Message: Process 1414245 (nginx) of user 1000 dumped core. Stack trace of thread 1414245: #0 0x00007ff596938285 __strlen_avx2 (libc.so.6 + 0x162285) openresty#1 0x00007ff596f623d2 lj_cf_ffi_string (libluajit-5.1.so.2 + 0x523d2) openresty#2 0x00007ff596f1bb4b lj_BC_FUNCC (libluajit-5.1.so.2 + 0xbb4b) openresty#3 0x00007ff596f74223 lua_pcall (libluajit-5.1.so.2 + 0x64223) openresty#4 0x00000000005044b7 n/a (/home/jiahao/work/org/lua-resty-core/work/nginx/sbin/nginx + 0x1044b7) ```
…t function end. Avoid inserting new parameters in the middle of the function to prevent core dumps when using old lua-resty-core with new lua-nginx-module. Example stack trace: ``` Message: Process 2199905 (nginx) of user 1000 dumped core. Stack trace of thread 2199905: #0 0x00007ffaf1e4b385 in OPENSSL_sk_num (st=st@entry=0xffffffff) at crypto/stack/stack.c:382 openresty#1 0x0000000000510aba in sk_X509_num (sk=0xffffffff) at /opt/ssl/include/openssl/x509.h:99 openresty#2 ngx_http_lua_ffi_ssl_verify_client (r=<optimized out>, client_certs=<optimized out>, trusted_certs=0xffffffff, depth=<optimized out>, err=0x0) at /home/jiahao/work/org/lua-nginx-module/src/ngx_http_lua_ssl_certby.c:1588 ```
…t_peer at function end. Avoid inserting new parameters in the middle of the function to prevent core dumps when using old lua-resty-core with new lua-nginx-module. Example stack trace: ``` Message: Process 1414245 (nginx) of user 1000 dumped core. Stack trace of thread 1414245: #0 0x00007ff596938285 __strlen_avx2 (libc.so.6 + 0x162285) openresty#1 0x00007ff596f623d2 lj_cf_ffi_string (libluajit-5.1.so.2 + 0x523d2) openresty#2 0x00007ff596f1bb4b lj_BC_FUNCC (libluajit-5.1.so.2 + 0xbb4b) openresty#3 0x00007ff596f74223 lua_pcall (libluajit-5.1.so.2 + 0x64223) openresty#4 0x00000000005044b7 n/a (/home/jiahao/work/org/lua-resty-core/work/nginx/sbin/nginx + 0x1044b7) ```
…t function end. Avoid inserting new parameters in the middle of the function to prevent core dumps when using old lua-resty-core with new lua-nginx-module. Example stack trace: ``` Message: Process 2199905 (nginx) of user 1000 dumped core. Stack trace of thread 2199905: #0 0x00007ffaf1e4b385 in OPENSSL_sk_num (st=st@entry=0xffffffff) at crypto/stack/stack.c:382 openresty#1 0x0000000000510aba in sk_X509_num (sk=0xffffffff) at /opt/ssl/include/openssl/x509.h:99 openresty#2 ngx_http_lua_ffi_ssl_verify_client (r=<optimized out>, client_certs=<optimized out>, trusted_certs=0xffffffff, depth=<optimized out>, err=0x0) at /home/jiahao/work/org/lua-nginx-module/src/ngx_http_lua_ssl_certby.c:1588 ```
…t_peer at function end. Avoid inserting new parameters in the middle of the function to prevent core dumps when using old lua-resty-core with new lua-nginx-module. Example stack trace: ``` Message: Process 1414245 (nginx) of user 1000 dumped core. Stack trace of thread 1414245: #0 0x00007ff596938285 __strlen_avx2 (libc.so.6 + 0x162285) openresty#1 0x00007ff596f623d2 lj_cf_ffi_string (libluajit-5.1.so.2 + 0x523d2) openresty#2 0x00007ff596f1bb4b lj_BC_FUNCC (libluajit-5.1.so.2 + 0xbb4b) openresty#3 0x00007ff596f74223 lua_pcall (libluajit-5.1.so.2 + 0x64223) openresty#4 0x00000000005044b7 n/a (/home/jiahao/work/org/lua-resty-core/work/nginx/sbin/nginx + 0x1044b7) ```
…t function end. Avoid inserting new parameters in the middle of the function to prevent core dumps when using old lua-resty-core with new lua-nginx-module. Example stack trace: ``` Message: Process 2199905 (nginx) of user 1000 dumped core. Stack trace of thread 2199905: #0 0x00007ffaf1e4b385 in OPENSSL_sk_num (st=st@entry=0xffffffff) at crypto/stack/stack.c:382 openresty#1 0x0000000000510aba in sk_X509_num (sk=0xffffffff) at /opt/ssl/include/openssl/x509.h:99 openresty#2 ngx_http_lua_ffi_ssl_verify_client (r=<optimized out>, client_certs=<optimized out>, trusted_certs=0xffffffff, depth=<optimized out>, err=0x0) at /home/jiahao/work/org/lua-nginx-module/src/ngx_http_lua_ssl_certby.c:1588 ```
I am using the lua nginx module successfully with the standard lua interpreter. I tried it with luajit version 2.0 beta 5, but nginx crashes with a segmentation fault. What information can I provide to help resolve this issue? I'm running it on Snow Leopard Server. I'll try it with other versions of luajit to see if I can narrow down the problem. What versions of luajit have you tested with? Which platforms?
The text was updated successfully, but these errors were encountered: