| Subject: | Segfault due to logic typo in _get_fd_entry |
Hi there,
Ran into a segfault on startup with metabase-relayd when using
POE::XS::Loop::EPoll with https URLs, after some tracing it appears to
be this piece of code:
$ gdb `which perl`
GNU gdb (Ubuntu/Linaro 7.4-2012.04-0ubuntu2) 7.4-2012.04
Copyright (C) 2012 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later
<http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show
copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
For bug reporting instructions, please see:
<http://bugs.launchpad.net/gdb-linaro/>...
Reading symbols from /home/tom/perl5/perlbrew/perls/perl-
5.16.1/bin/perl...(no debugging symbols found)...done.
(gdb) set args /home/tom/perl5/perlbrew/perls/perl-5.16.1/bin/metabase-
relayd --idfile ~/.cpantesters/metabase_id.json --url
https://metabase.cpantesters.org/api/v1/ --port 9323 --address 0.0.0.0
(gdb) run
Starting program: /home/tom/perl5/perlbrew/perls/perl-5.16.1/bin/perl
/home/tom/perl5/perlbrew/perls/perl-5.16.1/bin/metabase-relayd --idfile
~/.cpantesters/metabase_id.json --url
https://metabase.cpantesters.org/api/v1/ --port 9323 --address 0.0.0.0
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-
gnu/libthread_db.so.1".
Running metabase-relayd with options:
url https://metabase.cpantesters.org/api/v1/
dbfile /home/tom/.metabase/relay.db
idfile /home/tom/.cpantesters/metabase_id.json
address 0.0.0.0
port 9323
Program received signal SIGSEGV, Segmentation fault.
0x00007ffff518e909 in _get_fd_entry (fd=18797888) at EPoll.xs:194
194 return fd_lookup[fd];
(gdb) print fd
$1 = 18797888
(gdb) print fd_lookup
$2 = (int *) 0x1234ee0
(gdb) list
189 static int
190 _get_fd_entry(int fd) {
191 if (fd < 0 && fd >= fd_lookup_count)
192 return -1;
193
194 return fd_lookup[fd];
195 }
196
197 static int
198 _make_fd_entry(int fd) {
which the attached patch should resolve. Haven't had a chance to put
together a proper test case for it, unfortunately, but with this applied
so far I've had no recurrence of the original issue.
best regards,
Tom
| Subject: | poe-xs-loop-epoll_get_fd_entry_fix.patch |
diff -uNPr POE-XS-Loop-EPoll-1.001-dE3Ae7/EPoll.xs POE-XS-Loop-EPoll-1.001-get_fd_entry_fix/EPoll.xs
--- POE-XS-Loop-EPoll-1.001-dE3Ae7/EPoll.xs 2010-03-07 02:11:43.000000000 +0000
+++ POE-XS-Loop-EPoll-1.001-get_fd_entry_fix/EPoll.xs 2012-08-15 15:16:40.572777652 +0100
@@ -188,7 +188,7 @@
static int
_get_fd_entry(int fd) {
- if (fd < 0 && fd >= fd_lookup_count)
+ if (fd < 0 || fd >= fd_lookup_count)
return -1;
return fd_lookup[fd];