ID: 40098 Updated by: [EMAIL PROTECTED] Reported By: wharmby at uk dot ibm dot com Status: Closed Bug Type: Streams related Operating System: Linux RHEL4 PHP Version: 5CVS-2007-01-11 (snap) Assigned To: iliaa New Comment:
Fixed in CVS, thanks. Previous Comments: ------------------------------------------------------------------------ [2007-01-12 08:57:36] wharmby at uk dot ibm dot com Hi Ilia Thanks for applying the fix for this defect. However, in adjusting the fix to deal with the SYSCONF errors I believe you have introduced a storage leak. The code as it stands does not free "pwbuf" if the call to getpwnam_r is successful after it has finished extracting the data from it. Here is a patch for the missing code built against latest snapshot for 5.2: http://www.pastebin.ca/313940 Regards Andy ------------------------------------------------------------------------ [2007-01-12 01:46:21] [EMAIL PROTECTED] This bug has been fixed in CVS. Snapshots of the sources are packaged every three hours; this change will be in the next snapshot. You can grab the snapshot at http://snaps.php.net/. Thank you for the report, and for helping us make PHP better. ------------------------------------------------------------------------ [2007-01-11 11:59:44] wharmby at uk dot ibm dot com Description: ------------ The current implementation of php_fopen_primary_script() uses the non-reentrant getpwnam() rather than the re-entrant getpwnam_r(). Therefore calling it on a Linux ZTS enabled build could lead to unpredictable/undesirable results. The code should use the re-entrant version if it is available. The following patch which were built against the latest snapshot (Jan 11 2007, 0730 GMT) modifies the code in main/fopen_wrappers.c to use the re-entrant getpwnam_r if its available: http://pastebin.ca/312969 Fix tested on Linux RHEL4 with Apache 2.0.52 and user_dir set in php.ini The modified code was then driven by specifying the ~user_id in the URL e.g http://localhost/~andy/phpinfo.php. Reproduce code: --------------- Problem found by code inspection. As with most thread safety issues difficult to produce a simple testcase which will show a reproducible crash but current Linux executable is clearly not reentrant and therefore not thread safe. Expected result: ---------------- N/A Actual result: -------------- N/A ------------------------------------------------------------------------ -- Edit this bug report at http://bugs.php.net/?id=40098&edit=1