php.net |  support |  documentation |  report a bug |  advanced search |  search howto |  statistics |  random bug |  login
Bug #73619 Crash or zend_mm_heap_corrupted error in opcache with OwnCloud 7 update
Submitted: 2016-11-28 21:14 UTC Modified: 2016-11-28 21:21 UTC
From: adamw at happyassassin dot net Assigned:
Status: Closed Package: opcache
PHP Version: 7.0.13 OS: Fedora 20
Private report: No CVE-ID: None
View Developer Edit
Welcome! If you don't have a Git account, you can't do anything here.
If you reported this bug, you can edit this bug over here.
(description)
Block user comment
Status: Assign to:
Package:
Bug Type:
Summary:
From: adamw at happyassassin dot net
New email:
PHP Version: OS:

 

 [2016-11-28 21:14 UTC] adamw at happyassassin dot net
Description:
------------
See downstream bug report:

https://bugzilla.redhat.com/show_bug.cgi?id=1398474

SELinux blocks web servers from what it calls 'execmem' actions by default; as I understand it, this means execution of writeable memory, which is considered a potential avenue for exploitation and not to be necessary for well-written code.

With PHP 7.0.13's default configuration, where just-in-time compilation of PCREs is enabled, Fedora 25 systems see a flood of these 'execmem' denials. Disabling the PCRE JIT feature with `pcre.jit=0` makes them go away.

There is some more discussion of this at https://unix.stackexchange.com/questions/287831/how-to-work-out-why-apache-is-attempting-execmem , where several people traced this problem to the PCRE JIT code using gdb and backtraces are available. These seem to run through _pcre_jit_compile / pcre_study / pcre_get_compiled_regex_cache / php_do_pcre_match .


Patches

Pull Requests

History

AllCommentsChangesGit/SVN commitsRelated reports
 [2016-11-28 21:21 UTC] adamw at happyassassin dot net
-Status: Open +Status: Closed
 [2016-11-28 21:21 UTC] adamw at happyassassin dot net
I have absolutely no idea how this wound up as a dupe of https://bugs.php.net/bug.php?id=67687 , as I filed it with completely different content. Can only imagine it's some kind of bug to do with using the same password for both bugs.
 
PHP Copyright © 2001-2024 The PHP Group
All rights reserved.
Last updated: Sun Dec 22 05:01:30 2024 UTC