You are viewing a plain text version of this content. The canonical link for it is here.
Posted to bugs@httpd.apache.org by bu...@apache.org on 2007/05/02 19:57:40 UTC
DO NOT REPLY [Bug 42321] New: - LFS support on 32bit OSes is broken
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=42321>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.
http://issues.apache.org/bugzilla/show_bug.cgi?id=42321
Summary: LFS support on 32bit OSes is broken
Product: Apache httpd-2
Version: 2.0.59
Platform: Sun
OS/Version: All
Status: NEW
Severity: major
Priority: P2
Component: Build
AssignedTo: bugs@httpd.apache.org
ReportedBy: pirzyk+apache@uiuc.edu
Installed the version of Apache that comes with Solaris 10. It is based on
httpd-2.0.52,55, and 58, depending on the patch level. They added LFS support
with the patch (120543-09), which also brought apache up to version 2.0.58. We
have an in house developed auth module (mod_bluestem,
http://www-dev.cites.uiuc.edu/mod_bluestem/) that break when the patch is
installed. The problem is the request_rec structure is corrupted when calling
any of the bluestem hooks. Specifically, r->per_dir_config is set to NULL. In
my debugging with the Sun Apache source code, the vanilla httpd-2.0.58 source
code, I found that the configure script for srclib/apr has a bug in it.
supirzyk@nani:/scratch/httpd-2.0.58/srclib/apr
9>diff -Nru configure.in.orig configure.in
--- configure.in.orig 2006-03-16 04:04:04.000000000 -0600
+++ configure.in 2007-05-02 10:50:46.687788000 -0500
@@ -1199,7 +1199,7 @@
if test "$ac_cv_type_off_t" = "yes"; then
APR_CHECK_SIZEOF_EXTENDED([#include <sys/types.h>], off_t, 8)
AC_MSG_CHECKING([which type to use for apr_off_t])
- if test "${ac_cv_sizeof_off_t}${ac_cv_sizeof_long}" = "44"; then
+ if test "${ac_cv_sizeof_off_t}${ac_cv_sizeof_long}" != "44"; then
# Special case: off_t may change size with _FILE_OFFSET_BITS
# on 32-bit systems with LFS support. To avoid compatibility
# with other software which may export _FILE_OFFSET_BITS,
I have verified the problem exists in 2.0.59 as well.
--
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
---------------------------------------------------------------------
To unsubscribe, e-mail: bugs-unsubscribe@httpd.apache.org
For additional commands, e-mail: bugs-help@httpd.apache.org
DO NOT REPLY [Bug 42321] - LFS support on 32bit OSes is broken
Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=42321>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.
http://issues.apache.org/bugzilla/show_bug.cgi?id=42321
------- Additional Comments From jorton@redhat.com 2007-05-03 02:38 -------
Ah, yes, the comment is wrong; APR 0.9.x sets apr_off_t to "long" for the cases
where other packages may redefine _FILE_OFFSET_BITS and change the definition of
off_t. I've updated it: http://svn.apache.org/viewvc?view=rev&rev=534764
--
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
---------------------------------------------------------------------
To unsubscribe, e-mail: bugs-unsubscribe@httpd.apache.org
For additional commands, e-mail: bugs-help@httpd.apache.org
DO NOT REPLY [Bug 42321] - LFS support on 32bit OSes is broken
Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=42321>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.
http://issues.apache.org/bugzilla/show_bug.cgi?id=42321
pirzyk+apache@uiuc.edu changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Resolution|INVALID |
------- Additional Comments From pirzyk+apache@uiuc.edu 2007-05-02 12:25 -------
After recomping mod_bluestem, per note 2, the problem still exists. Why do you set off_t to long when
off_t and long are the same. It is a no-op and a waste of code for configure. If it is supposed to be
that way, then there still is a MAJOR bug in apache. r->per_dir_config should NEVER be NULL when
calling modules. The macro, ap_get_module_config never test to see of r->per_dir_config is NULL.
The second point, per the comments in configure.in
# Special case: off_t may change size with _FILE_OFFSET_BITS
# on 32-bit systems with LFS support. To avoid compatibility
# with other software which may export _FILE_OFFSET_BITS,
# hard-code apr_off_t to long.
This means that _FILE_OFFSET_BITS is set to something other than a long. On a 32 bit OS, that is 4
bytes. The test case will be true if off_t == long. The comments do not say this, the comments imply
that off_t != long. That makes the test wrong.
--
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
---------------------------------------------------------------------
To unsubscribe, e-mail: bugs-unsubscribe@httpd.apache.org
For additional commands, e-mail: bugs-help@httpd.apache.org
DO NOT REPLY [Bug 42321] - LFS support on 32bit OSes is broken
Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=42321>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.
http://issues.apache.org/bugzilla/show_bug.cgi?id=42321
jorton@redhat.com changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |INVALID
------- Additional Comments From jorton@redhat.com 2007-05-02 11:50 -------
That bit of configure code works as intended. APR 0.9.x does not (and cannot)
support LFS - adding such support will change the ABI. It looks like Sun knew
this; see "note 2":
http://sunsolve.sun.com/search/document.do?assetkey=1-21-120543-09-1
--
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
---------------------------------------------------------------------
To unsubscribe, e-mail: bugs-unsubscribe@httpd.apache.org
For additional commands, e-mail: bugs-help@httpd.apache.org
DO NOT REPLY [Bug 42321] - LFS support on 32bit OSes is broken
Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=42321>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.
http://issues.apache.org/bugzilla/show_bug.cgi?id=42321
pirzyk+apache@uiuc.edu changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|REOPENED |RESOLVED
Resolution| |INVALID
------- Additional Comments From pirzyk+apache@uiuc.edu 2007-05-02 14:47 -------
Found that mod_bluestem is not honoring CPPFLAGS as defined by apxs.
--
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
---------------------------------------------------------------------
To unsubscribe, e-mail: bugs-unsubscribe@httpd.apache.org
For additional commands, e-mail: bugs-help@httpd.apache.org