Re: FreeRADIUS 2 not listening on right port

2008-06-18 Thread doc74

Just installed 2.0.4 with MySQL Support from Ubuntu Repositories and got the
same problem:
(http://packages.ubuntu.com/de/intrepid/i386/freeradius/download and
http://packages.ubuntu.com/de/intrepid/i386/freeradius-mysql/download)

Listening on authentication address * port 1087
Listening on accounting address * port 1088
Listening on proxy address * port 1089
Ready to process requests.

next start:

Listening on authentication address * port 1090
Listening on accounting address * port 1091
Listening on proxy address * port 1093
Ready to process requests.

next start:

Listening on authentication address * port 1093
Listening on accounting address * port 1094
Listening on proxy address * port 1095
Ready to process requests.

..

Seems that the Port are increasing.

Greets

Chris
-- 
View this message in context: 
http://www.nabble.com/FreeRADIUS-2-not-listening-on-right-port-tp17258249p17982548.html
Sent from the FreeRadius - User mailing list archive at Nabble.com.

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-06-18 Thread Alan DeKok
doc74 wrote:
 Just installed 2.0.4 with MySQL Support from Ubuntu Repositories and got the
 same problem:

 Seems that the Port are increasing.

  Wow.  There's nothing in the server that remembers ports from one
execution to the next.

  I'm running FreeRADIUS on ubuntu 7.10, and I don't see this.  I've
also run it (again from source) on 8.04, and I don't see this.

  Try compiling it from source...

  Alan DeKok.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-06-18 Thread John Dennis

Alan DeKok wrote:

doc74 wrote:
  

Just installed 2.0.4 with MySQL Support from Ubuntu Repositories and got the
same problem:



  

Seems that the Port are increasing.



  Wow.  There's nothing in the server that remembers ports from one
execution to the next.

  I'm running FreeRADIUS on ubuntu 7.10, and I don't see this.  I've
also run it (again from source) on 8.04, and I don't see this.

  Try compiling it from source...
  

Recall this problem is the consequence of two independent factors:

* the C source code use of pointer aliasing
* the compiler used and optimization level

The 2.0.5 version has Alan's fixes for pointer aliasing, if you build 
from source using a version less than 2.0.5 then your results will be 
dependent on the compiler. Rather than relying on your local compiler 
being more forgiving it might be wise to use 2.0.5 which contain the 
fixes which provoked the problem in the first place.


--
John Dennis [EMAIL PROTECTED]

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Re: FreeRADIUS 2 not listening on right port

2008-05-29 Thread Alan DeKok
Phil Mayers wrote:
 For those not following the Fedora bug, it (or rather, it's dependency)
 has been closed by Ulrich Drepper. He seems to be saying that the
 FreeRadius code is incorrect and specifically that an invalid typecast
 is triggering the compiler to generate bad code:

  Interesting.  So GCC doesn't complain, and it generates bad assembly
for C code that it thinks is perfectly valid.

 https://bugzilla.redhat.com/show_bug.cgi?id=448743#c6
 
 Summary (as far as I can make out): Because the code looks like this:
...
 ...the compiler can't detect that the i value is used, and optimises
 the code touching it away.

  i.e.  GCC doesn't properly analyze the code that it optimizes, so it
generates the wrong optimizations.  GCC has a history of doing this...

 I'm not expert enough in the C99 standard to judge whether the comments
 at the above URL are correct or not; I suspect it's a matter of some
 controversy, and will back slowly away now... ;o)

  I think your comments about the underlying cause are correct.
Ulrich's comments about Posix are interesting:

 The invalid casts are forbidden by ISO C.  POSIX
 does not and cannot guarantee anything about this type of use of
 sockaddr_storage.

  To quote the Opengroup page:

http://www.opengroup.org/onlinepubs/009695399/basedefs/sys/socket.h.html

...
The sys/socket.h header shall define the sockaddr_storage structure.
This structure shall be:

* Large enough to accommodate all supported protocol-specific
address structures

* Aligned at an appropriate boundary so that pointers to it can be
cast as pointers to protocol-specific address structures and used to
access the fields of those structures without alignment problems
...

  i.e. the code in FreeRADIUS is correct.  It uses sockaddr_stoage as a
generic container for protocol-specific address structures.  It casts a
pointer to sockaddr_storage to a pointer to protocol-specific address
structures.  The recommendation to use a union to hold both
sockaddr_storage  sockaddr_in is... interesting.  If you have to do
that, WTF is the use of sockaddr_storage?

  Looking on the net, I also found a fair amount of code using
sockaddr_storage in the recommended Posix way.  It appears that
FreeRADIUS is OK, and just got hit by a GCC bug.

  i.e. for ANY code like this:

foo = (type_t *) bar;
switch (x) {
case 1:
foo-a = 1;
break;
...
}
function(bar);

  GCC will optimize away the line doing foo-a = 1;.  Nice.

  Alan DeKok.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-29 Thread John Dennis

Alan DeKok wrote:

Phil Mayers wrote:
  

For those not following the Fedora bug, it (or rather, it's dependency)
has been closed by Ulrich Drepper. He seems to be saying that the
FreeRadius code is incorrect and specifically that an invalid typecast
is triggering the compiler to generate bad code:



  
I wanted to understand the issues surrounding strict aliasing better. I 
found the following article to be well written, quite readable, and 
informative:


Understanding Strict Aliasing 
http://www.cellperformance.com/mike_acton/2006/06/understanding_strict_aliasing.html

http://www.cellperformance.com/mike_acton/2006/06/understanding_strict_aliasing.html

If one is to take away any simple concept and/or rule from the article 
it would have to be these 3 statements:


One pointer is said to /alias/ another pointer when both refer to the 
same location or object. 
In C99, it is /illegal/ to create an alias of a different type than the 
original.
Strict aliasing means that two objects of different types cannot refer 
to the same location in memory.


However, note that this is exactly what fr_socket is doing, salocal is 
a pointer to sockaddr_storage, sa is in one scope is a pointer to 
sockaddr_in and in another scope sa is a pointer to sockaddr_in6. Each 
of these 3 pointers point to exact same memory location, according to 
the above definitions this is a classic case of illegal pointer aliasing.


if (ipaddr-af == AF_INET) {
struct sockaddr_in *sa;

sa = (struct sockaddr_in *) salocal;
sa-sin_port = htons((uint16_t) port);
} else if (ipaddr-af == AF_INET6) {
struct sockaddr_in6 *sa;

sa = (struct sockaddr_in6 *) salocal;
sa-sin6_port = htons((uint16_t) port);
}



  Interesting.  So GCC doesn't complain, and it generates bad assembly
for C code that it thinks is perfectly valid.

  
Agreed, GCC should have flagged this. However, for what it's worth the 
above article states the compiler may not always be able to recognize 
various illegal constructs and programmers should not depend it. If and 
when GCC should emit warnings is a subject for debate, but I think we 
can all agree it would have been nicer if GCC had emitted a warning. 
It's a shame it didn't, but I don't think it changes the overarching 
issue. Also, in expediency I did not try building with the various GCC 
warning flags to see if I could get GCC to emit a warning for this case, 
perhaps GCC at some level of verbosity would have complained.

https://bugzilla.redhat.com/show_bug.cgi?id=448743#c6

Summary (as far as I can make out): Because the code looks like this:


...
  

...the compiler can't detect that the i value is used, and optimises
the code touching it away.



  
Yup, this is weird, but as I understand it illegal constructs produce 
undefined results, perhaps the fact GCC stored some LHS values but not 
others is just an example of undefined.

  i.e.  GCC doesn't properly analyze the code that it optimizes, so it
generates the wrong optimizations.  GCC has a history of doing this...

  

I'm not expert enough in the C99 standard to judge whether the comments
at the above URL are correct or not; I suspect it's a matter of some
controversy, and will back slowly away now... ;o)



  I think your comments about the underlying cause are correct.
Ulrich's comments about Posix are interesting:

  

The invalid casts are forbidden by ISO C.  POSIX
does not and cannot guarantee anything about this type of use of
sockaddr_storage.



  To quote the Opengroup page:

http://www.opengroup.org/onlinepubs/009695399/basedefs/sys/socket.h.html

...
The sys/socket.h header shall define the sockaddr_storage structure.
This structure shall be:

* Large enough to accommodate all supported protocol-specific
address structures

* Aligned at an appropriate boundary so that pointers to it can be
cast as pointers to protocol-specific address structures and used to
access the fields of those structures without alignment problems
...

  i.e. the code in FreeRADIUS is correct.
For what it's worth my reading of the above statements is different than 
yours. I read the above as sockaddr_storage provides a type whose size 
is sufficient to hold all sockaddr types while respecting the alignment 
requirement in any given sockaddr type. The discussion of pointer 
casting fails to make explicit such casting must still conform to the 
rules of C99. I do not read the second paragraph as meaning pointer 
aliasing shall be supported, something outside that scope of that document.


Proper type casting is discussed in the above article in the section: 
Casting through a union (1  2) which states The most commonly 
accepted method of converting one type of object to another is by using 
a union.


Note this is what both Jakub and Ulrich independently suggested in the 
bug report.


Also note that the suggested union contains a compatible member, which 
as I understand is necessary to satisfy the rules as well as providing 
enough 

Re: FreeRADIUS 2 not listening on right port

2008-05-29 Thread Alan DeKok
John Dennis wrote:
 I wanted to understand the issues surrounding strict aliasing better. I
 found the following article to be well written, quite readable, and
 informative:

  I found a NetBSD post with similar information:

http://mail-index.netbsd.org/tech-kern/2003/08/11/0001.html

 However, note that this is exactly what fr_socket is doing, salocal is
 a pointer to sockaddr_storage, sa is in one scope is a pointer to
 sockaddr_in and in another scope sa is a pointer to sockaddr_in6. Each
 of these 3 pointers point to exact same memory location, according to
 the above definitions this is a classic case of illegal pointer aliasing.

  Under ISO C89, and ISO C99.  But it appears to be legal under the
Posix description for struct sockaddr_storage.  Aren't standards
wonderful.

 It's a shame it didn't, but I don't think it changes the overarching
 issue. Also, in expediency I did not try building with the various GCC
 warning flags to see if I could get GCC to emit a warning for this case,
 perhaps GCC at some level of verbosity would have complained.

  I've done some builds with -fstrict-aliasing -Wstrict-aliasing=2.
Ignoring the casts on parameters to functions, I've fixed a number of
things, through the simple expedient of using memcpy() to copy from
(struct sockaddr_storage) to (struct sockaddr_in), and back.  It's ugly,
but it SHOULD work.

 For what it's worth my reading of the above statements is different than
 yours. I read the above as sockaddr_storage provides a type whose size
 is sufficient to hold all sockaddr types while respecting the alignment
 requirement in any given sockaddr type. The discussion of pointer
 casting fails to make explicit such casting must still conform to the
 rules of C99. I do not read the second paragraph as meaning pointer
 aliasing shall be supported, something outside that scope of that document.

  At the minimum, the Opengroup specs seem misleading.

 Proper type casting is discussed in the above article in the section:
 Casting through a union (1  2) which states The most commonly
 accepted method of converting one type of object to another is by using
 a union.
 
 Note this is what both Jakub and Ulrich independently suggested in the
 bug report.

  Which makes me wonder WTF is the use of 'struct sockaddr_storage'.
All of the documentation I've read indicates that it's *supposed* to be
the preferred structure to cast back and forth to 'struct
sockaddr_in'... but that's forbidden by ISO C89.

  You'd think that if the Posix people read the ISO C specs, they would
have created a 'struct sockaddr_union', which was defined to be the
union of all types, and which would have avoided these problems.

  Maybe I'm naive...

 I have created a patch for packet.c which uses a union (the patch is
 attached). I built freeradius 2.0.4 with the patch and -O2 optimization
 and the random port problem seems to have been resolved. Note, I have
 not looked though all the source code to see if there are other code
 constructs which may also require a fix.

  There's a number.

  Please try CVS head.  I've committed a whack of memcpy's, which SHOULD
avoid these issues.  (And avoid umpteen unions).

 Also, does anyone (Alan?) have any problems with updating the bug report
 with some of this email dialog to capture the issues for future readers?

  I'd like to say that for me, the Posix spec is clear: casting between
'struct sockaddr_storage' and other 'struct sockaddr_*' is explicitely
allowed.  Looking at code on the net, there are LOTS of programs doing this.

  What is a different about FreeRADIUS is that it's *accessing* the
pointer after the cast, and not just passing it to a function.  This
appears to be explicitly required by Posix:

  ...pointers to it can be cast as pointers to protocol-specific address
structures and used to access the fields of those structures without
alignment problems...

  i.e. TO ACCESS the fields.  ACCESS to me means ACCESS.  R/W ACCESS.

  Except that this is forbidden by ISO C89, and no one else does it.  sigh

  It serves me right for reading only half of the 10's of 1000's of
pages of specs.

  Alan DeKok.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-28 Thread John Dennis

John Dennis wrote:

Phil Mayers wrote:

Alan DeKok wrote:

Michael Griego wrote:
I did a little looking into this this evening.  This assessment 
looks to

be correct as it looks to be related to compiler optimizations.  With
the optimizations disabled in Make.inc, FreeRADIUS will start up on 
the
correct port.  For the fr_socket function, gcc appears to be 
optimizing
the arguments by sending them through the registers instead of the 
stack

frame, but the port argument is being clobbered (optimized out)
before the htons(port) call.  Specifically, according to a 
step-through

with GDB, after the first function call in fr_socket (which is to
socket()), the port variable is gone (optimized out).


  sigh  I've started testing the server with other compilers.  GCC is
getting too ugly for my liking.

  I'll put a note on the main web page.: DON'T USE -O2 ON FEDORA!


Please note, this bug only seems to be present in the F-9 (recently 
released version 9 of Fedora).

I can confirm this; I've opened a bug in the Fedora bugzilla:

https://bugzilla.redhat.com/show_bug.cgi?id=446864

...with any luck, the toolchain will get fixed - it's clearly not a 
FreeRadius bug, and I wonder what else it's broken...


FYI, I have rebuilt the F-9 RPM with optimization turned off and 
submitted it for an update (should hit the mirrors sometime soon).


Just a heads up. I have some time constraints at my end. The only 
thing I had time for before leaving until next Wednesday was to 
rebuild the package, I cannot not confirm at this moment if the 
resulting package resolves the issue, but given the comments in this 
thread it would appear to solve the problem in the near term. I 
thought it was worthwhile to push the new RPM out, even under these 
constraints.


When I return I will investigate the root cause of the failure 
further. Obviously this is a temporary workaround and if there are 
compiler issues it needs to get addressed as soon as possible.
I have investigated this further. It does appear as though there is a 
compiler bug based on my examination of the generated machine code, I 
have opened the following bug report against GCC. The bug report 
includes the preprocessed C code, the generated machine code, and the 
disassembly of the machine code. The bug report also points the exact 
line (compound statement) which seems to be omitted in the optimized 
machine code.


https://bugzilla.redhat.com/show_bug.cgi?id=448743

For the time being I will build the F-9 FreeRADIUS packages without 
optimization until this is resolved.


In the interim I would appreciate feedback on the new package.

The new package is: freeradius-2.0.3-3.fc9

My sincere thank you to everyone who contributed to diagnosing the 
problem and my apologies for not catching this earlier.





--
John Dennis [EMAIL PROTECTED]

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-28 Thread A . L . M . Buxey
Hi,

 Please note, this bug only seems to be present in the F-9 (recently 
 released version 9 of Fedora).

 For the time being I will build the F-9 FreeRADIUS packages without 
 optimization until this is resolved.

is it a case of this bug doing OTHER things to the codebase etc
or is it worth ONLY turning off optimization for the single
bit of affected code. turning off optimization for the routines
that deal with encryptions etc could be painful.

alan
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-28 Thread John Dennis

[EMAIL PROTECTED] wrote:

Hi,

  
Please note, this bug only seems to be present in the F-9 (recently 
released version 9 of Fedora).


For the time being I will build the F-9 FreeRADIUS packages without 
optimization until this is resolved.



is it a case of this bug doing OTHER things to the codebase etc
or is it worth ONLY turning off optimization for the single
bit of affected code. turning off optimization for the routines
that deal with encryptions etc could be painful.
  
This is a good question. The compiler folks have not responded yet so I 
don't know if this is an isolated issue in specific code or a more 
general problem which might rear it's head elsewhere. Unscientific 
hallway conversations have not turned up any other known F-9 compiler 
bugs, so this might be isolated to just one function. But in the absence 
of concrete information about the extent of the compiler malfunction the 
only prudent thing to do is to completely disable optimization on the 
belief correct program behaviour trumps performance, especially on a 
brand new OS release. When more concrete information is known I will 
consider rebuilding with different optimization.


If someone wants to be a guinea pig and run with optimization disabled 
only on packet.c and see if they have any anomalous behaviour then I'm 
sure we would all benefit from that experiment, however I cannot in good 
conscience push such a package into general distribution at this stage.


--
John Dennis [EMAIL PROTECTED]

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Re: FreeRADIUS 2 not listening on right port

2008-05-28 Thread A . L . M . Buxey
Hi,

 If someone wants to be a guinea pig and run with optimization disabled only 
 on packet.c and see if they have any anomalous behaviour then I'm sure we 
 would all benefit from that experiment, however I cannot in good conscience 
 push such a package into general distribution at this stage.

updates-testing ?  I thought that was the point of that 
repo ?

alan
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-28 Thread Phil Mayers

[EMAIL PROTECTED] wrote:

Hi,

Please note, this bug only seems to be present in the F-9 (recently 
released version 9 of Fedora).


For the time being I will build the F-9 FreeRADIUS packages without 
optimization until this is resolved.


is it a case of this bug doing OTHER things to the codebase etc
or is it worth ONLY turning off optimization for the single
bit of affected code. turning off optimization for the routines
that deal with encryptions etc could be painful.


All,

For those not following the Fedora bug, it (or rather, it's dependency) 
has been closed by Ulrich Drepper. He seems to be saying that the 
FreeRadius code is incorrect and specifically that an invalid typecast 
is triggering the compiler to generate bad code:


https://bugzilla.redhat.com/show_bug.cgi?id=448743#c6

Summary (as far as I can make out): Because the code looks like this:

struct addr

if v==4:
 i = (ip4*)addr
 i-port = xx

elif v==6:
 i = (ip6*)addr
 i-port = xx

bind(addr)

...the compiler can't detect that the i value is used, and optimises 
the code touching it away. The right way to do it is:


union {
  ip4 i4;
  ip6 i6;
} addr;
if v==4:
  addr.i4.port = xx
elif v==6:
  addr.i6.port = xx
bind(addr)

I believe the Python guys have had similar issues in the past; 
-fno-strict-aliasing might be required under gcc.


I'm not expert enough in the C99 standard to judge whether the comments 
at the above URL are correct or not; I suspect it's a matter of some 
controversy, and will back slowly away now... ;o)

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-16 Thread Phil Mayers

Alan DeKok wrote:

Michael Griego wrote:

I did a little looking into this this evening.  This assessment looks to
be correct as it looks to be related to compiler optimizations.  With
the optimizations disabled in Make.inc, FreeRADIUS will start up on the
correct port.  For the fr_socket function, gcc appears to be optimizing
the arguments by sending them through the registers instead of the stack
frame, but the port argument is being clobbered (optimized out)
before the htons(port) call.  Specifically, according to a step-through
with GDB, after the first function call in fr_socket (which is to
socket()), the port variable is gone (optimized out).


  sigh  I've started testing the server with other compilers.  GCC is
getting too ugly for my liking.

  I'll put a note on the main web page.: DON'T USE -O2 ON FEDORA!


I can confirm this; I've opened a bug in the Fedora bugzilla:

https://bugzilla.redhat.com/show_bug.cgi?id=446864

...with any luck, the toolchain will get fixed - it's clearly not a 
FreeRadius bug, and I wonder what else it's broken...

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


RE: FreeRADIUS 2 not listening on right port

2008-05-16 Thread Casartello, Thomas
Yeah I just compiled without optimization and its working fine now.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-Original Message-
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
.org] On Behalf Of Phil Mayers
Sent: Friday, May 16, 2008 10:49 AM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

Alan DeKok wrote:
 Michael Griego wrote:
 I did a little looking into this this evening.  This assessment looks
to
 be correct as it looks to be related to compiler optimizations.  With
 the optimizations disabled in Make.inc, FreeRADIUS will start up on
the
 correct port.  For the fr_socket function, gcc appears to be
optimizing
 the arguments by sending them through the registers instead of the
stack
 frame, but the port argument is being clobbered (optimized out)
 before the htons(port) call.  Specifically, according to a
step-through
 with GDB, after the first function call in fr_socket (which is to
 socket()), the port variable is gone (optimized out).
 
   sigh  I've started testing the server with other compilers.  GCC
is
 getting too ugly for my liking.
 
   I'll put a note on the main web page.: DON'T USE -O2 ON FEDORA!

I can confirm this; I've opened a bug in the Fedora bugzilla:

https://bugzilla.redhat.com/show_bug.cgi?id=446864

...with any luck, the toolchain will get fixed - it's clearly not a 
FreeRadius bug, and I wonder what else it's broken...
-
List info/subscribe/unsubscribe? See
http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-16 Thread Michael Griego

On May 16, 2008, at 9:49 AM, Phil Mayers wrote:

...with any luck, the toolchain will get fixed - it's clearly not a  
FreeRadius bug, and I wonder what else it's broken...



After discovering what the problem was, I immediately wondered the  
same thing myself.


--Mike
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-16 Thread John Dennis

Phil Mayers wrote:

Alan DeKok wrote:

Michael Griego wrote:

I did a little looking into this this evening.  This assessment looks to
be correct as it looks to be related to compiler optimizations.  With
the optimizations disabled in Make.inc, FreeRADIUS will start up on the
correct port.  For the fr_socket function, gcc appears to be optimizing
the arguments by sending them through the registers instead of the stack
frame, but the port argument is being clobbered (optimized out)
before the htons(port) call.  Specifically, according to a step-through
with GDB, after the first function call in fr_socket (which is to
socket()), the port variable is gone (optimized out).


  sigh  I've started testing the server with other compilers.  GCC is
getting too ugly for my liking.

  I'll put a note on the main web page.: DON'T USE -O2 ON FEDORA!


I can confirm this; I've opened a bug in the Fedora bugzilla:

https://bugzilla.redhat.com/show_bug.cgi?id=446864

...with any luck, the toolchain will get fixed - it's clearly not a 
FreeRadius bug, and I wonder what else it's broken...


FYI, I have rebuilt the F-9 RPM with optimization turned off and 
submitted it for an update (should hit the mirrors sometime soon).


Just a heads up. I have some time constraints at my end. The only thing 
I had time for before leaving until next Wednesday was to rebuild the 
package, I cannot not confirm at this moment if the resulting package 
resolves the issue, but given the comments in this thread it would 
appear to solve the problem in the near term. I thought it was 
worthwhile to push the new RPM out, even under these constraints.


When I return I will investigate the root cause of the failure further. 
Obviously this is a temporary workaround and if there are compiler 
issues it needs to get addressed as soon as possible.


In the interim I would appreciate feedback on the new package.

The new package is: freeradius-2.0.3-3.fc9

My sincere thank you to everyone who contributed to diagnosing the 
problem and my apologies for not catching this earlier.


--
John Dennis [EMAIL PROTECTED]
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-16 Thread Alan DeKok
John Dennis wrote:
 FYI, I have rebuilt the F-9 RPM with optimization turned off and
 submitted it for an update (should hit the mirrors sometime soon).

  Did I mention I love Redhat?

  Thanks.  Fast turn-around times are wonderful.

 My sincere thank you to everyone who contributed to diagnosing the
 problem and my apologies for not catching this earlier.

  If I can get access to a build machine, I'm sure I can do loads of
testing...

  Alan DeKok.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-16 Thread Marinko Tarlac

What did you expect from Fedora... To many bugs for my taste :)

Michael Griego wrote:

On May 16, 2008, at 9:49 AM, Phil Mayers wrote:

...with any luck, the toolchain will get fixed - it's clearly not a 
FreeRadius bug, and I wonder what else it's broken...



After discovering what the problem was, I immediately wondered the 
same thing myself.


--Mike
-
List info/subscribe/unsubscribe? See 
http://www.freeradius.org/list/users.html




-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


FreeRADIUS 2 not listening on right port

2008-05-15 Thread Casartello, Thomas
I just upgraded by FreeRADIUS server from the version 1 to version 2
family. I have the listen {} statements configured as follows:

radiusd:  Opening IP addresses and Ports 

listen {

type = auth

ipaddr = *

port = 1812

}

listen {

type = acct

ipaddr = *

port = 1813

}

main {

snmp = no

smux_password = 

snmp_write_access = no

}

Listening on authentication address * port 41045

Listening on accounting address * port 54893

Listening on proxy address * port 38374

Ready to process requests.

 

However as you can see if always listens on random ports. What am I
doing wrong? I am using version 2.0.2 which was distributed with Fedora
9.

 

Thomas E. Casartello, Jr.

Infrastructure Technician

Linux Specialist

Department of Information Technology

Westfield State College

Wilson 105-A

(413) 572-8245

E-Mail: [EMAIL PROTECTED]

 

Red Hat Certified Technician (RHCT)

 

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Casartello, Thomas
Compiling from source did NOT solve the problem.

 

Thomas E. Casartello, Jr.

Infrastructure Technician

Linux Specialist

Department of Information Technology

Westfield State College

Wilson 105-A

(413) 572-8245

E-Mail: [EMAIL PROTECTED]

 

Red Hat Certified Technician (RHCT)

 

From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
.org] On Behalf Of Casartello, Thomas
Sent: Thursday, May 15, 2008 1:16 PM
To: freeradius-users@lists.freeradius.org
Subject: FreeRADIUS 2 not listening on right port

 

I just upgraded by FreeRADIUS server from the version 1 to version 2
family. I have the listen {} statements configured as follows:

radiusd:  Opening IP addresses and Ports 

listen {

type = auth

ipaddr = *

port = 1812

}

listen {

type = acct

ipaddr = *

port = 1813

}

main {

snmp = no

smux_password = 

snmp_write_access = no

}

Listening on authentication address * port 41045

Listening on accounting address * port 54893

Listening on proxy address * port 38374

Ready to process requests.

 

However as you can see if always listens on random ports. What am I
doing wrong? I am using version 2.0.2 which was distributed with Fedora
9.

 

Thomas E. Casartello, Jr.

Infrastructure Technician

Linux Specialist

Department of Information Technology

Westfield State College

Wilson 105-A

(413) 572-8245

E-Mail: [EMAIL PROTECTED]

 

Red Hat Certified Technician (RHCT)

 

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Danner, Mearl
You're not running NAT/PAT through iptables are you?

It'll translate 1812/1813 inside to some high port/some high port outside.

Not sure how the server will pick that up. Maybe the port after translation.

If so you'll need to not port translate the radius ports. I can do it in a Pix, 
but haven't used iptables for translation in a long while.

Mearl

From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Casartello, 
Thomas
Sent: Thursday, May 15, 2008 12:31 PM
To: FreeRadius users mailing list
Subject: RE: FreeRADIUS 2 not listening on right port

Compiling from source did NOT solve the problem.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)

From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Casartello, 
Thomas
Sent: Thursday, May 15, 2008 1:16 PM
To: freeradius-users@lists.freeradius.org
Subject: FreeRADIUS 2 not listening on right port

I just upgraded by FreeRADIUS server from the version 1 to version 2 family. I 
have the listen {} statements configured as follows:
radiusd:  Opening IP addresses and Ports 
listen {
    type = auth
    ipaddr = *
    port = 1812
}
listen {
    type = acct
    ipaddr = *
    port = 1813
}
main {
    snmp = no
    smux_password = 
    snmp_write_access = no
}
Listening on authentication address * port 41045
Listening on accounting address * port 54893
Listening on proxy address * port 38374
Ready to process requests.

However as you can see if always listens on random ports. What am I doing 
wrong? I am using version 2.0.2 which was distributed with Fedora 9.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Casartello, Thomas
No I am not doing any kind of NAT. I actually have IPTables disabled right now.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)

-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Danner, Mearl
Sent: Thursday, May 15, 2008 1:42 PM
To: FreeRadius users mailing list
Subject: RE: FreeRADIUS 2 not listening on right port

You're not running NAT/PAT through iptables are you?

It'll translate 1812/1813 inside to some high port/some high port outside.

Not sure how the server will pick that up. Maybe the port after translation.

If so you'll need to not port translate the radius ports. I can do it in a Pix, 
but haven't used iptables for translation in a long while.

Mearl

From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Casartello, 
Thomas
Sent: Thursday, May 15, 2008 12:31 PM
To: FreeRadius users mailing list
Subject: RE: FreeRADIUS 2 not listening on right port

Compiling from source did NOT solve the problem.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)

From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Casartello, 
Thomas
Sent: Thursday, May 15, 2008 1:16 PM
To: freeradius-users@lists.freeradius.org
Subject: FreeRADIUS 2 not listening on right port

I just upgraded by FreeRADIUS server from the version 1 to version 2 family. I 
have the listen {} statements configured as follows:
radiusd:  Opening IP addresses and Ports 
listen {
    type = auth
    ipaddr = *
    port = 1812
}
listen {
    type = acct
    ipaddr = *
    port = 1813
}
main {
    snmp = no
    smux_password = 
    snmp_write_access = no
}
Listening on authentication address * port 41045
Listening on accounting address * port 54893
Listening on proxy address * port 38374
Ready to process requests.

However as you can see if always listens on random ports. What am I doing 
wrong? I am using version 2.0.2 which was distributed with Fedora 9.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Danner, Mearl
Have you tried binding to a specific IP address rather than *?

 -Original Message-
 From: freeradius-users-
 [EMAIL PROTECTED] [mailto:freeradius-
 [EMAIL PROTECTED] On Behalf Of
 Casartello, Thomas
 Sent: Thursday, May 15, 2008 12:44 PM
 To: FreeRadius users mailing list
 Subject: RE: FreeRADIUS 2 not listening on right port
 
 No I am not doing any kind of NAT. I actually have IPTables disabled
 right now.
 
 Thomas E. Casartello, Jr.
 Infrastructure Technician
 Linux Specialist
 Department of Information Technology
 Westfield State College
 Wilson 105-A
 (413) 572-8245
 E-Mail: [EMAIL PROTECTED]
 
 Red Hat Certified Technician (RHCT)
 
 -Original Message-
 From: freeradius-users-
 [EMAIL PROTECTED] [mailto:freeradius-
 [EMAIL PROTECTED] On Behalf Of
 Danner, Mearl
 Sent: Thursday, May 15, 2008 1:42 PM
 To: FreeRadius users mailing list
 Subject: RE: FreeRADIUS 2 not listening on right port
 
 You're not running NAT/PAT through iptables are you?
 
 It'll translate 1812/1813 inside to some high port/some high port
 outside.
 
 Not sure how the server will pick that up. Maybe the port after
 translation.
 
 If so you'll need to not port translate the radius ports. I can do it
 in a Pix, but haven't used iptables for translation in a long while.
 
 Mearl
 
 From: freeradius-users-
 [EMAIL PROTECTED] [mailto:freeradius-
 [EMAIL PROTECTED] On Behalf Of
 Casartello, Thomas
 Sent: Thursday, May 15, 2008 12:31 PM
 To: FreeRadius users mailing list
 Subject: RE: FreeRADIUS 2 not listening on right port
 
 Compiling from source did NOT solve the problem.
 
 Thomas E. Casartello, Jr.
 Infrastructure Technician
 Linux Specialist
 Department of Information Technology
 Westfield State College
 Wilson 105-A
 (413) 572-8245
 E-Mail: [EMAIL PROTECTED]
 
 Red Hat Certified Technician (RHCT)
 
 From: freeradius-users-
 [EMAIL PROTECTED] [mailto:freeradius-
 [EMAIL PROTECTED] On Behalf Of
 Casartello, Thomas
 Sent: Thursday, May 15, 2008 1:16 PM
 To: freeradius-users@lists.freeradius.org
 Subject: FreeRADIUS 2 not listening on right port
 
 I just upgraded by FreeRADIUS server from the version 1 to version 2
 family. I have the listen {} statements configured as follows:
 radiusd:  Opening IP addresses and Ports 
 listen {
     type = auth
     ipaddr = *
     port = 1812
 }
 listen {
     type = acct
     ipaddr = *
     port = 1813
 }
 main {
     snmp = no
     smux_password = 
     snmp_write_access = no
 }
 Listening on authentication address * port 41045
 Listening on accounting address * port 54893
 Listening on proxy address * port 38374
 Ready to process requests.
 
 However as you can see if always listens on random ports. What am I
 doing wrong? I am using version 2.0.2 which was distributed with Fedora
 9.
 
 Thomas E. Casartello, Jr.
 Infrastructure Technician
 Linux Specialist
 Department of Information Technology
 Westfield State College
 Wilson 105-A
 (413) 572-8245
 E-Mail: [EMAIL PROTECTED]
 
 Red Hat Certified Technician (RHCT)
 
 
 -
 List info/subscribe/unsubscribe? See
 http://www.freeradius.org/list/users.html
 
 -
 List info/subscribe/unsubscribe? See
 http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Casartello, Thomas
Yes. Same result. I went back to 1.1.7 on the same box and its working fine now.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Danner, Mearl
Sent: Thursday, May 15, 2008 2:01 PM
To: FreeRadius users mailing list
Subject: RE: FreeRADIUS 2 not listening on right port

Have you tried binding to a specific IP address rather than *?

 -Original Message-
 From: freeradius-users-
 [EMAIL PROTECTED] [mailto:freeradius-
 [EMAIL PROTECTED] On Behalf Of
 Casartello, Thomas
 Sent: Thursday, May 15, 2008 12:44 PM
 To: FreeRadius users mailing list
 Subject: RE: FreeRADIUS 2 not listening on right port
 
 No I am not doing any kind of NAT. I actually have IPTables disabled
 right now.
 
 Thomas E. Casartello, Jr.
 Infrastructure Technician
 Linux Specialist
 Department of Information Technology
 Westfield State College
 Wilson 105-A
 (413) 572-8245
 E-Mail: [EMAIL PROTECTED]
 
 Red Hat Certified Technician (RHCT)
 
 -Original Message-
 From: freeradius-users-
 [EMAIL PROTECTED] [mailto:freeradius-
 [EMAIL PROTECTED] On Behalf Of
 Danner, Mearl
 Sent: Thursday, May 15, 2008 1:42 PM
 To: FreeRadius users mailing list
 Subject: RE: FreeRADIUS 2 not listening on right port
 
 You're not running NAT/PAT through iptables are you?
 
 It'll translate 1812/1813 inside to some high port/some high port
 outside.
 
 Not sure how the server will pick that up. Maybe the port after
 translation.
 
 If so you'll need to not port translate the radius ports. I can do it
 in a Pix, but haven't used iptables for translation in a long while.
 
 Mearl
 
 From: freeradius-users-
 [EMAIL PROTECTED] [mailto:freeradius-
 [EMAIL PROTECTED] On Behalf Of
 Casartello, Thomas
 Sent: Thursday, May 15, 2008 12:31 PM
 To: FreeRadius users mailing list
 Subject: RE: FreeRADIUS 2 not listening on right port
 
 Compiling from source did NOT solve the problem.
 
 Thomas E. Casartello, Jr.
 Infrastructure Technician
 Linux Specialist
 Department of Information Technology
 Westfield State College
 Wilson 105-A
 (413) 572-8245
 E-Mail: [EMAIL PROTECTED]
 
 Red Hat Certified Technician (RHCT)
 
 From: freeradius-users-
 [EMAIL PROTECTED] [mailto:freeradius-
 [EMAIL PROTECTED] On Behalf Of
 Casartello, Thomas
 Sent: Thursday, May 15, 2008 1:16 PM
 To: freeradius-users@lists.freeradius.org
 Subject: FreeRADIUS 2 not listening on right port
 
 I just upgraded by FreeRADIUS server from the version 1 to version 2
 family. I have the listen {} statements configured as follows:
 radiusd:  Opening IP addresses and Ports 
 listen {
     type = auth
     ipaddr = *
     port = 1812
 }
 listen {
     type = acct
     ipaddr = *
     port = 1813
 }
 main {
     snmp = no
     smux_password = 
     snmp_write_access = no
 }
 Listening on authentication address * port 41045
 Listening on accounting address * port 54893
 Listening on proxy address * port 38374
 Ready to process requests.
 
 However as you can see if always listens on random ports. What am I
 doing wrong? I am using version 2.0.2 which was distributed with Fedora
 9.
 
 Thomas E. Casartello, Jr.
 Infrastructure Technician
 Linux Specialist
 Department of Information Technology
 Westfield State College
 Wilson 105-A
 (413) 572-8245
 E-Mail: [EMAIL PROTECTED]
 
 Red Hat Certified Technician (RHCT)
 
 
 -
 List info/subscribe/unsubscribe? See
 http://www.freeradius.org/list/users.html
 
 -
 List info/subscribe/unsubscribe? See
 http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Alan DeKok
Casartello, Thomas wrote:
 Compiling from source did NOT solve the problem.

  It looks like Fedora is broken.

  The server code does this:

  if (port == 0) {
call system function to look up radius port in /etc/services
if (found ) {
port = port found in /etc/services
} else {
   port = 1812
}
  }

  The only way I can see it choosing random ports is if the lookup in
/etc/services returns found, with a random port.

  I suggest hard-coding the port numbers (1812/1813) into the listen
sections.  Maybe also see if 'radius and radacct are defined in
/etc/services.

  Alan DeKok.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Casartello, Thomas
I tried hardcoding them in the listen section. Same result.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-Original Message-
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
.org] On Behalf Of Alan DeKok
Sent: Thursday, May 15, 2008 2:16 PM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

Casartello, Thomas wrote:
 Compiling from source did NOT solve the problem.

  It looks like Fedora is broken.

  The server code does this:

  if (port == 0) {
call system function to look up radius port in /etc/services
if (found ) {
port = port found in /etc/services
} else {
   port = 1812
}
  }

  The only way I can see it choosing random ports is if the lookup in
/etc/services returns found, with a random port.

  I suggest hard-coding the port numbers (1812/1813) into the listen
sections.  Maybe also see if 'radius and radacct are defined in
/etc/services.

  Alan DeKok.
-
List info/subscribe/unsubscribe? See
http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Alan DeKok
Casartello, Thomas wrote:
 I tried hardcoding them in the listen section. Same result.

  Weird.

  My guess, then, is that it seems to be a problem with the specific GCC
version on Fedora.

  Please try the attached patch.  If it doesn't work, then the only way
to fix it is for me to get an SSH login to a fedora machine.

  Oh, and 2.0.4. works on Ubuntu, Debian, *BSD, Solaris...

  Alan DeKok.
Index: src/lib/packet.c
===
RCS file: /source/radiusd/src/lib/packet.c,v
retrieving revision 1.20
diff -u -r1.20 packet.c
--- src/lib/packet.c	1 Jan 2008 17:29:12 -	1.20
+++ src/lib/packet.c	15 May 2008 19:34:22 -
@@ -175,6 +175,7 @@
 int fr_socket(fr_ipaddr_t *ipaddr, int port)
 {
 	int sockfd;
+	uint16_t sport;
 	struct sockaddr_storage salocal;
 	socklen_t	salen;
 
@@ -185,6 +186,7 @@
 
 	sockfd = socket(ipaddr-af, SOCK_DGRAM, 0);
 	if (sockfd  0) {
+		librad_log(cannot open socket: %s, strerror(errno));
 		return sockfd;
 	}
 
@@ -194,10 +196,13 @@
 	 */
 	if (udpfromto_init(sockfd) != 0) {
 		close(sockfd);
+		librad_log(cannot initialize udpfromto: %s, strerror(errno));
 		return -1;
 	}
 #endif
 
+	sport = port;
+	sport = htons(sport);
 	memset(salocal, 0, sizeof(salocal));
 	if (ipaddr-af == AF_INET) {
 		struct sockaddr_in *sa;
@@ -205,7 +210,7 @@
 		sa = (struct sockaddr_in *) salocal;
 		sa-sin_family = AF_INET;
 		sa-sin_addr = ipaddr-ipaddr.ip4addr;
-		sa-sin_port = htons((uint16_t) port);
+		sa-sin_port = sport;
 		salen = sizeof(*sa);
 
 #ifdef HAVE_STRUCT_SOCKADDR_IN6
@@ -215,7 +220,7 @@
 		sa = (struct sockaddr_in6 *) salocal;
 		sa-sin6_family = AF_INET6;
 		sa-sin6_addr = ipaddr-ipaddr.ip6addr;
-		sa-sin6_port = htons((uint16_t) port);
+		sa-sin6_port = sport;
 		salen = sizeof(*sa);
 
 #if 1
@@ -242,6 +247,7 @@
 
 	if (bind(sockfd, (struct sockaddr *) salocal, salen)  0) {
 		close(sockfd);
+		librad_log(cannot bind socket: %s, strerror(errno));
 		return -1;
 	}
 
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread A . L . M . Buxey
Hi,
 I tried hardcoding them in the listen section. Same result.

TBH, I've compiled release and CVS versions of freeradius 1.1.x
and 2.0.x on centos, fedora core, RHEL3, ubuntu 7 and 8
and have never seen this issue before. 

you running SELinux or some sort of security tool?

alan
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread A . L . M . Buxey
Hi,
 Casartello, Thomas wrote:
  I tried hardcoding them in the listen section. Same result.

64bit machine?

alan
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Hoggins!

Hi,

Exact same problem here... Really thinking about reverting to v1.x

Casartello, Thomas a écrit :

I tried hardcoding them in the listen section. Same result.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-Original Message-
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
.org] On Behalf Of Alan DeKok
Sent: Thursday, May 15, 2008 2:16 PM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

Casartello, Thomas wrote:
  

Compiling from source did NOT solve the problem.



  It looks like Fedora is broken.

  The server code does this:

  if (port == 0) {
call system function to look up radius port in /etc/services
if (found ) {
port = port found in /etc/services
} else {
   port = 1812
}
  }

  The only way I can see it choosing random ports is if the lookup in
/etc/services returns found, with a random port.

  I suggest hard-coding the port numbers (1812/1813) into the listen
sections.  Maybe also see if 'radius and radacct are defined in
/etc/services.

  Alan DeKok.
-
List info/subscribe/unsubscribe? See
http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
  

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Hoggins!
I'm running FC9, by the way... maybe that explains this sudden amount of 
same problems, since the FC9 release was on tuesday.


Casartello, Thomas a écrit :

I tried hardcoding them in the listen section. Same result.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-Original Message-
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
.org] On Behalf Of Alan DeKok
Sent: Thursday, May 15, 2008 2:16 PM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

Casartello, Thomas wrote:
  

Compiling from source did NOT solve the problem.



  It looks like Fedora is broken.

  The server code does this:

  if (port == 0) {
call system function to look up radius port in /etc/services
if (found ) {
port = port found in /etc/services
} else {
   port = 1812
}
  }

  The only way I can see it choosing random ports is if the lookup in
/etc/services returns found, with a random port.

  I suggest hard-coding the port numbers (1812/1813) into the listen
sections.  Maybe also see if 'radius and radacct are defined in
/etc/services.

  Alan DeKok.
-
List info/subscribe/unsubscribe? See
http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
  

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Casartello, Thomas
As am I.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Hoggins!
Sent: Thursday, May 15, 2008 4:05 PM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

I'm running FC9, by the way... maybe that explains this sudden amount of 
same problems, since the FC9 release was on tuesday.

Casartello, Thomas a écrit :
 I tried hardcoding them in the listen section. Same result.

 Thomas E. Casartello, Jr.
 Infrastructure Technician
 Linux Specialist
 Department of Information Technology
 Westfield State College
 Wilson 105-A
 (413) 572-8245
 E-Mail: [EMAIL PROTECTED]

 Red Hat Certified Technician (RHCT)


 -Original Message-
 From:
 [EMAIL PROTECTED]
 [mailto:[EMAIL PROTECTED]
 .org] On Behalf Of Alan DeKok
 Sent: Thursday, May 15, 2008 2:16 PM
 To: FreeRadius users mailing list
 Subject: Re: FreeRADIUS 2 not listening on right port

 Casartello, Thomas wrote:
   
 Compiling from source did NOT solve the problem.
 

   It looks like Fedora is broken.

   The server code does this:

   if (port == 0) {
 call system function to look up radius port in /etc/services
 if (found ) {
 port = port found in /etc/services
 } else {
port = 1812
 }
   }

   The only way I can see it choosing random ports is if the lookup in
 /etc/services returns found, with a random port.

   I suggest hard-coding the port numbers (1812/1813) into the listen
 sections.  Maybe also see if 'radius and radacct are defined in
 /etc/services.

   Alan DeKok.
 -
 List info/subscribe/unsubscribe? See
 http://www.freeradius.org/list/users.html

 -
 List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
   
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Casartello, Thomas
No luck on that patch. I'll try to get you a login sometime over the
next couple days.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-Original Message-
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
.org] On Behalf Of Alan DeKok
Sent: Thursday, May 15, 2008 3:32 PM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

Casartello, Thomas wrote:
 I tried hardcoding them in the listen section. Same result.

  Weird.

  My guess, then, is that it seems to be a problem with the specific GCC
version on Fedora.

  Please try the attached patch.  If it doesn't work, then the only way
to fix it is for me to get an SSH login to a fedora machine.

  Oh, and 2.0.4. works on Ubuntu, Debian, *BSD, Solaris...

  Alan DeKok.

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Alan DeKok
Hoggins! wrote:
 I'm running FC9, by the way... maybe that explains this sudden amount of
 same problems, since the FC9 release was on tuesday.

  Maybe someone running FC9 could try debugging the problem.

  I haven't run a redhat-based system for *years*.

  Since this works on every other system on the planet, it sounds *very*
much like an issue in FC9.

  Alan DeKok.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread A . L . M . Buxey
Hi,
 I'm running FC9, by the way... maybe that explains this sudden amount of 
 same problems, since the FC9 release was on tuesday.

yep. havent tested FC9 - wonder what they've changed to make such
a change in port behaviour..

alan
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread A . L . M . Buxey
Hi,

   Maybe someone running FC9 could try debugging the problem.

as, no doubt, one of my systems will be FC9 in a short
while I could look att his - what exactly should I be
looking for? i'll dig around for the new features
and changes they've made.


alan
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Casartello, Thomas
Install the freeradius rpm or install from source. It basically binds to
a random port no matter what you do in the config files. Freeradius
1.1.7 works fine in Fedora 9. I'm going to try using 2.0.4 on Fedora 8
box.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)

-Original Message-
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
.org] On Behalf Of [EMAIL PROTECTED]
Sent: Thursday, May 15, 2008 5:00 PM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

Hi,

   Maybe someone running FC9 could try debugging the problem.

as, no doubt, one of my systems will be FC9 in a short
while I could look att his - what exactly should I be
looking for? i'll dig around for the new features
and changes they've made.


alan
-
List info/subscribe/unsubscribe? See
http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Casartello, Thomas
Fedora 9 did do a pretty big gcc version jump. Fedora 8 used 4.1.2,
while 9 uses 4.3.0. BTW I tested it in Fedora 8 and it worked fine, so
it's definitely a 9 issue.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-Original Message-
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
.org] On Behalf Of Alan DeKok
Sent: Thursday, May 15, 2008 4:22 PM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

Hoggins! wrote:
 I'm running FC9, by the way... maybe that explains this sudden amount
of
 same problems, since the FC9 release was on tuesday.

  Maybe someone running FC9 could try debugging the problem.

  I haven't run a redhat-based system for *years*.

  Since this works on every other system on the planet, it sounds *very*
much like an issue in FC9.

  Alan DeKok.
-
List info/subscribe/unsubscribe? See
http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Hoggins!
Shouldn't the maintainer of the specific FC9 freeradius package be aware 
of this critical issue ?

I guess a newer release is for very soon.

Casartello, Thomas a écrit :

Fedora 9 did do a pretty big gcc version jump. Fedora 8 used 4.1.2,
while 9 uses 4.3.0. BTW I tested it in Fedora 8 and it worked fine, so
it's definitely a 9 issue.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-Original Message-
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
.org] On Behalf Of Alan DeKok
Sent: Thursday, May 15, 2008 4:22 PM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

Hoggins! wrote:
  

I'm running FC9, by the way... maybe that explains this sudden amount


of
  

same problems, since the FC9 release was on tuesday.



  Maybe someone running FC9 could try debugging the problem.

  I haven't run a redhat-based system for *years*.

  Since this works on every other system on the planet, it sounds *very*
much like an issue in FC9.

  Alan DeKok.
-
List info/subscribe/unsubscribe? See
http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
  

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


RE: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Casartello, Thomas
It's not just the Fedora package. Even if you compile the latest freeradius 
from source it still has the problem.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)

-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Hoggins!
Sent: Thursday, May 15, 2008 8:21 PM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

Shouldn't the maintainer of the specific FC9 freeradius package be aware 
of this critical issue ?
I guess a newer release is for very soon.

Casartello, Thomas a écrit :
 Fedora 9 did do a pretty big gcc version jump. Fedora 8 used 4.1.2,
 while 9 uses 4.3.0. BTW I tested it in Fedora 8 and it worked fine, so
 it's definitely a 9 issue.

 Thomas E. Casartello, Jr.
 Infrastructure Technician
 Linux Specialist
 Department of Information Technology
 Westfield State College
 Wilson 105-A
 (413) 572-8245
 E-Mail: [EMAIL PROTECTED]

 Red Hat Certified Technician (RHCT)


 -Original Message-
 From:
 [EMAIL PROTECTED]
 [mailto:[EMAIL PROTECTED]
 .org] On Behalf Of Alan DeKok
 Sent: Thursday, May 15, 2008 4:22 PM
 To: FreeRadius users mailing list
 Subject: Re: FreeRADIUS 2 not listening on right port

 Hoggins! wrote:
   
 I'm running FC9, by the way... maybe that explains this sudden amount
 
 of
   
 same problems, since the FC9 release was on tuesday.
 

   Maybe someone running FC9 could try debugging the problem.

   I haven't run a redhat-based system for *years*.

   Since this works on every other system on the planet, it sounds *very*
 much like an issue in FC9.

   Alan DeKok.
 -
 List info/subscribe/unsubscribe? See
 http://www.freeradius.org/list/users.html

 -
 List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
   
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Michael Griego
I did a little looking into this this evening.  This assessment looks  
to be correct as it looks to be related to compiler optimizations.   
With the optimizations disabled in Make.inc, FreeRADIUS will start up  
on the correct port.  For the fr_socket function, gcc appears to be  
optimizing the arguments by sending them through the registers instead  
of the stack frame, but the port argument is being clobbered  
(optimized out) before the htons(port) call.  Specifically,  
according to a step-through with GDB, after the first function call in  
fr_socket (which is to socket()), the port variable is gone  
(optimized out).


--Mike

On May 15, 2008, at 4:30 PM, Casartello, Thomas wrote:


Fedora 9 did do a pretty big gcc version jump. Fedora 8 used 4.1.2,
while 9 uses 4.3.0. BTW I tested it in Fedora 8 and it worked fine, so
it's definitely a 9 issue.

Thomas E. Casartello, Jr.
Infrastructure Technician
Linux Specialist
Department of Information Technology
Westfield State College
Wilson 105-A
(413) 572-8245
E-Mail: [EMAIL PROTECTED]

Red Hat Certified Technician (RHCT)


-Original Message-
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
.org] On Behalf Of Alan DeKok
Sent: Thursday, May 15, 2008 4:22 PM
To: FreeRadius users mailing list
Subject: Re: FreeRADIUS 2 not listening on right port

Hoggins! wrote:

I'm running FC9, by the way... maybe that explains this sudden amount

of

same problems, since the FC9 release was on tuesday.


 Maybe someone running FC9 could try debugging the problem.

 I haven't run a redhat-based system for *years*.

 Since this works on every other system on the planet, it sounds  
*very*

much like an issue in FC9.

 Alan DeKok.
-
List info/subscribe/unsubscribe? See
http://www.freeradius.org/list/users.html

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: FreeRADIUS 2 not listening on right port

2008-05-15 Thread Alan DeKok
Michael Griego wrote:
 I did a little looking into this this evening.  This assessment looks to
 be correct as it looks to be related to compiler optimizations.  With
 the optimizations disabled in Make.inc, FreeRADIUS will start up on the
 correct port.  For the fr_socket function, gcc appears to be optimizing
 the arguments by sending them through the registers instead of the stack
 frame, but the port argument is being clobbered (optimized out)
 before the htons(port) call.  Specifically, according to a step-through
 with GDB, after the first function call in fr_socket (which is to
 socket()), the port variable is gone (optimized out).

  sigh  I've started testing the server with other compilers.  GCC is
getting too ugly for my liking.

  I'll put a note on the main web page.: DON'T USE -O2 ON FEDORA!

  Alan DeKok.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html