segfault when rlm_perl used
Hi, i build FR 2.1.7 on slackware 13.0 (glibc-2.9) with perl 5.10.0. every time, i kill radiusd, message like radiusd[5754]: segfault at 0 ip b73e2213 sp bfb86dac error 4 in libc-2.9.so[b736b000+15a000] appears in the log. This is not good bad, but worst when radiusd crashed on run-time. Then in logs appears message like: Oct 13 21:58:53 rs kernel: radiusd[11441]: segfault at 20004 ip b7478636 sp b502bcb0 error 4 in rlm_perl-2.1.7.so[b73cb000+15b000] Oct 14 22:09:56 rs kernel: radiusd[17687]: segfault at 8 ip b731e35d sp b6f61ce0 error 4 in rlm_perl-2.1.7.so[b71fd000+15b000] Oct 14 22:14:06 rs kernel: radiusd[18374]: segfault at 8 ip b73d735d sp b701ace0 error 4 in rlm_perl-2.1.7.so[b72b6000+15b000] Oct 15 18:07:58 rs kernel: radiusd[23858]: segfault at 8 ip b745935d sp ae89cce0 error 4 in rlm_perl-2.1.7.so[b7338000+15b000] Oct 15 18:08:56 rs kernel: radiusd[23896]: segfault at c ip b74d400e sp ae919c90 error 4[b73b5000+15b000] Oct 15 18:09:55 rs kernel: radiusd[24042]: segfault at 8 ip b736935d sp b07acce0 error 4 in rlm_perl-2.1.7.so[b7248000+15b000] when i comment out all perl statement in config, any segfault (in rlm_perl-2.1.7.so and in libc-2.9.so) disappears. Regards, Michael. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re[2]: segfault when rlm_perl used
Здравствуйте, FreeRadius! Вы писали 16 октября 2009 г., 18:03:17: On Oct 16, 2009, at 2:57 PM, Michael Chernyakhovsky wrote: Oct 13 21:58:53 rs kernel: radiusd[11441]: segfault at 20004 ip b7478636 sp b502bcb0 error 4 in rlm_perl-2.1.7.so[b73cb000+15b000] Oct 14 22:09:56 rs kernel: radiusd[17687]: segfault at 8 ip b731e35d sp b6f61ce0 error 4 in rlm_perl-2.1.7.so[b71fd000+15b000] Oct 14 22:14:06 rs kernel: radiusd[18374]: segfault at 8 ip b73d735d sp b701ace0 error 4 in rlm_perl-2.1.7.so[b72b6000+15b000] Oct 15 18:07:58 rs kernel: radiusd[23858]: segfault at 8 ip b745935d sp ae89cce0 error 4 in rlm_perl-2.1.7.so[b7338000+15b000] Oct 15 18:08:56 rs kernel: radiusd[23896]: segfault at c ip b74d400e sp ae919c90 error 4[b73b5000+15b000] Oct 15 18:09:55 rs kernel: radiusd[24042]: segfault at 8 ip b736935d sp b07acce0 error 4 in rlm_perl-2.1.7.so[b7248000+15b000] Could you include the output of gdb, as suggested in doc/bugs? Bug 31 created: https://bugs.freeradius.org/bugzilla/show_bug.cgi?id=31 - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html -- С уважением, Михаил Черняховский, НП Магинфоцентр, г. Магнитогорск. тел. (3519) 49-69-00, моб. +7 902 896 2872 - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re[2]: redundant ippools using rlm_ippool in v1.1.3
Здравствуйте, FreeRadius!
Вы писали 16 сентября 2008 г., 23:23:13:
The patch works as promised in version 1.1.3 if applied. The
post-auth section in radiusd.conf doesn't though.
As suggested by Michael with Kostas' patch applied you have to pass
DEFAULT as Pool-Name check attribute in 'users'.
You should to set Pool-Name to DEFAULT in 'users'. like this
DEFAULT Huntgroup-Name == lns0x, Pool-Name := DEFAULT,
Post-Auth-Type := LNS0X
Fall-Through = Yes
DEFAULT Huntgroup-Name == lns0y, Pool-Name := DEFAULT,
Post-Auth-Type := LNS0Y
Fall-Through = Yes
The post-auth section as posted by me does not work:
Post-Auth-Type LNS0X {
redundant {
lns0x-00
lns0x-01
}
}
what about simple:
Post-Auth-Type LNS0X {
lns0x-00
lns0x-01
}
Post-Auth-Type LNS0Y {
lns0y-00
lns0y-01
}
lns0x-01 is for unknown reasons not called. It has to be:
are lns0x-00 have no free entries?
in my case second pool used only when first does't give free ip.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re[2]: rlm_detail bug when locking = yes
Nicolas wrote:
we need close outfd right before return RLM_MODULE_FAIL; at line 236.
or get the fixes from CVS head which solve many other problems
in rlm_detail. (for example file renamed or unlinked by another
program)
Everyone will save a considerable amount of time if you could backport
the changes from CVS to 1.1.2, run tests to check the new code, and
send unified diff to the list.
http://freeradius.org/radiusd/doc/DIFFS
i try to examine snapshot-20060616
btw, there are old bug in rlm_detail.c.
279:if ((outfd = open(buffer, O_WRONLY | O_APPEND | O_CREAT,
280: inst-detailperm)) 0) {
281:radlog(L_ERR, rlm_detail: Couldn't open file %s: %s,
282: buffer, strerror(errno));
283:return RLM_MODULE_FAIL;
284:}
286:/*
287: * If we fail to aquire the filelock in 80 tries
288: * (approximately two seconds) we bail out.
289: */
290:if (inst-locking) {
291:lseek(outfd, 0L, SEEK_SET);
292:if (rad_lockfd_nonblock(outfd, 0) 0) {
293:close(outfd);
294:tv.tv_sec = 0;
295:tv.tv_usec = 25000;
296:select(0, NULL, NULL, NULL, tv);
297:lock_count++;
298:continue;
299:}
open() executed once. but while we failed to get lock at first try,
close(outfd) executed and all the subsequent attempts to lock file
fails
i comment line 293 out.
After this i havn't got any messages like
Error: rlm_detail: Failed to aquire filelock for
/var/log/radius/radacct/detail, giving up
in log!
so now rlm_detail with lock=yes and radrelay all works fine.
Mike.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re[2]: rlm_detail bug when locking = yes
open() executed once. but while we failed to get lock at first try,
close(outfd) executed and all the subsequent attempts to lock file
fails
i comment line 293 out.
I don't understand why this helps. We open the file at the beginning
of the do..while() loop, so it should be fine.
err. sorry.
source of snapshot are like to freeradius-1.1.2
(rlm_detail.c,v1.37.2.1.2.1 2006/04/19). i miss that open() was moved
inside loop :(
BTW if you remove the close() line 293 you're leaking one file descriptor
each time the lock fails.
i use stable 1.1.2, not snapshot.
there open() called before do-while loop in this version.
close() executed every iteration.
so i have to comment close() in block
if (rad_lockfd_nonblock(outfd, 0) 0) {
close(outfd);
tv.tv_sec = 0;
tv.tv_usec = 25000;
select(0, NULL, NULL, NULL, tv);
lock_count++;
}
for correct work.
sorry again.
Mike.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
rlm_detail bug when locking = yes
hello, everybody.
my radius use a rlm_perl module, wich open and write some
informations to files. also it use rlm_detail but without locking.
2 days ago i add second rlm_detail to config with option
locking = yes.
since i got 2 files with corrupted datas which accessed by rlm_perl.
i think bug is in rlm_detail.c. there is no close() after successful
open and failed filelock!
198:/*
199: * Open create the file, with the given permissions.
200: */
201:if ((outfd = open(buffer, O_WRONLY | O_APPEND | O_CREAT,
202: inst-detailperm)) 0) {
203:radlog(L_ERR, rlm_detail: Couldn't open file %s: %s,
204: buffer, strerror(errno));
205:return RLM_MODULE_FAIL;
206:}
...
233:if (!locked inst-locking lock_count = 80) {
234:radlog(L_ERR, rlm_detail: Failed to aquire filelock for %s,
givingup,
235: buffer);
236:return RLM_MODULE_FAIL;
237:}
we need close outfd right before return RLM_MODULE_FAIL; at line
236.
Mike.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
rlm_detail and radrelay concurrence
hello, everyone
i use radrelay
there are errors in log from rlm_detail like
Error: rlm_detail: Couldn't open file /var/log/radius/radacct/detail-relay: Bad
file descriptor
while examine rlm_detail.c i found two places in it.
first while open (create if need) detail-file (line 204 rlm_detail.c).
second while do converting the FD to FP (line 243) after aquiring
filelock of detail file.
Bad file description error appear because radrelay
can remove detail file while rad_detail trying to aquire filelock, I
think (line 655 of radrelay.c).
rlm_detail.c:
201:if ((outfd = open(buffer, O_WRONLY | O_APPEND | O_CREAT,
202: inst-detailperm)) 0) {
203:radlog(L_ERR, rlm_detail: Couldn't open file %s: %s,
204: buffer, strerror(errno));
205:return RLM_MODULE_FAIL;
206:}
208:/*
209: * If we're not using locking, we'll just pass straight though
210: * the while loop.
211: * If we fail to aquire the filelock in 80 tries (approximately
212: * two seconds) we bail out.
213: */
214:locked = 0;
215:lock_count = 0;
216:do {
231:} while (!locked inst-locking lock_count 80);
...
239:/*
240: * Convert the FD to FP. The FD is no longer valid
241: * after this operation.
242: */
243:if ((outfp = fdopen(outfd, a)) == NULL) {
244:radlog(L_ERR, rlm_detail: Couldn't open file %s: %s,
245: buffer, strerror(errno));
246:if (inst-locking) {
247:lseek(outfd, 0L, SEEK_SET);
248:rad_unlockfd(outfd, 0);
249:DEBUG(rlm_detail: Released filelock);
250:}
251:close(outfd);
253:return RLM_MODULE_FAIL;
254:}
radrelay.c:
655:if (detail_move(r_args-detail, work) == 0) {
656:if (debug_flag 0)
657:fprintf(stderr, Moving %s to %s\n,
658:r_args-detail, work);
659:/*
660: * rlm_detail might still write
661: * something to detail.work if
662: * it opens detail before it is
663: * renamed (race condition)
664: */
665:ms_sleep(1000);
666:state = STATE_BACKLOG;
667:}
668:fpos = ftell(fp);
669:fseek(fp, 0L, SEEK_SET);
670:rad_unlockfd(fileno(fp), 0);
671:fseek(fp, fpos, SEEK_SET);
so detail_move can be called by radrelay while rad_detail already
opens detail (line 201) have outfd and try to aquire lock. after radrelay made
rad_unlockfd(fileno(fp), 0) (line 670), rlm_detail try to fdopen (line
243) and will get error.
As workaround we can give to rlm_detail recreate detail...
--- rlm_detail.c-orig 2006-06-15 12:09:43.0 +0600
+++ rlm_detail.c2006-06-15 12:22:49.0 +0600
@@ -119,6 +119,7 @@
struct stat st;
int locked;
int lock_count;
+ int opencount=0;
struct timeval tv;
REALM *proxy_realm;
charproxy_buffer[16];
@@ -195,6 +196,8 @@
*p = '/';
} /* else there was no directory delimiter. */
+
+reopen:
/*
* Open create the file, with the given permissions.
*/
@@ -241,8 +244,9 @@
* after this operation.
*/
if ((outfp = fdopen(outfd, a)) == NULL) {
- radlog(L_ERR, rlm_detail: Couldn't open file %s: %s,
- buffer, strerror(errno));
+ radlog(L_ERR, rlm_detail: Couldn't open file %s: %s. %s,
+ buffer, strerror(errno),
+ (opencount==0)?try reopen to create:return
MODULE_FAIL);
if (inst-locking) {
lseek(outfd, 0L, SEEK_SET);
rad_unlockfd(outfd, 0);
@@ -250,7 +254,10 @@
}
close(outfd);
- return RLM_MODULE_FAIL;
+ if (opencount++ == 0)
+ goto reopen;
+ else
+ return RLM_MODULE_FAIL;
}
/*
btw, why to sleep 1000ms on [EMAIL PROTECTED]
how rlm_detail can write something to detail.work, while file is locked
by radreplay?
Mike.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
