2.6.7 and khelper

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

First of I have no idea what khelper actually does, but it seems to make a problem on my box.

I have a Debian/unstable box here (the same one that has these "fast clock problems with 2.6.7-mm1) and every night after the syslog restart the process with the id "4", which is khelper is reported to be respawning to fast.

Now is this a kernel issue, or more a issue of my system and some jobs that might disrupt this?

- --
Clemens Schwaighofer - IT Engineer & System Administration
==========================================================
TEQUILA\Japan, 6-17-2 Ginza Chuo-ku, Tokyo 104-8167, JAPAN
Tel: +81-(0)3-3545-7703 Fax: +81-(0)3-3545-7343
http://www.tequila.co.jp
==========================================================
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA23bwjBz/yQjBxz8RAlICAKDJDZWc7kFf3Kk907cUYdpUzRtoGACgpiJm vVs8NTMKABMkufRM6NwLifM=
=yG2m
-----END PGP SIGNATURE-----

It is a kernel thread which is used for making calls out to userspace applications from within the kernel.

Strange. I assume that what's happening is that the children of khelper are being created and are dying, and init is somehow seeing this happen. Maybe SIGCHLD, probably via wait4(). Perhaps init should be changed to not complain about processes which it did't parent. But then, that should already be the case.

Could you please apply the below debug patch, then send us all the relevant syslog output, including the messages from init?

Thanks.

diff -puN kernel/kmod.c~khelper-child-sequence kernel/kmod.c
@@ -155,6 +155,9 @@ static int ____call_usermodehelper(void 
 	struct subprocess_info *sub_info = data;
 	int retval;
 	cpumask_t mask = CPU_MASK_ALL;
+	static int call_umh_id;
+
+	sprintf(current->comm, "call_umh%d\n", call_umh_id++);
 
 	/* Unblock all signals. */
 	flush_signals(current);

@@ -182,6 +185,9 @@ static int wait_for_helper(void *data)
 	struct subprocess_info *sub_info = data;
 	pid_t pid;
 	struct k_sigaction sa;
+	static int khelper_id;
+
+	sprintf(current->comm, "waiter%d", khelper_id++);
 
 	/* Install a handler: if SIGCLD isn't handled sys_wait4 won't
 	 * populate the status, but will return -ECHILD. */

_

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Andrew Morton wrote:

| Strange.  I assume that what's happening is that the children of khelper
| are being created and are dying, and init is somehow seeing this happen.
| Maybe SIGCHLD, probably via wait4().  Perhaps init should be changed
to not
| complain about processes which it did't parent.  But then, that should
| already be the case.
|
|
| Could you please apply the below debug patch, then send us all the

relevant
| syslog output, including the messages from init?

I have applied the patch against a vanilla 2.6.7, but until now i haven't got any output. But after the next crontab run, I'll see that. I'll push that later.

- --
Clemens Schwaighofer - IT Engineer & System Administration
==========================================================
TEQUILA\Japan, 6-17-2 Ginza Chuo-ku, Tokyo 104-8167, JAPAN
Tel: +81-(0)3-3545-7703 Fax: +81-(0)3-3545-7343
http://www.tequila.co.jp
==========================================================
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA2/iUjBz/yQjBxz8RAit0AJ9pt1gKbEfyvTQRQuAjN15Wsj1MugCgr6I+ wqHSBXyxgsCbDaKVHSVIs1I=
=6wlJ
-----END PGP SIGNATURE-----

Umm, that ID refers to /etc/inittab, not to the process ID.

Matthias Urlichs

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Matthias Urlichs wrote:

| Hi, Andrew Morton wrote:
|
|
|>>I have a Debian/unstable box here (the same one that has these "fast
|>>clock problems with 2.6.7-mm1) and every night after the syslog restart
|>>the process with the id "4", which is khelper is reported to be
|>>respawning to fast.
|>
|>Strange.  I assume that what's happening is that the children of khelper
|>are being created and are dying,
|
|
| Umm, that ID refers to /etc/inittab, not to the process ID.

hmm, whatever it really was, thought it happend 4 times in a row, I couldn't trigger it the whole weekend ...

- --
Clemens Schwaighofer - IT Engineer & System Administration
==========================================================
TEQUILA\Japan, 6-17-2 Ginza Chuo-ku, Tokyo 104-8167, JAPAN
Tel: +81-(0)3-3545-7703 Fax: +81-(0)3-3545-7343
http://www.tequila.co.jp
==========================================================
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA32ZXjBz/yQjBxz8RAoBNAKDS+ZwnDDaPj8+POWzvI6/EULPXngCcDwBd aq4om4MF50YPTyfgDl4NU7E=
=CxLJ
-----END PGP SIGNATURE-----