Re: [3.1.6] Deadlock upon recursive call_usermodehelper_exec().

[Tetsuo Handa]

Additional info. I added debug printk()

--- a/lib/kobject_uevent.c
+++ b/lib/kobject_uevent.c
@@ -305,8 +305,13 @@ int kobject_uevent_env(struct kobject *k
 		if (retval)
 			goto exit;
 
+		printk(KERN_INFO "*** Calling call_usermodehelper(\"%s\")\n",
+		       argv[0]);
 		retval = call_usermodehelper(argv[0], argv,
 					     env->envp, UMH_WAIT_EXEC);
+		printk(KERN_INFO
+		       "*** Returned from call_usermodehelper(\"%s\")\n",
+		       argv[0]);
 	}
 
 exit:
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -1435,7 +1435,14 @@ int search_binary_handler(struct linux_b
 				break; /* -ENOEXEC */
 			if (try)
 				break; /* -ENOEXEC */
+			printk(KERN_INFO
+			       "*** Calling request_module(\"binfmt-%04x\")\n",
+			       *(unsigned short *)(&bprm->buf[2]));
 			request_module("binfmt-%04x", *(unsigned short *)(&bprm->buf[2]));
+			printk(KERN_INFO
+			       "*** Returned from "
+			       "request_module(\"binfmt-%04x\")\n",
+			       *(unsigned short *)(&bprm->buf[2]));
 		}
 #else
 		break;

and got below output just before deadlock.

*** Calling call_usermodehelper("/sbin/hotplug")
*** Calling request_module("binfmt-0000")

So, this deadlock occurs when call_usermodehelper(UMH_WAIT_EXEC) is called and
request_module() is also called from search_binary_handler() when processing
call_usermodehelper(UMH_WAIT_EXEC).

This would not hanppen for normal condition, but 100% happens when the
userspace helper program is damaged.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/