Re: problem with starting 2.6.25-rc1 and latest git

[Jean Delvare]

Hi Mariusz,

On Tue, 26 Feb 2008 17:04:46 +0100, Mariusz Kozlowski wrote:
> I can reproduce this here on rc3 as well. Weird - on git tree before
> rc3 was released this didn't happen. rc2 and rc1 are broken. I also
> don't know why nobody else see this but us :) And not sure how
> to debug it some more or provide more accurate information.

Here's a quick and dirty debugging patch I wrote in an attempt to find
out what was going on. Unfortunately, the problem never happened to me
when this debug patch was applied. Maybe you'll have more "luck"?

 kernel/module.c |    4 ++++
 1 file changed, 4 insertions(+)

--- linux-2.6.25-rc1.orig/kernel/module.c	2008-02-14 09:16:15.000000000 +0100
+++ linux-2.6.25-rc1/kernel/module.c	2008-02-14 23:20:22.000000000 +0100
@@ -2153,8 +2153,10 @@ sys_init_module(void __user *umod,
 		return PTR_ERR(mod);
 	}
 
+	printk(KERN_WARNING "loading module %s, checkpoint 1\n", mod->name);
 	/* Drop lock so they can recurse */
 	mutex_unlock(&module_mutex);
+	printk(KERN_WARNING "loading module %s, checkpoint 2\n", mod->name);
 
 	blocking_notifier_call_chain(&module_notify_list,
 			MODULE_STATE_COMING, mod);
@@ -2177,6 +2179,7 @@ sys_init_module(void __user *umod,
 
 	/* Now it's a first class citizen! */
 	mutex_lock(&module_mutex);
+	printk(KERN_WARNING "loading module %s, checkpoint 3\n", mod->name);
 	mod->state = MODULE_STATE_LIVE;
 	/* Drop initial reference. */
 	module_put(mod);
@@ -2186,6 +2189,7 @@ sys_init_module(void __user *umod,
 	mod->init_size = 0;
 	mod->init_text_size = 0;
 	mutex_unlock(&module_mutex);
+	printk(KERN_WARNING "loading module %s, checkpoint 4\n", mod->name);
 	wake_up(&module_wq);
 
 	return 0;


-- 
Jean Delvare
--
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/