Re: [PATCH] module: return error when mod_sysfs_init() failed

["Akinobu Mita" <akinobu.mita@gmail.com>]

2007/9/29, Greg KH <gregkh@suse.de>:
> > Index: 2.6-git/kernel/module.c
> > ===================================================================
> > --- 2.6-git.orig/kernel/module.c
> > +++ 2.6-git/kernel/module.c
> > @@ -1782,7 +1782,8 @@ static struct module *load_module(void _
> >       module_unload_init(mod);
> >
> >       /* Initialize kobject, so we can reference it. */
> > -     if (mod_sysfs_init(mod) != 0)
> > +     err = mod_sysfs_init(mod);
> > +     if (err)
> >               goto cleanup;
>
> I must be still asleep this morning, but I think this patch does the
> exact same thing as the original code does, right?  Otherwise, this
> code would always be failing.
>
> Or do I just need to go get my morning coffee to wake up and see the
> problem here?

Hello,

In the original code, the "err" is zero before goto cleanup.
This "err" will be the return value of load_module().

load_module() is the function which returns error as pointer
and the expression IS_ERR(NULL) is false. So the caller of load_module()
cannot catch that error.

I found this problem when I was running the fault injection test
script in Documentation/fault-injection/fault-injection.txt with
random module.

#!/bin/bash

FAILTYPE=failslab
echo Y > /debug/$FAILTYPE/task-filter
echo 10 > /debug/$FAILTYPE/probability
echo 100 > /debug/$FAILTYPE/interval
echo -1 > /debug/$FAILTYPE/times
echo 0 > /debug/$FAILTYPE/space
echo 2 > /debug/$FAILTYPE/verbose
echo 0 > /debug/$FAILTYPE/ignore-gfp-wait

faulty_system()
{
        bash -c "echo 1 > /proc/self/make-it-fail && exec $*"
}

if [ $# -eq 0 ]
then
        echo "Usage: $0 modulename [ modulename ... ]"
        exit 1
fi

for m in $*
do
        echo inserting $m...
        faulty_system modprobe $m

        echo removing $m...
        faulty_system modprobe -r $m
done