Re: [PATCH] firmware: dmi-sysfs: Fix null-ptr-deref in dmi_sysfs_register_handle

[Chen Zhongjin]

Hi,

On 2022/11/11 1:49, Greg KH wrote:
> On Mon, Oct 17, 2022 at 05:53:42PM +0800, Chen Zhongjin wrote:
> > KASAN reported a null-ptr-deref error:
> >
> > KASAN: null-ptr-deref in range [0x0000000000000008-0x000000000000000f]
> > CPU: 0 PID: 1373 Comm: modprobe
> > Hardware name: QEMU Standard PC (i440FX + PIIX, 1996)
> > RIP: 0010:dmi_sysfs_entry_release
> > ...
> > Call Trace:
> >   <TASK>
> >   kobject_put
> >   dmi_sysfs_register_handle (drivers/firmware/dmi-sysfs.c:540) dmi_sysfs
> >   dmi_decode_table (drivers/firmware/dmi_scan.c:133)
> >   dmi_walk (drivers/firmware/dmi_scan.c:1115)
> >   dmi_sysfs_init (drivers/firmware/dmi-sysfs.c:149) dmi_sysfs
> >   do_one_initcall (init/main.c:1296)
> >   ...
> > Kernel panic - not syncing: Fatal exception
> > Kernel Offset: 0x4000000 from 0xffffffff81000000
> > ---[ end Kernel panic - not syncing: Fatal exception ]---
> >
> > It is because previous patch added kobject_put() to release the memory
> > which will call  dmi_sysfs_entry_release() and list_del().
> >
> > However, list_add_tail(entry->list) is called after the error block,
> > so the list_head is uninitialized and cannot be deleted.
> >
> > Because entry is allocated by kzalloc() so the list.prev is NULL in
> > the error path. Check it in dmi_sysfs_entry_release() to avoid
> > deleting uninitialized list_head.
> >
> > Fixes: 660ba678f999 ("firmware: dmi-sysfs: Fix memory leak in dmi_sysfs_register_handle")
> >
> > Signed-off-by: Chen Zhongjin <chenzhongjin@xxxxxxxxxx>
> > ---
> >   drivers/firmware/dmi-sysfs.c | 9 ++++++---
> >   1 file changed, 6 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/firmware/dmi-sysfs.c b/drivers/firmware/dmi-sysfs.c
> > index 66727ad3361b..f8815eeed00c 100644
> > --- a/drivers/firmware/dmi-sysfs.c
> > +++ b/drivers/firmware/dmi-sysfs.c
> > @@ -557,9 +557,12 @@ static void dmi_sysfs_entry_release(struct kobject *kobj)
> >   {
> >   	struct dmi_sysfs_entry *entry = to_entry(kobj);
> >   
> > -	spin_lock(&entry_list_lock);
> > -	list_del(&entry->list);
> > -	spin_unlock(&entry_list_lock);
> > +	if (entry->list.prev != NULL) {
> You should not be poking around in a lock structure like this at all.
> Also the lock isn't held, so how do you know this is going to work?
>
> I suggest fixing up the original patch, perhaps reverting that instead?

It makes sense, but in fact the original patch did fix the memory leak.

I'm going to move kobject_put behind list_add_tail to keep consistency, 
because only after that the
entry initialization is complete.

Will send v2, thanks for your review!


Best,

Chen

> thanks,
>
> greg k-h