Re: [PATCH] Forcing subclasses to have same name pointer as their parent class

Re: [PATCH] Forcing subclasses to have same name pointer as their parent class

[Boqun Feng]

Hi,

On Thu, Jul 04, 2024 at 03:32:24AM +0300, botta633 wrote:
> Preventing Lockdep_set_subclass from creating a new instance of the 
> string literal. Hence, we will always have the class->name. This
> prevents kernel panics when locking up a lock class while comparing
> class locks and class names.

Good catch! Thanks.

> 
> 

Please remove the extra blank line here.

> Signed-off-by: botta633 <bottaawesome633@gmail.com>

Do you mind putting your real name here? Besides, IIUC, this is fixing:

	https://syzkaller.appspot.com/bug?extid=7f4a6f7f7051474e40ad		

, right? If so, there are some more things:

*	Copy ext4 and syzkaller people, so that you could get more
	tests.

*	Since this is a bug fix, could you please figure out which
	commit introduces the issue, so that you can put a correct
	"Fixes:" tag along with your signed-off-by?

*	Since the issue was reported by syzkaller, you should put their
	"Reported-by" tag, they have an example in the website I paste
	above.

*	Please also Cc stable mail list so that the fix can be
	backported, you can find the information on "Cc: stable" tag at:

		https://docs.kernel.org/process/stable-kernel-rules.html

*	Last but not the least, could you try to add a test case in
	lib/locking-selftest.c to ensure the issue you fixed won't
	happen again? This could be tricky, since you will need to fight
	against the compiler to generate two string literals with the
	same content.

[Cc ext4 and syzkaller]

Regards,
Boqun

> ---
>  include/linux/lockdep.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h
> index 08b0d1d9d78b..df8fa5929de7 100644
> --- a/include/linux/lockdep.h
> +++ b/include/linux/lockdep.h
> @@ -173,7 +173,7 @@ static inline void lockdep_init_map(struct lockdep_map *lock, const char *name,
>  			      (lock)->dep_map.lock_type)
>  
>  #define lockdep_set_subclass(lock, sub)					\
> -	lockdep_init_map_type(&(lock)->dep_map, #lock, (lock)->dep_map.key, sub,\
> +	lockdep_init_map_type(&(lock)->dep_map, (lock)->dep_map.name, (lock)->dep_map.key, sub,\
>  			      (lock)->dep_map.wait_type_inner,		\
>  			      (lock)->dep_map.wait_type_outer,		\
>  			      (lock)->dep_map.lock_type)
> -- 
> 2.45.2
> 

[PATCH] Forcing subclasses to have same name pointer as their parent class

[botta633]

Preventing Lockdep_set_subclass from creating a new instance of the 
string literal. Hence, we will always have the class->name. This
prevents kernel panics when locking up a lock class while comparing
class locks and class names.


Signed-off-by: botta633 <bottaawesome633@gmail.com>
---
 include/linux/lockdep.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h
index 08b0d1d9d78b..df8fa5929de7 100644
--- a/include/linux/lockdep.h
+++ b/include/linux/lockdep.h
@@ -173,7 +173,7 @@ static inline void lockdep_init_map(struct lockdep_map *lock, const char *name,
 			      (lock)->dep_map.lock_type)
 
 #define lockdep_set_subclass(lock, sub)					\
-	lockdep_init_map_type(&(lock)->dep_map, #lock, (lock)->dep_map.key, sub,\
+	lockdep_init_map_type(&(lock)->dep_map, (lock)->dep_map.name, (lock)->dep_map.key, sub,\
 			      (lock)->dep_map.wait_type_inner,		\
 			      (lock)->dep_map.wait_type_outer,		\
 			      (lock)->dep_map.lock_type)
-- 
2.45.2

Re: [PATCH] Forcing subclasses to have same name pointer as their parent class

[Boqun Feng]

On Tue, Jul 09, 2024 at 02:26:05AM +0300, ahmed Ehab wrote:
> Hi,
>  thanks for the great feedback.
> 
> 
> 
> 
> * Last but not the least, could you try to add a test case in
> lib/locking-selftest.c to ensure the issue you fixed won't        happen
> again? This could be tricky, since you will need to fight        against
> the compiler to generate two string literals with the        same content.*
> I added a test case, but I am not sure how to run this test file as it
> doesn't seem to be part of the kselftests. I compiled it successfully by
> setting the lock debugging option but couldn't get it to run.
> 

My usual approach is: set CONFIG_DEBUG_LOCKING_API_SELFTESTS=y, compile
a kernel, and boot it in qemu, the test result will show in serial log
of the VM. Hope that helps.

Regards,
Boqun

> Regards,
> Ahmed
> 
> On Thu, Jul 4, 2024 at 1:46 AM Boqun Feng <boqun.feng@gmail.com> wrote:
> 
> > Hi,
> >
> > On Thu, Jul 04, 2024 at 03:32:24AM +0300, botta633 wrote:
> > > Preventing Lockdep_set_subclass from creating a new instance of the
> > > string literal. Hence, we will always have the class->name. This
> > > prevents kernel panics when locking up a lock class while comparing
> > > class locks and class names.
> >
> > Good catch! Thanks.
> >
> > >
> > >
> >
> > Please remove the extra blank line here.
> >
> > > Signed-off-by: botta633 <bottaawesome633@gmail.com>
> >
> > Do you mind putting your real name here? Besides, IIUC, this is fixing:
> >
> >         https://syzkaller.appspot.com/bug?extid=7f4a6f7f7051474e40ad
> >
> >
> > , right? If so, there are some more things:
> >
> > *       Copy ext4 and syzkaller people, so that you could get more
> >         tests.
> >
> > *       Since this is a bug fix, could you please figure out which
> >         commit introduces the issue, so that you can put a correct
> >         "Fixes:" tag along with your signed-off-by?
> >
> > *       Since the issue was reported by syzkaller, you should put their
> >         "Reported-by" tag, they have an example in the website I paste
> >         above.
> >
> > *       Please also Cc stable mail list so that the fix can be
> >         backported, you can find the information on "Cc: stable" tag at:
> >
> >                 https://docs.kernel.org/process/stable-kernel-rules.html
> >
> > *       Last but not the least, could you try to add a test case in
> >         lib/locking-selftest.c to ensure the issue you fixed won't
> >         happen again? This could be tricky, since you will need to fight
> >         against the compiler to generate two string literals with the
> >         same content.
> >
> > [Cc ext4 and syzkaller]
> >
> > Regards,
> > Boqun
> >
> > > ---
> > >  include/linux/lockdep.h | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h
> > > index 08b0d1d9d78b..df8fa5929de7 100644
> > > --- a/include/linux/lockdep.h
> > > +++ b/include/linux/lockdep.h
> > > @@ -173,7 +173,7 @@ static inline void lockdep_init_map(struct
> > lockdep_map *lock, const char *name,
> > >                             (lock)->dep_map.lock_type)
> > >
> > >  #define lockdep_set_subclass(lock, sub)
> >       \
> > > -     lockdep_init_map_type(&(lock)->dep_map, #lock,
> > (lock)->dep_map.key, sub,\
> > > +     lockdep_init_map_type(&(lock)->dep_map, (lock)->dep_map.name,
> > (lock)->dep_map.key, sub,\
> > >                             (lock)->dep_map.wait_type_inner,          \
> > >                             (lock)->dep_map.wait_type_outer,          \
> > >                             (lock)->dep_map.lock_type)
> > > --
> > > 2.45.2
> > >
> >