Re: [patch V5 01/32] mm/slab: Remove broken stack trace storage

[Vlastimil Babka <vbabka@suse.cz>]

On 4/16/19 4:10 PM, Thomas Gleixner wrote:
> kstack_end() is broken on interrupt stacks as they are not guaranteed to be
> sized THREAD_SIZE and THREAD_SIZE aligned.
> 
> As SLAB seems not to be used much with debugging enabled and might just go
> away completely according to:
> 
>   https://lkml.kernel.org/r/612f9b99-a75b-6aeb-cf92-7dc5421cd950@suse.cz
> 
> just remove the bogus code instead of trying to fix it.
> 
> Fixes: 98eb235b7feb ("[PATCH] page unmapping debug") - History tree
> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Pekka Enberg <penberg@kernel.org>
> Cc: linux-mm@kvack.org

Acked-by: Vlastimil Babka <vbabka@suse.cz>

Thanks.

> ---
> V5: Remove the cruft.
> V4: Make it actually work
> V2: Made the code simpler to understand (Andy)
> ---
>  mm/slab.c |   22 +++-------------------
>  1 file changed, 3 insertions(+), 19 deletions(-)
> 
> --- a/mm/slab.c
> +++ b/mm/slab.c
> @@ -1470,33 +1470,17 @@ static bool is_debug_pagealloc_cache(str
>  static void store_stackinfo(struct kmem_cache *cachep, unsigned long *addr,
>  			    unsigned long caller)
>  {
> -	int size = cachep->object_size;
> +	int size = cachep->object_size / sizeof(unsigned long);
>  
>  	addr = (unsigned long *)&((char *)addr)[obj_offset(cachep)];
>  
> -	if (size < 5 * sizeof(unsigned long))
> +	if (size < 4)
>  		return;
>  
>  	*addr++ = 0x12345678;
>  	*addr++ = caller;
>  	*addr++ = smp_processor_id();
> -	size -= 3 * sizeof(unsigned long);
> -	{
> -		unsigned long *sptr = &caller;
> -		unsigned long svalue;
> -
> -		while (!kstack_end(sptr)) {
> -			svalue = *sptr++;
> -			if (kernel_text_address(svalue)) {
> -				*addr++ = svalue;
> -				size -= sizeof(unsigned long);
> -				if (size <= sizeof(unsigned long))
> -					break;
> -			}
> -		}
> -
> -	}
> -	*addr++ = 0x87654321;
> +	*addr = 0x87654321;
>  }
>  
>  static void slab_kernel_map(struct kmem_cache *cachep, void *objp,
>