Re: [PATCH 5/6] BC: kernel memory accounting (core)

[Kirill Korotaev <dev@sw.ru>]

Oleg Nesterov wrote:
> On 08/23, Kirill Korotaev wrote:
> 
>>+int bc_slab_charge(kmem_cache_t *cachep, void *objp, gfp_t flags)
>>+{
>>+	unsigned int size;
>>+	struct beancounter *bc, **slab_bcp;
>>+
>>+	bc = get_exec_bc();
>>+	if (bc == NULL)
>>+		return 0;
> 
> 
> Is it possible to .exec_bc == NULL ?
> 
> If yes, why do we need init_bc? We can do 'set_exec_bc(NULL)' in __do_IRQ()
> instead.
no, exec_bc can't be NULL. thanks for catching old hunks which historically exist
due to old times when host system was not accounted (bc was NULL).

Thanks,
Kirill