Re: [mm/slub] ae107fa919: BUG:unable_to_handle_page_fault_for_address

[Vlastimil Babka]

On 3/12/22 14:08, Vlastimil Babka wrote:
> On 3/12/22 13:06, Mike Rapoport wrote:
>>>>> Ideas? implementing something like kmem_cache_init_early() again?
>>>>
>>>> I think we could simply make CONFIG_SLUB_DEBUG_ON select/depend on
>>>> STACKDEPOT_ALWAYS_INIT?
>>>
>>> Oh, sounds better.
>>>
>>> If we make CONFIG_SLUB_DEBUG_ON select STACK_DEPOT_ALWAYS_INIT,
>>> that is simple solution. but stackdepot will be initialized on
>>> slub_debug=- too.
> 
> Valid point.
> 
>>> But I think no one will set CONFIG_SLUB_DEBUG_ON=y if not debugging...
>> 
>> If memory wasted by stack_table is a real concern, we may free it after
>> parsing slub_debug or add a condition taking into account
>> CONFIG_SLUB_DEBUG_ON and slub_debug=- to the 
>> 
>> 	if (slub_debug & SLAB_STORE_USER)
>> 		stack_depot_want_early_init();
>> 
>> But I agree that if somebody runs a kernel with CONFIG_SLUB_DEBUG_ON=y, the
>> goal is to have slub debugging on, so making CONFIG_SLUB_DEBUG_ON select
>> STACK_DEPOT_ALWAYS_INIT totally makes sense to me.
> 
> Agree. We had to decouple stackdpepot init from CONFIG_SLUB_DEBUG
> because almost everyone enables CONFIG_SLUB_DEBUG by default, but on
> smaller systems doesn't want to always pay the price even if unused.
> But I don't expect anyone to enable CONFIG_SLUB_DEBUG_ON and then run
> with slub_debug=- by default, and complain. If anyone like that exists,
> they could also add the stackdepot_disable parameter.

Pushed fixed version to
https://git.kernel.org/cgit/linux/kernel/git/vbabka/linux.git
slub-stackdepot-v3r1 for now
As it's too late to target 5.18 now, will post v3 properly after merge
window and add to slab-next towards 5.19 then.

>>> I don't think making CONFIG_SLUB_DEBUG_ON depend on
>>> CONFIG_STACKDEPOT_ALWAYS_INIT is good solution. only KASAN selects it.
>>>
>>> -- 
>>> Thank you, You are awesome!
>>> Hyeonggon :-)
>> 
>