Re: [RFC V2 3/3] s390/mm: Define arch_get_mappable_range()

[Anshuman Khandual <anshuman.khandual@arm.com>]

On 12/7/20 2:33 PM, David Hildenbrand wrote:
> On 07.12.20 05:38, Anshuman Khandual wrote:
>>
>>
>> On 12/3/20 5:31 PM, David Hildenbrand wrote:
>>> On 03.12.20 12:51, Heiko Carstens wrote:
>>>> On Thu, Dec 03, 2020 at 06:03:00AM +0530, Anshuman Khandual wrote:
>>>>>>> diff --git a/arch/s390/mm/extmem.c b/arch/s390/mm/extmem.c
>>>>>>> index 5060956b8e7d..cc055a78f7b6 100644
>>>>>>> --- a/arch/s390/mm/extmem.c
>>>>>>> +++ b/arch/s390/mm/extmem.c
>>>>>>> @@ -337,6 +337,11 @@ __segment_load (char *name, int do_nonshared, unsigned long *addr, unsigned long
>>>>>>>  		goto out_free_resource;
>>>>>>>  	}
>>>>>>>  
>>>>>>> +	if (seg->end + 1 > VMEM_MAX_PHYS || seg->end + 1 < seg->start_addr) {
>>>>>>> +		rc = -ERANGE;
>>>>>>> +		goto out_resource;
>>>>>>> +	}
>>>>>>> +
>>>>>>>  	rc = vmem_add_mapping(seg->start_addr, seg->end - seg->start_addr + 1);
>>>>>>>  	if (rc)
>>>>>>>  		goto out_resource;
>>>>>>> diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c
>>>>>>> index b239f2ba93b0..06dddcc0ce06 100644
>>>>>>> --- a/arch/s390/mm/vmem.c
>>>>>>> +++ b/arch/s390/mm/vmem.c
>>>>>>> @@ -532,14 +532,19 @@ void vmem_remove_mapping(unsigned long start, unsigned long size)
>>>>>>>  	mutex_unlock(&vmem_mutex);
>>>>>>>  }
>>>>>>>  
>>>>>>> +struct range arch_get_mappable_range(void)
>>>>>>> +{
>>>>>>> +	struct range memhp_range;
>>>>>>> +
>>>>>>> +	memhp_range.start = 0;
>>>>>>> +	memhp_range.end =  VMEM_MAX_PHYS;
>>>>>>> +	return memhp_range;
>>>>>>> +}
>>>>>>> +
>>>>>>>  int vmem_add_mapping(unsigned long start, unsigned long size)
>>>>>>>  {
>>>>>>>  	int ret;
>>>>>>>  
>>>>>>> -	if (start + size > VMEM_MAX_PHYS ||
>>>>>>> -	    start + size < start)
>>>>>>> -		return -ERANGE;
>>>>>>> -
>>>>>>
>>>>>> I really fail to see how this could be considered an improvement for
>>>>>> s390. Especially I do not like that the (central) range check is now
>>>>>> moved to the caller (__segment_load). Which would mean potential
>>>>>> additional future callers would have to duplicate that code as well.
>>>>>
>>>>> The physical range check is being moved to the generic hotplug code
>>>>> via arch_get_mappable_range() instead, making the existing check in
>>>>> vmem_add_mapping() redundant. Dropping the check there necessitates
>>>>> adding back a similar check in __segment_load(). Otherwise there
>>>>> will be a loss of functionality in terms of range check.
>>>>>
>>>>> May be we could just keep this existing check in vmem_add_mapping()
>>>>> as well in order avoid this movement but then it would be redundant
>>>>> check in every hotplug path.
>>>>>
>>>>> So I guess the choice is to either have redundant range checks in
>>>>> all hotplug paths or future internal callers of vmem_add_mapping()
>>>>> take care of the range check.
>>>>
>>>> The problem I have with this current approach from an architecture
>>>> perspective: we end up having two completely different methods which
>>>> are doing the same and must be kept in sync. This might be obvious
>>>> looking at this patch, but I'm sure this will go out-of-sync (aka
>>>> broken) sooner or later.
>>>
>>> Exactly, there should be one function only that was the whole idea of
>>> arch_get_mappable_range().
>>>
>>>>
>>>> Therefore I would really like to see a single method to do the range
>>>> checking. Maybe you could add a callback into architecture code, so
>>>> that such an architecture specific function could also be used
>>>> elsewhere. Dunno.
>>>>
>>>
>>> I think we can just switch to using "memhp_range_allowed()" here then
>>> after implementing arch_get_mappable_range().
>>>
>>> Doesn't hurt to double check in vmem_add_mapping() - especially to keep
>>> extmem working without changes. At least for callers of memory hotplug
>>> it's then clear which values actually won't fail deep down in arch code.
>>
>> But there is a small problem here. memhp_range_allowed() is now defined
>> and available with CONFIG_MEMORY_HOTPLUG where as vmem_add_mapping() and
>> __segment_load() are generally available without any config dependency.
>> So if CONFIG_MEMORY_HOTPLUG is not enabled there will be a build failure
>> in vmem_add_mapping() for memhp_range_allowed() symbol.
>>
>> We could just move VM_BUG_ON(!memhp_range_allowed(start, size, 1)) check
>> from vmem_add_mapping() to arch_add_memory() like on arm64 platform. But
>> then __segment_load() would need that additional new check to compensate
>> as proposed earlier.
>>
>> Also leaving vmem_add_mapping() and __segment_load() unchanged will cause
>> the address range check to be called three times on the hotplug path i.e
>>
>> 1. register_memory_resource()
>> 2. arch_add_memory()
>> 3. vmem_add_mapping()
>>
>> Moving memhp_range_allowed() check inside arch_add_memory() seems better
>> and consistent with arm64. Also in the future, any platform which choose
>> to override arch_get_mappable() will have this additional VM_BUG_ON() in
>> their arch_add_memory().
> 
> Yeah, it might not make sense to add these checks all over the place.
> The important part is that
> 
> 1. There is a check somewhere (and if it's deep down in arch code)
> 2. There is an obvious way for callers to find out what valid values are.
> 
> 
> I guess it would be good enough to
> 
> a) Factor out getting arch ranges into arch_get_mappable_range()
> b) Provide memhp_get_pluggable_range()

Have posted V1 earlier in the day which hopefully accommodates all previous
suggestions but otherwise do let me know if anything else still needs to be
improved upon.

https://lore.kernel.org/linux-mm/1607400978-31595-1-git-send-email-anshuman.khandual@arm.com/

> 
> Both changes only make sense with an in-tree user. I'm planning on using
> this functionality in virtio-mem code. I can pickup your patches, drop
> the superfluous checks, and use it from virtio-mem code. Makese sense
> (BTW, looks like we'll see aarch64 support for virtio-mem soon)?

I have not been following virtio-mem closely. But is there something pending
on arm64 platform which prevents virtio-mem enablement ?