Re: [PATCH v3] lib: stackdepot: Add support to configure STACK_HASH_SIZE

[Vijayanand Jitta]

On 12/22/2020 1:59 AM, Minchan Kim wrote:
> On Mon, Dec 21, 2020 at 04:04:09PM +0100, Alexander Potapenko wrote:
>> On Mon, Dec 21, 2020 at 12:15 PM Vijayanand Jitta <vjitta@xxxxxxxxxxxxxx> wrote:
>>>
>>>
>>>
>>> On 12/18/2020 2:10 PM, Vijayanand Jitta wrote:
>>>>
>>>>
>>>> On 12/17/2020 4:24 PM, Alexander Potapenko wrote:
>>>>>>> Can you provide an example of a use case in which the user wants to
>>>>>>> use the stack depot of a smaller size without disabling it completely,
>>>>>>> and that size cannot be configured statically?
>>>>>>> As far as I understand, for the page owner example you gave it's
>>>>>>> sufficient to provide a switch that can disable the stack depot if
>>>>>>> page_owner=off.
>>>>>>>
>>>>>> There are two use cases here,
>>>>>>
>>>>>> 1. We don't want to consume memory when page_owner=off ,boolean flag
>>>>>> would work here.
>>>>>>
>>>>>> 2. We would want to enable page_owner on low ram devices but we don't
>>>>>> want stack depot to consume 8 MB of memory, so for this case we would
>>>>>> need a configurable stack_hash_size so that we can still use page_owner
>>>>>> with lower memory consumption.
>>>>>>
>>>>>> So, a configurable stack_hash_size would work for both these use cases,
>>>>>> we can set it to '0' for first case and set the required size for the
>>>>>> second case.
>>>>>
>>>>> Will a combined solution with a boolean boot-time flag and a static
>>>>> CONFIG_STACKDEPOT_HASH_SIZE work for these cases?
>>>>> I suppose low-memory devices have a separate kernel config anyway?
>>>>>
>>>>
>>>> Yes, the combined solution will also work but i think having a single
>>>> run time config is simpler instead of having two things to configure.
>>>>
>>>
>>> To add to it we started of with a CONFIG first, after the comments from
>>> Minchan (https://lkml.org/lkml/2020/11/3/2121) we decided to switch to
>>> run time param.
>>>
>>> Quoting Minchan's comments below:
>>>
>>> "
>>> 1. When we don't use page_owner, we don't want to waste any memory for
>>> stackdepot hash array.
>>> 2. When we use page_owner, we want to have reasonable stackdeport hash array
>>>
>>> With this configuration, it couldn't meet since we always need to
>>> reserve a reasonable size for the array.
>>> Can't we make the hash size as a kernel parameter?
>>> With it, we could use it like this.
>>>
>>> 1. page_owner=off, stackdepot_stack_hash=0 -> no more wasted memory
>>> when we don't use page_owner
>>> 2. page_owner=on, stackdepot_stack_hash=8M -> reasonable hash size
>>> when we use page_owner.
>>> "
>>
>> Minchan, what do you think about making the hash size itself a static
>> parameter, while letting the user disable stackdepot completely at
>> runtime?
>> As noted before, I am concerned that moving a low-level configuration
>> bit (which essentially means "save 8Mb - (1 << stackdepot_stack_hash)
>> of static memory") to the boot parameters will be unused by most
>> admins and may actually trick them into thinking they reduce the
>> overall stackdepot memory consumption noticeably.
>> I also suppose device vendors may prefer setting a fixed (maybe
>> non-default) hash size for low-memory devices rather than letting the
>> admins increase it.
> 
> I am totally fine if we could save the static memory alloation when
> the page_owner is not used.
> 
> IOW, page_owner=disable, stackdepot=disable will not consume the 8M
> memory.
> When we want to use page_owner, we could just do like this
> 
> 	page_owner=enable, stackdepot=enable
> 
> (Maybe we need something to make warning if stackdepot is disabled
> but someone want to use it, for example, KASAN?)
> 
> Vijayanand, If we could work this this, should we still need the
> config option, then? 
> 

Michan, We would still need config option so that we can reduce the
memory consumption on low ram devices using config.

Alex, On this,
"I also suppose device vendors may prefer setting a fixed (maybe
non-default) hash size for low-memory devices rather than letting the
admins increase it."
I see kernel param swiotlb does similar thing i.e; '0' to disable and
set a value to configure size.

I am fine with either of the approaches,

1. I can split this patch into two
   i)  A bool variable to enable/disable stack depot.
   ii) A config for the size.

(or)

2. A run time param - '0' to disable and set a valid size to enable.

Let me know your comments.
-- 
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a
member of Code Aurora Forum, hosted by The Linux Foundation