Re: [PATCH v1 04/11] mm/memremap: add ZONE_DEVICE support for compound pages

From: Joao Martins <joao.m.martins@oracle.com>
To: Matthew Wilcox <willy@infradead.org>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>,
	Linux MM <linux-mm@kvack.org>,
	linux-nvdimm <linux-nvdimm@lists.01.org>,
	Jason Gunthorpe <jgg@ziepe.ca>, Jane Chu <jane.chu@oracle.com>,
	Muchun Song <songmuchun@bytedance.com>,
	Mike Kravetz <mike.kravetz@oracle.com>,
	Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [PATCH v1 04/11] mm/memremap: add ZONE_DEVICE support for compound pages
Date: Thu, 6 May 2021 13:15:23 +0100	[thread overview]
Message-ID: <8df871e2-3c36-5fba-c98f-85caef7261b7@oracle.com> (raw)
In-Reply-To: <20210506114357.GA388843@casper.infradead.org>

On 5/6/21 12:43 PM, Matthew Wilcox wrote:
> On Thu, May 06, 2021 at 11:23:25AM +0100, Joao Martins wrote:
>>>> I think it is ok for dax/nvdimm to continue to maintain their align
>>>> value because it should be ok to have 4MB align if the device really
>>>> wanted. However, when it goes to map that alignment with
>>>> memremap_pages() it can pick a mode. For example, it's already the
>>>> case that dax->align == 1GB is mapped with DEVMAP_PTE today, so
>>>> they're already separate concepts that can stay separate.
>>>
>>> devdax namespace with align of 1G implies we expect to map them with 1G
>>> pte entries? I didn't follow when you say we map them today with
>>> DEVMAP_PTE entries.
>>
>> This sort of confusion is largelly why Dan is suggesting a @geometry for naming rather
>> than @align (which traditionally refers to page tables entry sizes in pagemap-related stuff).
>>
>> DEVMAP_{PTE,PMD,PUD} refers to the representation of metadata in base pages (DEVMAP_PTE),
>> compound pages of PMD order (DEVMAP_PMD) or compound pages of PUD order (DEVMAP_PUD).
>>
>> So, today:
>>
>> * namespaces with align of 1G would use *struct pages of order-0* (DEVMAP_PTE) backed with
>> PMD entries in the direct map.
>> * namespaces with align of 2M would use *struct pages of order-0* (DEVMAP_PTE) backed with
>> PMD entries in the direct map.
>>
>> After this series:
>>
>> * namespaces with align of 1G would use *compound struct pages of order-30* (DEVMAP_PUD)
>> backed with PMD entries in the direct map.
> 
> order-18
> 
Right, thanks for the correction.

Forgot to subtract PAGE_SIZE order (12).

>> * namespaces with align of 1G would use *compound struct pages of order-21* (DEVMAP_PMD)
>> backed with PTE entries in the direct map.
> 
> i think you mean "align of 2M", and order-9.
> 
True.

> (note that these numbers are/can be different for powerpc since it
> supports different TLB page sizes.  please don't get locked into x86
> sizes, and don't ignore the benefits *to software* of managing memory
> in sizes other than just those supported by the hardware).
> 
I am not ignoring that either that I think.

The page size looking values above is also what the consumer of this (device-dax) allows
you to use as @align, hence why you see DEVMAP_PTE, PMD and PUD as the sizes.
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org