[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <Ylb6gvRJ27BCPv1z@casper.infradead.org>
Date: Wed, 13 Apr 2022 17:29:54 +0100
From: Matthew Wilcox <willy@...radead.org>
To: Anshuman Khandual <anshuman.khandual@....com>
Cc: inux-mm@...ck.org, Andrew Morton <akpm@...ux-foundation.org>,
Will Deacon <will@...nel.org>,
"Aneesh Kumar K.V" <aneesh.kumar@...ux.ibm.com>,
Nick Piggin <npiggin@...il.com>,
Peter Zijlstra <peterz@...radead.org>,
Arnd Bergmann <arnd@...db.de>, linux-arch@...r.kernel.org,
linux-mm@...ck.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] tlb/hugetlb: Add framework to handle PGDIR_SIZE HugeTLB
pages
On Wed, Apr 13, 2022 at 03:37:14PM +0530, Anshuman Khandual wrote:
> Change tlb_remove_huge_tlb_entry() to accommodate larger PGDIR_SIZE HugeTLB
> pages via adding a new helper tlb_flush_pgd_range(). While here also update
> struct mmu_gather as required, that is add a new member cleared_pgds.
Is this just completionist junk, or are there really CPUs that support
PGDIR sized pages? We already have support for p4d (512GB), even though
at least Intel don't support pages larger than 1GB. I think the largest
size page I've heard of is 16GB on POWER. PGDIR_SIZE would be 256TB on
x86 which just seems ludicrous to be talking about supporting.
> ---
> This applies on v5.18-rc2, some earlier context could be found here
>
> https://lore.kernel.org/all/20220406112124.GD2731@worktop.programming.kicks-ass.net/
Read it, doesn't talk about any real users of this.
Powered by blists - more mailing lists