[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <aFRpRnXi3tP_iHEJ@arm.com>
Date: Thu, 19 Jun 2025 20:47:18 +0100
From: Catalin Marinas <catalin.marinas@....com>
To: Mikołaj Lenczewski <miko.lenczewski@....com>
Cc: ryan.roberts@....com, yang@...amperecomputing.com, will@...nel.org,
jean-philippe@...aro.org, robin.murphy@....com, joro@...tes.org,
maz@...nel.org, oliver.upton@...ux.dev, joey.gouly@....com,
james.morse@....com, broonie@...nel.org, ardb@...nel.org,
baohua@...nel.org, suzuki.poulose@....com, david@...hat.com,
jgg@...pe.ca, nicolinc@...dia.com, jsnitsel@...hat.com,
mshavit@...gle.com, kevin.tian@...el.com,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
iommu@...ts.linux.dev
Subject: Re: [PATCH v7 4/4] arm64/mm: Elide tlbi in contpte_convert() under
BBML2
On Thu, Jun 19, 2025 at 08:29:24PM +0100, Catalin Marinas wrote:
> On Tue, Jun 17, 2025 at 09:51:04AM +0000, Mikołaj Lenczewski wrote:
> > + if (!system_supports_bbml2_noabort())
> > + __flush_tlb_range(&vma, start_addr, addr, PAGE_SIZE, true, 3);
> >
> > __set_ptes(mm, start_addr, start_ptep, pte, CONT_PTES);
>
> Eliding the TLBI here is all good but looking at the overall set_ptes(),
> why do we bother with unfold+fold for BBML2? Can we not just change
> them in place without going through __ptep_get_and_clear()?
Ah, it's unlikely that we'd be able to fold them back if only one pte in
the range was modified, so this optimisation would very rarely/never
trigger. So, for this patch:
Reviewed-by: Catalin Marinas <catalin.marinas@....com>
Powered by blists - more mailing lists