[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20240611121219.GP791043@ziepe.ca>
Date: Tue, 11 Jun 2024 09:12:19 -0300
From: Jason Gunthorpe <jgg@...pe.ca>
To: Tomasz Jeznach <tjeznach@...osinc.com>
Cc: Zong Li <zong.li@...ive.com>, Joerg Roedel <joro@...tes.org>,
Will Deacon <will@...nel.org>, Robin Murphy <robin.murphy@....com>,
Paul Walmsley <paul.walmsley@...ive.com>,
Palmer Dabbelt <palmer@...belt.com>,
Albert Ou <aou@...s.berkeley.edu>,
Anup Patel <apatel@...tanamicro.com>,
Sunil V L <sunilvl@...tanamicro.com>,
Nick Kossifidis <mick@....forth.gr>,
Sebastien Boeuf <seb@...osinc.com>,
Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzk+dt@...nel.org>,
Conor Dooley <conor+dt@...nel.org>, devicetree@...r.kernel.org,
iommu@...ts.linux.dev, linux-riscv@...ts.infradead.org,
linux-kernel@...r.kernel.org, linux@...osinc.com,
Lu Baolu <baolu.lu@...ux.intel.com>, Jim Shu <jim.shu@...ive.com>,
Vincent Chen <vincent.chen@...ive.com>
Subject: Re: [PATCH v6 5/7] iommu/riscv: Device directory management.
On Mon, Jun 10, 2024 at 07:00:34PM -0700, Tomasz Jeznach wrote:
> For now, I'll change the implementation to assume negative caching for
> DDTE and will follow up with device tree / driver updates to make the
> invalidation optional when revised specifications will be available.
Is there a reason to make it optional? It seems like it doesn't have
any performance downside to just always invalidate, attachment is not
a critical path operation.
I could see making something like negative PTE invalidation optional
as that is more performance path..
Jason
Powered by blists - more mailing lists