[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <40cdea465fef49a8a337b48e2a748138c66a08eb.camel@icenowy.me>
Date: Sat, 01 Apr 2023 09:47:05 +0800
From: Icenowy Zheng <uwu@...nowy.me>
To: Conor Dooley <conor@...nel.org>,
"Lad, Prabhakar" <prabhakar.csengg@...il.com>
Cc: Arnd Bergmann <arnd@...db.de>,
Conor Dooley <conor.dooley@...rochip.com>,
Geert Uytterhoeven <geert+renesas@...der.be>,
Heiko Stuebner <heiko@...ech.de>, Guo Ren <guoren@...nel.org>,
Andrew Jones <ajones@...tanamicro.com>,
Paul Walmsley <paul.walmsley@...ive.com>,
Palmer Dabbelt <palmer@...belt.com>,
Albert Ou <aou@...s.berkeley.edu>,
Samuel Holland <samuel@...lland.org>,
linux-riscv@...ts.infradead.org, Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-renesas-soc@...r.kernel.org,
Biju Das <biju.das.jz@...renesas.com>,
Lad Prabhakar <prabhakar.mahadev-lad.rj@...renesas.com>
Subject: Re: [PATCH v7 0/6] RISC-V non-coherent function pointer based CMO +
non-coherent DMA support for AX45MP
在 2023-03-31星期五的 21:15 +0100,Conor Dooley写道:
> On Fri, Mar 31, 2023 at 08:09:16PM +0000, Lad, Prabhakar wrote:
> > Hi Conor,
> >
> > On Fri, Mar 31, 2023 at 7:05 PM Conor Dooley <conor@...nel.org>
> > wrote:
> > >
> > > On Thu, Mar 30, 2023 at 09:42:11PM +0100, Prabhakar wrote:
> > >
> > > > - This series requires testing on Cores with zicbom and T-Head
> > > > SoCs
> > >
> > > I don't actually know if there are Zicbom parts, may need to test
> > > that
> > > on QEMU.
> > > I had to revert unrelated content to boot, but my D1 NFS setup
> > > seems to
> > > work fine with these changes, so where it is relevant:
> > > Tested-by: Conor Dooley <conor.dooley@...rochip.com> # tyre-
> > > kicking on D1
> > >
> > Thank you for testing this. By any chance did you compare the
> > performance?
>
> No, just tyre kicking. Icenowy had some benchmark for it IIRC, I
> think
> mining some coin or w/e. +CC them.
I previously tested the function pointer based CMO, it do not affect
the performance beyond the measurement error. Maybe it's because CMO
operations are only done at the start and end of DMA operations.
My previous test system is LiteX + OpenC906.
Powered by blists - more mailing lists