Received-SPF: pass (google.com: domain of linux-kernel+bounces-136634-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Precedence: bulk
Mime-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset=UTF-8
Date: Tue, 09 Apr 2024 12:09:17 +0200
Message-Id: <D0FID4G7T4N0.194WATCMTOUEY@bootlin.com>
Cc: "Rob Herring" <robh+dt@kernel.org>, "Krzysztof Kozlowski"
 <krzysztof.kozlowski+dt@linaro.org>, "Conor Dooley" <conor+dt@kernel.org>,
 "Vaishnav Achath" <vaishnav.a@ti.com>, "Thomas Bogendoerfer"
 <tsbogend@alpha.franken.de>, "Rob Herring" <robh@kernel.org>,
 <linux-spi@vger.kernel.org>, <devicetree@vger.kernel.org>,
 <linux-kernel@vger.kernel.org>, <linux-mips@vger.kernel.org>, "Vladimir
 Kondratiev" <vladimir.kondratiev@mobileye.com>, "Gregory CLEMENT"
 <gregory.clement@bootlin.com>, "Thomas Petazzoni"
 <thomas.petazzoni@bootlin.com>, "Tawfik Bayouk"
 <tawfik.bayouk@mobileye.com>
To: "Mark Brown" <broonie@kernel.org>
From: =?utf-8?q?Th=C3=A9o_Lebrun?= <theo.lebrun@bootlin.com>
Subject: Re: [PATCH v2 08/11] spi: cadence-qspi: add early busywait to
 cqspi_wait_for_bit()
References: <20240405-cdns-qspi-mbly-v2-0-956679866d6d@bootlin.com>
 <20240405-cdns-qspi-mbly-v2-8-956679866d6d@bootlin.com>
 <1f7087ad-824e-47fe-9953-ed5152c8f18f@sirena.org.uk>
 <D0ETJXNHOUKQ.C368FEZVM9DD@bootlin.com>
 <9cb48440-c71e-4a73-8104-4780f0e98e72@sirena.org.uk>
In-Reply-To: <9cb48440-c71e-4a73-8104-4780f0e98e72@sirena.org.uk>

Hello,

On Mon Apr 8, 2024 at 6:40 PM CEST, Mark Brown wrote:
> On Mon, Apr 08, 2024 at 04:42:43PM +0200, Th=C3=A9o Lebrun wrote:
> > On Mon Apr 8, 2024 at 4:16 PM CEST, Mark Brown wrote:
> > > On Fri, Apr 05, 2024 at 05:02:18PM +0200, Th=C3=A9o Lebrun wrote:
>
> > > > The reason is to avoid hrtimer interrupts on the system. All read
> > > > operations take less than 100=C2=B5s.
>
> > > Why would this be platform specific, this seems like a very standard
> > > optimisation technique?
>
> > It does not make sense if you know that all read operations take more
> > than 100=C2=B5s. I preferred being conservative. If you confirm it make=
s
> > sense I'll remove the quirk.
>
> It does seem plausible at least, and the time could be made a tuneable
> with quirks or otherwise if that's needed.  I think I'd expect the MIPS
> platform you're working with to be towards the lower end of performance
> for systems that are new enough to have this hardware.

Next revision will do the same busywait behavior unconditionally then.

Thanks!

--
Th=C3=A9o Lebrun, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com