[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20160309022915.GB3898@sirena.org.uk>
Date: Wed, 9 Mar 2016 09:29:15 +0700
From: Mark Brown <broonie@...nel.org>
To: Ming Lei <tom.leiming@...il.com>
Cc: Stephen Boyd <stephen.boyd@...aro.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
linux-arm@...ts.infradead.org,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Robin Murphy <robin.murphy@....com>,
Laura Abbott <labbott@...hat.com>,
Arnd Bergmann <arnd@...db.de>,
Marek Szyprowski <m.szyprowski@...sung.com>,
Andrew Morton <akpm@...ux-foundation.org>,
Catalin Marinas <catalin.marinas@....com>,
Will Deacon <will.deacon@....com>,
Vikram Mulukutla <markivx@...eaurora.org>,
Mimi Zohar <zohar@...ux.vnet.ibm.com>
Subject: Re: [RFC/PATCH 4/4] firmware: Support requesting firmware directly
into DMA memory
On Tue, Mar 08, 2016 at 09:42:17PM +0800, Ming Lei wrote:
> On Tue, Mar 8, 2016 at 5:22 PM, Stephen Boyd <stephen.boyd@...aro.org> wrote:
> > This design creates needless memory pressure and delays loading
> > because we have to copy from kernel memory to somewhere else.
> Given firmware request can't be a frequent operation, I don't think it is
> a big deal about the so called memory pressure and delay.
Boot time is a very important metric for some embedded product classes
and is obviously one of the most common times when we need to load lots
of firmware. Some mobile platforms have some individual firmwares that
are getting on for 100M in size, holding a copy of one of them in RAM is
a substantial proportion of the available memory in a system with 1-2G
total especially when the system is also busy with other things.
Download attachment "signature.asc" of type "application/pgp-signature" (474 bytes)
Powered by blists - more mailing lists