[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20181108152530.GA27198@infradead.org>
Date: Thu, 8 Nov 2018 07:25:30 -0800
From: Christoph Hellwig <hch@...radead.org>
To: Joakim Tjernlund <Joakim.Tjernlund@...inera.com>
Cc: "danielwa@...co.com" <danielwa@...co.com>,
"dwmw2@...radead.org" <dwmw2@...radead.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"nkela@...co.com" <nkela@...co.com>,
"linux-mtd@...ts.infradead.org" <linux-mtd@...ts.infradead.org>,
"xe-linux-external@...co.com" <xe-linux-external@...co.com>
Subject: Re: [PATCH] jffs2: implement mount option to configure endianness
On Wed, Nov 07, 2018 at 06:41:47PM +0000, Joakim Tjernlund wrote:
> > Certainly not. I'm not sure which architectures do have Spectre V2
> > mitigations which make indirect branches expensive now... perhaps there is
> > no intersection with the cases where we really care about JFFS2 being
> > CPU-bound?
>
> All that passing of a new extra arg and extra if .. elif .. does cost too.
I know of no current microarchitecture where an indirect call is cheaper
than a simple branch, which can be very well predicted.
Powered by blists - more mailing lists