[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAMzpN2hDJ_dJMvDMkLzTdtyxEr6zZnd9ZhKxNLfRKMx0_qKZUw@mail.gmail.com>
Date: Tue, 22 Mar 2016 10:40:32 -0400
From: Brian Gerst <brgerst@...il.com>
To: Andi Kleen <andi@...stfloor.org>
Cc: Andy Lutomirski <luto@...capital.net>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
X86 ML <x86@...nel.org>, Andi Kleen <ak@...ux.intel.com>
Subject: Re: [PATCH 2/9] x86: Add support for rd/wr fs/gs base
On Mon, Mar 21, 2016 at 6:15 PM, Andi Kleen <andi@...stfloor.org> wrote:
> On Mon, Mar 21, 2016 at 03:08:46PM -0700, Andy Lutomirski wrote:
>> On Mon, Mar 21, 2016 at 3:05 PM, Andi Kleen <andi@...stfloor.org> wrote:
>> >> Please add a patch before this one that renames gs to gsbase. This is
>> >> unreadable as is.
>> >
>> > I investigated this now, and it's not straight forward because on 32bit
>> > ->gs is actually gsindex, not gsbase. So with a straight rename you
>> > would end up with index in base on 32bit, which would be confusing.
>> >
>>
>> I would take this inconsistency as a reason why this needs to be fixed
>> before any further changes are made.
>
> The patchkit doesn't change anything for 32bit, so any inconsistency
> in 32bit is completely orthogonal.
The inconsistency is with the 64-bit code. fs/gs imply the %fs and
%gs registers (the selector index), not the base.
The rename should be:
fs -> fsbase
gs -> gsbase
fsindex -> fs
gsindex -> gs
--
Brian Gerst
Powered by blists - more mailing lists