[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20160314073803.22c90d69@arm.com>
Date: Mon, 14 Mar 2016 07:38:03 +0000
From: Marc Zyngier <marc.zyngier@....com>
To: Pratyush Anand <panand@...hat.com>
Cc: David Long <dave.long@...aro.org>,
Catalin Marinas <catalin.marinas@....com>,
Will Deacon <will.deacon@....com>,
Sandeepa Prabhu <sandeepa.s.prabhu@...il.com>,
William Cohen <wcohen@...hat.com>,
"Steve Capper" <steve.capper@...aro.org>,
<linux-arm-kernel@...ts.infradead.org>,
<linux-kernel@...r.kernel.org>,
Dave P Martin <Dave.Martin@....com>,
"Mark Rutland" <mark.rutland@....com>,
Robin Murphy <Robin.Murphy@....com>,
"Ard Biesheuvel" <ard.biesheuvel@...aro.org>,
Jens Wiklander <jens.wiklander@...aro.org>,
Christoffer Dall <christoffer.dall@...aro.org>,
Alex Bennée
<alex.bennee@...aro.org>, Yang Shi <yang.shi@...aro.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Viresh Kumar <viresh.kumar@...aro.org>,
"Suzuki K. Poulose" <suzuki.poulose@....com>,
Kees Cook <keescook@...omium.org>,
Zi Shen Lim <zlim.lnx@...il.com>,
John Blackwood <john.blackwood@...r.com>,
Feng Kan <fkan@....com>,
Balamurugan Shanmugam <bshanmugam@....com>,
James Morse <james.morse@....com>,
Vladimir Murzin <Vladimir.Murzin@....com>,
"Mark Salyzyn" <salyzyn@...roid.com>,
Petr Mladek <pmladek@...e.com>,
Andrew Morton <akpm@...ux-foundation.org>,
Mark Brown <broonie@...nel.org>
Subject: Re: [PATCH v11 4/9] arm64: add conditional instruction simulation
support
On Mon, 14 Mar 2016 09:34:55 +0530
Pratyush Anand <panand@...hat.com> wrote:
Hi Pratyush,
> On 13/03/2016:12:09:03 PM, Marc Zyngier wrote:
> > On Wed, 9 Mar 2016 00:32:18 -0500
> > David Long <dave.long@...aro.org> wrote:
> >
> > > +pstate_check_t * const opcode_condition_checks[16] = {
> > > + __check_eq, __check_ne, __check_cs, __check_cc,
> > > + __check_mi, __check_pl, __check_vs, __check_vc,
> > > + __check_hi, __check_ls, __check_ge, __check_lt,
> > > + __check_gt, __check_le, __check_al, __check_al
> >
> > The very last entry seems wrong, or is at least the opposite of what
> > the current code has. It should be something called __check_nv(), and
> > always return false (condition code NEVER).
>
> May be __check_nv() name is more appropriate as per definition, but shouldn't it
> still return true, because TRM says:
> "The condition code NV exists only to provide a valid disassembly of the 0b1111
> encoding, otherwise its behavior is identical to AL"
Indeed, I missed that. But this interpretation is for the A64
instruction set, and this array is also used by the new
arm32_check_condition. The condition code table for A32 seems to
completely ignore the 0b1111 code (there is simply no entry for it), and
it is only in the ConditionHolds pseudocode that you can see how this
is actually special-cased.
So I'm fine leaving the code as it is, but a comment and a pointer to
the ARMv8 ARM wouldn't go amiss.
Thanks,
M.
--
Jazz is not dead. It just smells funny.
Powered by blists - more mailing lists