Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Vineet Gupta <Vineet.Gupta1@synopsys.com>
To:     Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>,
        "linux-snps-arc@lists.infradead.org" 
        <linux-snps-arc@lists.infradead.org>
CC:     "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "Alexey Brodkin" <Alexey.Brodkin@synopsys.com>,
        Peter Zijlstra <peterz@infradead.org>,
        "Will Deacon" <will.deacon@arm.com>,
        Boqun Feng <boqun.feng@gmail.com>
Subject: Re: [PATCH] ARC: atomic64: fix atomic64_add_unless function
Thread-Topic: [PATCH] ARC: atomic64: fix atomic64_add_unless function
Thread-Index: AQHUMY2jvyJle0DCLEGtK0ktniYvNg==
Date:   Tue, 14 Aug 2018 13:42:23 +0000
Message-ID: <C2D7FE5348E1B147BCA15975FBA23075012B08CB78@us01wembx1.internal.synopsys.com>
References: <20180811160856.24936-1-Eugeniy.Paltsev@synopsys.com>
Accept-Language: en-US
Content-Language: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 08/11/2018 09:09 AM, Eugeniy Paltsev wrote:=0A=
> Current implementation of 'atomic64_add_unless' function=0A=
> (and hence 'atomic64_inc_not_zero') return incorrect value=0A=
> if lover 32 bits of compared 64-bit number are equal and=0A=
> higher 32 bits aren't.=0A=
>=0A=
> For in following example atomic64_add_unless must return '1'=0A=
> but it actually returns '0':=0A=
> --------->8---------=0A=
> atomic64_t val =3D ATOMIC64_INIT(0x4444000000000000LL);=0A=
> int ret =3D atomic64_add_unless(&val, 1LL, 0LL)=0A=
> --------->8---------=0A=
>=0A=
> This happens because we write '0' to returned variable regardless=0A=
> of higher 32 bits comparison result.=0A=
>=0A=
> So fix it.=0A=
>=0A=
> NOTE:=0A=
>  this change was tested with atomic64_test.=0A=
>=0A=
> Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>=0A=
=0A=
LGTM. Curious, was this from code review or did u actually run into this ?=
=0A=
=0A=
Thx,=0A=
-Vineet=0A=
=0A=
> ---=0A=
>  arch/arc/include/asm/atomic.h | 2 +-=0A=
>  1 file changed, 1 insertion(+), 1 deletion(-)=0A=
>=0A=
> diff --git a/arch/arc/include/asm/atomic.h b/arch/arc/include/asm/atomic.=
h=0A=
> index 11859287c52a..e840cb1763b2 100644=0A=
> --- a/arch/arc/include/asm/atomic.h=0A=
> +++ b/arch/arc/include/asm/atomic.h=0A=
> @@ -578,11 +578,11 @@ static inline int atomic64_add_unless(atomic64_t *v=
, long long a, long long u)=0A=
>  =0A=
>  	__asm__ __volatile__(=0A=
>  	"1:	llockd  %0, [%2]	\n"=0A=
> -	"	mov	%1, 1		\n"=0A=
>  	"	brne	%L0, %L4, 2f	# continue to add since v !=3D u \n"=0A=
>  	"	breq.d	%H0, %H4, 3f	# return since v =3D=3D u \n"=0A=
>  	"	mov	%1, 0		\n"=0A=
>  	"2:				\n"=0A=
> +	"	mov	%1, 1		\n"=0A=
>  	"	add.f   %L0, %L0, %L3	\n"=0A=
>  	"	adc     %H0, %H0, %H3	\n"=0A=
>  	"	scondd  %0, [%2]	\n"=0A=
=0A=