Received-SPF: pass (google.com: domain of linux-kernel+bounces-29211-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Precedence: bulk
MIME-Version: 1.0
References: <20240110102102.61587-1-tudor.ambarus@linaro.org>
 <20240110102102.61587-12-tudor.ambarus@linaro.org> <CAPLW+4mKBwsc9VLrGTd2k6d0n-K9TZAjH6M8trcK3Av8TQ2Ngg@mail.gmail.com>
 <6b6b1512-18a2-48bd-b284-8f4deff84309@linaro.org>
In-Reply-To: <6b6b1512-18a2-48bd-b284-8f4deff84309@linaro.org>
From: Sam Protsenko <semen.protsenko@linaro.org>
Date: Wed, 17 Jan 2024 10:24:21 -0600
Message-ID: <CAPLW+4=k-hrJP5oDVG9+-XmAbNhpRoe-jnjphkk0dnbEH=r9vw@mail.gmail.com>
Subject: Re: [PATCH 11/18] tty: serial: samsung: don't compare with zero an if
 (bitwise expression)
To: Tudor Ambarus <tudor.ambarus@linaro.org>
Cc: krzysztof.kozlowski@linaro.org, alim.akhtar@samsung.com, 
	gregkh@linuxfoundation.org, jirislaby@kernel.org, 
	linux-arm-kernel@lists.infradead.org, linux-samsung-soc@vger.kernel.org, 
	linux-kernel@vger.kernel.org, linux-serial@vger.kernel.org, 
	andre.draszik@linaro.org, peter.griffin@linaro.org, kernel-team@android.com, 
	willmcvicker@google.com
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Wed, Jan 17, 2024 at 9:41=E2=80=AFAM Tudor Ambarus <tudor.ambarus@linaro=
org> wrote:
>
>
>
> On 1/16/24 18:38, Sam Protsenko wrote:
> > On Wed, Jan 10, 2024 at 4:24=E2=80=AFAM Tudor Ambarus <tudor.ambarus@li=
naro.org> wrote:
> >>
> >> Since an if tests the numeric value of an expression, certain coding
> >> shortcuts can be used. The most obvious one is writing
> >>     if (expression)
> >> instead of
> >>     if (expression !=3D 0)
> >>
> >> Since our case is a bitwise expression, it's more natural and clear to
> >> use the ``if (expression)`` shortcut.
> >
> > Maybe the author of this code:
> >
> >     (ufstat & info->tx_fifomask) !=3D 0
> >
> > just wanted to outline (logically) that the result of this bitwise
> > operation produces FIFO length, which he checks to have non-zero
> > length? Mechanically of course it doesn't matter much, and I guess
>
> that's a bitwise AND with the fifo mask to check if the fifo is empty or
> not, it doesn't care about the length, just if the fifo is empty. IOW if
> any of those bits are set, the fifo is not empty. I think not comparing
> with zero explicitly is better. At the same time I'm fine dropping the
> patch as well. So please tell me if you want me to reword the commit
> message or drop the patch entirely.
>

I'm not opposed to this patch, just don't have any preference in this
case. But the patch is ok with me.

> > everyone can understand what's going on there even without '!=3D 0'
> > part. But it looks quite intentional to me, because in the same 'if'
> > block the author uses this as well:
> >
> >     (ufstat & info->tx_fifofull)
>
> tx_fifofull is just a bit in the register, in my case BIT(24). If that
> bit is one, the fifo is full. Not comparing with zero is fine here, as
> we're interested just in that bit/flag.
>
> >
> > without any comparison operators.
> >
> >>
> >> Signed-off-by: Tudor Ambarus <tudor.ambarus@linaro.org>
> >> ---
> >>  drivers/tty/serial/samsung_tty.c | 3 +--
> >>  1 file changed, 1 insertion(+), 2 deletions(-)
> >>
> >> diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/sam=
sung_tty.c
> >> index dbbe6b8e3ceb..f2413da14b1d 100644
> >> --- a/drivers/tty/serial/samsung_tty.c
> >> +++ b/drivers/tty/serial/samsung_tty.c
> >> @@ -988,8 +988,7 @@ static unsigned int s3c24xx_serial_tx_empty(struct=
 uart_port *port)
> >>         u32 ufcon =3D rd_regl(port, S3C2410_UFCON);
> >>
> >>         if (ufcon & S3C2410_UFCON_FIFOMODE) {
> >> -               if ((ufstat & info->tx_fifomask) !=3D 0 ||
> >> -                   (ufstat & info->tx_fifofull))
> >> +               if ((ufstat & info->tx_fifomask) || (ufstat & info->tx=
_fifofull))
> >
> > Does this line fit into 80 characters? If no, please rework it so it
>
> it fits
>

Just checked, and it's 1 character off (so it has length of 81
characters). I know it's not a strong rule in kernel anymore, but I
like it personally. If you are going to fix that, be free to add:

Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>

> > does. I guess it's also possible to get rid of superfluous braces
> > there, but then the code might look confusing, and I'm not sure if
> > checkpatch would be ok with that.
> >
>
> I find it better with the braces.
>
> Thanks!
> ta