[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAC5umygy=BBSF_v_vVNnrZ9_1LCpJW0V212yxVqmGUUF-kOOSA@mail.gmail.com>
Date: Thu, 27 Sep 2018 00:49:37 +0900
From: Akinobu Mita <akinobu.mita@...il.com>
To: Arnd Bergmann <arnd@...db.de>
Cc: Sakari Ailus <sakari.ailus@...ux.intel.com>,
Mauro Carvalho Chehab <mchehab@...nel.org>,
Sylwester Nawrocki <s.nawrocki@...sung.com>,
Hans Verkuil <hans.verkuil@...co.com>,
wsa+renesas@...g-engineering.com, gustavo@...eddedor.com,
linux-media@...r.kernel.org, LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] media: ov9650: avoid maybe-uninitialized warnings
2018年9月26日(水) 21:51 Arnd Bergmann <arnd@...db.de>:
>
> The regmap change causes multiple warnings like
>
> drivers/media/i2c/ov9650.c: In function 'ov965x_g_volatile_ctrl':
> drivers/media/i2c/ov9650.c:889:29: error: 'reg2' may be used uninitialized in this function [-Werror=maybe-uninitialized]
> exposure = ((reg2 & 0x3f) << 10) | (reg1 << 2) |
> ~~~~~~~~~~~~~~~^~~~~~
>
> It is apparently hard for the compiler to see here if ov965x_read()
> returned successfully or not. Besides, we have a v4l2_dbg() statement
> that prints an uninitialized value if regmap_read() fails.
>
> Adding an 'else' clause avoids the ambiguity.
>
> Fixes: 361f3803adfe ("media: ov9650: use SCCB regmap")
> Signed-off-by: Arnd Bergmann <arnd@...db.de>
Looks good.
Reviewed-by: Akinobu Mita <akinobu.mita@...il.com>
> ---
> drivers/media/i2c/ov9650.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/media/i2c/ov9650.c b/drivers/media/i2c/ov9650.c
> index 3c9e6798d14b..77944da31de1 100644
> --- a/drivers/media/i2c/ov9650.c
> +++ b/drivers/media/i2c/ov9650.c
> @@ -433,6 +433,8 @@ static int ov965x_read(struct ov965x *ov965x, u8 addr, u8 *val)
> ret = regmap_read(ov965x->regmap, addr, &buf);
> if (!ret)
> *val = buf;
> + else
> + *val = -1;
>
> v4l2_dbg(2, debug, &ov965x->sd, "%s: 0x%02x @ 0x%02x. (%d)\n",
> __func__, *val, addr, ret);
> --
> 2.18.0
>
Powered by blists - more mailing lists