Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Liam Beguin <liambeguin@gmail.com>
To:     liambeguin@gmail.com, peda@axentia.se, jic23@kernel.org,
        andy.shevchenko@gmail.com, lars@metafoo.de
Cc:     linux-kernel@vger.kernel.org, linux-iio@vger.kernel.org,
        devicetree@vger.kernel.org, robh+dt@kernel.org
Subject: [PATCH v15 04/10] iio: afe: rescale: fix accuracy for small fractional scales
Date:   Sat, 12 Feb 2022 21:57:33 -0500
Message-Id: <20220213025739.2561834-5-liambeguin@gmail.com>
In-Reply-To: <20220213025739.2561834-1-liambeguin@gmail.com>
References: <20220213025739.2561834-1-liambeguin@gmail.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

The approximation caused by integer divisions can be costly on smaller
scale values since the decimal part is significant compared to the
integer part. Switch to an IIO_VAL_INT_PLUS_NANO scale type in such
cases to maintain accuracy.

Signed-off-by: Liam Beguin <liambeguin@gmail.com>
Reviewed-by: Peter Rosin <peda@axentia.se>
---
 drivers/iio/afe/iio-rescale.c | 20 +++++++++++++++++---
 1 file changed, 17 insertions(+), 3 deletions(-)

diff --git a/drivers/iio/afe/iio-rescale.c b/drivers/iio/afe/iio-rescale.c
index 8eaf766e28cc..5d78f0cf47d2 100644
--- a/drivers/iio/afe/iio-rescale.c
+++ b/drivers/iio/afe/iio-rescale.c
@@ -24,7 +24,7 @@ int rescale_process_scale(struct rescale *rescale, int scale_type,
 			  int *val, int *val2)
 {
 	s64 tmp;
-	s32 rem;
+	s32 rem, rem2;
 	u32 mult;
 	u32 neg;
 
@@ -43,9 +43,23 @@ int rescale_process_scale(struct rescale *rescale, int scale_type,
 		tmp = (s64)*val * 1000000000LL;
 		tmp = div_s64(tmp, rescale->denominator);
 		tmp *= rescale->numerator;
-		tmp = div_s64(tmp, 1000000000LL);
+
+		tmp = div_s64_rem(tmp, 1000000000LL, &rem);
 		*val = tmp;
-		return scale_type;
+
+		if (!rem)
+			return scale_type;
+
+		tmp = 1 << *val2;
+
+		rem2 = *val % (int)tmp;
+		*val = *val / (int)tmp;
+
+		*val2 = rem / (int)tmp;
+		if (rem2)
+			*val2 += div_s64((s64)rem2 * 1000000000LL, tmp);
+
+		return IIO_VAL_INT_PLUS_NANO;
 	case IIO_VAL_INT_PLUS_NANO:
 	case IIO_VAL_INT_PLUS_MICRO:
 		mult = scale_type == IIO_VAL_INT_PLUS_NANO ? 1000000000L : 1000000L;
-- 
2.35.1.4.g5d01301f2b86