Commit 81595ed4 authored Sep 07, 2021 by Robert Lubos Committed by Christopher Friedt Sep 16, 2021

net: lwm2m: Fix binary to float32_value_t conversion



The helper function to conver 32-bit binary float value to
float32_value_t incorrectly identified the "hidden" bit, resulting in
invalid conversion when TLV encoding was used to write a resource
(the output value was divided by 2).

Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>

parent 36a72dd1

subsys/net/lib/lwm2m/lwm2m_util.c

+2 −2

Original line number	Diff line number	Diff line
		@@ -197,8 +197,8 @@ int lwm2m_b32_to_f32(uint8_t b32, size_t len, float32_value_t f32)
		/* remove bias */
		e -= 127;

		/* enable "hidden" fraction bit 23 which is always 1 */
		f = ((int32_t)1 << 22);
		/* enable "hidden" fraction bit 24 which is always 1 */
		f = ((int32_t)1 << 23);
		/* calc fraction: bits 22-0 */
		f += ((int32_t)(b32[1] & 0x7F) << 16);
		f += ((int32_t)b32[2] << 8);

Admin message