qcom-charger: fg-util: add float decode function

Some SRAM registers are encoded using a floating point representation.
Add a function to decode these registers into signed micro-unit
integers.

The exponent and mantissa are signed integers represented by two's
complement, and the exponent value is offset by -9. This is a half
float representation stored as:

EEEEE MMMMMMMMMMM

Where E are the exponent bits, and M are the mantissa bits.

To decode this representation the following formula is applied:
2^(exponent - 9) * mantissa

Change-Id: Id9f28a0eeb2a904aca41eb46d0215d80287e0b88
Signed-off-by: Nicholas Troast <ntroast@codeaurora.org>

0 files changed, 0 insertions, 0 deletions