From 065835767dd877d8b0e2ccdbe5b915c5e22bd62a Mon Sep 17 00:00:00 2001 From: dakkar Date: Sat, 9 Jun 2018 12:55:53 +0100 Subject: working CHIP sender --- trasmitter/README.rst.txt | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) (limited to 'trasmitter/README.rst.txt') diff --git a/trasmitter/README.rst.txt b/trasmitter/README.rst.txt index d2fd957..78200a4 100644 --- a/trasmitter/README.rst.txt +++ b/trasmitter/README.rst.txt @@ -38,12 +38,5 @@ sysfs interface`_, seem to be fast enough for our purposes. See the ``sender-chip.pl`` test program. -The timing via the sysfs interface is awful. - -It looks like we can drive the pins (not the XIO ones) via -memory-mapped registers: the sunxi-tools ``pio.c`` has most of the -code needed. - -We could write a C program to just send the signal train. Look at -``clock.c`` for a way to handle fast / real-time scheduling and -delays. +If we keep the ``*/value`` file open, we get good enough timing. The +``sender-chip.c`` program works. -- cgit v1.2.3