view blobs/patches/main-pirelli.patch @ 624:012028896cfb

FFS dev.c, Leonardo target: Fujitsu MB84VF5F5F4J2 #if 0'ed out The FFS code we got from TI/Openmoko had a stanza for "Fujitsu MB84VF5F5F4J2 stacked device", using a fake device ID code that would need to be patched manually into cfgffs.c (suppressing and overriding autodetection) and using an FFS base address in the nCS2 bank, indicating that this FFS config was probably meant for the MCP version of Leonardo which allows for 16 MiB flash with a second bank on nCS2. We previously had this FFS config stanza conditionalized under CONFIG_TARGET_LEONARDO because the base address contained therein is invalid for other targets, but now that we actually have a Leonardo build target in FC Magnetite, I realize that the better approach is to #if 0 out this stanza altogether: it is already non-functional because it uses a fake device ID code, thus it is does not add support for more Leonardo board variants, instead it is just noise.
author Mychaela Falconia <falcon@freecalypso.org>
date Sun, 22 Dec 2019 21:24:29 +0000
parents 8dd671b7d41e
children
line wrap: on
line source

# This patch applies to the Init_Target() function in the init.obj module in
# main.lib; the present version is for making TCS211 run on the Pirelli.
# This patch sets the same memory and peripheral chip select timings and
# widths as Pirelli's fw.

[init.obj]

# value goes into nCS0, nCS1 and nCS3 config registers
.text 66 A4
# the nCS2 setting in our original blob is already correct for the Pirelli

# value goes into nCS4 config reg
.text 72 A7

# nop out the write into 0x02700000

.text 128 C0
.text 129 46