layer1: introduce concept of a 'l1 completion'
authorHarald Welte <laforge@gnumonks.org>
Fri, 7 May 2010 14:46:47 +0000 (16:46 +0200)
committerHarald Welte <laforge@gnumonks.org>
Mon, 17 May 2010 07:23:35 +0000 (09:23 +0200)
commitea3b3826f2c8650d36d033e99a7af08c05ef562a
tree1e490b0828cd373be6e412b2d6589c0e1e0d9e9d
parent452bc184700841744884e4675db5e3bc520198d5
layer1: introduce concept of a 'l1 completion'

The idea is that the L1S part can schedule a completion handler which
will then execute in the asynchronous L1A part.  This should keep the
FIQ priority L1S extremely short, deferring most of the work into
the L1A part that runs in regular process context.
src/target/firmware/include/layer1/async.h
src/target/firmware/include/layer1/sync.h
src/target/firmware/layer1/async.c
src/target/firmware/layer1/sync.c