Hi,
The following set of patches apply on top of the Kevin's pm-wip-opp
branch. What I have tried to do in this set of patches are:
(Not in patch-set order)
* OPP layer internals have moved to list based implementation.
* The OPP layer APIs have been changed. The search APIs have been
reduced to one instead of opp_find_{exact|floor|ceil}.
* OPP book-keeping is now done inside OPP layer. We do not have to
maintain pointers to {mpu|dsp|l3}_opp, outside this layer.
* removed omap_opp_def as this is very similar to omap_opp.
* Cleaned up the SRF framework to use new OPP APIs.
* Removed VDD1,2 OPP resources and instead introduced voltage resources.
This results in leaner code.
* L3 frequency change now happens from cpufreq notifier mechanism.
* cpufreq driver now honors the CPUFREQ{H|L} flags.
* uv to vsel precision loss is handled cleanly.
Verified this on zoom2 with and without lock debugging.
Some output from cpufreq translation statistics.
#
cat /sys/devices/system/cpu/cpu0/cpufreq/stats/trans_table
From :
To
600000 550000 500000 250000 125000
600000: 0 6804 4536 4536 4535
550000: 4536 0 6804 4536 4535
500000: 4537 4536 0 6804 4535
250000: 4536 4536 4536 0 6802
125000: 6802 4535 4535 4535 0
diffstat output!
mach-omap2/pm.h | 17 +
mach-omap2/pm34xx.c | 79 ++++--
mach-omap2/resource34xx.c | 542 ++++++++++++++-----------------------------
mach-omap2/resource34xx.h | 62 ++--
mach-omap2/smartreflex.c | 285 +++++++++++-----------
mach-omap2/smartreflex.h | 16 -
plat-omap/common.c | 6
plat-omap/cpu-omap.c | 73 +++++
plat-omap/include/plat/io.h | 1
plat-omap/include/plat/opp.h | 265 +++++----------------
plat-omap/omap-pm-noop.c | 35 --
plat-omap/omap-pm-srf.c | 38 ---
plat-omap/opp.c | 497 +++++++++++++++++++++------------------
plat-omap/opp_twl_tps.c | 11
14 files changed, 851 insertions(+), 1076 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
