[PATCH 00/10] Hopefully last major suspend work
andy at openmoko.com
Mon Jun 16 00:17:05 CEST 2008
The following series gets me really smooth fast suspend (since a
previously posted patch, with no white flashes) and now resume.
The backlight sequencing is modified by the patchset to actually
work, so backlight only comes up after the LCM has been brought
out of deep sleep. That means that there is no white flash any
more in resume. In addition, the patchset adds to pcf50633
platform a setting for the backlight ramp speed on resume, which
I set to take something around 1s to come to full brightness.
Previously, when resuming the framebuffer contents were often
degraded with noise, and / or exciting horizontal lines. I
took apart the suspend resume action of glamo-core and fixed
the sequencing and lack of "engine management" that caused this --
glamo framebuffer contents are now untouched by suspend / resume
and come back cleanly.
Sean and Mike reported more mixed results that I get with
previous patches, the patchset includes a major change to lock
out pcf IRQ service completely from start of pcf suspend until
completion of pcf resume. To the extent the pcf irq work was
implicated in their problems, this should do something about
them. It also makes the other work functions in pcf50633
There are also more cleanups of pcf50633 including moving out
any mention of GTAxx in the code into platform.
Anyway allowing for any continuing problems found by other
testers, basically this patchset with the others makes GTA02
suspend look like it should on a cool phone now.
Andy Green (10):
arch/arm/mach-s3c2440/mach-gta02.c | 30 +++
drivers/base/core.c | 5 +
drivers/i2c/busses/i2c-s3c2410.c | 1
drivers/i2c/chips/pcf50633.c | 356 +++++++++++++++++++++++-------------
drivers/mfd/glamo/glamo-core.c | 269 +++++++++++++--------------
drivers/mfd/glamo/glamo-core.h | 2
drivers/mfd/glamo/glamo-regs.h | 6 +
include/linux/pcf50633.h | 10 +
8 files changed, 408 insertions(+), 271 deletions(-)
More information about the openmoko-kernel