On Mon, 2016-02-01 at 17:04 +0200, Jarkko Nikula wrote:

On 01/15/2016 08:38 AM, Fu, Zhonghui wrote:

quoted

Now, PM core supports asynchronous suspend/resume mode for devices
during system suspend/resume, and the power state transition of one
device may be completed in separate kernel thread. PM core ensures
all power state transition dependency between devices. This patch
enables designware i2c controllers to suspend/resume
asynchronously.
This will take advantage of multicore and improve system
suspend/resume
speed. After enabling all i2c devices, i2c adapters and i2c
controllers
on ASUS T100TA tablet, the system suspend-to-idle time is reduced
to
about 510ms from 750ms, and the system resume time is reduced to
about
790ms from 900ms.

Signed-off-by: Zhonghui Fu <redacted>
---
Changes in v2:
- Move the device_enable_async_suspend() call into i2c_dw_proble()

  drivers/i2c/busses/i2c-designware-core.c |    1 +
  1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/drivers/i2c/busses/i2c-designware-core.c

b/drivers/i2c/busses/i2c-designware-core.c
index ba9732c..5d6ad27 100644

--- a/drivers/i2c/busses/i2c-designware-core.c
+++ b/drivers/i2c/busses/i2c-designware-core.c

@@ -861,6 +861,7 @@ int i2c_dw_probe(struct dw_i2c_dev *dev)

  	init_completion(&dev->cmd_complete);
  	mutex_init(&dev->lock);
+	device_enable_async_suspend(dev->dev);

I'm aware Andy had concerns about about stability and you also
mention 
in another thread:

"if enable all LPSS devices suspend/resume asynchronously, the
system 
can't resume sometimes on ASUS T100TA(BayTrail-T SoC). But, I have 
verified that the system can resume normally every time if enable
only 
i2c controller async mode and let other LPSS devices in sync mode on 
ASUS T100TA.".

What I'm thinking would this change move things forward and help to
find 
out what potential underlying issues there are in LPSS code.

I'm fine with this I2C side change but can we assure we don't 
accidentally enable the same on other LPSS devices before finding
the 
root cause for the crash?

Besides that we have to be really aware about DMA power related fix
introduced in v4.5-rc1 in acpi_lpss.c [1]. So, I would like to see a
wide testing especially on Intel Baytrail / Braswell platforms before
enabling it.

[1] http://comments.gmane.org/gmane.linux.acpi.devel/80263

-- 
Andy Shevchenko [off-list ref]
Intel Finland Oy