On Tue 30 Sep 12:59 PDT 2014, Andy Gross wrote:

On Mon, Sep 29, 2014 at 06:53:24PM -0700, Bjorn Andersson wrote:

quoted

On Mon 29 Sep 15:00 PDT 2014, Andy Gross wrote:

quoted

The runtime pm calls need to be done before populating the children via the
i2c_add_adapter call.  If this is not done, a child can run into issues trying
to do i2c read/writes due to the pm_runtime_sync failing.

May I ask in what case this would fail?  I thought we tested this as we found
the faulty error check after calling pm_runtime_get_sync().

This is a different kind of failure.   If during probe, the children do some I2C
activity, they will encounter issues with the runtime_sync due to the pm_runtime
not being initialized.  However, once the qup probe completes, everything works
fine.

The original runtime_sync issue revolved around the runtime_sync return value
being misinterpreted due to the incorrect check.

Yeah, I just wondered why my testing didn't trigger this. But I suspect that
the answer is simply that the client I ran with did not do i2c accesses from
probe.

Regards,
Bjorn