Re: [PATCH v2] i2c: i2c-qcom-geni: Serve transfer during early resume stage

[Mukesh Kumar Savaliya]

Thanks Bjorn.

On 11/19/2024 8:29 PM, Bjorn Andersson wrote:
> On Tue, Nov 19, 2024 at 8:29 AM Mukesh Kumar Savaliya
> <quic_msavaliy@xxxxxxxxxxx> wrote:
> > On 10/11/2024 11:28 PM, Bjorn Andersson wrote:
> > > On Fri, Oct 11, 2024 at 05:47:57PM +0530, Mukesh Kumar Savaliya wrote:
> [..]
> > > > pm_runtime_get_sync() function fails during PM early resume and returning
> > > > -EACCES because runtime PM for the device is disabled at the early stage
> > > > causing i2c transfer to fail. Make changes to serve transfer with forced
> > > > resume.
> > > >
> > > > Few i2c clients like PCI OR touch may request i2c transfers during early
> > > > resume stage. In order to serve transfer request do :
> > >
> > > This problem description is too generic. I am not aware of any use case
> > > upstream where PCI or touch might need to perform i2c transfers during
> > > early resume; your commit message should educate me.
> > yes, it's generic as of now since we have an internal usecase with PCI
> > is yet to be enabled in upstream. Not tied up with any usecase in
> > upstream, i just heard recently.
> >
> > Provided the scenario is generic and possible by any client, can this
> > code change be reviewed or shall be kept on halt till PCI usecase gets
> > enabled ?
>
> If this is a valid scenario in the upstream kernel, yes. If it solves
> a problem only manifesting itself based on a downstream design then
> you need to exactly describe that scenario so that reviewers can
> decide if this is a problem with the upstream kernel or your
> downstream design.
The exact scenario from PCIe switch is as below.  Hope it describes the 
usecase at high level to understand PCIe client -> PCIe Driver -> I2C 
driver.

PCIe switch needs certain configurations through i2c after powering on 
the switch, as part of suspend we are using suspend_noirq() to turn off 
the switch because some PCIe clients do some transfers till 
suspend_noirq() phase. And as part of resume_noirq(), we will enable the 
power to the switch and configures the switch again through i2c.


> Regards,
> Bjonr