On 11/30/17 2:56 PM, Dave Gerlach wrote:
> This is a resend of v5 of this series found here [1]. It introduces
> relocatable PM handlers for the emif that are copied to sram and
> run from there during low power mode entry.
>
> The patches still have the previous ACKs but have a small change to
> accomodate a change made by Tony in commit cd57dc5a2099 ("ARM: dts:
> Add missing hwmod related nodes for am33xx"). If there are objections
> to this let me know ASAP.
>
> Now that a hwmod is present for the am335x EMIF, on probe fail the call to
> pm_runtime_put_sync causes the board to hang. In fact, this emif driver should
> never alter the PM state of the hardware at all through normal kernel calls, it
> is the job of the suspend handlers that are added, that is the whole point of
> this driver. Because of this, I have dropped all runtime pm calls, as any
> change to the PM state while the kernel is running is dangerous as we may shut
> of the memory controller. It makes the most sense just to drop runtime PM from
> the driver entirely. Besides that patch is unchanged.
>
> This code is required for low-power modes to work on AM335x and AM437x and a
> forthcoming PM series for those platforms will depend on this series. After
> both this and the PM series are reviewed I will send the necessary device tree
> changes for both, but in the meantime all remaining patches for am335x and
> am437x PM can be found here [2].
>
Applied
* [email protected] <[email protected]> [171203 04:03]:
>
>
> On 11/30/17 2:56 PM, Dave Gerlach wrote:
> > This is a resend of v5 of this series found here [1]. It introduces
> > relocatable PM handlers for the emif that are copied to sram and
> > run from there during low power mode entry.
> >
> > The patches still have the previous ACKs but have a small change to
> > accomodate a change made by Tony in commit cd57dc5a2099 ("ARM: dts:
> > Add missing hwmod related nodes for am33xx"). If there are objections
> > to this let me know ASAP.
> >
> > Now that a hwmod is present for the am335x EMIF, on probe fail the call to
> > pm_runtime_put_sync causes the board to hang. In fact, this emif driver should
> > never alter the PM state of the hardware at all through normal kernel calls, it
> > is the job of the suspend handlers that are added, that is the whole point of
> > this driver. Because of this, I have dropped all runtime pm calls, as any
> > change to the PM state while the kernel is running is dangerous as we may shut
> > of the memory controller. It makes the most sense just to drop runtime PM from
> > the driver entirely. Besides that patch is unchanged.
> >
> > This code is required for low-power modes to work on AM335x and AM437x and a
> > forthcoming PM series for those platforms will depend on this series. After
> > both this and the PM series are reviewed I will send the necessary device tree
> > changes for both, but in the meantime all remaining patches for am335x and
> > am437x PM can be found here [2].
> >
> Applied
OK, do you have some immutable commit for these I can pull in too for
the related SoC changes?
Regards,
Tony
On 12/3/2017 8:36 AM, Tony Lindgren wrote:
> * [email protected] <[email protected]> [171203 04:03]:
>>
>>
>> On 11/30/17 2:56 PM, Dave Gerlach wrote:
>>> This is a resend of v5 of this series found here [1]. It introduces
>>> relocatable PM handlers for the emif that are copied to sram and
>>> run from there during low power mode entry.
>>>
>>> The patches still have the previous ACKs but have a small change to
>>> accomodate a change made by Tony in commit cd57dc5a2099 ("ARM: dts:
>>> Add missing hwmod related nodes for am33xx"). If there are objections
>>> to this let me know ASAP.
>>>
>>> Now that a hwmod is present for the am335x EMIF, on probe fail the call to
>>> pm_runtime_put_sync causes the board to hang. In fact, this emif driver should
>>> never alter the PM state of the hardware at all through normal kernel calls, it
>>> is the job of the suspend handlers that are added, that is the whole point of
>>> this driver. Because of this, I have dropped all runtime pm calls, as any
>>> change to the PM state while the kernel is running is dangerous as we may shut
>>> of the memory controller. It makes the most sense just to drop runtime PM from
>>> the driver entirely. Besides that patch is unchanged.
>>>
>>> This code is required for low-power modes to work on AM335x and AM437x and a
>>> forthcoming PM series for those platforms will depend on this series. After
>>> both this and the PM series are reviewed I will send the necessary device tree
>>> changes for both, but in the meantime all remaining patches for am335x and
>>> am437x PM can be found here [2].
>>>
>> Applied
>
> OK, do you have some immutable commit for these I can pull in too for
> the related SoC changes?
>
Hopefully I don't have do any changes to below branch.
git://git.kernel.org/pub/scm/linux/kernel/git/ssantosh/linux-keystone.git
for_4.16/drivers-soc