| Message ID | 20190223093232.29003-3-lokeshvutla@ti.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | soc: ti: k3-am654: Allow for exclsive request of devices | expand |
On 15:02-20190223, Lokesh Vutla wrote: > TISCI protocol supports for enabling the device with exclusive > permissions. Certain remoteproc devices or some shared devices > across VM doesn't wants to request devices with this flag set. > So add support for getting this information from DT by increasing > the power-domain cells to 2. > > Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com> > --- > Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt | 7 +++++-- > 1 file changed, 5 insertions(+), 2 deletions(-) > > diff --git a/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt b/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt > index f7b00a7c0f68..5fdda7475023 100644 > --- a/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt > +++ b/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt > @@ -19,8 +19,11 @@ child of the pmmc node. > Required Properties: > -------------------- > - compatible: should be "ti,sci-pm-domain" > -- #power-domain-cells: Must be 1 so that an id can be provided in each > - device node. > +- #power-domain-cells: Can be one of the following: > + 1: Containing the device id of each node > + 2: First entry should be device id > + Second entry should be 1 or 0. Use 1 for enabling > + the device with exclusive permissions set else 0. Rob, any thoughts?
On Sat, Feb 23, 2019 at 03:02:29PM +0530, Lokesh Vutla wrote: > TISCI protocol supports for enabling the device with exclusive > permissions. Certain remoteproc devices or some shared devices > across VM doesn't wants to request devices with this flag set. The wording here is weird. You want to add exclusive support, but you are talking about devices which don't want exclusive support. > So add support for getting this information from DT by increasing > the power-domain cells to 2. Seems like this would be racy. Whomever wants exclusive access would have to be the first to request a device. Seems like whether devices are able to be shared or not would be known because generally h/w has to be designed for that. > > Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com> > --- > Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt | 7 +++++-- > 1 file changed, 5 insertions(+), 2 deletions(-) > > diff --git a/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt b/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt > index f7b00a7c0f68..5fdda7475023 100644 > --- a/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt > +++ b/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt > @@ -19,8 +19,11 @@ child of the pmmc node. > Required Properties: > -------------------- > - compatible: should be "ti,sci-pm-domain" > -- #power-domain-cells: Must be 1 so that an id can be provided in each > - device node. > +- #power-domain-cells: Can be one of the following: > + 1: Containing the device id of each node > + 2: First entry should be device id > + Second entry should be 1 or 0. Use 1 for enabling > + the device with exclusive permissions set else 0. > > Example (K2G): > ------------- > -- > 2.19.2 >
On 10:43-20190312, Rob Herring wrote: > On Sat, Feb 23, 2019 at 03:02:29PM +0530, Lokesh Vutla wrote: > > TISCI protocol supports for enabling the device with exclusive > > permissions. Certain remoteproc devices or some shared devices > > across VM doesn't wants to request devices with this flag set. > > The wording here is weird. You want to add exclusive support, but you > are talking about devices which don't want exclusive support. > > > So add support for getting this information from DT by increasing > > the power-domain cells to 2. > > Seems like this would be racy. Whomever wants exclusive access would > have to be the first to request a device. Seems like whether devices are > able to be shared or not would be known because generally h/w has to be > designed for that. Rob, I wonder if it is better to represent this as exclusive power domain or shared power domain - because that is what the firmware in effect makes OS think the hardware is? Two power domain nodes in dts, instead of introducing property as part of cells? Device may belong to a shared domain or an exclusive domain depending on the SoC variations.
diff --git a/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt b/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt index f7b00a7c0f68..5fdda7475023 100644 --- a/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt +++ b/Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt @@ -19,8 +19,11 @@ child of the pmmc node. Required Properties: -------------------- - compatible: should be "ti,sci-pm-domain" -- #power-domain-cells: Must be 1 so that an id can be provided in each - device node. +- #power-domain-cells: Can be one of the following: + 1: Containing the device id of each node + 2: First entry should be device id + Second entry should be 1 or 0. Use 1 for enabling + the device with exclusive permissions set else 0. Example (K2G): -------------
TISCI protocol supports for enabling the device with exclusive permissions. Certain remoteproc devices or some shared devices across VM doesn't wants to request devices with this flag set. So add support for getting this information from DT by increasing the power-domain cells to 2. Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com> --- Documentation/devicetree/bindings/soc/ti/sci-pm-domain.txt | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)