Message ID | 20190912113916.20093-1-t-kristo@ti.com |
---|---|
Headers | show |
Series | soc: ti: add OMAP PRM driver (for reset) | expand |
* Tero Kristo <t-kristo@ti.com> [190912 11:40]: > @@ -565,6 +566,12 @@ void omap_pcs_legacy_init(int irq, void (*rearm)(void)) > pcs_pdata.rearm = rearm; > } > > +static struct ti_prm_platform_data ti_prm_pdata = { > + .clkdm_deny_idle = clkdm_deny_idle, > + .clkdm_allow_idle = clkdm_allow_idle, > + .clkdm_lookup = clkdm_lookup, > +}; > + > /* > * GPIOs for TWL are initialized by the I2C bus and need custom > * handing until DSS has device tree bindings. > @@ -664,6 +671,11 @@ static struct of_dev_auxdata omap_auxdata_lookup[] = { > /* Common auxdata */ > OF_DEV_AUXDATA("ti,sysc", 0, NULL, &ti_sysc_pdata), > OF_DEV_AUXDATA("pinctrl-single", 0, NULL, &pcs_pdata), > + OF_DEV_AUXDATA("ti,omap4-prm-inst", 0, NULL, &ti_prm_pdata), > + OF_DEV_AUXDATA("ti,omap5-prm-inst", 0, NULL, &ti_prm_pdata), > + OF_DEV_AUXDATA("ti,dra7-prm-inst", 0, NULL, &ti_prm_pdata), > + OF_DEV_AUXDATA("ti,am3-prm-inst", 0, NULL, &ti_prm_pdata), > + OF_DEV_AUXDATA("ti,am4-prm-inst", 0, NULL, &ti_prm_pdata), > { /* sentinel */ }, > }; Hmm I think I already commented on this.. Just one entry please: OF_DEV_AUXDATA("ti,omap-prm-inst", 0, NULL, &ti_prm_pdata), As the auxdata is the same for all of them. Note that all the dts files need to have also the generic compatible "ti,omap-prm-inst" after the SoC specific one. Regards, Tony
On 12/09/2019 20:09, Tony Lindgren wrote: > * Tero Kristo <t-kristo@ti.com> [190912 11:40]: >> @@ -565,6 +566,12 @@ void omap_pcs_legacy_init(int irq, void (*rearm)(void)) >> pcs_pdata.rearm = rearm; >> } >> >> +static struct ti_prm_platform_data ti_prm_pdata = { >> + .clkdm_deny_idle = clkdm_deny_idle, >> + .clkdm_allow_idle = clkdm_allow_idle, >> + .clkdm_lookup = clkdm_lookup, >> +}; >> + >> /* >> * GPIOs for TWL are initialized by the I2C bus and need custom >> * handing until DSS has device tree bindings. >> @@ -664,6 +671,11 @@ static struct of_dev_auxdata omap_auxdata_lookup[] = { >> /* Common auxdata */ >> OF_DEV_AUXDATA("ti,sysc", 0, NULL, &ti_sysc_pdata), >> OF_DEV_AUXDATA("pinctrl-single", 0, NULL, &pcs_pdata), >> + OF_DEV_AUXDATA("ti,omap4-prm-inst", 0, NULL, &ti_prm_pdata), >> + OF_DEV_AUXDATA("ti,omap5-prm-inst", 0, NULL, &ti_prm_pdata), >> + OF_DEV_AUXDATA("ti,dra7-prm-inst", 0, NULL, &ti_prm_pdata), >> + OF_DEV_AUXDATA("ti,am3-prm-inst", 0, NULL, &ti_prm_pdata), >> + OF_DEV_AUXDATA("ti,am4-prm-inst", 0, NULL, &ti_prm_pdata), >> { /* sentinel */ }, >> }; > > Hmm I think I already commented on this.. Just one entry please: > > OF_DEV_AUXDATA("ti,omap-prm-inst", 0, NULL, &ti_prm_pdata), > > As the auxdata is the same for all of them. Note that all the > dts files need to have also the generic compatible > "ti,omap-prm-inst" after the SoC specific one. Ok that should be fine, sorry for missing it out. I can update this in the next rev, just need to update the dt binding also. -Tero -- Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
* Tero Kristo <t-kristo@ti.com> [190919 12:33]: > The parent clockdomain for reset must be in force wakeup mode, otherwise > the reset may never complete. Add pdata quirks for this purpose for PRM > driver. > > Signed-off-by: Tero Kristo <t-kristo@ti.com> > --- > v6: replaced platform specific compatibles with a single common > compatible check Thanks looks good to me: Acked-by: Tony Lindgren <tony@atomide.com>
* Tero Kristo <t-kristo@ti.com> [190912 04:39]: > Hi, > > V5 of the series, re-sent the whole series as one patch was dropped. > Changes compared to v3/v4: > > - removed dependency towards clock driver (patch #5 was completely > dropped compared to v3/v4) > - dropped clocks property from dt binding > - re-added the pdata patch which was accidentally dropped out (it has > dependency towards this series.) > > The new implementation (without clock driver dependency) relies on the > bus driver to sequence events properly, otherwise some timeouts will > occur either at clock driver or reset driver end. With the two updated patches seems like we're done with this series? If so, I suggest either Santosh or me sets up an immutable branch against v5.3 or v5.4-rc1 that we all can merge in. I will need it for the related dts changes at least. Regards, Tony
On 9/20/19 7:28 AM, Tony Lindgren wrote: > * Tero Kristo <t-kristo@ti.com> [190912 04:39]: >> Hi, >> >> V5 of the series, re-sent the whole series as one patch was dropped. >> Changes compared to v3/v4: >> >> - removed dependency towards clock driver (patch #5 was completely >> dropped compared to v3/v4) >> - dropped clocks property from dt binding >> - re-added the pdata patch which was accidentally dropped out (it has >> dependency towards this series.) >> >> The new implementation (without clock driver dependency) relies on the >> bus driver to sequence events properly, otherwise some timeouts will >> occur either at clock driver or reset driver end. > > With the two updated patches seems like we're done with this > series? > > If so, I suggest either Santosh or me sets up an immutable > branch against v5.3 or v5.4-rc1 that we all can merge in. > I will need it for the related dts changes at least. > I will pick this up Tony and apply it once v5.4-rc1 is out. Regards, Santosh
* santosh.shilimkar@oracle.com <santosh.shilimkar@oracle.com> [190920 19:20]: > On 9/20/19 7:28 AM, Tony Lindgren wrote: > > * Tero Kristo <t-kristo@ti.com> [190912 04:39]: > > > Hi, > > > > > > V5 of the series, re-sent the whole series as one patch was dropped. > > > Changes compared to v3/v4: > > > > > > - removed dependency towards clock driver (patch #5 was completely > > > dropped compared to v3/v4) > > > - dropped clocks property from dt binding > > > - re-added the pdata patch which was accidentally dropped out (it has > > > dependency towards this series.) > > > > > > The new implementation (without clock driver dependency) relies on the > > > bus driver to sequence events properly, otherwise some timeouts will > > > occur either at clock driver or reset driver end. > > > > With the two updated patches seems like we're done with this > > series? > > > > If so, I suggest either Santosh or me sets up an immutable > > branch against v5.3 or v5.4-rc1 that we all can merge in. > > I will need it for the related dts changes at least. > > > I will pick this up Tony and apply it once v5.4-rc1 is out. OK thanks. Tony