RE: [PATCH 2/3] OMAP3: PM: Configure PRM setup times from board files
-Original Message- From: Kevin Hilman [mailto:khil...@deeprootsystems.com] Sent: Wednesday, October 14, 2009 1:05 AM To: Nayak, Rajendra Cc: linux-omap@vger.kernel.org Subject: Re: [PATCH 2/3] OMAP3: PM: Configure PRM setup times from board files Rajendra Nayak rna...@ti.com writes: The setup times to be programmed in the PRM module on OMAP (for clksetup, voltsetup etc) are board specific. They depend heavily on the PMIC used and even on different boards with the same PMIC, they vary based on the sleep/wake sequence used, system clock speed et al. This patch makes it possible for these setup values to be configured from different board files. Signed-off-by: Rajendra Nayak rna...@ti.com [...] diff --git a/arch/arm/mach-omap2/pm34xx.c b/arch/arm/mach-omap2/pm34xx.c index 2242d23..6f2fb51 100644 --- a/arch/arm/mach-omap2/pm34xx.c +++ b/arch/arm/mach-omap2/pm34xx.c @@ -1084,6 +1084,9 @@ int omap3_pm_set_suspend_state(struct powerdomain *pwrdm, int state) void omap3_set_prm_setup_vc(struct prm_setup_vc *setup_vc) { +if (!setup_vc) +return; + prm_setup.clksetup = setup_vc-clksetup; prm_setup.voltsetup_time1 = setup_vc-voltsetup_time1; prm_setup.voltsetup_time2 = setup_vc-voltsetup_time2; @@ -1285,13 +1288,15 @@ static void __init configure_vc(void) void omap3_pm_early_init(struct omap_opp *mpu_opps, struct omap_opp *dsp_opps, - struct omap_opp *l3_opps) + struct omap_opp *l3_opps, + struct prm_setup_vc *setup_times) { prm_clear_mod_reg_bits(OMAP3430_OFFMODE_POL, OMAP3430_GR_MOD, OMAP3_PRM_POLCTRL_OFFSET); configure_vc(); omap_pm_if_early_init(mpu_opps, dsp_opps, l3_opps); +omap3_set_prm_setup_vc(setup_times); I think there's an ordering problem here since the configure_vc() which uses the values is called before the setup_vc(). Yes, the ordering seems to be certainly wrong. I will fix that in the updated patch-set. Also, if setup_times is passed as NULL, configure_vc() will be writing wrong values with undefined results to the VC. I guess not, since it uses default values from the prm_setup table defined in pm34xx.c which are not optimal but certainally not wrong. No? Same is the case with cpuidle params as well. Also, if my proposed omap3_pm_init_vc() approach is taken and it is optional, some sort of defaults should probably be set. Kevin -- To unsubscribe from this list: send the line unsubscribe linux-omap in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 2/3] OMAP3: PM: Configure PRM setup times from board files
Nayak, Rajendra rna...@ti.com writes: [...] diff --git a/arch/arm/mach-omap2/pm34xx.c b/arch/arm/mach-omap2/pm34xx.c index 2242d23..6f2fb51 100644 --- a/arch/arm/mach-omap2/pm34xx.c +++ b/arch/arm/mach-omap2/pm34xx.c @@ -1084,6 +1084,9 @@ int omap3_pm_set_suspend_state(struct powerdomain *pwrdm, int state) void omap3_set_prm_setup_vc(struct prm_setup_vc *setup_vc) { + if (!setup_vc) + return; + prm_setup.clksetup = setup_vc-clksetup; prm_setup.voltsetup_time1 = setup_vc-voltsetup_time1; prm_setup.voltsetup_time2 = setup_vc-voltsetup_time2; @@ -1285,13 +1288,15 @@ static void __init configure_vc(void) void omap3_pm_early_init(struct omap_opp *mpu_opps, struct omap_opp *dsp_opps, -struct omap_opp *l3_opps) +struct omap_opp *l3_opps, +struct prm_setup_vc *setup_times) { prm_clear_mod_reg_bits(OMAP3430_OFFMODE_POL, OMAP3430_GR_MOD, OMAP3_PRM_POLCTRL_OFFSET); configure_vc(); omap_pm_if_early_init(mpu_opps, dsp_opps, l3_opps); + omap3_set_prm_setup_vc(setup_times); I think there's an ordering problem here since the configure_vc() which uses the values is called before the setup_vc(). Yes, the ordering seems to be certainly wrong. I will fix that in the updated patch-set. Also, if setup_times is passed as NULL, configure_vc() will be writing wrong values with undefined results to the VC. I guess not, since it uses default values from the prm_setup table defined in pm34xx.c which are not optimal but certainally not wrong. No? Same is the case with cpuidle params as well. Ah, you're right. I saw the defaults in cpuidle, but not in pm34xx.c. That should be fine. Then, having these defaults will make the init call optional and only needed in boards that want to override. Thanks, Kevin -- To unsubscribe from this list: send the line unsubscribe linux-omap in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 2/3] OMAP3: PM: Configure PRM setup times from board files
Rajendra Nayak rna...@ti.com writes: The setup times to be programmed in the PRM module on OMAP (for clksetup, voltsetup etc) are board specific. They depend heavily on the PMIC used and even on different boards with the same PMIC, they vary based on the sleep/wake sequence used, system clock speed et al. This patch makes it possible for these setup values to be configured from different board files. Signed-off-by: Rajendra Nayak rna...@ti.com [...] diff --git a/arch/arm/mach-omap2/pm34xx.c b/arch/arm/mach-omap2/pm34xx.c index 2242d23..6f2fb51 100644 --- a/arch/arm/mach-omap2/pm34xx.c +++ b/arch/arm/mach-omap2/pm34xx.c @@ -1084,6 +1084,9 @@ int omap3_pm_set_suspend_state(struct powerdomain *pwrdm, int state) void omap3_set_prm_setup_vc(struct prm_setup_vc *setup_vc) { + if (!setup_vc) + return; + prm_setup.clksetup = setup_vc-clksetup; prm_setup.voltsetup_time1 = setup_vc-voltsetup_time1; prm_setup.voltsetup_time2 = setup_vc-voltsetup_time2; @@ -1285,13 +1288,15 @@ static void __init configure_vc(void) void omap3_pm_early_init(struct omap_opp *mpu_opps, struct omap_opp *dsp_opps, - struct omap_opp *l3_opps) + struct omap_opp *l3_opps, + struct prm_setup_vc *setup_times) { prm_clear_mod_reg_bits(OMAP3430_OFFMODE_POL, OMAP3430_GR_MOD, OMAP3_PRM_POLCTRL_OFFSET); configure_vc(); omap_pm_if_early_init(mpu_opps, dsp_opps, l3_opps); + omap3_set_prm_setup_vc(setup_times); I think there's an ordering problem here since the configure_vc() which uses the values is called before the setup_vc(). Also, if setup_times is passed as NULL, configure_vc() will be writing wrong values with undefined results to the VC. Also, if my proposed omap3_pm_init_vc() approach is taken and it is optional, some sort of defaults should probably be set. Kevin -- To unsubscribe from this list: send the line unsubscribe linux-omap in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html