From: Kevin Hilman <khilman@deeprootsystems.com>
To: Thara Gopinath <thara@ti.com>
Cc: linux-omap@vger.kernel.org, paul@pwsan.com, nm@ti.com,
b-cousson@ti.com, vishwanath.bs@ti.com, sawant@ti.com
Subject: Re: [PATCH 15/16] OMAP3: PM: VP force update method of voltage scaling
Date: Tue, 02 Mar 2010 16:58:37 -0800 [thread overview]
Message-ID: <87ljeap6f6.fsf@deeprootsystems.com> (raw)
In-Reply-To: <1267003757-22456-16-git-send-email-thara@ti.com> (Thara Gopinath's message of "Wed\, 24 Feb 2010 14\:59\:16 +0530")
Thara Gopinath <thara@ti.com> writes:
> This patch introduces VP force update method of voltage scaling
> and enables it by default. The older method of vc bypass is now
> configuratble through a menu config option. VP force update is the
> hardware recommended method of voltage scaling.
>
> Signed-off-by: Thara Gopinath <thara@ti.com>
This should be a runtime option intead of a compile-time option.
Default should be HW recommended method, but board code should be able
to override.
Only did a quick glance at this, an some minor coding style issues below...
> ---
> arch/arm/mach-omap2/voltage.c | 149 +++++++++++++++++++++++++++++++++++++++--
> arch/arm/mach-omap2/voltage.h | 1 +
> arch/arm/plat-omap/Kconfig | 19 +++++
> 3 files changed, 162 insertions(+), 7 deletions(-)
>
> diff --git a/arch/arm/mach-omap2/voltage.c b/arch/arm/mach-omap2/voltage.c
> index 49167c0..4f325af 100644
> --- a/arch/arm/mach-omap2/voltage.c
> +++ b/arch/arm/mach-omap2/voltage.c
> @@ -29,7 +29,8 @@
> #include "prm-regbits-34xx.h"
> #include "voltage.h"
>
> -#define VP_IDLE_TIMEOUT 200
> +#define VP_IDLE_TIMEOUT 200
> +#define VP_TRANXDONE_TIMEOUT 300
>
> /*
> * OMAP3 Voltage controller SR parameters. TODO: Pass this info as part of
> @@ -70,6 +71,7 @@ struct vp_reg_info {
> u32 vp_vddmin;
> u32 vp_vddmax;
> u32 vp_timeout;
> + u32 vp_tranxdone_status;
> };
> static struct vp_reg_info *vp_reg;
> /*
> @@ -299,6 +301,8 @@ static void __init vp_reg_configure(int vp_id)
> OMAP3430_VDDMIN_SHIFT);
> vp_reg[vp_id].vp_vddmax = (OMAP3_VP1_VLIMITTO_VDDMAX <<
> OMAP3430_VDDMAX_SHIFT);
> + vp_reg[vp_id].vp_tranxdone_status =
> + OMAP3430_VP1_TRANXDONE_ST;
> } else if (vp_id == VP2) {
> /*
> * OMAP3430 has error gain varying btw higher and
> @@ -312,6 +316,8 @@ static void __init vp_reg_configure(int vp_id)
> OMAP3430_VDDMIN_SHIFT);
> vp_reg[vp_id].vp_vddmax = (OMAP3_VP2_VLIMITTO_VDDMAX <<
> OMAP3430_VDDMAX_SHIFT);
> + vp_reg[vp_id].vp_tranxdone_status =
> + OMAP3430_VP2_TRANXDONE_ST;
> } else {
> pr_warning("Voltage processor%d does not exisit\
> in OMAP3 \n", vp_id);
> @@ -346,6 +352,131 @@ static void __init vp_reg_configure(int vp_id)
> /* TODO Extend this for OMAP4 ?? Or need a separate file */
> }
>
> +#ifdef CONFIG_OMAP_VOLT_VPFORCEUPDATE
> +/* VP force update method of voltage scaling */
> +static int vp_forceupdate_scale_voltage(u32 vdd, u8 target_vsel,
> + u8 current_vsel)
> +{
> + u32 smps_steps = 0, smps_delay = 0;
> + u32 vpconfig;
> + int timeout = 0;
> + int vp_id = vdd - 1;
> +
> + if (!((vdd == VDD1_OPP) || (vdd == VDD2_OPP))) {
> + pr_warning("Wrong vdd id passed to vp forceupdate\n");
> + return false;
> + }
> +
> + smps_steps = abs(target_vsel - current_vsel);
> +
> + /* OMAP3430 has errorgain varying btw higher and lower opp's */
> + if (cpu_is_omap34xx()) {
> + if (vdd == VDD1_OPP) {
> + u32 vc_cmdval0;
> +
> + vc_cmdval0 = voltage_read_reg(vc_reg.vc_cmdval0_reg);
> + vc_cmdval0 &= ~VC_CMD_ON_MASK;
> + vc_cmdval0 |= (target_vsel << VC_CMD_ON_SHIFT);
> + voltage_write_reg(vc_reg.vc_cmdval0_reg, vc_cmdval0);
> + vp_reg[vp_id].vp_errorgain = (((get_vdd1_opp() > 2) ?
> + (OMAP3_VP_CONFIG_ERRORGAIN_HIGHOPP) :
> + (OMAP3_VP_CONFIG_ERRORGAIN_LOWOPP)) <<
> + OMAP3430_ERRORGAIN_SHIFT);
> + } else if (vdd == VDD2_OPP) {
> + u32 vc_cmdval1;
> +
> + vc_cmdval1 = voltage_read_reg(vc_reg.vc_cmdval1_reg);
> + vc_cmdval1 &= ~VC_CMD_ON_MASK;
> + vc_cmdval1 |= (target_vsel << VC_CMD_ON_SHIFT);
> + voltage_write_reg(vc_reg.vc_cmdval1_reg, vc_cmdval1);
> + vp_reg[vp_id].vp_errorgain = (((get_vdd2_opp() > 2) ?
> + (OMAP3_VP_CONFIG_ERRORGAIN_HIGHOPP) :
> + (OMAP3_VP_CONFIG_ERRORGAIN_LOWOPP)) <<
> + OMAP3430_ERRORGAIN_SHIFT);
> + }
> + }
> +
> + /* Clear all pending TransactionDone interrupt/status. Typical latency
> + * is <3us
> + */
> + while (timeout++ < VP_TRANXDONE_TIMEOUT) {
> + voltage_write_reg(PRM_IRQSTATUS_REG,
> + vp_reg[vp_id].vp_tranxdone_status);
> + if (!(voltage_read_reg(PRM_IRQSTATUS_REG) &
> + vp_reg[vp_id].vp_tranxdone_status))
> + break;
> + udelay(1);
> + }
> +
> + if (timeout >= VP_TRANXDONE_TIMEOUT) {
> + pr_warning("VP%d TRANXDONE timeout exceeded. Voltage change \
> + aborted", vdd);
> + return false;
> + }
> +
> + /* Configure for VP-Force Update */
> + vpconfig = voltage_read_reg(vp_reg[vp_id].vp_offs.vp_vpconfig_reg);
> + vpconfig &= ~(VP_CONFIG_INITVDD | VP_FORCEUPDATE |
> + VP_INITVOLTAGE_MASK | VP_ERRORGAIN_MASK);
> + vpconfig |= ((target_vsel << VP_INITVOLTAGE_SHIFT) |
> + vp_reg[vp_id].vp_errorgain);
> + voltage_write_reg(vp_reg[vp_id].vp_offs.vp_vpconfig_reg, vpconfig);
> +
> + /* Trigger initVDD value copy to voltage processor */
> + vpconfig |= VP_CONFIG_INITVDD;
> + voltage_write_reg(vp_reg[vp_id].vp_offs.vp_vpconfig_reg, vpconfig);
> +
> + /* Force update of voltage */
> + vpconfig |= VP_FORCEUPDATE;
> + voltage_write_reg(vp_reg[vp_id].vp_offs.vp_vpconfig_reg, vpconfig);
> +
> + timeout = 0;
> + /* Wait for TransactionDone. Typical latency is <200us.
> + * Depends on SMPSWAITTIMEMIN/MAX and voltage change
> + */
multi-line comment style
> + while ((timeout++ < VP_TRANXDONE_TIMEOUT) &&
> + (!(voltage_read_reg(PRM_IRQSTATUS_REG) &
> + vp_reg[vp_id].vp_tranxdone_status)))
> + udelay(1);
> +
> + if (timeout >= VP_TRANXDONE_TIMEOUT)
> + pr_warning("VP%d TRANXDONE timeout exceeded. TRANXDONE never \
> + got set after the voltage update.Serious error!!!!\n",
> + vdd);
> +
> + /* Wait for voltage to settle with SW wait-loop */
> + smps_delay = ((smps_steps * 125) / 40) + 2;
> + udelay(smps_delay);
> +
> + /* Disable TransactionDone interrupt , clear all status, clear
> + * control registers
> + */
muti-line comment
> + timeout = 0;
> + while (timeout++ < VP_TRANXDONE_TIMEOUT) {
> + voltage_write_reg(PRM_IRQSTATUS_REG,
> + vp_reg[vp_id].vp_tranxdone_status);
> + if (!(voltage_read_reg(PRM_IRQSTATUS_REG) &
> + vp_reg[vp_id].vp_tranxdone_status))
> + break;
> + udelay(1);
> + }
> + if (timeout >= VP_TRANXDONE_TIMEOUT)
> + pr_warning("VP%d TRANXDONE timeout exceeded while trying to \
> + clear the TRANXDONE status\n", vdd);
> +
> + vpconfig = voltage_read_reg(vp_reg[vp_id].vp_offs.vp_vpconfig_reg);
> + /* Clear initVDD copy trigger bit */
> + vpconfig &= ~VP_CONFIG_INITVDD;
> + voltage_write_reg(vp_reg[vp_id].vp_offs.vp_vpconfig_reg, vpconfig);
> + /* Clear force bit */
> + vpconfig &= ~VP_FORCEUPDATE;
> + voltage_write_reg(vp_reg[vp_id].vp_offs.vp_vpconfig_reg, vpconfig);
> +
> + return true;
> +}
> +#endif
> +
> +#ifdef CONFIG_OMAP_VOLT_VCBYPASS
> /**
> * vc_bypass_scale_voltage - VC bypass method of voltage scaling
> */
> @@ -437,7 +568,7 @@ static int vc_bypass_scale_voltage(u32 vdd, u8 target_vsel, u8 current_vsel)
> udelay(smps_delay);
> return true;
> }
> -
> +#endif
>
> static void __init init_voltageprocessors(void)
> {
> @@ -474,13 +605,16 @@ void omap_voltageprocessor_enable(int vp_id)
> if (voltage_read_reg(vp_reg[vp_id - 1].vp_offs.vp_vpconfig_reg) &
> VP_CONFIG_VPENABLE)
> return;
> +
> +#ifdef CONFIG_OMAP_VOLT_VCBYPASS
> /*
> * This latching is required only if VC bypass method is used for
> * voltage scaling during dvfs.
> */
> vp_latch_vsel(vp_id - 1);
> - vpconfig = voltage_read_reg(vp_reg[vp_id - 1].vp_offs.vp_vpconfig_reg);
> +#endif
> /* Enable VP */
> + vpconfig = voltage_read_reg(vp_reg[vp_id - 1].vp_offs.vp_vpconfig_reg);
> voltage_write_reg(vp_reg[vp_id - 1].vp_offs.vp_vpconfig_reg,
> vpconfig | VP_CONFIG_VPENABLE);
> }
> @@ -529,11 +663,12 @@ void omap_voltageprocessor_disable(int vp_id)
> * for a particular voltage domain during dvfs or any other situation.
> */
> int omap_voltage_scale(int vdd, u8 target_vsel, u8 current_vsel)
> -{ /*
> - * TODO add VP force update method of voltage scaling
> - * and choose btw the two
> - */
> +{
> +#ifdef CONFIG_OMAP_VOLT_VCBYPASS
> return vc_bypass_scale_voltage(vdd, target_vsel, current_vsel);
> +#elif CONFIG_OMAP_VOLT_VPFORCEUPDATE
> + return vp_forceupdate_scale_voltage(vdd, target_vsel, current_vsel);
> +#endif
> }
>
> /**
> diff --git a/arch/arm/mach-omap2/voltage.h b/arch/arm/mach-omap2/voltage.h
> index c3203c9..615bde6 100644
> --- a/arch/arm/mach-omap2/voltage.h
> +++ b/arch/arm/mach-omap2/voltage.h
> @@ -17,6 +17,7 @@
> extern int get_vdd1_opp(void);
> extern int get_vdd2_opp(void);
>
> +#define PRM_IRQSTATUS_REG OMAP3430_PRM_IRQSTATUS_MPU
> /* Generic VP definitions. Need to be redefined for OMAP4 */
> #define VP_CONFIG_TIMEOUTEN OMAP3430_TIMEOUTEN
> #define VP_CONFIG_INITVDD OMAP3430_INITVDD
> diff --git a/arch/arm/plat-omap/Kconfig b/arch/arm/plat-omap/Kconfig
> index 9d286e6..1a611a9 100644
> --- a/arch/arm/plat-omap/Kconfig
> +++ b/arch/arm/plat-omap/Kconfig
> @@ -92,6 +92,25 @@ config OMAP_SMARTREFLEX_TESTING
>
> WARNING: Enabling this option may cause your device to hang!
>
> +choice
> + prompt "Choose Voltage Scale method"
> + depends on ARCH_OMAP3 && PM
> + default OMAP_VOLT_VPFORCEUPDATE
> +
> +config OMAP_VOLT_VPFORCEUPDATE
> + bool "Voltage scaling via VP force update method"
> + help
> + Say Y if you want to enable VP force update method
> + of voltage scaling. This is the h/w recomended way
> + of voltage scaling in OMAP3.
> +
> +config OMAP_VOLT_VCBYPASS
> + bool "Voltage Scale via Voltage controller in bypass"
> + help
> + Say Y if you want to enable VC Bypass method of voltage scaling.
> + Not the default recommended method.
> +endchoice
> +
> config OMAP_RESET_CLOCKS
> bool "Reset unused clocks during boot"
> depends on ARCH_OMAP
> --
> 1.7.0.rc1.33.g07cf0f
next prev parent reply other threads:[~2010-03-03 0:58 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-02-24 9:29 [PATCH 00/16] OMAP3: PM: Smartreflex and voltage revamp Thara Gopinath
2010-02-24 9:29 ` [PATCH 01/16] OMAP3: PM: Adding hwmod data for Smartreflex Thara Gopinath
2010-02-24 9:29 ` [PATCH 02/16] OMAP3: PM: Create list to keep track of various smartreflex instances Thara Gopinath
2010-02-24 9:29 ` [PATCH 03/16] OMAP3: PM: Convert smartreflex driver into a platform driver using hwmods and omap-device layer Thara Gopinath
2010-02-24 9:29 ` [PATCH 04/16] OMAP3: PM: Move smartreflex autocompensation enable disable hooks to PM debugfs Thara Gopinath
2010-02-24 9:29 ` [PATCH 05/16] OMAP3: PM: Export get_vdd1_opp and get_vdd2_opp from shared resource framework Thara Gopinath
2010-02-24 9:29 ` [PATCH 06/16] OMAP3: PM: Smartreflex class related changes for smartreflex.c Thara Gopinath
2010-02-24 9:29 ` [PATCH 07/16] OMAP3: PM: Adding smartreflex class 3 driver Thara Gopinath
2010-02-24 9:29 ` [PATCH 08/16] OMAP3: PM: Disabling Smartreflex across both frequency and voltage scaling during DVFS Thara Gopinath
2010-02-24 9:29 ` [PATCH 09/16] OMAP3: PM: Creating separate files for handling OMAP3 voltage related operations Thara Gopinath
2010-02-24 9:29 ` [PATCH 10/16] OMAP3: PM: Cleaning up of smartreflex header file Thara Gopinath
2010-02-24 9:29 ` [PATCH 11/16] OMAP3: PM: Configurations for Smartreflex Class 2 and Smartreflex Class 3 Thara Gopinath
2010-02-24 9:29 ` [PATCH 12/16] OMAP3: PM: Support for enabling smartreflex autocompensation by default Thara Gopinath
2010-02-24 9:29 ` [PATCH 13/16] OMAP3: PM: Correcting accessing of ERRCONFIG register in smartreflex.c Thara Gopinath
2010-02-24 9:29 ` [PATCH 14/16] OMAP3: PM: Implement latest h/w recommendations for SR and VP registers and SR VP enable disable sequence Thara Gopinath
2010-02-24 9:29 ` [PATCH 15/16] OMAP3: PM: VP force update method of voltage scaling Thara Gopinath
2010-02-24 9:29 ` [PATCH 16/16] OMAP3: PM: Enabling Smartreflex Class 3 driver by default in pm defconfig Thara Gopinath
2010-03-03 0:58 ` Kevin Hilman [this message]
2010-03-05 15:22 ` [PATCH 15/16] OMAP3: PM: VP force update method of voltage scaling Gopinath, Thara
2010-03-05 15:26 ` Felipe Contreras
2010-03-05 15:30 ` Gopinath, Thara
2010-03-05 16:22 ` Felipe Contreras
2010-03-05 18:17 ` Snipping irrelevant text from a discussion (was: "RE: [PATCH 15/16] OMAP3: PM: VP force update method of voltage scaling") Aguirre, Sergio
2010-03-05 21:18 ` Felipe Contreras
2010-03-09 1:42 ` Tony Lindgren
2010-03-09 6:51 ` Felipe Balbi
2010-03-09 15:21 ` Aguirre, Sergio
2010-03-09 18:21 ` Tony Lindgren
2010-03-03 0:54 ` [PATCH 14/16] OMAP3: PM: Implement latest h/w recommendations for SR and VP registers and SR VP enable disable sequence Kevin Hilman
2010-03-03 0:48 ` [PATCH 12/16] OMAP3: PM: Support for enabling smartreflex autocompensation by default Kevin Hilman
2010-03-05 15:20 ` Gopinath, Thara
2010-03-03 0:37 ` [PATCH 11/16] OMAP3: PM: Configurations for Smartreflex Class 2 and Smartreflex Class 3 Kevin Hilman
2010-03-05 15:12 ` Gopinath, Thara
2010-03-05 19:20 ` Kevin Hilman
2010-03-02 20:02 ` [PATCH 09/16] OMAP3: PM: Creating separate files for handling OMAP3 voltage related operations Kevin Hilman
2010-03-05 15:17 ` Gopinath, Thara
2010-03-02 19:36 ` [PATCH 07/16] OMAP3: PM: Adding smartreflex class 3 driver Kevin Hilman
2010-03-05 15:03 ` Gopinath, Thara
2010-03-05 19:12 ` Kevin Hilman
2010-03-02 18:44 ` [PATCH 06/16] OMAP3: PM: Smartreflex class related changes for smartreflex.c Kevin Hilman
2010-03-05 15:00 ` Gopinath, Thara
2010-03-05 18:29 ` Kevin Hilman
2010-03-02 23:37 ` Kevin Hilman
2010-03-02 23:52 ` Kevin Hilman
2010-03-05 15:18 ` Gopinath, Thara
2010-03-05 18:30 ` Kevin Hilman
2010-03-02 18:28 ` [PATCH 04/16] OMAP3: PM: Move smartreflex autocompensation enable disable hooks to PM debugfs Kevin Hilman
2010-02-25 2:39 ` [PATCH 03/16] OMAP3: PM: Convert smartreflex driver into a platform driver using hwmods and omap-device layer ambresh
2010-03-02 18:28 ` Kevin Hilman
2010-03-05 14:26 ` Gopinath, Thara
2010-03-12 9:48 ` Gopinath, Thara
2010-03-13 0:36 ` Kevin Hilman
2010-03-15 19:00 ` Tony Lindgren
2010-03-03 0:02 ` Kevin Hilman
2010-02-25 1:42 ` [PATCH 02/16] OMAP3: PM: Create list to keep track of various smartreflex instances ambresh
2010-03-02 17:40 ` Kevin Hilman
2010-02-26 23:21 ` Mike Turquette
2010-03-02 17:39 ` Kevin Hilman
2010-02-24 16:52 ` [PATCH 01/16] OMAP3: PM: Adding hwmod data for Smartreflex Mike Turquette
2010-03-06 0:45 ` Kevin Hilman
2010-03-06 0:58 ` [PATCH 00/16] OMAP3: PM: Smartreflex and voltage revamp Kevin Hilman
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87ljeap6f6.fsf@deeprootsystems.com \
--to=khilman@deeprootsystems.com \
--cc=b-cousson@ti.com \
--cc=linux-omap@vger.kernel.org \
--cc=nm@ti.com \
--cc=paul@pwsan.com \
--cc=sawant@ti.com \
--cc=thara@ti.com \
--cc=vishwanath.bs@ti.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox