From: David Brownell Subject: Re: [patch 2.6.29-rc3-git 1/2] regulator: twl4030 regulators Date: Mon, 23 Feb 2009 23:25:10 -0800 Message-ID: <200902232325.10447.david-b@pacbell.net> References: <200902081037.06645.david-b@pacbell.net> <20090224005536.GC3601@sirena.org.uk> <200902231822.27422.david-b@pacbell.net> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Received: from smtp115.sbc.mail.sp1.yahoo.com ([69.147.64.88]:45950 "HELO smtp115.sbc.mail.sp1.yahoo.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1751852AbZBXHZO (ORCPT ); Tue, 24 Feb 2009 02:25:14 -0500 In-Reply-To: <200902231822.27422.david-b@pacbell.net> Content-Disposition: inline Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: Mark Brown Cc: Liam Girdwood , lkml , OMAP , Pierre Ossman On Monday 23 February 2009, David Brownell wrote: > > There's also fun and games to be had with accuracy once you=20 > > start looking too closely at the discrete voltages. >=20 > Yes; the patch I sent is explicitly making those available. >=20 > But I ignored issues like "+/- 3% accurate output" for LDOs > (or switchers) ... if anyone really needs to address them, > patches will be needed. =A0For now I only care that a 3.1 Volt > output can match both MMC_VDD_30_31 and MMC_VDD_31_32! ;) And -- for kicks -- here's one notion of what it might look like to have the MMC stack support the regulator framework. - Dave =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Prototype glue between MMC and regulator stacks ... compiles, and mmc_regulator_get_ocrmask() passed sanity testing. NOTES: - The MMC core does't call mmc_regulator_set_ocr() because hosts may need to do that in conjunction with updating I/O voltage. Case in point, MMC1 on omap_hsmmc ... where the host driver must update MMC1_HCTL.SDVS and PBIAS registers in addition to the regulator, supporting 1.8V or 3.0V voltage ranges. (MMC2 and MMC3 use external level shifting for Vdd !=3D 1.8V.) Likewise, using eMMC "managed NAND" solutions, powerup includes not both Vcc ("vdd" to Linux, e.g. 3.0V) and an I/O interface rail VccQ (e.g. 1.8V). The JEDEC spec for eMMC requires VccQ powerup after Vcc, and powerdown before it. - The "vdd" supply name isn't fixed, since platforms may need to use more than one I/O supply. Case in point, MMC1 on omap_hsmmc (again) ... where a second supply is needed to kick in 8-bit I/O using DAT4..DAT7 signals. That would not be handled quite like VccQ, since it's only used for 8-bit I/O widths (MMCplus cards, some eMMC, etc). =20 --- drivers/mmc/core/Kconfig | 8 +++ drivers/mmc/core/core.c | 98 ++++++++++++++++++++++++++++++++++++++= +++++++ include/linux/mmc/host.h | 3 + 3 files changed, 109 insertions(+) --- a/drivers/mmc/core/Kconfig +++ b/drivers/mmc/core/Kconfig @@ -14,3 +14,11 @@ config MMC_UNSAFE_RESUME This option is usually just for embedded systems which use a MMC/SD card for rootfs. Most people should say N here. =20 +config MMC_REGULATOR + bool + depends on REGULATOR + default y + help + Select this to provide some helper utilities to access the + "vdd" (card) voltage supply associated with an MMC/SD slot. + --- a/drivers/mmc/core/core.c +++ b/drivers/mmc/core/core.c @@ -21,6 +21,7 @@ #include #include #include +#include =20 #include #include @@ -523,6 +524,103 @@ u32 mmc_vddrange_to_ocrmask(int vdd_min, } EXPORT_SYMBOL(mmc_vddrange_to_ocrmask); =20 +#ifdef CONFIG_MMC_REGULATOR + +/** + * mmc_regulator_get_ocrmask - return mask of supported voltages + * @host: mmc host whose supply will be consulted + * @supply: supply voltage to use; "vdd" if NULL + * + * This returns either a negative errno, or a mask of voltages + * that can be provided to MMC/SD/SDIO devices using the specified + * host's "vdd" supply. + */ +int mmc_regulator_get_ocrmask(struct mmc_host *host, const char *suppl= y) +{ + int result =3D 0; + struct regulator *reg; + int count; + int i; + + reg =3D regulator_get(host->parent, supply ? : "vdd"); + if (IS_ERR(reg)) + return PTR_ERR(reg); + + count =3D regulator_count_voltages(reg); + if (count < 0) { + result =3D count; + goto done; + } + + for (i =3D 0; i < count; i++) { + int vdd_uV; + int vdd_mV; + + vdd_uV =3D regulator_list_voltage(reg, i); + if (vdd_uV <=3D 0) + continue; + + vdd_mV =3D vdd_uV / 1000; + result |=3D mmc_vddrange_to_ocrmask(vdd_mV, vdd_mV); + } + +done: + regulator_put(reg); + return result; +} +EXPORT_SYMBOL(mmc_regulator_get_ocrmask); + +/** + * mmc_regulator_set_ocr - set regulator to match host->ios voltage + * @host: mmc host whose supply voltage will be changed + * @supply: supply voltage to use; "vdd" if NULL + * + * MMC host drivers may use this to enable or disable a regulator + * using a particular supply voltage. This would normally be + * called from the set_ios() method, possibly as part of updating + * digital interfaces to support that voltage. + */ +int mmc_regulator_set_ocr(struct mmc_host *host, const char *supply) +{ + int result =3D 0; + struct regulator *reg; + int min_mV, max_mV; + int enabled; + + reg =3D regulator_get(host->parent, supply ? : "vdd"); + if (IS_ERR(reg)) + return PTR_ERR(reg); + enabled =3D regulator_is_enabled(reg); + if (WARN(enabled < 0, "%s: regulator_is_enabled --> %d\n", + mmc_hostname(host), enabled)) + enabled =3D !host->ios.vdd; + + if (host->ios.vdd) { + int tmp; + + tmp =3D host->ios.vdd - ilog2(MMC_VDD_165_195); + if (tmp =3D=3D 0) { + min_mV =3D 1650; + max_mV =3D 1950; + } else { + min_mV =3D 2000 + tmp * 100; + max_mV =3D min_mV + 100; + } + + result =3D regulator_set_voltage(reg, min_mV * 1000, max_mV * 1000); + if (result =3D=3D 0 && !enabled) + result =3D regulator_enable(reg); + } else if (enabled) { + result =3D regulator_disable(reg); + } + + regulator_put(reg); + return result; +} +EXPORT_SYMBOL(mmc_regulator_set_ocr); + +#endif + /* * Mask off any voltages we don't support and select * the lowest voltage --- a/include/linux/mmc/host.h +++ b/include/linux/mmc/host.h @@ -192,5 +192,8 @@ static inline void mmc_signal_sdio_irq(s wake_up_process(host->sdio_irq_thread); } =20 +int mmc_regulator_get_ocrmask(struct mmc_host *host, const char *suppl= y); +int mmc_regulator_set_ocr(struct mmc_host *host, const char *supply); + #endif =20 -- To unsubscribe from this list: send the line "unsubscribe linux-omap" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html