From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S934945AbdJQKIs (ORCPT ); Tue, 17 Oct 2017 06:08:48 -0400 Received: from metis.ext.4.pengutronix.de ([92.198.50.35]:52485 "EHLO metis.ext.4.pengutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933518AbdJQKIq (ORCPT ); Tue, 17 Oct 2017 06:08:46 -0400 Message-ID: <1508234898.6854.6.camel@pengutronix.de> Subject: Re: [PATCH] reset: meson: add level reset support for GX SoC family From: Philipp Zabel To: Neil Armstrong Cc: linux-amlogic@lists.infradead.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org Date: Tue, 17 Oct 2017 12:08:18 +0200 In-Reply-To: <1508167573-17396-1-git-send-email-narmstrong@baylibre.com> References: <1508167573-17396-1-git-send-email-narmstrong@baylibre.com> Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.22.6-1+deb9u1 Mime-Version: 1.0 Content-Transfer-Encoding: 7bit X-SA-Exim-Connect-IP: 2001:67c:670:100:3ad5:47ff:feaf:1a17 X-SA-Exim-Mail-From: p.zabel@pengutronix.de X-SA-Exim-Scanned: No (on metis.ext.pengutronix.de); SAEximRunCond expanded to false X-PTX-Original-Recipient: linux-kernel@vger.kernel.org Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Neil, On Mon, 2017-10-16 at 17:26 +0200, Neil Armstrong wrote: > The Amlogic GX SoC family embeds alternate registers to drive the reset > levels next to the pulse registers. > > This patch adds support for level reset handling on the GX family only. > > The Meson8 family has an alternate way to handle level reset. > > Signed-off-by: Neil Armstrong thank you for the patch, comments below: > --- > drivers/reset/reset-meson.c | 57 +++++++++++++++++++++++++++++++++++++++++---- > 1 file changed, 53 insertions(+), 4 deletions(-) > > diff --git a/drivers/reset/reset-meson.c b/drivers/reset/reset-meson.c > index a8b915e..d55e440 100644 > --- a/drivers/reset/reset-meson.c > +++ b/drivers/reset/reset-meson.c > @@ -62,9 +62,11 @@ > #include > #include > #include > +#include > > #define REG_COUNT 8 > #define BITS_PER_REG 32 > +#define LEVEL_OFFSET 0x7c > > struct meson_reset { > void __iomem *reg_base; > @@ -88,18 +90,61 @@ static int meson_reset_reset(struct reset_controller_dev *rcdev, > return 0; > } > > -static const struct reset_control_ops meson_reset_ops = { > +static int meson_reset_level(struct reset_controller_dev *rcdev, > + unsigned long id, bool assert) > +{ > + struct meson_reset *data = > + container_of(rcdev, struct meson_reset, rcdev); > + unsigned int bank = id / BITS_PER_REG; > + unsigned int offset = id % BITS_PER_REG; > + void __iomem *reg_addr = data->reg_base + LEVEL_OFFSET + (bank << 2); > + u32 reg; > + > + if (bank >= REG_COUNT) > + return -EINVAL; This check is not necessary. The same check is in meson_reset_reset, which I didn't notice last time. of_reset_simple_xlate, the default rcdev->of_xlate implementation, already guarantees id < rcdev->nr_resets. And since nr_resets is set to REG_COUNT * BITS_PER_REG, we know that id < REG_COUNT * BITS_PER_REG and thus bank <= id / BITS_PER_REG < REG_COUNT. > + reg = readl(reg_addr); > + if (assert) > + writel(reg & ~BIT(offset), reg_addr); > + else > + writel(reg | BIT(offset), reg_addr); These read-modify-write operations must be protected by a spinlock. > + > + return 0; > +} > + > +static int meson_reset_assert(struct reset_controller_dev *rcdev, > + unsigned long id) > +{ > + return meson_reset_level(rcdev, id, true); > +} > + > +static int meson_reset_deassert(struct reset_controller_dev *rcdev, > + unsigned long id) > +{ > + return meson_reset_level(rcdev, id, false); > +} > + > +static const struct reset_control_ops meson_reset_meson8_ops = { > + .reset = meson_reset_reset, > +}; > + > +static const struct reset_control_ops meson_reset_gx_ops = { > .reset = meson_reset_reset, > + .assert = meson_reset_assert, > + .deassert = meson_reset_deassert, > }; > > static const struct of_device_id meson_reset_dt_ids[] = { > - { .compatible = "amlogic,meson8b-reset", }, > - { .compatible = "amlogic,meson-gxbb-reset", }, > + { .compatible = "amlogic,meson8b-reset", > + .data = (void *) &meson_reset_meson8_ops, }, > + { .compatible = "amlogic,meson-gxbb-reset", > + .data = (void *) &meson_reset_gx_ops, }, > { /* sentinel */ }, > }; of_device_id.data ist const void *, so there is no need to cast here. > static int meson_reset_probe(struct platform_device *pdev) > { > + const struct reset_control_ops *ops; > struct meson_reset *data; > struct resource *res; > > @@ -107,6 +152,10 @@ static int meson_reset_probe(struct platform_device *pdev) > if (!data) > return -ENOMEM; > > + ops = of_device_get_match_data(&pdev->dev); > + if (!ops) > + return -EINVAL; > + > res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > data->reg_base = devm_ioremap_resource(&pdev->dev, res); > if (IS_ERR(data->reg_base)) > @@ -116,7 +165,7 @@ static int meson_reset_probe(struct platform_device *pdev) > > data->rcdev.owner = THIS_MODULE; > data->rcdev.nr_resets = REG_COUNT * BITS_PER_REG; > - data->rcdev.ops = &meson_reset_ops; > + data->rcdev.ops = ops; > data->rcdev.of_node = pdev->dev.of_node; > > return devm_reset_controller_register(&pdev->dev, &data->rcdev); regards Philipp