Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751778AbbD2SbL (ORCPT ); Wed, 29 Apr 2015 14:31:11 -0400 Received: from mail-wg0-f52.google.com ([74.125.82.52]:33578 "EHLO mail-wg0-f52.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751186AbbD2SbI (ORCPT ); Wed, 29 Apr 2015 14:31:08 -0400 Date: Wed, 29 Apr 2015 19:30:59 +0100 From: Lee Jones To: Andrew Bresticker Cc: Thierry Reding , Stephen Warren , Alexandre Courbot , "linux-tegra@vger.kernel.org" , "linux-usb@vger.kernel.org" , "devicetree@vger.kernel.org" , "linux-arm-kernel@lists.infradead.org" , "linux-kernel@vger.kernel.org" , Samuel Ortiz Subject: Re: [PATCH V7 5/9] mfd: Add driver for NVIDIA Tegra XUSB Message-ID: <20150429183059.GA9169@x1> References: <1430174242-29465-1-git-send-email-abrestic@chromium.org> <1430174242-29465-6-git-send-email-abrestic@chromium.org> <20150429092348.GQ9169@x1> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 7611 Lines: 233 On Wed, 29 Apr 2015, Andrew Bresticker wrote: > Lee, > > On Wed, Apr 29, 2015 at 2:23 AM, Lee Jones wrote: > > On Mon, 27 Apr 2015, Andrew Bresticker wrote: > > > >> Add an MFD driver for the XUSB host complex found on NVIDIA Tegra124 > >> and later SoCs. > >> > >> Signed-off-by: Andrew Bresticker > >> Cc: Samuel Ortiz > >> Cc: Lee Jones > > >> --- /dev/null > >> +++ b/drivers/mfd/tegra-xusb.c > > >> +struct tegra_xusb_soc_data { > >> + struct mfd_cell *devs; > >> + unsigned int num_devs; > >> +}; > >> + > >> +static struct resource tegra_xhci_resources[] = { > >> + { > >> + .name = "host", > >> + .flags = IORESOURCE_IRQ, > >> + }, > >> + { > >> + .name = "xhci", > >> + .flags = IORESOURCE_MEM, > >> + }, > >> + { > >> + .name = "ipfs", > >> + .flags = IORESOURCE_MEM, > >> + }, > >> +}; > >> + > >> +static struct resource tegra_xusb_mbox_resources[] = { > >> + { > >> + .name = "smi", > >> + .flags = IORESOURCE_IRQ, > >> + }, > >> +}; > > > > DEFINE_RES_IRQ_NAMED() > > > >> +static struct mfd_cell tegra124_xusb_devs[] = { > >> + { > >> + .name = "tegra-xhci", > >> + .of_compatible = "nvidia,tegra124-xhci", > >> + }, > >> + { > >> + .name = "tegra-xusb-mbox", > >> + .of_compatible = "nvidia,tegra124-xusb-mbox", > >> + }, > >> +}; > >> + > >> +static const struct tegra_xusb_soc_data tegra124_xusb_data = { > >> + .devs = tegra124_xusb_devs, > >> + .num_devs = ARRAY_SIZE(tegra124_xusb_devs), > >> +}; > >> + > >> +static const struct of_device_id tegra_xusb_of_match[] = { > >> + { .compatible = "nvidia,tegra124-xusb", .data = &tegra124_xusb_data }, > > > > Yuk! Why are you mixing platform data and DT in this way? > > > > Why can't you just stick all of this in DT? > > I assume you mean the resources? The compatible strings will at least > need to be SoC-specific since they will change from SoC to SoC. > > >> + {}, > >> +}; > >> + > >> +MODULE_DEVICE_TABLE(of, tegra_xusb_of_match); > >> +static struct regmap_config tegra_fpci_regmap_config = { > >> + .reg_bits = 32, > >> + .val_bits = 32, > >> + .reg_stride = 4, > >> +}; > >> + > >> +static int tegra_xusb_probe(struct platform_device *pdev) > >> +{ > >> + const struct tegra_xusb_soc_data *soc; > >> + const struct of_device_id *match; > >> + struct tegra_xusb *xusb; > >> + struct resource *res; > >> + void __iomem *fpci_base; > >> + int irq, ret; > >> + > >> + xusb = devm_kzalloc(&pdev->dev, sizeof(*xusb), GFP_KERNEL); > >> + if (!xusb) > >> + return -ENOMEM; > >> + platform_set_drvdata(pdev, xusb); > >> + > >> + match = of_match_node(tegra_xusb_of_match, pdev->dev.of_node); > >> + soc = match->data; > >> + > >> + irq = platform_get_irq_byname(pdev, "host"); > >> + if (irq < 0) > >> + return irq; > >> + tegra_xhci_resources[0].start = irq; > >> + tegra_xhci_resources[0].end = irq; > >> + > >> + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "xhci"); > >> + if (!res) > >> + return -ENODEV; > >> + tegra_xhci_resources[1].start = res->start; > >> + tegra_xhci_resources[1].end = res->end; > >> + > >> + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "ipfs"); > >> + if (!res) > >> + return -ENODEV; > >> + tegra_xhci_resources[2].start = res->start; > >> + tegra_xhci_resources[2].end = res->end; > >> + > >> + soc->devs[0].resources = tegra_xhci_resources; > >> + soc->devs[0].num_resources = ARRAY_SIZE(tegra_xhci_resources); > >> + > >> + irq = platform_get_irq_byname(pdev, "smi"); > >> + if (irq < 0) > >> + return irq; > >> + tegra_xusb_mbox_resources[0].start = irq; > >> + tegra_xusb_mbox_resources[0].end = irq; > >> + > >> + soc->devs[1].resources = tegra_xusb_mbox_resources; > >> + soc->devs[1].num_resources = ARRAY_SIZE(tegra_xusb_mbox_resources); > >> + > >> + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "fpci"); > >> + fpci_base = devm_ioremap_resource(&pdev->dev, res); > >> + if (IS_ERR(fpci_base)) > >> + return PTR_ERR(fpci_base); > > > > This stuff is not good. > > > > Either let MFD handle it with mfd_cells or stick all of this stuff in > > DT and parse it from the child devices. > > I'm not sure what exactly you mean by "let MFD handle it with > mfd_cells" here - is that not what I'm doing now? Anyway I think that > leaves two ways of representing this in DT. > > Either have the MFD take a single IOMEM resource and divide it up > statically within the driver: > > usb@0,70090000 { > compatible = "nvidia,tegra124-xusb"; > reg = <0x0 0x70090000 0x0 0xa000>; > > usb-host { > compatible = "nvidia,tegra124-xhci"; > interrupts = <...>; > ... > }: > > mailbox { > compatible = "nvidia,tegra124-xusb-mbox"; > interrupts = <...>; > ... > }; > }; > > ... or have the MFD take only the shared FPCI resource and have the > sub-devices parse the rest from DT: > > usb@0,70098000 { > compatible = "nvidia,tegra124-xusb"; > reg = <0x0 0x70098000 0x0 0x1000>; > ranges; > > usb-host@0,70090000 { > compatible = "nvidia,tegra124-xhci"; > reg = <0x0 0x70090000 0x0 0x8000>, > <0x0 0x70099000 0x0 0x1000>; > interrupts = <...>; > ... > }: > > mailbox { > compatible = "nvidia,tegra124-xusb-mbox"; > interrupts = <...>; > ... > }; > }; > > I don't have a strong preference here, but I think the former more > accurately represents the resource hierarchy. Since Thierry requested > the use of an MFD, I'd like him to weigh in on this - Thierry? The latter is what I had in mind. > >> + tegra_fpci_regmap_config.max_register = res->end - res->start - 3; > >> + xusb->fpci_regs = devm_regmap_init_mmio(&pdev->dev, fpci_base, > >> + &tegra_fpci_regmap_config); > >> + if (IS_ERR(xusb->fpci_regs)) { > >> + ret = PTR_ERR(xusb->fpci_regs); > >> + dev_err(&pdev->dev, "Failed to init regmap: %d\n", ret); > >> + return ret; > >> + } > >> + > >> + ret = mfd_add_devices(&pdev->dev, -1, soc->devs, soc->num_devs, > >> + NULL, 0, NULL); > >> + if (ret) { > >> + dev_err(&pdev->dev, "Failed to add MFD devices: %d\n", ret); > >> + return ret; > >> + } > >> + > >> + return 0; > >> +} > > >> --- /dev/null > >> +++ b/include/soc/tegra/xusb.h > > >> +struct tegra_xusb { > >> + struct regmap *fpci_regs; > > > > Are you going to add to this? > > No, I don't have any plans to add to this struct. Then you don't require a struct. :) -- Lee Jones Linaro STMicroelectronics Landing Team Lead Linaro.org │ Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/