From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 8FFD6C6FA89 for ; Sat, 3 Sep 2022 17:44:08 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:In-Reply-To:MIME-Version:References: Message-ID:Subject:Cc:To:From:Date:Reply-To:Content-ID:Content-Description: Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID: List-Owner; bh=1pzzdT2VBQ7ebamIBs9HcexmQ+9Jbc3WOE8WUthWsXM=; b=Kd6OFkTR38gv8q Hc1PlMS7B8XlPl9Hv7Mw7k5gdC77UYWMYRcr5R6bK9fU1G6e52SlAo38bNIPyZuFB0L/8xnmjvy0v luDzEA/jWS/MVYX8GtKgUwo4tcAWuQjJyw7qwLwUGMSRhu+MrmpfhxP6aGKtc4ql7AHTcLUmP5cFV uDycy3uvO0NvxxyQFRqJueAch+S2JUPULyt4uK6vU4xgITCG1bFjjoWY5kBIluq4lOjsC7PeWhW4G Of46Z+U16vW8Ezy6uMGgZUlE4Tv2gH2n6fCTompJsCW05uu5pEmOc2hNUggL4Nds0bIAjGZD9eReP M2ZDF4Vtd2lslLi7+TtA==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.94.2 #2 (Red Hat Linux)) id 1oUXAT-00ER3f-FH; Sat, 03 Sep 2022 17:42:53 +0000 Received: from mga03.intel.com ([134.134.136.65]) by bombadil.infradead.org with esmtps (Exim 4.94.2 #2 (Red Hat Linux)) id 1oUXAQ-00EQz1-GI for linux-arm-kernel@lists.infradead.org; Sat, 03 Sep 2022 17:42:52 +0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1662226970; x=1693762970; h=date:from:to:cc:subject:message-id:references: mime-version:in-reply-to; bh=dmMI3hgnJrDyySmSfUKCyzdmMyWHvL+eCHVnZB0KYmU=; b=fi9orKeFGiOotfPZfteuZXPW6YkXUH8EK7FNBiiY/QPTNS+bhvZPSaVo S+sw/9toBaAg6a1D2dTuJ7a0s0DycE0Fs6mkbMd6r68t3GkWogpxjqWqZ p6faYXpKsBFaUm0q2I99n/GNnV5zKuEUA4ca9WlkYZXtmt8wF8bo47Cj8 SduCBysVAKyZZow/x8ZJFb/D0boQQbbKlUlelMCwH6DEG5AXqU6uXH7zI t8ngRkvGvNJTqRIPIUVaRg8dDlKTbEXREFaFZnvkPQF5p3RECday0adBh o5Da72cPlBVRJL6vEeHszzNtRIlLwCUowlDVqnvTdGzE/ugYHdTVInKK3 A==; X-IronPort-AV: E=McAfee;i="6500,9779,10459"; a="297471681" X-IronPort-AV: E=Sophos;i="5.93,287,1654585200"; d="scan'208";a="297471681" Received: from fmsmga008.fm.intel.com ([10.253.24.58]) by orsmga103.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Sep 2022 10:42:44 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.93,287,1654585200"; d="scan'208";a="674751116" Received: from yilunxu-optiplex-7050.sh.intel.com (HELO localhost) ([10.239.159.165]) by fmsmga008.fm.intel.com with ESMTP; 03 Sep 2022 10:42:39 -0700 Date: Sun, 4 Sep 2022 01:33:15 +0800 From: Xu Yilun To: "Manne, Nava kishore" Cc: "git (AMD-Xilinx)" , "robh+dt@kernel.org" , "krzysztof.kozlowski+dt@linaro.org" , "michal.simek@xilinx.com" , "mdf@kernel.org" , "hao.wu@intel.com" , "trix@redhat.com" , "p.zabel@pengutronix.de" , "gregkh@linuxfoundation.org" , "ronak.jain@xilinx.com" , "rajan.vaja@xilinx.com" , "abhyuday.godhasara@xilinx.com" , "piyush.mehta@xilinx.com" , "lakshmi.sai.krishna.potthuri@xilinx.com" , "harsha.harsha@xilinx.com" , "linus.walleij@linaro.org" , "nava.manne@xilinx.com" , "devicetree@vger.kernel.org" , "linux-arm-kernel@lists.infradead.org" , "linux-kernel@vger.kernel.org" , "linux-fpga@vger.kernel.org" Subject: Re: [PATCH 4/4] fpga: zynqmp: Add afi config driver Message-ID: References: <20220824035542.706433-1-nava.kishore.manne@amd.com> <20220824035542.706433-5-nava.kishore.manne@amd.com> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20220903_104250_678398_DE13233A X-CRM114-Status: GOOD ( 46.55 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org On 2022-08-30 at 09:19:52 +0000, Manne, Nava kishore wrote: > Hi Yilun, > > Please find my response inline. > > > -----Original Message----- > > From: Xu Yilun > > Sent: Saturday, August 27, 2022 5:56 PM > > To: Manne, Nava kishore > > Cc: git (AMD-Xilinx) ; robh+dt@kernel.org; > > krzysztof.kozlowski+dt@linaro.org; michal.simek@xilinx.com; > > mdf@kernel.org; hao.wu@intel.com; trix@redhat.com; > > p.zabel@pengutronix.de; gregkh@linuxfoundation.org; > > ronak.jain@xilinx.com; rajan.vaja@xilinx.com; > > abhyuday.godhasara@xilinx.com; piyush.mehta@xilinx.com; > > lakshmi.sai.krishna.potthuri@xilinx.com; harsha.harsha@xilinx.com; > > linus.walleij@linaro.org; nava.manne@xilinx.com; > > devicetree@vger.kernel.org; linux-arm-kernel@lists.infradead.org; linux- > > kernel@vger.kernel.org; linux-fpga@vger.kernel.org > > Subject: Re: [PATCH 4/4] fpga: zynqmp: Add afi config driver > > > > CAUTION: This message has originated from an External Source. Please use > > proper judgment and caution when opening attachments, clicking links, or > > responding to this email. > > > > > > On 2022-08-24 at 09:25:42 +0530, Nava kishore Manne wrote: > > > Add zynqmp AXI FIFO interface(AFI) config driver. This is useful for > > > the configuration of the PS-PL interface on Zynq US+ MPSoC platform. > > > > Please help illustrate how to use the device for FPGA reprogramming, why it > > should be implemented as an FPGA bridge. > > > > From the code I actually didn't see any operation that gates the fpga-region > > from other part of the machine. > > > > The Zynq UltraScale MPSoC family consists of a system-on-chip (SoC) style integrated processing system (PS) > and a Programmable Logic (PL) unit, providing an extensible and flexible SoC solution on a single die. > Xilinx Zynq US+ MPSoC connect the PS to the programmable logic (PL) through the AXI port. > This AXI port helps to establish the data path between the PS and PL (Here AXI Interface act as a Gating between PS and PL) > and this AXI port configuration vary from design to design. In-order to establish the proper communication path between > PS and PL (Full region), the AXI port data path should be configured with proper values priories to load the full region. > > Will update the description in v2. > > > > > > > Signed-off-by: Nava kishore Manne > > > --- > > > MAINTAINERS | 6 ++ > > > drivers/fpga/Kconfig | 13 +++ > > > drivers/fpga/Makefile | 1 + > > > drivers/fpga/zynqmp-afi.c | 211 > > > ++++++++++++++++++++++++++++++++++++++ > > > 4 files changed, 231 insertions(+) > > > create mode 100644 drivers/fpga/zynqmp-afi.c > > > > > > diff --git a/MAINTAINERS b/MAINTAINERS index > > > 20ffac651214..957e753e6406 100644 > > > --- a/MAINTAINERS > > > +++ b/MAINTAINERS > > > @@ -8032,6 +8032,12 @@ F: Documentation/fpga/ > > > F: drivers/fpga/ > > > F: include/linux/fpga/ > > > > > > +FPGA ZYNQMP PS-PL BRIDGE DRIVER > > > +M: Nava kishore Manne > > > +S: Supported > > > +F: Documentation/devicetree/bindings/fpga/xlnx,zynqmp-afi-fpga.yaml > > > +F: drivers/fpga/zynqmp-afi.c > > > + > > > INTEL MAX10 BMC SECURE UPDATES > > > M: Russ Weight > > > L: linux-fpga@vger.kernel.org > > > diff --git a/drivers/fpga/Kconfig b/drivers/fpga/Kconfig index > > > 6c416955da53..c08794d30fb5 100644 > > > --- a/drivers/fpga/Kconfig > > > +++ b/drivers/fpga/Kconfig > > > @@ -130,6 +130,19 @@ config XILINX_PR_DECOUPLER > > > reconfiguration, preventing the system deadlock that can > > > occur if AXI transactions are interrupted by DFX. > > > > > > +config ZYNQMP_AFI > > > + tristate "Xilinx ZYNQMP AFI support" > > > + depends on FPGA_BRIDGE > > > + help > > > + Say Y to enable drivers to handle the PS-PL clocks configurations > > > + and PS-PL Bus-width. Xilinx Zynq US+ MPSoC connect the PS to the > > > + programmable logic (PL) through the AXI port. This AXI port helps > > > + to establish the data path between the PS and PL. > > > + In-order to establish the proper communication path between PS and > > PL, > > > + the AXI port data path should be configured with the proper Bus- > > width > > > + values and it will also handles the PS-PL reset signals to reset the > > > + PL domain. > > > > Same concern, please describe its relationship to FPGA reprogramming. > > > > Same as above. > Will update the description in v2. > > > > + > > > config FPGA_REGION > > > tristate "FPGA Region" > > > depends on FPGA_BRIDGE > > > diff --git a/drivers/fpga/Makefile b/drivers/fpga/Makefile index > > > 42ae8b58abce..94cfe60972db 100644 > > > --- a/drivers/fpga/Makefile > > > +++ b/drivers/fpga/Makefile > > > @@ -31,6 +31,7 @@ obj-$(CONFIG_FPGA_BRIDGE) += fpga-bridge.o > > > obj-$(CONFIG_SOCFPGA_FPGA_BRIDGE) += altera-hps2fpga.o altera- > > fpga2sdram.o > > > obj-$(CONFIG_ALTERA_FREEZE_BRIDGE) += altera-freeze-bridge.o > > > obj-$(CONFIG_XILINX_PR_DECOUPLER) += xilinx-pr-decoupler.o > > > +obj-$(CONFIG_ZYNQMP_AFI) += zynqmp-afi.o > > > > > > # High Level Interfaces > > > obj-$(CONFIG_FPGA_REGION) += fpga-region.o > > > diff --git a/drivers/fpga/zynqmp-afi.c b/drivers/fpga/zynqmp-afi.c new > > > file mode 100644 index 000000000000..bc975d304039 > > > --- /dev/null > > > +++ b/drivers/fpga/zynqmp-afi.c > > > @@ -0,0 +1,211 @@ > > > +// SPDX-License-Identifier: GPL-2.0 > > > +/* > > > + * Copyright (C) 2022 Xilinx, Inc. > > > + */ > > > + > > > +#include > > > +#include #include > > > + #include #include > > > + #include #include > > > + #include #include > > > + > > > + > > > +/* Registers and special values for doing register-based operations */ > > > +#define AFI_RDCHAN_CTRL_OFFSET 0x00 > > > +#define AFI_WRCHAN_CTRL_OFFSET 0x14 > > > +#define AFI_BUSWIDTH_MASK BIT(0) > > > + > > > +/** > > > + * struct zynqmp_afi - AFI register description. > > > + * @dev: device that owns this. > > > + * @of_node: Device Tree overlay. > > > + * @resets: Pointer to the reset control for ps-pl resets. > > > + */ > > > +struct zynqmp_afi { > > > + struct device *dev; > > > + struct device_node *of_node; > > > + struct reset_control *resets; > > > +}; > > > + > > > +/** > > > + * struct zynqmp_afi_configreg - AFI configuration registers info. > > > + * @reg: Name of the AFI configuration register. > > > + * @id: Register index value. > > > + */ > > > +struct zynqmp_afi_configreg { > > > + char *reg; > > > + u32 id; > > > +}; > > > + > > > +static struct zynqmp_afi_configreg afi_cfgreg[] = { > > > + {.reg = "xlnx,afi-fm0-rd-bus-width", .id = AFIFM0_RDCTRL}, > > > + {.reg = "xlnx,afi-fm1-rd-bus-width", .id = AFIFM1_RDCTRL}, > > > + {.reg = "xlnx,afi-fm2-rd-bus-width", .id = AFIFM2_RDCTRL}, > > > + {.reg = "xlnx,afi-fm3-rd-bus-width", .id = AFIFM3_RDCTRL}, > > > + {.reg = "xlnx,afi-fm4-rd-bus-width", .id = AFIFM4_RDCTRL}, > > > + {.reg = "xlnx,afi-fm5-rd-bus-width", .id = AFIFM5_RDCTRL}, > > > + {.reg = "xlnx,afi-fm6-rd-bus-width", .id = AFIFM6_RDCTRL}, > > > + {.reg = "xlnx,afi-fm0-wr-bus-width", .id = AFIFM0_WRCTRL}, > > > + {.reg = "xlnx,afi-fm1-wr-bus-width", .id = AFIFM1_WRCTRL}, > > > + {.reg = "xlnx,afi-fm2-wr-bus-width", .id = AFIFM2_WRCTRL}, > > > + {.reg = "xlnx,afi-fm3-wr-bus-width", .id = AFIFM3_WRCTRL}, > > > + {.reg = "xlnx,afi-fm4-wr-bus-width", .id = AFIFM4_WRCTRL}, > > > + {.reg = "xlnx,afi-fm5-wr-bus-width", .id = AFIFM5_WRCTRL}, > > > + {.reg = "xlnx,afi-fm6-wr-bus-width", .id = AFIFM6_WRCTRL}, > > > + {.reg = "xlnx,afi-fs-ss0-bus-width", .id = AFIFS}, > > > + {.reg = "xlnx,afi-fs-ss2-bus-width", .id = AFIFS_SS2}, > > > + {} > > > +}; > > > + > > > +static int zynqmp_afi_config(struct zynqmp_afi *afi_data) { > > > + struct zynqmp_afi_configreg *cfgptr = afi_cfgreg; > > > + struct device_node *np = afi_data->of_node; > > > + u32 afi_ss0_val, afi_ss1_val, bus_width; > > > + int ret; > > > + > > > + while (cfgptr->reg) { > > > + ret = of_property_read_u32(np, cfgptr->reg, &bus_width); > > > + if (!ret) { > > > + if (cfgptr->id == AFIFS_SS2) { > > > + if (bus_width == 32) > > > + ret = zynqmp_pm_afi(AFIFS_SS2, > > > + AFIFS_SS_BUS_WIDTH_32_CONFIG_VAL); > > > + else if (bus_width == 64) > > > + ret = zynqmp_pm_afi(AFIFS_SS2, > > > + > > AFIFS_SS0_SS2_BUS_WIDTH_64_CONFIG_VAL); > > > + else if (bus_width == 128) > > > + ret = zynqmp_pm_afi(AFIFS_SS2, > > > + > > AFIFS_SS0_SS2_BUS_WIDTH_128_CONFIG_VAL); > > > + else > > > + return -EINVAL; > > > + } else if (cfgptr->id == AFIFS) { > > > + if (bus_width == 32) > > > + afi_ss0_val = AFIFS_SS_BUS_WIDTH_32_CONFIG_VAL; > > > + else if (bus_width == 64) > > > + afi_ss0_val = > > AFIFS_SS0_SS2_BUS_WIDTH_64_CONFIG_VAL; > > > + else if (bus_width == 128) > > > + afi_ss0_val = > > AFIFS_SS0_SS2_BUS_WIDTH_128_CONFIG_VAL; > > > + else > > > + return -EINVAL; > > > + > > > + ret = of_property_read_u32(np, "xlnx,afi-fs-ss1-bus- > > width", > > > + &bus_width); > > > + if (!ret) { > > > + if (bus_width == 32) > > > + afi_ss1_val = > > AFIFS_SS_BUS_WIDTH_32_CONFIG_VAL; > > > + else if (bus_width == 64) > > > + afi_ss1_val = > > AFIFS_SS1_BUS_WIDTH_64_CONFIG_VAL; > > > + else if (bus_width == 128) > > > + afi_ss1_val = > > AFIFS_SS1_BUS_WIDTH_128_CONFIG_VAL; > > > + else > > > + return -EINVAL; > > > + > > > + ret = zynqmp_pm_afi(AFIFS, afi_ss1_val | > > afi_ss0_val); > > > + } > > > + } else { > > > + if (bus_width == 32) > > > + ret = zynqmp_pm_afi(cfgptr->id, > > > + AFIFM_BUS_WIDTH_32_CONFIG_VAL); > > > + else if (bus_width == 64) > > > + ret = zynqmp_pm_afi(cfgptr->id, > > > + AFIFM_BUS_WIDTH_64_CONFIG_VAL); > > > + else if (bus_width == 128) > > > + ret = zynqmp_pm_afi(cfgptr->id, > > > + AFIFM_BUS_WIDTH_128_CONFIG_VAL); > > > + else > > > + return -EINVAL; > > > + } > > > + } > > > + cfgptr++; > > > + } > > > + > > > + return 0; > > > +} > > > + > > > +static int zynqmp_afi_enable_set(struct fpga_bridge *bridge, bool > > > +enable) { > > > + struct device_node *overlay = bridge->info->overlay; > > > + struct zynqmp_afi *priv = bridge->priv; > > > + int ret = 0; > > > + > > > + if (enable) { > > > + reset_control_reset(priv->resets); > > > + return 0; > > > + } > > > + > > > + of_node_get(overlay); > > > + priv->of_node = > > > + of_find_node_with_property(overlay, > > > + "xlnx,afi-fm0-rd-bus-width"); > > > > I'm a little confused here. The fpga_image_info.overlay is for fpga-region, but > > from your binding doc this property is for this afi device. You want to add > > another overlay targeting afi dev node during reprograming? > > > > Yes, it uses Overlay's(overlay targeting afi dev node). As I said above for design-to-design AFI configs I have concern on it. We should limit the overlay usage only on fpga_region, this is because we are changing the hardware in this region during reprograming. Adding overlays for other nodes could be easily abused, any changes could be made even not related to the fpga region. So if any cross device configurations should be made during reprograming, it should be initiated by fpga region driver through normal kernel APIs. Thanks, Yilun > relevant to the full region may change and these configs must be set prior to program the full region > once this overlay integrated to the live tree the existing afi config values will be replace with the new one's > > > > + if (priv->of_node) > > > + ret = zynqmp_afi_config(priv); > > > + of_node_put(priv->of_node); > > > + > > > + return ret; > > > +} > > > + > > > +static const struct fpga_bridge_ops zynqmp_afi_br_ops = { > > > + .enable_set = zynqmp_afi_enable_set, }; > > > + > > > +static const struct of_device_id zynqmp_afi_of_match[] = { > > > + { .compatible = "xlnx,zynqmp-afi-fpga" }, > > > + { }, > > > +}; > > > +MODULE_DEVICE_TABLE(of, zynqmp_afi_of_match); > > > + > > > +static int zynqmp_afi_probe(struct platform_device *pdev) { > > > + struct device *dev = &pdev->dev; > > > + struct zynqmp_afi *priv; > > > + struct fpga_bridge *br; > > > + > > > + priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL); > > > + if (!priv) > > > + return -ENOMEM; > > > + > > > + priv->dev = dev; > > > + > > > + priv->resets = > > devm_reset_control_array_get_optional_exclusive(&pdev->dev); > > > + if (IS_ERR(priv->resets)) > > > + return PTR_ERR(priv->resets); > > > + > > > + br = fpga_bridge_register(dev, "Xilinx ZynqMP AFI", > > > + &zynqmp_afi_br_ops, priv); > > > + if (IS_ERR(br)) { > > > + dev_err(dev, "unable to register Xilinx ZynqMP AFI"); > > > > Need a "\n" at the end? > > > > Will fix. > > Regards, > Navakishore. > _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel