On Thu, Feb 06, 2020 at 01:45:57PM +0000, Z.q. Hou wrote:
> Hi Andrew,
>
> Thanks a lot for your comments!
>
> > -----Original Message-----
> > From: Andrew Murray <andrew.murray@xxxxxxx>
> > Sent: 2020年1月13日 20:03
> > To: Z.q. Hou <zhiqiang.hou@xxxxxxx>
> > Cc: linux-pci@xxxxxxxxxxxxxxx; linux-arm-kernel@xxxxxxxxxxxxxxxxxxx;
> > devicetree@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> > bhelgaas@xxxxxxxxxx; robh+dt@xxxxxxxxxx; arnd@xxxxxxxx;
> > mark.rutland@xxxxxxx; l.subrahmanya@xxxxxxxxxxxxxx;
> > shawnguo@xxxxxxxxxx; m.karthikeyan@xxxxxxxxxxxxxx; Leo Li
> > <leoyang.li@xxxxxxx>; lorenzo.pieralisi@xxxxxxx;
> > catalin.marinas@xxxxxxx; will.deacon@xxxxxxx; Mingkai Hu
> > <mingkai.hu@xxxxxxx>; M.h. Lian <minghuan.lian@xxxxxxx>; Xiaowei Bao
> > <xiaowei.bao@xxxxxxx>
> > Subject: Re: [PATCHv9 10/12] PCI: mobiveil: Add PCIe Gen4 RC driver for NXP
> > Layerscape SoCs
> >
> > On Wed, Nov 20, 2019 at 03:46:23AM +0000, Z.q. Hou wrote:
> > > From: Hou Zhiqiang <Zhiqiang.Hou@xxxxxxx>
> > >
> > > This PCIe controller is based on the Mobiveil GPEX IP, which is
> > > compatible with the PCI Express™ Base Specification, Revision 4.0.
> > >
> > > Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@xxxxxxx>
> > > Reviewed-by: Minghuan Lian <Minghuan.Lian@xxxxxxx>
> > > ---
> > > drivers/pci/controller/mobiveil/Kconfig | 10 +
> > > drivers/pci/controller/mobiveil/Makefile | 1 +
> > > .../mobiveil/pcie-layerscape-gen4.c | 274
> > ++++++++++++++++++
> > > .../pci/controller/mobiveil/pcie-mobiveil.h | 16 +-
> > > 4 files changed, 299 insertions(+), 2 deletions(-) create mode
> > > 100644 drivers/pci/controller/mobiveil/pcie-layerscape-gen4.c
> > >
> > > diff --git a/drivers/pci/controller/mobiveil/Kconfig
> > > b/drivers/pci/controller/mobiveil/Kconfig
> > > index 64343c07bfed..c823be8dab1c 100644
> > > --- a/drivers/pci/controller/mobiveil/Kconfig
> > > +++ b/drivers/pci/controller/mobiveil/Kconfig
> > > @@ -21,4 +21,14 @@ config PCIE_MOBIVEIL_PLAT
> > > Soft IP. It has up to 8 outbound and inbound windows
> > > for address translation and it is a PCIe Gen4 IP.
> > >
> > > +config PCIE_LAYERSCAPE_GEN4
> > > + bool "Freescale Layerscape PCIe Gen4 controller"
> > > + depends on PCI
> > > + depends on OF && (ARM64 || ARCH_LAYERSCAPE)
> > > + depends on PCI_MSI_IRQ_DOMAIN
> > > + select PCIE_MOBIVEIL_HOST
> > > + help
> > > + Say Y here if you want PCIe Gen4 controller support on
> > > + Layerscape SoCs. The PCIe controller can work in RC or
> > > + EP mode according to RCW[HOST_AGT_PEX] setting.
> >
> > I think you can remove the last sentence - it doesn't give any value to users of
> > KConfig.
>
> OK, will remove it in v10.
>
> >
> >
> > > endmenu
> > > diff --git a/drivers/pci/controller/mobiveil/Makefile
> > > b/drivers/pci/controller/mobiveil/Makefile
> > > index 9fb6d1c6504d..99d879de32d6 100644
> > > --- a/drivers/pci/controller/mobiveil/Makefile
> > > +++ b/drivers/pci/controller/mobiveil/Makefile
> > > @@ -2,3 +2,4 @@
> > > obj-$(CONFIG_PCIE_MOBIVEIL) += pcie-mobiveil.o
> > > obj-$(CONFIG_PCIE_MOBIVEIL_HOST) += pcie-mobiveil-host.o
> > > obj-$(CONFIG_PCIE_MOBIVEIL_PLAT) += pcie-mobiveil-plat.o
> > > +obj-$(CONFIG_PCIE_LAYERSCAPE_GEN4) += pcie-layerscape-gen4.o
> > > diff --git a/drivers/pci/controller/mobiveil/pcie-layerscape-gen4.c
> > > b/drivers/pci/controller/mobiveil/pcie-layerscape-gen4.c
> > > new file mode 100644
> > > index 000000000000..6c0d3e2532db
> > > --- /dev/null
> > > +++ b/drivers/pci/controller/mobiveil/pcie-layerscape-gen4.c
> > > @@ -0,0 +1,274 @@
> > > +// SPDX-License-Identifier: GPL-2.0
> > > +/*
> > > + * PCIe Gen4 host controller driver for NXP Layerscape SoCs
> > > + *
> > > + * Copyright 2019 NXP
> > > + *
> > > + * Author: Zhiqiang Hou <Zhiqiang.Hou@xxxxxxx> */
> > > +
> > > +#include <linux/kernel.h>
> > > +#include <linux/interrupt.h>
> > > +#include <linux/init.h>
> > > +#include <linux/of_pci.h>
> > > +#include <linux/of_platform.h>
> > > +#include <linux/of_irq.h>
> > > +#include <linux/of_address.h>
> > > +#include <linux/pci.h>
> > > +#include <linux/platform_device.h>
> > > +#include <linux/resource.h>
> > > +#include <linux/mfd/syscon.h>
> > > +#include <linux/regmap.h>
> > > +
> > > +#include "pcie-mobiveil.h"
> > > +
> > > +/* LUT and PF control registers */
> > > +#define PCIE_LUT_OFF 0x80000
> > > +#define PCIE_PF_OFF 0xc0000
> > > +#define PCIE_PF_INT_STAT 0x18
> > > +#define PF_INT_STAT_PABRST BIT(31)
> > > +
> > > +#define PCIE_PF_DBG 0x7fc
> > > +#define PF_DBG_LTSSM_MASK 0x3f
> > > +#define PF_DBG_LTSSM_L0 0x2d /* L0 state */
> > > +#define PF_DBG_WE BIT(31)
> > > +#define PF_DBG_PABR BIT(27)
> > > +
> > > +#define to_ls_pcie_g4(x) platform_get_drvdata((x)->pdev)
> > > +
> > > +struct ls_pcie_g4 {
> > > + struct mobiveil_pcie pci;
> > > + struct delayed_work dwork;
> > > + int irq;
> > > +};
> > > +
> > > +static inline u32 ls_pcie_g4_lut_readl(struct ls_pcie_g4 *pcie, u32
> > > +off) {
> > > + return ioread32(pcie->pci.csr_axi_slave_base + PCIE_LUT_OFF + off);
> > > +}
> > > +
> > > +static inline void ls_pcie_g4_lut_writel(struct ls_pcie_g4 *pcie,
> > > + u32 off, u32 val)
> > > +{
> > > + iowrite32(val, pcie->pci.csr_axi_slave_base + PCIE_LUT_OFF + off); }
> > > +
> > > +static inline u32 ls_pcie_g4_pf_readl(struct ls_pcie_g4 *pcie, u32
> > > +off) {
> > > + return ioread32(pcie->pci.csr_axi_slave_base + PCIE_PF_OFF + off); }
> > > +
> > > +static inline void ls_pcie_g4_pf_writel(struct ls_pcie_g4 *pcie,
> > > + u32 off, u32 val)
> > > +{
> > > + iowrite32(val, pcie->pci.csr_axi_slave_base + PCIE_PF_OFF + off); }
> > > +
> > > +static bool ls_pcie_g4_is_bridge(struct ls_pcie_g4 *pcie) {
> > > + struct mobiveil_pcie *mv_pci = &pcie->pci;
> > > + u32 header_type;
> > > +
> > > + header_type = mobiveil_csr_readb(mv_pci, PCI_HEADER_TYPE);
> > > + header_type &= 0x7f;
> > > +
> > > + return header_type == PCI_HEADER_TYPE_BRIDGE; }
> > > +
> > > +static int ls_pcie_g4_link_up(struct mobiveil_pcie *pci) {
> > > + struct ls_pcie_g4 *pcie = to_ls_pcie_g4(pci);
> > > + u32 state;
> > > +
> > > + state = ls_pcie_g4_pf_readl(pcie, PCIE_PF_DBG);
> > > + state = state & PF_DBG_LTSSM_MASK;
> > > +
> > > + if (state == PF_DBG_LTSSM_L0)
> > > + return 1;
> > > +
> > > + return 0;
> > > +}
> > > +
> > > +static void ls_pcie_g4_disable_interrupt(struct ls_pcie_g4 *pcie) {
> > > + struct mobiveil_pcie *mv_pci = &pcie->pci;
> > > +
> > > + mobiveil_csr_writel(mv_pci, 0, PAB_INTP_AMBA_MISC_ENB); }
> > > +
> > > +static void ls_pcie_g4_enable_interrupt(struct ls_pcie_g4 *pcie) {
> > > + struct mobiveil_pcie *mv_pci = &pcie->pci;
> > > + u32 val;
> > > +
> > > + /* Clear the interrupt status */
> > > + mobiveil_csr_writel(mv_pci, 0xffffffff, PAB_INTP_AMBA_MISC_STAT);
> > > +
> > > + val = PAB_INTP_INTX_MASK | PAB_INTP_MSI | PAB_INTP_RESET |
> > > + PAB_INTP_PCIE_UE | PAB_INTP_IE_PMREDI | PAB_INTP_IE_EC;
> > > + mobiveil_csr_writel(mv_pci, val, PAB_INTP_AMBA_MISC_ENB); }
> > > +
> > > +static void ls_pcie_g4_reinit_hw(struct ls_pcie_g4 *pcie) {
> > > + struct mobiveil_pcie *mv_pci = &pcie->pci;
> > > + struct device *dev = &mv_pci->pdev->dev;
> > > + u32 val, act_stat;
> > > + int to = 100;
> > > +
> > > + /* Poll for pab_csb_reset to set and PAB activity to clear */
> > > + do {
> > > + usleep_range(10, 15);
> > > + val = ls_pcie_g4_pf_readl(pcie, PCIE_PF_INT_STAT);
> > > + act_stat = mobiveil_csr_readl(mv_pci, PAB_ACTIVITY_STAT);
> > > + } while (((val & PF_INT_STAT_PABRST) == 0 || act_stat) && to--);
> > > + if (to < 0) {
> > > + dev_err(dev, "Poll PABRST&PABACT timeout\n");
> > > + return;
> >
> > If a timeout happens here - the caller has no idea this has happened and yet
> > the following work doesn't get done. Isn't this a problem?
>
> Will change the return value type to 'int' in v10, so that the caller can know the fail.
>
> >
> > > + }
> > > +
> > > + /* clear PEX_RESET bit in PEX_PF0_DBG register */
> > > + val = ls_pcie_g4_pf_readl(pcie, PCIE_PF_DBG);
> > > + val |= PF_DBG_WE;
> > > + ls_pcie_g4_pf_writel(pcie, PCIE_PF_DBG, val);
> > > +
> > > + val = ls_pcie_g4_pf_readl(pcie, PCIE_PF_DBG);
> > > + val |= PF_DBG_PABR;
> > > + ls_pcie_g4_pf_writel(pcie, PCIE_PF_DBG, val);
> > > +
> > > + val = ls_pcie_g4_pf_readl(pcie, PCIE_PF_DBG);
> > > + val &= ~PF_DBG_WE;
> > > + ls_pcie_g4_pf_writel(pcie, PCIE_PF_DBG, val);
> > > +
> > > + mobiveil_host_init(mv_pci, true);
> >
> > Can mobiveil_host_init fail?
>
> It should not fail, only register programming operations were left in this function.
>
> >
> > > +
> > > + to = 100;
> > > + while (!ls_pcie_g4_link_up(mv_pci) && to--)
> > > + usleep_range(200, 250);
> > > + if (to < 0)
> > > + dev_err(dev, "PCIe link training timeout\n"); }
> > > +
> > > +static irqreturn_t ls_pcie_g4_isr(int irq, void *dev_id) {
> > > + struct ls_pcie_g4 *pcie = (struct ls_pcie_g4 *)dev_id;
> > > + struct mobiveil_pcie *mv_pci = &pcie->pci;
> > > + u32 val;
> > > +
> > > + val = mobiveil_csr_readl(mv_pci, PAB_INTP_AMBA_MISC_STAT);
> > > + if (!val)
> > > + return IRQ_NONE;
> > > +
> > > + if (val & PAB_INTP_RESET) {
> >
> > Can you explain why this is needed (perhaps also in the cover letter)?
>
> The hot reset will result in the RC crash, so need the ISR to reset the RC.
>
> >
> > > + ls_pcie_g4_disable_interrupt(pcie);
> > > + schedule_delayed_work(&pcie->dwork, msecs_to_jiffies(1));
> > > + }
> > > +
> > > + mobiveil_csr_writel(mv_pci, val, PAB_INTP_AMBA_MISC_STAT);
> > > +
> > > + return IRQ_HANDLED;
> > > +}
> > > +
> > > +static int ls_pcie_g4_interrupt_init(struct mobiveil_pcie *mv_pci) {
> > > + struct ls_pcie_g4 *pcie = to_ls_pcie_g4(mv_pci);
> > > + struct platform_device *pdev = mv_pci->pdev;
> > > + struct device *dev = &pdev->dev;
> > > + int ret;
> > > +
> > > + pcie->irq = platform_get_irq_byname(pdev, "intr");
> > > + if (pcie->irq < 0) {
> > > + dev_err(dev, "Can't get 'intr' IRQ, errno = %d\n", pcie->irq);
> > > + return pcie->irq;
> > > + }
> > > + ret = devm_request_irq(dev, pcie->irq, ls_pcie_g4_isr,
> > > + IRQF_SHARED, pdev->name, pcie);
> > > + if (ret) {
> > > + dev_err(dev, "Can't register PCIe IRQ, errno = %d\n", ret);
> > > + return ret;
> > > + }
> > > +
> > > + return 0;
> > > +}
> > > +
> > > +static void ls_pcie_g4_reset(struct work_struct *work) {
> > > + struct delayed_work *dwork = container_of(work, struct delayed_work,
> > > + work);
> > > + struct ls_pcie_g4 *pcie = container_of(dwork, struct ls_pcie_g4, dwork);
> > > + struct mobiveil_pcie *mv_pci = &pcie->pci;
> > > + u16 ctrl;
> > > +
> > > + ctrl = mobiveil_csr_readw(mv_pci, PCI_BRIDGE_CONTROL);
> > > + ctrl &= ~PCI_BRIDGE_CTL_BUS_RESET;
> > > + mobiveil_csr_writew(mv_pci, ctrl, PCI_BRIDGE_CONTROL);
> > > + ls_pcie_g4_reinit_hw(pcie);
> > > + ls_pcie_g4_enable_interrupt(pcie);
> > > +}
> > > +
> > > +static struct mobiveil_rp_ops ls_pcie_g4_rp_ops = {
> > > + .interrupt_init = ls_pcie_g4_interrupt_init, };
> > > +
> > > +static const struct mobiveil_pab_ops ls_pcie_g4_pab_ops = {
> > > + .link_up = ls_pcie_g4_link_up,
> > > +};
> > > +
> > > +static int __init ls_pcie_g4_probe(struct platform_device *pdev) {
> > > + struct device *dev = &pdev->dev;
> > > + struct pci_host_bridge *bridge;
> > > + struct mobiveil_pcie *mv_pci;
> > > + struct ls_pcie_g4 *pcie;
> > > + struct device_node *np = dev->of_node;
> > > + int ret;
> > > +
> > > + if (!of_parse_phandle(np, "msi-parent", 0)) {
> > > + dev_err(dev, "Failed to find msi-parent\n");
> > > + return -EINVAL;
> > > + }
> > > +
> > > + bridge = devm_pci_alloc_host_bridge(dev, sizeof(*pcie));
> > > + if (!bridge)
> > > + return -ENOMEM;
> > > +
> > > + pcie = pci_host_bridge_priv(bridge);
> > > + mv_pci = &pcie->pci;
> > > +
> > > + mv_pci->pdev = pdev;
> > > + mv_pci->ops = &ls_pcie_g4_pab_ops;
> > > + mv_pci->rp.ops = &ls_pcie_g4_rp_ops;
> > > + mv_pci->rp.bridge = bridge;
> > > +
> > > + platform_set_drvdata(pdev, pcie);
> > > +
> > > + INIT_DELAYED_WORK(&pcie->dwork, ls_pcie_g4_reset);
> > > +
> > > + ret = mobiveil_pcie_host_probe(mv_pci);
> > > + if (ret) {
> > > + dev_err(dev, "Fail to probe\n");
> > > + return ret;
> > > + }
> > > +
> > > + if (!ls_pcie_g4_is_bridge(pcie))
> >
> > Is this a check that could apply to all host bridge drivers and thus live in
> > mobiveil_pcie_host_probe?
>
> Yes, will do in v10.
>
> >
> > > + return -ENODEV;
> > > +
> > > + ls_pcie_g4_enable_interrupt(pcie);
> >
> > Is there an issue here in that we enable interrupts *after* telling the kernel
> > about our controller? (Same applies for bailing if the IP isn't a bridge).
>
> Andrew, I don't understand the issue, can you help to explain?
If I recall correctly mobiveil_pcie_host_probe tells the kernel there is a PCI host
bridge and allows it to start enumerating the tree - at this point surely interrupts
may be expected - however we don't enable them until after this point. I'd assume
we'd need to get the hardware into a state where it can handle interrupts before
telling the kernel it can use this host bridge.
Likewise is ls_pcie_g4_is_bridge returns false we fail the probe yet the kernel may
already be enumerating the bus.
Thanks,
Andrew Murray
>
> Thanks,
> Zhiqiang
>
> >
> > Thanks,
> >
> > Andrew Murray
> >
> > > +
> > > + return 0;
> > > +}
> > > +
> > > +static const struct of_device_id ls_pcie_g4_of_match[] = {
> > > + { .compatible = "fsl,lx2160a-pcie", },
> > > + { },
> > > +};
> > > +
> > > +static struct platform_driver ls_pcie_g4_driver = {
> > > + .driver = {
> > > + .name = "layerscape-pcie-gen4",
> > > + .of_match_table = ls_pcie_g4_of_match,
> > > + .suppress_bind_attrs = true,
> > > + },
> > > +};
> > > +
> > > +builtin_platform_driver_probe(ls_pcie_g4_driver, ls_pcie_g4_probe);
> > > diff --git a/drivers/pci/controller/mobiveil/pcie-mobiveil.h
> > > b/drivers/pci/controller/mobiveil/pcie-mobiveil.h
> > > index 750a7fd95bc1..c57a68d2bac4 100644
> > > --- a/drivers/pci/controller/mobiveil/pcie-mobiveil.h
> > > +++ b/drivers/pci/controller/mobiveil/pcie-mobiveil.h
> > > @@ -43,6 +43,8 @@
> > > #define PAGE_LO_MASK 0x3ff
> > > #define PAGE_SEL_OFFSET_SHIFT 10
> > >
> > > +#define PAB_ACTIVITY_STAT 0x81c
> > > +
> > > #define PAB_AXI_PIO_CTRL 0x0840
> > > #define APIO_EN_MASK 0xf
> > >
> > > @@ -51,8 +53,18 @@
> > >
> > > #define PAB_INTP_AMBA_MISC_ENB 0x0b0c
> > > #define PAB_INTP_AMBA_MISC_STAT 0x0b1c
> > > -#define PAB_INTP_INTX_MASK 0x01e0
> > > -#define PAB_INTP_MSI_MASK 0x8
> > > +#define PAB_INTP_RESET BIT(1)
> > > +#define PAB_INTP_MSI BIT(3)
> > > +#define PAB_INTP_INTA BIT(5)
> > > +#define PAB_INTP_INTB BIT(6)
> > > +#define PAB_INTP_INTC BIT(7)
> > > +#define PAB_INTP_INTD BIT(8)
> > > +#define PAB_INTP_PCIE_UE BIT(9)
> > > +#define PAB_INTP_IE_PMREDI BIT(29)
> > > +#define PAB_INTP_IE_EC BIT(30)
> > > +#define PAB_INTP_MSI_MASK PAB_INTP_MSI
> > > +#define PAB_INTP_INTX_MASK (PAB_INTP_INTA |
> > PAB_INTP_INTB |\
> > > + PAB_INTP_INTC | PAB_INTP_INTD)
> > >
> > > #define PAB_AXI_AMAP_CTRL(win) PAB_REG_ADDR(0x0ba0,
> > win)
> > > #define WIN_ENABLE_SHIFT 0
> > > --
> > > 2.17.1
> > >
