diff mbox

[5/5] PCI: cadence: add EndPoint Controller driver for Cadence PCIe controller

Message ID 297fa17e12cf0f2fb223c05eeb18570707ff5bf1.1511439189.git.cyrille.pitchen@free-electrons.com (mailing list archive)
State New, archived
Delegated to: Bjorn Helgaas
Headers show

Commit Message

Cyrille Pitchen Nov. 23, 2017, 3:01 p.m. UTC
This patch adds support to the Cadence PCIe controller in endpoint mode.

Signed-off-by: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
---
 drivers/pci/cadence/Kconfig           |   9 +
 drivers/pci/cadence/Makefile          |   1 +
 drivers/pci/cadence/pcie-cadence-ep.c | 553 ++++++++++++++++++++++++++++++++++
 3 files changed, 563 insertions(+)
 create mode 100644 drivers/pci/cadence/pcie-cadence-ep.c

Comments

Lorenzo Pieralisi Dec. 1, 2017, 12:20 p.m. UTC | #1
On Thu, Nov 23, 2017 at 04:01:50PM +0100, Cyrille Pitchen wrote:
> This patch adds support to the Cadence PCIe controller in endpoint mode.

Please add a brief description to the log to describe the most salient
features.

> Signed-off-by: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
> ---
>  drivers/pci/cadence/Kconfig           |   9 +
>  drivers/pci/cadence/Makefile          |   1 +
>  drivers/pci/cadence/pcie-cadence-ep.c | 553 ++++++++++++++++++++++++++++++++++
>  3 files changed, 563 insertions(+)
>  create mode 100644 drivers/pci/cadence/pcie-cadence-ep.c
> 
> diff --git a/drivers/pci/cadence/Kconfig b/drivers/pci/cadence/Kconfig
> index 120306cae2aa..b2e6af71f39e 100644
> --- a/drivers/pci/cadence/Kconfig
> +++ b/drivers/pci/cadence/Kconfig
> @@ -21,4 +21,13 @@ config PCIE_CADENCE_HOST
>  	  mode. This PCIe controller may be embedded into many different vendors
>  	  SoCs.
>  
> +config PCIE_CADENCE_EP
> +	bool "Cadence PCIe endpoint controller"
> +	depends on PCI_ENDPOINT
> +	select PCIE_CADENCE
> +	help
> +	  Say Y here if you want to support the Cadence PCIe  controller in
> +	  endpoint mode. This PCIe controller may be embedded into many
> +	  different vendors SoCs.
> +
>  endif # PCI_CADENCE
> diff --git a/drivers/pci/cadence/Makefile b/drivers/pci/cadence/Makefile
> index d57d192d2595..61e9c8d6839d 100644
> --- a/drivers/pci/cadence/Makefile
> +++ b/drivers/pci/cadence/Makefile
> @@ -1,2 +1,3 @@
>  obj-$(CONFIG_PCIE_CADENCE) += pcie-cadence.o
>  obj-$(CONFIG_PCIE_CADENCE_HOST) += pcie-cadence-host.o
> +obj-$(CONFIG_PCIE_CADENCE_EP) += pcie-cadence-ep.o
> diff --git a/drivers/pci/cadence/pcie-cadence-ep.c b/drivers/pci/cadence/pcie-cadence-ep.c
> new file mode 100644
> index 000000000000..a1d761101a9c
> --- /dev/null
> +++ b/drivers/pci/cadence/pcie-cadence-ep.c
> @@ -0,0 +1,553 @@
> +/*
> + * Cadence PCIe host controller driver.

You should update this comment.

> + *
> + * Copyright (c) 2017 Cadence
> + *
> + * Author: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 as
> + * published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program.  If not, see <http://www.gnu.org/licenses/>.
> + */
> +#include <linux/kernel.h>
> +#include <linux/of.h>
> +#include <linux/pci-epc.h>
> +#include <linux/platform_device.h>
> +#include <linux/pm_runtime.h>
> +#include <linux/sizes.h>
> +#include <linux/delay.h>

Nit: alphabetical order.

> +#include "pcie-cadence.h"
> +
> +#define CDNS_PCIE_EP_MIN_APERTURE		128	/* 128 bytes */
> +#define CDNS_PCIE_EP_IRQ_PCI_ADDR_NONE		0x1
> +#define CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY	0x3
> +
> +/**
> + * struct cdns_pcie_ep_data - hardware specific data
> + * @max_regions: maximum nmuber of regions supported by hardware

s/nmuber/number

> + */
> +struct cdns_pcie_ep_data {
> +	size_t				max_regions;
> +};
> +
> +/**
> + * struct cdns_pcie_ep - private data for this PCIe endpoint controller driver
> + * @pcie: Cadence PCIe controller
> + * @data: pointer to a 'struct cdns_pcie_data'
> + */
> +struct cdns_pcie_ep {
> +	struct cdns_pcie		pcie;
> +	const struct cdns_pcie_ep_data	*data;
> +	struct pci_epc			*epc;
> +	unsigned long			ob_region_map;
> +	phys_addr_t			*ob_addr;
> +	phys_addr_t			irq_phys_addr;
> +	void __iomem			*irq_cpu_addr;
> +	u64				irq_pci_addr;
> +	u8				irq_pending;
> +};
> +
> +static int cdns_pcie_ep_write_header(struct pci_epc *epc,
> +				     struct pci_epf_header *hdr)
> +{
> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	u8 fn = 0;
> +
> +	if (fn == 0) {

I think there is some code to retrieve fn missing here.

> +		u32 id = CDNS_PCIE_LM_ID_VENDOR(hdr->vendorid) |
> +			 CDNS_PCIE_LM_ID_SUBSYS(hdr->subsys_vendor_id);
> +		cdns_pcie_writel(pcie, CDNS_PCIE_LM_ID, id);
> +	}
> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_DEVICE_ID, hdr->deviceid);
> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_REVISION_ID, hdr->revid);
> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_CLASS_PROG, hdr->progif_code);
> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_CLASS_DEVICE,
> +			       hdr->subclass_code | hdr->baseclass_code << 8);
> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_CACHE_LINE_SIZE,
> +			       hdr->cache_line_size);
> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_SUBSYSTEM_ID, hdr->subsys_id);
> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_INTERRUPT_PIN, hdr->interrupt_pin);
> +
> +	return 0;
> +}
> +
> +static int cdns_pcie_ep_set_bar(struct pci_epc *epc, enum pci_barno bar,
> +				dma_addr_t bar_phys, size_t size, int flags)
> +{
> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	u32 addr0, addr1, reg, cfg, b, aperture, ctrl;
> +	u8 fn = 0;
> +	u64 sz;
> +
> +	/* BAR size is 2^(aperture + 7) */
> +	sz = max_t(size_t, size, CDNS_PCIE_EP_MIN_APERTURE);
> +	sz = 1ULL << fls64(sz - 1);
> +	aperture = ilog2(sz) - 7; /* 128B -> 0, 256B -> 1, 512B -> 2, ... */
> +
> +	if ((flags & PCI_BASE_ADDRESS_SPACE) == PCI_BASE_ADDRESS_SPACE_IO) {
> +		ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_IO_32BITS;
> +	} else {
> +		bool is_prefetch = !!(flags & PCI_BASE_ADDRESS_MEM_PREFETCH);
> +		bool is_64bits = sz > SZ_2G;
> +
> +		if (is_64bits && (bar & 1))
> +			return -EINVAL;
> +
> +		switch (is_64bits << 1 | is_prefetch) {

I would not mind implementing this as a nested if-else, I am not a big
fan of using bool this way.

> +		case 0:
> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_MEM_32BITS;
> +			break;
> +
> +		case 1:
> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_PREFETCH_MEM_32BITS;
> +			break;
> +
> +		case 2:
> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_MEM_64BITS;
> +			break;
> +
> +		case 3:
> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_PREFETCH_MEM_64BITS;
> +			break;
> +		}
> +	}
> +
> +	addr0 = lower_32_bits(bar_phys);
> +	addr1 = upper_32_bits(bar_phys);
> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar),
> +			 addr0);
> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar),
> +			 addr1);

Is fn always 0 ?

> +	if (bar < BAR_4) {
> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG0(fn);
> +		b = bar;
> +	} else {
> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG1(fn);
> +		b = bar - BAR_4;
> +	}
> +
> +	cfg = cdns_pcie_readl(pcie, reg);
> +	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
> +		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
> +	cfg |= (CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE(b, aperture) |
> +		CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL(b, ctrl));
> +	cdns_pcie_writel(pcie, reg, cfg);
> +
> +	return 0;
> +}
> +
> +static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, enum pci_barno bar)
> +{
> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	u32 reg, cfg, b, ctrl;
> +	u8 fn = 0;
> +
> +	if (bar < BAR_4) {
> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG0(fn);
> +		b = bar;
> +	} else {
> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG1(fn);
> +		b = bar - BAR_4;
> +	}
> +
> +	ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_DISABLED;
> +	cfg = cdns_pcie_readl(pcie, reg);
> +	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
> +		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
> +	cfg |= CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(ctrl);
> +	cdns_pcie_writel(pcie, reg, cfg);
> +
> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar), 0);
> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar), 0);
> +}
> +
> +static int cdns_pcie_ep_map_addr(struct pci_epc *epc, phys_addr_t addr,
> +				 u64 pci_addr, size_t size)
> +{
> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	u32 r;
> +
> +	r = find_first_zero_bit(&ep->ob_region_map, sizeof(ep->ob_region_map));

Second argument must be in bits not bytes.

https://marc.info/?l=linux-pci&m=151179781225513&w=2

> +	if (r >= ep->data->max_regions - 1) {
> +		dev_err(&epc->dev, "no free outbound region\n");
> +		return -EINVAL;
> +	}
> +
> +	cdns_pcie_set_outbound_region(pcie, r, false, addr, pci_addr, size);
> +
> +	set_bit(r, &ep->ob_region_map);
> +	ep->ob_addr[r] = addr;
> +
> +	return 0;
> +}
> +
> +static void cdns_pcie_ep_unmap_addr(struct pci_epc *epc, phys_addr_t addr)
> +{
> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	u32 r;
> +
> +	for (r = 0; r < ep->data->max_regions - 1; r++)
> +		if (ep->ob_addr[r] == addr)
> +			break;
> +
> +	if (r >= ep->data->max_regions - 1)

== ?

> +		return;
> +
> +	cdns_pcie_reset_outbound_region(pcie, r);
> +
> +	ep->ob_addr[r] = 0;
> +	clear_bit(r, &ep->ob_region_map);
> +}
> +
> +static int cdns_pcie_ep_set_msi(struct pci_epc *epc, u8 mmc)
> +{
> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
> +	u16 flags;
> +	u8 fn = 0;
> +
> +	/* Validate the ID of the MSI Capability structure. */
> +	if (cdns_pcie_ep_fn_readb(pcie, fn, cap) != PCI_CAP_ID_MSI)
> +		return -EINVAL;
> +
> +	/*
> +	 * Set the Multiple Message Capable bitfield into the Message Control
> +	 * register.
> +	 */
> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
> +	flags = (flags & ~PCI_MSI_FLAGS_QMASK) | (mmc << 1);
> +	flags |= PCI_MSI_FLAGS_64BIT;
> +	flags &= ~PCI_MSI_FLAGS_MASKBIT;
> +	cdns_pcie_ep_fn_writew(pcie, fn, cap + PCI_MSI_FLAGS, flags);
> +
> +	return 0;
> +}
> +
> +static int cdns_pcie_ep_get_msi(struct pci_epc *epc)
> +{
> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
> +	u16 flags, mmc, mme;
> +	u8 fn = 0;
> +
> +	/* Validate the ID of the MSI Capability structure. */
> +	if (cdns_pcie_ep_fn_readb(pcie, fn, cap) != PCI_CAP_ID_MSI)
> +		return -EINVAL;
> +
> +	/* Validate that the MSI feature is actually enabled. */
> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
> +	if (!(flags & PCI_MSI_FLAGS_ENABLE))
> +		return -EINVAL;
> +
> +	/*
> +	 * Get the Multiple Message Enable bitfield from the Message Control
> +	 * register.
> +	 */
> +	mmc = (flags & PCI_MSI_FLAGS_QMASK) >> 1;
> +	mme = (flags & PCI_MSI_FLAGS_QSIZE) >> 4;
> +	if (mme > mmc)
> +		mme = mmc;
> +	if (mme > 5)
> +		mme = 5;

You should comment on what this 5 means and why it is fine to cap mme.

> +
> +	return mme;
> +}
> +
> +static void cdns_pcie_ep_assert_intx(struct cdns_pcie_ep *ep, u8 fn,
> +				     u8 intx, bool is_asserted)
> +{
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	u32 r = ep->data->max_regions - 1;
> +	u32 offset;
> +	u16 status;
> +	u8 msg_code;
> +
> +	intx &= 3;
> +
> +	/* Set the outbound region if needed. */
> +	if (unlikely(ep->irq_pci_addr != CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY)) {
> +		/* Last region was reserved for IRQ writes. */
> +		cdns_pcie_set_outbound_region_for_normal_msg(pcie, r,
> +							     ep->irq_phys_addr);
> +		ep->irq_pci_addr = CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY;
> +	}
> +
> +	if (is_asserted) {
> +		ep->irq_pending |= BIT(intx);
> +		msg_code = MSG_CODE_ASSERT_INTA + intx;
> +	} else {
> +		ep->irq_pending &= ~BIT(intx);
> +		msg_code = MSG_CODE_DEASSERT_INTA + intx;
> +	}
> +
> +	status = cdns_pcie_ep_fn_readw(pcie, fn, PCI_STATUS);
> +	if (((status & PCI_STATUS_INTERRUPT) != 0) ^ (ep->irq_pending != 0)) {
> +		status ^= PCI_STATUS_INTERRUPT;
> +		cdns_pcie_ep_fn_writew(pcie, fn, PCI_STATUS, status);
> +	}
> +
> +	offset = CDNS_PCIE_NORMAL_MSG_ROUTING(MSG_ROUTING_LOCAL) |
> +		 CDNS_PCIE_NORMAL_MSG_CODE(msg_code) |
> +		 CDNS_PCIE_MSG_NO_DATA;
> +	writel(0, ep->irq_cpu_addr + offset);
> +}
> +
> +static int cdns_pcie_ep_send_legacy_irq(struct cdns_pcie_ep *ep, u8 fn, u8 intx)
> +{
> +	u16 cmd;
> +
> +	cmd = cdns_pcie_ep_fn_readw(&ep->pcie, fn, PCI_COMMAND);
> +	if (cmd & PCI_COMMAND_INTX_DISABLE)
> +		return -EINVAL;
> +
> +	cdns_pcie_ep_assert_intx(ep, fn, intx, true);
> +	mdelay(1);

Add a comment please to explain the mdelay value.

> +	cdns_pcie_ep_assert_intx(ep, fn, intx, false);
> +	return 0;
> +}
> +
> +static int cdns_pcie_ep_raise_irq(struct pci_epc *epc,
> +				  enum pci_epc_irq_type type, u8 interrupt_num)
> +{
> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
> +	u16 flags, mmc, mme, data, data_mask;
> +	u8 msi_count;
> +	u64 pci_addr, pci_addr_mask = 0xff;
> +	u8 fn = 0;
> +
> +	/* Handle legacy IRQ. */
> +	if (type == PCI_EPC_IRQ_LEGACY)
> +		return cdns_pcie_ep_send_legacy_irq(ep, fn, 0);
> +
> +	/* Otherwise MSI. */
> +	if (type != PCI_EPC_IRQ_MSI)
> +		return -EINVAL;
> +
> +	/* Check whether the MSI feature has been enabled by the PCI host. */
> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
> +	if (!(flags & PCI_MSI_FLAGS_ENABLE))
> +		return -EINVAL;
> +
> +	/* Get the number of enabled MSIs */
> +	mmc = (flags & PCI_MSI_FLAGS_QMASK) >> 1;
> +	mme = (flags & PCI_MSI_FLAGS_QSIZE) >> 4;
> +	if (mme > mmc)
> +		mme = mmc;
> +	if (mme > 5)
> +		mme = 5;

Same comment as above.

> +	msi_count = 1 << mme;
> +	if (!interrupt_num || interrupt_num > msi_count)
> +		return -EINVAL;
> +
> +	/* Compute the data value to be written. */
> +	data_mask = msi_count - 1;
> +	data = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_DATA_64);
> +	data = (data & ~data_mask) | ((interrupt_num - 1) & data_mask);
> +
> +	/* Get the PCI address where to write the data into. */
> +	pci_addr = cdns_pcie_ep_fn_readl(pcie, fn, cap + PCI_MSI_ADDRESS_HI);
> +	pci_addr <<= 32;
> +	pci_addr |= cdns_pcie_ep_fn_readl(pcie, fn, cap + PCI_MSI_ADDRESS_LO);
> +	pci_addr &= GENMASK_ULL(63, 2);
> +
> +	/* Set the outbound region if needed. */
> +	if (unlikely(ep->irq_pci_addr != pci_addr)) {
> +		/* Last region was reserved for IRQ writes. */
> +		cdns_pcie_set_outbound_region(pcie, ep->data->max_regions - 1,
> +					      false,
> +					      ep->irq_phys_addr,
> +					      pci_addr & ~pci_addr_mask,
> +					      pci_addr_mask + 1);
> +		ep->irq_pci_addr = pci_addr;
> +	}
> +	writew(data, ep->irq_cpu_addr + (pci_addr & pci_addr_mask));
> +
> +	return 0;
> +}
> +
> +static int cdns_pcie_ep_start(struct pci_epc *epc)
> +{
> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	struct pci_epf *epf;
> +	u32 cfg;
> +	u8 fn = 0;
> +
> +	/* Enable this endpoint function. */
> +	cfg = cdns_pcie_readl(pcie, CDNS_PCIE_LM_EP_FUNC_CFG);
> +	cfg |= BIT(fn);
> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, cfg);
> +
> +	/*
> +	 * Already linked-up: don't call directly pci_epc_linkup() because we've
> +	 * already locked the epc->lock.
> +	 */
> +	list_for_each_entry(epf, &epc->pci_epf, list)
> +		pci_epf_linkup(epf);
> +
> +	return 0;
> +}
> +
> +static void cdns_pcie_ep_stop(struct pci_epc *epc)
> +{
> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> +	struct cdns_pcie *pcie = &ep->pcie;
> +	u32 cfg;
> +	u8 fn = 0;
> +
> +	/* Disable this endpoint function (function 0 can't be disabled). */

I do not understand this comment and how it applies to the code,
in other words fn is always 0 here (so it can't be disabled)
I do not understand what this code is there for.

> +	cfg = cdns_pcie_readl(pcie, CDNS_PCIE_LM_EP_FUNC_CFG);
> +	cfg &= ~BIT(fn);
> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, cfg);
> +}
> +
> +static const struct pci_epc_ops cdns_pcie_epc_ops = {
> +	.write_header	= cdns_pcie_ep_write_header,
> +	.set_bar	= cdns_pcie_ep_set_bar,
> +	.clear_bar	= cdns_pcie_ep_clear_bar,
> +	.map_addr	= cdns_pcie_ep_map_addr,
> +	.unmap_addr	= cdns_pcie_ep_unmap_addr,
> +	.set_msi	= cdns_pcie_ep_set_msi,
> +	.get_msi	= cdns_pcie_ep_get_msi,
> +	.raise_irq	= cdns_pcie_ep_raise_irq,
> +	.start		= cdns_pcie_ep_start,
> +	.stop		= cdns_pcie_ep_stop,
> +};
> +
> +static const struct cdns_pcie_ep_data cdns_pcie_ep_data = {
> +	.max_regions	= 16,
> +};

As I mentioned in patch 3, should this be set-up with DT ?

Thanks,
Lorenzo

> +
> +static const struct of_device_id cdns_pcie_ep_of_match[] = {
> +	{ .compatible = "cdns,cdns-pcie-ep",
> +	  .data = &cdns_pcie_ep_data },
> +
> +	{ },
> +};
> +
> +static int cdns_pcie_ep_probe(struct platform_device *pdev)
> +{
> +	struct device *dev = &pdev->dev;
> +	struct device_node *np = dev->of_node;
> +	const struct of_device_id *of_id;
> +	struct cdns_pcie_ep *ep;
> +	struct cdns_pcie *pcie;
> +	struct pci_epc *epc;
> +	struct resource *res;
> +	size_t max_regions;
> +	int ret;
> +
> +	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> +	if (!ep)
> +		return -ENOMEM;
> +
> +	platform_set_drvdata(pdev, ep);
> +
> +	pcie = &ep->pcie;
> +	pcie->is_rc = false;
> +
> +	of_id = of_match_node(cdns_pcie_ep_of_match, np);
> +	ep->data = (const struct cdns_pcie_ep_data *)of_id->data;
> +
> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "reg");
> +	pcie->reg_base = devm_ioremap_resource(dev, res);
> +	if (IS_ERR(pcie->reg_base)) {
> +		dev_err(dev, "missing \"reg\"\n");
> +		return PTR_ERR(pcie->reg_base);
> +	}
> +
> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "mem");
> +	if (!res) {
> +		dev_err(dev, "missing \"mem\"\n");
> +		return -EINVAL;
> +	}
> +	pcie->mem_res = res;
> +
> +	max_regions = ep->data->max_regions;
> +	ep->ob_addr = devm_kzalloc(dev, max_regions * sizeof(*ep->ob_addr),
> +				   GFP_KERNEL);
> +	if (!ep->ob_addr)
> +		return -ENOMEM;
> +
> +	pm_runtime_enable(dev);
> +	ret = pm_runtime_get_sync(dev);
> +	if (ret < 0) {
> +		dev_err(dev, "pm_runtime_get_sync() failed\n");
> +		goto err_get_sync;
> +	}
> +
> +	/* Disable all but function 0 (anyway BIT(0) is hardwired to 1). */
> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, BIT(0));
> +
> +	epc = devm_pci_epc_create(dev, &cdns_pcie_epc_ops);
> +	if (IS_ERR(epc)) {
> +		dev_err(dev, "failed to create epc device\n");
> +		ret = PTR_ERR(epc);
> +		goto err_init;
> +	}
> +
> +	ep->epc = epc;
> +	epc_set_drvdata(epc, ep);
> +
> +	ret = of_property_read_u8(np, "max-functions", &epc->max_functions);
> +	if (ret < 0)
> +		epc->max_functions = 1;
> +
> +	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> +			       resource_size(pcie->mem_res));
> +	if (ret < 0) {
> +		dev_err(dev, "failed to initialize the memory space\n");
> +		goto err_init;
> +	}
> +
> +	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> +						  SZ_128K);
> +	if (!ep->irq_cpu_addr) {
> +		dev_err(dev, "failed to reserve memory space for MSI\n");
> +		ret = -ENOMEM;
> +		goto free_epc_mem;
> +	}
> +	ep->irq_pci_addr = CDNS_PCIE_EP_IRQ_PCI_ADDR_NONE;
> +
> +	return 0;
> +
> + free_epc_mem:
> +	pci_epc_mem_exit(epc);
> +
> + err_init:
> +	pm_runtime_put_sync(dev);
> +
> + err_get_sync:
> +	pm_runtime_disable(dev);
> +
> +	return ret;
> +}
> +
> +static struct platform_driver cdns_pcie_ep_driver = {
> +	.driver = {
> +		.name = "cdns-pcie-ep",
> +		.of_match_table = cdns_pcie_ep_of_match,
> +	},
> +	.probe = cdns_pcie_ep_probe,
> +};
> +builtin_platform_driver(cdns_pcie_ep_driver);
> -- 
> 2.11.0
>
Cyrille Pitchen Dec. 4, 2017, 2:56 p.m. UTC | #2
Hi Lorenzo,

Le 01/12/2017 à 13:20, Lorenzo Pieralisi a écrit :
> On Thu, Nov 23, 2017 at 04:01:50PM +0100, Cyrille Pitchen wrote:
>> This patch adds support to the Cadence PCIe controller in endpoint mode.
> 
> Please add a brief description to the log to describe the most salient
> features.
> 
>> Signed-off-by: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>> ---
>>  drivers/pci/cadence/Kconfig           |   9 +
>>  drivers/pci/cadence/Makefile          |   1 +
>>  drivers/pci/cadence/pcie-cadence-ep.c | 553 ++++++++++++++++++++++++++++++++++
>>  3 files changed, 563 insertions(+)
>>  create mode 100644 drivers/pci/cadence/pcie-cadence-ep.c
>>
>> diff --git a/drivers/pci/cadence/Kconfig b/drivers/pci/cadence/Kconfig
>> index 120306cae2aa..b2e6af71f39e 100644
>> --- a/drivers/pci/cadence/Kconfig
>> +++ b/drivers/pci/cadence/Kconfig
>> @@ -21,4 +21,13 @@ config PCIE_CADENCE_HOST
>>  	  mode. This PCIe controller may be embedded into many different vendors
>>  	  SoCs.
>>  
>> +config PCIE_CADENCE_EP
>> +	bool "Cadence PCIe endpoint controller"
>> +	depends on PCI_ENDPOINT
>> +	select PCIE_CADENCE
>> +	help
>> +	  Say Y here if you want to support the Cadence PCIe  controller in
>> +	  endpoint mode. This PCIe controller may be embedded into many
>> +	  different vendors SoCs.
>> +
>>  endif # PCI_CADENCE
>> diff --git a/drivers/pci/cadence/Makefile b/drivers/pci/cadence/Makefile
>> index d57d192d2595..61e9c8d6839d 100644
>> --- a/drivers/pci/cadence/Makefile
>> +++ b/drivers/pci/cadence/Makefile
>> @@ -1,2 +1,3 @@
>>  obj-$(CONFIG_PCIE_CADENCE) += pcie-cadence.o
>>  obj-$(CONFIG_PCIE_CADENCE_HOST) += pcie-cadence-host.o
>> +obj-$(CONFIG_PCIE_CADENCE_EP) += pcie-cadence-ep.o
>> diff --git a/drivers/pci/cadence/pcie-cadence-ep.c b/drivers/pci/cadence/pcie-cadence-ep.c
>> new file mode 100644
>> index 000000000000..a1d761101a9c
>> --- /dev/null
>> +++ b/drivers/pci/cadence/pcie-cadence-ep.c
>> @@ -0,0 +1,553 @@
>> +/*
>> + * Cadence PCIe host controller driver.
> 
> You should update this comment.
> 
>> + *
>> + * Copyright (c) 2017 Cadence
>> + *
>> + * Author: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + *
>> + * This program is distributed in the hope that it will be useful,
>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> + * GNU General Public License for more details.
>> + *
>> + * You should have received a copy of the GNU General Public License along with
>> + * this program.  If not, see <http://www.gnu.org/licenses/>.
>> + */
>> +#include <linux/kernel.h>
>> +#include <linux/of.h>
>> +#include <linux/pci-epc.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/pm_runtime.h>
>> +#include <linux/sizes.h>
>> +#include <linux/delay.h>
> 
> Nit: alphabetical order.
> 
>> +#include "pcie-cadence.h"
>> +
>> +#define CDNS_PCIE_EP_MIN_APERTURE		128	/* 128 bytes */
>> +#define CDNS_PCIE_EP_IRQ_PCI_ADDR_NONE		0x1
>> +#define CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY	0x3
>> +
>> +/**
>> + * struct cdns_pcie_ep_data - hardware specific data
>> + * @max_regions: maximum nmuber of regions supported by hardware
> 
> s/nmuber/number
> 
>> + */
>> +struct cdns_pcie_ep_data {
>> +	size_t				max_regions;
>> +};
>> +
>> +/**
>> + * struct cdns_pcie_ep - private data for this PCIe endpoint controller driver
>> + * @pcie: Cadence PCIe controller
>> + * @data: pointer to a 'struct cdns_pcie_data'
>> + */
>> +struct cdns_pcie_ep {
>> +	struct cdns_pcie		pcie;
>> +	const struct cdns_pcie_ep_data	*data;
>> +	struct pci_epc			*epc;
>> +	unsigned long			ob_region_map;
>> +	phys_addr_t			*ob_addr;
>> +	phys_addr_t			irq_phys_addr;
>> +	void __iomem			*irq_cpu_addr;
>> +	u64				irq_pci_addr;
>> +	u8				irq_pending;
>> +};
>> +
>> +static int cdns_pcie_ep_write_header(struct pci_epc *epc,
>> +				     struct pci_epf_header *hdr)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u8 fn = 0;
>> +
>> +	if (fn == 0) {
> 
> I think there is some code to retrieve fn missing here.
>

I've used some "u8 fn = 0;" when implementing every handler of
'struct pci_epc_ops' because the current API of the EPC library supports
only one function for now but I guess this will change soon.

So I've implement each handler as if a new 'fn' parameter were added to the
handler prototype. I guess it would make sense to add such a parameter to
the pci_epc_*() functions in pci-epc-core.c.
 
>> +		u32 id = CDNS_PCIE_LM_ID_VENDOR(hdr->vendorid) |
>> +			 CDNS_PCIE_LM_ID_SUBSYS(hdr->subsys_vendor_id);
>> +		cdns_pcie_writel(pcie, CDNS_PCIE_LM_ID, id);
>> +	}
>> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_DEVICE_ID, hdr->deviceid);
>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_REVISION_ID, hdr->revid);
>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_CLASS_PROG, hdr->progif_code);
>> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_CLASS_DEVICE,
>> +			       hdr->subclass_code | hdr->baseclass_code << 8);
>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_CACHE_LINE_SIZE,
>> +			       hdr->cache_line_size);
>> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_SUBSYSTEM_ID, hdr->subsys_id);
>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_INTERRUPT_PIN, hdr->interrupt_pin);
>> +
>> +	return 0;
>> +}
>> +
>> +static int cdns_pcie_ep_set_bar(struct pci_epc *epc, enum pci_barno bar,
>> +				dma_addr_t bar_phys, size_t size, int flags)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 addr0, addr1, reg, cfg, b, aperture, ctrl;
>> +	u8 fn = 0;
>> +	u64 sz;
>> +
>> +	/* BAR size is 2^(aperture + 7) */
>> +	sz = max_t(size_t, size, CDNS_PCIE_EP_MIN_APERTURE);
>> +	sz = 1ULL << fls64(sz - 1);
>> +	aperture = ilog2(sz) - 7; /* 128B -> 0, 256B -> 1, 512B -> 2, ... */
>> +
>> +	if ((flags & PCI_BASE_ADDRESS_SPACE) == PCI_BASE_ADDRESS_SPACE_IO) {
>> +		ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_IO_32BITS;
>> +	} else {
>> +		bool is_prefetch = !!(flags & PCI_BASE_ADDRESS_MEM_PREFETCH);
>> +		bool is_64bits = sz > SZ_2G;
>> +
>> +		if (is_64bits && (bar & 1))
>> +			return -EINVAL;
>> +
>> +		switch (is_64bits << 1 | is_prefetch) {
> 
> I would not mind implementing this as a nested if-else, I am not a big
> fan of using bool this way.
>

It did it this way to cope with the 80 columns rule but I will try to do it
with nested if-elst statements. The macro names I've chosen are quite long!

>> +		case 0:
>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_MEM_32BITS;
>> +			break;
>> +
>> +		case 1:
>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_PREFETCH_MEM_32BITS;
>> +			break;
>> +
>> +		case 2:
>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_MEM_64BITS;
>> +			break;
>> +
>> +		case 3:
>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_PREFETCH_MEM_64BITS;
>> +			break;
>> +		}
>> +	}
>> +
>> +	addr0 = lower_32_bits(bar_phys);
>> +	addr1 = upper_32_bits(bar_phys);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar),
>> +			 addr0);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar),
>> +			 addr1);
> 
> Is fn always 0 ?
>

If the API of the EPC library is updated, fn could be something else but 0
later but for now, fn is actually always 0.

>> +	if (bar < BAR_4) {
>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG0(fn);
>> +		b = bar;
>> +	} else {
>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG1(fn);
>> +		b = bar - BAR_4;
>> +	}
>> +
>> +	cfg = cdns_pcie_readl(pcie, reg);
>> +	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
>> +		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
>> +	cfg |= (CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE(b, aperture) |
>> +		CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL(b, ctrl));
>> +	cdns_pcie_writel(pcie, reg, cfg);
>> +
>> +	return 0;
>> +}
>> +
>> +static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, enum pci_barno bar)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 reg, cfg, b, ctrl;
>> +	u8 fn = 0;
>> +
>> +	if (bar < BAR_4) {
>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG0(fn);
>> +		b = bar;
>> +	} else {
>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG1(fn);
>> +		b = bar - BAR_4;
>> +	}
>> +
>> +	ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_DISABLED;
>> +	cfg = cdns_pcie_readl(pcie, reg);
>> +	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
>> +		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
>> +	cfg |= CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(ctrl);
>> +	cdns_pcie_writel(pcie, reg, cfg);
>> +
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar), 0);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar), 0);
>> +}
>> +
>> +static int cdns_pcie_ep_map_addr(struct pci_epc *epc, phys_addr_t addr,
>> +				 u64 pci_addr, size_t size)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 r;
>> +
>> +	r = find_first_zero_bit(&ep->ob_region_map, sizeof(ep->ob_region_map));
> 
> Second argument must be in bits not bytes.
> 
> https://marc.info/?l=linux-pci&m=151179781225513&w=2
>

OK, thanks!

>> +	if (r >= ep->data->max_regions - 1) {
>> +		dev_err(&epc->dev, "no free outbound region\n");
>> +		return -EINVAL;
>> +	}
>> +
>> +	cdns_pcie_set_outbound_region(pcie, r, false, addr, pci_addr, size);
>> +
>> +	set_bit(r, &ep->ob_region_map);
>> +	ep->ob_addr[r] = addr;
>> +
>> +	return 0;
>> +}
>> +
>> +static void cdns_pcie_ep_unmap_addr(struct pci_epc *epc, phys_addr_t addr)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 r;
>> +
>> +	for (r = 0; r < ep->data->max_regions - 1; r++)
>> +		if (ep->ob_addr[r] == addr)
>> +			break;
>> +
>> +	if (r >= ep->data->max_regions - 1)
> 
> == ?
>

I will change that.

>> +		return;
>> +
>> +	cdns_pcie_reset_outbound_region(pcie, r);
>> +
>> +	ep->ob_addr[r] = 0;
>> +	clear_bit(r, &ep->ob_region_map);
>> +}
>> +
>> +static int cdns_pcie_ep_set_msi(struct pci_epc *epc, u8 mmc)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
>> +	u16 flags;
>> +	u8 fn = 0;
>> +
>> +	/* Validate the ID of the MSI Capability structure. */
>> +	if (cdns_pcie_ep_fn_readb(pcie, fn, cap) != PCI_CAP_ID_MSI)
>> +		return -EINVAL;
>> +
>> +	/*
>> +	 * Set the Multiple Message Capable bitfield into the Message Control
>> +	 * register.
>> +	 */
>> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
>> +	flags = (flags & ~PCI_MSI_FLAGS_QMASK) | (mmc << 1);
>> +	flags |= PCI_MSI_FLAGS_64BIT;
>> +	flags &= ~PCI_MSI_FLAGS_MASKBIT;
>> +	cdns_pcie_ep_fn_writew(pcie, fn, cap + PCI_MSI_FLAGS, flags);
>> +
>> +	return 0;
>> +}
>> +
>> +static int cdns_pcie_ep_get_msi(struct pci_epc *epc)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
>> +	u16 flags, mmc, mme;
>> +	u8 fn = 0;
>> +
>> +	/* Validate the ID of the MSI Capability structure. */
>> +	if (cdns_pcie_ep_fn_readb(pcie, fn, cap) != PCI_CAP_ID_MSI)
>> +		return -EINVAL;
>> +
>> +	/* Validate that the MSI feature is actually enabled. */
>> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
>> +	if (!(flags & PCI_MSI_FLAGS_ENABLE))
>> +		return -EINVAL;
>> +
>> +	/*
>> +	 * Get the Multiple Message Enable bitfield from the Message Control
>> +	 * register.
>> +	 */
>> +	mmc = (flags & PCI_MSI_FLAGS_QMASK) >> 1;
>> +	mme = (flags & PCI_MSI_FLAGS_QSIZE) >> 4;
>> +	if (mme > mmc)
>> +		mme = mmc;
>> +	if (mme > 5)
>> +		mme = 5;
> 
> You should comment on what this 5 means and why it is fine to cap mme.
> 
>> +
>> +	return mme;
>> +}
>> +
>> +static void cdns_pcie_ep_assert_intx(struct cdns_pcie_ep *ep, u8 fn,
>> +				     u8 intx, bool is_asserted)
>> +{
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 r = ep->data->max_regions - 1;
>> +	u32 offset;
>> +	u16 status;
>> +	u8 msg_code;
>> +
>> +	intx &= 3;
>> +
>> +	/* Set the outbound region if needed. */
>> +	if (unlikely(ep->irq_pci_addr != CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY)) {
>> +		/* Last region was reserved for IRQ writes. */
>> +		cdns_pcie_set_outbound_region_for_normal_msg(pcie, r,
>> +							     ep->irq_phys_addr);
>> +		ep->irq_pci_addr = CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY;
>> +	}
>> +
>> +	if (is_asserted) {
>> +		ep->irq_pending |= BIT(intx);
>> +		msg_code = MSG_CODE_ASSERT_INTA + intx;
>> +	} else {
>> +		ep->irq_pending &= ~BIT(intx);
>> +		msg_code = MSG_CODE_DEASSERT_INTA + intx;
>> +	}
>> +
>> +	status = cdns_pcie_ep_fn_readw(pcie, fn, PCI_STATUS);
>> +	if (((status & PCI_STATUS_INTERRUPT) != 0) ^ (ep->irq_pending != 0)) {
>> +		status ^= PCI_STATUS_INTERRUPT;
>> +		cdns_pcie_ep_fn_writew(pcie, fn, PCI_STATUS, status);
>> +	}
>> +
>> +	offset = CDNS_PCIE_NORMAL_MSG_ROUTING(MSG_ROUTING_LOCAL) |
>> +		 CDNS_PCIE_NORMAL_MSG_CODE(msg_code) |
>> +		 CDNS_PCIE_MSG_NO_DATA;
>> +	writel(0, ep->irq_cpu_addr + offset);
>> +}
>> +
>> +static int cdns_pcie_ep_send_legacy_irq(struct cdns_pcie_ep *ep, u8 fn, u8 intx)
>> +{
>> +	u16 cmd;
>> +
>> +	cmd = cdns_pcie_ep_fn_readw(&ep->pcie, fn, PCI_COMMAND);
>> +	if (cmd & PCI_COMMAND_INTX_DISABLE)
>> +		return -EINVAL;
>> +
>> +	cdns_pcie_ep_assert_intx(ep, fn, intx, true);
>> +	mdelay(1);
> 
> Add a comment please to explain the mdelay value.
>

Honestly, I don't know for this value: I've just adapted the source code
from what was already done by dra7xx_pcie_raise_legacy_irq() from
drivers/pci/dwc/pci-dra7xx.c.

>> +	cdns_pcie_ep_assert_intx(ep, fn, intx, false);
>> +	return 0;
>> +}
>> +
>> +static int cdns_pcie_ep_raise_irq(struct pci_epc *epc,
>> +				  enum pci_epc_irq_type type, u8 interrupt_num)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
>> +	u16 flags, mmc, mme, data, data_mask;
>> +	u8 msi_count;
>> +	u64 pci_addr, pci_addr_mask = 0xff;
>> +	u8 fn = 0;
>> +
>> +	/* Handle legacy IRQ. */
>> +	if (type == PCI_EPC_IRQ_LEGACY)
>> +		return cdns_pcie_ep_send_legacy_irq(ep, fn, 0);
>> +
>> +	/* Otherwise MSI. */
>> +	if (type != PCI_EPC_IRQ_MSI)
>> +		return -EINVAL;
>> +
>> +	/* Check whether the MSI feature has been enabled by the PCI host. */
>> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
>> +	if (!(flags & PCI_MSI_FLAGS_ENABLE))
>> +		return -EINVAL;
>> +
>> +	/* Get the number of enabled MSIs */
>> +	mmc = (flags & PCI_MSI_FLAGS_QMASK) >> 1;
>> +	mme = (flags & PCI_MSI_FLAGS_QSIZE) >> 4;
>> +	if (mme > mmc)
>> +		mme = mmc;
>> +	if (mme > 5)
>> +		mme = 5;
> 
> Same comment as above.
> 
>> +	msi_count = 1 << mme;
>> +	if (!interrupt_num || interrupt_num > msi_count)
>> +		return -EINVAL;
>> +
>> +	/* Compute the data value to be written. */
>> +	data_mask = msi_count - 1;
>> +	data = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_DATA_64);
>> +	data = (data & ~data_mask) | ((interrupt_num - 1) & data_mask);
>> +
>> +	/* Get the PCI address where to write the data into. */
>> +	pci_addr = cdns_pcie_ep_fn_readl(pcie, fn, cap + PCI_MSI_ADDRESS_HI);
>> +	pci_addr <<= 32;
>> +	pci_addr |= cdns_pcie_ep_fn_readl(pcie, fn, cap + PCI_MSI_ADDRESS_LO);
>> +	pci_addr &= GENMASK_ULL(63, 2);
>> +
>> +	/* Set the outbound region if needed. */
>> +	if (unlikely(ep->irq_pci_addr != pci_addr)) {
>> +		/* Last region was reserved for IRQ writes. */
>> +		cdns_pcie_set_outbound_region(pcie, ep->data->max_regions - 1,
>> +					      false,
>> +					      ep->irq_phys_addr,
>> +					      pci_addr & ~pci_addr_mask,
>> +					      pci_addr_mask + 1);
>> +		ep->irq_pci_addr = pci_addr;
>> +	}
>> +	writew(data, ep->irq_cpu_addr + (pci_addr & pci_addr_mask));
>> +
>> +	return 0;
>> +}
>> +
>> +static int cdns_pcie_ep_start(struct pci_epc *epc)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	struct pci_epf *epf;
>> +	u32 cfg;
>> +	u8 fn = 0;
>> +
>> +	/* Enable this endpoint function. */
>> +	cfg = cdns_pcie_readl(pcie, CDNS_PCIE_LM_EP_FUNC_CFG);
>> +	cfg |= BIT(fn);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, cfg);
>> +
>> +	/*
>> +	 * Already linked-up: don't call directly pci_epc_linkup() because we've
>> +	 * already locked the epc->lock.
>> +	 */
>> +	list_for_each_entry(epf, &epc->pci_epf, list)
>> +		pci_epf_linkup(epf);
>> +
>> +	return 0;
>> +}
>> +
>> +static void cdns_pcie_ep_stop(struct pci_epc *epc)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 cfg;
>> +	u8 fn = 0;
>> +
>> +	/* Disable this endpoint function (function 0 can't be disabled). */
> 
> I do not understand this comment and how it applies to the code,
> in other words fn is always 0 here (so it can't be disabled)
> I do not understand what this code is there for.
>

fn == 0 for the same reason as for other 'struct pci_epc_ops' handlers.
I added this comment above because BIT(0) is hard-wired to 1: in other word
function 0 is the only one that can't be disabled.

So the current code actually does nothing. I've implemented anyway still
hoping that the API would change so function other than 0 could be used.

>> +	cfg = cdns_pcie_readl(pcie, CDNS_PCIE_LM_EP_FUNC_CFG);
>> +	cfg &= ~BIT(fn);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, cfg);
>> +}
>> +
>> +static const struct pci_epc_ops cdns_pcie_epc_ops = {
>> +	.write_header	= cdns_pcie_ep_write_header,
>> +	.set_bar	= cdns_pcie_ep_set_bar,
>> +	.clear_bar	= cdns_pcie_ep_clear_bar,
>> +	.map_addr	= cdns_pcie_ep_map_addr,
>> +	.unmap_addr	= cdns_pcie_ep_unmap_addr,
>> +	.set_msi	= cdns_pcie_ep_set_msi,
>> +	.get_msi	= cdns_pcie_ep_get_msi,
>> +	.raise_irq	= cdns_pcie_ep_raise_irq,
>> +	.start		= cdns_pcie_ep_start,
>> +	.stop		= cdns_pcie_ep_stop,
>> +};
>> +
>> +static const struct cdns_pcie_ep_data cdns_pcie_ep_data = {
>> +	.max_regions	= 16,
>> +};
> 
> As I mentioned in patch 3, should this be set-up with DT ?
>

OK, I'll move it as a DT property.

Best regards,

Cyrille

> Thanks,
> Lorenzo
> 
>> +
>> +static const struct of_device_id cdns_pcie_ep_of_match[] = {
>> +	{ .compatible = "cdns,cdns-pcie-ep",
>> +	  .data = &cdns_pcie_ep_data },
>> +
>> +	{ },
>> +};
>> +
>> +static int cdns_pcie_ep_probe(struct platform_device *pdev)
>> +{
>> +	struct device *dev = &pdev->dev;
>> +	struct device_node *np = dev->of_node;
>> +	const struct of_device_id *of_id;
>> +	struct cdns_pcie_ep *ep;
>> +	struct cdns_pcie *pcie;
>> +	struct pci_epc *epc;
>> +	struct resource *res;
>> +	size_t max_regions;
>> +	int ret;
>> +
>> +	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>> +	if (!ep)
>> +		return -ENOMEM;
>> +
>> +	platform_set_drvdata(pdev, ep);
>> +
>> +	pcie = &ep->pcie;
>> +	pcie->is_rc = false;
>> +
>> +	of_id = of_match_node(cdns_pcie_ep_of_match, np);
>> +	ep->data = (const struct cdns_pcie_ep_data *)of_id->data;
>> +
>> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "reg");
>> +	pcie->reg_base = devm_ioremap_resource(dev, res);
>> +	if (IS_ERR(pcie->reg_base)) {
>> +		dev_err(dev, "missing \"reg\"\n");
>> +		return PTR_ERR(pcie->reg_base);
>> +	}
>> +
>> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "mem");
>> +	if (!res) {
>> +		dev_err(dev, "missing \"mem\"\n");
>> +		return -EINVAL;
>> +	}
>> +	pcie->mem_res = res;
>> +
>> +	max_regions = ep->data->max_regions;
>> +	ep->ob_addr = devm_kzalloc(dev, max_regions * sizeof(*ep->ob_addr),
>> +				   GFP_KERNEL);
>> +	if (!ep->ob_addr)
>> +		return -ENOMEM;
>> +
>> +	pm_runtime_enable(dev);
>> +	ret = pm_runtime_get_sync(dev);
>> +	if (ret < 0) {
>> +		dev_err(dev, "pm_runtime_get_sync() failed\n");
>> +		goto err_get_sync;
>> +	}
>> +
>> +	/* Disable all but function 0 (anyway BIT(0) is hardwired to 1). */
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, BIT(0));
>> +
>> +	epc = devm_pci_epc_create(dev, &cdns_pcie_epc_ops);
>> +	if (IS_ERR(epc)) {
>> +		dev_err(dev, "failed to create epc device\n");
>> +		ret = PTR_ERR(epc);
>> +		goto err_init;
>> +	}
>> +
>> +	ep->epc = epc;
>> +	epc_set_drvdata(epc, ep);
>> +
>> +	ret = of_property_read_u8(np, "max-functions", &epc->max_functions);
>> +	if (ret < 0)
>> +		epc->max_functions = 1;
>> +
>> +	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
>> +			       resource_size(pcie->mem_res));
>> +	if (ret < 0) {
>> +		dev_err(dev, "failed to initialize the memory space\n");
>> +		goto err_init;
>> +	}
>> +
>> +	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
>> +						  SZ_128K);
>> +	if (!ep->irq_cpu_addr) {
>> +		dev_err(dev, "failed to reserve memory space for MSI\n");
>> +		ret = -ENOMEM;
>> +		goto free_epc_mem;
>> +	}
>> +	ep->irq_pci_addr = CDNS_PCIE_EP_IRQ_PCI_ADDR_NONE;
>> +
>> +	return 0;
>> +
>> + free_epc_mem:
>> +	pci_epc_mem_exit(epc);
>> +
>> + err_init:
>> +	pm_runtime_put_sync(dev);
>> +
>> + err_get_sync:
>> +	pm_runtime_disable(dev);
>> +
>> +	return ret;
>> +}
>> +
>> +static struct platform_driver cdns_pcie_ep_driver = {
>> +	.driver = {
>> +		.name = "cdns-pcie-ep",
>> +		.of_match_table = cdns_pcie_ep_of_match,
>> +	},
>> +	.probe = cdns_pcie_ep_probe,
>> +};
>> +builtin_platform_driver(cdns_pcie_ep_driver);
>> -- 
>> 2.11.0
>>
>
Kishon Vijay Abraham I Dec. 5, 2017, 9:19 a.m. UTC | #3
Hi,

On Friday 01 December 2017 05:50 PM, Lorenzo Pieralisi wrote:
> On Thu, Nov 23, 2017 at 04:01:50PM +0100, Cyrille Pitchen wrote:
>> This patch adds support to the Cadence PCIe controller in endpoint mode.
> 
> Please add a brief description to the log to describe the most salient
> features.
> 
>> Signed-off-by: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>> ---
>>  drivers/pci/cadence/Kconfig           |   9 +
>>  drivers/pci/cadence/Makefile          |   1 +
>>  drivers/pci/cadence/pcie-cadence-ep.c | 553 ++++++++++++++++++++++++++++++++++
>>  3 files changed, 563 insertions(+)
>>  create mode 100644 drivers/pci/cadence/pcie-cadence-ep.c
>>
>> diff --git a/drivers/pci/cadence/Kconfig b/drivers/pci/cadence/Kconfig
>> index 120306cae2aa..b2e6af71f39e 100644
>> --- a/drivers/pci/cadence/Kconfig
>> +++ b/drivers/pci/cadence/Kconfig
>> @@ -21,4 +21,13 @@ config PCIE_CADENCE_HOST
>>  	  mode. This PCIe controller may be embedded into many different vendors
>>  	  SoCs.
>>  
>> +config PCIE_CADENCE_EP
>> +	bool "Cadence PCIe endpoint controller"
>> +	depends on PCI_ENDPOINT
>> +	select PCIE_CADENCE
>> +	help
>> +	  Say Y here if you want to support the Cadence PCIe  controller in
>> +	  endpoint mode. This PCIe controller may be embedded into many
>> +	  different vendors SoCs.
>> +
>>  endif # PCI_CADENCE
>> diff --git a/drivers/pci/cadence/Makefile b/drivers/pci/cadence/Makefile
>> index d57d192d2595..61e9c8d6839d 100644
>> --- a/drivers/pci/cadence/Makefile
>> +++ b/drivers/pci/cadence/Makefile
>> @@ -1,2 +1,3 @@
>>  obj-$(CONFIG_PCIE_CADENCE) += pcie-cadence.o
>>  obj-$(CONFIG_PCIE_CADENCE_HOST) += pcie-cadence-host.o
>> +obj-$(CONFIG_PCIE_CADENCE_EP) += pcie-cadence-ep.o
>> diff --git a/drivers/pci/cadence/pcie-cadence-ep.c b/drivers/pci/cadence/pcie-cadence-ep.c
>> new file mode 100644
>> index 000000000000..a1d761101a9c
>> --- /dev/null
>> +++ b/drivers/pci/cadence/pcie-cadence-ep.c
>> @@ -0,0 +1,553 @@
>> +/*
>> + * Cadence PCIe host controller driver.
> 
> You should update this comment.
> 
>> + *
>> + * Copyright (c) 2017 Cadence
>> + *
>> + * Author: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + *
>> + * This program is distributed in the hope that it will be useful,
>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> + * GNU General Public License for more details.
>> + *
>> + * You should have received a copy of the GNU General Public License along with
>> + * this program.  If not, see <http://www.gnu.org/licenses/>.
>> + */
>> +#include <linux/kernel.h>
>> +#include <linux/of.h>
>> +#include <linux/pci-epc.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/pm_runtime.h>
>> +#include <linux/sizes.h>
>> +#include <linux/delay.h>
> 
> Nit: alphabetical order.
> 
>> +#include "pcie-cadence.h"
>> +
>> +#define CDNS_PCIE_EP_MIN_APERTURE		128	/* 128 bytes */
>> +#define CDNS_PCIE_EP_IRQ_PCI_ADDR_NONE		0x1
>> +#define CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY	0x3
>> +
>> +/**
>> + * struct cdns_pcie_ep_data - hardware specific data
>> + * @max_regions: maximum nmuber of regions supported by hardware
> 
> s/nmuber/number
> 
>> + */
>> +struct cdns_pcie_ep_data {
>> +	size_t				max_regions;
>> +};
>> +
>> +/**
>> + * struct cdns_pcie_ep - private data for this PCIe endpoint controller driver
>> + * @pcie: Cadence PCIe controller
>> + * @data: pointer to a 'struct cdns_pcie_data'
>> + */
>> +struct cdns_pcie_ep {
>> +	struct cdns_pcie		pcie;
>> +	const struct cdns_pcie_ep_data	*data;
>> +	struct pci_epc			*epc;
>> +	unsigned long			ob_region_map;
>> +	phys_addr_t			*ob_addr;
>> +	phys_addr_t			irq_phys_addr;
>> +	void __iomem			*irq_cpu_addr;
>> +	u64				irq_pci_addr;
>> +	u8				irq_pending;
>> +};
>> +
>> +static int cdns_pcie_ep_write_header(struct pci_epc *epc,
>> +				     struct pci_epf_header *hdr)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u8 fn = 0;
>> +
>> +	if (fn == 0) {
> 
> I think there is some code to retrieve fn missing here.

hmm.. the endpoint core has to send the function number which right now it's
not doing though it has the function number info in pci_epf.
> 
>> +		u32 id = CDNS_PCIE_LM_ID_VENDOR(hdr->vendorid) |
>> +			 CDNS_PCIE_LM_ID_SUBSYS(hdr->subsys_vendor_id);
>> +		cdns_pcie_writel(pcie, CDNS_PCIE_LM_ID, id);
>> +	}
>> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_DEVICE_ID, hdr->deviceid);
>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_REVISION_ID, hdr->revid);
>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_CLASS_PROG, hdr->progif_code);
>> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_CLASS_DEVICE,
>> +			       hdr->subclass_code | hdr->baseclass_code << 8);
>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_CACHE_LINE_SIZE,
>> +			       hdr->cache_line_size);
>> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_SUBSYSTEM_ID, hdr->subsys_id);
>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_INTERRUPT_PIN, hdr->interrupt_pin);
>> +
>> +	return 0;
>> +}
>> +
>> +static int cdns_pcie_ep_set_bar(struct pci_epc *epc, enum pci_barno bar,
>> +				dma_addr_t bar_phys, size_t size, int flags)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 addr0, addr1, reg, cfg, b, aperture, ctrl;
>> +	u8 fn = 0;

Here too endpoint core should send the function number..
>> +	u64 sz;
>> +
>> +	/* BAR size is 2^(aperture + 7) */
>> +	sz = max_t(size_t, size, CDNS_PCIE_EP_MIN_APERTURE);
>> +	sz = 1ULL << fls64(sz - 1);
>> +	aperture = ilog2(sz) - 7; /* 128B -> 0, 256B -> 1, 512B -> 2, ... */
>> +
>> +	if ((flags & PCI_BASE_ADDRESS_SPACE) == PCI_BASE_ADDRESS_SPACE_IO) {
>> +		ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_IO_32BITS;
>> +	} else {
>> +		bool is_prefetch = !!(flags & PCI_BASE_ADDRESS_MEM_PREFETCH);
>> +		bool is_64bits = sz > SZ_2G;
>> +
>> +		if (is_64bits && (bar & 1))
>> +			return -EINVAL;
>> +
>> +		switch (is_64bits << 1 | is_prefetch) {
> 
> I would not mind implementing this as a nested if-else, I am not a big
> fan of using bool this way.
> 
>> +		case 0:
>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_MEM_32BITS;
>> +			break;
>> +
>> +		case 1:
>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_PREFETCH_MEM_32BITS;
>> +			break;
>> +
>> +		case 2:
>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_MEM_64BITS;
>> +			break;
>> +
>> +		case 3:
>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_PREFETCH_MEM_64BITS;
>> +			break;
>> +		}
>> +	}
>> +
>> +	addr0 = lower_32_bits(bar_phys);
>> +	addr1 = upper_32_bits(bar_phys);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar),
>> +			 addr0);

It would be nice if you can have defines for CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0
included in this patch rather than "PCI: cadence: Add host driver for Cadence
PCIe controller". All EP specific functions in header file should be included here.
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar),
>> +			 addr1);
> 
> Is fn always 0 ?
> 
>> +	if (bar < BAR_4) {
>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG0(fn);
>> +		b = bar;
>> +	} else {
>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG1(fn);
>> +		b = bar - BAR_4;
>> +	}
>> +
>> +	cfg = cdns_pcie_readl(pcie, reg);
>> +	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
>> +		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
>> +	cfg |= (CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE(b, aperture) |
>> +		CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL(b, ctrl));
>> +	cdns_pcie_writel(pcie, reg, cfg);
>> +
>> +	return 0;
>> +}
>> +
>> +static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, enum pci_barno bar)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 reg, cfg, b, ctrl;
>> +	u8 fn = 0;

Here too endpoint core should send the function number..
>> +
>> +	if (bar < BAR_4) {
>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG0(fn);
>> +		b = bar;
>> +	} else {
>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG1(fn);
>> +		b = bar - BAR_4;
>> +	}
>> +
>> +	ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_DISABLED;
>> +	cfg = cdns_pcie_readl(pcie, reg);
>> +	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
>> +		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
>> +	cfg |= CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(ctrl);
>> +	cdns_pcie_writel(pcie, reg, cfg);
>> +
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar), 0);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar), 0);
>> +}
>> +
>> +static int cdns_pcie_ep_map_addr(struct pci_epc *epc, phys_addr_t addr,
>> +				 u64 pci_addr, size_t size)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 r;
>> +
>> +	r = find_first_zero_bit(&ep->ob_region_map, sizeof(ep->ob_region_map));
> 
> Second argument must be in bits not bytes.
> 
> https://marc.info/?l=linux-pci&m=151179781225513&w=2
> 
>> +	if (r >= ep->data->max_regions - 1) {
>> +		dev_err(&epc->dev, "no free outbound region\n");
>> +		return -EINVAL;
>> +	}
>> +
>> +	cdns_pcie_set_outbound_region(pcie, r, false, addr, pci_addr, size);
>> +
>> +	set_bit(r, &ep->ob_region_map);
>> +	ep->ob_addr[r] = addr;
>> +
>> +	return 0;
>> +}
>> +
>> +static void cdns_pcie_ep_unmap_addr(struct pci_epc *epc, phys_addr_t addr)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 r;
>> +
>> +	for (r = 0; r < ep->data->max_regions - 1; r++)
>> +		if (ep->ob_addr[r] == addr)
>> +			break;
>> +
>> +	if (r >= ep->data->max_regions - 1)
> 
> == ?
> 
>> +		return;
>> +
>> +	cdns_pcie_reset_outbound_region(pcie, r);
>> +
>> +	ep->ob_addr[r] = 0;
>> +	clear_bit(r, &ep->ob_region_map);
>> +}
>> +
>> +static int cdns_pcie_ep_set_msi(struct pci_epc *epc, u8 mmc)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
>> +	u16 flags;
>> +	u8 fn = 0;
>> +
>> +	/* Validate the ID of the MSI Capability structure. */
>> +	if (cdns_pcie_ep_fn_readb(pcie, fn, cap) != PCI_CAP_ID_MSI)
>> +		return -EINVAL;
>> +
>> +	/*
>> +	 * Set the Multiple Message Capable bitfield into the Message Control
>> +	 * register.
>> +	 */
>> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
>> +	flags = (flags & ~PCI_MSI_FLAGS_QMASK) | (mmc << 1);
>> +	flags |= PCI_MSI_FLAGS_64BIT;
>> +	flags &= ~PCI_MSI_FLAGS_MASKBIT;

Any reason why "Per-vector masking capable" is reset?
>> +	cdns_pcie_ep_fn_writew(pcie, fn, cap + PCI_MSI_FLAGS, flags);
>> +
>> +	return 0;
>> +}
>> +
>> +static int cdns_pcie_ep_get_msi(struct pci_epc *epc)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
>> +	u16 flags, mmc, mme;
>> +	u8 fn = 0;
>> +
>> +	/* Validate the ID of the MSI Capability structure. */
>> +	if (cdns_pcie_ep_fn_readb(pcie, fn, cap) != PCI_CAP_ID_MSI)
>> +		return -EINVAL;
>> +
>> +	/* Validate that the MSI feature is actually enabled. */
>> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
>> +	if (!(flags & PCI_MSI_FLAGS_ENABLE))
>> +		return -EINVAL;
>> +
>> +	/*
>> +	 * Get the Multiple Message Enable bitfield from the Message Control
>> +	 * register.
>> +	 */
>> +	mmc = (flags & PCI_MSI_FLAGS_QMASK) >> 1;
>> +	mme = (flags & PCI_MSI_FLAGS_QSIZE) >> 4;
>> +	if (mme > mmc)
>> +		mme = mmc;
>> +	if (mme > 5)
>> +		mme = 5;

I'm not sure if both these above checks are required..
> 
> You should comment on what this 5 means and why it is fine to cap mme.
> 
>> +
>> +	return mme;
>> +}
>> +
>> +static void cdns_pcie_ep_assert_intx(struct cdns_pcie_ep *ep, u8 fn,
>> +				     u8 intx, bool is_asserted)
>> +{
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 r = ep->data->max_regions - 1;
>> +	u32 offset;
>> +	u16 status;
>> +	u8 msg_code;
>> +
>> +	intx &= 3;
>> +
>> +	/* Set the outbound region if needed. */
>> +	if (unlikely(ep->irq_pci_addr != CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY)) {
>> +		/* Last region was reserved for IRQ writes. */
>> +		cdns_pcie_set_outbound_region_for_normal_msg(pcie, r,
>> +							     ep->irq_phys_addr);
>> +		ep->irq_pci_addr = CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY;
>> +	}
>> +
>> +	if (is_asserted) {
>> +		ep->irq_pending |= BIT(intx);
>> +		msg_code = MSG_CODE_ASSERT_INTA + intx;
>> +	} else {
>> +		ep->irq_pending &= ~BIT(intx);
>> +		msg_code = MSG_CODE_DEASSERT_INTA + intx;
>> +	}
>> +
>> +	status = cdns_pcie_ep_fn_readw(pcie, fn, PCI_STATUS);
>> +	if (((status & PCI_STATUS_INTERRUPT) != 0) ^ (ep->irq_pending != 0)) {
>> +		status ^= PCI_STATUS_INTERRUPT;
>> +		cdns_pcie_ep_fn_writew(pcie, fn, PCI_STATUS, status);
>> +	}

here you are setting the PCI_STATUS_INTERRUPT even before sending the ASSERT
message.
>> +
>> +	offset = CDNS_PCIE_NORMAL_MSG_ROUTING(MSG_ROUTING_LOCAL) |
>> +		 CDNS_PCIE_NORMAL_MSG_CODE(msg_code) |
>> +		 CDNS_PCIE_MSG_NO_DATA;
>> +	writel(0, ep->irq_cpu_addr + offset);
>> +}
>> +
>> +static int cdns_pcie_ep_send_legacy_irq(struct cdns_pcie_ep *ep, u8 fn, u8 intx)
>> +{
>> +	u16 cmd;
>> +
>> +	cmd = cdns_pcie_ep_fn_readw(&ep->pcie, fn, PCI_COMMAND);
>> +	if (cmd & PCI_COMMAND_INTX_DISABLE)
>> +		return -EINVAL;
>> +
>> +	cdns_pcie_ep_assert_intx(ep, fn, intx, true);
>> +	mdelay(1);
> 
> Add a comment please to explain the mdelay value.
> 
>> +	cdns_pcie_ep_assert_intx(ep, fn, intx, false);
>> +	return 0;
>> +}
>> +
>> +static int cdns_pcie_ep_raise_irq(struct pci_epc *epc,
>> +				  enum pci_epc_irq_type type, u8 interrupt_num)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
>> +	u16 flags, mmc, mme, data, data_mask;
>> +	u8 msi_count;
>> +	u64 pci_addr, pci_addr_mask = 0xff;
>> +	u8 fn = 0;
>> +
>> +	/* Handle legacy IRQ. */
>> +	if (type == PCI_EPC_IRQ_LEGACY)
>> +		return cdns_pcie_ep_send_legacy_irq(ep, fn, 0);
>> +
>> +	/* Otherwise MSI. */
>> +	if (type != PCI_EPC_IRQ_MSI)
>> +		return -EINVAL;

MSI-X?
>> +
>> +	/* Check whether the MSI feature has been enabled by the PCI host. */
>> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
>> +	if (!(flags & PCI_MSI_FLAGS_ENABLE))
>> +		return -EINVAL;
>> +
>> +	/* Get the number of enabled MSIs */
>> +	mmc = (flags & PCI_MSI_FLAGS_QMASK) >> 1;
>> +	mme = (flags & PCI_MSI_FLAGS_QSIZE) >> 4;
>> +	if (mme > mmc)
>> +		mme = mmc;
>> +	if (mme > 5)
>> +		mme = 5;
> 
> Same comment as above.
> 
>> +	msi_count = 1 << mme;
>> +	if (!interrupt_num || interrupt_num > msi_count)
>> +		return -EINVAL;
>> +
>> +	/* Compute the data value to be written. */
>> +	data_mask = msi_count - 1;
>> +	data = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_DATA_64);
>> +	data = (data & ~data_mask) | ((interrupt_num - 1) & data_mask);
>> +
>> +	/* Get the PCI address where to write the data into. */
>> +	pci_addr = cdns_pcie_ep_fn_readl(pcie, fn, cap + PCI_MSI_ADDRESS_HI);
>> +	pci_addr <<= 32;
>> +	pci_addr |= cdns_pcie_ep_fn_readl(pcie, fn, cap + PCI_MSI_ADDRESS_LO);
>> +	pci_addr &= GENMASK_ULL(63, 2);
>> +
>> +	/* Set the outbound region if needed. */
>> +	if (unlikely(ep->irq_pci_addr != pci_addr)) {
>> +		/* Last region was reserved for IRQ writes. */
>> +		cdns_pcie_set_outbound_region(pcie, ep->data->max_regions - 1,
>> +					      false,
>> +					      ep->irq_phys_addr,
>> +					      pci_addr & ~pci_addr_mask,
>> +					      pci_addr_mask + 1);
>> +		ep->irq_pci_addr = pci_addr;
>> +	}
>> +	writew(data, ep->irq_cpu_addr + (pci_addr & pci_addr_mask));
>> +
>> +	return 0;
>> +}
>> +
>> +static int cdns_pcie_ep_start(struct pci_epc *epc)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	struct pci_epf *epf;
>> +	u32 cfg;
>> +	u8 fn = 0;
>> +
>> +	/* Enable this endpoint function. */
>> +	cfg = cdns_pcie_readl(pcie, CDNS_PCIE_LM_EP_FUNC_CFG);
>> +	cfg |= BIT(fn);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, cfg);
>> +
>> +	/*
>> +	 * Already linked-up: don't call directly pci_epc_linkup() because we've
>> +	 * already locked the epc->lock.
>> +	 */

Not sure what you mean by linked-up here?
>> +	list_for_each_entry(epf, &epc->pci_epf, list)
>> +		pci_epf_linkup(epf);

>> +
>> +	return 0;
>> +}
>> +
>> +static void cdns_pcie_ep_stop(struct pci_epc *epc)
>> +{
>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>> +	struct cdns_pcie *pcie = &ep->pcie;
>> +	u32 cfg;
>> +	u8 fn = 0;
>> +
>> +	/* Disable this endpoint function (function 0 can't be disabled). */
> 
> I do not understand this comment and how it applies to the code,
> in other words fn is always 0 here (so it can't be disabled)
> I do not understand what this code is there for.
> 
>> +	cfg = cdns_pcie_readl(pcie, CDNS_PCIE_LM_EP_FUNC_CFG);
>> +	cfg &= ~BIT(fn);
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, cfg);
>> +}
>> +
>> +static const struct pci_epc_ops cdns_pcie_epc_ops = {
>> +	.write_header	= cdns_pcie_ep_write_header,
>> +	.set_bar	= cdns_pcie_ep_set_bar,
>> +	.clear_bar	= cdns_pcie_ep_clear_bar,
>> +	.map_addr	= cdns_pcie_ep_map_addr,
>> +	.unmap_addr	= cdns_pcie_ep_unmap_addr,
>> +	.set_msi	= cdns_pcie_ep_set_msi,
>> +	.get_msi	= cdns_pcie_ep_get_msi,
>> +	.raise_irq	= cdns_pcie_ep_raise_irq,
>> +	.start		= cdns_pcie_ep_start,
>> +	.stop		= cdns_pcie_ep_stop,
>> +};
>> +
>> +static const struct cdns_pcie_ep_data cdns_pcie_ep_data = {
>> +	.max_regions	= 16,
>> +};
> 
> As I mentioned in patch 3, should this be set-up with DT ?
> 
> Thanks,
> Lorenzo
> 
>> +
>> +static const struct of_device_id cdns_pcie_ep_of_match[] = {
>> +	{ .compatible = "cdns,cdns-pcie-ep",
>> +	  .data = &cdns_pcie_ep_data },
>> +
>> +	{ },
>> +};
>> +
>> +static int cdns_pcie_ep_probe(struct platform_device *pdev)
>> +{
>> +	struct device *dev = &pdev->dev;
>> +	struct device_node *np = dev->of_node;
>> +	const struct of_device_id *of_id;
>> +	struct cdns_pcie_ep *ep;
>> +	struct cdns_pcie *pcie;
>> +	struct pci_epc *epc;
>> +	struct resource *res;
>> +	size_t max_regions;
>> +	int ret;
>> +
>> +	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>> +	if (!ep)
>> +		return -ENOMEM;
>> +
>> +	platform_set_drvdata(pdev, ep);
>> +
>> +	pcie = &ep->pcie;
>> +	pcie->is_rc = false;
>> +
>> +	of_id = of_match_node(cdns_pcie_ep_of_match, np);
>> +	ep->data = (const struct cdns_pcie_ep_data *)of_id->data;
>> +
>> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "reg");
>> +	pcie->reg_base = devm_ioremap_resource(dev, res);
>> +	if (IS_ERR(pcie->reg_base)) {
>> +		dev_err(dev, "missing \"reg\"\n");
>> +		return PTR_ERR(pcie->reg_base);
>> +	}
>> +
>> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "mem");
>> +	if (!res) {
>> +		dev_err(dev, "missing \"mem\"\n");
>> +		return -EINVAL;
>> +	}
>> +	pcie->mem_res = res;
>> +
>> +	max_regions = ep->data->max_regions;
>> +	ep->ob_addr = devm_kzalloc(dev, max_regions * sizeof(*ep->ob_addr),
>> +				   GFP_KERNEL);
>> +	if (!ep->ob_addr)
>> +		return -ENOMEM;
>> +
>> +	pm_runtime_enable(dev);
>> +	ret = pm_runtime_get_sync(dev);
>> +	if (ret < 0) {
>> +		dev_err(dev, "pm_runtime_get_sync() failed\n");
>> +		goto err_get_sync;
>> +	}
>> +
>> +	/* Disable all but function 0 (anyway BIT(0) is hardwired to 1). */
>> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, BIT(0));

why disable all functions?
>> +
>> +	epc = devm_pci_epc_create(dev, &cdns_pcie_epc_ops);
>> +	if (IS_ERR(epc)) {
>> +		dev_err(dev, "failed to create epc device\n");
>> +		ret = PTR_ERR(epc);
>> +		goto err_init;
>> +	}
>> +
>> +	ep->epc = epc;
>> +	epc_set_drvdata(epc, ep);
>> +
>> +	ret = of_property_read_u8(np, "max-functions", &epc->max_functions);
>> +	if (ret < 0)
>> +		epc->max_functions = 1;
>> +
>> +	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
>> +			       resource_size(pcie->mem_res));
>> +	if (ret < 0) {
>> +		dev_err(dev, "failed to initialize the memory space\n");
>> +		goto err_init;
>> +	}
>> +
>> +	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
>> +						  SZ_128K);

Any reason why you chose SZ_128K?

Thanks
Kishon
Philippe Ombredanne Dec. 7, 2017, 10:05 a.m. UTC | #4
Cyrille,


On Tue, Dec 5, 2017 at 10:19 AM, Kishon Vijay Abraham I <kishon@ti.com> wrote:
> Hi,
>
> On Friday 01 December 2017 05:50 PM, Lorenzo Pieralisi wrote:
>> On Thu, Nov 23, 2017 at 04:01:50PM +0100, Cyrille Pitchen wrote:
>>> This patch adds support to the Cadence PCIe controller in endpoint mode.
>>
>> Please add a brief description to the log to describe the most salient
>> features.
>>
>>> Signed-off-by: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>>> ---
>>>  drivers/pci/cadence/Kconfig           |   9 +
>>>  drivers/pci/cadence/Makefile          |   1 +
>>>  drivers/pci/cadence/pcie-cadence-ep.c | 553 ++++++++++++++++++++++++++++++++++
>>>  3 files changed, 563 insertions(+)
>>>  create mode 100644 drivers/pci/cadence/pcie-cadence-ep.c
>>>
>>> diff --git a/drivers/pci/cadence/Kconfig b/drivers/pci/cadence/Kconfig
>>> index 120306cae2aa..b2e6af71f39e 100644
>>> --- a/drivers/pci/cadence/Kconfig
>>> +++ b/drivers/pci/cadence/Kconfig
>>> @@ -21,4 +21,13 @@ config PCIE_CADENCE_HOST
>>>        mode. This PCIe controller may be embedded into many different vendors
>>>        SoCs.
>>>
>>> +config PCIE_CADENCE_EP
>>> +    bool "Cadence PCIe endpoint controller"
>>> +    depends on PCI_ENDPOINT
>>> +    select PCIE_CADENCE
>>> +    help
>>> +      Say Y here if you want to support the Cadence PCIe  controller in
>>> +      endpoint mode. This PCIe controller may be embedded into many
>>> +      different vendors SoCs.
>>> +
>>>  endif # PCI_CADENCE
>>> diff --git a/drivers/pci/cadence/Makefile b/drivers/pci/cadence/Makefile
>>> index d57d192d2595..61e9c8d6839d 100644
>>> --- a/drivers/pci/cadence/Makefile
>>> +++ b/drivers/pci/cadence/Makefile
>>> @@ -1,2 +1,3 @@
>>>  obj-$(CONFIG_PCIE_CADENCE) += pcie-cadence.o
>>>  obj-$(CONFIG_PCIE_CADENCE_HOST) += pcie-cadence-host.o
>>> +obj-$(CONFIG_PCIE_CADENCE_EP) += pcie-cadence-ep.o
>>> diff --git a/drivers/pci/cadence/pcie-cadence-ep.c b/drivers/pci/cadence/pcie-cadence-ep.c
>>> new file mode 100644
>>> index 000000000000..a1d761101a9c
>>> --- /dev/null
>>> +++ b/drivers/pci/cadence/pcie-cadence-ep.c
>>> @@ -0,0 +1,553 @@
>>> +/*
>>> + * Cadence PCIe host controller driver.
>>
>> You should update this comment.
>>
>>> + *
>>> + * Copyright (c) 2017 Cadence
>>> + *
>>> + * Author: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>>> + *
>>> + * This program is free software; you can redistribute it and/or modify
>>> + * it under the terms of the GNU General Public License version 2 as
>>> + * published by the Free Software Foundation.
>>> + *
>>> + * This program is distributed in the hope that it will be useful,
>>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>>> + * GNU General Public License for more details.
>>> + *
>>> + * You should have received a copy of the GNU General Public License along with
>>> + * this program.  If not, see <http://www.gnu.org/licenses/>.
>>> + */

Would you consider using the new SPDX ids instead of this long legalese blurb?
You can  check tglx's doc patches for details and Linus comments on
why he wants it a certain way with C++ // comments.

Here this could come out much streamlined with this top level comment block:

// SPDX-License-Identifier: GPL-2.0
// Copyright (c) 2017 Cadence
// Cadence PCIe host controller driver.
// Author: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>

This is only a suggestion, but everyone kinda likes it when the
licensing boilerplate is minimal and out of the visual way in order to
focus is on what matters most: the code!
Thank you for your kind consideration.
Cyrille Pitchen Dec. 13, 2017, 4:03 p.m. UTC | #5
Hi Philippe,

Le 07/12/2017 à 11:05, Philippe Ombredanne a écrit :
> Cyrille,
> 
> 
> On Tue, Dec 5, 2017 at 10:19 AM, Kishon Vijay Abraham I <kishon@ti.com> wrote:
>> Hi,
>>
>> On Friday 01 December 2017 05:50 PM, Lorenzo Pieralisi wrote:
>>> On Thu, Nov 23, 2017 at 04:01:50PM +0100, Cyrille Pitchen wrote:
>>>> This patch adds support to the Cadence PCIe controller in endpoint mode.
>>>
>>> Please add a brief description to the log to describe the most salient
>>> features.
>>>
>>>> Signed-off-by: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>>>> ---
>>>>  drivers/pci/cadence/Kconfig           |   9 +
>>>>  drivers/pci/cadence/Makefile          |   1 +
>>>>  drivers/pci/cadence/pcie-cadence-ep.c | 553 ++++++++++++++++++++++++++++++++++
>>>>  3 files changed, 563 insertions(+)
>>>>  create mode 100644 drivers/pci/cadence/pcie-cadence-ep.c
>>>>
>>>> diff --git a/drivers/pci/cadence/Kconfig b/drivers/pci/cadence/Kconfig
>>>> index 120306cae2aa..b2e6af71f39e 100644
>>>> --- a/drivers/pci/cadence/Kconfig
>>>> +++ b/drivers/pci/cadence/Kconfig
>>>> @@ -21,4 +21,13 @@ config PCIE_CADENCE_HOST
>>>>        mode. This PCIe controller may be embedded into many different vendors
>>>>        SoCs.
>>>>
>>>> +config PCIE_CADENCE_EP
>>>> +    bool "Cadence PCIe endpoint controller"
>>>> +    depends on PCI_ENDPOINT
>>>> +    select PCIE_CADENCE
>>>> +    help
>>>> +      Say Y here if you want to support the Cadence PCIe  controller in
>>>> +      endpoint mode. This PCIe controller may be embedded into many
>>>> +      different vendors SoCs.
>>>> +
>>>>  endif # PCI_CADENCE
>>>> diff --git a/drivers/pci/cadence/Makefile b/drivers/pci/cadence/Makefile
>>>> index d57d192d2595..61e9c8d6839d 100644
>>>> --- a/drivers/pci/cadence/Makefile
>>>> +++ b/drivers/pci/cadence/Makefile
>>>> @@ -1,2 +1,3 @@
>>>>  obj-$(CONFIG_PCIE_CADENCE) += pcie-cadence.o
>>>>  obj-$(CONFIG_PCIE_CADENCE_HOST) += pcie-cadence-host.o
>>>> +obj-$(CONFIG_PCIE_CADENCE_EP) += pcie-cadence-ep.o
>>>> diff --git a/drivers/pci/cadence/pcie-cadence-ep.c b/drivers/pci/cadence/pcie-cadence-ep.c
>>>> new file mode 100644
>>>> index 000000000000..a1d761101a9c
>>>> --- /dev/null
>>>> +++ b/drivers/pci/cadence/pcie-cadence-ep.c
>>>> @@ -0,0 +1,553 @@
>>>> +/*
>>>> + * Cadence PCIe host controller driver.
>>>
>>> You should update this comment.
>>>
>>>> + *
>>>> + * Copyright (c) 2017 Cadence
>>>> + *
>>>> + * Author: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>>>> + *
>>>> + * This program is free software; you can redistribute it and/or modify
>>>> + * it under the terms of the GNU General Public License version 2 as
>>>> + * published by the Free Software Foundation.
>>>> + *
>>>> + * This program is distributed in the hope that it will be useful,
>>>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>>>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>>>> + * GNU General Public License for more details.
>>>> + *
>>>> + * You should have received a copy of the GNU General Public License along with
>>>> + * this program.  If not, see <http://www.gnu.org/licenses/>.
>>>> + */
> 
> Would you consider using the new SPDX ids instead of this long legalese blurb?
> You can  check tglx's doc patches for details and Linus comments on
> why he wants it a certain way with C++ // comments.
> 
> Here this could come out much streamlined with this top level comment block:
> 
> // SPDX-License-Identifier: GPL-2.0
> // Copyright (c) 2017 Cadence
> // Cadence PCIe host controller driver.
> // Author: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
> 
> This is only a suggestion, but everyone kinda likes it when the
> licensing boilerplate is minimal and out of the visual way in order to
> focus is on what matters most: the code!
> Thank you for your kind consideration.
> 

I don't mind changing for your proposal if this is the preferred way.
I'm updating the series.

Best regards,

Cyrille
Cyrille Pitchen Dec. 13, 2017, 4:50 p.m. UTC | #6
Hi Kishon,

Le 05/12/2017 à 10:19, Kishon Vijay Abraham I a écrit :
> Hi,
> 
> On Friday 01 December 2017 05:50 PM, Lorenzo Pieralisi wrote:
>> On Thu, Nov 23, 2017 at 04:01:50PM +0100, Cyrille Pitchen wrote:
>>> This patch adds support to the Cadence PCIe controller in endpoint mode.
>>
>> Please add a brief description to the log to describe the most salient
>> features.
>>
>>> Signed-off-by: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>>> ---
>>>  drivers/pci/cadence/Kconfig           |   9 +
>>>  drivers/pci/cadence/Makefile          |   1 +
>>>  drivers/pci/cadence/pcie-cadence-ep.c | 553 ++++++++++++++++++++++++++++++++++
>>>  3 files changed, 563 insertions(+)
>>>  create mode 100644 drivers/pci/cadence/pcie-cadence-ep.c
>>>
>>> diff --git a/drivers/pci/cadence/Kconfig b/drivers/pci/cadence/Kconfig
>>> index 120306cae2aa..b2e6af71f39e 100644
>>> --- a/drivers/pci/cadence/Kconfig
>>> +++ b/drivers/pci/cadence/Kconfig
>>> @@ -21,4 +21,13 @@ config PCIE_CADENCE_HOST
>>>  	  mode. This PCIe controller may be embedded into many different vendors
>>>  	  SoCs.
>>>  
>>> +config PCIE_CADENCE_EP
>>> +	bool "Cadence PCIe endpoint controller"
>>> +	depends on PCI_ENDPOINT
>>> +	select PCIE_CADENCE
>>> +	help
>>> +	  Say Y here if you want to support the Cadence PCIe  controller in
>>> +	  endpoint mode. This PCIe controller may be embedded into many
>>> +	  different vendors SoCs.
>>> +
>>>  endif # PCI_CADENCE
>>> diff --git a/drivers/pci/cadence/Makefile b/drivers/pci/cadence/Makefile
>>> index d57d192d2595..61e9c8d6839d 100644
>>> --- a/drivers/pci/cadence/Makefile
>>> +++ b/drivers/pci/cadence/Makefile
>>> @@ -1,2 +1,3 @@
>>>  obj-$(CONFIG_PCIE_CADENCE) += pcie-cadence.o
>>>  obj-$(CONFIG_PCIE_CADENCE_HOST) += pcie-cadence-host.o
>>> +obj-$(CONFIG_PCIE_CADENCE_EP) += pcie-cadence-ep.o
>>> diff --git a/drivers/pci/cadence/pcie-cadence-ep.c b/drivers/pci/cadence/pcie-cadence-ep.c
>>> new file mode 100644
>>> index 000000000000..a1d761101a9c
>>> --- /dev/null
>>> +++ b/drivers/pci/cadence/pcie-cadence-ep.c
>>> @@ -0,0 +1,553 @@
>>> +/*
>>> + * Cadence PCIe host controller driver.
>>
>> You should update this comment.
>>
>>> + *
>>> + * Copyright (c) 2017 Cadence
>>> + *
>>> + * Author: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>>> + *
>>> + * This program is free software; you can redistribute it and/or modify
>>> + * it under the terms of the GNU General Public License version 2 as
>>> + * published by the Free Software Foundation.
>>> + *
>>> + * This program is distributed in the hope that it will be useful,
>>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>>> + * GNU General Public License for more details.
>>> + *
>>> + * You should have received a copy of the GNU General Public License along with
>>> + * this program.  If not, see <http://www.gnu.org/licenses/>.
>>> + */
>>> +#include <linux/kernel.h>
>>> +#include <linux/of.h>
>>> +#include <linux/pci-epc.h>
>>> +#include <linux/platform_device.h>
>>> +#include <linux/pm_runtime.h>
>>> +#include <linux/sizes.h>
>>> +#include <linux/delay.h>
>>
>> Nit: alphabetical order.
>>
>>> +#include "pcie-cadence.h"
>>> +
>>> +#define CDNS_PCIE_EP_MIN_APERTURE		128	/* 128 bytes */
>>> +#define CDNS_PCIE_EP_IRQ_PCI_ADDR_NONE		0x1
>>> +#define CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY	0x3
>>> +
>>> +/**
>>> + * struct cdns_pcie_ep_data - hardware specific data
>>> + * @max_regions: maximum nmuber of regions supported by hardware
>>
>> s/nmuber/number
>>
>>> + */
>>> +struct cdns_pcie_ep_data {
>>> +	size_t				max_regions;
>>> +};
>>> +
>>> +/**
>>> + * struct cdns_pcie_ep - private data for this PCIe endpoint controller driver
>>> + * @pcie: Cadence PCIe controller
>>> + * @data: pointer to a 'struct cdns_pcie_data'
>>> + */
>>> +struct cdns_pcie_ep {
>>> +	struct cdns_pcie		pcie;
>>> +	const struct cdns_pcie_ep_data	*data;
>>> +	struct pci_epc			*epc;
>>> +	unsigned long			ob_region_map;
>>> +	phys_addr_t			*ob_addr;
>>> +	phys_addr_t			irq_phys_addr;
>>> +	void __iomem			*irq_cpu_addr;
>>> +	u64				irq_pci_addr;
>>> +	u8				irq_pending;
>>> +};
>>> +
>>> +static int cdns_pcie_ep_write_header(struct pci_epc *epc,
>>> +				     struct pci_epf_header *hdr)
>>> +{
>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	u8 fn = 0;
>>> +
>>> +	if (fn == 0) {
>>
>> I think there is some code to retrieve fn missing here.
> 
> hmm.. the endpoint core has to send the function number which right now it's
> not doing though it has the function number info in pci_epf.

Would it be OK if I add a new patch in the next series adding a
'struct pcie_epf *epf' as a 2nd argument to all handlers in the
'struct pcie_epc_ops'? This way I could have access to epf->func_no as needed.

Best regards,

Cyrille

>>
>>> +		u32 id = CDNS_PCIE_LM_ID_VENDOR(hdr->vendorid) |
>>> +			 CDNS_PCIE_LM_ID_SUBSYS(hdr->subsys_vendor_id);
>>> +		cdns_pcie_writel(pcie, CDNS_PCIE_LM_ID, id);
>>> +	}
>>> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_DEVICE_ID, hdr->deviceid);
>>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_REVISION_ID, hdr->revid);
>>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_CLASS_PROG, hdr->progif_code);
>>> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_CLASS_DEVICE,
>>> +			       hdr->subclass_code | hdr->baseclass_code << 8);
>>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_CACHE_LINE_SIZE,
>>> +			       hdr->cache_line_size);
>>> +	cdns_pcie_ep_fn_writew(pcie, fn, PCI_SUBSYSTEM_ID, hdr->subsys_id);
>>> +	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_INTERRUPT_PIN, hdr->interrupt_pin);
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static int cdns_pcie_ep_set_bar(struct pci_epc *epc, enum pci_barno bar,
>>> +				dma_addr_t bar_phys, size_t size, int flags)
>>> +{
>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	u32 addr0, addr1, reg, cfg, b, aperture, ctrl;
>>> +	u8 fn = 0;
> 
> Here too endpoint core should send the function number..
>>> +	u64 sz;
>>> +
>>> +	/* BAR size is 2^(aperture + 7) */
>>> +	sz = max_t(size_t, size, CDNS_PCIE_EP_MIN_APERTURE);
>>> +	sz = 1ULL << fls64(sz - 1);
>>> +	aperture = ilog2(sz) - 7; /* 128B -> 0, 256B -> 1, 512B -> 2, ... */
>>> +
>>> +	if ((flags & PCI_BASE_ADDRESS_SPACE) == PCI_BASE_ADDRESS_SPACE_IO) {
>>> +		ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_IO_32BITS;
>>> +	} else {
>>> +		bool is_prefetch = !!(flags & PCI_BASE_ADDRESS_MEM_PREFETCH);
>>> +		bool is_64bits = sz > SZ_2G;
>>> +
>>> +		if (is_64bits && (bar & 1))
>>> +			return -EINVAL;
>>> +
>>> +		switch (is_64bits << 1 | is_prefetch) {
>>
>> I would not mind implementing this as a nested if-else, I am not a big
>> fan of using bool this way.
>>
>>> +		case 0:
>>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_MEM_32BITS;
>>> +			break;
>>> +
>>> +		case 1:
>>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_PREFETCH_MEM_32BITS;
>>> +			break;
>>> +
>>> +		case 2:
>>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_MEM_64BITS;
>>> +			break;
>>> +
>>> +		case 3:
>>> +			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_PREFETCH_MEM_64BITS;
>>> +			break;
>>> +		}
>>> +	}
>>> +
>>> +	addr0 = lower_32_bits(bar_phys);
>>> +	addr1 = upper_32_bits(bar_phys);
>>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar),
>>> +			 addr0);
> 
> It would be nice if you can have defines for CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0
> included in this patch rather than "PCI: cadence: Add host driver for Cadence
> PCIe controller". All EP specific functions in header file should be included here.
>>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar),
>>> +			 addr1);
>>
>> Is fn always 0 ?
>>
>>> +	if (bar < BAR_4) {
>>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG0(fn);
>>> +		b = bar;
>>> +	} else {
>>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG1(fn);
>>> +		b = bar - BAR_4;
>>> +	}
>>> +
>>> +	cfg = cdns_pcie_readl(pcie, reg);
>>> +	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
>>> +		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
>>> +	cfg |= (CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE(b, aperture) |
>>> +		CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL(b, ctrl));
>>> +	cdns_pcie_writel(pcie, reg, cfg);
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, enum pci_barno bar)
>>> +{
>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	u32 reg, cfg, b, ctrl;
>>> +	u8 fn = 0;
> 
> Here too endpoint core should send the function number..
>>> +
>>> +	if (bar < BAR_4) {
>>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG0(fn);
>>> +		b = bar;
>>> +	} else {
>>> +		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG1(fn);
>>> +		b = bar - BAR_4;
>>> +	}
>>> +
>>> +	ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_DISABLED;
>>> +	cfg = cdns_pcie_readl(pcie, reg);
>>> +	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
>>> +		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
>>> +	cfg |= CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(ctrl);
>>> +	cdns_pcie_writel(pcie, reg, cfg);
>>> +
>>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar), 0);
>>> +	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar), 0);
>>> +}
>>> +
>>> +static int cdns_pcie_ep_map_addr(struct pci_epc *epc, phys_addr_t addr,
>>> +				 u64 pci_addr, size_t size)
>>> +{
>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	u32 r;
>>> +
>>> +	r = find_first_zero_bit(&ep->ob_region_map, sizeof(ep->ob_region_map));
>>
>> Second argument must be in bits not bytes.
>>
>> https://marc.info/?l=linux-pci&m=151179781225513&w=2
>>
>>> +	if (r >= ep->data->max_regions - 1) {
>>> +		dev_err(&epc->dev, "no free outbound region\n");
>>> +		return -EINVAL;
>>> +	}
>>> +
>>> +	cdns_pcie_set_outbound_region(pcie, r, false, addr, pci_addr, size);
>>> +
>>> +	set_bit(r, &ep->ob_region_map);
>>> +	ep->ob_addr[r] = addr;
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static void cdns_pcie_ep_unmap_addr(struct pci_epc *epc, phys_addr_t addr)
>>> +{
>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	u32 r;
>>> +
>>> +	for (r = 0; r < ep->data->max_regions - 1; r++)
>>> +		if (ep->ob_addr[r] == addr)
>>> +			break;
>>> +
>>> +	if (r >= ep->data->max_regions - 1)
>>
>> == ?
>>
>>> +		return;
>>> +
>>> +	cdns_pcie_reset_outbound_region(pcie, r);
>>> +
>>> +	ep->ob_addr[r] = 0;
>>> +	clear_bit(r, &ep->ob_region_map);
>>> +}
>>> +
>>> +static int cdns_pcie_ep_set_msi(struct pci_epc *epc, u8 mmc)
>>> +{
>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
>>> +	u16 flags;
>>> +	u8 fn = 0;
>>> +
>>> +	/* Validate the ID of the MSI Capability structure. */
>>> +	if (cdns_pcie_ep_fn_readb(pcie, fn, cap) != PCI_CAP_ID_MSI)
>>> +		return -EINVAL;
>>> +
>>> +	/*
>>> +	 * Set the Multiple Message Capable bitfield into the Message Control
>>> +	 * register.
>>> +	 */
>>> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
>>> +	flags = (flags & ~PCI_MSI_FLAGS_QMASK) | (mmc << 1);
>>> +	flags |= PCI_MSI_FLAGS_64BIT;
>>> +	flags &= ~PCI_MSI_FLAGS_MASKBIT;
> 
> Any reason why "Per-vector masking capable" is reset?
>>> +	cdns_pcie_ep_fn_writew(pcie, fn, cap + PCI_MSI_FLAGS, flags);
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static int cdns_pcie_ep_get_msi(struct pci_epc *epc)
>>> +{
>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
>>> +	u16 flags, mmc, mme;
>>> +	u8 fn = 0;
>>> +
>>> +	/* Validate the ID of the MSI Capability structure. */
>>> +	if (cdns_pcie_ep_fn_readb(pcie, fn, cap) != PCI_CAP_ID_MSI)
>>> +		return -EINVAL;
>>> +
>>> +	/* Validate that the MSI feature is actually enabled. */
>>> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
>>> +	if (!(flags & PCI_MSI_FLAGS_ENABLE))
>>> +		return -EINVAL;
>>> +
>>> +	/*
>>> +	 * Get the Multiple Message Enable bitfield from the Message Control
>>> +	 * register.
>>> +	 */
>>> +	mmc = (flags & PCI_MSI_FLAGS_QMASK) >> 1;
>>> +	mme = (flags & PCI_MSI_FLAGS_QSIZE) >> 4;
>>> +	if (mme > mmc)
>>> +		mme = mmc;
>>> +	if (mme > 5)
>>> +		mme = 5;
> 
> I'm not sure if both these above checks are required..
>>
>> You should comment on what this 5 means and why it is fine to cap mme.
>>
>>> +
>>> +	return mme;
>>> +}
>>> +
>>> +static void cdns_pcie_ep_assert_intx(struct cdns_pcie_ep *ep, u8 fn,
>>> +				     u8 intx, bool is_asserted)
>>> +{
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	u32 r = ep->data->max_regions - 1;
>>> +	u32 offset;
>>> +	u16 status;
>>> +	u8 msg_code;
>>> +
>>> +	intx &= 3;
>>> +
>>> +	/* Set the outbound region if needed. */
>>> +	if (unlikely(ep->irq_pci_addr != CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY)) {
>>> +		/* Last region was reserved for IRQ writes. */
>>> +		cdns_pcie_set_outbound_region_for_normal_msg(pcie, r,
>>> +							     ep->irq_phys_addr);
>>> +		ep->irq_pci_addr = CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY;
>>> +	}
>>> +
>>> +	if (is_asserted) {
>>> +		ep->irq_pending |= BIT(intx);
>>> +		msg_code = MSG_CODE_ASSERT_INTA + intx;
>>> +	} else {
>>> +		ep->irq_pending &= ~BIT(intx);
>>> +		msg_code = MSG_CODE_DEASSERT_INTA + intx;
>>> +	}
>>> +
>>> +	status = cdns_pcie_ep_fn_readw(pcie, fn, PCI_STATUS);
>>> +	if (((status & PCI_STATUS_INTERRUPT) != 0) ^ (ep->irq_pending != 0)) {
>>> +		status ^= PCI_STATUS_INTERRUPT;
>>> +		cdns_pcie_ep_fn_writew(pcie, fn, PCI_STATUS, status);
>>> +	}
> 
> here you are setting the PCI_STATUS_INTERRUPT even before sending the ASSERT
> message.
>>> +
>>> +	offset = CDNS_PCIE_NORMAL_MSG_ROUTING(MSG_ROUTING_LOCAL) |
>>> +		 CDNS_PCIE_NORMAL_MSG_CODE(msg_code) |
>>> +		 CDNS_PCIE_MSG_NO_DATA;
>>> +	writel(0, ep->irq_cpu_addr + offset);
>>> +}
>>> +
>>> +static int cdns_pcie_ep_send_legacy_irq(struct cdns_pcie_ep *ep, u8 fn, u8 intx)
>>> +{
>>> +	u16 cmd;
>>> +
>>> +	cmd = cdns_pcie_ep_fn_readw(&ep->pcie, fn, PCI_COMMAND);
>>> +	if (cmd & PCI_COMMAND_INTX_DISABLE)
>>> +		return -EINVAL;
>>> +
>>> +	cdns_pcie_ep_assert_intx(ep, fn, intx, true);
>>> +	mdelay(1);
>>
>> Add a comment please to explain the mdelay value.
>>
>>> +	cdns_pcie_ep_assert_intx(ep, fn, intx, false);
>>> +	return 0;
>>> +}
>>> +
>>> +static int cdns_pcie_ep_raise_irq(struct pci_epc *epc,
>>> +				  enum pci_epc_irq_type type, u8 interrupt_num)
>>> +{
>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
>>> +	u16 flags, mmc, mme, data, data_mask;
>>> +	u8 msi_count;
>>> +	u64 pci_addr, pci_addr_mask = 0xff;
>>> +	u8 fn = 0;
>>> +
>>> +	/* Handle legacy IRQ. */
>>> +	if (type == PCI_EPC_IRQ_LEGACY)
>>> +		return cdns_pcie_ep_send_legacy_irq(ep, fn, 0);
>>> +
>>> +	/* Otherwise MSI. */
>>> +	if (type != PCI_EPC_IRQ_MSI)
>>> +		return -EINVAL;
> 
> MSI-X?
>>> +
>>> +	/* Check whether the MSI feature has been enabled by the PCI host. */
>>> +	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
>>> +	if (!(flags & PCI_MSI_FLAGS_ENABLE))
>>> +		return -EINVAL;
>>> +
>>> +	/* Get the number of enabled MSIs */
>>> +	mmc = (flags & PCI_MSI_FLAGS_QMASK) >> 1;
>>> +	mme = (flags & PCI_MSI_FLAGS_QSIZE) >> 4;
>>> +	if (mme > mmc)
>>> +		mme = mmc;
>>> +	if (mme > 5)
>>> +		mme = 5;
>>
>> Same comment as above.
>>
>>> +	msi_count = 1 << mme;
>>> +	if (!interrupt_num || interrupt_num > msi_count)
>>> +		return -EINVAL;
>>> +
>>> +	/* Compute the data value to be written. */
>>> +	data_mask = msi_count - 1;
>>> +	data = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_DATA_64);
>>> +	data = (data & ~data_mask) | ((interrupt_num - 1) & data_mask);
>>> +
>>> +	/* Get the PCI address where to write the data into. */
>>> +	pci_addr = cdns_pcie_ep_fn_readl(pcie, fn, cap + PCI_MSI_ADDRESS_HI);
>>> +	pci_addr <<= 32;
>>> +	pci_addr |= cdns_pcie_ep_fn_readl(pcie, fn, cap + PCI_MSI_ADDRESS_LO);
>>> +	pci_addr &= GENMASK_ULL(63, 2);
>>> +
>>> +	/* Set the outbound region if needed. */
>>> +	if (unlikely(ep->irq_pci_addr != pci_addr)) {
>>> +		/* Last region was reserved for IRQ writes. */
>>> +		cdns_pcie_set_outbound_region(pcie, ep->data->max_regions - 1,
>>> +					      false,
>>> +					      ep->irq_phys_addr,
>>> +					      pci_addr & ~pci_addr_mask,
>>> +					      pci_addr_mask + 1);
>>> +		ep->irq_pci_addr = pci_addr;
>>> +	}
>>> +	writew(data, ep->irq_cpu_addr + (pci_addr & pci_addr_mask));
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static int cdns_pcie_ep_start(struct pci_epc *epc)
>>> +{
>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	struct pci_epf *epf;
>>> +	u32 cfg;
>>> +	u8 fn = 0;
>>> +
>>> +	/* Enable this endpoint function. */
>>> +	cfg = cdns_pcie_readl(pcie, CDNS_PCIE_LM_EP_FUNC_CFG);
>>> +	cfg |= BIT(fn);
>>> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, cfg);
>>> +
>>> +	/*
>>> +	 * Already linked-up: don't call directly pci_epc_linkup() because we've
>>> +	 * already locked the epc->lock.
>>> +	 */
> 
> Not sure what you mean by linked-up here?
>>> +	list_for_each_entry(epf, &epc->pci_epf, list)
>>> +		pci_epf_linkup(epf);
> 
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static void cdns_pcie_ep_stop(struct pci_epc *epc)
>>> +{
>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>> +	u32 cfg;
>>> +	u8 fn = 0;
>>> +
>>> +	/* Disable this endpoint function (function 0 can't be disabled). */
>>
>> I do not understand this comment and how it applies to the code,
>> in other words fn is always 0 here (so it can't be disabled)
>> I do not understand what this code is there for.
>>
>>> +	cfg = cdns_pcie_readl(pcie, CDNS_PCIE_LM_EP_FUNC_CFG);
>>> +	cfg &= ~BIT(fn);
>>> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, cfg);
>>> +}
>>> +
>>> +static const struct pci_epc_ops cdns_pcie_epc_ops = {
>>> +	.write_header	= cdns_pcie_ep_write_header,
>>> +	.set_bar	= cdns_pcie_ep_set_bar,
>>> +	.clear_bar	= cdns_pcie_ep_clear_bar,
>>> +	.map_addr	= cdns_pcie_ep_map_addr,
>>> +	.unmap_addr	= cdns_pcie_ep_unmap_addr,
>>> +	.set_msi	= cdns_pcie_ep_set_msi,
>>> +	.get_msi	= cdns_pcie_ep_get_msi,
>>> +	.raise_irq	= cdns_pcie_ep_raise_irq,
>>> +	.start		= cdns_pcie_ep_start,
>>> +	.stop		= cdns_pcie_ep_stop,
>>> +};
>>> +
>>> +static const struct cdns_pcie_ep_data cdns_pcie_ep_data = {
>>> +	.max_regions	= 16,
>>> +};
>>
>> As I mentioned in patch 3, should this be set-up with DT ?
>>
>> Thanks,
>> Lorenzo
>>
>>> +
>>> +static const struct of_device_id cdns_pcie_ep_of_match[] = {
>>> +	{ .compatible = "cdns,cdns-pcie-ep",
>>> +	  .data = &cdns_pcie_ep_data },
>>> +
>>> +	{ },
>>> +};
>>> +
>>> +static int cdns_pcie_ep_probe(struct platform_device *pdev)
>>> +{
>>> +	struct device *dev = &pdev->dev;
>>> +	struct device_node *np = dev->of_node;
>>> +	const struct of_device_id *of_id;
>>> +	struct cdns_pcie_ep *ep;
>>> +	struct cdns_pcie *pcie;
>>> +	struct pci_epc *epc;
>>> +	struct resource *res;
>>> +	size_t max_regions;
>>> +	int ret;
>>> +
>>> +	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>>> +	if (!ep)
>>> +		return -ENOMEM;
>>> +
>>> +	platform_set_drvdata(pdev, ep);
>>> +
>>> +	pcie = &ep->pcie;
>>> +	pcie->is_rc = false;
>>> +
>>> +	of_id = of_match_node(cdns_pcie_ep_of_match, np);
>>> +	ep->data = (const struct cdns_pcie_ep_data *)of_id->data;
>>> +
>>> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "reg");
>>> +	pcie->reg_base = devm_ioremap_resource(dev, res);
>>> +	if (IS_ERR(pcie->reg_base)) {
>>> +		dev_err(dev, "missing \"reg\"\n");
>>> +		return PTR_ERR(pcie->reg_base);
>>> +	}
>>> +
>>> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "mem");
>>> +	if (!res) {
>>> +		dev_err(dev, "missing \"mem\"\n");
>>> +		return -EINVAL;
>>> +	}
>>> +	pcie->mem_res = res;
>>> +
>>> +	max_regions = ep->data->max_regions;
>>> +	ep->ob_addr = devm_kzalloc(dev, max_regions * sizeof(*ep->ob_addr),
>>> +				   GFP_KERNEL);
>>> +	if (!ep->ob_addr)
>>> +		return -ENOMEM;
>>> +
>>> +	pm_runtime_enable(dev);
>>> +	ret = pm_runtime_get_sync(dev);
>>> +	if (ret < 0) {
>>> +		dev_err(dev, "pm_runtime_get_sync() failed\n");
>>> +		goto err_get_sync;
>>> +	}
>>> +
>>> +	/* Disable all but function 0 (anyway BIT(0) is hardwired to 1). */
>>> +	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, BIT(0));
> 
> why disable all functions?
>>> +
>>> +	epc = devm_pci_epc_create(dev, &cdns_pcie_epc_ops);
>>> +	if (IS_ERR(epc)) {
>>> +		dev_err(dev, "failed to create epc device\n");
>>> +		ret = PTR_ERR(epc);
>>> +		goto err_init;
>>> +	}
>>> +
>>> +	ep->epc = epc;
>>> +	epc_set_drvdata(epc, ep);
>>> +
>>> +	ret = of_property_read_u8(np, "max-functions", &epc->max_functions);
>>> +	if (ret < 0)
>>> +		epc->max_functions = 1;
>>> +
>>> +	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
>>> +			       resource_size(pcie->mem_res));
>>> +	if (ret < 0) {
>>> +		dev_err(dev, "failed to initialize the memory space\n");
>>> +		goto err_init;
>>> +	}
>>> +
>>> +	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
>>> +						  SZ_128K);
> 
> Any reason why you chose SZ_128K?
> 
> Thanks
> Kishon
>
Cyrille Pitchen Dec. 14, 2017, 5:03 p.m. UTC | #7
Le 13/12/2017 à 17:50, Cyrille Pitchen a écrit :
> Hi Kishon,
> 
> Le 05/12/2017 à 10:19, Kishon Vijay Abraham I a écrit :
>> Hi,
>>
>> On Friday 01 December 2017 05:50 PM, Lorenzo Pieralisi wrote:
>>> On Thu, Nov 23, 2017 at 04:01:50PM +0100, Cyrille Pitchen wrote:
>>>> This patch adds support to the Cadence PCIe controller in endpoint mode.
>>>
>>> Please add a brief description to the log to describe the most salient
>>> features.
>>>
>>>> Signed-off-by: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>>>> ---
>>>>  drivers/pci/cadence/Kconfig           |   9 +
>>>>  drivers/pci/cadence/Makefile          |   1 +
>>>>  drivers/pci/cadence/pcie-cadence-ep.c | 553 ++++++++++++++++++++++++++++++++++
>>>>  3 files changed, 563 insertions(+)
>>>>  create mode 100644 drivers/pci/cadence/pcie-cadence-ep.c
[...]
>>>> +static int cdns_pcie_ep_write_header(struct pci_epc *epc,
>>>> +				     struct pci_epf_header *hdr)
>>>> +{
>>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>>> +	u8 fn = 0;
>>>> +
>>>> +	if (fn == 0) {
>>>
>>> I think there is some code to retrieve fn missing here.
>>
>> hmm.. the endpoint core has to send the function number which right now it's
>> not doing though it has the function number info in pci_epf.
> 
> Would it be OK if I add a new patch in the next series adding a
> 'struct pcie_epf *epf' as a 2nd argument to all handlers in the
> 'struct pcie_epc_ops'? This way I could have access to epf->func_no as needed.
>

Except for pci_epc_start() and pci_epc_stop(), both only called from
pci_epc_start_store(), I don't have trouble getting the epf value to be passed
as a 2nd argument to all other handlers in 'struct pcie_epc_ops'.

Now my next question is: is it better to keep the 'struct pci_epc *epc' as
the 1st argument of all those handlers or do you prefer me to remove it as
the value can always be retrieved from epf->epc, since now we provide epf as
a new argument ?

I have no personal preference. Please let me know your choice :)
 
> Best regards,
> 
> Cyrille
>
Kishon Vijay Abraham I Dec. 15, 2017, 5:49 a.m. UTC | #8
Hi Cyrille,

On Thursday 14 December 2017 10:33 PM, Cyrille Pitchen wrote:
> Le 13/12/2017 à 17:50, Cyrille Pitchen a écrit :
>> Hi Kishon,
>>
>> Le 05/12/2017 à 10:19, Kishon Vijay Abraham I a écrit :
>>> Hi,
>>>
>>> On Friday 01 December 2017 05:50 PM, Lorenzo Pieralisi wrote:
>>>> On Thu, Nov 23, 2017 at 04:01:50PM +0100, Cyrille Pitchen wrote:
>>>>> This patch adds support to the Cadence PCIe controller in endpoint mode.
>>>>
>>>> Please add a brief description to the log to describe the most salient
>>>> features.
>>>>
>>>>> Signed-off-by: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>>>>> ---
>>>>>  drivers/pci/cadence/Kconfig           |   9 +
>>>>>  drivers/pci/cadence/Makefile          |   1 +
>>>>>  drivers/pci/cadence/pcie-cadence-ep.c | 553 ++++++++++++++++++++++++++++++++++
>>>>>  3 files changed, 563 insertions(+)
>>>>>  create mode 100644 drivers/pci/cadence/pcie-cadence-ep.c
> [...]
>>>>> +static int cdns_pcie_ep_write_header(struct pci_epc *epc,
>>>>> +				     struct pci_epf_header *hdr)
>>>>> +{
>>>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>>>> +	u8 fn = 0;
>>>>> +
>>>>> +	if (fn == 0) {
>>>>
>>>> I think there is some code to retrieve fn missing here.
>>>
>>> hmm.. the endpoint core has to send the function number which right now it's
>>> not doing though it has the function number info in pci_epf.
>>
>> Would it be OK if I add a new patch in the next series adding a
>> 'struct pcie_epf *epf' as a 2nd argument to all handlers in the
>> 'struct pcie_epc_ops'? This way I could have access to epf->func_no as needed.

I prefer we just pass the func_no as the second argument. Do you see a problem
with that?
>>
> 
> Except for pci_epc_start() and pci_epc_stop(), both only called from
> pci_epc_start_store(), I don't have trouble getting the epf value to be passed
> as a 2nd argument to all other handlers in 'struct pcie_epc_ops'.

pci_epc_start()/pci_epc_stop() is used to start/stop the end point controller
as a whole and shouldn't need epf.
> 
> Now my next question is: is it better to keep the 'struct pci_epc *epc' as
> the 1st argument of all those handlers or do you prefer me to remove it as
> the value can always be retrieved from epf->epc, since now we provide epf as
> a new argument ?

Do we really need to pass epf when func_no is all we need?

Thanks
Kishon
Cyrille Pitchen Dec. 15, 2017, 11:49 a.m. UTC | #9
Hi Kishon,

Le 15/12/2017 à 06:49, Kishon Vijay Abraham I a écrit :
> Hi Cyrille,
> 
> On Thursday 14 December 2017 10:33 PM, Cyrille Pitchen wrote:
>> Le 13/12/2017 à 17:50, Cyrille Pitchen a écrit :
>>> Hi Kishon,
>>>
>>> Le 05/12/2017 à 10:19, Kishon Vijay Abraham I a écrit :
>>>> Hi,
>>>>
>>>> On Friday 01 December 2017 05:50 PM, Lorenzo Pieralisi wrote:
>>>>> On Thu, Nov 23, 2017 at 04:01:50PM +0100, Cyrille Pitchen wrote:
>>>>>> This patch adds support to the Cadence PCIe controller in endpoint mode.
>>>>>
>>>>> Please add a brief description to the log to describe the most salient
>>>>> features.
>>>>>
>>>>>> Signed-off-by: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
>>>>>> ---
>>>>>>  drivers/pci/cadence/Kconfig           |   9 +
>>>>>>  drivers/pci/cadence/Makefile          |   1 +
>>>>>>  drivers/pci/cadence/pcie-cadence-ep.c | 553 ++++++++++++++++++++++++++++++++++
>>>>>>  3 files changed, 563 insertions(+)
>>>>>>  create mode 100644 drivers/pci/cadence/pcie-cadence-ep.c
>> [...]
>>>>>> +static int cdns_pcie_ep_write_header(struct pci_epc *epc,
>>>>>> +				     struct pci_epf_header *hdr)
>>>>>> +{
>>>>>> +	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>>>>> +	struct cdns_pcie *pcie = &ep->pcie;
>>>>>> +	u8 fn = 0;
>>>>>> +
>>>>>> +	if (fn == 0) {
>>>>>
>>>>> I think there is some code to retrieve fn missing here.
>>>>
>>>> hmm.. the endpoint core has to send the function number which right now it's
>>>> not doing though it has the function number info in pci_epf.
>>>
>>> Would it be OK if I add a new patch in the next series adding a
>>> 'struct pcie_epf *epf' as a 2nd argument to all handlers in the
>>> 'struct pcie_epc_ops'? This way I could have access to epf->func_no as needed.
> 
> I prefer we just pass the func_no as the second argument. Do you see a problem
> with that?
>>>
>>
>> Except for pci_epc_start() and pci_epc_stop(), both only called from
>> pci_epc_start_store(), I don't have trouble getting the epf value to be passed
>> as a 2nd argument to all other handlers in 'struct pcie_epc_ops'.
> 
> pci_epc_start()/pci_epc_stop() is used to start/stop the end point controller
> as a whole and shouldn't need epf.
>>
>> Now my next question is: is it better to keep the 'struct pci_epc *epc' as
>> the 1st argument of all those handlers or do you prefer me to remove it as
>> the value can always be retrieved from epf->epc, since now we provide epf as
>> a new argument ?
> 
> Do we really need to pass epf when func_no is all we need?
>

No, func_no alone is currently all I need so it's perfectly fine for me.
I've just wanted to ask before to be sure I implement the expected solution
before submitting :)

Thanks!

Cyrille

 
> Thanks
> Kishon
>
diff mbox

Patch

diff --git a/drivers/pci/cadence/Kconfig b/drivers/pci/cadence/Kconfig
index 120306cae2aa..b2e6af71f39e 100644
--- a/drivers/pci/cadence/Kconfig
+++ b/drivers/pci/cadence/Kconfig
@@ -21,4 +21,13 @@  config PCIE_CADENCE_HOST
 	  mode. This PCIe controller may be embedded into many different vendors
 	  SoCs.
 
+config PCIE_CADENCE_EP
+	bool "Cadence PCIe endpoint controller"
+	depends on PCI_ENDPOINT
+	select PCIE_CADENCE
+	help
+	  Say Y here if you want to support the Cadence PCIe  controller in
+	  endpoint mode. This PCIe controller may be embedded into many
+	  different vendors SoCs.
+
 endif # PCI_CADENCE
diff --git a/drivers/pci/cadence/Makefile b/drivers/pci/cadence/Makefile
index d57d192d2595..61e9c8d6839d 100644
--- a/drivers/pci/cadence/Makefile
+++ b/drivers/pci/cadence/Makefile
@@ -1,2 +1,3 @@ 
 obj-$(CONFIG_PCIE_CADENCE) += pcie-cadence.o
 obj-$(CONFIG_PCIE_CADENCE_HOST) += pcie-cadence-host.o
+obj-$(CONFIG_PCIE_CADENCE_EP) += pcie-cadence-ep.o
diff --git a/drivers/pci/cadence/pcie-cadence-ep.c b/drivers/pci/cadence/pcie-cadence-ep.c
new file mode 100644
index 000000000000..a1d761101a9c
--- /dev/null
+++ b/drivers/pci/cadence/pcie-cadence-ep.c
@@ -0,0 +1,553 @@ 
+/*
+ * Cadence PCIe host controller driver.
+ *
+ * Copyright (c) 2017 Cadence
+ *
+ * Author: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+#include <linux/kernel.h>
+#include <linux/of.h>
+#include <linux/pci-epc.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/sizes.h>
+#include <linux/delay.h>
+
+#include "pcie-cadence.h"
+
+#define CDNS_PCIE_EP_MIN_APERTURE		128	/* 128 bytes */
+#define CDNS_PCIE_EP_IRQ_PCI_ADDR_NONE		0x1
+#define CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY	0x3
+
+/**
+ * struct cdns_pcie_ep_data - hardware specific data
+ * @max_regions: maximum nmuber of regions supported by hardware
+ */
+struct cdns_pcie_ep_data {
+	size_t				max_regions;
+};
+
+/**
+ * struct cdns_pcie_ep - private data for this PCIe endpoint controller driver
+ * @pcie: Cadence PCIe controller
+ * @data: pointer to a 'struct cdns_pcie_data'
+ */
+struct cdns_pcie_ep {
+	struct cdns_pcie		pcie;
+	const struct cdns_pcie_ep_data	*data;
+	struct pci_epc			*epc;
+	unsigned long			ob_region_map;
+	phys_addr_t			*ob_addr;
+	phys_addr_t			irq_phys_addr;
+	void __iomem			*irq_cpu_addr;
+	u64				irq_pci_addr;
+	u8				irq_pending;
+};
+
+static int cdns_pcie_ep_write_header(struct pci_epc *epc,
+				     struct pci_epf_header *hdr)
+{
+	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
+	struct cdns_pcie *pcie = &ep->pcie;
+	u8 fn = 0;
+
+	if (fn == 0) {
+		u32 id = CDNS_PCIE_LM_ID_VENDOR(hdr->vendorid) |
+			 CDNS_PCIE_LM_ID_SUBSYS(hdr->subsys_vendor_id);
+		cdns_pcie_writel(pcie, CDNS_PCIE_LM_ID, id);
+	}
+	cdns_pcie_ep_fn_writew(pcie, fn, PCI_DEVICE_ID, hdr->deviceid);
+	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_REVISION_ID, hdr->revid);
+	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_CLASS_PROG, hdr->progif_code);
+	cdns_pcie_ep_fn_writew(pcie, fn, PCI_CLASS_DEVICE,
+			       hdr->subclass_code | hdr->baseclass_code << 8);
+	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_CACHE_LINE_SIZE,
+			       hdr->cache_line_size);
+	cdns_pcie_ep_fn_writew(pcie, fn, PCI_SUBSYSTEM_ID, hdr->subsys_id);
+	cdns_pcie_ep_fn_writeb(pcie, fn, PCI_INTERRUPT_PIN, hdr->interrupt_pin);
+
+	return 0;
+}
+
+static int cdns_pcie_ep_set_bar(struct pci_epc *epc, enum pci_barno bar,
+				dma_addr_t bar_phys, size_t size, int flags)
+{
+	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
+	struct cdns_pcie *pcie = &ep->pcie;
+	u32 addr0, addr1, reg, cfg, b, aperture, ctrl;
+	u8 fn = 0;
+	u64 sz;
+
+	/* BAR size is 2^(aperture + 7) */
+	sz = max_t(size_t, size, CDNS_PCIE_EP_MIN_APERTURE);
+	sz = 1ULL << fls64(sz - 1);
+	aperture = ilog2(sz) - 7; /* 128B -> 0, 256B -> 1, 512B -> 2, ... */
+
+	if ((flags & PCI_BASE_ADDRESS_SPACE) == PCI_BASE_ADDRESS_SPACE_IO) {
+		ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_IO_32BITS;
+	} else {
+		bool is_prefetch = !!(flags & PCI_BASE_ADDRESS_MEM_PREFETCH);
+		bool is_64bits = sz > SZ_2G;
+
+		if (is_64bits && (bar & 1))
+			return -EINVAL;
+
+		switch (is_64bits << 1 | is_prefetch) {
+		case 0:
+			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_MEM_32BITS;
+			break;
+
+		case 1:
+			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_PREFETCH_MEM_32BITS;
+			break;
+
+		case 2:
+			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_MEM_64BITS;
+			break;
+
+		case 3:
+			ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_PREFETCH_MEM_64BITS;
+			break;
+		}
+	}
+
+	addr0 = lower_32_bits(bar_phys);
+	addr1 = upper_32_bits(bar_phys);
+	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar),
+			 addr0);
+	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar),
+			 addr1);
+
+	if (bar < BAR_4) {
+		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG0(fn);
+		b = bar;
+	} else {
+		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG1(fn);
+		b = bar - BAR_4;
+	}
+
+	cfg = cdns_pcie_readl(pcie, reg);
+	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
+		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
+	cfg |= (CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE(b, aperture) |
+		CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL(b, ctrl));
+	cdns_pcie_writel(pcie, reg, cfg);
+
+	return 0;
+}
+
+static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, enum pci_barno bar)
+{
+	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
+	struct cdns_pcie *pcie = &ep->pcie;
+	u32 reg, cfg, b, ctrl;
+	u8 fn = 0;
+
+	if (bar < BAR_4) {
+		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG0(fn);
+		b = bar;
+	} else {
+		reg = CDNS_PCIE_LM_EP_FUNC_BAR_CFG1(fn);
+		b = bar - BAR_4;
+	}
+
+	ctrl = CDNS_PCIE_LM_BAR_CFG_CTRL_DISABLED;
+	cfg = cdns_pcie_readl(pcie, reg);
+	cfg &= ~(CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_APERTURE_MASK(b) |
+		 CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(b));
+	cfg |= CDNS_PCIE_LM_EP_FUNC_BAR_CFG_BAR_CTRL_MASK(ctrl);
+	cdns_pcie_writel(pcie, reg, cfg);
+
+	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR0(fn, bar), 0);
+	cdns_pcie_writel(pcie, CDNS_PCIE_AT_IB_EP_FUNC_BAR_ADDR1(fn, bar), 0);
+}
+
+static int cdns_pcie_ep_map_addr(struct pci_epc *epc, phys_addr_t addr,
+				 u64 pci_addr, size_t size)
+{
+	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
+	struct cdns_pcie *pcie = &ep->pcie;
+	u32 r;
+
+	r = find_first_zero_bit(&ep->ob_region_map, sizeof(ep->ob_region_map));
+	if (r >= ep->data->max_regions - 1) {
+		dev_err(&epc->dev, "no free outbound region\n");
+		return -EINVAL;
+	}
+
+	cdns_pcie_set_outbound_region(pcie, r, false, addr, pci_addr, size);
+
+	set_bit(r, &ep->ob_region_map);
+	ep->ob_addr[r] = addr;
+
+	return 0;
+}
+
+static void cdns_pcie_ep_unmap_addr(struct pci_epc *epc, phys_addr_t addr)
+{
+	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
+	struct cdns_pcie *pcie = &ep->pcie;
+	u32 r;
+
+	for (r = 0; r < ep->data->max_regions - 1; r++)
+		if (ep->ob_addr[r] == addr)
+			break;
+
+	if (r >= ep->data->max_regions - 1)
+		return;
+
+	cdns_pcie_reset_outbound_region(pcie, r);
+
+	ep->ob_addr[r] = 0;
+	clear_bit(r, &ep->ob_region_map);
+}
+
+static int cdns_pcie_ep_set_msi(struct pci_epc *epc, u8 mmc)
+{
+	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
+	struct cdns_pcie *pcie = &ep->pcie;
+	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
+	u16 flags;
+	u8 fn = 0;
+
+	/* Validate the ID of the MSI Capability structure. */
+	if (cdns_pcie_ep_fn_readb(pcie, fn, cap) != PCI_CAP_ID_MSI)
+		return -EINVAL;
+
+	/*
+	 * Set the Multiple Message Capable bitfield into the Message Control
+	 * register.
+	 */
+	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
+	flags = (flags & ~PCI_MSI_FLAGS_QMASK) | (mmc << 1);
+	flags |= PCI_MSI_FLAGS_64BIT;
+	flags &= ~PCI_MSI_FLAGS_MASKBIT;
+	cdns_pcie_ep_fn_writew(pcie, fn, cap + PCI_MSI_FLAGS, flags);
+
+	return 0;
+}
+
+static int cdns_pcie_ep_get_msi(struct pci_epc *epc)
+{
+	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
+	struct cdns_pcie *pcie = &ep->pcie;
+	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
+	u16 flags, mmc, mme;
+	u8 fn = 0;
+
+	/* Validate the ID of the MSI Capability structure. */
+	if (cdns_pcie_ep_fn_readb(pcie, fn, cap) != PCI_CAP_ID_MSI)
+		return -EINVAL;
+
+	/* Validate that the MSI feature is actually enabled. */
+	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
+	if (!(flags & PCI_MSI_FLAGS_ENABLE))
+		return -EINVAL;
+
+	/*
+	 * Get the Multiple Message Enable bitfield from the Message Control
+	 * register.
+	 */
+	mmc = (flags & PCI_MSI_FLAGS_QMASK) >> 1;
+	mme = (flags & PCI_MSI_FLAGS_QSIZE) >> 4;
+	if (mme > mmc)
+		mme = mmc;
+	if (mme > 5)
+		mme = 5;
+
+	return mme;
+}
+
+static void cdns_pcie_ep_assert_intx(struct cdns_pcie_ep *ep, u8 fn,
+				     u8 intx, bool is_asserted)
+{
+	struct cdns_pcie *pcie = &ep->pcie;
+	u32 r = ep->data->max_regions - 1;
+	u32 offset;
+	u16 status;
+	u8 msg_code;
+
+	intx &= 3;
+
+	/* Set the outbound region if needed. */
+	if (unlikely(ep->irq_pci_addr != CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY)) {
+		/* Last region was reserved for IRQ writes. */
+		cdns_pcie_set_outbound_region_for_normal_msg(pcie, r,
+							     ep->irq_phys_addr);
+		ep->irq_pci_addr = CDNS_PCIE_EP_IRQ_PCI_ADDR_LEGACY;
+	}
+
+	if (is_asserted) {
+		ep->irq_pending |= BIT(intx);
+		msg_code = MSG_CODE_ASSERT_INTA + intx;
+	} else {
+		ep->irq_pending &= ~BIT(intx);
+		msg_code = MSG_CODE_DEASSERT_INTA + intx;
+	}
+
+	status = cdns_pcie_ep_fn_readw(pcie, fn, PCI_STATUS);
+	if (((status & PCI_STATUS_INTERRUPT) != 0) ^ (ep->irq_pending != 0)) {
+		status ^= PCI_STATUS_INTERRUPT;
+		cdns_pcie_ep_fn_writew(pcie, fn, PCI_STATUS, status);
+	}
+
+	offset = CDNS_PCIE_NORMAL_MSG_ROUTING(MSG_ROUTING_LOCAL) |
+		 CDNS_PCIE_NORMAL_MSG_CODE(msg_code) |
+		 CDNS_PCIE_MSG_NO_DATA;
+	writel(0, ep->irq_cpu_addr + offset);
+}
+
+static int cdns_pcie_ep_send_legacy_irq(struct cdns_pcie_ep *ep, u8 fn, u8 intx)
+{
+	u16 cmd;
+
+	cmd = cdns_pcie_ep_fn_readw(&ep->pcie, fn, PCI_COMMAND);
+	if (cmd & PCI_COMMAND_INTX_DISABLE)
+		return -EINVAL;
+
+	cdns_pcie_ep_assert_intx(ep, fn, intx, true);
+	mdelay(1);
+	cdns_pcie_ep_assert_intx(ep, fn, intx, false);
+	return 0;
+}
+
+static int cdns_pcie_ep_raise_irq(struct pci_epc *epc,
+				  enum pci_epc_irq_type type, u8 interrupt_num)
+{
+	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
+	struct cdns_pcie *pcie = &ep->pcie;
+	u32 cap = CDNS_PCIE_EP_FUNC_MSI_CAP_OFFSET;
+	u16 flags, mmc, mme, data, data_mask;
+	u8 msi_count;
+	u64 pci_addr, pci_addr_mask = 0xff;
+	u8 fn = 0;
+
+	/* Handle legacy IRQ. */
+	if (type == PCI_EPC_IRQ_LEGACY)
+		return cdns_pcie_ep_send_legacy_irq(ep, fn, 0);
+
+	/* Otherwise MSI. */
+	if (type != PCI_EPC_IRQ_MSI)
+		return -EINVAL;
+
+	/* Check whether the MSI feature has been enabled by the PCI host. */
+	flags = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_FLAGS);
+	if (!(flags & PCI_MSI_FLAGS_ENABLE))
+		return -EINVAL;
+
+	/* Get the number of enabled MSIs */
+	mmc = (flags & PCI_MSI_FLAGS_QMASK) >> 1;
+	mme = (flags & PCI_MSI_FLAGS_QSIZE) >> 4;
+	if (mme > mmc)
+		mme = mmc;
+	if (mme > 5)
+		mme = 5;
+
+	msi_count = 1 << mme;
+	if (!interrupt_num || interrupt_num > msi_count)
+		return -EINVAL;
+
+	/* Compute the data value to be written. */
+	data_mask = msi_count - 1;
+	data = cdns_pcie_ep_fn_readw(pcie, fn, cap + PCI_MSI_DATA_64);
+	data = (data & ~data_mask) | ((interrupt_num - 1) & data_mask);
+
+	/* Get the PCI address where to write the data into. */
+	pci_addr = cdns_pcie_ep_fn_readl(pcie, fn, cap + PCI_MSI_ADDRESS_HI);
+	pci_addr <<= 32;
+	pci_addr |= cdns_pcie_ep_fn_readl(pcie, fn, cap + PCI_MSI_ADDRESS_LO);
+	pci_addr &= GENMASK_ULL(63, 2);
+
+	/* Set the outbound region if needed. */
+	if (unlikely(ep->irq_pci_addr != pci_addr)) {
+		/* Last region was reserved for IRQ writes. */
+		cdns_pcie_set_outbound_region(pcie, ep->data->max_regions - 1,
+					      false,
+					      ep->irq_phys_addr,
+					      pci_addr & ~pci_addr_mask,
+					      pci_addr_mask + 1);
+		ep->irq_pci_addr = pci_addr;
+	}
+	writew(data, ep->irq_cpu_addr + (pci_addr & pci_addr_mask));
+
+	return 0;
+}
+
+static int cdns_pcie_ep_start(struct pci_epc *epc)
+{
+	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
+	struct cdns_pcie *pcie = &ep->pcie;
+	struct pci_epf *epf;
+	u32 cfg;
+	u8 fn = 0;
+
+	/* Enable this endpoint function. */
+	cfg = cdns_pcie_readl(pcie, CDNS_PCIE_LM_EP_FUNC_CFG);
+	cfg |= BIT(fn);
+	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, cfg);
+
+	/*
+	 * Already linked-up: don't call directly pci_epc_linkup() because we've
+	 * already locked the epc->lock.
+	 */
+	list_for_each_entry(epf, &epc->pci_epf, list)
+		pci_epf_linkup(epf);
+
+	return 0;
+}
+
+static void cdns_pcie_ep_stop(struct pci_epc *epc)
+{
+	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
+	struct cdns_pcie *pcie = &ep->pcie;
+	u32 cfg;
+	u8 fn = 0;
+
+	/* Disable this endpoint function (function 0 can't be disabled). */
+	cfg = cdns_pcie_readl(pcie, CDNS_PCIE_LM_EP_FUNC_CFG);
+	cfg &= ~BIT(fn);
+	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, cfg);
+}
+
+static const struct pci_epc_ops cdns_pcie_epc_ops = {
+	.write_header	= cdns_pcie_ep_write_header,
+	.set_bar	= cdns_pcie_ep_set_bar,
+	.clear_bar	= cdns_pcie_ep_clear_bar,
+	.map_addr	= cdns_pcie_ep_map_addr,
+	.unmap_addr	= cdns_pcie_ep_unmap_addr,
+	.set_msi	= cdns_pcie_ep_set_msi,
+	.get_msi	= cdns_pcie_ep_get_msi,
+	.raise_irq	= cdns_pcie_ep_raise_irq,
+	.start		= cdns_pcie_ep_start,
+	.stop		= cdns_pcie_ep_stop,
+};
+
+static const struct cdns_pcie_ep_data cdns_pcie_ep_data = {
+	.max_regions	= 16,
+};
+
+static const struct of_device_id cdns_pcie_ep_of_match[] = {
+	{ .compatible = "cdns,cdns-pcie-ep",
+	  .data = &cdns_pcie_ep_data },
+
+	{ },
+};
+
+static int cdns_pcie_ep_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *np = dev->of_node;
+	const struct of_device_id *of_id;
+	struct cdns_pcie_ep *ep;
+	struct cdns_pcie *pcie;
+	struct pci_epc *epc;
+	struct resource *res;
+	size_t max_regions;
+	int ret;
+
+	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
+	if (!ep)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, ep);
+
+	pcie = &ep->pcie;
+	pcie->is_rc = false;
+
+	of_id = of_match_node(cdns_pcie_ep_of_match, np);
+	ep->data = (const struct cdns_pcie_ep_data *)of_id->data;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "reg");
+	pcie->reg_base = devm_ioremap_resource(dev, res);
+	if (IS_ERR(pcie->reg_base)) {
+		dev_err(dev, "missing \"reg\"\n");
+		return PTR_ERR(pcie->reg_base);
+	}
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "mem");
+	if (!res) {
+		dev_err(dev, "missing \"mem\"\n");
+		return -EINVAL;
+	}
+	pcie->mem_res = res;
+
+	max_regions = ep->data->max_regions;
+	ep->ob_addr = devm_kzalloc(dev, max_regions * sizeof(*ep->ob_addr),
+				   GFP_KERNEL);
+	if (!ep->ob_addr)
+		return -ENOMEM;
+
+	pm_runtime_enable(dev);
+	ret = pm_runtime_get_sync(dev);
+	if (ret < 0) {
+		dev_err(dev, "pm_runtime_get_sync() failed\n");
+		goto err_get_sync;
+	}
+
+	/* Disable all but function 0 (anyway BIT(0) is hardwired to 1). */
+	cdns_pcie_writel(pcie, CDNS_PCIE_LM_EP_FUNC_CFG, BIT(0));
+
+	epc = devm_pci_epc_create(dev, &cdns_pcie_epc_ops);
+	if (IS_ERR(epc)) {
+		dev_err(dev, "failed to create epc device\n");
+		ret = PTR_ERR(epc);
+		goto err_init;
+	}
+
+	ep->epc = epc;
+	epc_set_drvdata(epc, ep);
+
+	ret = of_property_read_u8(np, "max-functions", &epc->max_functions);
+	if (ret < 0)
+		epc->max_functions = 1;
+
+	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
+			       resource_size(pcie->mem_res));
+	if (ret < 0) {
+		dev_err(dev, "failed to initialize the memory space\n");
+		goto err_init;
+	}
+
+	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
+						  SZ_128K);
+	if (!ep->irq_cpu_addr) {
+		dev_err(dev, "failed to reserve memory space for MSI\n");
+		ret = -ENOMEM;
+		goto free_epc_mem;
+	}
+	ep->irq_pci_addr = CDNS_PCIE_EP_IRQ_PCI_ADDR_NONE;
+
+	return 0;
+
+ free_epc_mem:
+	pci_epc_mem_exit(epc);
+
+ err_init:
+	pm_runtime_put_sync(dev);
+
+ err_get_sync:
+	pm_runtime_disable(dev);
+
+	return ret;
+}
+
+static struct platform_driver cdns_pcie_ep_driver = {
+	.driver = {
+		.name = "cdns-pcie-ep",
+		.of_match_table = cdns_pcie_ep_of_match,
+	},
+	.probe = cdns_pcie_ep_probe,
+};
+builtin_platform_driver(cdns_pcie_ep_driver);