Re: [PATCH v5 1/3] PCI: altera: Add Stratix 10 PCIe support

[Lorenzo Pieralisi]

On Tue, Feb 26, 2019 at 05:15:46PM +0800, Ley Foon Tan wrote:
> Add PCIe Root Port support for Stratix 10 device.
> 
> Main differences compare with PCIe Root Port IP on Cyclone V
> and Arria 10 devices:
> 
> - HIP interface to access Root Port configuration register.
> - TLP programming flow:
>   - One REG0 register
>   - Don't need to check alignment
> 
> Signed-off-by: Ley Foon Tan <ley.foon.tan@xxxxxxxxx>
> ---
>  drivers/pci/controller/pcie-altera.c | 266 ++++++++++++++++++++++++---
>  1 file changed, 242 insertions(+), 24 deletions(-)
> 
> diff --git a/drivers/pci/controller/pcie-altera.c b/drivers/pci/controller/pcie-altera.c
> index 7d05e51205b3..b3c05f2f309b 100644
> --- a/drivers/pci/controller/pcie-altera.c
> +++ b/drivers/pci/controller/pcie-altera.c
> @@ -11,6 +11,7 @@
>  #include <linux/irqchip/chained_irq.h>
>  #include <linux/init.h>
>  #include <linux/of_address.h>
> +#include <linux/of_device.h>
>  #include <linux/of_irq.h>
>  #include <linux/of_pci.h>
>  #include <linux/pci.h>
> @@ -37,7 +38,12 @@
>  #define RP_LTSSM_MASK			0x1f
>  #define LTSSM_L0			0xf
>  
> -#define PCIE_CAP_OFFSET			0x80
> +#define S10_RP_TX_CNTRL			0x2004
> +#define S10_RP_RXCPL_REG		0x2008
> +#define S10_RP_RXCPL_STATUS		0x200C
> +#define S10_RP_CFG_ADDR(pcie, reg)	\
> +	(((pcie)->hip_base) + (reg) + (1 << 20))
> +
>  /* TLP configuration type 0 and 1 */
>  #define TLP_FMTTYPE_CFGRD0		0x04	/* Configuration Read Type 0 */
>  #define TLP_FMTTYPE_CFGWR0		0x44	/* Configuration Write Type 0 */
> @@ -49,18 +55,19 @@
>  #define RP_DEVFN			0
>  #define TLP_REQ_ID(bus, devfn)		(((bus) << 8) | (devfn))
>  #define TLP_CFGRD_DW0(pcie, bus)					\
> -    ((((bus == pcie->root_bus_nr) ? TLP_FMTTYPE_CFGRD0			\
> -				    : TLP_FMTTYPE_CFGRD1) << 24) |	\
> -     TLP_PAYLOAD_SIZE)
> +	((((bus == pcie->root_bus_nr) ? pcie->pcie_data->cfgrd0		\
> +				: pcie->pcie_data->cfgrd1) << 24) |	\
> +				TLP_PAYLOAD_SIZE)
>  #define TLP_CFGWR_DW0(pcie, bus)					\
> -    ((((bus == pcie->root_bus_nr) ? TLP_FMTTYPE_CFGWR0			\
> -				    : TLP_FMTTYPE_CFGWR1) << 24) |	\
> -     TLP_PAYLOAD_SIZE)
> +	((((bus == pcie->root_bus_nr) ? pcie->pcie_data->cfgwr0		\
> +				: pcie->pcie_data->cfgwr1) << 24) |	\
> +				TLP_PAYLOAD_SIZE)
>  #define TLP_CFG_DW1(pcie, tag, be)	\
> -    (((TLP_REQ_ID(pcie->root_bus_nr,  RP_DEVFN)) << 16) | (tag << 8) | (be))
> +	(((TLP_REQ_ID(pcie->root_bus_nr,  RP_DEVFN)) << 16) | (tag << 8) | (be))
>  #define TLP_CFG_DW2(bus, devfn, offset)	\
>  				(((bus) << 24) | ((devfn) << 16) | (offset))
>  #define TLP_COMP_STATUS(s)		(((s) >> 13) & 7)
> +#define TLP_BYTE_COUNT(s)		(((s) >> 0) & 0xfff)
>  #define TLP_HDR_SIZE			3
>  #define TLP_LOOP			500
>  
> @@ -69,14 +76,47 @@
>  
>  #define DWORD_MASK			3
>  
> +#define S10_TLP_FMTTYPE_CFGRD0		0x05
> +#define S10_TLP_FMTTYPE_CFGRD1		0x04
> +#define S10_TLP_FMTTYPE_CFGWR0		0x45
> +#define S10_TLP_FMTTYPE_CFGWR1		0x44
> +
> +enum altera_pcie_version {
> +	ALTERA_PCIE_V1 = 0,
> +	ALTERA_PCIE_V2,
> +};
> +
>  struct altera_pcie {
>  	struct platform_device	*pdev;
> -	void __iomem		*cra_base;	/* DT Cra */
> +	void __iomem		*cra_base;
> +	void __iomem		*hip_base;
>  	int			irq;
>  	u8			root_bus_nr;
>  	struct irq_domain	*irq_domain;
>  	struct resource		bus_range;
>  	struct list_head	resources;
> +	const struct altera_pcie_data	*pcie_data;
> +};
> +
> +struct altera_pcie_ops {
> +	int (*tlp_read_pkt)(struct altera_pcie *pcie, u32 *value);
> +	void (*tlp_write_pkt)(struct altera_pcie *pcie, u32 *headers,
> +			      u32 data, bool align);
> +	bool (*get_link_status)(struct altera_pcie *pcie);
> +	int (*rp_read_cfg)(struct altera_pcie *pcie, int where,
> +			   int size, u32 *value);
> +	int (*rp_write_cfg)(struct altera_pcie *pcie, u8 busno,
> +			    int where, int size, u32 value);
> +};
> +
> +struct altera_pcie_data {
> +	const struct altera_pcie_ops *ops;
> +	enum altera_pcie_version version;
> +	u32 cap_offset;		/* PCIe capability structure register offset */
> +	u32 cfgrd0;
> +	u32 cfgrd1;
> +	u32 cfgwr0;
> +	u32 cfgwr1;
>  };
>  
>  struct tlp_rp_regpair_t {
> @@ -101,6 +141,15 @@ static bool altera_pcie_link_up(struct altera_pcie *pcie)
>  	return !!((cra_readl(pcie, RP_LTSSM) & RP_LTSSM_MASK) == LTSSM_L0);
>  }
>  
> +static bool s10_altera_pcie_link_up(struct altera_pcie *pcie)
> +{
> +	void __iomem *addr = S10_RP_CFG_ADDR(pcie,
> +				   pcie->pcie_data->cap_offset +
> +				   PCI_EXP_LNKSTA);
> +
> +	return !!(readw(addr) & PCI_EXP_LNKSTA_DLLLA);
> +}
> +
>  /*
>   * Altera PCIe port uses BAR0 of RC's configuration space as the translation
>   * from PCI bus to native BUS.  Entire DDR region is mapped into PCIe space
> @@ -128,12 +177,18 @@ static void tlp_write_tx(struct altera_pcie *pcie,
>  	cra_writel(pcie, tlp_rp_regdata->ctrl, RP_TX_CNTRL);
>  }
>  
> +static void s10_tlp_write_tx(struct altera_pcie *pcie, u32 reg0, u32 ctrl)
> +{
> +	cra_writel(pcie, reg0, RP_TX_REG0);
> +	cra_writel(pcie, ctrl, S10_RP_TX_CNTRL);
> +}
> +
>  static bool altera_pcie_valid_device(struct altera_pcie *pcie,
>  				     struct pci_bus *bus, int dev)
>  {
>  	/* If there is no link, then there is no device */
>  	if (bus->number != pcie->root_bus_nr) {
> -		if (!altera_pcie_link_up(pcie))
> +		if (!pcie->pcie_data->ops->get_link_status(pcie))
>  			return false;
>  	}
>  
> @@ -183,6 +238,49 @@ static int tlp_read_packet(struct altera_pcie *pcie, u32 *value)
>  	return PCIBIOS_DEVICE_NOT_FOUND;
>  }
>  
> +static int s10_tlp_read_packet(struct altera_pcie *pcie, u32 *value)
> +{
> +	int i;
> +	u32 ctrl;
> +	u32 comp_status;
> +	u32 dw[4];
> +	u32 count = 0;
> +
> +	for (i = 0; i < TLP_LOOP; i++) {
> +		ctrl = cra_readl(pcie, S10_RP_RXCPL_STATUS);
> +		if (!(ctrl & RP_RXCPL_SOP))
> +			continue;
> +
> +		/* Read first DW */
> +		dw[count++] = cra_readl(pcie, S10_RP_RXCPL_REG);

Maybe I still do not understand it, but why this is not:

dw[0] = cra_readl(pcie, S10_RP_RXCPL_REG);

and

count = 1;

above ?

> +
> +		/* Poll for EOP */
> +		for (i = 0; i < TLP_LOOP; i++) {
> +			if (count >= ARRAY_SIZE(dw))

Can't this check be part of the loop condition ? And should not
we get out of this function when this inner loop terminate ?

> +				return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +			ctrl = cra_readl(pcie, S10_RP_RXCPL_STATUS);
> +			dw[count++] = cra_readl(pcie, S10_RP_RXCPL_REG);
> +			if (ctrl & RP_RXCPL_EOP) {
> +				comp_status = TLP_COMP_STATUS(dw[1]);
> +				if (comp_status)
> +					return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +				if (value &&
> +				    TLP_BYTE_COUNT(dw[1]) == sizeof(u32) &&
> +				    count == 4)
> +					*value = dw[3];
> +
> +				return PCIBIOS_SUCCESSFUL;
> +			}
> +		}
> +
> +		udelay(5);

When should this delay be executed ? If we get here should not we bail
out of the loop ?

I would like to try to merge these patches but I need a clarification
on this code first, I can update the code myself for the sake of making
progress.

Lorenzo

> +	}
> +
> +	return PCIBIOS_DEVICE_NOT_FOUND;
> +}
> +
>  static void tlp_write_packet(struct altera_pcie *pcie, u32 *headers,
>  			     u32 data, bool align)
>  {
> @@ -210,6 +308,15 @@ static void tlp_write_packet(struct altera_pcie *pcie, u32 *headers,
>  	tlp_write_tx(pcie, &tlp_rp_regdata);
>  }
>  
> +static void s10_tlp_write_packet(struct altera_pcie *pcie, u32 *headers,
> +				 u32 data, bool dummy)
> +{
> +	s10_tlp_write_tx(pcie, headers[0], RP_TX_SOP);
> +	s10_tlp_write_tx(pcie, headers[1], 0);
> +	s10_tlp_write_tx(pcie, headers[2], 0);
> +	s10_tlp_write_tx(pcie, data, RP_TX_EOP);
> +}
> +
>  static int tlp_cfg_dword_read(struct altera_pcie *pcie, u8 bus, u32 devfn,
>  			      int where, u8 byte_en, u32 *value)
>  {
> @@ -219,9 +326,9 @@ static int tlp_cfg_dword_read(struct altera_pcie *pcie, u8 bus, u32 devfn,
>  	headers[1] = TLP_CFG_DW1(pcie, TLP_READ_TAG, byte_en);
>  	headers[2] = TLP_CFG_DW2(bus, devfn, where);
>  
> -	tlp_write_packet(pcie, headers, 0, false);
> +	pcie->pcie_data->ops->tlp_write_pkt(pcie, headers, 0, false);
>  
> -	return tlp_read_packet(pcie, value);
> +	return pcie->pcie_data->ops->tlp_read_pkt(pcie, value);
>  }
>  
>  static int tlp_cfg_dword_write(struct altera_pcie *pcie, u8 bus, u32 devfn,
> @@ -236,11 +343,13 @@ static int tlp_cfg_dword_write(struct altera_pcie *pcie, u8 bus, u32 devfn,
>  
>  	/* check alignment to Qword */
>  	if ((where & 0x7) == 0)
> -		tlp_write_packet(pcie, headers, value, true);
> +		pcie->pcie_data->ops->tlp_write_pkt(pcie, headers,
> +						    value, true);
>  	else
> -		tlp_write_packet(pcie, headers, value, false);
> +		pcie->pcie_data->ops->tlp_write_pkt(pcie, headers,
> +						    value, false);
>  
> -	ret = tlp_read_packet(pcie, NULL);
> +	ret = pcie->pcie_data->ops->tlp_read_pkt(pcie, NULL);
>  	if (ret != PCIBIOS_SUCCESSFUL)
>  		return ret;
>  
> @@ -254,6 +363,53 @@ static int tlp_cfg_dword_write(struct altera_pcie *pcie, u8 bus, u32 devfn,
>  	return PCIBIOS_SUCCESSFUL;
>  }
>  
> +static int s10_rp_read_cfg(struct altera_pcie *pcie, int where,
> +			   int size, u32 *value)
> +{
> +	void *addr = S10_RP_CFG_ADDR(pcie, where);
> +
> +	switch (size) {
> +	case 1:
> +		*value = readb(addr);
> +		break;
> +	case 2:
> +		*value = readw(addr);
> +		break;
> +	default:
> +		*value = readl(addr);
> +		break;
> +	}
> +
> +	return PCIBIOS_SUCCESSFUL;
> +}
> +
> +static int s10_rp_write_cfg(struct altera_pcie *pcie, u8 busno,
> +			    int where, int size, u32 value)
> +{
> +	void *addr = S10_RP_CFG_ADDR(pcie, where);
> +
> +	switch (size) {
> +	case 1:
> +		writeb(value, addr);
> +		break;
> +	case 2:
> +		writew(value, addr);
> +		break;
> +	default:
> +		writel(value, addr);
> +		break;
> +	}
> +
> +	/*
> +	 * Monitor changes to PCI_PRIMARY_BUS register on root port
> +	 * and update local copy of root bus number accordingly.
> +	 */
> +	if (busno == pcie->root_bus_nr && where == PCI_PRIMARY_BUS)
> +		pcie->root_bus_nr = value & 0xff;
> +
> +	return PCIBIOS_SUCCESSFUL;
> +}
> +
>  static int _altera_pcie_cfg_read(struct altera_pcie *pcie, u8 busno,
>  				 unsigned int devfn, int where, int size,
>  				 u32 *value)
> @@ -262,6 +418,10 @@ static int _altera_pcie_cfg_read(struct altera_pcie *pcie, u8 busno,
>  	u32 data;
>  	u8 byte_en;
>  
> +	if (busno == pcie->root_bus_nr && pcie->pcie_data->ops->rp_read_cfg)
> +		return pcie->pcie_data->ops->rp_read_cfg(pcie, where,
> +							 size, value);
> +
>  	switch (size) {
>  	case 1:
>  		byte_en = 1 << (where & 3);
> @@ -302,6 +462,10 @@ static int _altera_pcie_cfg_write(struct altera_pcie *pcie, u8 busno,
>  	u32 shift = 8 * (where & 3);
>  	u8 byte_en;
>  
> +	if (busno == pcie->root_bus_nr && pcie->pcie_data->ops->rp_write_cfg)
> +		return pcie->pcie_data->ops->rp_write_cfg(pcie, busno,
> +						     where, size, value);
> +
>  	switch (size) {
>  	case 1:
>  		data32 = (value & 0xff) << shift;
> @@ -365,7 +529,8 @@ static int altera_read_cap_word(struct altera_pcie *pcie, u8 busno,
>  	int ret;
>  
>  	ret = _altera_pcie_cfg_read(pcie, busno, devfn,
> -				    PCIE_CAP_OFFSET + offset, sizeof(*value),
> +				    pcie->pcie_data->cap_offset + offset,
> +				    sizeof(*value),
>  				    &data);
>  	*value = data;
>  	return ret;
> @@ -375,7 +540,8 @@ static int altera_write_cap_word(struct altera_pcie *pcie, u8 busno,
>  				 unsigned int devfn, int offset, u16 value)
>  {
>  	return _altera_pcie_cfg_write(pcie, busno, devfn,
> -				      PCIE_CAP_OFFSET + offset, sizeof(value),
> +				      pcie->pcie_data->cap_offset + offset,
> +				      sizeof(value),
>  				      value);
>  }
>  
> @@ -403,7 +569,7 @@ static void altera_wait_link_retrain(struct altera_pcie *pcie)
>  	/* Wait for link is up */
>  	start_jiffies = jiffies;
>  	for (;;) {
> -		if (altera_pcie_link_up(pcie))
> +		if (pcie->pcie_data->ops->get_link_status(pcie))
>  			break;
>  
>  		if (time_after(jiffies, start_jiffies + LINK_UP_TIMEOUT)) {
> @@ -418,7 +584,7 @@ static void altera_pcie_retrain(struct altera_pcie *pcie)
>  {
>  	u16 linkcap, linkstat, linkctl;
>  
> -	if (!altera_pcie_link_up(pcie))
> +	if (!pcie->pcie_data->ops->get_link_status(pcie))
>  		return;
>  
>  	/*
> @@ -540,12 +706,20 @@ static int altera_pcie_parse_dt(struct altera_pcie *pcie)
>  	struct device *dev = &pcie->pdev->dev;
>  	struct platform_device *pdev = pcie->pdev;
>  	struct resource *cra;
> +	struct resource *hip;
>  
>  	cra = platform_get_resource_byname(pdev, IORESOURCE_MEM, "Cra");
>  	pcie->cra_base = devm_ioremap_resource(dev, cra);
>  	if (IS_ERR(pcie->cra_base))
>  		return PTR_ERR(pcie->cra_base);
>  
> +	if (pcie->pcie_data->version == ALTERA_PCIE_V2) {
> +		hip = platform_get_resource_byname(pdev, IORESOURCE_MEM, "Hip");
> +		pcie->hip_base = devm_ioremap_resource(&pdev->dev, hip);
> +		if (IS_ERR(pcie->hip_base))
> +			return PTR_ERR(pcie->hip_base);
> +	}
> +
>  	/* setup IRQ */
>  	pcie->irq = platform_get_irq(pdev, 0);
>  	if (pcie->irq < 0) {
> @@ -562,6 +736,48 @@ static void altera_pcie_host_init(struct altera_pcie *pcie)
>  	altera_pcie_retrain(pcie);
>  }
>  
> +static const struct altera_pcie_ops altera_pcie_ops_1_0 = {
> +	.tlp_read_pkt = tlp_read_packet,
> +	.tlp_write_pkt = tlp_write_packet,
> +	.get_link_status = altera_pcie_link_up,
> +};
> +
> +static const struct altera_pcie_ops altera_pcie_ops_2_0 = {
> +	.tlp_read_pkt = s10_tlp_read_packet,
> +	.tlp_write_pkt = s10_tlp_write_packet,
> +	.get_link_status = s10_altera_pcie_link_up,
> +	.rp_read_cfg = s10_rp_read_cfg,
> +	.rp_write_cfg = s10_rp_write_cfg,
> +};
> +
> +static const struct altera_pcie_data altera_pcie_1_0_data = {
> +	.ops = &altera_pcie_ops_1_0,
> +	.cap_offset = 0x80,
> +	.version = ALTERA_PCIE_V1,
> +	.cfgrd0 = TLP_FMTTYPE_CFGRD0,
> +	.cfgrd1 = TLP_FMTTYPE_CFGRD1,
> +	.cfgwr0 = TLP_FMTTYPE_CFGWR0,
> +	.cfgwr1 = TLP_FMTTYPE_CFGWR1,
> +};
> +
> +static const struct altera_pcie_data altera_pcie_2_0_data = {
> +	.ops = &altera_pcie_ops_2_0,
> +	.version = ALTERA_PCIE_V2,
> +	.cap_offset = 0x70,
> +	.cfgrd0 = S10_TLP_FMTTYPE_CFGRD0,
> +	.cfgrd1 = S10_TLP_FMTTYPE_CFGRD1,
> +	.cfgwr0 = S10_TLP_FMTTYPE_CFGWR0,
> +	.cfgwr1 = S10_TLP_FMTTYPE_CFGWR1,
> +};
> +
> +static const struct of_device_id altera_pcie_of_match[] = {
> +	{.compatible = "altr,pcie-root-port-1.0",
> +	 .data = &altera_pcie_1_0_data },
> +	{.compatible = "altr,pcie-root-port-2.0",
> +	 .data = &altera_pcie_2_0_data },
> +	{},
> +};
> +
>  static int altera_pcie_probe(struct platform_device *pdev)
>  {
>  	struct device *dev = &pdev->dev;
> @@ -570,6 +786,7 @@ static int altera_pcie_probe(struct platform_device *pdev)
>  	struct pci_bus *child;
>  	struct pci_host_bridge *bridge;
>  	int ret;
> +	const struct of_device_id *match;
>  
>  	bridge = devm_pci_alloc_host_bridge(dev, sizeof(*pcie));
>  	if (!bridge)
> @@ -578,6 +795,12 @@ static int altera_pcie_probe(struct platform_device *pdev)
>  	pcie = pci_host_bridge_priv(bridge);
>  	pcie->pdev = pdev;
>  
> +	match = of_match_device(altera_pcie_of_match, &pdev->dev);
> +	if (!match)
> +		return -ENODEV;
> +
> +	pcie->pcie_data = match->data;
> +
>  	ret = altera_pcie_parse_dt(pcie);
>  	if (ret) {
>  		dev_err(dev, "Parsing DT failed\n");
> @@ -628,11 +851,6 @@ static int altera_pcie_probe(struct platform_device *pdev)
>  	return ret;
>  }
>  
> -static const struct of_device_id altera_pcie_of_match[] = {
> -	{ .compatible = "altr,pcie-root-port-1.0", },
> -	{},
> -};
> -
>  static struct platform_driver altera_pcie_driver = {
>  	.probe		= altera_pcie_probe,
>  	.driver = {
> -- 
> 2.19.0
>