Merge git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6: (110 commits)
  sh: i2c-sh7760: Replase from ctrl_* to __raw_*
  sh: clkfwk: Shuffle around to match the intc split up.
  sh: clkfwk: modify for_each_frequency end condition
  sh: fix clk_get() error handling
  sh: clkfwk: Fix fault in frequency iterator.
  sh: clkfwk: Add a helper for rate rounding by divisor ranges.
  sh: clkfwk: Abstract rate rounding helper.
  sh: clkfwk: support clock remapping.
  sh: pci: Convert to upper/lower_32_bits() helpers.
  sh: mach-sdk7786: Add support for the FPGA SRAM.
  sh: Provide a generic SRAM pool for tiny memories.
  sh: pci: Support secondary FPGA-driven PCIe clocks on SDK7786.
  sh: pci: Support slot 4 routing on SDK7786.
  sh: Fix up PMB locking.
  sh: mach-sdk7786: Add support for fpga gpios.
  sh: use pr_fmt for clock framework, too.
  sh: remove name and id from struct clk
  sh: free-without-alloc fix for sh_mobile_lcdcfb
  sh: perf: Set up perf_max_events.
  sh: perf: Support SH-X3 hardware counters.
  ...

Fix up trivial conflicts (perf_max_events got removed) in arch/sh/kernel/perf_event.c

1 files changed, 195 insertions, 56 deletions

diff --git a/arch/sh/drivers/pci/pcie-sh7786.c b/arch/sh/drivers/pci/pcie-sh7786.c
index 68cb9b0ac9d2..96e9b058aa1d 100644
--- a/arch/sh/drivers/pci/pcie-sh7786.c
+++ b/arch/sh/drivers/pci/pcie-sh7786.c
@@ -13,11 +13,14 @@
 #include <linux/io.h>
 #include <linux/delay.h>
 #include <linux/slab.h>
+#include <linux/clk.h>
+#include <linux/sh_clk.h>
 #include "pcie-sh7786.h"
 #include <asm/sizes.h>
 
 struct sh7786_pcie_port {
 	struct pci_channel	*hose;
+	struct clk		*fclk, phy_clk;
 	unsigned int		index;
 	int			endpoint;
 	int			link;
@@ -51,6 +54,7 @@ static struct resource sh7786_pci0_resources[] = {
 		.name	= "PCIe0 MEM 2",
 		.start	= 0xfe100000,
 		.end	= 0xfe100000 + SZ_1M - 1,
+		.flags	= IORESOURCE_MEM,
 	},
 };
 
@@ -74,6 +78,7 @@ static struct resource sh7786_pci1_resources[] = {
 		.name	= "PCIe1 MEM 2",
 		.start	= 0xfe300000,
 		.end	= 0xfe300000 + SZ_1M - 1,
+		.flags	= IORESOURCE_MEM,
 	},
 };
 
@@ -82,6 +87,7 @@ static struct resource sh7786_pci2_resources[] = {
 		.name	= "PCIe2 IO",
 		.start	= 0xfc800000,
 		.end	= 0xfc800000 + SZ_4M - 1,
+		.flags	= IORESOURCE_IO,
 	}, {
 		.name	= "PCIe2 MEM 0",
 		.start	= 0x80000000,
@@ -96,6 +102,7 @@ static struct resource sh7786_pci2_resources[] = {
 		.name	= "PCIe2 MEM 2",
 		.start	= 0xfcd00000,
 		.end	= 0xfcd00000 + SZ_1M - 1,
+		.flags	= IORESOURCE_MEM,
 	},
 };
 
@@ -117,7 +124,29 @@ static struct pci_channel sh7786_pci_channels[] = {
 	DEFINE_CONTROLLER(0xfcc00000, 2),
 };
 
-static int phy_wait_for_ack(struct pci_channel *chan)
+static struct clk fixed_pciexclkp = {
+	.rate = 100000000,	/* 100 MHz reference clock */
+};
+
+static void __devinit sh7786_pci_fixup(struct pci_dev *dev)
+{
+	/*
+	 * Prevent enumeration of root complex resources.
+	 */
+	if (pci_is_root_bus(dev->bus) && dev->devfn == 0) {
+		int i;
+
+		for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
+			dev->resource[i].start	= 0;
+			dev->resource[i].end	= 0;
+			dev->resource[i].flags	= 0;
+		}
+	}
+}
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_SH7786,
+			 sh7786_pci_fixup);
+
+static int __init phy_wait_for_ack(struct pci_channel *chan)
 {
 	unsigned int timeout = 100;
 
@@ -131,7 +160,7 @@ static int phy_wait_for_ack(struct pci_channel *chan)
 	return -ETIMEDOUT;
 }
 
-static int pci_wait_for_irq(struct pci_channel *chan, unsigned int mask)
+static int __init pci_wait_for_irq(struct pci_channel *chan, unsigned int mask)
 {
 	unsigned int timeout = 100;
 
@@ -145,19 +174,14 @@ static int pci_wait_for_irq(struct pci_channel *chan, unsigned int mask)
 	return -ETIMEDOUT;
 }
 
-static void phy_write_reg(struct pci_channel *chan, unsigned int addr,
-			  unsigned int lane, unsigned int data)
+static void __init phy_write_reg(struct pci_channel *chan, unsigned int addr,
+				 unsigned int lane, unsigned int data)
 {
-	unsigned long phyaddr, ctrl;
+	unsigned long phyaddr;
 
 	phyaddr = (1 << BITS_CMD) + ((lane & 0xf) << BITS_LANE) +
 			((addr & 0xff) << BITS_ADR);
 
-	/* Enable clock */
-	ctrl = pci_read_reg(chan, SH4A_PCIEPHYCTLR);
-	ctrl |= (1 << BITS_CKE);
-	pci_write_reg(chan, ctrl, SH4A_PCIEPHYCTLR);
-
 	/* Set write data */
 	pci_write_reg(chan, data, SH4A_PCIEPHYDOUTR);
 	pci_write_reg(chan, phyaddr, SH4A_PCIEPHYADRR);
@@ -165,20 +189,74 @@ static void phy_write_reg(struct pci_channel *chan, unsigned int addr,
 	phy_wait_for_ack(chan);
 
 	/* Clear command */
+	pci_write_reg(chan, 0, SH4A_PCIEPHYDOUTR);
 	pci_write_reg(chan, 0, SH4A_PCIEPHYADRR);
 
 	phy_wait_for_ack(chan);
+}
 
-	/* Disable clock */
-	ctrl = pci_read_reg(chan, SH4A_PCIEPHYCTLR);
-	ctrl &= ~(1 << BITS_CKE);
-	pci_write_reg(chan, ctrl, SH4A_PCIEPHYCTLR);
+static int __init pcie_clk_init(struct sh7786_pcie_port *port)
+{
+	struct pci_channel *chan = port->hose;
+	struct clk *clk;
+	char fclk_name[16];
+	int ret;
+
+	/*
+	 * First register the fixed clock
+	 */
+	ret = clk_register(&fixed_pciexclkp);
+	if (unlikely(ret != 0))
+		return ret;
+
+	/*
+	 * Grab the port's function clock, which the PHY clock depends
+	 * on. clock lookups don't help us much at this point, since no
+	 * dev_id is available this early. Lame.
+	 */
+	snprintf(fclk_name, sizeof(fclk_name), "pcie%d_fck", port->index);
+
+	port->fclk = clk_get(NULL, fclk_name);
+	if (IS_ERR(port->fclk)) {
+		ret = PTR_ERR(port->fclk);
+		goto err_fclk;
+	}
+
+	clk_enable(port->fclk);
+
+	/*
+	 * And now, set up the PHY clock
+	 */
+	clk = &port->phy_clk;
+
+	memset(clk, 0, sizeof(struct clk));
+
+	clk->parent = &fixed_pciexclkp;
+	clk->enable_reg = (void __iomem *)(chan->reg_base + SH4A_PCIEPHYCTLR);
+	clk->enable_bit = BITS_CKE;
+
+	ret = sh_clk_mstp32_register(clk, 1);
+	if (unlikely(ret < 0))
+		goto err_phy;
+
+	return 0;
+
+err_phy:
+	clk_disable(port->fclk);
+	clk_put(port->fclk);
+err_fclk:
+	clk_unregister(&fixed_pciexclkp);
+
+	return ret;
 }
 
-static int phy_init(struct pci_channel *chan)
+static int __init phy_init(struct sh7786_pcie_port *port)
 {
+	struct pci_channel *chan = port->hose;
 	unsigned int timeout = 100;
 
+	clk_enable(&port->phy_clk);
+
 	/* Initialize the phy */
 	phy_write_reg(chan, 0x60, 0xf, 0x004b008b);
 	phy_write_reg(chan, 0x61, 0xf, 0x00007b41);
@@ -187,9 +265,13 @@ static int phy_init(struct pci_channel *chan)
 	phy_write_reg(chan, 0x66, 0xf, 0x00000010);
 	phy_write_reg(chan, 0x74, 0xf, 0x0007001c);
 	phy_write_reg(chan, 0x79, 0xf, 0x01fc000d);
+	phy_write_reg(chan, 0xb0, 0xf, 0x00000610);
 
 	/* Deassert Standby */
-	phy_write_reg(chan, 0x67, 0xf, 0x00000400);
+	phy_write_reg(chan, 0x67, 0x1, 0x00000400);
+
+	/* Disable clock */
+	clk_disable(&port->phy_clk);
 
 	while (timeout--) {
 		if (pci_read_reg(chan, SH4A_PCIEPHYSR))
@@ -201,22 +283,33 @@ static int phy_init(struct pci_channel *chan)
 	return -ETIMEDOUT;
 }
 
-static int pcie_init(struct sh7786_pcie_port *port)
+static void __init pcie_reset(struct sh7786_pcie_port *port)
+{
+	struct pci_channel *chan = port->hose;
+
+	pci_write_reg(chan, 1, SH4A_PCIESRSTR);
+	pci_write_reg(chan, 0, SH4A_PCIETCTLR);
+	pci_write_reg(chan, 0, SH4A_PCIESRSTR);
+	pci_write_reg(chan, 0, SH4A_PCIETXVC0SR);
+}
+
+static int __init pcie_init(struct sh7786_pcie_port *port)
 {
 	struct pci_channel *chan = port->hose;
 	unsigned int data;
 	phys_addr_t memphys;
 	size_t memsize;
-	int ret, i;
+	int ret, i, win;
 
 	/* Begin initialization */
-	pci_write_reg(chan, 0, SH4A_PCIETCTLR);
+	pcie_reset(port);
 
-	/* Initialize as type1. */
-	data = pci_read_reg(chan, SH4A_PCIEPCICONF3);
-	data &= ~(0x7f << 16);
-	data |= PCI_HEADER_TYPE_BRIDGE << 16;
-	pci_write_reg(chan, data, SH4A_PCIEPCICONF3);
+	/*
+	 * Initial header for port config space is type 1, set the device
+	 * class to match. Hardware takes care of propagating the IDSETR
+	 * settings, so there is no need to bother with a quirk.
+	 */
+	pci_write_reg(chan, PCI_CLASS_BRIDGE_PCI << 16, SH4A_PCIEIDSETR1);
 
 	/* Initialize default capabilities. */
 	data = pci_read_reg(chan, SH4A_PCIEEXPCAP0);
@@ -268,30 +361,33 @@ static int pcie_init(struct sh7786_pcie_port *port)
 	 * LAR1/LAMR1.
 	 */
 	if (memsize > SZ_512M) {
-		__raw_writel(memphys + SZ_512M, chan->reg_base + SH4A_PCIELAR1);
-		__raw_writel(((memsize - SZ_512M) - SZ_256) | 1,
-			     chan->reg_base + SH4A_PCIELAMR1);
+		pci_write_reg(chan, memphys + SZ_512M, SH4A_PCIELAR1);
+		pci_write_reg(chan, ((memsize - SZ_512M) - SZ_256) | 1,
+			      SH4A_PCIELAMR1);
 		memsize = SZ_512M;
 	} else {
 		/*
 		 * Otherwise just zero it out and disable it.
 		 */
-		__raw_writel(0, chan->reg_base + SH4A_PCIELAR1);
-		__raw_writel(0, chan->reg_base + SH4A_PCIELAMR1);
+		pci_write_reg(chan, 0, SH4A_PCIELAR1);
+		pci_write_reg(chan, 0, SH4A_PCIELAMR1);
 	}
 
 	/*
 	 * LAR0/LAMR0 covers up to the first 512MB, which is enough to
 	 * cover all of lowmem on most platforms.
 	 */
-	__raw_writel(memphys, chan->reg_base + SH4A_PCIELAR0);
-	__raw_writel((memsize - SZ_256) | 1, chan->reg_base + SH4A_PCIELAMR0);
+	pci_write_reg(chan, memphys, SH4A_PCIELAR0);
+	pci_write_reg(chan, (memsize - SZ_256) | 1, SH4A_PCIELAMR0);
 
 	/* Finish initialization */
 	data = pci_read_reg(chan, SH4A_PCIETCTLR);
 	data |= 0x1;
 	pci_write_reg(chan, data, SH4A_PCIETCTLR);
 
+	/* Let things settle down a bit.. */
+	mdelay(100);
+
 	/* Enable DL_Active Interrupt generation */
 	data = pci_read_reg(chan, SH4A_PCIEDLINTENR);
 	data |= PCIEDLINTENR_DLL_ACT_ENABLE;
@@ -302,9 +398,12 @@ static int pcie_init(struct sh7786_pcie_port *port)
 	data |= PCIEMACCTLR_SCR_DIS | (0xff << 16);
 	pci_write_reg(chan, data, SH4A_PCIEMACCTLR);
 
+	/*
+	 * This will timeout if we don't have a link, but we permit the
+	 * port to register anyways in order to support hotplug on future
+	 * hardware.
+	 */
 	ret = pci_wait_for_irq(chan, MASK_INT_TX_CTRL);
-	if (unlikely(ret != 0))
-		return -ENODEV;
 
 	data = pci_read_reg(chan, SH4A_PCIEPCICONF1);
 	data &= ~(PCI_STATUS_DEVSEL_MASK << 16);
@@ -317,35 +416,48 @@ static int pcie_init(struct sh7786_pcie_port *port)
 
 	wmb();
 
-	data = pci_read_reg(chan, SH4A_PCIEMACSR);
-	printk(KERN_NOTICE "PCI: PCIe#%d link width %d\n",
-	       port->index, (data >> 20) & 0x3f);
-
+	if (ret == 0) {
+		data = pci_read_reg(chan, SH4A_PCIEMACSR);
+		printk(KERN_NOTICE "PCI: PCIe#%d x%d link detected\n",
+		       port->index, (data >> 20) & 0x3f);
+	} else
+		printk(KERN_NOTICE "PCI: PCIe#%d link down\n",
+		       port->index);
 
-	for (i = 0; i < chan->nr_resources; i++) {
+	for (i = win = 0; i < chan->nr_resources; i++) {
 		struct resource *res = chan->resources + i;
 		resource_size_t size;
-		u32 enable_mask;
+		u32 mask;
 
-		pci_write_reg(chan, 0x00000000, SH4A_PCIEPTCTLR(i));
+		/*
+		 * We can't use the 32-bit mode windows in legacy 29-bit
+		 * mode, so just skip them entirely.
+		 */
+		if ((res->flags & IORESOURCE_MEM_32BIT) && __in_29bit_mode())
+			continue;
 
-		size = resource_size(res);
+		pci_write_reg(chan, 0x00000000, SH4A_PCIEPTCTLR(win));
 
 		/*
 		 * The PAMR mask is calculated in units of 256kB, which
 		 * keeps things pretty simple.
 		 */
-		__raw_writel(((roundup_pow_of_two(size) / SZ_256K) - 1) << 18,
-			     chan->reg_base + SH4A_PCIEPAMR(i));
+		size = resource_size(res);
+		mask = (roundup_pow_of_two(size) / SZ_256K) - 1;
+		pci_write_reg(chan, mask << 18, SH4A_PCIEPAMR(win));
 
-		pci_write_reg(chan, 0x00000000, SH4A_PCIEPARH(i));
-		pci_write_reg(chan, 0x00000000, SH4A_PCIEPARL(i));
+		pci_write_reg(chan, upper_32_bits(res->start),
+			      SH4A_PCIEPARH(win));
+		pci_write_reg(chan, lower_32_bits(res->start),
+			      SH4A_PCIEPARL(win));
 
-		enable_mask = MASK_PARE;
+		mask = MASK_PARE;
 		if (res->flags & IORESOURCE_IO)
-			enable_mask |= MASK_SPC;
+			mask |= MASK_SPC;
+
+		pci_write_reg(chan, mask, SH4A_PCIEPTCTLR(win));
 
-		pci_write_reg(chan, enable_mask, SH4A_PCIEPTCTLR(i));
+		win++;
 	}
 
 	return 0;
@@ -356,26 +468,33 @@ int __init pcibios_map_platform_irq(struct pci_dev *pdev, u8 slot, u8 pin)
         return 71;
 }
 
-static int sh7786_pcie_core_init(void)
+static int __init sh7786_pcie_core_init(void)
 {
 	/* Return the number of ports */
 	return test_mode_pin(MODE_PIN12) ? 3 : 2;
 }
 
-static int __devinit sh7786_pcie_init_hw(struct sh7786_pcie_port *port)
+static int __init sh7786_pcie_init_hw(struct sh7786_pcie_port *port)
 {
 	int ret;
 
-	ret = phy_init(port->hose);
-	if (unlikely(ret < 0))
-		return ret;
-
 	/*
 	 * Check if we are configured in endpoint or root complex mode,
 	 * this is a fixed pin setting that applies to all PCIe ports.
 	 */
 	port->endpoint = test_mode_pin(MODE_PIN11);
 
+	/*
+	 * Setup clocks, needed both for PHY and PCIe registers.
+	 */
+	ret = pcie_clk_init(port);
+	if (unlikely(ret < 0))
+		return ret;
+
+	ret = phy_init(port);
+	if (unlikely(ret < 0))
+		return ret;
+
 	ret = pcie_init(port);
 	if (unlikely(ret < 0))
 		return ret;
@@ -390,9 +509,10 @@ static struct sh7786_pcie_hwops sh7786_65nm_pcie_hwops __initdata = {
 
 static int __init sh7786_pcie_init(void)
 {
+	struct clk *platclk;
 	int ret = 0, i;
 
-	printk(KERN_NOTICE "PCI: Starting intialization.\n");
+	printk(KERN_NOTICE "PCI: Starting initialization.\n");
 
 	sh7786_pcie_hwops = &sh7786_65nm_pcie_hwops;
 
@@ -407,6 +527,22 @@ static int __init sh7786_pcie_init(void)
 	if (unlikely(!sh7786_pcie_ports))
 		return -ENOMEM;
 
+	/*
+	 * Fetch any optional platform clock associated with this block.
+	 *
+	 * This is a rather nasty hack for boards with spec-mocking FPGAs
+	 * that have a secondary set of clocks outside of the on-chip
+	 * ones that need to be accounted for before there is any chance
+	 * of touching the existing MSTP bits or CPG clocks.
+	 */
+	platclk = clk_get(NULL, "pcie_plat_clk");
+	if (IS_ERR(platclk)) {
+		/* Sane hardware should probably get a WARN_ON.. */
+		platclk = NULL;
+	}
+
+	clk_enable(platclk);
+
 	printk(KERN_NOTICE "PCI: probing %d ports.\n", nr_ports);
 
 	for (i = 0; i < nr_ports; i++) {
@@ -419,8 +555,11 @@ static int __init sh7786_pcie_init(void)
 		ret |= sh7786_pcie_hwops->port_init_hw(port);
 	}
 
-	if (unlikely(ret))
+	if (unlikely(ret)) {
+		clk_disable(platclk);
+		clk_put(platclk);
 		return ret;
+	}
 
 	return 0;
 }