From fa3959f457109cc7d082b86ea6daae927982815b Mon Sep 17 00:00:00 2001 From: Lennert Buytenhek Date: Thu, 24 Apr 2008 01:27:02 +0200 Subject: [PATCH 01/32] mv643xx_eth: get rid of static variables, allow multiple instances Move mv643xx_eth's static state (ethernet register block base address and MII management interface spinlock) into a struct hanging off the shared platform device. This is necessary to support chips that contain multiple mv643xx_eth silicon blocks. Signed-off-by: Lennert Buytenhek Acked-by: Nicolas Pitre Signed-off-by: Dale Farnsworth --- arch/arm/mach-orion5x/common.c | 2 + arch/powerpc/platforms/chrp/pegasos_eth.c | 4 ++ arch/powerpc/sysdev/mv64x60_dev.c | 2 + arch/ppc/syslib/mv64x60.c | 3 ++ drivers/net/mv643xx_eth.c | 62 ++++++++++++++++------- include/linux/mv643xx_eth.h | 3 ++ 6 files changed, 59 insertions(+), 17 deletions(-) diff --git a/arch/arm/mach-orion5x/common.c b/arch/arm/mach-orion5x/common.c index 439c7784af02..2bedf71659cb 100644 --- a/arch/arm/mach-orion5x/common.c +++ b/arch/arm/mach-orion5x/common.c @@ -223,7 +223,9 @@ static struct platform_device orion5x_eth = { void __init orion5x_eth_init(struct mv643xx_eth_platform_data *eth_data) { + eth_data->shared = &orion5x_eth_shared; orion5x_eth.dev.platform_data = eth_data; + platform_device_register(&orion5x_eth_shared); platform_device_register(&orion5x_eth); } diff --git a/arch/powerpc/platforms/chrp/pegasos_eth.c b/arch/powerpc/platforms/chrp/pegasos_eth.c index 5bcc58d9a4dd..130ff72d99dd 100644 --- a/arch/powerpc/platforms/chrp/pegasos_eth.c +++ b/arch/powerpc/platforms/chrp/pegasos_eth.c @@ -58,7 +58,9 @@ static struct resource mv643xx_eth0_resources[] = { static struct mv643xx_eth_platform_data eth0_pd = { + .shared = &mv643xx_eth_shared_device, .port_number = 0, + .tx_sram_addr = PEGASOS2_SRAM_BASE_ETH0, .tx_sram_size = PEGASOS2_SRAM_TXRING_SIZE, .tx_queue_size = PEGASOS2_SRAM_TXRING_SIZE/16, @@ -88,7 +90,9 @@ static struct resource mv643xx_eth1_resources[] = { }; static struct mv643xx_eth_platform_data eth1_pd = { + .shared = &mv643xx_eth_shared_device, .port_number = 1, + .tx_sram_addr = PEGASOS2_SRAM_BASE_ETH1, .tx_sram_size = PEGASOS2_SRAM_TXRING_SIZE, .tx_queue_size = PEGASOS2_SRAM_TXRING_SIZE/16, diff --git a/arch/powerpc/sysdev/mv64x60_dev.c b/arch/powerpc/sysdev/mv64x60_dev.c index 41af1223e2a0..a132e0de8ca5 100644 --- a/arch/powerpc/sysdev/mv64x60_dev.c +++ b/arch/powerpc/sysdev/mv64x60_dev.c @@ -239,6 +239,8 @@ static int __init mv64x60_eth_device_setup(struct device_node *np, int id, memset(&pdata, 0, sizeof(pdata)); + pdata.shared = shared_pdev; + prop = of_get_property(np, "reg", NULL); if (!prop) return -ENODEV; diff --git a/arch/ppc/syslib/mv64x60.c b/arch/ppc/syslib/mv64x60.c index 90fe904d3614..418f3053de52 100644 --- a/arch/ppc/syslib/mv64x60.c +++ b/arch/ppc/syslib/mv64x60.c @@ -341,6 +341,7 @@ static struct resource mv64x60_eth0_resources[] = { }; static struct mv643xx_eth_platform_data eth0_pd = { + .shared = &mv64x60_eth_shared_device; .port_number = 0, }; @@ -366,6 +367,7 @@ static struct resource mv64x60_eth1_resources[] = { }; static struct mv643xx_eth_platform_data eth1_pd = { + .shared = &mv64x60_eth_shared_device; .port_number = 1, }; @@ -391,6 +393,7 @@ static struct resource mv64x60_eth2_resources[] = { }; static struct mv643xx_eth_platform_data eth2_pd = { + .shared = &mv64x60_eth_shared_device; .port_number = 2, }; diff --git a/drivers/net/mv643xx_eth.c b/drivers/net/mv643xx_eth.c index 381b36e5f64c..eebf0d288e36 100644 --- a/drivers/net/mv643xx_eth.c +++ b/drivers/net/mv643xx_eth.c @@ -507,7 +507,15 @@ struct mv643xx_mib_counters { u32 late_collision; }; +struct mv643xx_shared_private { + void __iomem *eth_base; + + /* used to protect SMI_REG, which is shared across ports */ + spinlock_t phy_lock; +}; + struct mv643xx_private { + struct mv643xx_shared_private *shared; int port_num; /* User Ethernet port number */ u32 rx_sram_addr; /* Base address of rx sram area */ @@ -614,19 +622,14 @@ static const struct ethtool_ops mv643xx_ethtool_ops; static char mv643xx_driver_name[] = "mv643xx_eth"; static char mv643xx_driver_version[] = "1.0"; -static void __iomem *mv643xx_eth_base; - -/* used to protect SMI_REG, which is shared across ports */ -static DEFINE_SPINLOCK(mv643xx_eth_phy_lock); - static inline u32 rdl(struct mv643xx_private *mp, int offset) { - return readl(mv643xx_eth_base + offset); + return readl(mp->shared->eth_base + offset); } static inline void wrl(struct mv643xx_private *mp, int offset, u32 data) { - writel(data, mv643xx_eth_base + offset); + writel(data, mp->shared->eth_base + offset); } /* @@ -1827,6 +1830,11 @@ static int mv643xx_eth_probe(struct platform_device *pdev) return -ENODEV; } + if (pd->shared == NULL) { + printk(KERN_ERR "No mv643xx_eth_platform_data->shared\n"); + return -ENODEV; + } + dev = alloc_etherdev(sizeof(struct mv643xx_private)); if (!dev) return -ENOMEM; @@ -1877,6 +1885,7 @@ static int mv643xx_eth_probe(struct platform_device *pdev) spin_lock_init(&mp->lock); + mp->shared = platform_get_drvdata(pd->shared); port_num = mp->port_num = pd->port_number; /* set default config values */ @@ -1986,27 +1995,46 @@ static int mv643xx_eth_remove(struct platform_device *pdev) static int mv643xx_eth_shared_probe(struct platform_device *pdev) { static int mv643xx_version_printed = 0; + struct mv643xx_shared_private *msp; struct resource *res; + int ret; if (!mv643xx_version_printed++) printk(KERN_NOTICE "MV-643xx 10/100/1000 Ethernet Driver\n"); + ret = -EINVAL; res = platform_get_resource(pdev, IORESOURCE_MEM, 0); if (res == NULL) - return -ENODEV; + goto out; - mv643xx_eth_base = ioremap(res->start, res->end - res->start + 1); - if (mv643xx_eth_base == NULL) - return -ENOMEM; + ret = -ENOMEM; + msp = kmalloc(sizeof(*msp), GFP_KERNEL); + if (msp == NULL) + goto out; + memset(msp, 0, sizeof(*msp)); + + msp->eth_base = ioremap(res->start, res->end - res->start + 1); + if (msp->eth_base == NULL) + goto out_free; + + spin_lock_init(&msp->phy_lock); + + platform_set_drvdata(pdev, msp); return 0; +out_free: + kfree(msp); +out: + return ret; } static int mv643xx_eth_shared_remove(struct platform_device *pdev) { - iounmap(mv643xx_eth_base); - mv643xx_eth_base = NULL; + struct mv643xx_shared_private *msp = platform_get_drvdata(pdev); + + iounmap(msp->eth_base); + kfree(msp); return 0; } @@ -2911,7 +2939,7 @@ static void eth_port_read_smi_reg(struct mv643xx_private *mp, int i; /* the SMI register is a shared resource */ - spin_lock_irqsave(&mv643xx_eth_phy_lock, flags); + spin_lock_irqsave(&mp->shared->phy_lock, flags); /* wait for the SMI register to become available */ for (i = 0; rdl(mp, SMI_REG) & ETH_SMI_BUSY; i++) { @@ -2936,7 +2964,7 @@ static void eth_port_read_smi_reg(struct mv643xx_private *mp, *value = rdl(mp, SMI_REG) & 0xffff; out: - spin_unlock_irqrestore(&mv643xx_eth_phy_lock, flags); + spin_unlock_irqrestore(&mp->shared->phy_lock, flags); } /* @@ -2969,7 +2997,7 @@ static void eth_port_write_smi_reg(struct mv643xx_private *mp, phy_addr = ethernet_phy_get(mp); /* the SMI register is a shared resource */ - spin_lock_irqsave(&mv643xx_eth_phy_lock, flags); + spin_lock_irqsave(&mp->shared->phy_lock, flags); /* wait for the SMI register to become available */ for (i = 0; rdl(mp, SMI_REG) & ETH_SMI_BUSY; i++) { @@ -2983,7 +3011,7 @@ static void eth_port_write_smi_reg(struct mv643xx_private *mp, wrl(mp, SMI_REG, (phy_addr << 16) | (phy_reg << 21) | ETH_SMI_OPCODE_WRITE | (value & 0xffff)); out: - spin_unlock_irqrestore(&mv643xx_eth_phy_lock, flags); + spin_unlock_irqrestore(&mp->shared->phy_lock, flags); } /* diff --git a/include/linux/mv643xx_eth.h b/include/linux/mv643xx_eth.h index 30e11aa3c1c9..2d59855b61c1 100644 --- a/include/linux/mv643xx_eth.h +++ b/include/linux/mv643xx_eth.h @@ -1,6 +1,7 @@ /* * MV-643XX ethernet platform device data definition file. */ + #ifndef __LINUX_MV643XX_ETH_H #define __LINUX_MV643XX_ETH_H @@ -13,7 +14,9 @@ #define MV643XX_ETH_BASE_ADDR_ENABLE_REG 0x2290 struct mv643xx_eth_platform_data { + struct platform_device *shared; int port_number; + u16 force_phy_addr; /* force override if phy_addr == 0 */ u16 phy_addr; From f2ce825d2a89b30af14fa577298fecaab7bc9504 Mon Sep 17 00:00:00 2001 From: Lennert Buytenhek Date: Thu, 24 Apr 2008 01:27:17 +0200 Subject: [PATCH 02/32] mv643xx_eth: mbus decode window support Make it possible to pass mbus_dram_target_info to the mv643xx_eth driver via the platform data, and make the mv643xx_eth driver program the window registers based on this data if it is passed in. Signed-off-by: Lennert Buytenhek Reviewed-by: Tzachi Perelstein Acked-by: Russell King Signed-off-by: Dale Farnsworth --- drivers/net/mv643xx_eth.c | 51 +++++++++++++++++++++++++++++++++++++ include/linux/mv643xx_eth.h | 6 +++++ 2 files changed, 57 insertions(+) diff --git a/drivers/net/mv643xx_eth.c b/drivers/net/mv643xx_eth.c index eebf0d288e36..aabf1c60946d 100644 --- a/drivers/net/mv643xx_eth.c +++ b/drivers/net/mv643xx_eth.c @@ -91,6 +91,11 @@ */ #define PHY_ADDR_REG 0x0000 #define SMI_REG 0x0004 +#define WINDOW_BASE(i) (0x0200 + ((i) << 3)) +#define WINDOW_SIZE(i) (0x0204 + ((i) << 3)) +#define WINDOW_REMAP_HIGH(i) (0x0280 + ((i) << 2)) +#define WINDOW_BAR_ENABLE 0x0290 +#define WINDOW_PROTECT(i) (0x0294 + ((i) << 4)) /* * Per-port registers. @@ -512,6 +517,8 @@ struct mv643xx_shared_private { /* used to protect SMI_REG, which is shared across ports */ spinlock_t phy_lock; + + u32 win_protect; }; struct mv643xx_private { @@ -1888,6 +1895,9 @@ static int mv643xx_eth_probe(struct platform_device *pdev) mp->shared = platform_get_drvdata(pd->shared); port_num = mp->port_num = pd->port_number; + if (mp->shared->win_protect) + wrl(mp, WINDOW_PROTECT(port_num), mp->shared->win_protect); + /* set default config values */ eth_port_uc_addr_get(mp, dev->dev_addr); mp->rx_ring_size = PORT_DEFAULT_RECEIVE_QUEUE_SIZE; @@ -1992,9 +2002,44 @@ static int mv643xx_eth_remove(struct platform_device *pdev) return 0; } +static void mv643xx_eth_conf_mbus_windows(struct mv643xx_shared_private *msp, + struct mbus_dram_target_info *dram) +{ + void __iomem *base = msp->eth_base; + u32 win_enable; + u32 win_protect; + int i; + + for (i = 0; i < 6; i++) { + writel(0, base + WINDOW_BASE(i)); + writel(0, base + WINDOW_SIZE(i)); + if (i < 4) + writel(0, base + WINDOW_REMAP_HIGH(i)); + } + + win_enable = 0x3f; + win_protect = 0; + + for (i = 0; i < dram->num_cs; i++) { + struct mbus_dram_window *cs = dram->cs + i; + + writel((cs->base & 0xffff0000) | + (cs->mbus_attr << 8) | + dram->mbus_dram_target_id, base + WINDOW_BASE(i)); + writel((cs->size - 1) & 0xffff0000, base + WINDOW_SIZE(i)); + + win_enable &= ~(1 << i); + win_protect |= 3 << (2 * i); + } + + writel(win_enable, base + WINDOW_BAR_ENABLE); + msp->win_protect = win_protect; +} + static int mv643xx_eth_shared_probe(struct platform_device *pdev) { static int mv643xx_version_printed = 0; + struct mv643xx_eth_shared_platform_data *pd = pdev->dev.platform_data; struct mv643xx_shared_private *msp; struct resource *res; int ret; @@ -2021,6 +2066,12 @@ static int mv643xx_eth_shared_probe(struct platform_device *pdev) platform_set_drvdata(pdev, msp); + /* + * (Re-)program MBUS remapping windows if we are asked to. + */ + if (pd != NULL && pd->dram != NULL) + mv643xx_eth_conf_mbus_windows(msp, pd->dram); + return 0; out_free: diff --git a/include/linux/mv643xx_eth.h b/include/linux/mv643xx_eth.h index 2d59855b61c1..4801b02b444e 100644 --- a/include/linux/mv643xx_eth.h +++ b/include/linux/mv643xx_eth.h @@ -5,6 +5,8 @@ #ifndef __LINUX_MV643XX_ETH_H #define __LINUX_MV643XX_ETH_H +#include + #define MV643XX_ETH_SHARED_NAME "mv643xx_eth_shared" #define MV643XX_ETH_NAME "mv643xx_eth" #define MV643XX_ETH_SHARED_REGS 0x2000 @@ -13,6 +15,10 @@ #define MV643XX_ETH_SIZE_REG_4 0x2224 #define MV643XX_ETH_BASE_ADDR_ENABLE_REG 0x2290 +struct mv643xx_eth_shared_platform_data { + struct mbus_dram_target_info *dram; +}; + struct mv643xx_eth_platform_data { struct platform_device *shared; int port_number; From c416a41f99be190e1f558cb06f70ddd560ce8b4b Mon Sep 17 00:00:00 2001 From: Lennert Buytenhek Date: Thu, 24 Apr 2008 01:27:32 +0200 Subject: [PATCH 03/32] mv643xx_eth: configurable t_clk Make t_clk configurable via platform device data (with the current hardcoded value, 133 MHz, being the default), as it varies across different chip families. Signed-off-by: Lennert Buytenhek Acked-by: Nicolas Pitre Signed-off-by: Dale Farnsworth --- drivers/net/mv643xx_eth.c | 17 +++++++++-------- include/linux/mv643xx_eth.h | 1 + 2 files changed, 10 insertions(+), 8 deletions(-) diff --git a/drivers/net/mv643xx_eth.c b/drivers/net/mv643xx_eth.c index aabf1c60946d..8bd41e4e88a9 100644 --- a/drivers/net/mv643xx_eth.c +++ b/drivers/net/mv643xx_eth.c @@ -519,6 +519,8 @@ struct mv643xx_shared_private { spinlock_t phy_lock; u32 win_protect; + + unsigned int t_clk; }; struct mv643xx_private { @@ -1129,7 +1131,6 @@ static irqreturn_t mv643xx_eth_int_handler(int irq, void *dev_id) * * INPUT: * struct mv643xx_private *mp Ethernet port - * unsigned int t_clk t_clk of the MV-643xx chip in HZ units * unsigned int delay Delay in usec * * OUTPUT: @@ -1140,10 +1141,10 @@ static irqreturn_t mv643xx_eth_int_handler(int irq, void *dev_id) * */ static unsigned int eth_port_set_rx_coal(struct mv643xx_private *mp, - unsigned int t_clk, unsigned int delay) + unsigned int delay) { unsigned int port_num = mp->port_num; - unsigned int coal = ((t_clk / 1000000) * delay) / 64; + unsigned int coal = ((mp->shared->t_clk / 1000000) * delay) / 64; /* Set RX Coalescing mechanism */ wrl(mp, SDMA_CONFIG_REG(port_num), @@ -1168,7 +1169,6 @@ static unsigned int eth_port_set_rx_coal(struct mv643xx_private *mp, * * INPUT: * struct mv643xx_private *mp Ethernet port - * unsigned int t_clk t_clk of the MV-643xx chip in HZ units * unsigned int delay Delay in uSeconds * * OUTPUT: @@ -1179,9 +1179,9 @@ static unsigned int eth_port_set_rx_coal(struct mv643xx_private *mp, * */ static unsigned int eth_port_set_tx_coal(struct mv643xx_private *mp, - unsigned int t_clk, unsigned int delay) + unsigned int delay) { - unsigned int coal = ((t_clk / 1000000) * delay) / 64; + unsigned int coal = ((mp->shared->t_clk / 1000000) * delay) / 64; /* Set TX Coalescing mechanism */ wrl(mp, TX_FIFO_URGENT_THRESHOLD_REG(mp->port_num), coal << 4); @@ -1423,11 +1423,11 @@ static int mv643xx_eth_open(struct net_device *dev) #ifdef MV643XX_COAL mp->rx_int_coal = - eth_port_set_rx_coal(mp, 133000000, MV643XX_RX_COAL); + eth_port_set_rx_coal(mp, MV643XX_RX_COAL); #endif mp->tx_int_coal = - eth_port_set_tx_coal(mp, 133000000, MV643XX_TX_COAL); + eth_port_set_tx_coal(mp, MV643XX_TX_COAL); /* Unmask phy and link status changes interrupts */ wrl(mp, INTERRUPT_EXTEND_MASK_REG(port_num), ETH_INT_UNMASK_ALL_EXT); @@ -2063,6 +2063,7 @@ static int mv643xx_eth_shared_probe(struct platform_device *pdev) goto out_free; spin_lock_init(&msp->phy_lock); + msp->t_clk = (pd != NULL && pd->t_clk != 0) ? pd->t_clk : 133000000; platform_set_drvdata(pdev, msp); diff --git a/include/linux/mv643xx_eth.h b/include/linux/mv643xx_eth.h index 4801b02b444e..9f3a6032ff2e 100644 --- a/include/linux/mv643xx_eth.h +++ b/include/linux/mv643xx_eth.h @@ -17,6 +17,7 @@ struct mv643xx_eth_shared_platform_data { struct mbus_dram_target_info *dram; + unsigned int t_clk; }; struct mv643xx_eth_platform_data { From 240e4419e0cfcba737883b637ec2bdcc071ea03d Mon Sep 17 00:00:00 2001 From: Lennert Buytenhek Date: Thu, 24 Apr 2008 01:27:44 +0200 Subject: [PATCH 04/32] mv643xx_eth: shorten shared platform driver name Change the MV643XX_ETH_SHARED_NAME platform driver name to something shorter than 19 characters, so that we can register multiple (otherwise we end up with sysfs conflicts since all instances will map to "mv643xx_eth_shared." as there is a 20-char sysfs file name limit.) Signed-off-by: Lennert Buytenhek Acked-by: Nicolas Pitre Signed-off-by: Dale Farnsworth --- include/linux/mv643xx_eth.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/include/linux/mv643xx_eth.h b/include/linux/mv643xx_eth.h index 9f3a6032ff2e..66dc9571922a 100644 --- a/include/linux/mv643xx_eth.h +++ b/include/linux/mv643xx_eth.h @@ -7,8 +7,8 @@ #include -#define MV643XX_ETH_SHARED_NAME "mv643xx_eth_shared" -#define MV643XX_ETH_NAME "mv643xx_eth" +#define MV643XX_ETH_SHARED_NAME "mv643xx_eth" +#define MV643XX_ETH_NAME "mv643xx_eth_port" #define MV643XX_ETH_SHARED_REGS 0x2000 #define MV643XX_ETH_SHARED_REGS_SIZE 0x2000 #define MV643XX_ETH_BAR_4 0x2220 From ce4e2e4558903ef92edf1ab4e09b0b338a09fd61 Mon Sep 17 00:00:00 2001 From: Lennert Buytenhek Date: Thu, 24 Apr 2008 01:29:59 +0200 Subject: [PATCH 05/32] mv643xx_eth: inter-mv643xx SMI port sharing There exist chips with up to four mv643xx_eth silicon blocks but only one external SMI (MII management) interface -- the SMI logic of the first block is shared by all the blocks. Handle this by allowing a per-port override of which mv643xx_eth_shared's SMI registers (and spinlock) to use. Signed-off-by: Lennert Buytenhek Acked-by: Nicolas Pitre Signed-off-by: Dale Farnsworth --- drivers/net/mv643xx_eth.c | 38 +++++++++++++++++++++---------------- include/linux/mv643xx_eth.h | 2 ++ 2 files changed, 24 insertions(+), 16 deletions(-) diff --git a/drivers/net/mv643xx_eth.c b/drivers/net/mv643xx_eth.c index 8bd41e4e88a9..b7915cdcc6a5 100644 --- a/drivers/net/mv643xx_eth.c +++ b/drivers/net/mv643xx_eth.c @@ -527,6 +527,8 @@ struct mv643xx_private { struct mv643xx_shared_private *shared; int port_num; /* User Ethernet port number */ + struct mv643xx_shared_private *shared_smi; + u32 rx_sram_addr; /* Base address of rx sram area */ u32 rx_sram_size; /* Size of rx sram area */ u32 tx_sram_addr; /* Base address of tx sram area */ @@ -1898,6 +1900,10 @@ static int mv643xx_eth_probe(struct platform_device *pdev) if (mp->shared->win_protect) wrl(mp, WINDOW_PROTECT(port_num), mp->shared->win_protect); + mp->shared_smi = mp->shared; + if (pd->shared_smi != NULL) + mp->shared_smi = platform_get_drvdata(pd->shared_smi); + /* set default config values */ eth_port_uc_addr_get(mp, dev->dev_addr); mp->rx_ring_size = PORT_DEFAULT_RECEIVE_QUEUE_SIZE; @@ -2986,15 +2992,16 @@ static void eth_port_reset(struct mv643xx_private *mp) static void eth_port_read_smi_reg(struct mv643xx_private *mp, unsigned int phy_reg, unsigned int *value) { + void __iomem *smi_reg = mp->shared_smi->eth_base + SMI_REG; int phy_addr = ethernet_phy_get(mp); unsigned long flags; int i; /* the SMI register is a shared resource */ - spin_lock_irqsave(&mp->shared->phy_lock, flags); + spin_lock_irqsave(&mp->shared_smi->phy_lock, flags); /* wait for the SMI register to become available */ - for (i = 0; rdl(mp, SMI_REG) & ETH_SMI_BUSY; i++) { + for (i = 0; readl(smi_reg) & ETH_SMI_BUSY; i++) { if (i == PHY_WAIT_ITERATIONS) { printk("%s: PHY busy timeout\n", mp->dev->name); goto out; @@ -3002,11 +3009,11 @@ static void eth_port_read_smi_reg(struct mv643xx_private *mp, udelay(PHY_WAIT_MICRO_SECONDS); } - wrl(mp, SMI_REG, - (phy_addr << 16) | (phy_reg << 21) | ETH_SMI_OPCODE_READ); + writel((phy_addr << 16) | (phy_reg << 21) | ETH_SMI_OPCODE_READ, + smi_reg); /* now wait for the data to be valid */ - for (i = 0; !(rdl(mp, SMI_REG) & ETH_SMI_READ_VALID); i++) { + for (i = 0; !(readl(smi_reg) & ETH_SMI_READ_VALID); i++) { if (i == PHY_WAIT_ITERATIONS) { printk("%s: PHY read timeout\n", mp->dev->name); goto out; @@ -3014,9 +3021,9 @@ static void eth_port_read_smi_reg(struct mv643xx_private *mp, udelay(PHY_WAIT_MICRO_SECONDS); } - *value = rdl(mp, SMI_REG) & 0xffff; + *value = readl(smi_reg) & 0xffff; out: - spin_unlock_irqrestore(&mp->shared->phy_lock, flags); + spin_unlock_irqrestore(&mp->shared_smi->phy_lock, flags); } /* @@ -3042,17 +3049,16 @@ out: static void eth_port_write_smi_reg(struct mv643xx_private *mp, unsigned int phy_reg, unsigned int value) { - int phy_addr; - int i; + void __iomem *smi_reg = mp->shared_smi->eth_base + SMI_REG; + int phy_addr = ethernet_phy_get(mp); unsigned long flags; - - phy_addr = ethernet_phy_get(mp); + int i; /* the SMI register is a shared resource */ - spin_lock_irqsave(&mp->shared->phy_lock, flags); + spin_lock_irqsave(&mp->shared_smi->phy_lock, flags); /* wait for the SMI register to become available */ - for (i = 0; rdl(mp, SMI_REG) & ETH_SMI_BUSY; i++) { + for (i = 0; readl(smi_reg) & ETH_SMI_BUSY; i++) { if (i == PHY_WAIT_ITERATIONS) { printk("%s: PHY busy timeout\n", mp->dev->name); goto out; @@ -3060,10 +3066,10 @@ static void eth_port_write_smi_reg(struct mv643xx_private *mp, udelay(PHY_WAIT_MICRO_SECONDS); } - wrl(mp, SMI_REG, (phy_addr << 16) | (phy_reg << 21) | - ETH_SMI_OPCODE_WRITE | (value & 0xffff)); + writel((phy_addr << 16) | (phy_reg << 21) | + ETH_SMI_OPCODE_WRITE | (value & 0xffff), smi_reg); out: - spin_unlock_irqrestore(&mp->shared->phy_lock, flags); + spin_unlock_irqrestore(&mp->shared_smi->phy_lock, flags); } /* diff --git a/include/linux/mv643xx_eth.h b/include/linux/mv643xx_eth.h index 66dc9571922a..a15cdd4a8e58 100644 --- a/include/linux/mv643xx_eth.h +++ b/include/linux/mv643xx_eth.h @@ -24,6 +24,8 @@ struct mv643xx_eth_platform_data { struct platform_device *shared; int port_number; + struct platform_device *shared_smi; + u16 force_phy_addr; /* force override if phy_addr == 0 */ u16 phy_addr; From 822973ba79fd5a5b711270c2de7196c6b50c6687 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 2 May 2008 17:49:37 -0700 Subject: [PATCH 06/32] bonding: Do not call free_netdev for already registered device. If the call to bond_create_sysfs_entry in bond_create fails, the proper rollback is to call unregister_netdevice, not free_netdev. Otherwise - kernel BUG at net/core/dev.c:4057! Checked with artificial failures injected into bond_create_sysfs_entry. Pavel's original patch modified by Jay Vosburgh to move code around for clarity (remove goto-hopping within the unwind block). Signed-off-by: Pavel Emelyanov Signed-off-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- drivers/net/bonding/bond_main.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 6425603bc379..5509732d3f9d 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -4936,7 +4936,9 @@ int bond_create(char *name, struct bond_params *params, struct bonding **newbond if (res < 0) { rtnl_lock(); down_write(&bonding_rwsem); - goto out_bond; + bond_deinit(bond_dev); + unregister_netdevice(bond_dev); + goto out_rtnl; } return 0; From c4ebc66a1a8e3576322a9f47f0d06ec3c96a08d7 Mon Sep 17 00:00:00 2001 From: Jay Vosburgh Date: Fri, 2 May 2008 17:49:38 -0700 Subject: [PATCH 07/32] bonding: fix error unwind in bonding_store_bonds Fixed an error unwind in bonding_store_bonds that didn't release the locks it held, and consolidated unwinds into a common block at the end of the function. Bug reported by Pavel Emelyanov , who provided a different fix. Signed-off-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- drivers/net/bonding/bond_sysfs.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c index 979c2d05ff9c..68c41a00d93d 100644 --- a/drivers/net/bonding/bond_sysfs.c +++ b/drivers/net/bonding/bond_sysfs.c @@ -146,29 +146,29 @@ static ssize_t bonding_store_bonds(struct class *cls, const char *buffer, size_t ": Unable remove bond %s due to open references.\n", ifname); res = -EPERM; - goto out; + goto out_unlock; } printk(KERN_INFO DRV_NAME ": %s is being deleted...\n", bond->dev->name); bond_destroy(bond); - up_write(&bonding_rwsem); - rtnl_unlock(); - goto out; + goto out_unlock; } printk(KERN_ERR DRV_NAME ": unable to delete non-existent bond %s\n", ifname); res = -ENODEV; - up_write(&bonding_rwsem); - rtnl_unlock(); - goto out; + goto out_unlock; } err_no_cmd: printk(KERN_ERR DRV_NAME ": no command found in bonding_masters. Use +ifname or -ifname.\n"); - res = -EPERM; + return -EPERM; + +out_unlock: + up_write(&bonding_rwsem); + rtnl_unlock(); /* Always return either count or an error. If you return 0, you'll * get called forever, which is bad. From ae68c39819ddf30549652962768a50edae5eec6f Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 2 May 2008 17:49:39 -0700 Subject: [PATCH 08/32] bonding: Deadlock between bonding_store_bonds and bond_destroy_sysfs. The sysfs layer has an internal protection, that ensures, that all the process sitting inside ->sore/->show callback exits before the appropriate entry is unregistered (the calltraces are rather big, but I can provide them if required). On the other hand, bonding takes rtnl_lock in a) the bonding_store_bonds, i.e. in ->store callback, b) module exit before calling the sysfs unregister routines. Thus, the classical AB-BA deadlock may occur. To reproduce run # while :; do modprobe bonding; rmmod bonding; done and # while :; do echo '+bond%d' > /sys/class/net/bonding_masters ; done in parallel. The fix is to move the bond_destroy_sysfs out of the rtnl_lock, but _before_ bond_free_all to make sure no bonding devices exist after module unload. Signed-off-by: Pavel Emelyanov Acked-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- drivers/net/bonding/bond_main.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 5509732d3f9d..e41b3e57260c 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -4992,9 +4992,10 @@ err: destroy_workqueue(bond->wq); } + bond_destroy_sysfs(); + rtnl_lock(); bond_free_all(); - bond_destroy_sysfs(); rtnl_unlock(); out: return res; @@ -5006,9 +5007,10 @@ static void __exit bonding_exit(void) unregister_netdevice_notifier(&bond_netdev_notifier); unregister_inetaddr_notifier(&bond_inetaddr_notifier); + bond_destroy_sysfs(); + rtnl_lock(); bond_free_all(); - bond_destroy_sysfs(); rtnl_unlock(); } From 569f0c4d909c7f73de634abcdc36344cb72de36a Mon Sep 17 00:00:00 2001 From: Jay Vosburgh Date: Fri, 2 May 2008 18:06:02 -0700 Subject: [PATCH 09/32] bonding: fix enslavement error unwinds As part of: commit c2edacf80e155ef54ae4774379d461b60896bc2e Author: Jay Vosburgh Date: Mon Jul 9 10:42:47 2007 -0700 bonding / ipv6: no addrconf for slaves separately from master two steps were rearranged in the enslavement process: netdev_set_master is now before the call to dev_open to open the slave. This patch updates the error cases and unwind process at the end of bond_enslave to match the new order. Without this patch, it is possible for the enslavement to fail, but leave the slave with IFF_SLAVE set in its flags. Signed-off-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- drivers/net/bonding/bond_main.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index e41b3e57260c..50a40e433154 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -1425,13 +1425,13 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev) res = netdev_set_master(slave_dev, bond_dev); if (res) { dprintk("Error %d calling netdev_set_master\n", res); - goto err_close; + goto err_restore_mac; } /* open the slave since the application closed it */ res = dev_open(slave_dev); if (res) { dprintk("Openning slave %s failed\n", slave_dev->name); - goto err_restore_mac; + goto err_unset_master; } new_slave->dev = slave_dev; @@ -1444,7 +1444,7 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev) */ res = bond_alb_init_slave(bond, new_slave); if (res) { - goto err_unset_master; + goto err_close; } } @@ -1619,7 +1619,7 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev) res = bond_create_slave_symlinks(bond_dev, slave_dev); if (res) - goto err_unset_master; + goto err_close; printk(KERN_INFO DRV_NAME ": %s: enslaving %s as a%s interface with a%s link.\n", @@ -1631,12 +1631,12 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev) return 0; /* Undo stages on error */ -err_unset_master: - netdev_set_master(slave_dev, NULL); - err_close: dev_close(slave_dev); +err_unset_master: + netdev_set_master(slave_dev, NULL); + err_restore_mac: if (!bond->params.fail_over_mac) { memcpy(addr.sa_data, new_slave->perm_hwaddr, ETH_ALEN); From f162b9d58273a9a5747211133c8ccb2de5cf5ff2 Mon Sep 17 00:00:00 2001 From: Andy Fleming Date: Fri, 2 May 2008 13:00:30 -0500 Subject: [PATCH 10/32] gianfar: Fix a locking bug in gianfar's sysfs code During sparse cleanup, found a locking bug. Some of the sysfs functions were acquiring a lock, and then returning in the event of an error. We rearrange the code so that the lock is released in error conditions, too. Signed-off-by: Andy Fleming Signed-off-by: Jeff Garzik --- drivers/net/gianfar.c | 5 +++-- drivers/net/gianfar.h | 3 +++ drivers/net/gianfar_sysfs.c | 10 ++++++---- 3 files changed, 12 insertions(+), 6 deletions(-) diff --git a/drivers/net/gianfar.c b/drivers/net/gianfar.c index 587afe7be689..6f22f068d6ee 100644 --- a/drivers/net/gianfar.c +++ b/drivers/net/gianfar.c @@ -138,6 +138,7 @@ static int gfar_poll(struct napi_struct *napi, int budget); static void gfar_netpoll(struct net_device *dev); #endif int gfar_clean_rx_ring(struct net_device *dev, int rx_work_limit); +static int gfar_clean_tx_ring(struct net_device *dev); static int gfar_process_frame(struct net_device *dev, struct sk_buff *skb, int length); static void gfar_vlan_rx_register(struct net_device *netdev, struct vlan_group *grp); @@ -1141,7 +1142,7 @@ static int gfar_close(struct net_device *dev) } /* Changes the mac address if the controller is not running. */ -int gfar_set_mac_address(struct net_device *dev) +static int gfar_set_mac_address(struct net_device *dev) { gfar_set_mac_for_addr(dev, 0, dev->dev_addr); @@ -1260,7 +1261,7 @@ static void gfar_timeout(struct net_device *dev) } /* Interrupt Handler for Transmit complete */ -int gfar_clean_tx_ring(struct net_device *dev) +static int gfar_clean_tx_ring(struct net_device *dev) { struct txbd8 *bdp; struct gfar_private *priv = netdev_priv(dev); diff --git a/drivers/net/gianfar.h b/drivers/net/gianfar.h index fd487be3993e..27f37c81e52c 100644 --- a/drivers/net/gianfar.h +++ b/drivers/net/gianfar.h @@ -782,5 +782,8 @@ extern void gfar_halt(struct net_device *dev); extern void gfar_phy_test(struct mii_bus *bus, struct phy_device *phydev, int enable, u32 regnum, u32 read); void gfar_init_sysfs(struct net_device *dev); +int gfar_local_mdio_write(struct gfar_mii __iomem *regs, int mii_id, + int regnum, u16 value); +int gfar_local_mdio_read(struct gfar_mii __iomem *regs, int mii_id, int regnum); #endif /* __GIANFAR_H */ diff --git a/drivers/net/gianfar_sysfs.c b/drivers/net/gianfar_sysfs.c index 230878b94190..5116f68e01b9 100644 --- a/drivers/net/gianfar_sysfs.c +++ b/drivers/net/gianfar_sysfs.c @@ -103,10 +103,10 @@ static ssize_t gfar_set_rx_stash_size(struct device *dev, spin_lock_irqsave(&priv->rxlock, flags); if (length > priv->rx_buffer_size) - return count; + goto out; if (length == priv->rx_stash_size) - return count; + goto out; priv->rx_stash_size = length; @@ -125,6 +125,7 @@ static ssize_t gfar_set_rx_stash_size(struct device *dev, gfar_write(&priv->regs->attr, temp); +out: spin_unlock_irqrestore(&priv->rxlock, flags); return count; @@ -154,10 +155,10 @@ static ssize_t gfar_set_rx_stash_index(struct device *dev, spin_lock_irqsave(&priv->rxlock, flags); if (index > priv->rx_stash_size) - return count; + goto out; if (index == priv->rx_stash_index) - return count; + goto out; priv->rx_stash_index = index; @@ -166,6 +167,7 @@ static ssize_t gfar_set_rx_stash_index(struct device *dev, temp |= ATTRELI_EI(index); gfar_write(&priv->regs->attreli, flags); +out: spin_unlock_irqrestore(&priv->rxlock, flags); return count; From 9b9a8bfc8dfbe09dc57f274e32e8b06151abbad7 Mon Sep 17 00:00:00 2001 From: Andy Fleming Date: Fri, 2 May 2008 13:00:51 -0500 Subject: [PATCH 11/32] phylib: Fix some sparse warnings Declared some things static, declared some things in the header. Signed-off-by: Andy Fleming Signed-off-by: Jeff Garzik --- drivers/net/phy/phy.c | 2 +- include/linux/phy.h | 5 +++++ 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c index 3c18bb594957..45cc2914d347 100644 --- a/drivers/net/phy/phy.c +++ b/drivers/net/phy/phy.c @@ -547,7 +547,7 @@ static void phy_force_reduction(struct phy_device *phydev) * Must not be called from interrupt context, or while the * phydev->lock is held. */ -void phy_error(struct phy_device *phydev) +static void phy_error(struct phy_device *phydev) { mutex_lock(&phydev->lock); phydev->state = PHY_HALTED; diff --git a/include/linux/phy.h b/include/linux/phy.h index 02df20f085fe..7224c4099a28 100644 --- a/include/linux/phy.h +++ b/include/linux/phy.h @@ -412,6 +412,8 @@ int mdiobus_register(struct mii_bus *bus); void mdiobus_unregister(struct mii_bus *bus); void phy_sanitize_settings(struct phy_device *phydev); int phy_stop_interrupts(struct phy_device *phydev); +int phy_enable_interrupts(struct phy_device *phydev); +int phy_disable_interrupts(struct phy_device *phydev); static inline int phy_read_status(struct phy_device *phydev) { return phydev->drv->read_status(phydev); @@ -447,5 +449,8 @@ int phy_register_fixup_for_uid(u32 phy_uid, u32 phy_uid_mask, int (*run)(struct phy_device *)); int phy_scan_fixups(struct phy_device *phydev); +int __init mdio_bus_init(void); +void mdio_bus_exit(void); + extern struct bus_type mdio_bus_type; #endif /* __PHY_H */ From 6fee40e9b8155a4af904d69765c96c00f975acf5 Mon Sep 17 00:00:00 2001 From: Andy Fleming Date: Fri, 2 May 2008 13:01:23 -0500 Subject: [PATCH 12/32] ucc_geth: Fix a bunch of sparse warnings ucc_geth didn't have anything marked as __iomem. It was also inconsistent with its use of in/out accessors (using them sometimes, not using them other times). Cleaning this up cuts the warnings down from hundreds to just over a dozen. Signed-off-by: Andy Fleming Signed-off-by: Jeff Garzik --- drivers/net/ucc_geth.c | 276 +++++++++++++++++---------------- drivers/net/ucc_geth.h | 48 +++--- drivers/net/ucc_geth_ethtool.c | 6 - drivers/net/ucc_geth_mii.c | 4 +- 4 files changed, 174 insertions(+), 160 deletions(-) diff --git a/drivers/net/ucc_geth.c b/drivers/net/ucc_geth.c index 281ce3d39532..b3bbb2dc5cfa 100644 --- a/drivers/net/ucc_geth.c +++ b/drivers/net/ucc_geth.c @@ -62,7 +62,6 @@ #endif /* UGETH_VERBOSE_DEBUG */ #define UGETH_MSG_DEFAULT (NETIF_MSG_IFUP << 1 ) - 1 -void uec_set_ethtool_ops(struct net_device *netdev); static DEFINE_SPINLOCK(ugeth_lock); @@ -216,7 +215,8 @@ static struct list_head *dequeue(struct list_head *lh) } } -static struct sk_buff *get_new_skb(struct ucc_geth_private *ugeth, u8 *bd) +static struct sk_buff *get_new_skb(struct ucc_geth_private *ugeth, + u8 __iomem *bd) { struct sk_buff *skb = NULL; @@ -236,21 +236,22 @@ static struct sk_buff *get_new_skb(struct ucc_geth_private *ugeth, u8 *bd) skb->dev = ugeth->dev; - out_be32(&((struct qe_bd *)bd)->buf, + out_be32(&((struct qe_bd __iomem *)bd)->buf, dma_map_single(NULL, skb->data, ugeth->ug_info->uf_info.max_rx_buf_length + UCC_GETH_RX_DATA_BUF_ALIGNMENT, DMA_FROM_DEVICE)); - out_be32((u32 *)bd, (R_E | R_I | (in_be32((u32 *)bd) & R_W))); + out_be32((u32 __iomem *)bd, + (R_E | R_I | (in_be32((u32 __iomem*)bd) & R_W))); return skb; } static int rx_bd_buffer_set(struct ucc_geth_private *ugeth, u8 rxQ) { - u8 *bd; + u8 __iomem *bd; u32 bd_status; struct sk_buff *skb; int i; @@ -259,7 +260,7 @@ static int rx_bd_buffer_set(struct ucc_geth_private *ugeth, u8 rxQ) i = 0; do { - bd_status = in_be32((u32*)bd); + bd_status = in_be32((u32 __iomem *)bd); skb = get_new_skb(ugeth, bd); if (!skb) /* If can not allocate data buffer, @@ -277,7 +278,7 @@ static int rx_bd_buffer_set(struct ucc_geth_private *ugeth, u8 rxQ) } static int fill_init_enet_entries(struct ucc_geth_private *ugeth, - volatile u32 *p_start, + u32 *p_start, u8 num_entries, u32 thread_size, u32 thread_alignment, @@ -316,7 +317,7 @@ static int fill_init_enet_entries(struct ucc_geth_private *ugeth, } static int return_init_enet_entries(struct ucc_geth_private *ugeth, - volatile u32 *p_start, + u32 *p_start, u8 num_entries, enum qe_risc_allocation risc, int skip_page_for_first_entry) @@ -326,21 +327,22 @@ static int return_init_enet_entries(struct ucc_geth_private *ugeth, int snum; for (i = 0; i < num_entries; i++) { + u32 val = *p_start; + /* Check that this entry was actually valid -- needed in case failed in allocations */ - if ((*p_start & ENET_INIT_PARAM_RISC_MASK) == risc) { + if ((val & ENET_INIT_PARAM_RISC_MASK) == risc) { snum = - (u32) (*p_start & ENET_INIT_PARAM_SNUM_MASK) >> + (u32) (val & ENET_INIT_PARAM_SNUM_MASK) >> ENET_INIT_PARAM_SNUM_SHIFT; qe_put_snum((u8) snum); if (!((i == 0) && skip_page_for_first_entry)) { /* First entry of Rx does not have page */ init_enet_offset = - (in_be32(p_start) & - ENET_INIT_PARAM_PTR_MASK); + (val & ENET_INIT_PARAM_PTR_MASK); qe_muram_free(init_enet_offset); } - *(p_start++) = 0; /* Just for cosmetics */ + *p_start++ = 0; } } @@ -349,7 +351,7 @@ static int return_init_enet_entries(struct ucc_geth_private *ugeth, #ifdef DEBUG static int dump_init_enet_entries(struct ucc_geth_private *ugeth, - volatile u32 *p_start, + u32 __iomem *p_start, u8 num_entries, u32 thread_size, enum qe_risc_allocation risc, @@ -360,11 +362,13 @@ static int dump_init_enet_entries(struct ucc_geth_private *ugeth, int snum; for (i = 0; i < num_entries; i++) { + u32 val = in_be32(p_start); + /* Check that this entry was actually valid -- needed in case failed in allocations */ - if ((*p_start & ENET_INIT_PARAM_RISC_MASK) == risc) { + if ((val & ENET_INIT_PARAM_RISC_MASK) == risc) { snum = - (u32) (*p_start & ENET_INIT_PARAM_SNUM_MASK) >> + (u32) (val & ENET_INIT_PARAM_SNUM_MASK) >> ENET_INIT_PARAM_SNUM_SHIFT; qe_put_snum((u8) snum); if (!((i == 0) && skip_page_for_first_entry)) { @@ -440,7 +444,7 @@ static int hw_add_addr_in_paddr(struct ucc_geth_private *ugeth, static int hw_clear_addr_in_paddr(struct ucc_geth_private *ugeth, u8 paddr_num) { - struct ucc_geth_82xx_address_filtering_pram *p_82xx_addr_filt; + struct ucc_geth_82xx_address_filtering_pram __iomem *p_82xx_addr_filt; if (!(paddr_num < NUM_OF_PADDRS)) { ugeth_warn("%s: Illagel paddr_num.", __FUNCTION__); @@ -448,7 +452,7 @@ static int hw_clear_addr_in_paddr(struct ucc_geth_private *ugeth, u8 paddr_num) } p_82xx_addr_filt = - (struct ucc_geth_82xx_address_filtering_pram *) ugeth->p_rx_glbl_pram-> + (struct ucc_geth_82xx_address_filtering_pram __iomem *) ugeth->p_rx_glbl_pram-> addressfiltering; /* Writing address ff.ff.ff.ff.ff.ff disables address @@ -463,11 +467,11 @@ static int hw_clear_addr_in_paddr(struct ucc_geth_private *ugeth, u8 paddr_num) static void hw_add_addr_in_hash(struct ucc_geth_private *ugeth, u8 *p_enet_addr) { - struct ucc_geth_82xx_address_filtering_pram *p_82xx_addr_filt; + struct ucc_geth_82xx_address_filtering_pram __iomem *p_82xx_addr_filt; u32 cecr_subblock; p_82xx_addr_filt = - (struct ucc_geth_82xx_address_filtering_pram *) ugeth->p_rx_glbl_pram-> + (struct ucc_geth_82xx_address_filtering_pram __iomem *) ugeth->p_rx_glbl_pram-> addressfiltering; cecr_subblock = @@ -487,7 +491,7 @@ static void hw_add_addr_in_hash(struct ucc_geth_private *ugeth, static void magic_packet_detection_enable(struct ucc_geth_private *ugeth) { struct ucc_fast_private *uccf; - struct ucc_geth *ug_regs; + struct ucc_geth __iomem *ug_regs; u32 maccfg2, uccm; uccf = ugeth->uccf; @@ -507,7 +511,7 @@ static void magic_packet_detection_enable(struct ucc_geth_private *ugeth) static void magic_packet_detection_disable(struct ucc_geth_private *ugeth) { struct ucc_fast_private *uccf; - struct ucc_geth *ug_regs; + struct ucc_geth __iomem *ug_regs; u32 maccfg2, uccm; uccf = ugeth->uccf; @@ -538,13 +542,13 @@ static void get_statistics(struct ucc_geth_private *ugeth, rx_firmware_statistics, struct ucc_geth_hardware_statistics *hardware_statistics) { - struct ucc_fast *uf_regs; - struct ucc_geth *ug_regs; + struct ucc_fast __iomem *uf_regs; + struct ucc_geth __iomem *ug_regs; struct ucc_geth_tx_firmware_statistics_pram *p_tx_fw_statistics_pram; struct ucc_geth_rx_firmware_statistics_pram *p_rx_fw_statistics_pram; ug_regs = ugeth->ug_regs; - uf_regs = (struct ucc_fast *) ug_regs; + uf_regs = (struct ucc_fast __iomem *) ug_regs; p_tx_fw_statistics_pram = ugeth->p_tx_fw_statistics_pram; p_rx_fw_statistics_pram = ugeth->p_rx_fw_statistics_pram; @@ -1132,9 +1136,9 @@ static void dump_regs(struct ucc_geth_private *ugeth) } #endif /* DEBUG */ -static void init_default_reg_vals(volatile u32 *upsmr_register, - volatile u32 *maccfg1_register, - volatile u32 *maccfg2_register) +static void init_default_reg_vals(u32 __iomem *upsmr_register, + u32 __iomem *maccfg1_register, + u32 __iomem *maccfg2_register) { out_be32(upsmr_register, UCC_GETH_UPSMR_INIT); out_be32(maccfg1_register, UCC_GETH_MACCFG1_INIT); @@ -1148,7 +1152,7 @@ static int init_half_duplex_params(int alt_beb, u8 alt_beb_truncation, u8 max_retransmissions, u8 collision_window, - volatile u32 *hafdup_register) + u32 __iomem *hafdup_register) { u32 value = 0; @@ -1180,7 +1184,7 @@ static int init_inter_frame_gap_params(u8 non_btb_cs_ipg, u8 non_btb_ipg, u8 min_ifg, u8 btb_ipg, - volatile u32 *ipgifg_register) + u32 __iomem *ipgifg_register) { u32 value = 0; @@ -1215,9 +1219,9 @@ int init_flow_control_params(u32 automatic_flow_control_mode, int tx_flow_control_enable, u16 pause_period, u16 extension_field, - volatile u32 *upsmr_register, - volatile u32 *uempr_register, - volatile u32 *maccfg1_register) + u32 __iomem *upsmr_register, + u32 __iomem *uempr_register, + u32 __iomem *maccfg1_register) { u32 value = 0; @@ -1243,8 +1247,8 @@ int init_flow_control_params(u32 automatic_flow_control_mode, static int init_hw_statistics_gathering_mode(int enable_hardware_statistics, int auto_zero_hardware_statistics, - volatile u32 *upsmr_register, - volatile u16 *uescr_register) + u32 __iomem *upsmr_register, + u16 __iomem *uescr_register) { u32 upsmr_value = 0; u16 uescr_value = 0; @@ -1270,12 +1274,12 @@ static int init_hw_statistics_gathering_mode(int enable_hardware_statistics, static int init_firmware_statistics_gathering_mode(int enable_tx_firmware_statistics, int enable_rx_firmware_statistics, - volatile u32 *tx_rmon_base_ptr, + u32 __iomem *tx_rmon_base_ptr, u32 tx_firmware_statistics_structure_address, - volatile u32 *rx_rmon_base_ptr, + u32 __iomem *rx_rmon_base_ptr, u32 rx_firmware_statistics_structure_address, - volatile u16 *temoder_register, - volatile u32 *remoder_register) + u16 __iomem *temoder_register, + u32 __iomem *remoder_register) { /* Note: this function does not check if */ /* the parameters it receives are NULL */ @@ -1307,8 +1311,8 @@ static int init_mac_station_addr_regs(u8 address_byte_0, u8 address_byte_3, u8 address_byte_4, u8 address_byte_5, - volatile u32 *macstnaddr1_register, - volatile u32 *macstnaddr2_register) + u32 __iomem *macstnaddr1_register, + u32 __iomem *macstnaddr2_register) { u32 value = 0; @@ -1344,7 +1348,7 @@ static int init_mac_station_addr_regs(u8 address_byte_0, } static int init_check_frame_length_mode(int length_check, - volatile u32 *maccfg2_register) + u32 __iomem *maccfg2_register) { u32 value = 0; @@ -1360,7 +1364,7 @@ static int init_check_frame_length_mode(int length_check, } static int init_preamble_length(u8 preamble_length, - volatile u32 *maccfg2_register) + u32 __iomem *maccfg2_register) { u32 value = 0; @@ -1376,7 +1380,7 @@ static int init_preamble_length(u8 preamble_length, static int init_rx_parameters(int reject_broadcast, int receive_short_frames, - int promiscuous, volatile u32 *upsmr_register) + int promiscuous, u32 __iomem *upsmr_register) { u32 value = 0; @@ -1403,7 +1407,7 @@ static int init_rx_parameters(int reject_broadcast, } static int init_max_rx_buff_len(u16 max_rx_buf_len, - volatile u16 *mrblr_register) + u16 __iomem *mrblr_register) { /* max_rx_buf_len value must be a multiple of 128 */ if ((max_rx_buf_len == 0) @@ -1415,8 +1419,8 @@ static int init_max_rx_buff_len(u16 max_rx_buf_len, } static int init_min_frame_len(u16 min_frame_length, - volatile u16 *minflr_register, - volatile u16 *mrblr_register) + u16 __iomem *minflr_register, + u16 __iomem *mrblr_register) { u16 mrblr_value = 0; @@ -1431,8 +1435,8 @@ static int init_min_frame_len(u16 min_frame_length, static int adjust_enet_interface(struct ucc_geth_private *ugeth) { struct ucc_geth_info *ug_info; - struct ucc_geth *ug_regs; - struct ucc_fast *uf_regs; + struct ucc_geth __iomem *ug_regs; + struct ucc_fast __iomem *uf_regs; int ret_val; u32 upsmr, maccfg2, tbiBaseAddress; u16 value; @@ -1517,8 +1521,8 @@ static int adjust_enet_interface(struct ucc_geth_private *ugeth) static void adjust_link(struct net_device *dev) { struct ucc_geth_private *ugeth = netdev_priv(dev); - struct ucc_geth *ug_regs; - struct ucc_fast *uf_regs; + struct ucc_geth __iomem *ug_regs; + struct ucc_fast __iomem *uf_regs; struct phy_device *phydev = ugeth->phydev; unsigned long flags; int new_state = 0; @@ -1678,9 +1682,9 @@ static int ugeth_graceful_stop_rx(struct ucc_geth_private * ugeth) uccf = ugeth->uccf; /* Clear acknowledge bit */ - temp = ugeth->p_rx_glbl_pram->rxgstpack; + temp = in_8(&ugeth->p_rx_glbl_pram->rxgstpack); temp &= ~GRACEFUL_STOP_ACKNOWLEDGE_RX; - ugeth->p_rx_glbl_pram->rxgstpack = temp; + out_8(&ugeth->p_rx_glbl_pram->rxgstpack, temp); /* Keep issuing command and checking acknowledge bit until it is asserted, according to spec */ @@ -1692,7 +1696,7 @@ static int ugeth_graceful_stop_rx(struct ucc_geth_private * ugeth) qe_issue_cmd(QE_GRACEFUL_STOP_RX, cecr_subblock, QE_CR_PROTOCOL_ETHERNET, 0); - temp = ugeth->p_rx_glbl_pram->rxgstpack; + temp = in_8(&ugeth->p_rx_glbl_pram->rxgstpack); } while (!(temp & GRACEFUL_STOP_ACKNOWLEDGE_RX)); uccf->stopped_rx = 1; @@ -1991,19 +1995,20 @@ static int ugeth_82xx_filtering_clear_all_addr_in_hash(struct ucc_geth_private * enum enet_addr_type enet_addr_type) { - struct ucc_geth_82xx_address_filtering_pram *p_82xx_addr_filt; + struct ucc_geth_82xx_address_filtering_pram __iomem *p_82xx_addr_filt; struct ucc_fast_private *uccf; enum comm_dir comm_dir; struct list_head *p_lh; u16 i, num; - u32 *addr_h, *addr_l; + u32 __iomem *addr_h; + u32 __iomem *addr_l; u8 *p_counter; uccf = ugeth->uccf; p_82xx_addr_filt = - (struct ucc_geth_82xx_address_filtering_pram *) ugeth->p_rx_glbl_pram-> - addressfiltering; + (struct ucc_geth_82xx_address_filtering_pram __iomem *) + ugeth->p_rx_glbl_pram->addressfiltering; if (enet_addr_type == ENET_ADDR_TYPE_GROUP) { addr_h = &(p_82xx_addr_filt->gaddr_h); @@ -2079,7 +2084,7 @@ static int ugeth_82xx_filtering_clear_addr_in_paddr(struct ucc_geth_private *uge static void ucc_geth_memclean(struct ucc_geth_private *ugeth) { u16 i, j; - u8 *bd; + u8 __iomem *bd; if (!ugeth) return; @@ -2154,8 +2159,8 @@ static void ucc_geth_memclean(struct ucc_geth_private *ugeth) for (j = 0; j < ugeth->ug_info->bdRingLenTx[i]; j++) { if (ugeth->tx_skbuff[i][j]) { dma_unmap_single(NULL, - ((struct qe_bd *)bd)->buf, - (in_be32((u32 *)bd) & + in_be32(&((struct qe_bd __iomem *)bd)->buf), + (in_be32((u32 __iomem *)bd) & BD_LENGTH_MASK), DMA_TO_DEVICE); dev_kfree_skb_any(ugeth->tx_skbuff[i][j]); @@ -2182,7 +2187,7 @@ static void ucc_geth_memclean(struct ucc_geth_private *ugeth) for (j = 0; j < ugeth->ug_info->bdRingLenRx[i]; j++) { if (ugeth->rx_skbuff[i][j]) { dma_unmap_single(NULL, - ((struct qe_bd *)bd)->buf, + in_be32(&((struct qe_bd __iomem *)bd)->buf), ugeth->ug_info-> uf_info.max_rx_buf_length + UCC_GETH_RX_DATA_BUF_ALIGNMENT, @@ -2218,8 +2223,8 @@ static void ucc_geth_set_multi(struct net_device *dev) { struct ucc_geth_private *ugeth; struct dev_mc_list *dmi; - struct ucc_fast *uf_regs; - struct ucc_geth_82xx_address_filtering_pram *p_82xx_addr_filt; + struct ucc_fast __iomem *uf_regs; + struct ucc_geth_82xx_address_filtering_pram __iomem *p_82xx_addr_filt; int i; ugeth = netdev_priv(dev); @@ -2228,14 +2233,14 @@ static void ucc_geth_set_multi(struct net_device *dev) if (dev->flags & IFF_PROMISC) { - uf_regs->upsmr |= UPSMR_PRO; + out_be32(&uf_regs->upsmr, in_be32(&uf_regs->upsmr) | UPSMR_PRO); } else { - uf_regs->upsmr &= ~UPSMR_PRO; + out_be32(&uf_regs->upsmr, in_be32(&uf_regs->upsmr)&~UPSMR_PRO); p_82xx_addr_filt = - (struct ucc_geth_82xx_address_filtering_pram *) ugeth-> + (struct ucc_geth_82xx_address_filtering_pram __iomem *) ugeth-> p_rx_glbl_pram->addressfiltering; if (dev->flags & IFF_ALLMULTI) { @@ -2270,7 +2275,7 @@ static void ucc_geth_set_multi(struct net_device *dev) static void ucc_geth_stop(struct ucc_geth_private *ugeth) { - struct ucc_geth *ug_regs = ugeth->ug_regs; + struct ucc_geth __iomem *ug_regs = ugeth->ug_regs; struct phy_device *phydev = ugeth->phydev; u32 tempval; @@ -2419,20 +2424,20 @@ static int ucc_struct_init(struct ucc_geth_private *ugeth) return -ENOMEM; } - ugeth->ug_regs = (struct ucc_geth *) ioremap(uf_info->regs, sizeof(struct ucc_geth)); + ugeth->ug_regs = (struct ucc_geth __iomem *) ioremap(uf_info->regs, sizeof(struct ucc_geth)); return 0; } static int ucc_geth_startup(struct ucc_geth_private *ugeth) { - struct ucc_geth_82xx_address_filtering_pram *p_82xx_addr_filt; - struct ucc_geth_init_pram *p_init_enet_pram; + struct ucc_geth_82xx_address_filtering_pram __iomem *p_82xx_addr_filt; + struct ucc_geth_init_pram __iomem *p_init_enet_pram; struct ucc_fast_private *uccf; struct ucc_geth_info *ug_info; struct ucc_fast_info *uf_info; - struct ucc_fast *uf_regs; - struct ucc_geth *ug_regs; + struct ucc_fast __iomem *uf_regs; + struct ucc_geth __iomem *ug_regs; int ret_val = -EINVAL; u32 remoder = UCC_GETH_REMODER_INIT; u32 init_enet_pram_offset, cecr_subblock, command, maccfg1; @@ -2440,7 +2445,8 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) u16 temoder = UCC_GETH_TEMODER_INIT; u16 test; u8 function_code = 0; - u8 *bd, *endOfRing; + u8 __iomem *bd; + u8 __iomem *endOfRing; u8 numThreadsRxNumerical, numThreadsTxNumerical; ugeth_vdbg("%s: IN", __FUNCTION__); @@ -2602,11 +2608,11 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) if (UCC_GETH_TX_BD_RING_ALIGNMENT > 4) align = UCC_GETH_TX_BD_RING_ALIGNMENT; ugeth->tx_bd_ring_offset[j] = - kmalloc((u32) (length + align), GFP_KERNEL); + (u32) kmalloc((u32) (length + align), GFP_KERNEL); if (ugeth->tx_bd_ring_offset[j] != 0) ugeth->p_tx_bd_ring[j] = - (void*)((ugeth->tx_bd_ring_offset[j] + + (u8 __iomem *)((ugeth->tx_bd_ring_offset[j] + align) & ~(align - 1)); } else if (uf_info->bd_mem_part == MEM_PART_MURAM) { ugeth->tx_bd_ring_offset[j] = @@ -2614,7 +2620,7 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) UCC_GETH_TX_BD_RING_ALIGNMENT); if (!IS_ERR_VALUE(ugeth->tx_bd_ring_offset[j])) ugeth->p_tx_bd_ring[j] = - (u8 *) qe_muram_addr(ugeth-> + (u8 __iomem *) qe_muram_addr(ugeth-> tx_bd_ring_offset[j]); } if (!ugeth->p_tx_bd_ring[j]) { @@ -2626,8 +2632,8 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) return -ENOMEM; } /* Zero unused end of bd ring, according to spec */ - memset(ugeth->p_tx_bd_ring[j] + - ug_info->bdRingLenTx[j] * sizeof(struct qe_bd), 0, + memset_io((void __iomem *)(ugeth->p_tx_bd_ring[j] + + ug_info->bdRingLenTx[j] * sizeof(struct qe_bd)), 0, length - ug_info->bdRingLenTx[j] * sizeof(struct qe_bd)); } @@ -2639,10 +2645,10 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) if (UCC_GETH_RX_BD_RING_ALIGNMENT > 4) align = UCC_GETH_RX_BD_RING_ALIGNMENT; ugeth->rx_bd_ring_offset[j] = - kmalloc((u32) (length + align), GFP_KERNEL); + (u32) kmalloc((u32) (length + align), GFP_KERNEL); if (ugeth->rx_bd_ring_offset[j] != 0) ugeth->p_rx_bd_ring[j] = - (void*)((ugeth->rx_bd_ring_offset[j] + + (u8 __iomem *)((ugeth->rx_bd_ring_offset[j] + align) & ~(align - 1)); } else if (uf_info->bd_mem_part == MEM_PART_MURAM) { ugeth->rx_bd_ring_offset[j] = @@ -2650,7 +2656,7 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) UCC_GETH_RX_BD_RING_ALIGNMENT); if (!IS_ERR_VALUE(ugeth->rx_bd_ring_offset[j])) ugeth->p_rx_bd_ring[j] = - (u8 *) qe_muram_addr(ugeth-> + (u8 __iomem *) qe_muram_addr(ugeth-> rx_bd_ring_offset[j]); } if (!ugeth->p_rx_bd_ring[j]) { @@ -2685,14 +2691,14 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) bd = ugeth->confBd[j] = ugeth->txBd[j] = ugeth->p_tx_bd_ring[j]; for (i = 0; i < ug_info->bdRingLenTx[j]; i++) { /* clear bd buffer */ - out_be32(&((struct qe_bd *)bd)->buf, 0); + out_be32(&((struct qe_bd __iomem *)bd)->buf, 0); /* set bd status and length */ - out_be32((u32 *)bd, 0); + out_be32((u32 __iomem *)bd, 0); bd += sizeof(struct qe_bd); } bd -= sizeof(struct qe_bd); /* set bd status and length */ - out_be32((u32 *)bd, T_W); /* for last BD set Wrap bit */ + out_be32((u32 __iomem *)bd, T_W); /* for last BD set Wrap bit */ } /* Init Rx bds */ @@ -2717,14 +2723,14 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) bd = ugeth->rxBd[j] = ugeth->p_rx_bd_ring[j]; for (i = 0; i < ug_info->bdRingLenRx[j]; i++) { /* set bd status and length */ - out_be32((u32 *)bd, R_I); + out_be32((u32 __iomem *)bd, R_I); /* clear bd buffer */ - out_be32(&((struct qe_bd *)bd)->buf, 0); + out_be32(&((struct qe_bd __iomem *)bd)->buf, 0); bd += sizeof(struct qe_bd); } bd -= sizeof(struct qe_bd); /* set bd status and length */ - out_be32((u32 *)bd, R_W); /* for last BD set Wrap bit */ + out_be32((u32 __iomem *)bd, R_W); /* for last BD set Wrap bit */ } /* @@ -2744,10 +2750,10 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) return -ENOMEM; } ugeth->p_tx_glbl_pram = - (struct ucc_geth_tx_global_pram *) qe_muram_addr(ugeth-> + (struct ucc_geth_tx_global_pram __iomem *) qe_muram_addr(ugeth-> tx_glbl_pram_offset); /* Zero out p_tx_glbl_pram */ - memset(ugeth->p_tx_glbl_pram, 0, sizeof(struct ucc_geth_tx_global_pram)); + memset_io((void __iomem *)ugeth->p_tx_glbl_pram, 0, sizeof(struct ucc_geth_tx_global_pram)); /* Fill global PRAM */ @@ -2768,7 +2774,7 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) } ugeth->p_thread_data_tx = - (struct ucc_geth_thread_data_tx *) qe_muram_addr(ugeth-> + (struct ucc_geth_thread_data_tx __iomem *) qe_muram_addr(ugeth-> thread_dat_tx_offset); out_be32(&ugeth->p_tx_glbl_pram->tqptr, ugeth->thread_dat_tx_offset); @@ -2779,7 +2785,8 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) /* iphoffset */ for (i = 0; i < TX_IP_OFFSET_ENTRY_MAX; i++) - ugeth->p_tx_glbl_pram->iphoffset[i] = ug_info->iphoffset[i]; + out_8(&ugeth->p_tx_glbl_pram->iphoffset[i], + ug_info->iphoffset[i]); /* SQPTR */ /* Size varies with number of Tx queues */ @@ -2797,7 +2804,7 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) } ugeth->p_send_q_mem_reg = - (struct ucc_geth_send_queue_mem_region *) qe_muram_addr(ugeth-> + (struct ucc_geth_send_queue_mem_region __iomem *) qe_muram_addr(ugeth-> send_q_mem_reg_offset); out_be32(&ugeth->p_tx_glbl_pram->sqptr, ugeth->send_q_mem_reg_offset); @@ -2841,25 +2848,26 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) } ugeth->p_scheduler = - (struct ucc_geth_scheduler *) qe_muram_addr(ugeth-> + (struct ucc_geth_scheduler __iomem *) qe_muram_addr(ugeth-> scheduler_offset); out_be32(&ugeth->p_tx_glbl_pram->schedulerbasepointer, ugeth->scheduler_offset); /* Zero out p_scheduler */ - memset(ugeth->p_scheduler, 0, sizeof(struct ucc_geth_scheduler)); + memset_io((void __iomem *)ugeth->p_scheduler, 0, sizeof(struct ucc_geth_scheduler)); /* Set values in scheduler */ out_be32(&ugeth->p_scheduler->mblinterval, ug_info->mblinterval); out_be16(&ugeth->p_scheduler->nortsrbytetime, ug_info->nortsrbytetime); - ugeth->p_scheduler->fracsiz = ug_info->fracsiz; - ugeth->p_scheduler->strictpriorityq = ug_info->strictpriorityq; - ugeth->p_scheduler->txasap = ug_info->txasap; - ugeth->p_scheduler->extrabw = ug_info->extrabw; + out_8(&ugeth->p_scheduler->fracsiz, ug_info->fracsiz); + out_8(&ugeth->p_scheduler->strictpriorityq, + ug_info->strictpriorityq); + out_8(&ugeth->p_scheduler->txasap, ug_info->txasap); + out_8(&ugeth->p_scheduler->extrabw, ug_info->extrabw); for (i = 0; i < NUM_TX_QUEUES; i++) - ugeth->p_scheduler->weightfactor[i] = - ug_info->weightfactor[i]; + out_8(&ugeth->p_scheduler->weightfactor[i], + ug_info->weightfactor[i]); /* Set pointers to cpucount registers in scheduler */ ugeth->p_cpucount[0] = &(ugeth->p_scheduler->cpucount0); @@ -2890,10 +2898,10 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) return -ENOMEM; } ugeth->p_tx_fw_statistics_pram = - (struct ucc_geth_tx_firmware_statistics_pram *) + (struct ucc_geth_tx_firmware_statistics_pram __iomem *) qe_muram_addr(ugeth->tx_fw_statistics_pram_offset); /* Zero out p_tx_fw_statistics_pram */ - memset(ugeth->p_tx_fw_statistics_pram, + memset_io((void __iomem *)ugeth->p_tx_fw_statistics_pram, 0, sizeof(struct ucc_geth_tx_firmware_statistics_pram)); } @@ -2930,10 +2938,10 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) return -ENOMEM; } ugeth->p_rx_glbl_pram = - (struct ucc_geth_rx_global_pram *) qe_muram_addr(ugeth-> + (struct ucc_geth_rx_global_pram __iomem *) qe_muram_addr(ugeth-> rx_glbl_pram_offset); /* Zero out p_rx_glbl_pram */ - memset(ugeth->p_rx_glbl_pram, 0, sizeof(struct ucc_geth_rx_global_pram)); + memset_io((void __iomem *)ugeth->p_rx_glbl_pram, 0, sizeof(struct ucc_geth_rx_global_pram)); /* Fill global PRAM */ @@ -2953,7 +2961,7 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) } ugeth->p_thread_data_rx = - (struct ucc_geth_thread_data_rx *) qe_muram_addr(ugeth-> + (struct ucc_geth_thread_data_rx __iomem *) qe_muram_addr(ugeth-> thread_dat_rx_offset); out_be32(&ugeth->p_rx_glbl_pram->rqptr, ugeth->thread_dat_rx_offset); @@ -2976,10 +2984,10 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) return -ENOMEM; } ugeth->p_rx_fw_statistics_pram = - (struct ucc_geth_rx_firmware_statistics_pram *) + (struct ucc_geth_rx_firmware_statistics_pram __iomem *) qe_muram_addr(ugeth->rx_fw_statistics_pram_offset); /* Zero out p_rx_fw_statistics_pram */ - memset(ugeth->p_rx_fw_statistics_pram, 0, + memset_io((void __iomem *)ugeth->p_rx_fw_statistics_pram, 0, sizeof(struct ucc_geth_rx_firmware_statistics_pram)); } @@ -3000,7 +3008,7 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) } ugeth->p_rx_irq_coalescing_tbl = - (struct ucc_geth_rx_interrupt_coalescing_table *) + (struct ucc_geth_rx_interrupt_coalescing_table __iomem *) qe_muram_addr(ugeth->rx_irq_coalescing_tbl_offset); out_be32(&ugeth->p_rx_glbl_pram->intcoalescingptr, ugeth->rx_irq_coalescing_tbl_offset); @@ -3069,11 +3077,11 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) } ugeth->p_rx_bd_qs_tbl = - (struct ucc_geth_rx_bd_queues_entry *) qe_muram_addr(ugeth-> + (struct ucc_geth_rx_bd_queues_entry __iomem *) qe_muram_addr(ugeth-> rx_bd_qs_tbl_offset); out_be32(&ugeth->p_rx_glbl_pram->rbdqptr, ugeth->rx_bd_qs_tbl_offset); /* Zero out p_rx_bd_qs_tbl */ - memset(ugeth->p_rx_bd_qs_tbl, + memset_io((void __iomem *)ugeth->p_rx_bd_qs_tbl, 0, ug_info->numQueuesRx * (sizeof(struct ucc_geth_rx_bd_queues_entry) + sizeof(struct ucc_geth_rx_prefetched_bds))); @@ -3133,7 +3141,7 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) &ugeth->p_rx_glbl_pram->remoder); /* function code register */ - ugeth->p_rx_glbl_pram->rstate = function_code; + out_8(&ugeth->p_rx_glbl_pram->rstate, function_code); /* initialize extended filtering */ if (ug_info->rxExtendedFiltering) { @@ -3160,7 +3168,7 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) } ugeth->p_exf_glbl_param = - (struct ucc_geth_exf_global_pram *) qe_muram_addr(ugeth-> + (struct ucc_geth_exf_global_pram __iomem *) qe_muram_addr(ugeth-> exf_glbl_param_offset); out_be32(&ugeth->p_rx_glbl_pram->exfGlobalParam, ugeth->exf_glbl_param_offset); @@ -3175,7 +3183,7 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) ugeth_82xx_filtering_clear_addr_in_paddr(ugeth, (u8) j); p_82xx_addr_filt = - (struct ucc_geth_82xx_address_filtering_pram *) ugeth-> + (struct ucc_geth_82xx_address_filtering_pram __iomem *) ugeth-> p_rx_glbl_pram->addressfiltering; ugeth_82xx_filtering_clear_all_addr_in_hash(ugeth, @@ -3307,17 +3315,21 @@ static int ucc_geth_startup(struct ucc_geth_private *ugeth) return -ENOMEM; } p_init_enet_pram = - (struct ucc_geth_init_pram *) qe_muram_addr(init_enet_pram_offset); + (struct ucc_geth_init_pram __iomem *) qe_muram_addr(init_enet_pram_offset); /* Copy shadow InitEnet command parameter structure into PRAM */ - p_init_enet_pram->resinit1 = ugeth->p_init_enet_param_shadow->resinit1; - p_init_enet_pram->resinit2 = ugeth->p_init_enet_param_shadow->resinit2; - p_init_enet_pram->resinit3 = ugeth->p_init_enet_param_shadow->resinit3; - p_init_enet_pram->resinit4 = ugeth->p_init_enet_param_shadow->resinit4; + out_8(&p_init_enet_pram->resinit1, + ugeth->p_init_enet_param_shadow->resinit1); + out_8(&p_init_enet_pram->resinit2, + ugeth->p_init_enet_param_shadow->resinit2); + out_8(&p_init_enet_pram->resinit3, + ugeth->p_init_enet_param_shadow->resinit3); + out_8(&p_init_enet_pram->resinit4, + ugeth->p_init_enet_param_shadow->resinit4); out_be16(&p_init_enet_pram->resinit5, ugeth->p_init_enet_param_shadow->resinit5); - p_init_enet_pram->largestexternallookupkeysize = - ugeth->p_init_enet_param_shadow->largestexternallookupkeysize; + out_8(&p_init_enet_pram->largestexternallookupkeysize, + ugeth->p_init_enet_param_shadow->largestexternallookupkeysize); out_be32(&p_init_enet_pram->rgftgfrxglobal, ugeth->p_init_enet_param_shadow->rgftgfrxglobal); for (i = 0; i < ENET_INIT_PARAM_MAX_ENTRIES_RX; i++) @@ -3371,7 +3383,7 @@ static int ucc_geth_start_xmit(struct sk_buff *skb, struct net_device *dev) #ifdef CONFIG_UGETH_TX_ON_DEMAND struct ucc_fast_private *uccf; #endif - u8 *bd; /* BD pointer */ + u8 __iomem *bd; /* BD pointer */ u32 bd_status; u8 txQ = 0; @@ -3383,7 +3395,7 @@ static int ucc_geth_start_xmit(struct sk_buff *skb, struct net_device *dev) /* Start from the next BD that should be filled */ bd = ugeth->txBd[txQ]; - bd_status = in_be32((u32 *)bd); + bd_status = in_be32((u32 __iomem *)bd); /* Save the skb pointer so we can free it later */ ugeth->tx_skbuff[txQ][ugeth->skb_curtx[txQ]] = skb; @@ -3393,7 +3405,7 @@ static int ucc_geth_start_xmit(struct sk_buff *skb, struct net_device *dev) 1) & TX_RING_MOD_MASK(ugeth->ug_info->bdRingLenTx[txQ]); /* set up the buffer descriptor */ - out_be32(&((struct qe_bd *)bd)->buf, + out_be32(&((struct qe_bd __iomem *)bd)->buf, dma_map_single(NULL, skb->data, skb->len, DMA_TO_DEVICE)); /* printk(KERN_DEBUG"skb->data is 0x%x\n",skb->data); */ @@ -3401,7 +3413,7 @@ static int ucc_geth_start_xmit(struct sk_buff *skb, struct net_device *dev) bd_status = (bd_status & T_W) | T_R | T_I | T_L | skb->len; /* set bd status and length */ - out_be32((u32 *)bd, bd_status); + out_be32((u32 __iomem *)bd, bd_status); dev->trans_start = jiffies; @@ -3441,7 +3453,7 @@ static int ucc_geth_start_xmit(struct sk_buff *skb, struct net_device *dev) static int ucc_geth_rx(struct ucc_geth_private *ugeth, u8 rxQ, int rx_work_limit) { struct sk_buff *skb; - u8 *bd; + u8 __iomem *bd; u16 length, howmany = 0; u32 bd_status; u8 *bdBuffer; @@ -3454,11 +3466,11 @@ static int ucc_geth_rx(struct ucc_geth_private *ugeth, u8 rxQ, int rx_work_limit /* collect received buffers */ bd = ugeth->rxBd[rxQ]; - bd_status = in_be32((u32 *)bd); + bd_status = in_be32((u32 __iomem *)bd); /* while there are received buffers and BD is full (~R_E) */ while (!((bd_status & (R_E)) || (--rx_work_limit < 0))) { - bdBuffer = (u8 *) in_be32(&((struct qe_bd *)bd)->buf); + bdBuffer = (u8 *) in_be32(&((struct qe_bd __iomem *)bd)->buf); length = (u16) ((bd_status & BD_LENGTH_MASK) - 4); skb = ugeth->rx_skbuff[rxQ][ugeth->skb_currx[rxQ]]; @@ -3516,7 +3528,7 @@ static int ucc_geth_rx(struct ucc_geth_private *ugeth, u8 rxQ, int rx_work_limit else bd += sizeof(struct qe_bd); - bd_status = in_be32((u32 *)bd); + bd_status = in_be32((u32 __iomem *)bd); } ugeth->rxBd[rxQ] = bd; @@ -3527,11 +3539,11 @@ static int ucc_geth_tx(struct net_device *dev, u8 txQ) { /* Start from the next BD that should be filled */ struct ucc_geth_private *ugeth = netdev_priv(dev); - u8 *bd; /* BD pointer */ + u8 __iomem *bd; /* BD pointer */ u32 bd_status; bd = ugeth->confBd[txQ]; - bd_status = in_be32((u32 *)bd); + bd_status = in_be32((u32 __iomem *)bd); /* Normal processing. */ while ((bd_status & T_R) == 0) { @@ -3561,7 +3573,7 @@ static int ucc_geth_tx(struct net_device *dev, u8 txQ) bd += sizeof(struct qe_bd); else bd = ugeth->p_tx_bd_ring[txQ]; - bd_status = in_be32((u32 *)bd); + bd_status = in_be32((u32 __iomem *)bd); } ugeth->confBd[txQ] = bd; return 0; diff --git a/drivers/net/ucc_geth.h b/drivers/net/ucc_geth.h index 9f8b7580a3a4..abc0e2242634 100644 --- a/drivers/net/ucc_geth.h +++ b/drivers/net/ucc_geth.h @@ -700,8 +700,8 @@ struct ucc_geth_82xx_address_filtering_pram { u32 iaddr_l; /* individual address filter, low */ u32 gaddr_h; /* group address filter, high */ u32 gaddr_l; /* group address filter, low */ - struct ucc_geth_82xx_enet_address taddr; - struct ucc_geth_82xx_enet_address paddr[NUM_OF_PADDRS]; + struct ucc_geth_82xx_enet_address __iomem taddr; + struct ucc_geth_82xx_enet_address __iomem paddr[NUM_OF_PADDRS]; u8 res0[0x40 - 0x38]; } __attribute__ ((packed)); @@ -1186,40 +1186,40 @@ struct ucc_geth_private { struct ucc_fast_private *uccf; struct net_device *dev; struct napi_struct napi; - struct ucc_geth *ug_regs; + struct ucc_geth __iomem *ug_regs; struct ucc_geth_init_pram *p_init_enet_param_shadow; - struct ucc_geth_exf_global_pram *p_exf_glbl_param; + struct ucc_geth_exf_global_pram __iomem *p_exf_glbl_param; u32 exf_glbl_param_offset; - struct ucc_geth_rx_global_pram *p_rx_glbl_pram; + struct ucc_geth_rx_global_pram __iomem *p_rx_glbl_pram; u32 rx_glbl_pram_offset; - struct ucc_geth_tx_global_pram *p_tx_glbl_pram; + struct ucc_geth_tx_global_pram __iomem *p_tx_glbl_pram; u32 tx_glbl_pram_offset; - struct ucc_geth_send_queue_mem_region *p_send_q_mem_reg; + struct ucc_geth_send_queue_mem_region __iomem *p_send_q_mem_reg; u32 send_q_mem_reg_offset; - struct ucc_geth_thread_data_tx *p_thread_data_tx; + struct ucc_geth_thread_data_tx __iomem *p_thread_data_tx; u32 thread_dat_tx_offset; - struct ucc_geth_thread_data_rx *p_thread_data_rx; + struct ucc_geth_thread_data_rx __iomem *p_thread_data_rx; u32 thread_dat_rx_offset; - struct ucc_geth_scheduler *p_scheduler; + struct ucc_geth_scheduler __iomem *p_scheduler; u32 scheduler_offset; - struct ucc_geth_tx_firmware_statistics_pram *p_tx_fw_statistics_pram; + struct ucc_geth_tx_firmware_statistics_pram __iomem *p_tx_fw_statistics_pram; u32 tx_fw_statistics_pram_offset; - struct ucc_geth_rx_firmware_statistics_pram *p_rx_fw_statistics_pram; + struct ucc_geth_rx_firmware_statistics_pram __iomem *p_rx_fw_statistics_pram; u32 rx_fw_statistics_pram_offset; - struct ucc_geth_rx_interrupt_coalescing_table *p_rx_irq_coalescing_tbl; + struct ucc_geth_rx_interrupt_coalescing_table __iomem *p_rx_irq_coalescing_tbl; u32 rx_irq_coalescing_tbl_offset; - struct ucc_geth_rx_bd_queues_entry *p_rx_bd_qs_tbl; + struct ucc_geth_rx_bd_queues_entry __iomem *p_rx_bd_qs_tbl; u32 rx_bd_qs_tbl_offset; - u8 *p_tx_bd_ring[NUM_TX_QUEUES]; + u8 __iomem *p_tx_bd_ring[NUM_TX_QUEUES]; u32 tx_bd_ring_offset[NUM_TX_QUEUES]; - u8 *p_rx_bd_ring[NUM_RX_QUEUES]; + u8 __iomem *p_rx_bd_ring[NUM_RX_QUEUES]; u32 rx_bd_ring_offset[NUM_RX_QUEUES]; - u8 *confBd[NUM_TX_QUEUES]; - u8 *txBd[NUM_TX_QUEUES]; - u8 *rxBd[NUM_RX_QUEUES]; + u8 __iomem *confBd[NUM_TX_QUEUES]; + u8 __iomem *txBd[NUM_TX_QUEUES]; + u8 __iomem *rxBd[NUM_RX_QUEUES]; int badFrame[NUM_RX_QUEUES]; u16 cpucount[NUM_TX_QUEUES]; - volatile u16 *p_cpucount[NUM_TX_QUEUES]; + u16 __iomem *p_cpucount[NUM_TX_QUEUES]; int indAddrRegUsed[NUM_OF_PADDRS]; u8 paddr[NUM_OF_PADDRS][ENET_NUM_OCTETS_PER_ADDRESS]; /* ethernet address */ u8 numGroupAddrInHash; @@ -1251,4 +1251,12 @@ struct ucc_geth_private { int oldlink; }; +void uec_set_ethtool_ops(struct net_device *netdev); +int init_flow_control_params(u32 automatic_flow_control_mode, + int rx_flow_control_enable, int tx_flow_control_enable, + u16 pause_period, u16 extension_field, + u32 __iomem *upsmr_register, u32 __iomem *uempr_register, + u32 __iomem *maccfg1_register); + + #endif /* __UCC_GETH_H__ */ diff --git a/drivers/net/ucc_geth_ethtool.c b/drivers/net/ucc_geth_ethtool.c index 9a9622c13e2b..299b7f176950 100644 --- a/drivers/net/ucc_geth_ethtool.c +++ b/drivers/net/ucc_geth_ethtool.c @@ -108,12 +108,6 @@ static char rx_fw_stat_gstrings[][ETH_GSTRING_LEN] = { #define UEC_TX_FW_STATS_LEN ARRAY_SIZE(tx_fw_stat_gstrings) #define UEC_RX_FW_STATS_LEN ARRAY_SIZE(rx_fw_stat_gstrings) -extern int init_flow_control_params(u32 automatic_flow_control_mode, - int rx_flow_control_enable, - int tx_flow_control_enable, u16 pause_period, - u16 extension_field, volatile u32 *upsmr_register, - volatile u32 *uempr_register, volatile u32 *maccfg1_register); - static int uec_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd) { diff --git a/drivers/net/ucc_geth_mii.c b/drivers/net/ucc_geth_mii.c index 2af490781005..940474736922 100644 --- a/drivers/net/ucc_geth_mii.c +++ b/drivers/net/ucc_geth_mii.c @@ -104,7 +104,7 @@ int uec_mdio_read(struct mii_bus *bus, int mii_id, int regnum) } /* Reset the MIIM registers, and wait for the bus to free */ -int uec_mdio_reset(struct mii_bus *bus) +static int uec_mdio_reset(struct mii_bus *bus) { struct ucc_mii_mng __iomem *regs = (void __iomem *)bus->priv; unsigned int timeout = PHY_INIT_TIMEOUT; @@ -240,7 +240,7 @@ reg_map_fail: return err; } -int uec_mdio_remove(struct of_device *ofdev) +static int uec_mdio_remove(struct of_device *ofdev) { struct device *device = &ofdev->dev; struct mii_bus *bus = dev_get_drvdata(device); From afd8e39919c913993ac2f9984af8a9ba21c63d27 Mon Sep 17 00:00:00 2001 From: Anton Vorontsov Date: Tue, 29 Apr 2008 19:53:13 +0400 Subject: [PATCH 13/32] uli526x: initialize the hardware prior to requesting interrupts The firmware on MPC8610HPCD boards enables ULI ethernet and leaves it in some funky state before booting Linux. For drivers, it's always good idea to (re)initialize the hardware prior to requesting interrupts. This patch fixes the following oops: Oops: Kernel access of bad area, sig: 11 [#1] MPC86xx HPCD NIP: c0172820 LR: c017287c CTR: 00000000 [...] NIP [c0172820] allocate_rx_buffer+0x2c/0xb0 LR [c017287c] allocate_rx_buffer+0x88/0xb0 Call Trace: [df82bdc0] [c017287c] allocate_rx_buffer+0x88/0xb0 (unreliable) [df82bde0] [c0173000] uli526x_interrupt+0xe4/0x49c [df82be20] [c0045418] request_irq+0xf0/0x114 [df82be50] [c01737b0] uli526x_open+0x48/0x160 [df82be70] [c0201184] dev_open+0xb0/0xe8 [df82be80] [c0200104] dev_change_flags+0x90/0x1bc [df82bea0] [c035fab0] ip_auto_config+0x214/0xef4 [df82bf60] [c03421c8] kernel_init+0xc4/0x2ac [df82bff0] [c0010834] kernel_thread+0x44/0x60 Instruction dump: 4e800020 9421ffe0 7c0802a6 bfa10014 7c7e1b78 90010024 80030060 83e30054 2b80002f 419d0078 3fa0c039 48000058 <907f0010> 80630088 2f830000 419e0014 Signed-off-by: Anton Vorontsov Signed-off-by: Jeff Garzik --- drivers/net/tulip/uli526x.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/net/tulip/uli526x.c b/drivers/net/tulip/uli526x.c index a59c1f224aa8..1f077ac9b0e0 100644 --- a/drivers/net/tulip/uli526x.c +++ b/drivers/net/tulip/uli526x.c @@ -434,10 +434,6 @@ static int uli526x_open(struct net_device *dev) ULI526X_DBUG(0, "uli526x_open", 0); - ret = request_irq(dev->irq, &uli526x_interrupt, IRQF_SHARED, dev->name, dev); - if (ret) - return ret; - /* system variable init */ db->cr6_data = CR6_DEFAULT | uli526x_cr6_user_set; db->tx_packet_cnt = 0; @@ -456,6 +452,10 @@ static int uli526x_open(struct net_device *dev) /* Initialize ULI526X board */ uli526x_init(dev); + ret = request_irq(dev->irq, &uli526x_interrupt, IRQF_SHARED, dev->name, dev); + if (ret) + return ret; + /* Active System Interface */ netif_wake_queue(dev); From e284e5c6601cbb16e48854be26aa57a8fa844e35 Mon Sep 17 00:00:00 2001 From: Anton Vorontsov Date: Tue, 29 Apr 2008 19:53:18 +0400 Subject: [PATCH 14/32] uli526x: fix endianness issues in the setup frame This patch fixes uli526x driver's issues on a PowerPC boards: uli chip is unable to receive the packets. It appears that send_frame_filter prepares the setup frame in the endianness unsafe manner. On a big endian machines we should shift the address nibble by two bytes. Signed-off-by: Anton Vorontsov Signed-off-by: Jeff Garzik --- drivers/net/tulip/uli526x.c | 30 ++++++++++++++++++------------ 1 file changed, 18 insertions(+), 12 deletions(-) diff --git a/drivers/net/tulip/uli526x.c b/drivers/net/tulip/uli526x.c index 1f077ac9b0e0..2511ca7a12aa 100644 --- a/drivers/net/tulip/uli526x.c +++ b/drivers/net/tulip/uli526x.c @@ -1368,6 +1368,12 @@ static void update_cr6(u32 cr6_data, unsigned long ioaddr) * This setup frame initialize ULI526X address filter mode */ +#ifdef __BIG_ENDIAN +#define FLT_SHIFT 16 +#else +#define FLT_SHIFT 0 +#endif + static void send_filter_frame(struct net_device *dev, int mc_cnt) { struct uli526x_board_info *db = netdev_priv(dev); @@ -1384,27 +1390,27 @@ static void send_filter_frame(struct net_device *dev, int mc_cnt) /* Node address */ addrptr = (u16 *) dev->dev_addr; - *suptr++ = addrptr[0]; - *suptr++ = addrptr[1]; - *suptr++ = addrptr[2]; + *suptr++ = addrptr[0] << FLT_SHIFT; + *suptr++ = addrptr[1] << FLT_SHIFT; + *suptr++ = addrptr[2] << FLT_SHIFT; /* broadcast address */ - *suptr++ = 0xffff; - *suptr++ = 0xffff; - *suptr++ = 0xffff; + *suptr++ = 0xffff << FLT_SHIFT; + *suptr++ = 0xffff << FLT_SHIFT; + *suptr++ = 0xffff << FLT_SHIFT; /* fit the multicast address */ for (mcptr = dev->mc_list, i = 0; i < mc_cnt; i++, mcptr = mcptr->next) { addrptr = (u16 *) mcptr->dmi_addr; - *suptr++ = addrptr[0]; - *suptr++ = addrptr[1]; - *suptr++ = addrptr[2]; + *suptr++ = addrptr[0] << FLT_SHIFT; + *suptr++ = addrptr[1] << FLT_SHIFT; + *suptr++ = addrptr[2] << FLT_SHIFT; } for (; i<14; i++) { - *suptr++ = 0xffff; - *suptr++ = 0xffff; - *suptr++ = 0xffff; + *suptr++ = 0xffff << FLT_SHIFT; + *suptr++ = 0xffff << FLT_SHIFT; + *suptr++ = 0xffff << FLT_SHIFT; } /* prepare the setup frame */ From 97ac8caee238d2a81c23661916f7acd3a22c85fe Mon Sep 17 00:00:00 2001 From: Bruce Allan Date: Tue, 29 Apr 2008 09:16:05 -0700 Subject: [PATCH 15/32] e1000e: Add support for BM PHYs on ICH9 This patch adds support for the BM PHY, a new PHY model being used on ICH9-based implementations. This new PHY exposes issues in the ICH9 silicon when receiving jumbo frames large enough to use more than a certain part of the Rx FIFO, and this unfortunately breaks packet split jumbo receives. For this reason we re-introduce (for affected adapters only) the jumbo single-skb receive routine back so that people who do wish to use jumbo frames on these ich9 platforms can do so. Part of this problem has to do with CPU sleep states and to make sure that all the wake up timings are correctly we force them with the recently merged pm_qos infrastructure written by Mark Gross. (See http://lkml.org/lkml/2007/10/4/400). To make code read a bit easier we introduce a _IS_ICH flag so that we don't need to do mac type checks over the code. Signed-off-by: Bruce Allan Signed-off-by: Auke Kok Signed-off-by: Jeff Garzik --- drivers/net/e1000e/defines.h | 10 ++ drivers/net/e1000e/e1000.h | 7 +- drivers/net/e1000e/ethtool.c | 39 +++-- drivers/net/e1000e/hw.h | 22 +++ drivers/net/e1000e/ich8lan.c | 83 ++++++++- drivers/net/e1000e/netdev.c | 330 ++++++++++++++++++++++++++++++++++- drivers/net/e1000e/phy.c | 278 +++++++++++++++++++++++++++++ 7 files changed, 748 insertions(+), 21 deletions(-) diff --git a/drivers/net/e1000e/defines.h b/drivers/net/e1000e/defines.h index 2a53875cddbf..f823b8ba5785 100644 --- a/drivers/net/e1000e/defines.h +++ b/drivers/net/e1000e/defines.h @@ -648,6 +648,8 @@ #define IFE_E_PHY_ID 0x02A80330 #define IFE_PLUS_E_PHY_ID 0x02A80320 #define IFE_C_E_PHY_ID 0x02A80310 +#define BME1000_E_PHY_ID 0x01410CB0 +#define BME1000_E_PHY_ID_R2 0x01410CB1 /* M88E1000 Specific Registers */ #define M88E1000_PHY_SPEC_CTRL 0x10 /* PHY Specific Control Register */ @@ -701,6 +703,14 @@ #define M88EC018_EPSCR_DOWNSHIFT_COUNTER_MASK 0x0E00 #define M88EC018_EPSCR_DOWNSHIFT_COUNTER_5X 0x0800 +/* BME1000 PHY Specific Control Register */ +#define BME1000_PSCR_ENABLE_DOWNSHIFT 0x0800 /* 1 = enable downshift */ + + +#define PHY_PAGE_SHIFT 5 +#define PHY_REG(page, reg) (((page) << PHY_PAGE_SHIFT) | \ + ((reg) & MAX_PHY_REG_ADDRESS)) + /* * Bits... * 15-5: page diff --git a/drivers/net/e1000e/e1000.h b/drivers/net/e1000e/e1000.h index 38bfd0d261fe..d3bc6f8101fa 100644 --- a/drivers/net/e1000e/e1000.h +++ b/drivers/net/e1000e/e1000.h @@ -127,7 +127,7 @@ struct e1000_buffer { /* arrays of page information for packet split */ struct e1000_ps_page *ps_pages; }; - + struct page *page; }; struct e1000_ring { @@ -304,6 +304,7 @@ struct e1000_info { #define FLAG_HAS_CTRLEXT_ON_LOAD (1 << 5) #define FLAG_HAS_SWSM_ON_LOAD (1 << 6) #define FLAG_HAS_JUMBO_FRAMES (1 << 7) +#define FLAG_IS_ICH (1 << 9) #define FLAG_HAS_SMART_POWER_DOWN (1 << 11) #define FLAG_IS_QUAD_PORT_A (1 << 12) #define FLAG_IS_QUAD_PORT (1 << 13) @@ -386,6 +387,7 @@ extern void e1000e_set_kmrn_lock_loss_workaround_ich8lan(struct e1000_hw *hw, bool state); extern void e1000e_igp3_phy_powerdown_workaround_ich8lan(struct e1000_hw *hw); extern void e1000e_gig_downshift_workaround_ich8lan(struct e1000_hw *hw); +extern void e1000e_disable_gig_wol_ich8lan(struct e1000_hw *hw); extern s32 e1000e_check_for_copper_link(struct e1000_hw *hw); extern s32 e1000e_check_for_fiber_link(struct e1000_hw *hw); @@ -443,6 +445,9 @@ extern s32 e1000e_get_phy_info_m88(struct e1000_hw *hw); extern s32 e1000e_read_phy_reg_m88(struct e1000_hw *hw, u32 offset, u16 *data); extern s32 e1000e_write_phy_reg_m88(struct e1000_hw *hw, u32 offset, u16 data); extern enum e1000_phy_type e1000e_get_phy_type_from_id(u32 phy_id); +extern s32 e1000e_determine_phy_address(struct e1000_hw *hw); +extern s32 e1000e_write_phy_reg_bm(struct e1000_hw *hw, u32 offset, u16 data); +extern s32 e1000e_read_phy_reg_bm(struct e1000_hw *hw, u32 offset, u16 *data); extern void e1000e_phy_force_speed_duplex_setup(struct e1000_hw *hw, u16 *phy_ctrl); extern s32 e1000e_write_kmrn_reg(struct e1000_hw *hw, u32 offset, u16 data); extern s32 e1000e_read_kmrn_reg(struct e1000_hw *hw, u32 offset, u16 *data); diff --git a/drivers/net/e1000e/ethtool.c b/drivers/net/e1000e/ethtool.c index ce045acce63e..2bb6da057c40 100644 --- a/drivers/net/e1000e/ethtool.c +++ b/drivers/net/e1000e/ethtool.c @@ -803,8 +803,7 @@ static int e1000_reg_test(struct e1000_adapter *adapter, u64 *data) /* restore previous status */ ew32(STATUS, before); - if ((mac->type != e1000_ich8lan) && - (mac->type != e1000_ich9lan)) { + if (!(adapter->flags & FLAG_IS_ICH)) { REG_PATTERN_TEST(E1000_FCAL, 0xFFFFFFFF, 0xFFFFFFFF); REG_PATTERN_TEST(E1000_FCAH, 0x0000FFFF, 0xFFFFFFFF); REG_PATTERN_TEST(E1000_FCT, 0x0000FFFF, 0xFFFFFFFF); @@ -824,15 +823,13 @@ static int e1000_reg_test(struct e1000_adapter *adapter, u64 *data) REG_SET_AND_CHECK(E1000_RCTL, 0xFFFFFFFF, 0x00000000); - before = (((mac->type == e1000_ich8lan) || - (mac->type == e1000_ich9lan)) ? 0x06C3B33E : 0x06DFB3FE); + before = ((adapter->flags & FLAG_IS_ICH) ? 0x06C3B33E : 0x06DFB3FE); REG_SET_AND_CHECK(E1000_RCTL, before, 0x003FFFFB); REG_SET_AND_CHECK(E1000_TCTL, 0xFFFFFFFF, 0x00000000); REG_SET_AND_CHECK(E1000_RCTL, before, 0xFFFFFFFF); REG_PATTERN_TEST(E1000_RDBAL, 0xFFFFFFF0, 0xFFFFFFFF); - if ((mac->type != e1000_ich8lan) && - (mac->type != e1000_ich9lan)) + if (!(adapter->flags & FLAG_IS_ICH)) REG_PATTERN_TEST(E1000_TXCW, 0xC000FFFF, 0x0000FFFF); REG_PATTERN_TEST(E1000_TDBAL, 0xFFFFFFF0, 0xFFFFFFFF); REG_PATTERN_TEST(E1000_TIDV, 0x0000FFFF, 0x0000FFFF); @@ -911,9 +908,7 @@ static int e1000_intr_test(struct e1000_adapter *adapter, u64 *data) /* Test each interrupt */ for (i = 0; i < 10; i++) { - - if (((adapter->hw.mac.type == e1000_ich8lan) || - (adapter->hw.mac.type == e1000_ich9lan)) && i == 8) + if ((adapter->flags & FLAG_IS_ICH) && (i == 8)) continue; /* Interrupt to test */ @@ -1184,6 +1179,7 @@ static int e1000_integrated_phy_loopback(struct e1000_adapter *adapter) struct e1000_hw *hw = &adapter->hw; u32 ctrl_reg = 0; u32 stat_reg = 0; + u16 phy_reg = 0; hw->mac.autoneg = 0; @@ -1211,6 +1207,28 @@ static int e1000_integrated_phy_loopback(struct e1000_adapter *adapter) E1000_CTRL_SPD_100 |/* Force Speed to 100 */ E1000_CTRL_FD); /* Force Duplex to FULL */ break; + case e1000_phy_bm: + /* Set Default MAC Interface speed to 1GB */ + e1e_rphy(hw, PHY_REG(2, 21), &phy_reg); + phy_reg &= ~0x0007; + phy_reg |= 0x006; + e1e_wphy(hw, PHY_REG(2, 21), phy_reg); + /* Assert SW reset for above settings to take effect */ + e1000e_commit_phy(hw); + mdelay(1); + /* Force Full Duplex */ + e1e_rphy(hw, PHY_REG(769, 16), &phy_reg); + e1e_wphy(hw, PHY_REG(769, 16), phy_reg | 0x000C); + /* Set Link Up (in force link) */ + e1e_rphy(hw, PHY_REG(776, 16), &phy_reg); + e1e_wphy(hw, PHY_REG(776, 16), phy_reg | 0x0040); + /* Force Link */ + e1e_rphy(hw, PHY_REG(769, 16), &phy_reg); + e1e_wphy(hw, PHY_REG(769, 16), phy_reg | 0x0040); + /* Set Early Link Enable */ + e1e_rphy(hw, PHY_REG(769, 20), &phy_reg); + e1e_wphy(hw, PHY_REG(769, 20), phy_reg | 0x0400); + /* fall through */ default: /* force 1000, set loopback */ e1e_wphy(hw, PHY_CONTROL, 0x4140); @@ -1224,8 +1242,7 @@ static int e1000_integrated_phy_loopback(struct e1000_adapter *adapter) E1000_CTRL_SPD_1000 |/* Force Speed to 1000 */ E1000_CTRL_FD); /* Force Duplex to FULL */ - if ((adapter->hw.mac.type == e1000_ich8lan) || - (adapter->hw.mac.type == e1000_ich9lan)) + if (adapter->flags & FLAG_IS_ICH) ctrl_reg |= E1000_CTRL_SLU; /* Set Link Up */ } diff --git a/drivers/net/e1000e/hw.h b/drivers/net/e1000e/hw.h index a930e6d9cf02..74f263acb172 100644 --- a/drivers/net/e1000e/hw.h +++ b/drivers/net/e1000e/hw.h @@ -216,6 +216,21 @@ enum e1e_registers { #define IGP01E1000_PHY_LINK_HEALTH 0x13 /* PHY Link Health */ #define IGP02E1000_PHY_POWER_MGMT 0x19 /* Power Management */ #define IGP01E1000_PHY_PAGE_SELECT 0x1F /* Page Select */ +#define BM_PHY_PAGE_SELECT 22 /* Page Select for BM */ +#define IGP_PAGE_SHIFT 5 +#define PHY_REG_MASK 0x1F + +#define BM_WUC_PAGE 800 +#define BM_WUC_ADDRESS_OPCODE 0x11 +#define BM_WUC_DATA_OPCODE 0x12 +#define BM_WUC_ENABLE_PAGE 769 +#define BM_WUC_ENABLE_REG 17 +#define BM_WUC_ENABLE_BIT (1 << 2) +#define BM_WUC_HOST_WU_BIT (1 << 4) + +#define BM_WUC PHY_REG(BM_WUC_PAGE, 1) +#define BM_WUFC PHY_REG(BM_WUC_PAGE, 2) +#define BM_WUS PHY_REG(BM_WUC_PAGE, 3) #define IGP01E1000_PHY_PCS_INIT_REG 0x00B4 #define IGP01E1000_PHY_POLARITY_MASK 0x0078 @@ -331,10 +346,16 @@ enum e1e_registers { #define E1000_DEV_ID_ICH8_IFE_G 0x10C5 #define E1000_DEV_ID_ICH8_IGP_M 0x104D #define E1000_DEV_ID_ICH9_IGP_AMT 0x10BD +#define E1000_DEV_ID_ICH9_IGP_M_AMT 0x10F5 +#define E1000_DEV_ID_ICH9_IGP_M 0x10BF +#define E1000_DEV_ID_ICH9_IGP_M_V 0x10CB #define E1000_DEV_ID_ICH9_IGP_C 0x294C #define E1000_DEV_ID_ICH9_IFE 0x10C0 #define E1000_DEV_ID_ICH9_IFE_GT 0x10C3 #define E1000_DEV_ID_ICH9_IFE_G 0x10C2 +#define E1000_DEV_ID_ICH10_R_BM_LM 0x10CC +#define E1000_DEV_ID_ICH10_R_BM_LF 0x10CD +#define E1000_DEV_ID_ICH10_R_BM_V 0x10CE #define E1000_FUNC_1 1 @@ -378,6 +399,7 @@ enum e1000_phy_type { e1000_phy_gg82563, e1000_phy_igp_3, e1000_phy_ife, + e1000_phy_bm, }; enum e1000_bus_width { diff --git a/drivers/net/e1000e/ich8lan.c b/drivers/net/e1000e/ich8lan.c index 768485dbb2c6..9e38452a738c 100644 --- a/drivers/net/e1000e/ich8lan.c +++ b/drivers/net/e1000e/ich8lan.c @@ -38,6 +38,12 @@ * 82566DM Gigabit Network Connection * 82566MC Gigabit Network Connection * 82566MM Gigabit Network Connection + * 82567LM Gigabit Network Connection + * 82567LF Gigabit Network Connection + * 82567LM-2 Gigabit Network Connection + * 82567LF-2 Gigabit Network Connection + * 82567V-2 Gigabit Network Connection + * 82562GT-3 10/100 Network Connection */ #include @@ -198,6 +204,19 @@ static s32 e1000_init_phy_params_ich8lan(struct e1000_hw *hw) phy->addr = 1; phy->reset_delay_us = 100; + /* + * We may need to do this twice - once for IGP and if that fails, + * we'll set BM func pointers and try again + */ + ret_val = e1000e_determine_phy_address(hw); + if (ret_val) { + hw->phy.ops.write_phy_reg = e1000e_write_phy_reg_bm; + hw->phy.ops.read_phy_reg = e1000e_read_phy_reg_bm; + ret_val = e1000e_determine_phy_address(hw); + if (ret_val) + return ret_val; + } + phy->id = 0; while ((e1000_phy_unknown == e1000e_get_phy_type_from_id(phy->id)) && (i++ < 100)) { @@ -219,6 +238,13 @@ static s32 e1000_init_phy_params_ich8lan(struct e1000_hw *hw) phy->type = e1000_phy_ife; phy->autoneg_mask = E1000_ALL_NOT_GIG; break; + case BME1000_E_PHY_ID: + phy->type = e1000_phy_bm; + phy->autoneg_mask = AUTONEG_ADVERTISE_SPEED_DEFAULT; + hw->phy.ops.read_phy_reg = e1000e_read_phy_reg_bm; + hw->phy.ops.write_phy_reg = e1000e_write_phy_reg_bm; + hw->phy.ops.commit_phy = e1000e_phy_sw_reset; + break; default: return -E1000_ERR_PHY; break; @@ -664,6 +690,7 @@ static s32 e1000_get_phy_info_ich8lan(struct e1000_hw *hw) return e1000_get_phy_info_ife_ich8lan(hw); break; case e1000_phy_igp_3: + case e1000_phy_bm: return e1000e_get_phy_info_igp(hw); break; default: @@ -728,7 +755,7 @@ static s32 e1000_set_d0_lplu_state_ich8lan(struct e1000_hw *hw, bool active) s32 ret_val = 0; u16 data; - if (phy->type != e1000_phy_igp_3) + if (phy->type == e1000_phy_ife) return ret_val; phy_ctrl = er32(PHY_CTRL); @@ -1918,8 +1945,35 @@ static s32 e1000_setup_copper_link_ich8lan(struct e1000_hw *hw) ret_val = e1000e_copper_link_setup_igp(hw); if (ret_val) return ret_val; + } else if (hw->phy.type == e1000_phy_bm) { + ret_val = e1000e_copper_link_setup_m88(hw); + if (ret_val) + return ret_val; } + if (hw->phy.type == e1000_phy_ife) { + ret_val = e1e_rphy(hw, IFE_PHY_MDIX_CONTROL, ®_data); + if (ret_val) + return ret_val; + + reg_data &= ~IFE_PMC_AUTO_MDIX; + + switch (hw->phy.mdix) { + case 1: + reg_data &= ~IFE_PMC_FORCE_MDIX; + break; + case 2: + reg_data |= IFE_PMC_FORCE_MDIX; + break; + case 0: + default: + reg_data |= IFE_PMC_AUTO_MDIX; + break; + } + ret_val = e1e_wphy(hw, IFE_PHY_MDIX_CONTROL, reg_data); + if (ret_val) + return ret_val; + } return e1000e_setup_copper_link(hw); } @@ -2126,6 +2180,31 @@ void e1000e_gig_downshift_workaround_ich8lan(struct e1000_hw *hw) reg_data); } +/** + * e1000e_disable_gig_wol_ich8lan - disable gig during WoL + * @hw: pointer to the HW structure + * + * During S0 to Sx transition, it is possible the link remains at gig + * instead of negotiating to a lower speed. Before going to Sx, set + * 'LPLU Enabled' and 'Gig Disable' to force link speed negotiation + * to a lower speed. + * + * Should only be called for ICH9 devices. + **/ +void e1000e_disable_gig_wol_ich8lan(struct e1000_hw *hw) +{ + u32 phy_ctrl; + + if (hw->mac.type == e1000_ich9lan) { + phy_ctrl = er32(PHY_CTRL); + phy_ctrl |= E1000_PHY_CTRL_D0A_LPLU | + E1000_PHY_CTRL_GBE_DISABLE; + ew32(PHY_CTRL, phy_ctrl); + } + + return; +} + /** * e1000_cleanup_led_ich8lan - Restore the default LED operation * @hw: pointer to the HW structure @@ -2247,6 +2326,7 @@ static struct e1000_nvm_operations ich8_nvm_ops = { struct e1000_info e1000_ich8_info = { .mac = e1000_ich8lan, .flags = FLAG_HAS_WOL + | FLAG_IS_ICH | FLAG_RX_CSUM_ENABLED | FLAG_HAS_CTRLEXT_ON_LOAD | FLAG_HAS_AMT @@ -2262,6 +2342,7 @@ struct e1000_info e1000_ich8_info = { struct e1000_info e1000_ich9_info = { .mac = e1000_ich9lan, .flags = FLAG_HAS_JUMBO_FRAMES + | FLAG_IS_ICH | FLAG_HAS_WOL | FLAG_RX_CSUM_ENABLED | FLAG_HAS_CTRLEXT_ON_LOAD diff --git a/drivers/net/e1000e/netdev.c b/drivers/net/e1000e/netdev.c index 8991ab8911e2..8cbb40f3a506 100644 --- a/drivers/net/e1000e/netdev.c +++ b/drivers/net/e1000e/netdev.c @@ -43,10 +43,11 @@ #include #include #include +#include #include "e1000.h" -#define DRV_VERSION "0.2.1" +#define DRV_VERSION "0.3.3.3-k2" char e1000e_driver_name[] = "e1000e"; const char e1000e_driver_version[] = DRV_VERSION; @@ -340,6 +341,89 @@ no_buffers: } } +/** + * e1000_alloc_jumbo_rx_buffers - Replace used jumbo receive buffers + * @adapter: address of board private structure + * @rx_ring: pointer to receive ring structure + * @cleaned_count: number of buffers to allocate this pass + **/ + +static void e1000_alloc_jumbo_rx_buffers(struct e1000_adapter *adapter, + int cleaned_count) +{ + struct net_device *netdev = adapter->netdev; + struct pci_dev *pdev = adapter->pdev; + struct e1000_rx_desc *rx_desc; + struct e1000_ring *rx_ring = adapter->rx_ring; + struct e1000_buffer *buffer_info; + struct sk_buff *skb; + unsigned int i; + unsigned int bufsz = 256 - + 16 /* for skb_reserve */ - + NET_IP_ALIGN; + + i = rx_ring->next_to_use; + buffer_info = &rx_ring->buffer_info[i]; + + while (cleaned_count--) { + skb = buffer_info->skb; + if (skb) { + skb_trim(skb, 0); + goto check_page; + } + + skb = netdev_alloc_skb(netdev, bufsz); + if (unlikely(!skb)) { + /* Better luck next round */ + adapter->alloc_rx_buff_failed++; + break; + } + + /* Make buffer alignment 2 beyond a 16 byte boundary + * this will result in a 16 byte aligned IP header after + * the 14 byte MAC header is removed + */ + skb_reserve(skb, NET_IP_ALIGN); + + buffer_info->skb = skb; +check_page: + /* allocate a new page if necessary */ + if (!buffer_info->page) { + buffer_info->page = alloc_page(GFP_ATOMIC); + if (unlikely(!buffer_info->page)) { + adapter->alloc_rx_buff_failed++; + break; + } + } + + if (!buffer_info->dma) + buffer_info->dma = pci_map_page(pdev, + buffer_info->page, 0, + PAGE_SIZE, + PCI_DMA_FROMDEVICE); + + rx_desc = E1000_RX_DESC(*rx_ring, i); + rx_desc->buffer_addr = cpu_to_le64(buffer_info->dma); + + if (unlikely(++i == rx_ring->count)) + i = 0; + buffer_info = &rx_ring->buffer_info[i]; + } + + if (likely(rx_ring->next_to_use != i)) { + rx_ring->next_to_use = i; + if (unlikely(i-- == 0)) + i = (rx_ring->count - 1); + + /* Force memory writes to complete before letting h/w + * know there are new descriptors to fetch. (Only + * applicable for weak-ordered memory model archs, + * such as IA-64). */ + wmb(); + writel(i, adapter->hw.hw_addr + rx_ring->tail); + } +} + /** * e1000_clean_rx_irq - Send received data up the network stack; legacy * @adapter: board private structure @@ -782,6 +866,186 @@ next_desc: return cleaned; } +/** + * e1000_consume_page - helper function + **/ +static void e1000_consume_page(struct e1000_buffer *bi, struct sk_buff *skb, + u16 length) +{ + bi->page = NULL; + skb->len += length; + skb->data_len += length; + skb->truesize += length; +} + +/** + * e1000_clean_jumbo_rx_irq - Send received data up the network stack; legacy + * @adapter: board private structure + * + * the return value indicates whether actual cleaning was done, there + * is no guarantee that everything was cleaned + **/ + +static bool e1000_clean_jumbo_rx_irq(struct e1000_adapter *adapter, + int *work_done, int work_to_do) +{ + struct net_device *netdev = adapter->netdev; + struct pci_dev *pdev = adapter->pdev; + struct e1000_ring *rx_ring = adapter->rx_ring; + struct e1000_rx_desc *rx_desc, *next_rxd; + struct e1000_buffer *buffer_info, *next_buffer; + u32 length; + unsigned int i; + int cleaned_count = 0; + bool cleaned = false; + unsigned int total_rx_bytes=0, total_rx_packets=0; + + i = rx_ring->next_to_clean; + rx_desc = E1000_RX_DESC(*rx_ring, i); + buffer_info = &rx_ring->buffer_info[i]; + + while (rx_desc->status & E1000_RXD_STAT_DD) { + struct sk_buff *skb; + u8 status; + + if (*work_done >= work_to_do) + break; + (*work_done)++; + + status = rx_desc->status; + skb = buffer_info->skb; + buffer_info->skb = NULL; + + ++i; + if (i == rx_ring->count) + i = 0; + next_rxd = E1000_RX_DESC(*rx_ring, i); + prefetch(next_rxd); + + next_buffer = &rx_ring->buffer_info[i]; + + cleaned = true; + cleaned_count++; + pci_unmap_page(pdev, buffer_info->dma, PAGE_SIZE, + PCI_DMA_FROMDEVICE); + buffer_info->dma = 0; + + length = le16_to_cpu(rx_desc->length); + + /* errors is only valid for DD + EOP descriptors */ + if (unlikely((status & E1000_RXD_STAT_EOP) && + (rx_desc->errors & E1000_RXD_ERR_FRAME_ERR_MASK))) { + /* recycle both page and skb */ + buffer_info->skb = skb; + /* an error means any chain goes out the window + * too */ + if (rx_ring->rx_skb_top) + dev_kfree_skb(rx_ring->rx_skb_top); + rx_ring->rx_skb_top = NULL; + goto next_desc; + } + +#define rxtop rx_ring->rx_skb_top + if (!(status & E1000_RXD_STAT_EOP)) { + /* this descriptor is only the beginning (or middle) */ + if (!rxtop) { + /* this is the beginning of a chain */ + rxtop = skb; + skb_fill_page_desc(rxtop, 0, buffer_info->page, + 0, length); + } else { + /* this is the middle of a chain */ + skb_fill_page_desc(rxtop, + skb_shinfo(rxtop)->nr_frags, + buffer_info->page, 0, length); + /* re-use the skb, only consumed the page */ + buffer_info->skb = skb; + } + e1000_consume_page(buffer_info, rxtop, length); + goto next_desc; + } else { + if (rxtop) { + /* end of the chain */ + skb_fill_page_desc(rxtop, + skb_shinfo(rxtop)->nr_frags, + buffer_info->page, 0, length); + /* re-use the current skb, we only consumed the + * page */ + buffer_info->skb = skb; + skb = rxtop; + rxtop = NULL; + e1000_consume_page(buffer_info, skb, length); + } else { + /* no chain, got EOP, this buf is the packet + * copybreak to save the put_page/alloc_page */ + if (length <= copybreak && + skb_tailroom(skb) >= length) { + u8 *vaddr; + vaddr = kmap_atomic(buffer_info->page, + KM_SKB_DATA_SOFTIRQ); + memcpy(skb_tail_pointer(skb), vaddr, + length); + kunmap_atomic(vaddr, + KM_SKB_DATA_SOFTIRQ); + /* re-use the page, so don't erase + * buffer_info->page */ + skb_put(skb, length); + } else { + skb_fill_page_desc(skb, 0, + buffer_info->page, 0, + length); + e1000_consume_page(buffer_info, skb, + length); + } + } + } + + /* Receive Checksum Offload XXX recompute due to CRC strip? */ + e1000_rx_checksum(adapter, + (u32)(status) | + ((u32)(rx_desc->errors) << 24), + le16_to_cpu(rx_desc->csum), skb); + + /* probably a little skewed due to removing CRC */ + total_rx_bytes += skb->len; + total_rx_packets++; + + /* eth type trans needs skb->data to point to something */ + if (!pskb_may_pull(skb, ETH_HLEN)) { + ndev_err(netdev, "pskb_may_pull failed.\n"); + dev_kfree_skb(skb); + goto next_desc; + } + + e1000_receive_skb(adapter, netdev, skb, status, + rx_desc->special); + +next_desc: + rx_desc->status = 0; + + /* return some buffers to hardware, one at a time is too slow */ + if (unlikely(cleaned_count >= E1000_RX_BUFFER_WRITE)) { + adapter->alloc_rx_buf(adapter, cleaned_count); + cleaned_count = 0; + } + + /* use prefetched values */ + rx_desc = next_rxd; + buffer_info = next_buffer; + } + rx_ring->next_to_clean = i; + + cleaned_count = e1000_desc_unused(rx_ring); + if (cleaned_count) + adapter->alloc_rx_buf(adapter, cleaned_count); + + adapter->total_rx_bytes += total_rx_bytes; + adapter->total_rx_packets += total_rx_packets; + adapter->net_stats.rx_bytes += total_rx_bytes; + adapter->net_stats.rx_packets += total_rx_packets; + return cleaned; +} + /** * e1000_clean_rx_ring - Free Rx Buffers per Queue * @adapter: board private structure @@ -802,6 +1066,10 @@ static void e1000_clean_rx_ring(struct e1000_adapter *adapter) pci_unmap_single(pdev, buffer_info->dma, adapter->rx_buffer_len, PCI_DMA_FROMDEVICE); + else if (adapter->clean_rx == e1000_clean_jumbo_rx_irq) + pci_unmap_page(pdev, buffer_info->dma, + PAGE_SIZE, + PCI_DMA_FROMDEVICE); else if (adapter->clean_rx == e1000_clean_rx_irq_ps) pci_unmap_single(pdev, buffer_info->dma, adapter->rx_ps_bsize0, @@ -809,6 +1077,11 @@ static void e1000_clean_rx_ring(struct e1000_adapter *adapter) buffer_info->dma = 0; } + if (buffer_info->page) { + put_page(buffer_info->page); + buffer_info->page = NULL; + } + if (buffer_info->skb) { dev_kfree_skb(buffer_info->skb); buffer_info->skb = NULL; @@ -1755,10 +2028,12 @@ static void e1000_setup_rctl(struct e1000_adapter *adapter) * a lot of memory, since we allocate 3 pages at all times * per packet. */ - adapter->rx_ps_pages = 0; pages = PAGE_USE_COUNT(adapter->netdev->mtu); - if ((pages <= 3) && (PAGE_SIZE <= 16384) && (rctl & E1000_RCTL_LPE)) + if (!(adapter->flags & FLAG_IS_ICH) && (pages <= 3) && + (PAGE_SIZE <= 16384) && (rctl & E1000_RCTL_LPE)) adapter->rx_ps_pages = pages; + else + adapter->rx_ps_pages = 0; if (adapter->rx_ps_pages) { /* Configure extra packet-split registers */ @@ -1819,9 +2094,12 @@ static void e1000_configure_rx(struct e1000_adapter *adapter) sizeof(union e1000_rx_desc_packet_split); adapter->clean_rx = e1000_clean_rx_irq_ps; adapter->alloc_rx_buf = e1000_alloc_rx_buffers_ps; + } else if (adapter->netdev->mtu > ETH_FRAME_LEN + ETH_FCS_LEN) { + rdlen = rx_ring->count * sizeof(struct e1000_rx_desc); + adapter->clean_rx = e1000_clean_jumbo_rx_irq; + adapter->alloc_rx_buf = e1000_alloc_jumbo_rx_buffers; } else { - rdlen = rx_ring->count * - sizeof(struct e1000_rx_desc); + rdlen = rx_ring->count * sizeof(struct e1000_rx_desc); adapter->clean_rx = e1000_clean_rx_irq; adapter->alloc_rx_buf = e1000_alloc_rx_buffers; } @@ -1885,8 +2163,21 @@ static void e1000_configure_rx(struct e1000_adapter *adapter) * units), e.g. using jumbo frames when setting to E1000_ERT_2048 */ if ((adapter->flags & FLAG_HAS_ERT) && - (adapter->netdev->mtu > ETH_DATA_LEN)) - ew32(ERT, E1000_ERT_2048); + (adapter->netdev->mtu > ETH_DATA_LEN)) { + u32 rxdctl = er32(RXDCTL(0)); + ew32(RXDCTL(0), rxdctl | 0x3); + ew32(ERT, E1000_ERT_2048 | (1 << 13)); + /* + * With jumbo frames and early-receive enabled, excessive + * C4->C2 latencies result in dropped transactions. + */ + pm_qos_update_requirement(PM_QOS_CPU_DMA_LATENCY, + e1000e_driver_name, 55); + } else { + pm_qos_update_requirement(PM_QOS_CPU_DMA_LATENCY, + e1000e_driver_name, + PM_QOS_DEFAULT_VALUE); + } /* Enable Receives */ ew32(RCTL, rctl); @@ -2155,6 +2446,14 @@ void e1000e_reset(struct e1000_adapter *adapter) /* Allow time for pending master requests to run */ mac->ops.reset_hw(hw); + + /* + * For parts with AMT enabled, let the firmware know + * that the network interface is in control + */ + if ((adapter->flags & FLAG_HAS_AMT) && e1000e_check_mng_mode(hw)) + e1000_get_hw_control(adapter); + ew32(WUC, 0); if (mac->ops.init_hw(hw)) @@ -3469,6 +3768,8 @@ static int e1000_change_mtu(struct net_device *netdev, int new_mtu) * means we reserve 2 more, this pushes us to allocate from the next * larger slab size. * i.e. RXBUFFER_2048 --> size-4096 slab + * However with the new *_jumbo_rx* routines, jumbo receives will use + * fragmented skbs */ if (max_frame <= 256) @@ -3626,6 +3927,9 @@ static int e1000_suspend(struct pci_dev *pdev, pm_message_t state) ew32(CTRL_EXT, ctrl_ext); } + if (adapter->flags & FLAG_IS_ICH) + e1000e_disable_gig_wol_ich8lan(&adapter->hw); + /* Allow time for pending master requests to run */ e1000e_disable_pcie_master(&adapter->hw); @@ -4292,6 +4596,13 @@ static struct pci_device_id e1000_pci_tbl[] = { { PCI_VDEVICE(INTEL, E1000_DEV_ID_ICH9_IFE_GT), board_ich9lan }, { PCI_VDEVICE(INTEL, E1000_DEV_ID_ICH9_IGP_AMT), board_ich9lan }, { PCI_VDEVICE(INTEL, E1000_DEV_ID_ICH9_IGP_C), board_ich9lan }, + { PCI_VDEVICE(INTEL, E1000_DEV_ID_ICH9_IGP_M), board_ich9lan }, + { PCI_VDEVICE(INTEL, E1000_DEV_ID_ICH9_IGP_M_AMT), board_ich9lan }, + { PCI_VDEVICE(INTEL, E1000_DEV_ID_ICH9_IGP_M_V), board_ich9lan }, + + { PCI_VDEVICE(INTEL, E1000_DEV_ID_ICH10_R_BM_LM), board_ich9lan }, + { PCI_VDEVICE(INTEL, E1000_DEV_ID_ICH10_R_BM_LF), board_ich9lan }, + { PCI_VDEVICE(INTEL, E1000_DEV_ID_ICH10_R_BM_V), board_ich9lan }, { } /* terminate list */ }; @@ -4326,7 +4637,9 @@ static int __init e1000_init_module(void) printk(KERN_INFO "%s: Copyright (c) 1999-2008 Intel Corporation.\n", e1000e_driver_name); ret = pci_register_driver(&e1000_driver); - + pm_qos_add_requirement(PM_QOS_CPU_DMA_LATENCY, e1000e_driver_name, + PM_QOS_DEFAULT_VALUE); + return ret; } module_init(e1000_init_module); @@ -4340,6 +4653,7 @@ module_init(e1000_init_module); static void __exit e1000_exit_module(void) { pci_unregister_driver(&e1000_driver); + pm_qos_remove_requirement(PM_QOS_CPU_DMA_LATENCY, e1000e_driver_name); } module_exit(e1000_exit_module); diff --git a/drivers/net/e1000e/phy.c b/drivers/net/e1000e/phy.c index e102332a6bee..b133dcf0e950 100644 --- a/drivers/net/e1000e/phy.c +++ b/drivers/net/e1000e/phy.c @@ -34,6 +34,9 @@ static s32 e1000_get_phy_cfg_done(struct e1000_hw *hw); static s32 e1000_phy_force_speed_duplex(struct e1000_hw *hw); static s32 e1000_set_d0_lplu_state(struct e1000_hw *hw, bool active); static s32 e1000_wait_autoneg(struct e1000_hw *hw); +static u32 e1000_get_phy_addr_for_bm_page(u32 page, u32 reg); +static s32 e1000_access_phy_wakeup_reg_bm(struct e1000_hw *hw, u32 offset, + u16 *data, bool read); /* Cable length tables */ static const u16 e1000_m88_cable_length_table[] = @@ -465,6 +468,10 @@ s32 e1000e_copper_link_setup_m88(struct e1000_hw *hw) if (phy->disable_polarity_correction == 1) phy_data |= M88E1000_PSCR_POLARITY_REVERSAL; + /* Enable downshift on BM (disabled by default) */ + if (phy->type == e1000_phy_bm) + phy_data |= BME1000_PSCR_ENABLE_DOWNSHIFT; + ret_val = e1e_wphy(hw, M88E1000_PHY_SPEC_CTRL, phy_data); if (ret_val) return ret_val; @@ -1776,6 +1783,10 @@ enum e1000_phy_type e1000e_get_phy_type_from_id(u32 phy_id) case IFE_C_E_PHY_ID: phy_type = e1000_phy_ife; break; + case BME1000_E_PHY_ID: + case BME1000_E_PHY_ID_R2: + phy_type = e1000_phy_bm; + break; default: phy_type = e1000_phy_unknown; break; @@ -1783,6 +1794,273 @@ enum e1000_phy_type e1000e_get_phy_type_from_id(u32 phy_id) return phy_type; } +/** + * e1000e_determine_phy_address - Determines PHY address. + * @hw: pointer to the HW structure + * + * This uses a trial and error method to loop through possible PHY + * addresses. It tests each by reading the PHY ID registers and + * checking for a match. + **/ +s32 e1000e_determine_phy_address(struct e1000_hw *hw) +{ + s32 ret_val = -E1000_ERR_PHY_TYPE; + u32 phy_addr= 0; + u32 i = 0; + enum e1000_phy_type phy_type = e1000_phy_unknown; + + do { + for (phy_addr = 0; phy_addr < 4; phy_addr++) { + hw->phy.addr = phy_addr; + e1000e_get_phy_id(hw); + phy_type = e1000e_get_phy_type_from_id(hw->phy.id); + + /* + * If phy_type is valid, break - we found our + * PHY address + */ + if (phy_type != e1000_phy_unknown) { + ret_val = 0; + break; + } + } + i++; + } while ((ret_val != 0) && (i < 100)); + + return ret_val; +} + +/** + * e1000_get_phy_addr_for_bm_page - Retrieve PHY page address + * @page: page to access + * + * Returns the phy address for the page requested. + **/ +static u32 e1000_get_phy_addr_for_bm_page(u32 page, u32 reg) +{ + u32 phy_addr = 2; + + if ((page >= 768) || (page == 0 && reg == 25) || (reg == 31)) + phy_addr = 1; + + return phy_addr; +} + +/** + * e1000e_write_phy_reg_bm - Write BM PHY register + * @hw: pointer to the HW structure + * @offset: register offset to write to + * @data: data to write at register offset + * + * Acquires semaphore, if necessary, then writes the data to PHY register + * at the offset. Release any acquired semaphores before exiting. + **/ +s32 e1000e_write_phy_reg_bm(struct e1000_hw *hw, u32 offset, u16 data) +{ + s32 ret_val; + u32 page_select = 0; + u32 page = offset >> IGP_PAGE_SHIFT; + u32 page_shift = 0; + + /* Page 800 works differently than the rest so it has its own func */ + if (page == BM_WUC_PAGE) { + ret_val = e1000_access_phy_wakeup_reg_bm(hw, offset, &data, + false); + goto out; + } + + ret_val = hw->phy.ops.acquire_phy(hw); + if (ret_val) + goto out; + + hw->phy.addr = e1000_get_phy_addr_for_bm_page(page, offset); + + if (offset > MAX_PHY_MULTI_PAGE_REG) { + /* + * Page select is register 31 for phy address 1 and 22 for + * phy address 2 and 3. Page select is shifted only for + * phy address 1. + */ + if (hw->phy.addr == 1) { + page_shift = IGP_PAGE_SHIFT; + page_select = IGP01E1000_PHY_PAGE_SELECT; + } else { + page_shift = 0; + page_select = BM_PHY_PAGE_SELECT; + } + + /* Page is shifted left, PHY expects (page x 32) */ + ret_val = e1000e_write_phy_reg_mdic(hw, page_select, + (page << page_shift)); + if (ret_val) { + hw->phy.ops.release_phy(hw); + goto out; + } + } + + ret_val = e1000e_write_phy_reg_mdic(hw, MAX_PHY_REG_ADDRESS & offset, + data); + + hw->phy.ops.release_phy(hw); + +out: + return ret_val; +} + +/** + * e1000e_read_phy_reg_bm - Read BM PHY register + * @hw: pointer to the HW structure + * @offset: register offset to be read + * @data: pointer to the read data + * + * Acquires semaphore, if necessary, then reads the PHY register at offset + * and storing the retrieved information in data. Release any acquired + * semaphores before exiting. + **/ +s32 e1000e_read_phy_reg_bm(struct e1000_hw *hw, u32 offset, u16 *data) +{ + s32 ret_val; + u32 page_select = 0; + u32 page = offset >> IGP_PAGE_SHIFT; + u32 page_shift = 0; + + /* Page 800 works differently than the rest so it has its own func */ + if (page == BM_WUC_PAGE) { + ret_val = e1000_access_phy_wakeup_reg_bm(hw, offset, data, + true); + goto out; + } + + ret_val = hw->phy.ops.acquire_phy(hw); + if (ret_val) + goto out; + + hw->phy.addr = e1000_get_phy_addr_for_bm_page(page, offset); + + if (offset > MAX_PHY_MULTI_PAGE_REG) { + /* + * Page select is register 31 for phy address 1 and 22 for + * phy address 2 and 3. Page select is shifted only for + * phy address 1. + */ + if (hw->phy.addr == 1) { + page_shift = IGP_PAGE_SHIFT; + page_select = IGP01E1000_PHY_PAGE_SELECT; + } else { + page_shift = 0; + page_select = BM_PHY_PAGE_SELECT; + } + + /* Page is shifted left, PHY expects (page x 32) */ + ret_val = e1000e_write_phy_reg_mdic(hw, page_select, + (page << page_shift)); + if (ret_val) { + hw->phy.ops.release_phy(hw); + goto out; + } + } + + ret_val = e1000e_read_phy_reg_mdic(hw, MAX_PHY_REG_ADDRESS & offset, + data); + hw->phy.ops.release_phy(hw); + +out: + return ret_val; +} + +/** + * e1000_access_phy_wakeup_reg_bm - Read BM PHY wakeup register + * @hw: pointer to the HW structure + * @offset: register offset to be read or written + * @data: pointer to the data to read or write + * @read: determines if operation is read or write + * + * Acquires semaphore, if necessary, then reads the PHY register at offset + * and storing the retrieved information in data. Release any acquired + * semaphores before exiting. Note that procedure to read the wakeup + * registers are different. It works as such: + * 1) Set page 769, register 17, bit 2 = 1 + * 2) Set page to 800 for host (801 if we were manageability) + * 3) Write the address using the address opcode (0x11) + * 4) Read or write the data using the data opcode (0x12) + * 5) Restore 769_17.2 to its original value + **/ +static s32 e1000_access_phy_wakeup_reg_bm(struct e1000_hw *hw, u32 offset, + u16 *data, bool read) +{ + s32 ret_val; + u16 reg = ((u16)offset) & PHY_REG_MASK; + u16 phy_reg = 0; + u8 phy_acquired = 1; + + + ret_val = hw->phy.ops.acquire_phy(hw); + if (ret_val) { + phy_acquired = 0; + goto out; + } + + /* All operations in this function are phy address 1 */ + hw->phy.addr = 1; + + /* Set page 769 */ + e1000e_write_phy_reg_mdic(hw, IGP01E1000_PHY_PAGE_SELECT, + (BM_WUC_ENABLE_PAGE << IGP_PAGE_SHIFT)); + + ret_val = e1000e_read_phy_reg_mdic(hw, BM_WUC_ENABLE_REG, &phy_reg); + if (ret_val) + goto out; + + /* First clear bit 4 to avoid a power state change */ + phy_reg &= ~(BM_WUC_HOST_WU_BIT); + ret_val = e1000e_write_phy_reg_mdic(hw, BM_WUC_ENABLE_REG, phy_reg); + if (ret_val) + goto out; + + /* Write bit 2 = 1, and clear bit 4 to 769_17 */ + ret_val = e1000e_write_phy_reg_mdic(hw, BM_WUC_ENABLE_REG, + phy_reg | BM_WUC_ENABLE_BIT); + if (ret_val) + goto out; + + /* Select page 800 */ + ret_val = e1000e_write_phy_reg_mdic(hw, IGP01E1000_PHY_PAGE_SELECT, + (BM_WUC_PAGE << IGP_PAGE_SHIFT)); + + /* Write the page 800 offset value using opcode 0x11 */ + ret_val = e1000e_write_phy_reg_mdic(hw, BM_WUC_ADDRESS_OPCODE, reg); + if (ret_val) + goto out; + + if (read) { + /* Read the page 800 value using opcode 0x12 */ + ret_val = e1000e_read_phy_reg_mdic(hw, BM_WUC_DATA_OPCODE, + data); + } else { + /* Read the page 800 value using opcode 0x12 */ + ret_val = e1000e_write_phy_reg_mdic(hw, BM_WUC_DATA_OPCODE, + *data); + } + + if (ret_val) + goto out; + + /* + * Restore 769_17.2 to its original value + * Set page 769 + */ + e1000e_write_phy_reg_mdic(hw, IGP01E1000_PHY_PAGE_SELECT, + (BM_WUC_ENABLE_PAGE << IGP_PAGE_SHIFT)); + + /* Clear 769_17.2 */ + ret_val = e1000e_write_phy_reg_mdic(hw, BM_WUC_ENABLE_REG, phy_reg); + +out: + if (phy_acquired == 1) + hw->phy.ops.release_phy(hw); + return ret_val; +} + /** * e1000e_commit_phy - Soft PHY reset * @hw: pointer to the HW structure From 7ab267d4ecdad3032d6bb31619a2744fc2074b59 Mon Sep 17 00:00:00 2001 From: Jeff Garzik Date: Tue, 6 May 2008 12:16:24 -0400 Subject: [PATCH 16/32] fix warning in drivers/net/appletalk/cops.c MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit drivers/net/appletalk/cops.c: In function ‘cops_reset’: drivers/net/appletalk/cops.c:507: warning: comparison of distinct pointer types lacks a cast by replacing hand-woven msleep() with call to msleep() Signed-off-by: Jeff Garzik --- drivers/net/appletalk/cops.c | 16 +++++----------- 1 file changed, 5 insertions(+), 11 deletions(-) diff --git a/drivers/net/appletalk/cops.c b/drivers/net/appletalk/cops.c index 82e9a5bd0dd2..a0b4c8516073 100644 --- a/drivers/net/appletalk/cops.c +++ b/drivers/net/appletalk/cops.c @@ -499,19 +499,13 @@ static void cops_reset(struct net_device *dev, int sleep) { outb(0, ioaddr+DAYNA_RESET); /* Assert the reset port */ inb(ioaddr+DAYNA_RESET); /* Clear the reset */ - if(sleep) - { - long snap=jiffies; - - /* Let card finish initializing, about 1/3 second */ - while (time_before(jiffies, snap + HZ/3)) - schedule(); - } - else - mdelay(333); + if (sleep) + msleep(333); + else + mdelay(333); } + netif_wake_queue(dev); - return; } static void cops_load (struct net_device *dev) From aa807f79dad3d6a8e9b175d66418b0c5be1d5cd8 Mon Sep 17 00:00:00 2001 From: Gunnar Larisch Date: Mon, 5 May 2008 14:01:28 +0200 Subject: [PATCH 17/32] 3c980-TX needs EXTRA_PREAMBLE The ethernet card 3c980-TX needs a mdio_sync() to initialize the ethernet properly. This is forced by adding an EXTRA_PREAMBLE to its drv_flags. Without this, the driver did not reconnect after a link loss. Signed-off-by: Gunnar Larisch Acked-by: Steffen Klassert Signed-off-by: Jeff Garzik --- drivers/net/3c59x.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/3c59x.c b/drivers/net/3c59x.c index 6f8e7d4cf74d..e93c378b5b6e 100644 --- a/drivers/net/3c59x.c +++ b/drivers/net/3c59x.c @@ -319,7 +319,7 @@ static struct vortex_chip_info { {"3c920B-EMB-WNM (ATI Radeon 9100 IGP)", PCI_USES_MASTER, IS_TORNADO|HAS_MII|HAS_HWCKSM, 128, }, {"3c980 Cyclone", - PCI_USES_MASTER, IS_CYCLONE|HAS_HWCKSM, 128, }, + PCI_USES_MASTER, IS_CYCLONE|HAS_HWCKSM|EXTRA_PREAMBLE, 128, }, {"3c980C Python-T", PCI_USES_MASTER, IS_CYCLONE|HAS_NWAY|HAS_HWCKSM, 128, }, From 1daad055bfc928dfc8590664c455960059421151 Mon Sep 17 00:00:00 2001 From: Paulius Zaleckas Date: Mon, 5 May 2008 14:01:29 +0200 Subject: [PATCH 18/32] 3c59x: use netstats in net_device structure Use net_device_stats from net_device structure instead of local. Signed-off-by: Paulius Zaleckas Acked-by: Steffen Klassert Signed-off-by: Jeff Garzik --- drivers/net/3c59x.c | 71 ++++++++++++++++++++++----------------------- 1 file changed, 35 insertions(+), 36 deletions(-) diff --git a/drivers/net/3c59x.c b/drivers/net/3c59x.c index e93c378b5b6e..2edda8cc7f99 100644 --- a/drivers/net/3c59x.c +++ b/drivers/net/3c59x.c @@ -600,7 +600,6 @@ struct vortex_private { struct sk_buff* tx_skbuff[TX_RING_SIZE]; unsigned int cur_rx, cur_tx; /* The next free ring entry */ unsigned int dirty_rx, dirty_tx; /* The ring entries to be free()ed. */ - struct net_device_stats stats; /* Generic stats */ struct vortex_extra_stats xstats; /* NIC-specific extra stats */ struct sk_buff *tx_skb; /* Packet being eaten by bus master ctrl. */ dma_addr_t tx_skb_dma; /* Allocated DMA address for bus master ctrl DMA. */ @@ -1875,7 +1874,7 @@ static void vortex_tx_timeout(struct net_device *dev) issue_and_wait(dev, TxReset); - vp->stats.tx_errors++; + dev->stats.tx_errors++; if (vp->full_bus_master_tx) { printk(KERN_DEBUG "%s: Resetting the Tx ring pointer.\n", dev->name); if (vp->cur_tx - vp->dirty_tx > 0 && ioread32(ioaddr + DownListPtr) == 0) @@ -1887,7 +1886,7 @@ static void vortex_tx_timeout(struct net_device *dev) iowrite8(PKT_BUF_SZ>>8, ioaddr + TxFreeThreshold); iowrite16(DownUnstall, ioaddr + EL3_CMD); } else { - vp->stats.tx_dropped++; + dev->stats.tx_dropped++; netif_wake_queue(dev); } @@ -1928,8 +1927,8 @@ vortex_error(struct net_device *dev, int status) } dump_tx_ring(dev); } - if (tx_status & 0x14) vp->stats.tx_fifo_errors++; - if (tx_status & 0x38) vp->stats.tx_aborted_errors++; + if (tx_status & 0x14) dev->stats.tx_fifo_errors++; + if (tx_status & 0x38) dev->stats.tx_aborted_errors++; if (tx_status & 0x08) vp->xstats.tx_max_collisions++; iowrite8(0, ioaddr + TxStatus); if (tx_status & 0x30) { /* txJabber or txUnderrun */ @@ -2051,8 +2050,8 @@ vortex_start_xmit(struct sk_buff *skb, struct net_device *dev) if (vortex_debug > 2) printk(KERN_DEBUG "%s: Tx error, status %2.2x.\n", dev->name, tx_status); - if (tx_status & 0x04) vp->stats.tx_fifo_errors++; - if (tx_status & 0x38) vp->stats.tx_aborted_errors++; + if (tx_status & 0x04) dev->stats.tx_fifo_errors++; + if (tx_status & 0x38) dev->stats.tx_aborted_errors++; if (tx_status & 0x30) { issue_and_wait(dev, TxReset); } @@ -2350,7 +2349,7 @@ boomerang_interrupt(int irq, void *dev_id) } else { printk(KERN_DEBUG "boomerang_interrupt: no skb!\n"); } - /* vp->stats.tx_packets++; Counted below. */ + /* dev->stats.tx_packets++; Counted below. */ dirty_tx++; } vp->dirty_tx = dirty_tx; @@ -2409,12 +2408,12 @@ static int vortex_rx(struct net_device *dev) unsigned char rx_error = ioread8(ioaddr + RxErrors); if (vortex_debug > 2) printk(KERN_DEBUG " Rx error: status %2.2x.\n", rx_error); - vp->stats.rx_errors++; - if (rx_error & 0x01) vp->stats.rx_over_errors++; - if (rx_error & 0x02) vp->stats.rx_length_errors++; - if (rx_error & 0x04) vp->stats.rx_frame_errors++; - if (rx_error & 0x08) vp->stats.rx_crc_errors++; - if (rx_error & 0x10) vp->stats.rx_length_errors++; + dev->stats.rx_errors++; + if (rx_error & 0x01) dev->stats.rx_over_errors++; + if (rx_error & 0x02) dev->stats.rx_length_errors++; + if (rx_error & 0x04) dev->stats.rx_frame_errors++; + if (rx_error & 0x08) dev->stats.rx_crc_errors++; + if (rx_error & 0x10) dev->stats.rx_length_errors++; } else { /* The packet length: up to 4.5K!. */ int pkt_len = rx_status & 0x1fff; @@ -2446,7 +2445,7 @@ static int vortex_rx(struct net_device *dev) skb->protocol = eth_type_trans(skb, dev); netif_rx(skb); dev->last_rx = jiffies; - vp->stats.rx_packets++; + dev->stats.rx_packets++; /* Wait a limited time to go to next packet. */ for (i = 200; i >= 0; i--) if ( ! (ioread16(ioaddr + EL3_STATUS) & CmdInProgress)) @@ -2455,7 +2454,7 @@ static int vortex_rx(struct net_device *dev) } else if (vortex_debug > 0) printk(KERN_NOTICE "%s: No memory to allocate a sk_buff of " "size %d.\n", dev->name, pkt_len); - vp->stats.rx_dropped++; + dev->stats.rx_dropped++; } issue_and_wait(dev, RxDiscard); } @@ -2482,12 +2481,12 @@ boomerang_rx(struct net_device *dev) unsigned char rx_error = rx_status >> 16; if (vortex_debug > 2) printk(KERN_DEBUG " Rx error: status %2.2x.\n", rx_error); - vp->stats.rx_errors++; - if (rx_error & 0x01) vp->stats.rx_over_errors++; - if (rx_error & 0x02) vp->stats.rx_length_errors++; - if (rx_error & 0x04) vp->stats.rx_frame_errors++; - if (rx_error & 0x08) vp->stats.rx_crc_errors++; - if (rx_error & 0x10) vp->stats.rx_length_errors++; + dev->stats.rx_errors++; + if (rx_error & 0x01) dev->stats.rx_over_errors++; + if (rx_error & 0x02) dev->stats.rx_length_errors++; + if (rx_error & 0x04) dev->stats.rx_frame_errors++; + if (rx_error & 0x08) dev->stats.rx_crc_errors++; + if (rx_error & 0x10) dev->stats.rx_length_errors++; } else { /* The packet length: up to 4.5K!. */ int pkt_len = rx_status & 0x1fff; @@ -2529,7 +2528,7 @@ boomerang_rx(struct net_device *dev) } netif_rx(skb); dev->last_rx = jiffies; - vp->stats.rx_packets++; + dev->stats.rx_packets++; } entry = (++vp->cur_rx) % RX_RING_SIZE; } @@ -2591,7 +2590,7 @@ vortex_down(struct net_device *dev, int final_down) del_timer_sync(&vp->rx_oom_timer); del_timer_sync(&vp->timer); - /* Turn off statistics ASAP. We update vp->stats below. */ + /* Turn off statistics ASAP. We update dev->stats below. */ iowrite16(StatsDisable, ioaddr + EL3_CMD); /* Disable the receiver and transmitter. */ @@ -2728,7 +2727,7 @@ static struct net_device_stats *vortex_get_stats(struct net_device *dev) update_stats(ioaddr, dev); spin_unlock_irqrestore (&vp->lock, flags); } - return &vp->stats; + return &dev->stats; } /* Update statistics. @@ -2748,18 +2747,18 @@ static void update_stats(void __iomem *ioaddr, struct net_device *dev) /* Unlike the 3c5x9 we need not turn off stats updates while reading. */ /* Switch to the stats window, and read everything. */ EL3WINDOW(6); - vp->stats.tx_carrier_errors += ioread8(ioaddr + 0); - vp->stats.tx_heartbeat_errors += ioread8(ioaddr + 1); - vp->stats.tx_window_errors += ioread8(ioaddr + 4); - vp->stats.rx_fifo_errors += ioread8(ioaddr + 5); - vp->stats.tx_packets += ioread8(ioaddr + 6); - vp->stats.tx_packets += (ioread8(ioaddr + 9)&0x30) << 4; + dev->stats.tx_carrier_errors += ioread8(ioaddr + 0); + dev->stats.tx_heartbeat_errors += ioread8(ioaddr + 1); + dev->stats.tx_window_errors += ioread8(ioaddr + 4); + dev->stats.rx_fifo_errors += ioread8(ioaddr + 5); + dev->stats.tx_packets += ioread8(ioaddr + 6); + dev->stats.tx_packets += (ioread8(ioaddr + 9)&0x30) << 4; /* Rx packets */ ioread8(ioaddr + 7); /* Must read to clear */ /* Don't bother with register 9, an extension of registers 6&7. If we do use the 6&7 values the atomic update assumption above is invalid. */ - vp->stats.rx_bytes += ioread16(ioaddr + 10); - vp->stats.tx_bytes += ioread16(ioaddr + 12); + dev->stats.rx_bytes += ioread16(ioaddr + 10); + dev->stats.tx_bytes += ioread16(ioaddr + 12); /* Extra stats for get_ethtool_stats() */ vp->xstats.tx_multiple_collisions += ioread8(ioaddr + 2); vp->xstats.tx_single_collisions += ioread8(ioaddr + 3); @@ -2767,14 +2766,14 @@ static void update_stats(void __iomem *ioaddr, struct net_device *dev) EL3WINDOW(4); vp->xstats.rx_bad_ssd += ioread8(ioaddr + 12); - vp->stats.collisions = vp->xstats.tx_multiple_collisions + dev->stats.collisions = vp->xstats.tx_multiple_collisions + vp->xstats.tx_single_collisions + vp->xstats.tx_max_collisions; { u8 up = ioread8(ioaddr + 13); - vp->stats.rx_bytes += (up & 0x0f) << 16; - vp->stats.tx_bytes += (up & 0xf0) << 12; + dev->stats.rx_bytes += (up & 0x0f) << 16; + dev->stats.tx_bytes += (up & 0xf0) << 12; } EL3WINDOW(old_window >> 13); From 46fa06170d59b6b9951d09354829d85090f0d911 Mon Sep 17 00:00:00 2001 From: Bruce Robson Date: Fri, 2 May 2008 13:40:53 -0700 Subject: [PATCH 19/32] [netdrvr] eexpress: IPv6 fails - multicast problems Taken from http://bugzilla.kernel.org/show_bug.cgi?id=10577 I was unable to access a computer containing an Intel EtherExpress 16 network card using IPv6. I traced this to failure of neighbour discovery. When I used an "ip -6 neigh add" command, on the computer attempting access, to insert a binding between the IPv6 address of the computer with the Intel EtherExpress 16 network card and the card's ethernet address, I was able to access that computer using IPv6. Neighbour discovery requires working multicast. The driver sources file eexpress.c contains an approximately 30 line function eexp_setup_filter used when loading multicast addresses. I found 3 problems in this function 1) It wrote the number of multicast addresses to the card instead of the number of bytes in the multicast addresses. 2) When loading multiple multicast addresses it loaded the first one provided multiple times instead of loading each one once. 3) The setting of pointer 'data' from 'dmi->dmi_addr' occured before the test for the error situation of 'dmi' being NULL. Correcting these problems allows the computer with the Intel EtherExpress 16 network card to found by IPv6 neighbour discovery. p.s. There is some information on the Intel EtherExpress 16 at http://www.intel.com/support/etherexpress/vintage/sb/cs-013500.htm Datasheet for the Intel 82586 ethernet controller used by the card http://www.datasheetcatalog.com/datasheets_pdf/8/2/5/8/82586.shtml Signed-off-by: Bruce Robson Cc: Jeff Garzik Signed-off-by: Andrew Morton Signed-off-by: Jeff Garzik --- drivers/net/eexpress.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/drivers/net/eexpress.c b/drivers/net/eexpress.c index 2eb82aba4a8b..795c594a4b7c 100644 --- a/drivers/net/eexpress.c +++ b/drivers/net/eexpress.c @@ -202,7 +202,7 @@ static unsigned short start_code[] = { 0x0000,Cmd_MCast, 0x0076, /* link to next command */ #define CONF_NR_MULTICAST 0x44 - 0x0000, /* number of multicast addresses */ + 0x0000, /* number of bytes in multicast address(es) */ #define CONF_MULTICAST 0x46 0x0000, 0x0000, 0x0000, /* some addresses */ 0x0000, 0x0000, 0x0000, @@ -1569,7 +1569,7 @@ static void eexp_hw_init586(struct net_device *dev) static void eexp_setup_filter(struct net_device *dev) { - struct dev_mc_list *dmi = dev->mc_list; + struct dev_mc_list *dmi; unsigned short ioaddr = dev->base_addr; int count = dev->mc_count; int i; @@ -1580,9 +1580,9 @@ static void eexp_setup_filter(struct net_device *dev) } outw(CONF_NR_MULTICAST & ~31, ioaddr+SM_PTR); - outw(count, ioaddr+SHADOW(CONF_NR_MULTICAST)); - for (i = 0; i < count; i++) { - unsigned short *data = (unsigned short *)dmi->dmi_addr; + outw(6*count, ioaddr+SHADOW(CONF_NR_MULTICAST)); + for (i = 0, dmi = dev->mc_list; i < count; i++, dmi = dmi->next) { + unsigned short *data; if (!dmi) { printk(KERN_INFO "%s: too few multicast addresses\n", dev->name); break; @@ -1591,6 +1591,7 @@ static void eexp_setup_filter(struct net_device *dev) printk(KERN_INFO "%s: invalid multicast address length given.\n", dev->name); continue; } + data = (unsigned short *)dmi->dmi_addr; outw((CONF_MULTICAST+(6*i)) & ~31, ioaddr+SM_PTR); outw(data[0], ioaddr+SHADOW(CONF_MULTICAST+(6*i))); outw((CONF_MULTICAST+(6*i)+2) & ~31, ioaddr+SM_PTR); From a86e2cbe263c193a70b2e5c5a0c7e53ed39fc0ad Mon Sep 17 00:00:00 2001 From: Scott Wood Date: Fri, 2 May 2008 13:42:41 -0500 Subject: [PATCH 20/32] fs_enet: Fix a memory leak in fs_enet_mdio_probe There are more memory leaks in the !PPC_CPM_NEW_BINDING case, but that code will disappear soon along with arch/ppc. Reported by Daniel Marjamki at http://bugzilla.kernel.org/show_bug.cgi?id=10591 Signed-off-by: Scott Wood Signed-off-by: Jeff Garzik --- drivers/net/fs_enet/mii-fec.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/fs_enet/mii-fec.c b/drivers/net/fs_enet/mii-fec.c index ba75efc9f5b5..f0014cfbb275 100644 --- a/drivers/net/fs_enet/mii-fec.c +++ b/drivers/net/fs_enet/mii-fec.c @@ -194,7 +194,7 @@ static int __devinit fs_enet_mdio_probe(struct of_device *ofdev, ret = of_address_to_resource(ofdev->node, 0, &res); if (ret) - return ret; + goto out_res; snprintf(new_bus->id, MII_BUS_ID_SIZE, "%x", res.start); @@ -236,6 +236,7 @@ out_free_irqs: kfree(new_bus->irq); out_unmap_regs: iounmap(fec->fecp); +out_res: out_fec: kfree(fec); out_mii: From 01935d7d2c544a5dfc8313f79ed164d45115aa33 Mon Sep 17 00:00:00 2001 From: Don Fry Date: Tue, 29 Apr 2008 13:49:58 -0700 Subject: [PATCH 21/32] pcnet32: delete non NAPI code from driver. Delete the non-napi code from the driver and Kconfig. Tested x86_64. Apply at next open opportunity. Signed-off-by: Don Fry Signed-off-by: Jeff Garzik --- drivers/net/Kconfig | 14 ---------- drivers/net/pcnet32.c | 61 +++---------------------------------------- 2 files changed, 4 insertions(+), 71 deletions(-) diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig index af46341827f2..d27f54a2df77 100644 --- a/drivers/net/Kconfig +++ b/drivers/net/Kconfig @@ -1273,20 +1273,6 @@ config PCNET32 To compile this driver as a module, choose M here. The module will be called pcnet32. -config PCNET32_NAPI - bool "Use RX polling (NAPI)" - depends on PCNET32 - help - NAPI is a new driver API designed to reduce CPU and interrupt load - when the driver is receiving lots of packets from the card. It is - still somewhat experimental and thus not yet enabled by default. - - If your estimated Rx load is 10kpps or more, or if the card will be - deployed on potentially unfriendly networks (e.g. in a firewall), - then say Y here. - - If in doubt, say N. - config AMD8111_ETH tristate "AMD 8111 (new PCI lance) support" depends on NET_PCI && PCI diff --git a/drivers/net/pcnet32.c b/drivers/net/pcnet32.c index 4eb322e5273d..a1c454dbc164 100644 --- a/drivers/net/pcnet32.c +++ b/drivers/net/pcnet32.c @@ -22,12 +22,8 @@ *************************************************************************/ #define DRV_NAME "pcnet32" -#ifdef CONFIG_PCNET32_NAPI -#define DRV_VERSION "1.34-NAPI" -#else -#define DRV_VERSION "1.34" -#endif -#define DRV_RELDATE "14.Aug.2007" +#define DRV_VERSION "1.35" +#define DRV_RELDATE "21.Apr.2008" #define PFX DRV_NAME ": " static const char *const version = @@ -445,30 +441,24 @@ static struct pcnet32_access pcnet32_dwio = { static void pcnet32_netif_stop(struct net_device *dev) { -#ifdef CONFIG_PCNET32_NAPI struct pcnet32_private *lp = netdev_priv(dev); -#endif + dev->trans_start = jiffies; -#ifdef CONFIG_PCNET32_NAPI napi_disable(&lp->napi); -#endif netif_tx_disable(dev); } static void pcnet32_netif_start(struct net_device *dev) { -#ifdef CONFIG_PCNET32_NAPI struct pcnet32_private *lp = netdev_priv(dev); ulong ioaddr = dev->base_addr; u16 val; -#endif + netif_wake_queue(dev); -#ifdef CONFIG_PCNET32_NAPI val = lp->a.read_csr(ioaddr, CSR3); val &= 0x00ff; lp->a.write_csr(ioaddr, CSR3, val); napi_enable(&lp->napi); -#endif } /* @@ -911,11 +901,7 @@ static int pcnet32_loopback_test(struct net_device *dev, uint64_t * data1) rc = 1; /* default to fail */ if (netif_running(dev)) -#ifdef CONFIG_PCNET32_NAPI pcnet32_netif_stop(dev); -#else - pcnet32_close(dev); -#endif spin_lock_irqsave(&lp->lock, flags); lp->a.write_csr(ioaddr, CSR0, CSR0_STOP); /* stop the chip */ @@ -1046,7 +1032,6 @@ static int pcnet32_loopback_test(struct net_device *dev, uint64_t * data1) x = a->read_bcr(ioaddr, 32); /* reset internal loopback */ a->write_bcr(ioaddr, 32, (x & ~0x0002)); -#ifdef CONFIG_PCNET32_NAPI if (netif_running(dev)) { pcnet32_netif_start(dev); pcnet32_restart(dev, CSR0_NORMAL); @@ -1055,16 +1040,6 @@ static int pcnet32_loopback_test(struct net_device *dev, uint64_t * data1) lp->a.write_bcr(ioaddr, 20, 4); /* return to 16bit mode */ } spin_unlock_irqrestore(&lp->lock, flags); -#else - if (netif_running(dev)) { - spin_unlock_irqrestore(&lp->lock, flags); - pcnet32_open(dev); - } else { - pcnet32_purge_rx_ring(dev); - lp->a.write_bcr(ioaddr, 20, 4); /* return to 16bit mode */ - spin_unlock_irqrestore(&lp->lock, flags); - } -#endif return (rc); } /* end pcnet32_loopback_test */ @@ -1270,11 +1245,7 @@ static void pcnet32_rx_entry(struct net_device *dev, } dev->stats.rx_bytes += skb->len; skb->protocol = eth_type_trans(skb, dev); -#ifdef CONFIG_PCNET32_NAPI netif_receive_skb(skb); -#else - netif_rx(skb); -#endif dev->last_rx = jiffies; dev->stats.rx_packets++; return; @@ -1403,7 +1374,6 @@ static int pcnet32_tx(struct net_device *dev) return must_restart; } -#ifdef CONFIG_PCNET32_NAPI static int pcnet32_poll(struct napi_struct *napi, int budget) { struct pcnet32_private *lp = container_of(napi, struct pcnet32_private, napi); @@ -1442,7 +1412,6 @@ static int pcnet32_poll(struct napi_struct *napi, int budget) } return work_done; } -#endif #define PCNET32_REGS_PER_PHY 32 #define PCNET32_MAX_PHYS 32 @@ -1864,9 +1833,7 @@ pcnet32_probe1(unsigned long ioaddr, int shared, struct pci_dev *pdev) /* napi.weight is used in both the napi and non-napi cases */ lp->napi.weight = lp->rx_ring_size / 2; -#ifdef CONFIG_PCNET32_NAPI netif_napi_add(dev, &lp->napi, pcnet32_poll, lp->rx_ring_size / 2); -#endif if (fdx && !(lp->options & PCNET32_PORT_ASEL) && ((cards_found >= MAX_UNITS) || full_duplex[cards_found])) @@ -2297,9 +2264,7 @@ static int pcnet32_open(struct net_device *dev) goto err_free_ring; } -#ifdef CONFIG_PCNET32_NAPI napi_enable(&lp->napi); -#endif /* Re-initialize the PCNET32, and start it when done. */ lp->a.write_csr(ioaddr, 1, (lp->init_dma_addr & 0xffff)); @@ -2623,7 +2588,6 @@ pcnet32_interrupt(int irq, void *dev_id) dev->name, csr0); /* unlike for the lance, there is no restart needed */ } -#ifdef CONFIG_PCNET32_NAPI if (netif_rx_schedule_prep(dev, &lp->napi)) { u16 val; /* set interrupt masks */ @@ -2634,24 +2598,9 @@ pcnet32_interrupt(int irq, void *dev_id) __netif_rx_schedule(dev, &lp->napi); break; } -#else - pcnet32_rx(dev, lp->napi.weight); - if (pcnet32_tx(dev)) { - /* reset the chip to clear the error condition, then restart */ - lp->a.reset(ioaddr); - lp->a.write_csr(ioaddr, CSR4, 0x0915); /* auto tx pad */ - pcnet32_restart(dev, CSR0_START); - netif_wake_queue(dev); - } -#endif csr0 = lp->a.read_csr(ioaddr, CSR0); } -#ifndef CONFIG_PCNET32_NAPI - /* Set interrupt enable. */ - lp->a.write_csr(ioaddr, CSR0, CSR0_INTEN); -#endif - if (netif_msg_intr(lp)) printk(KERN_DEBUG "%s: exiting interrupt, csr0=%#4.4x.\n", dev->name, lp->a.read_csr(ioaddr, CSR0)); @@ -2670,9 +2619,7 @@ static int pcnet32_close(struct net_device *dev) del_timer_sync(&lp->watchdog_timer); netif_stop_queue(dev); -#ifdef CONFIG_PCNET32_NAPI napi_disable(&lp->napi); -#endif spin_lock_irqsave(&lp->lock, flags); From 1b3aa7afb60d34867eea5e73ee943b2a026fc47c Mon Sep 17 00:00:00 2001 From: Alan Cox Date: Tue, 29 Apr 2008 14:29:30 +0100 Subject: [PATCH 22/32] cxgb3: Use CAP_SYS_RAWIO for firmware Otherwise theoretically at least CAP_NET_ADMIN Reload new firmware Wait.. Firmware patches kernel So it should be CAY_SYS_RAWIO - not that I suspect this is in fact a credible attack vector! Signed-off-by: Alan Cox Signed-off-by: Jeff Garzik --- drivers/net/cxgb3/cxgb3_main.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/cxgb3/cxgb3_main.c b/drivers/net/cxgb3/cxgb3_main.c index 05e5f59e87fa..ce949d5fae39 100644 --- a/drivers/net/cxgb3/cxgb3_main.c +++ b/drivers/net/cxgb3/cxgb3_main.c @@ -1894,11 +1894,11 @@ static int cxgb_extension_ioctl(struct net_device *dev, void __user *useraddr) u8 *fw_data; struct ch_mem_range t; - if (!capable(CAP_NET_ADMIN)) + if (!capable(CAP_SYS_RAWIO)) return -EPERM; if (copy_from_user(&t, useraddr, sizeof(t))) return -EFAULT; - + /* Check t.len sanity ? */ fw_data = kmalloc(t.len, GFP_KERNEL); if (!fw_data) return -ENOMEM; From e410553fd35afd6d290b65e02dc501722406377d Mon Sep 17 00:00:00 2001 From: Joakim Tjernlund Date: Tue, 29 Apr 2008 13:03:57 +0200 Subject: [PATCH 23/32] ucc_geth: Don't use RX clock as TX clock. Commit 9fb1e350e16164d56990dde036ae9c0a2fd3f634, ucc_geth: use rx-clock-name and tx-clock-name device tree properties Introduced a typo that made the driver use the RX clock as TX clock, causing massive TX errors. Signed-off-by: Joakim Tjernlund Signed-off-by: Jeff Garzik --- drivers/net/ucc_geth.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ucc_geth.c b/drivers/net/ucc_geth.c index b3bbb2dc5cfa..ca0bdac07a78 100644 --- a/drivers/net/ucc_geth.c +++ b/drivers/net/ucc_geth.c @@ -3922,7 +3922,7 @@ static int ucc_geth_probe(struct of_device* ofdev, const struct of_device_id *ma return -EINVAL; } } else { - prop = of_get_property(np, "rx-clock", NULL); + prop = of_get_property(np, "tx-clock", NULL); if (!prop) { printk(KERN_ERR "ucc_geth: mising tx-clock-name property\n"); From f227ec3ca2b7be449fb2156e82b40cceed87a34a Mon Sep 17 00:00:00 2001 From: "Kok, Auke" Date: Tue, 29 Apr 2008 11:18:55 -0700 Subject: [PATCH 24/32] e1000e: don't return half-read eeprom on error On a read error, e1000e might have returned uninitialized block of eeprom data back to userspace. The convention is that 0xff is "empty", so mark the entire eeprom as empty in case of an error. Signed-off-by: Auke Kok Signed-off-by: Jeff Garzik --- drivers/net/e1000e/ethtool.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/net/e1000e/ethtool.c b/drivers/net/e1000e/ethtool.c index 2bb6da057c40..a14561f40db0 100644 --- a/drivers/net/e1000e/ethtool.c +++ b/drivers/net/e1000e/ethtool.c @@ -494,8 +494,12 @@ static int e1000_get_eeprom(struct net_device *netdev, for (i = 0; i < last_word - first_word + 1; i++) { ret_val = e1000_read_nvm(hw, first_word + i, 1, &eeprom_buff[i]); - if (ret_val) + if (ret_val) { + /* a read error occurred, throw away the + * result */ + memset(eeprom_buff, 0xff, sizeof(eeprom_buff)); break; + } } } From 62ab22278308a40bcb7f4079e9719ab8b7fe11b5 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ilpo=20J=C3=A4rvinen?= Date: Thu, 8 May 2008 01:09:11 -0700 Subject: [PATCH 25/32] tcp FRTO: SACK variant is errorneously used with NewReno MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Note: there's actually another bug in FRTO's SACK variant, which is the causing failure in NewReno case because of the error that's fixed here. I'll fix the SACK case separately (it's a separate bug really, though related, but in order to fix that I need to audit tp->snd_nxt usage a bit). There were two places where SACK variant of FRTO is getting incorrectly used even if SACK wasn't negotiated by the TCP flow. This leads to incorrect setting of frto_highmark with NewReno if a previous recovery was interrupted by another RTO. An eventual fallback to conventional recovery then incorrectly considers one or couple of segments as forward transmissions though they weren't, which then are not LOST marked during fallback making them "non-retransmittable" until the next RTO. In a bad case, those segments are really lost and are the only one left in the window. Thus TCP needs another RTO to continue. The next FRTO, however, could again repeat the same events making the progress of the TCP flow extremely slow. In order for these events to occur at all, FRTO must occur again in FRTOs step 3 while the key segments must be lost as well, which is not too likely in practice. It seems to most frequently with some small devices such as network printers that *seem* to accept TCP segments only in-order. In cases were key segments weren't lost, things get automatically resolved because those wrongly marked segments don't need to be retransmitted in order to continue. I found a reproducer after digging up relevant reports (few reports in total, none at netdev or lkml I know of), some cases seemed to indicate middlebox issues which seems now to be a false assumption some people had made. Bugzilla #10063 _might_ be related. Damon L. Chesser had a reproducable case and was kind enough to tcpdump it for me. With the tcpdump log it was quite trivial to figure out. Signed-off-by: Ilpo Järvinen Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 8ac15a604e08..26c936930e92 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -114,8 +114,6 @@ int sysctl_tcp_abc __read_mostly; #define FLAG_FORWARD_PROGRESS (FLAG_ACKED|FLAG_DATA_SACKED) #define FLAG_ANY_PROGRESS (FLAG_FORWARD_PROGRESS|FLAG_SND_UNA_ADVANCED) -#define IsSackFrto() (sysctl_tcp_frto == 0x2) - #define TCP_REMNANT (TCP_FLAG_FIN|TCP_FLAG_URG|TCP_FLAG_SYN|TCP_FLAG_PSH) #define TCP_HP_BITS (~(TCP_RESERVED_BITS|TCP_FLAG_PSH)) @@ -1686,6 +1684,11 @@ static inline void tcp_reset_reno_sack(struct tcp_sock *tp) tp->sacked_out = 0; } +static int tcp_is_sackfrto(const struct tcp_sock *tp) +{ + return (sysctl_tcp_frto == 0x2) && !tcp_is_reno(tp); +} + /* F-RTO can only be used if TCP has never retransmitted anything other than * head (SACK enhanced variant from Appendix B of RFC4138 is more robust here) */ @@ -1702,7 +1705,7 @@ int tcp_use_frto(struct sock *sk) if (icsk->icsk_mtup.probe_size) return 0; - if (IsSackFrto()) + if (tcp_is_sackfrto(tp)) return 1; /* Avoid expensive walking of rexmit queue if possible */ @@ -1792,7 +1795,7 @@ void tcp_enter_frto(struct sock *sk) /* Earlier loss recovery underway (see RFC4138; Appendix B). * The last condition is necessary at least in tp->frto_counter case. */ - if (IsSackFrto() && (tp->frto_counter || + if (tcp_is_sackfrto(tp) && (tp->frto_counter || ((1 << icsk->icsk_ca_state) & (TCPF_CA_Recovery|TCPF_CA_Loss))) && after(tp->high_seq, tp->snd_una)) { tp->frto_highmark = tp->high_seq; @@ -3124,7 +3127,7 @@ static int tcp_process_frto(struct sock *sk, int flag) return 1; } - if (!IsSackFrto() || tcp_is_reno(tp)) { + if (!tcp_is_sackfrto(tp)) { /* RFC4138 shortcoming in step 2; should also have case c): * ACK isn't duplicate nor advances window, e.g., opposite dir * data, winupdate From c67fa02799bccca3d2e16582493da6d57812ec01 Mon Sep 17 00:00:00 2001 From: "J.H.M. Dassen (Ray)" Date: Thu, 8 May 2008 01:11:04 -0700 Subject: [PATCH 26/32] net/ipv4: correct RFC 1122 section reference in comment RFC 1122 does not have a section 3.1.2.2. The requirement to silently discard datagrams with a bad checksum is in section 3.2.1.2 instead. Addresses http://bugzilla.kernel.org/show_bug.cgi?id=10611 Signed-off-by: J.H.M. Dassen (Ray) Signed-off-by: Andrew Morton Signed-off-by: David S. Miller --- net/ipv4/ip_input.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c index 7b4bad6d572f..ff77a4a7f9ec 100644 --- a/net/ipv4/ip_input.c +++ b/net/ipv4/ip_input.c @@ -397,7 +397,7 @@ int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, iph = ip_hdr(skb); /* - * RFC1122: 3.1.2.2 MUST silently discard any IP frame that fails the checksum. + * RFC1122: 3.2.1.2 MUST silently discard any IP frame that fails the checksum. * * Is the datagram acceptable? * From 7312096454b6cd71267eaa3d0efb408e449e9ff3 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Thu, 8 May 2008 01:13:31 -0700 Subject: [PATCH 27/32] macvlan: Fix memleak on device removal/crash on module removal As noticed by Ben Greear, macvlan crashes the kernel when unloading the module. The reason is that it tries to clean up the macvlan_port pointer on the macvlan device itself instead of the underlying device. A non-NULL pointer is taken as indication that the macvlan_handle_frame_hook is valid, when receiving the next packet on the underlying device it tries to call the NULL hook and crashes. Clean up the macvlan_port on the correct device to fix this. Signed-off-by; Patrick McHardy Tested-by: Ben Greear Signed-off-by: David S. Miller --- drivers/net/macvlan.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c index 2056cfc624dc..c36a03ae9bfb 100644 --- a/drivers/net/macvlan.c +++ b/drivers/net/macvlan.c @@ -450,7 +450,7 @@ static void macvlan_dellink(struct net_device *dev) unregister_netdevice(dev); if (list_empty(&port->vlans)) - macvlan_port_destroy(dev); + macvlan_port_destroy(port->dev); } static struct rtnl_link_ops macvlan_link_ops __read_mostly = { From ef75d49f116bccbb80bccd423ecf3cb86c4509a5 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Thu, 8 May 2008 01:15:21 -0700 Subject: [PATCH 28/32] netfilter: nf_conntrack_sip: restrict RTP expect flushing on error to last request Some Inovaphone PBXs exhibit very stange behaviour: when dialing for example "123", the device sends INVITE requests for "1", "12" and "123" back to back. The first requests will elicit error responses from the receiver, causing the SIP helper to flush the RTP expectations even though we might still see a positive response. Note the sequence number of the last INVITE request that contained a media description and only flush the expectations when receiving a negative response for that sequence number. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- include/linux/netfilter/nf_conntrack_sip.h | 1 + net/netfilter/nf_conntrack_sip.c | 22 +++++++++++++--------- 2 files changed, 14 insertions(+), 9 deletions(-) diff --git a/include/linux/netfilter/nf_conntrack_sip.h b/include/linux/netfilter/nf_conntrack_sip.h index 5da04e586a3f..23aa2ec6b7b7 100644 --- a/include/linux/netfilter/nf_conntrack_sip.h +++ b/include/linux/netfilter/nf_conntrack_sip.h @@ -7,6 +7,7 @@ struct nf_ct_sip_master { unsigned int register_cseq; + unsigned int invite_cseq; }; enum sip_expectation_classes { diff --git a/net/netfilter/nf_conntrack_sip.c b/net/netfilter/nf_conntrack_sip.c index 9f4900069561..2f9bbc058b48 100644 --- a/net/netfilter/nf_conntrack_sip.c +++ b/net/netfilter/nf_conntrack_sip.c @@ -870,6 +870,7 @@ static int process_sdp(struct sk_buff *skb, { enum ip_conntrack_info ctinfo; struct nf_conn *ct = nf_ct_get(skb, &ctinfo); + struct nf_conn_help *help = nfct_help(ct); unsigned int matchoff, matchlen; unsigned int mediaoff, medialen; unsigned int sdpoff; @@ -959,6 +960,9 @@ static int process_sdp(struct sk_buff *skb, if (nf_nat_sdp_session && ct->status & IPS_NAT_MASK) ret = nf_nat_sdp_session(skb, dptr, sdpoff, datalen, &rtp_addr); + if (ret == NF_ACCEPT && i > 0) + help->help.ct_sip_info.invite_cseq = cseq; + return ret; } static int process_invite_response(struct sk_buff *skb, @@ -967,14 +971,14 @@ static int process_invite_response(struct sk_buff *skb, { enum ip_conntrack_info ctinfo; struct nf_conn *ct = nf_ct_get(skb, &ctinfo); + struct nf_conn_help *help = nfct_help(ct); if ((code >= 100 && code <= 199) || (code >= 200 && code <= 299)) return process_sdp(skb, dptr, datalen, cseq); - else { + else if (help->help.ct_sip_info.invite_cseq == cseq) flush_expectations(ct, true); - return NF_ACCEPT; - } + return NF_ACCEPT; } static int process_update_response(struct sk_buff *skb, @@ -983,14 +987,14 @@ static int process_update_response(struct sk_buff *skb, { enum ip_conntrack_info ctinfo; struct nf_conn *ct = nf_ct_get(skb, &ctinfo); + struct nf_conn_help *help = nfct_help(ct); if ((code >= 100 && code <= 199) || (code >= 200 && code <= 299)) return process_sdp(skb, dptr, datalen, cseq); - else { + else if (help->help.ct_sip_info.invite_cseq == cseq) flush_expectations(ct, true); - return NF_ACCEPT; - } + return NF_ACCEPT; } static int process_prack_response(struct sk_buff *skb, @@ -999,14 +1003,14 @@ static int process_prack_response(struct sk_buff *skb, { enum ip_conntrack_info ctinfo; struct nf_conn *ct = nf_ct_get(skb, &ctinfo); + struct nf_conn_help *help = nfct_help(ct); if ((code >= 100 && code <= 199) || (code >= 200 && code <= 299)) return process_sdp(skb, dptr, datalen, cseq); - else { + else if (help->help.ct_sip_info.invite_cseq == cseq) flush_expectations(ct, true); - return NF_ACCEPT; - } + return NF_ACCEPT; } static int process_bye_request(struct sk_buff *skb, From f3261aff35cbc811fee0e23eaea277f1b7286eca Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Thu, 8 May 2008 01:16:04 -0700 Subject: [PATCH 29/32] netfilter: Kconfig: default DCCP/SCTP conntrack support to the protocol config values When conntrack and DCCP/SCTP protocols are enabled, chances are good that people also want DCCP/SCTP conntrack and NAT support. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/netfilter/Kconfig | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/net/netfilter/Kconfig b/net/netfilter/Kconfig index c1fc0f1a641c..aa8d80c35e28 100644 --- a/net/netfilter/Kconfig +++ b/net/netfilter/Kconfig @@ -90,6 +90,7 @@ config NF_CT_PROTO_DCCP tristate 'DCCP protocol connection tracking support (EXPERIMENTAL)' depends on EXPERIMENTAL && NF_CONNTRACK depends on NETFILTER_ADVANCED + default IP_DCCP help With this option enabled, the layer 3 independent connection tracking code will be able to do state tracking on DCCP connections. @@ -104,6 +105,7 @@ config NF_CT_PROTO_SCTP tristate 'SCTP protocol connection tracking support (EXPERIMENTAL)' depends on EXPERIMENTAL && NF_CONNTRACK depends on NETFILTER_ADVANCED + default IP_SCTP help With this option enabled, the layer 3 independent connection tracking code will be able to do state tracking on SCTP connections. @@ -532,6 +534,7 @@ config NETFILTER_XT_MATCH_DCCP tristate '"dccp" protocol match support' depends on NETFILTER_XTABLES depends on NETFILTER_ADVANCED + default IP_DCCP help With this option enabled, you will be able to use the iptables `dccp' match in order to match on DCCP source/destination ports @@ -725,6 +728,7 @@ config NETFILTER_XT_MATCH_SCTP tristate '"sctp" protocol match support (EXPERIMENTAL)' depends on NETFILTER_XTABLES && EXPERIMENTAL depends on NETFILTER_ADVANCED + default IP_SCTP help With this option enabled, you will be able to use the `sctp' match in order to match on SCTP source/destination ports From aca51397d01474f80cab8fc978559b45f2e453ad Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Thu, 8 May 2008 01:24:25 -0700 Subject: [PATCH 30/32] netns: Fix arbitrary net_device-s corruptions on net_ns stop. When a net namespace is destroyed, some devices (those, not killed on ns stop explicitly) are moved back to init_net. The problem, is that this net_ns change has one point of failure - the __dev_alloc_name() may be called if a name collision occurs (and this is easy to trigger). This allocator performs a likely-to-fail GFP_ATOMIC allocation to find a suitable number. Other possible conditions that may cause error (for device being ns local or not registered) are always false in this case. So, when this call fails, the device is unregistered. But this is *not* the right thing to do, since after this the device may be released (and kfree-ed) improperly. E. g. bridges require more actions (sysfs update, timer disarming, etc.), some other devices want to remove their private areas from lists, etc. I. e. arbitrary use-after-free cases may occur. The proposed fix is the following: since the only reason for the dev_change_net_namespace to fail is the name generation, we may give it a unique fall-back name w/o %d-s in it - the dev one, since ifindexes are still unique. So make this change, raise the failure-case printk loglevel to EMERG and replace the unregister_netdevice call with BUG(). [ Use snprintf() -DaveM ] Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/core/dev.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/net/core/dev.c b/net/core/dev.c index d334446a8eaf..4addaf0df96e 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4480,17 +4480,19 @@ static void __net_exit default_device_exit(struct net *net) rtnl_lock(); for_each_netdev_safe(net, dev, next) { int err; + char fb_name[IFNAMSIZ]; /* Ignore unmoveable devices (i.e. loopback) */ if (dev->features & NETIF_F_NETNS_LOCAL) continue; /* Push remaing network devices to init_net */ - err = dev_change_net_namespace(dev, &init_net, "dev%d"); + snprintf(fb_name, IFNAMSIZ, "dev%d", dev->ifindex); + err = dev_change_net_namespace(dev, &init_net, fb_name); if (err) { - printk(KERN_WARNING "%s: failed to move %s to init_net: %d\n", + printk(KERN_EMERG "%s: failed to move %s to init_net: %d\n", __func__, dev->name, err); - unregister_netdevice(dev); + BUG(); } } rtnl_unlock(); From c2ab7ac225e29006b7117d6a9fe8f3be8d98b0c2 Mon Sep 17 00:00:00 2001 From: Oliver Hartkopp Date: Thu, 8 May 2008 02:49:55 -0700 Subject: [PATCH 31/32] can: Fix can_send() handling on dev_queue_xmit() failures The tx packet counting and the local loopback of CAN frames should only happen in the case that the CAN frame has been enqueued to the netdevice tx queue successfully. Thanks to Andre Naujoks for reporting this issue. Signed-off-by: Oliver Hartkopp Signed-off-by: Urs Thuermann Signed-off-by: David S. Miller --- net/can/af_can.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-) diff --git a/net/can/af_can.c b/net/can/af_can.c index 2759b76f731c..7e8ca2836452 100644 --- a/net/can/af_can.c +++ b/net/can/af_can.c @@ -208,6 +208,7 @@ static int can_create(struct net *net, struct socket *sock, int protocol) */ int can_send(struct sk_buff *skb, int loop) { + struct sk_buff *newskb = NULL; int err; if (skb->dev->type != ARPHRD_CAN) { @@ -244,8 +245,7 @@ int can_send(struct sk_buff *skb, int loop) * If the interface is not capable to do loopback * itself, we do it here. */ - struct sk_buff *newskb = skb_clone(skb, GFP_ATOMIC); - + newskb = skb_clone(skb, GFP_ATOMIC); if (!newskb) { kfree_skb(skb); return -ENOMEM; @@ -254,7 +254,6 @@ int can_send(struct sk_buff *skb, int loop) newskb->sk = skb->sk; newskb->ip_summed = CHECKSUM_UNNECESSARY; newskb->pkt_type = PACKET_BROADCAST; - netif_rx(newskb); } } else { /* indication for the CAN driver: no loopback required */ @@ -266,11 +265,20 @@ int can_send(struct sk_buff *skb, int loop) if (err > 0) err = net_xmit_errno(err); + if (err) { + if (newskb) + kfree_skb(newskb); + return err; + } + + if (newskb) + netif_rx(newskb); + /* update statistics */ can_stats.tx_frames++; can_stats.tx_frames_delta++; - return err; + return 0; } EXPORT_SYMBOL(can_send); From e46b66bc42b6b1430b04cc5c207ecb2b2f4553dc Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Thu, 8 May 2008 02:53:17 -0700 Subject: [PATCH 32/32] net: Added ASSERT_RTNL() to dev_open() and dev_close(). dev_open() and dev_close() must be called holding the RTNL, since they call device functions and netdevice notifiers that are promised the RTNL. Signed-off-by: Ben Hutchings Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/net/core/dev.c b/net/core/dev.c index 4addaf0df96e..a1607bc0cd4c 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -994,6 +994,8 @@ int dev_open(struct net_device *dev) { int ret = 0; + ASSERT_RTNL(); + /* * Is it already up? */ @@ -1060,6 +1062,8 @@ int dev_open(struct net_device *dev) */ int dev_close(struct net_device *dev) { + ASSERT_RTNL(); + might_sleep(); if (!(dev->flags & IFF_UP))