2013-04-28 03:17:25

by Tony Prisk

[permalink] [raw]
Subject: [PATCH 0/4] Add support for velocity network driver on platform devices

The first three patches are general tidyup.
Patch #1 just alphabetizes the #includes to make it easier to read.
Patch #2 replaces vptr->dev with vptr->netdev, again for clarity in the code.
Patch #3 replaces the pci dma functions with the generic versions.

I have build tested these patches but don't have a PCI velocity to run-test it.
I can't see that they should introduce any problems as it is all renames and
function swapping.

Patch #4 adds support for the velocity driver on devicetree platform devices.
Binding document included. This patch is compile-tested for PCI, and boot
tested on a VIA APC8750.

I think it would be pertinent to get some tested-by's for PCI users.

Regards
Tony Prisk


Tony Prisk (4):
net: velocity: Alphabetize #includes
net: velocity: Rename vptr->dev to vptr->netdev
net: velocity: Convert to generic dma functions
net: velocity: Add platform device support to VIA velocity driver

.../devicetree/bindings/net/via-velocity.txt | 20 +
drivers/net/ethernet/via/Kconfig | 3 +-
drivers/net/ethernet/via/via-velocity.c | 612 ++++++++++++++------
drivers/net/ethernet/via/via-velocity.h | 35 +-
4 files changed, 480 insertions(+), 190 deletions(-)
create mode 100644 Documentation/devicetree/bindings/net/via-velocity.txt

--
1.7.9.5


2013-04-28 03:16:52

by Tony Prisk

[permalink] [raw]
Subject: [PATCH 1/4] net: velocity: Alphabetize #includes

Due to number of #includes, alphabetize them to help avoid duplicates.

Signed-off-by: Tony Prisk <[email protected]>
---
drivers/net/ethernet/via/via-velocity.c | 48 +++++++++++++++----------------
1 file changed, 24 insertions(+), 24 deletions(-)

diff --git a/drivers/net/ethernet/via/via-velocity.c b/drivers/net/ethernet/via/via-velocity.c
index 1bc7f9fd..eb2a023 100644
--- a/drivers/net/ethernet/via/via-velocity.c
+++ b/drivers/net/ethernet/via/via-velocity.c
@@ -42,40 +42,40 @@
*
*/

-#include <linux/module.h>
-#include <linux/types.h>
#include <linux/bitops.h>
-#include <linux/init.h>
-#include <linux/mm.h>
+#include <linux/crc-ccitt.h>
+#include <linux/crc32.h>
+#include <linux/delay.h>
#include <linux/errno.h>
+#include <linux/etherdevice.h>
+#include <linux/ethtool.h>
+#include <linux/if.h>
+#include <linux/if_arp.h>
+#include <linux/if_vlan.h>
+#include <linux/in.h>
+#include <linux/inetdevice.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
#include <linux/ioport.h>
-#include <linux/pci.h>
+#include <linux/ip.h>
#include <linux/kernel.h>
+#include <linux/mii.h>
+#include <linux/mm.h>
+#include <linux/module.h>
#include <linux/netdevice.h>
-#include <linux/etherdevice.h>
+#include <linux/pci.h>
+#include <linux/proc_fs.h>
+#include <linux/reboot.h>
#include <linux/skbuff.h>
-#include <linux/delay.h>
-#include <linux/timer.h>
#include <linux/slab.h>
-#include <linux/interrupt.h>
#include <linux/string.h>
-#include <linux/wait.h>
-#include <linux/io.h>
-#include <linux/if.h>
-#include <linux/uaccess.h>
-#include <linux/proc_fs.h>
-#include <linux/inetdevice.h>
-#include <linux/reboot.h>
-#include <linux/ethtool.h>
-#include <linux/mii.h>
-#include <linux/in.h>
-#include <linux/if_arp.h>
-#include <linux/if_vlan.h>
-#include <linux/ip.h>
#include <linux/tcp.h>
+#include <linux/timer.h>
+#include <linux/types.h>
+#include <linux/uaccess.h>
#include <linux/udp.h>
-#include <linux/crc-ccitt.h>
-#include <linux/crc32.h>
+#include <linux/wait.h>

#include "via-velocity.h"

--
1.7.9.5

2013-04-28 03:17:23

by Tony Prisk

[permalink] [raw]
Subject: [PATCH 4/4] net: velocity: Add platform device support to VIA velocity driver

Add support for the VIA Velocity network driver to be bound to a
OF created platform device.

Signed-off-by: Tony Prisk <[email protected]>
---
.../devicetree/bindings/net/via-velocity.txt | 20 +
drivers/net/ethernet/via/Kconfig | 3 +-
drivers/net/ethernet/via/via-velocity.c | 453 +++++++++++++++-----
drivers/net/ethernet/via/via-velocity.h | 30 +-
4 files changed, 398 insertions(+), 108 deletions(-)
create mode 100644 Documentation/devicetree/bindings/net/via-velocity.txt

diff --git a/Documentation/devicetree/bindings/net/via-velocity.txt b/Documentation/devicetree/bindings/net/via-velocity.txt
new file mode 100644
index 0000000..b3db469
--- /dev/null
+++ b/Documentation/devicetree/bindings/net/via-velocity.txt
@@ -0,0 +1,20 @@
+* VIA Velocity 10/100/1000 Network Controller
+
+Required properties:
+- compatible : Should be "via,velocity-vt6110"
+- reg : Address and length of the io space
+- interrupts : Should contain the controller interrupt line
+
+Optional properties:
+- no-eeprom : PCI network cards use an external EEPROM to store data. Embedded
+ devices quite often set this data in uboot and do not provide an eeprom.
+ Specify this option if you have no external eeprom.
+
+Examples:
+
[email protected] {
+ compatible = "via,velocity-vt6110";
+ reg = <0xd8004000 0x400>;
+ interrupts = <10>;
+ no-eeprom;
+};
diff --git a/drivers/net/ethernet/via/Kconfig b/drivers/net/ethernet/via/Kconfig
index 68a9ba6..6a87097 100644
--- a/drivers/net/ethernet/via/Kconfig
+++ b/drivers/net/ethernet/via/Kconfig
@@ -5,7 +5,6 @@
config NET_VENDOR_VIA
bool "VIA devices"
default y
- depends on PCI
---help---
If you have a network (Ethernet) card belonging to this class, say Y
and read the Ethernet-HOWTO, available from
@@ -45,7 +44,7 @@ config VIA_RHINE_MMIO

config VIA_VELOCITY
tristate "VIA Velocity support"
- depends on PCI
+ depends on (PCI || USE_OF)
select CRC32
select CRC_CCITT
select NET_CORE
diff --git a/drivers/net/ethernet/via/via-velocity.c b/drivers/net/ethernet/via/via-velocity.c
index a5fe892..f4b5ff1 100644
--- a/drivers/net/ethernet/via/via-velocity.c
+++ b/drivers/net/ethernet/via/via-velocity.c
@@ -65,7 +65,11 @@
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/netdevice.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/of_irq.h>
#include <linux/pci.h>
+#include <linux/platform_device.h>
#include <linux/proc_fs.h>
#include <linux/reboot.h>
#include <linux/skbuff.h>
@@ -84,6 +88,16 @@
static int velocity_nics;
static int msglevel = MSG_LEVEL_INFO;

+static void velocity_set_power_state(struct velocity_info *vptr, char state)
+{
+ void *addr = vptr->mac_regs;
+
+ if (vptr->bustype == BUS_PCI)
+ pci_set_power_state(vptr->pdev, state);
+ else
+ writeb(state, addr + 0x154);
+}
+
/**
* mac_get_cam_mask - Read a CAM mask
* @regs: register block for this velocity
@@ -362,12 +376,24 @@ static struct velocity_info_tbl chip_info_table[] = {
* Describe the PCI device identifiers that we support in this
* device driver. Used for hotplug autoloading.
*/
-static DEFINE_PCI_DEVICE_TABLE(velocity_id_table) = {
+#ifdef CONFIG_PCI
+static DEFINE_PCI_DEVICE_TABLE(velocity_pci_id_table) = {
{ PCI_DEVICE(PCI_VENDOR_ID_VIA, PCI_DEVICE_ID_VIA_612X) },
{ }
};

-MODULE_DEVICE_TABLE(pci, velocity_id_table);
+MODULE_DEVICE_TABLE(pci, velocity_pci_id_table);
+#endif
+
+/**
+ * Describe the OF device identifiers that we support in this
+ * device driver. Used for devicetree nodes.
+ */
+static struct of_device_id velocity_of_ids[] = {
+ { .compatible = "via,velocity-vt6110", .data = &chip_info_table[0] },
+ { /* Sentinel */ },
+};
+MODULE_DEVICE_TABLE(of, velocity_of_ids);

/**
* get_chip_name - identifier to name
@@ -386,29 +412,6 @@ static const char *get_chip_name(enum chip_type chip_id)
}

/**
- * velocity_remove1 - device unplug
- * @pdev: PCI device being removed
- *
- * Device unload callback. Called on an unplug or on module
- * unload for each active device that is present. Disconnects
- * the device from the network layer and frees all the resources
- */
-static void velocity_remove1(struct pci_dev *pdev)
-{
- struct net_device *dev = pci_get_drvdata(pdev);
- struct velocity_info *vptr = netdev_priv(dev);
-
- unregister_netdev(dev);
- iounmap(vptr->mac_regs);
- pci_release_regions(pdev);
- pci_disable_device(pdev);
- pci_set_drvdata(pdev, NULL);
- free_netdev(dev);
-
- velocity_nics--;
-}
-
-/**
* velocity_set_int_opt - parser for integer options
* @opt: pointer to option value
* @val: value the user requested (or -1 for default)
@@ -1179,6 +1182,14 @@ static void mii_init(struct velocity_info *vptr, u32 mii_status)
u16 BMCR;

switch (PHYID_GET_PHY_ID(vptr->phy_id)) {
+ case PHYID_ICPLUS_IP101A:
+ MII_REG_BITS_OFF((ADVERTISE_PAUSE_ASYM | ADVERTISE_PAUSE_CAP), MII_ADVERTISE, vptr->mac_regs);
+ if (vptr->mii_status & VELOCITY_DUPLEX_FULL)
+ MII_REG_BITS_ON(TCSR_ECHODIS, MII_SREVISION, vptr->mac_regs);
+ else
+ MII_REG_BITS_OFF(TCSR_ECHODIS, MII_SREVISION, vptr->mac_regs);
+ MII_REG_BITS_ON(PLED_LALBE, MII_TPISTATUS, vptr->mac_regs);
+ break;
case PHYID_CICADA_CS8201:
/*
* Reset to hardware default
@@ -1351,9 +1362,12 @@ static void velocity_init_registers(struct velocity_info *vptr,
velocity_soft_reset(vptr);
mdelay(5);

- mac_eeprom_reload(regs);
- for (i = 0; i < 6; i++)
- writeb(vptr->netdev->dev_addr[i], &(regs->PAR[i]));
+ if (!vptr->no_eeprom) {
+ mac_eeprom_reload(regs);
+ for (i = 0; i < 6; i++)
+ writeb(vptr->netdev->dev_addr[i],
+ &(regs->PAR[i]));
+ }

/*
* clear Pre_ACPI bit.
@@ -2231,15 +2245,15 @@ static int velocity_open(struct net_device *dev)
goto out;

/* Ensure chip is running */
- pci_set_power_state(vptr->pdev, PCI_D0);
+ velocity_set_power_state(vptr, PCI_D0);

velocity_init_registers(vptr, VELOCITY_INIT_COLD);

- ret = request_irq(vptr->pdev->irq, velocity_intr, IRQF_SHARED,
+ ret = request_irq(dev->irq, velocity_intr, IRQF_SHARED,
dev->name, dev);
if (ret < 0) {
/* Power down the chip */
- pci_set_power_state(vptr->pdev, PCI_D3hot);
+ velocity_set_power_state(vptr, PCI_D3hot);
velocity_free_rings(vptr);
goto out;
}
@@ -2411,7 +2425,7 @@ static int velocity_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
saving then we need to bring the device back up to talk to it */

if (!netif_running(dev))
- pci_set_power_state(vptr->pdev, PCI_D0);
+ velocity_set_power_state(vptr, PCI_D0);

switch (cmd) {
case SIOCGMIIPHY: /* Get address of MII PHY in use. */
@@ -2424,7 +2438,7 @@ static int velocity_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
ret = -EOPNOTSUPP;
}
if (!netif_running(dev))
- pci_set_power_state(vptr->pdev, PCI_D3hot);
+ velocity_set_power_state(vptr, PCI_D3hot);


return ret;
@@ -2490,7 +2504,7 @@ static int velocity_close(struct net_device *dev)
if (vptr->flags & VELOCITY_FLAGS_WOL_ENABLED)
velocity_get_ip(vptr);

- free_irq(vptr->pdev->irq, dev);
+ free_irq(dev->irq, dev);

velocity_free_rings(vptr);

@@ -2629,13 +2643,25 @@ static const struct net_device_ops velocity_netdev_ops = {
* Set up the initial velocity_info struct for the device that has been
* discovered.
*/
-static void velocity_init_info(struct pci_dev *pdev, struct velocity_info *vptr,
- const struct velocity_info_tbl *info)
+static void velocity_init_info(void *pdev,
+ struct velocity_info *vptr,
+ const struct velocity_info_tbl *info,
+ enum velocity_bus_type bustype)
{
+ struct pci_dev *pci_dev;
+ struct platform_device *platform_dev;
+
memset(vptr, 0, sizeof(struct velocity_info));

- vptr->dev = &pdev->dev;
- vptr->pdev = pdev;
+ if (bustype == BUS_PCI) {
+ pci_dev = pdev;
+ vptr->pdev = pci_dev;
+ vptr->dev = &pci_dev->dev;
+ } else {
+ platform_dev = pdev;
+ vptr->dev = &platform_dev->dev;
+ }
+
vptr->chip_id = info->chip_id;
vptr->tx.numq = info->txqueue;
vptr->multicast_limit = MCAM_SIZE;
@@ -2653,8 +2679,6 @@ static void velocity_init_info(struct pci_dev *pdev, struct velocity_info *vptr,
static int velocity_get_pci_info(struct velocity_info *vptr,
struct pci_dev *pdev)
{
- vptr->rev_id = pdev->revision;
-
pci_set_master(pdev);

vptr->ioaddr = pci_resource_start(pdev, 0);
@@ -2676,12 +2700,45 @@ static int velocity_get_pci_info(struct velocity_info *vptr,
dev_err(&pdev->dev, "region #1 is too small.\n");
return -EINVAL;
}
+
vptr->pdev = pdev;

return 0;
}

/**
+ * velocity_get_platform_info - retrieve platform info for device
+ * @vptr: velocity device
+ * @pdev: platform device it matches
+ *
+ * Retrieve the Platform configuration data that interests us
+ */
+static int velocity_get_platform_info(struct velocity_info *vptr,
+ struct platform_device *pdev)
+{
+ int ret;
+ struct resource res;
+
+ if (of_get_property(pdev->dev.of_node, "no-eeprom", NULL))
+ vptr->no_eeprom = 1;
+
+ ret = of_address_to_resource(pdev->dev.of_node, 0, &res);
+ if (ret) {
+ dev_err(&pdev->dev, "unable to find memory address\n");
+ return ret;
+ }
+
+ vptr->memaddr = res.start;
+
+ if (resource_size(&res) < VELOCITY_IO_SIZE) {
+ dev_err(&pdev->dev, "memory region is too small.\n");
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+/**
* velocity_print_info - per driver data
* @vptr: velocity
*
@@ -2705,42 +2762,59 @@ static u32 velocity_get_link(struct net_device *dev)
}

/**
- * velocity_found1 - set up discovered velocity card
+ * velocity_probe - set up discovered velocity device
* @pdev: PCI device
* @ent: PCI device table entry that matched
+ * @bustype: bus that device is connected to
*
* Configure a discovered adapter from scratch. Return a negative
* errno error code on failure paths.
*/
-static int velocity_found1(struct pci_dev *pdev,
- const struct pci_device_id *ent)
+static int velocity_probe(void *pdev,
+ const struct pci_device_id *ent,
+ enum velocity_bus_type bustype)
{
static int first = 1;
- struct net_device *dev;
+ struct pci_dev *pci_dev = NULL;
+ struct platform_device *platform_dev = NULL;
+ struct net_device *netdev;
+ struct device *dev;
+ const struct of_device_id *of_id;
int i;
const char *drv_string;
- const struct velocity_info_tbl *info = &chip_info_table[ent->driver_data];
+ const struct velocity_info_tbl *info;
struct velocity_info *vptr;
struct mac_regs __iomem *regs;
+ int pci = (bustype == BUS_PCI) ? 1 : 0;
int ret = -ENOMEM;

+ if (pci) {
+ pci_dev = pdev;
+ dev = &pci_dev->dev;
+ info = &chip_info_table[ent->driver_data];
+ } else {
+ platform_dev = pdev;
+ dev = &platform_dev->dev;
+ of_id = of_match_device(velocity_of_ids, dev);
+ info = of_id->data;
+ }
+
/* FIXME: this driver, like almost all other ethernet drivers,
* can support more than MAX_UNITS.
*/
if (velocity_nics >= MAX_UNITS) {
- dev_notice(&pdev->dev, "already found %d NICs.\n",
- velocity_nics);
+ dev_notice(dev, "already found %d NICs.\n", velocity_nics);
return -ENODEV;
}

- dev = alloc_etherdev(sizeof(struct velocity_info));
- if (!dev)
+ netdev = alloc_etherdev(sizeof(struct velocity_info));
+ if (!netdev)
goto out;

/* Chain it all together */

- SET_NETDEV_DEV(dev, &pdev->dev);
- vptr = netdev_priv(dev);
+ SET_NETDEV_DEV(netdev, dev);
+ vptr = netdev_priv(netdev);

if (first) {
printk(KERN_INFO "%s Ver. %s\n",
@@ -2750,26 +2824,42 @@ static int velocity_found1(struct pci_dev *pdev,
first = 0;
}

- velocity_init_info(pdev, vptr, info);
+ velocity_init_info(pdev, vptr, info, bustype);
+ vptr->bustype = bustype;
+ vptr->netdev = netdev;

- vptr->netdev = dev;
-
- ret = pci_enable_device(pdev);
- if (ret < 0)
- goto err_free_dev;
+ if (pci) {
+ ret = pci_enable_device(pci_dev);
+ if (ret < 0)
+ goto err_free_dev;
+ netdev->irq = pci_dev->irq;
+ } else {
+ netdev->irq =
+ irq_of_parse_and_map(platform_dev->dev.of_node, 0);

- ret = velocity_get_pci_info(vptr, pdev);
- if (ret < 0) {
- /* error message already printed */
- goto err_disable;
+ if (!netdev->irq) {
+ ret = -EINVAL;
+ goto err_free_dev;
+ }
}

- ret = pci_request_regions(pdev, VELOCITY_NAME);
- if (ret < 0) {
- dev_err(&pdev->dev, "No PCI resources.\n");
- goto err_disable;
+ if (pci) {
+ ret = velocity_get_pci_info(vptr, pci_dev);
+ if (ret < 0)
+ goto err_disable;
+
+ ret = pci_request_regions(pci_dev, VELOCITY_NAME);
+ if (ret < 0) {
+ dev_err(dev, "No PCI resources.\n");
+ goto err_disable;
+ }
+ } else {
+ ret = velocity_get_platform_info(vptr, platform_dev);
+ if (ret < 0)
+ goto err_disable;
}

+
regs = ioremap(vptr->memaddr, VELOCITY_IO_SIZE);
if (regs == NULL) {
ret = -EIO;
@@ -2778,13 +2868,18 @@ static int velocity_found1(struct pci_dev *pdev,

vptr->mac_regs = regs;

+ if (pci)
+ vptr->rev_id = pci_dev->revision;
+ else
+ vptr->rev_id = readb(&regs->CR0Set);
+
mac_wol_reset(regs);

for (i = 0; i < 6; i++)
- dev->dev_addr[i] = readb(&regs->PAR[i]);
+ netdev->dev_addr[i] = readb(&regs->PAR[i]);


- drv_string = dev_driver_string(&pdev->dev);
+ drv_string = dev_driver_string(dev);

velocity_get_options(&vptr->options, velocity_nics, drv_string);

@@ -2805,29 +2900,32 @@ static int velocity_found1(struct pci_dev *pdev,

vptr->phy_id = MII_GET_PHY_ID(vptr->mac_regs);

- dev->netdev_ops = &velocity_netdev_ops;
- dev->ethtool_ops = &velocity_ethtool_ops;
- netif_napi_add(dev, &vptr->napi, velocity_poll, VELOCITY_NAPI_WEIGHT);
+ netdev->netdev_ops = &velocity_netdev_ops;
+ netdev->ethtool_ops = &velocity_ethtool_ops;
+ netif_napi_add(netdev, &vptr->napi, velocity_poll, VELOCITY_NAPI_WEIGHT);

- dev->hw_features = NETIF_F_IP_CSUM | NETIF_F_SG | NETIF_F_HW_VLAN_TX;
- dev->features |= NETIF_F_HW_VLAN_TX | NETIF_F_HW_VLAN_FILTER |
+ netdev->hw_features = NETIF_F_IP_CSUM | NETIF_F_SG | NETIF_F_HW_VLAN_TX;
+ netdev->features |= NETIF_F_HW_VLAN_TX | NETIF_F_HW_VLAN_FILTER |
NETIF_F_HW_VLAN_RX | NETIF_F_IP_CSUM;

- ret = register_netdev(dev);
+ ret = register_netdev(netdev);
if (ret < 0)
goto err_iounmap;

- if (!velocity_get_link(dev)) {
- netif_carrier_off(dev);
+ if (!velocity_get_link(netdev)) {
+ netif_carrier_off(netdev);
vptr->mii_status |= VELOCITY_LINK_FAIL;
}

velocity_print_info(vptr);
- pci_set_drvdata(pdev, dev);
+ if (pci)
+ pci_set_drvdata(pci_dev, netdev);
+ else
+ platform_set_drvdata(platform_dev, netdev);

/* and leave the chip powered down */

- pci_set_power_state(pdev, PCI_D3hot);
+ velocity_set_power_state(vptr, PCI_D3hot);
velocity_nics++;
out:
return ret;
@@ -2835,14 +2933,79 @@ out:
err_iounmap:
iounmap(regs);
err_release_res:
- pci_release_regions(pdev);
+ if (pci)
+ pci_release_regions(pci_dev);
err_disable:
- pci_disable_device(pdev);
+ if (pci)
+ pci_disable_device(pci_dev);
err_free_dev:
- free_netdev(dev);
+ free_netdev(netdev);
goto out;
}

+/**
+ * velocity_remove - device unplug
+ * @pdev: PCI device being removed
+ * @bustype: bus that device is connected to
+ *
+ * Device unload callback. Called on an unplug or on module
+ * unload for each active device that is present. Disconnects
+ * the device from the network layer and frees all the resources
+ */
+static int velocity_remove(void *pdev, enum velocity_bus_type bustype)
+{
+ struct net_device *netdev;
+ struct velocity_info *vptr;
+ int pci = (bustype == BUS_PCI) ? 1 : 0;
+
+ if (pci)
+ netdev = pci_get_drvdata(pdev);
+ else
+ netdev = platform_get_drvdata(pdev);
+
+ vptr = netdev_priv(netdev);
+
+ unregister_netdev(netdev);
+ iounmap(vptr->mac_regs);
+
+ if (pci) {
+ pci_release_regions(pdev);
+ pci_disable_device(pdev);
+ pci_set_drvdata(pdev, NULL);
+ } else {
+ platform_set_drvdata(pdev, NULL);
+ }
+
+ free_netdev(netdev);
+
+ velocity_nics--;
+
+ return 0;
+}
+
+#ifdef CONFIG_PCI
+static int velocity_pci_probe(struct pci_dev *pdev,
+ const struct pci_device_id *ent)
+{
+ return velocity_probe(pdev, ent, BUS_PCI);
+}
+
+static void velocity_pci_remove(struct pci_dev *pdev)
+{
+ velocity_remove(pdev, BUS_PCI);
+}
+#endif
+
+static int velocity_platform_probe(struct platform_device *pdev)
+{
+ return velocity_probe(pdev, NULL, BUS_PLATFORM);
+}
+
+static int velocity_platform_remove(struct platform_device *pdev)
+{
+ return velocity_remove(pdev, BUS_PLATFORM);
+}
+
#ifdef CONFIG_PM
/**
* wol_calc_crc - WOL CRC
@@ -3000,11 +3163,20 @@ static void velocity_save_context(struct velocity_info *vptr, struct velocity_co

}

-static int velocity_suspend(struct pci_dev *pdev, pm_message_t state)
+static int velocity_suspend(void *pdev, pm_message_t state,
+ enum velocity_bus_type bustype)
{
- struct net_device *dev = pci_get_drvdata(pdev);
- struct velocity_info *vptr = netdev_priv(dev);
+ struct net_device *netdev;
+ struct velocity_info *vptr;
unsigned long flags;
+ int pci = (bustype == BUS_PCI) ? 1 : 0;
+
+ if (pci)
+ netdev = pci_get_drvdata(pdev)
+ else
+ netdev = platform_get_drvdata(pdev);
+
+ vptr = netdev_priv(netdev);

if (!netif_running(vptr->netdev))
return 0;
@@ -3012,20 +3184,24 @@ static int velocity_suspend(struct pci_dev *pdev, pm_message_t state)
netif_device_detach(vptr->netdev);

spin_lock_irqsave(&vptr->lock, flags);
- pci_save_state(pdev);
+ if (pci)
+ pci_save_state(pdev);

if (vptr->flags & VELOCITY_FLAGS_WOL_ENABLED) {
velocity_get_ip(vptr);
velocity_save_context(vptr, &vptr->context);
velocity_shutdown(vptr);
velocity_set_wol(vptr);
- pci_enable_wake(pdev, PCI_D3hot, 1);
- pci_set_power_state(pdev, PCI_D3hot);
+ if (pci)
+ pci_enable_wake(pdev, PCI_D3hot, 1);
+ velocity_set_power_state(vptr, PCI_D3hot);
} else {
velocity_save_context(vptr, &vptr->context);
velocity_shutdown(vptr);
- pci_disable_device(pdev);
- pci_set_power_state(pdev, pci_choose_state(pdev, state));
+ if (pci)
+ pci_disable_device(pdev);
+ velocity_set_power_state(vptr,
+ velocity_choose_state(pdev, state));
}

spin_unlock_irqrestore(&vptr->lock, flags);
@@ -3067,19 +3243,30 @@ static void velocity_restore_context(struct velocity_info *vptr, struct velocity
writeb(*((u8 *) (context->mac_reg + i)), ptr + i);
}

-static int velocity_resume(struct pci_dev *pdev)
+static int velocity_resume(void *pdev, enum velocity_bus_type bustype)
{
- struct net_device *dev = pci_get_drvdata(pdev);
- struct velocity_info *vptr = netdev_priv(dev);
+ struct net_device *netdev;
+ struct velocity_info *vptr;
unsigned long flags;
+ int pci = (bustype == BUS_PCI) ? 1 : 0;
int i;

+ if (pci)
+ netdev = pci_get_devdata(pdev);
+ else
+ netdev = platform_get_drvdata(pdev);
+
+ vptr = netdev_priv(netdev);
+
if (!netif_running(vptr->netdev))
return 0;

- pci_set_power_state(pdev, PCI_D0);
- pci_enable_wake(pdev, 0, 0);
- pci_restore_state(pdev);
+ velocity_set_power_state(vptr, PCI_D0);
+
+ if (pci) {
+ pci_enable_wake(pdev, 0, 0);
+ pci_restore_state(pdev);
+ }

mac_wol_reset(vptr->mac_regs);

@@ -3101,23 +3288,66 @@ static int velocity_resume(struct pci_dev *pdev)

return 0;
}
+
+#ifdef CONFIG_PCI
+static int velocity_pci_suspend(struct pci_dev *pdev, pm_message_t state)
+{
+ return velocity_suspend(pdev, state, BUS_PCI);
+};
+
+static int velocity_pci_resume(struct pci_dev *pdev)
+{
+ return velocity_resume(pdev, BUS_PCI);
+};
#endif

+static int velocity_platform_suspend(struct platform_device *pdev)
+{
+ return velocity_suspend(pdev, state, BUS_PLATFORM);
+}
+
+static int velocity_platform_resume(struct platform_device *pdev)
+{
+ return velocity_resume(pdev, BUS_PLATFORM);
+}
+
+static const struct dev_pm_ops velocity_platform_pm = {
+ .suspend = velocity_platform_suspend,
+ .resume = velocity_platform_resume,
+};
+
+#endif /* !CONFIG_PM */
+
/*
* Definition for our device driver. The PCI layer interface
* uses this to handle all our card discover and plugging
*/
-static struct pci_driver velocity_driver = {
+#ifdef CONFIG_PCI
+static struct pci_driver velocity_pci_driver = {
.name = VELOCITY_NAME,
- .id_table = velocity_id_table,
- .probe = velocity_found1,
- .remove = velocity_remove1,
+ .id_table = velocity_pci_id_table,
+ .probe = velocity_pci_probe,
+ .remove = velocity_pci_remove,
#ifdef CONFIG_PM
- .suspend = velocity_suspend,
- .resume = velocity_resume,
+ .suspend = velocity_pci_suspend,
+ .resume = velocity_pci_resume,
#endif
};
+#endif

+static struct platform_driver velocity_platform_driver = {
+ .probe = velocity_platform_probe,
+ .remove = velocity_platform_remove,
+#ifdef CONFIG_PM
+ .suspend = velocity_platform_suspend,
+ .resume = velocity_platform_resume,
+#endif
+ .driver = {
+ .name = "via-velocity",
+ .owner = THIS_MODULE,
+ .of_match_table = velocity_of_ids,
+ },
+};

/**
* velocity_ethtool_up - pre hook for ethtool
@@ -3130,7 +3360,7 @@ static int velocity_ethtool_up(struct net_device *dev)
{
struct velocity_info *vptr = netdev_priv(dev);
if (!netif_running(dev))
- pci_set_power_state(vptr->pdev, PCI_D0);
+ velocity_set_power_state(vptr, PCI_D0);
return 0;
}

@@ -3145,7 +3375,7 @@ static void velocity_ethtool_down(struct net_device *dev)
{
struct velocity_info *vptr = netdev_priv(dev);
if (!netif_running(dev))
- pci_set_power_state(vptr->pdev, PCI_D3hot);
+ velocity_set_power_state(vptr, PCI_D3hot);
}

static int velocity_get_settings(struct net_device *dev,
@@ -3267,7 +3497,11 @@ static void velocity_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo
struct velocity_info *vptr = netdev_priv(dev);
strlcpy(info->driver, VELOCITY_NAME, sizeof(info->driver));
strlcpy(info->version, VELOCITY_VERSION, sizeof(info->version));
- strlcpy(info->bus_info, pci_name(vptr->pdev), sizeof(info->bus_info));
+ if (vptr->bustype == BUS_PCI)
+ strlcpy(info->bus_info, pci_name(vptr->pdev),
+ sizeof(info->bus_info));
+ else
+ strlcpy(info->bus_info, "platform", sizeof(info->bus_info));
}

static void velocity_ethtool_get_wol(struct net_device *dev, struct ethtool_wolinfo *wol)
@@ -3560,9 +3794,15 @@ static int __init velocity_init_module(void)
int ret;

velocity_register_notifier();
- ret = pci_register_driver(&velocity_driver);
+#ifdef CONFIG_PCI
+ ret = pci_register_driver(&velocity_pci_driver);
if (ret < 0)
velocity_unregister_notifier();
+#endif
+ ret = platform_driver_register(&velocity_platform_driver);
+ if (ret < 0)
+ velocity_unregister_notifier();
+
return ret;
}

@@ -3577,7 +3817,10 @@ static int __init velocity_init_module(void)
static void __exit velocity_cleanup_module(void)
{
velocity_unregister_notifier();
- pci_unregister_driver(&velocity_driver);
+#ifdef CONFIG_PCI
+ pci_unregister_driver(&velocity_pci_driver);
+#endif
+ platform_driver_unregister(&velocity_platform_driver);
}

module_init(velocity_init_module);
diff --git a/drivers/net/ethernet/via/via-velocity.h b/drivers/net/ethernet/via/via-velocity.h
index c38bbae..7468765 100644
--- a/drivers/net/ethernet/via/via-velocity.h
+++ b/drivers/net/ethernet/via/via-velocity.h
@@ -1265,7 +1265,7 @@ struct velocity_context {
#define PHYID_VT3216_64BIT 0x000FC600UL
#define PHYID_MARVELL_1000 0x01410C50UL
#define PHYID_MARVELL_1000S 0x01410C40UL
-
+#define PHYID_ICPLUS_IP101A 0x02430C54UL
#define PHYID_REV_ID_MASK 0x0000000FUL

#define PHYID_GET_PHY_ID(i) ((i) & ~PHYID_REV_ID_MASK)
@@ -1433,10 +1433,38 @@ struct velocity_opt {

#define GET_RD_BY_IDX(vptr, idx) (vptr->rd_ring[idx])

+static inline char velocity_choose_state(struct device *dev,
+ pm_message_t state)
+
+{
+ switch (state.event) {
+ case PM_EVENT_ON:
+ return PCI_D0;
+ case PM_EVENT_FREEZE:
+ case PM_EVENT_PRETHAW:
+ /* REVISIT both freeze and pre-thaw "should" use D0 */
+ case PM_EVENT_SUSPEND:
+ case PM_EVENT_HIBERNATE:
+ return PCI_D3hot;
+ default:
+ dev_info(dev, "unrecognized suspend event %d\n", state.event);
+ BUG();
+ }
+
+ return PCI_D0;
+}
+
+enum velocity_bus_type {
+ BUS_PCI,
+ BUS_PLATFORM,
+};
+
struct velocity_info {
struct device *dev;
struct pci_dev *pdev;
struct net_device *netdev;
+ enum velocity_bus_type bustype;
+ int no_eeprom;

unsigned long active_vlans[BITS_TO_LONGS(VLAN_N_VID)];
u8 ip_addr[4];
--
1.7.9.5

2013-04-28 03:17:21

by Tony Prisk

[permalink] [raw]
Subject: [PATCH 2/4] net: velocity: Rename vptr->dev to vptr->netdev

Improve the clarity of the code in preparation for converting the
dma functions to generic versions, which require a struct device *.

This makes it possible to store a 'struct device *dev' in the
velocity_info structure.

Signed-off-by: Tony Prisk <[email protected]>
---
drivers/net/ethernet/via/via-velocity.c | 66 +++++++++++++++----------------
drivers/net/ethernet/via/via-velocity.h | 4 +-
2 files changed, 35 insertions(+), 35 deletions(-)

diff --git a/drivers/net/ethernet/via/via-velocity.c b/drivers/net/ethernet/via/via-velocity.c
index eb2a023..1168666 100644
--- a/drivers/net/ethernet/via/via-velocity.c
+++ b/drivers/net/ethernet/via/via-velocity.c
@@ -996,9 +996,9 @@ static void velocity_print_link_status(struct velocity_info *vptr)
{

if (vptr->mii_status & VELOCITY_LINK_FAIL) {
- VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: failed to detect cable link\n", vptr->dev->name);
+ VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: failed to detect cable link\n", vptr->netdev->name);
} else if (vptr->options.spd_dpx == SPD_DPX_AUTO) {
- VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: Link auto-negotiation", vptr->dev->name);
+ VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: Link auto-negotiation", vptr->netdev->name);

if (vptr->mii_status & VELOCITY_SPEED_1000)
VELOCITY_PRT(MSG_LEVEL_INFO, " speed 1000M bps");
@@ -1012,7 +1012,7 @@ static void velocity_print_link_status(struct velocity_info *vptr)
else
VELOCITY_PRT(MSG_LEVEL_INFO, " half duplex\n");
} else {
- VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: Link forced", vptr->dev->name);
+ VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: Link forced", vptr->netdev->name);
switch (vptr->options.spd_dpx) {
case SPD_DPX_1000_FULL:
VELOCITY_PRT(MSG_LEVEL_INFO, " speed 1000M bps full duplex\n");
@@ -1317,7 +1317,7 @@ static void velocity_init_registers(struct velocity_info *vptr,
case VELOCITY_INIT_RESET:
case VELOCITY_INIT_WOL:

- netif_stop_queue(vptr->dev);
+ netif_stop_queue(vptr->netdev);

/*
* Reset RX to prevent RX pointer not on the 4X location
@@ -1330,7 +1330,7 @@ static void velocity_init_registers(struct velocity_info *vptr,
if (velocity_set_media_mode(vptr, mii_status) != VELOCITY_LINK_CHANGE) {
velocity_print_link_status(vptr);
if (!(vptr->mii_status & VELOCITY_LINK_FAIL))
- netif_wake_queue(vptr->dev);
+ netif_wake_queue(vptr->netdev);
}

enable_flow_control_ability(vptr);
@@ -1352,7 +1352,7 @@ static void velocity_init_registers(struct velocity_info *vptr,

mac_eeprom_reload(regs);
for (i = 0; i < 6; i++)
- writeb(vptr->dev->dev_addr[i], &(regs->PAR[i]));
+ writeb(vptr->netdev->dev_addr[i], &(regs->PAR[i]));

/*
* clear Pre_ACPI bit.
@@ -1375,7 +1375,7 @@ static void velocity_init_registers(struct velocity_info *vptr,
/*
* Set packet filter: Receive directed and broadcast address
*/
- velocity_set_multi(vptr->dev);
+ velocity_set_multi(vptr->netdev);

/*
* Enable MII auto-polling
@@ -1402,14 +1402,14 @@ static void velocity_init_registers(struct velocity_info *vptr,
writel((CR0_DPOLL | CR0_TXON | CR0_RXON | CR0_STRT), &regs->CR0Set);

mii_status = velocity_get_opt_media_mode(vptr);
- netif_stop_queue(vptr->dev);
+ netif_stop_queue(vptr->netdev);

mii_init(vptr, mii_status);

if (velocity_set_media_mode(vptr, mii_status) != VELOCITY_LINK_CHANGE) {
velocity_print_link_status(vptr);
if (!(vptr->mii_status & VELOCITY_LINK_FAIL))
- netif_wake_queue(vptr->dev);
+ netif_wake_queue(vptr->netdev);
}

enable_flow_control_ability(vptr);
@@ -1472,7 +1472,7 @@ static int velocity_init_dma_rings(struct velocity_info *vptr)
rx_ring_size, &pool_dma);
if (!pool) {
dev_err(&pdev->dev, "%s : DMA memory allocation failed.\n",
- vptr->dev->name);
+ vptr->netdev->name);
return -ENOMEM;
}

@@ -1512,7 +1512,7 @@ static int velocity_alloc_rx_buf(struct velocity_info *vptr, int idx)
struct rx_desc *rd = &(vptr->rx.ring[idx]);
struct velocity_rd_info *rd_info = &(vptr->rx.info[idx]);

- rd_info->skb = netdev_alloc_skb(vptr->dev, vptr->rx.buf_sz + 64);
+ rd_info->skb = netdev_alloc_skb(vptr->netdev, vptr->rx.buf_sz + 64);
if (rd_info->skb == NULL)
return -ENOMEM;

@@ -1618,7 +1618,7 @@ static int velocity_init_rd_ring(struct velocity_info *vptr)

if (velocity_rx_refill(vptr) != vptr->options.numrx) {
VELOCITY_PRT(MSG_LEVEL_ERR, KERN_ERR
- "%s: failed to allocate RX buffer.\n", vptr->dev->name);
+ "%s: failed to allocate RX buffer.\n", vptr->netdev->name);
velocity_free_rd_ring(vptr);
goto out;
}
@@ -1807,7 +1807,7 @@ static void velocity_error(struct velocity_info *vptr, int status)
printk(KERN_ERR "TD structure error TDindex=%hx\n", readw(&regs->TDIdx[0]));
BYTE_REG_BITS_ON(TXESR_TDSTR, &regs->TXESR);
writew(TRDCSR_RUN, &regs->TDCSRClr);
- netif_stop_queue(vptr->dev);
+ netif_stop_queue(vptr->netdev);

/* FIXME: port over the pci_device_failed code and use it
here */
@@ -1848,10 +1848,10 @@ static void velocity_error(struct velocity_info *vptr, int status)

if (linked) {
vptr->mii_status &= ~VELOCITY_LINK_FAIL;
- netif_carrier_on(vptr->dev);
+ netif_carrier_on(vptr->netdev);
} else {
vptr->mii_status |= VELOCITY_LINK_FAIL;
- netif_carrier_off(vptr->dev);
+ netif_carrier_off(vptr->netdev);
}

velocity_print_link_status(vptr);
@@ -1865,9 +1865,9 @@ static void velocity_error(struct velocity_info *vptr, int status)
enable_mii_autopoll(regs);

if (vptr->mii_status & VELOCITY_LINK_FAIL)
- netif_stop_queue(vptr->dev);
+ netif_stop_queue(vptr->netdev);
else
- netif_wake_queue(vptr->dev);
+ netif_wake_queue(vptr->netdev);

}
if (status & ISR_MIBFI)
@@ -1892,7 +1892,7 @@ static int velocity_tx_srv(struct velocity_info *vptr)
int idx;
int works = 0;
struct velocity_td_info *tdinfo;
- struct net_device_stats *stats = &vptr->dev->stats;
+ struct net_device_stats *stats = &vptr->netdev->stats;

for (qnum = 0; qnum < vptr->tx.numq; qnum++) {
for (idx = vptr->tx.tail[qnum]; vptr->tx.used[qnum] > 0;
@@ -1937,9 +1937,9 @@ static int velocity_tx_srv(struct velocity_info *vptr)
* Look to see if we should kick the transmit network
* layer for more work.
*/
- if (netif_queue_stopped(vptr->dev) && (full == 0) &&
+ if (netif_queue_stopped(vptr->netdev) && (full == 0) &&
(!(vptr->mii_status & VELOCITY_LINK_FAIL))) {
- netif_wake_queue(vptr->dev);
+ netif_wake_queue(vptr->netdev);
}
return works;
}
@@ -1987,7 +1987,7 @@ static int velocity_rx_copy(struct sk_buff **rx_skb, int pkt_size,
if (pkt_size < rx_copybreak) {
struct sk_buff *new_skb;

- new_skb = netdev_alloc_skb_ip_align(vptr->dev, pkt_size);
+ new_skb = netdev_alloc_skb_ip_align(vptr->netdev, pkt_size);
if (new_skb) {
new_skb->ip_summed = rx_skb[0]->ip_summed;
skb_copy_from_linear_data(*rx_skb, new_skb->data, pkt_size);
@@ -2028,14 +2028,14 @@ static inline void velocity_iph_realign(struct velocity_info *vptr,
static int velocity_receive_frame(struct velocity_info *vptr, int idx)
{
void (*pci_action)(struct pci_dev *, dma_addr_t, size_t, int);
- struct net_device_stats *stats = &vptr->dev->stats;
+ struct net_device_stats *stats = &vptr->netdev->stats;
struct velocity_rd_info *rd_info = &(vptr->rx.info[idx]);
struct rx_desc *rd = &(vptr->rx.ring[idx]);
int pkt_len = le16_to_cpu(rd->rdesc0.len) & 0x3fff;
struct sk_buff *skb;

if (rd->rdesc0.RSR & (RSR_STP | RSR_EDP)) {
- VELOCITY_PRT(MSG_LEVEL_VERBOSE, KERN_ERR " %s : the received frame span multple RDs.\n", vptr->dev->name);
+ VELOCITY_PRT(MSG_LEVEL_VERBOSE, KERN_ERR " %s : the received frame span multple RDs.\n", vptr->netdev->name);
stats->rx_length_errors++;
return -EINVAL;
}
@@ -2073,7 +2073,7 @@ static int velocity_receive_frame(struct velocity_info *vptr, int idx)
PCI_DMA_FROMDEVICE);

skb_put(skb, pkt_len - 4);
- skb->protocol = eth_type_trans(skb, vptr->dev);
+ skb->protocol = eth_type_trans(skb, vptr->netdev);

if (rd->rdesc0.RSR & RSR_DETAG) {
u16 vid = swab16(le16_to_cpu(rd->rdesc1.PQTAG));
@@ -2098,7 +2098,7 @@ static int velocity_receive_frame(struct velocity_info *vptr, int idx)
*/
static int velocity_rx_srv(struct velocity_info *vptr, int budget_left)
{
- struct net_device_stats *stats = &vptr->dev->stats;
+ struct net_device_stats *stats = &vptr->netdev->stats;
int rd_curr = vptr->rx.curr;
int works = 0;

@@ -2290,7 +2290,7 @@ static int velocity_change_mtu(struct net_device *dev, int new_mtu)

if ((new_mtu < VELOCITY_MIN_MTU) || new_mtu > (VELOCITY_MAX_MTU)) {
VELOCITY_PRT(MSG_LEVEL_ERR, KERN_NOTICE "%s: Invalid MTU.\n",
- vptr->dev->name);
+ vptr->netdev->name);
ret = -EINVAL;
goto out_0;
}
@@ -2312,7 +2312,7 @@ static int velocity_change_mtu(struct net_device *dev, int new_mtu)
goto out_0;
}

- tmp_vptr->dev = dev;
+ tmp_vptr->netdev = dev;
tmp_vptr->pdev = vptr->pdev;
tmp_vptr->options = vptr->options;
tmp_vptr->tx.numq = vptr->tx.numq;
@@ -2690,7 +2690,7 @@ static int velocity_get_pci_info(struct velocity_info *vptr,
*/
static void velocity_print_info(struct velocity_info *vptr)
{
- struct net_device *dev = vptr->dev;
+ struct net_device *dev = vptr->netdev;

printk(KERN_INFO "%s: %s\n", dev->name, get_chip_name(vptr->chip_id));
printk(KERN_INFO "%s: Ethernet Address: %pM\n",
@@ -2753,7 +2753,7 @@ static int velocity_found1(struct pci_dev *pdev,

velocity_init_info(pdev, vptr, info);

- vptr->dev = dev;
+ vptr->netdev = dev;

ret = pci_enable_device(pdev);
if (ret < 0)
@@ -3007,10 +3007,10 @@ static int velocity_suspend(struct pci_dev *pdev, pm_message_t state)
struct velocity_info *vptr = netdev_priv(dev);
unsigned long flags;

- if (!netif_running(vptr->dev))
+ if (!netif_running(vptr->netdev))
return 0;

- netif_device_detach(vptr->dev);
+ netif_device_detach(vptr->netdev);

spin_lock_irqsave(&vptr->lock, flags);
pci_save_state(pdev);
@@ -3075,7 +3075,7 @@ static int velocity_resume(struct pci_dev *pdev)
unsigned long flags;
int i;

- if (!netif_running(vptr->dev))
+ if (!netif_running(vptr->netdev))
return 0;

pci_set_power_state(pdev, PCI_D0);
@@ -3098,7 +3098,7 @@ static int velocity_resume(struct pci_dev *pdev)

mac_enable_int(vptr->mac_regs);
spin_unlock_irqrestore(&vptr->lock, flags);
- netif_device_attach(vptr->dev);
+ netif_device_attach(vptr->netdev);

return 0;
}
diff --git a/drivers/net/ethernet/via/via-velocity.h b/drivers/net/ethernet/via/via-velocity.h
index 4cb9f13..ff8d7828 100644
--- a/drivers/net/ethernet/via/via-velocity.h
+++ b/drivers/net/ethernet/via/via-velocity.h
@@ -1435,7 +1435,7 @@ struct velocity_opt {

struct velocity_info {
struct pci_dev *pdev;
- struct net_device *dev;
+ struct net_device *netdev;

unsigned long active_vlans[BITS_TO_LONGS(VLAN_N_VID)];
u8 ip_addr[4];
@@ -1514,7 +1514,7 @@ static inline int velocity_get_ip(struct velocity_info *vptr)
int res = -ENOENT;

rcu_read_lock();
- in_dev = __in_dev_get_rcu(vptr->dev);
+ in_dev = __in_dev_get_rcu(vptr->netdev);
if (in_dev != NULL) {
ifa = (struct in_ifaddr *) in_dev->ifa_list;
if (ifa != NULL) {
--
1.7.9.5

2013-04-28 03:17:19

by Tony Prisk

[permalink] [raw]
Subject: [PATCH 3/4] net: velocity: Convert to generic dma functions

Remove the pci_* dma functions and replace with the more generic
versions.

In preparation of adding platform support, a new struct device *dev
is added to struct velocity_info which can be used by both the pci
and platform code.

Signed-off-by: Tony Prisk <[email protected]>
---
drivers/net/ethernet/via/via-velocity.c | 51 +++++++++++++++----------------
drivers/net/ethernet/via/via-velocity.h | 1 +
2 files changed, 26 insertions(+), 26 deletions(-)

diff --git a/drivers/net/ethernet/via/via-velocity.c b/drivers/net/ethernet/via/via-velocity.c
index 1168666..a5fe892 100644
--- a/drivers/net/ethernet/via/via-velocity.c
+++ b/drivers/net/ethernet/via/via-velocity.c
@@ -46,6 +46,7 @@
#include <linux/crc-ccitt.h>
#include <linux/crc32.h>
#include <linux/delay.h>
+#include <linux/dma-mapping.h>
#include <linux/errno.h>
#include <linux/etherdevice.h>
#include <linux/ethtool.h>
@@ -1457,7 +1458,6 @@ static int velocity_init_dma_rings(struct velocity_info *vptr)
struct velocity_opt *opt = &vptr->options;
const unsigned int rx_ring_size = opt->numrx * sizeof(struct rx_desc);
const unsigned int tx_ring_size = opt->numtx * sizeof(struct tx_desc);
- struct pci_dev *pdev = vptr->pdev;
dma_addr_t pool_dma;
void *pool;
unsigned int i;
@@ -1465,13 +1465,13 @@ static int velocity_init_dma_rings(struct velocity_info *vptr)
/*
* Allocate all RD/TD rings a single pool.
*
- * pci_alloc_consistent() fulfills the requirement for 64 bytes
+ * dma_alloc_coherent() fulfills the requirement for 64 bytes
* alignment
*/
- pool = pci_alloc_consistent(pdev, tx_ring_size * vptr->tx.numq +
- rx_ring_size, &pool_dma);
+ pool = dma_alloc_coherent(vptr->dev, tx_ring_size * vptr->tx.numq +
+ rx_ring_size, &pool_dma, GFP_ATOMIC);
if (!pool) {
- dev_err(&pdev->dev, "%s : DMA memory allocation failed.\n",
+ dev_err(vptr->dev, "%s : DMA memory allocation failed.\n",
vptr->netdev->name);
return -ENOMEM;
}
@@ -1522,8 +1522,8 @@ static int velocity_alloc_rx_buf(struct velocity_info *vptr, int idx)
*/
skb_reserve(rd_info->skb,
64 - ((unsigned long) rd_info->skb->data & 63));
- rd_info->skb_dma = pci_map_single(vptr->pdev, rd_info->skb->data,
- vptr->rx.buf_sz, PCI_DMA_FROMDEVICE);
+ rd_info->skb_dma = dma_map_single(vptr->dev, rd_info->skb->data,
+ vptr->rx.buf_sz, DMA_FROM_DEVICE);

/*
* Fill in the descriptor to match
@@ -1586,8 +1586,8 @@ static void velocity_free_rd_ring(struct velocity_info *vptr)

if (!rd_info->skb)
continue;
- pci_unmap_single(vptr->pdev, rd_info->skb_dma, vptr->rx.buf_sz,
- PCI_DMA_FROMDEVICE);
+ dma_unmap_single(vptr->dev, rd_info->skb_dma, vptr->rx.buf_sz,
+ DMA_FROM_DEVICE);
rd_info->skb_dma = 0;

dev_kfree_skb(rd_info->skb);
@@ -1668,7 +1668,7 @@ static void velocity_free_dma_rings(struct velocity_info *vptr)
const int size = vptr->options.numrx * sizeof(struct rx_desc) +
vptr->options.numtx * sizeof(struct tx_desc) * vptr->tx.numq;

- pci_free_consistent(vptr->pdev, size, vptr->rx.ring, vptr->rx.pool_dma);
+ dma_free_coherent(vptr->dev, size, vptr->rx.ring, vptr->rx.pool_dma);
}

static int velocity_init_rings(struct velocity_info *vptr, int mtu)
@@ -1725,8 +1725,8 @@ static void velocity_free_tx_buf(struct velocity_info *vptr,
pktlen = max_t(size_t, pktlen,
td->td_buf[i].size & ~TD_QUEUE);

- pci_unmap_single(vptr->pdev, tdinfo->skb_dma[i],
- le16_to_cpu(pktlen), PCI_DMA_TODEVICE);
+ dma_unmap_single(vptr->dev, tdinfo->skb_dma[i],
+ le16_to_cpu(pktlen), DMA_TO_DEVICE);
}
}
dev_kfree_skb_irq(skb);
@@ -1748,8 +1748,8 @@ static void velocity_free_td_ring_entry(struct velocity_info *vptr,
if (td_info->skb) {
for (i = 0; i < td_info->nskb_dma; i++) {
if (td_info->skb_dma[i]) {
- pci_unmap_single(vptr->pdev, td_info->skb_dma[i],
- td_info->skb->len, PCI_DMA_TODEVICE);
+ dma_unmap_single(vptr->dev, td_info->skb_dma[i],
+ td_info->skb->len, DMA_TO_DEVICE);
td_info->skb_dma[i] = 0;
}
}
@@ -2027,7 +2027,6 @@ static inline void velocity_iph_realign(struct velocity_info *vptr,
*/
static int velocity_receive_frame(struct velocity_info *vptr, int idx)
{
- void (*pci_action)(struct pci_dev *, dma_addr_t, size_t, int);
struct net_device_stats *stats = &vptr->netdev->stats;
struct velocity_rd_info *rd_info = &(vptr->rx.info[idx]);
struct rx_desc *rd = &(vptr->rx.ring[idx]);
@@ -2045,8 +2044,8 @@ static int velocity_receive_frame(struct velocity_info *vptr, int idx)

skb = rd_info->skb;

- pci_dma_sync_single_for_cpu(vptr->pdev, rd_info->skb_dma,
- vptr->rx.buf_sz, PCI_DMA_FROMDEVICE);
+ dma_sync_single_for_cpu(vptr->dev, rd_info->skb_dma,
+ vptr->rx.buf_sz, DMA_FROM_DEVICE);

/*
* Drop frame not meeting IEEE 802.3
@@ -2059,19 +2058,18 @@ static int velocity_receive_frame(struct velocity_info *vptr, int idx)
}
}

- pci_action = pci_dma_sync_single_for_device;
-
velocity_rx_csum(rd, skb);

if (velocity_rx_copy(&skb, pkt_len, vptr) < 0) {
velocity_iph_realign(vptr, skb, pkt_len);
- pci_action = pci_unmap_single;
rd_info->skb = NULL;
+ dma_unmap_single(vptr->dev, rd_info->skb_dma, vptr->rx.buf_sz,
+ DMA_FROM_DEVICE);
+ } else {
+ dma_sync_single_for_device(vptr->dev, rd_info->skb_dma,
+ vptr->rx.buf_sz, DMA_FROM_DEVICE);
}

- pci_action(vptr->pdev, rd_info->skb_dma, vptr->rx.buf_sz,
- PCI_DMA_FROMDEVICE);
-
skb_put(skb, pkt_len - 4);
skb->protocol = eth_type_trans(skb, vptr->netdev);

@@ -2548,7 +2546,8 @@ static netdev_tx_t velocity_xmit(struct sk_buff *skb,
* add it to the transmit ring.
*/
tdinfo->skb = skb;
- tdinfo->skb_dma[0] = pci_map_single(vptr->pdev, skb->data, pktlen, PCI_DMA_TODEVICE);
+ tdinfo->skb_dma[0] = dma_map_single(vptr->dev, skb->data, pktlen,
+ DMA_TO_DEVICE);
td_ptr->tdesc0.len = cpu_to_le16(pktlen);
td_ptr->td_buf[0].pa_low = cpu_to_le32(tdinfo->skb_dma[0]);
td_ptr->td_buf[0].pa_high = 0;
@@ -2558,7 +2557,7 @@ static netdev_tx_t velocity_xmit(struct sk_buff *skb,
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
const skb_frag_t *frag = &skb_shinfo(skb)->frags[i];

- tdinfo->skb_dma[i + 1] = skb_frag_dma_map(&vptr->pdev->dev,
+ tdinfo->skb_dma[i + 1] = skb_frag_dma_map(vptr->dev,
frag, 0,
skb_frag_size(frag),
DMA_TO_DEVICE);
@@ -2635,6 +2634,7 @@ static void velocity_init_info(struct pci_dev *pdev, struct velocity_info *vptr,
{
memset(vptr, 0, sizeof(struct velocity_info));

+ vptr->dev = &pdev->dev;
vptr->pdev = pdev;
vptr->chip_id = info->chip_id;
vptr->tx.numq = info->txqueue;
@@ -2742,7 +2742,6 @@ static int velocity_found1(struct pci_dev *pdev,
SET_NETDEV_DEV(dev, &pdev->dev);
vptr = netdev_priv(dev);

-
if (first) {
printk(KERN_INFO "%s Ver. %s\n",
VELOCITY_FULL_DRV_NAM, VELOCITY_VERSION);
diff --git a/drivers/net/ethernet/via/via-velocity.h b/drivers/net/ethernet/via/via-velocity.h
index ff8d7828..c38bbae 100644
--- a/drivers/net/ethernet/via/via-velocity.h
+++ b/drivers/net/ethernet/via/via-velocity.h
@@ -1434,6 +1434,7 @@ struct velocity_opt {
#define GET_RD_BY_IDX(vptr, idx) (vptr->rd_ring[idx])

struct velocity_info {
+ struct device *dev;
struct pci_dev *pdev;
struct net_device *netdev;

--
1.7.9.5

2013-04-28 03:31:39

by Tony Prisk

[permalink] [raw]
Subject: Re: [PATCH 3/4] net: velocity: Convert to generic dma functions

On 28/04/13 15:16, Tony Prisk wrote:
> Remove the pci_* dma functions and replace with the more generic
> versions.
>
> In preparation of adding platform support, a new struct device *dev
> is added to struct velocity_info which can be used by both the pci
> and platform code.
>
> Signed-off-by: Tony Prisk <[email protected]>
> ---
> drivers/net/ethernet/via/via-velocity.c | 51 +++++++++++++++----------------
> drivers/net/ethernet/via/via-velocity.h | 1 +
> 2 files changed, 26 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/net/ethernet/via/via-velocity.c b/drivers/net/ethernet/via/via-velocity.c
> index 1168666..a5fe892 100644
> --- a/drivers/net/ethernet/via/via-velocity.c
> +++ b/drivers/net/ethernet/via/via-velocity.c
> @@ -46,6 +46,7 @@
> #include <linux/crc-ccitt.h>
> #include <linux/crc32.h>
> #include <linux/delay.h>
> +#include <linux/dma-mapping.h>
> #include <linux/errno.h>
> #include <linux/etherdevice.h>
> #include <linux/ethtool.h>
> @@ -1457,7 +1458,6 @@ static int velocity_init_dma_rings(struct velocity_info *vptr)
> struct velocity_opt *opt = &vptr->options;
> const unsigned int rx_ring_size = opt->numrx * sizeof(struct rx_desc);
> const unsigned int tx_ring_size = opt->numtx * sizeof(struct tx_desc);
> - struct pci_dev *pdev = vptr->pdev;
> dma_addr_t pool_dma;
> void *pool;
> unsigned int i;
> @@ -1465,13 +1465,13 @@ static int velocity_init_dma_rings(struct velocity_info *vptr)
> /*
> * Allocate all RD/TD rings a single pool.
> *
> - * pci_alloc_consistent() fulfills the requirement for 64 bytes
> + * dma_alloc_coherent() fulfills the requirement for 64 bytes
> * alignment
> */
> - pool = pci_alloc_consistent(pdev, tx_ring_size * vptr->tx.numq +
> - rx_ring_size, &pool_dma);
> + pool = dma_alloc_coherent(vptr->dev, tx_ring_size * vptr->tx.numq +
> + rx_ring_size, &pool_dma, GFP_ATOMIC);
> if (!pool) {
> - dev_err(&pdev->dev, "%s : DMA memory allocation failed.\n",
> + dev_err(vptr->dev, "%s : DMA memory allocation failed.\n",
> vptr->netdev->name);
> return -ENOMEM;
> }
> @@ -1522,8 +1522,8 @@ static int velocity_alloc_rx_buf(struct velocity_info *vptr, int idx)
> */
> skb_reserve(rd_info->skb,
> 64 - ((unsigned long) rd_info->skb->data & 63));
> - rd_info->skb_dma = pci_map_single(vptr->pdev, rd_info->skb->data,
> - vptr->rx.buf_sz, PCI_DMA_FROMDEVICE);
> + rd_info->skb_dma = dma_map_single(vptr->dev, rd_info->skb->data,
> + vptr->rx.buf_sz, DMA_FROM_DEVICE);
>
> /*
> * Fill in the descriptor to match
> @@ -1586,8 +1586,8 @@ static void velocity_free_rd_ring(struct velocity_info *vptr)
>
> if (!rd_info->skb)
> continue;
> - pci_unmap_single(vptr->pdev, rd_info->skb_dma, vptr->rx.buf_sz,
> - PCI_DMA_FROMDEVICE);
> + dma_unmap_single(vptr->dev, rd_info->skb_dma, vptr->rx.buf_sz,
> + DMA_FROM_DEVICE);
> rd_info->skb_dma = 0;
>
> dev_kfree_skb(rd_info->skb);
> @@ -1668,7 +1668,7 @@ static void velocity_free_dma_rings(struct velocity_info *vptr)
> const int size = vptr->options.numrx * sizeof(struct rx_desc) +
> vptr->options.numtx * sizeof(struct tx_desc) * vptr->tx.numq;
>
> - pci_free_consistent(vptr->pdev, size, vptr->rx.ring, vptr->rx.pool_dma);
> + dma_free_coherent(vptr->dev, size, vptr->rx.ring, vptr->rx.pool_dma);
> }
>
> static int velocity_init_rings(struct velocity_info *vptr, int mtu)
> @@ -1725,8 +1725,8 @@ static void velocity_free_tx_buf(struct velocity_info *vptr,
> pktlen = max_t(size_t, pktlen,
> td->td_buf[i].size & ~TD_QUEUE);
>
> - pci_unmap_single(vptr->pdev, tdinfo->skb_dma[i],
> - le16_to_cpu(pktlen), PCI_DMA_TODEVICE);
> + dma_unmap_single(vptr->dev, tdinfo->skb_dma[i],
> + le16_to_cpu(pktlen), DMA_TO_DEVICE);
> }
> }
> dev_kfree_skb_irq(skb);
> @@ -1748,8 +1748,8 @@ static void velocity_free_td_ring_entry(struct velocity_info *vptr,
> if (td_info->skb) {
> for (i = 0; i < td_info->nskb_dma; i++) {
> if (td_info->skb_dma[i]) {
> - pci_unmap_single(vptr->pdev, td_info->skb_dma[i],
> - td_info->skb->len, PCI_DMA_TODEVICE);
> + dma_unmap_single(vptr->dev, td_info->skb_dma[i],
> + td_info->skb->len, DMA_TO_DEVICE);
> td_info->skb_dma[i] = 0;
> }
> }
> @@ -2027,7 +2027,6 @@ static inline void velocity_iph_realign(struct velocity_info *vptr,
> */
> static int velocity_receive_frame(struct velocity_info *vptr, int idx)
> {
> - void (*pci_action)(struct pci_dev *, dma_addr_t, size_t, int);
> struct net_device_stats *stats = &vptr->netdev->stats;
> struct velocity_rd_info *rd_info = &(vptr->rx.info[idx]);
> struct rx_desc *rd = &(vptr->rx.ring[idx]);
> @@ -2045,8 +2044,8 @@ static int velocity_receive_frame(struct velocity_info *vptr, int idx)
>
> skb = rd_info->skb;
>
> - pci_dma_sync_single_for_cpu(vptr->pdev, rd_info->skb_dma,
> - vptr->rx.buf_sz, PCI_DMA_FROMDEVICE);
> + dma_sync_single_for_cpu(vptr->dev, rd_info->skb_dma,
> + vptr->rx.buf_sz, DMA_FROM_DEVICE);
>
> /*
> * Drop frame not meeting IEEE 802.3
> @@ -2059,19 +2058,18 @@ static int velocity_receive_frame(struct velocity_info *vptr, int idx)
> }
> }
>
> - pci_action = pci_dma_sync_single_for_device;
> -
> velocity_rx_csum(rd, skb);
>
> if (velocity_rx_copy(&skb, pkt_len, vptr) < 0) {
> velocity_iph_realign(vptr, skb, pkt_len);
> - pci_action = pci_unmap_single;
> rd_info->skb = NULL;
> + dma_unmap_single(vptr->dev, rd_info->skb_dma, vptr->rx.buf_sz,
> + DMA_FROM_DEVICE);
> + } else {
> + dma_sync_single_for_device(vptr->dev, rd_info->skb_dma,
> + vptr->rx.buf_sz, DMA_FROM_DEVICE);
> }
>
> - pci_action(vptr->pdev, rd_info->skb_dma, vptr->rx.buf_sz,
> - PCI_DMA_FROMDEVICE);
> -
> skb_put(skb, pkt_len - 4);
> skb->protocol = eth_type_trans(skb, vptr->netdev);
>
> @@ -2548,7 +2546,8 @@ static netdev_tx_t velocity_xmit(struct sk_buff *skb,
> * add it to the transmit ring.
> */
> tdinfo->skb = skb;
> - tdinfo->skb_dma[0] = pci_map_single(vptr->pdev, skb->data, pktlen, PCI_DMA_TODEVICE);
> + tdinfo->skb_dma[0] = dma_map_single(vptr->dev, skb->data, pktlen,
> + DMA_TO_DEVICE);
> td_ptr->tdesc0.len = cpu_to_le16(pktlen);
> td_ptr->td_buf[0].pa_low = cpu_to_le32(tdinfo->skb_dma[0]);
> td_ptr->td_buf[0].pa_high = 0;
> @@ -2558,7 +2557,7 @@ static netdev_tx_t velocity_xmit(struct sk_buff *skb,
> for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
> const skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
>
> - tdinfo->skb_dma[i + 1] = skb_frag_dma_map(&vptr->pdev->dev,
> + tdinfo->skb_dma[i + 1] = skb_frag_dma_map(vptr->dev,
> frag, 0,
> skb_frag_size(frag),
> DMA_TO_DEVICE);
> @@ -2635,6 +2634,7 @@ static void velocity_init_info(struct pci_dev *pdev, struct velocity_info *vptr,
> {
> memset(vptr, 0, sizeof(struct velocity_info));
>
> + vptr->dev = &pdev->dev;
> vptr->pdev = pdev;
> vptr->chip_id = info->chip_id;
> vptr->tx.numq = info->txqueue;
> @@ -2742,7 +2742,6 @@ static int velocity_found1(struct pci_dev *pdev,
> SET_NETDEV_DEV(dev, &pdev->dev);
> vptr = netdev_priv(dev);
>
> -
> if (first) {
> printk(KERN_INFO "%s Ver. %s\n",
> VELOCITY_FULL_DRV_NAM, VELOCITY_VERSION);
> diff --git a/drivers/net/ethernet/via/via-velocity.h b/drivers/net/ethernet/via/via-velocity.h
> index ff8d7828..c38bbae 100644
> --- a/drivers/net/ethernet/via/via-velocity.h
> +++ b/drivers/net/ethernet/via/via-velocity.h
> @@ -1434,6 +1434,7 @@ struct velocity_opt {
> #define GET_RD_BY_IDX(vptr, idx) (vptr->rd_ring[idx])
>
> struct velocity_info {
> + struct device *dev;
> struct pci_dev *pdev;
> struct net_device *netdev;
>
I noticed I forgot to add the return value checking for the new
functions - will do for v2.

Regards
Tony Prisk

2013-04-28 05:21:54

by David Miller

[permalink] [raw]
Subject: Re: [PATCH 1/4] net: velocity: Alphabetize #includes

From: Tony Prisk <[email protected]>
Date: Sun, 28 Apr 2013 15:16:43 +1200

> Due to number of #includes, alphabetize them to help avoid duplicates.
>
> Signed-off-by: Tony Prisk <[email protected]>

Needless churn, I refuse to apply this.

2013-04-28 09:39:41

by Tony Prisk

[permalink] [raw]
Subject: Re: [PATCH 1/4] net: velocity: Alphabetize #includes

On 28/04/13 17:21, David Miller wrote:
> From: Tony Prisk <[email protected]>
> Date: Sun, 28 Apr 2013 15:16:43 +1200
>
>> Due to number of #includes, alphabetize them to help avoid duplicates.
>>
>> Signed-off-by: Tony Prisk <[email protected]>
> Needless churn, I refuse to apply this.
No problem - I'll drop it.
Tony Prisk