* Re: [PATCH] virtio: use __GFP_NOWARN for try_fill_recv in virtnet_poll
From: Rusty Russell @ 2011-02-15 10:11 UTC (permalink / raw)
To: Michal Hocko; +Cc: Michael S. Tsirkin, virtualization, netdev, linux-kernel
In-Reply-To: <20110215093527.GB8341@tiehlicka.suse.cz>
On Tue, 15 Feb 2011 08:05:27 pm Michal Hocko wrote:
> Hi,
> we have started seeing a lot of allocator messages complaining about
> failed allocations from virtnet_poll in soft IRQ. Could you consider the
> following patch, please?
Do we really want to silence this? Isn't warning about it kind of the
point? Your network is probably sucking if this happens...
Cheers,
Rusty.
^ permalink raw reply
* Re: [PATCH 1/1] ARC VMAC ethernet driver.
From: Eric Dumazet @ 2011-02-15 10:02 UTC (permalink / raw)
To: Andreas Fenkart; +Cc: davem, netdev
In-Reply-To: <1297762288-9263-1-git-send-email-afenkart@gmail.com>
Le mardi 15 février 2011 à 10:31 +0100, Andreas Fenkart a écrit :
> Signed-off-by: Andreas Fenkart <afenkart@gmail.com>
> + processed++;
> + skb->dev = dev;
eth_type_trans() already sets "skb->dev = dev;"
> + skb->protocol = eth_type_trans(skb, dev);
> + ap->stats.rx_packets++;
Hmm, why dont you use dev->stats internal structure ? No need to
maintain a shadow in ap->stats.
> + ap->stats.rx_bytes += skb->len;
> + dev->last_rx = jiffies;
/* last_rx = jiffies; not needed anymore */
> + netif_rx(skb);
A NAPI driver should use netif_receive_skb(), not netif_rx()
^ permalink raw reply
* [PATCH] virtio: use __GFP_NOWARN for try_fill_recv in virtnet_poll
From: Michal Hocko @ 2011-02-15 9:35 UTC (permalink / raw)
To: Rusty Russell; +Cc: Michael S. Tsirkin, virtualization, netdev, linux-kernel
Hi,
we have started seeing a lot of allocator messages complaining about
failed allocations from virtnet_poll in soft IRQ. Could you consider the
following patch, please?
The patch is based on 2.6.38-rc4.
---
>From aabc19f22915dafeac0f1f6aa7cb7e49a8021ba1 Mon Sep 17 00:00:00 2001
From: Michal Hocko <mhocko@suse.cz>
Date: Tue, 15 Feb 2011 10:20:59 +0100
Subject: [PATCH] virtio: use __GFP_NOWARN for try_fill_recv in virtnet_poll
virtnet_poll is called from soft IRQ and it tries to allocate GFP_ATOMIC
memory (through try_fill_recv). This allocation can fail and we are
falling back to schedule_delayed_work in that case.
Let's add __GFP_NOWARN to the allocation flags to get rid of the
allocator complains for failed allocations:
[22798.508903] The following is only an harmless informational message.
[22798.508909] Unless you get a _continuous_flood_ of these messages it means
[22798.508911] everything is working fine. Allocations from irqs cannot be
[22798.508913] perfectly reliable and the kernel is designed to handle that.
[22798.508917] loop3: page allocation failure. order:0, mode:0x20, alloc_flags:0x30 pflags:0x80208040
Signed-off-by: Michal Hocko <mhocko@suse.cz>
---
drivers/net/virtio_net.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 90a23e4..aea1e51 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -477,7 +477,7 @@ again:
}
if (vi->num < vi->max / 2) {
- if (!try_fill_recv(vi, GFP_ATOMIC))
+ if (!try_fill_recv(vi, GFP_ATOMIC|__GFP_NOWARN))
schedule_delayed_work(&vi->refill, 0);
}
--
1.7.2.3
--
Michal Hocko
SUSE Labs
SUSE LINUX s.r.o.
Lihovarska 1060/12
190 00 Praha 9
Czech Republic
^ permalink raw reply related
* [PATCH 1/1] ARC VMAC ethernet driver.
From: Andreas Fenkart @ 2011-02-15 9:31 UTC (permalink / raw)
To: davem; +Cc: netdev, Andreas Fenkart
In-Reply-To: <20101208.090040.39174008.davem@davemloft.net>
Signed-off-by: Andreas Fenkart <afenkart@gmail.com>
---
drivers/net/Kconfig | 10 +
drivers/net/Makefile | 1 +
drivers/net/arcvmac.c | 1495 +++++++++++++++++++++++++++++++++++++++++++++++++
drivers/net/arcvmac.h | 266 +++++++++
4 files changed, 1772 insertions(+), 0 deletions(-)
diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index 0382332..ab239da 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -251,6 +251,16 @@ config AX88796_93CX6
help
Select this if your platform comes with an external 93CX6 eeprom.
+config ARCVMAC
+ tristate "ARC VMAC ethernet support"
+ depends on HAS_DMA
+ select MII
+ select PHYLIB
+ select CRC32
+
+ help
+ MAC present on Zoran Quatro43XX
+
config MACE
tristate "MACE (Power Mac ethernet) support"
depends on PPC_PMAC && PPC32
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index b90738d..059e253 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -140,6 +140,7 @@ obj-$(CONFIG_ULTRA32) += smc-ultra32.o 8390.o
obj-$(CONFIG_E2100) += e2100.o 8390.o
obj-$(CONFIG_ES3210) += es3210.o 8390.o
obj-$(CONFIG_LNE390) += lne390.o 8390.o
+obj-$(CONFIG_ARCVMAC) += arcvmac.o
obj-$(CONFIG_NE3210) += ne3210.o 8390.o
obj-$(CONFIG_SB1250_MAC) += sb1250-mac.o
obj-$(CONFIG_B44) += b44.o
diff --git a/drivers/net/arcvmac.c b/drivers/net/arcvmac.c
new file mode 100644
index 0000000..099f4c8
--- /dev/null
+++ b/drivers/net/arcvmac.c
@@ -0,0 +1,1495 @@
+/*
+ * linux/arch/arc/drivers/arcvmac.c
+ *
+ * Copyright (C) 2003-2006 Codito Technologies, for linux-2.4 port
+ * Copyright (C) 2006-2007 Celunite Inc, for linux-2.6 port
+ * Copyright (C) 2007-2008 Sagem Communications, Fehmi Hafsi
+ * Copyright (C) 2009-2011 Sagem Communications, Andreas Fenkart
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ *
+ * external PHY support based on dnet.c
+ * ring management based on bcm63xx_enet.c
+ */
+
+#include <linux/clk.h>
+#include <linux/crc32.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/etherdevice.h>
+#include <linux/init.h>
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/netdevice.h>
+#include <linux/phy.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/sched.h>
+#include <linux/types.h>
+
+#include "arcvmac.h"
+
+/* Register access macros */
+#define vmac_writel(port, value, reg) \
+ writel(cpu_to_le32(value), (port)->regs + VMAC_##reg)
+#define vmac_readl(port, reg) le32_to_cpu(readl((port)->regs + VMAC_##reg))
+
+static int get_register_map(struct vmac_priv *ap);
+static int put_register_map(struct vmac_priv *ap);
+
+static unsigned char *read_mac_reg(struct net_device *dev,
+ unsigned char hwaddr[ETH_ALEN])
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ unsigned mac_lo, mac_hi;
+
+ WARN_ON(!hwaddr);
+ mac_lo = vmac_readl(ap, ADDRL);
+ mac_hi = vmac_readl(ap, ADDRH);
+
+ hwaddr[0] = (mac_lo >> 0) & 0xff;
+ hwaddr[1] = (mac_lo >> 8) & 0xff;
+ hwaddr[2] = (mac_lo >> 16) & 0xff;
+ hwaddr[3] = (mac_lo >> 24) & 0xff;
+ hwaddr[4] = (mac_hi >> 0) & 0xff;
+ hwaddr[5] = (mac_hi >> 8) & 0xff;
+ return hwaddr;
+}
+
+static void write_mac_reg(struct net_device *dev, unsigned char* hwaddr)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ unsigned mac_lo, mac_hi;
+
+ mac_lo = hwaddr[3] << 24 | hwaddr[2] << 16 | hwaddr[1] << 8 |
+ hwaddr[0];
+ mac_hi = hwaddr[5] << 8 | hwaddr[4];
+
+ vmac_writel(ap, mac_lo, ADDRL);
+ vmac_writel(ap, mac_hi, ADDRH);
+}
+
+static void vmac_mdio_xmit(struct vmac_priv *ap, unsigned val)
+{
+ init_completion(&ap->mdio_complete);
+ vmac_writel(ap, val, MDIO_DATA);
+ wait_for_completion(&ap->mdio_complete);
+}
+
+static int vmac_mdio_read(struct mii_bus *bus, int phy_id, int phy_reg)
+{
+ struct vmac_priv *vmac = bus->priv;
+ unsigned int val;
+
+ /* only 5 bits allowed for phy-addr and reg_offset */
+ WARN_ON(phy_id & ~0x1f || phy_reg & ~0x1f);
+
+ val = MDIO_BASE | MDIO_OP_READ;
+ val |= phy_id << 23 | phy_reg << 18;
+ vmac_mdio_xmit(vmac, val);
+
+ val = vmac_readl(vmac, MDIO_DATA);
+ return val & MDIO_DATA_MASK;
+}
+
+static int vmac_mdio_write(struct mii_bus *bus, int phy_id, int phy_reg,
+ u16 value)
+{
+ struct vmac_priv *vmac = bus->priv;
+ unsigned int val;
+
+ /* only 5 bits allowed for phy-addr and reg_offset */
+ WARN_ON(phy_id & ~0x1f || phy_reg & ~0x1f);
+
+ val = MDIO_BASE | MDIO_OP_WRITE;
+ val |= phy_id << 23 | phy_reg << 18;
+ val |= (value & MDIO_DATA_MASK);
+ vmac_mdio_xmit(vmac, val);
+
+ return 0;
+}
+
+static void vmac_handle_link_change(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ struct phy_device *phydev = ap->phy_dev;
+ unsigned long flags;
+ int report_change = 0;
+
+ spin_lock_irqsave(&ap->lock, flags);
+
+ if (phydev->duplex != ap->duplex) {
+ unsigned tmp;
+
+ tmp = vmac_readl(ap, ENABLE);
+
+ if (phydev->duplex)
+ tmp |= ENFL_MASK;
+ else
+ tmp &= ~ENFL_MASK;
+
+ vmac_writel(ap, tmp, ENABLE);
+
+ ap->duplex = phydev->duplex;
+ report_change = 1;
+ }
+
+ if (phydev->speed != ap->speed) {
+ ap->speed = phydev->speed;
+ report_change = 1;
+ }
+
+ if (phydev->link != ap->link) {
+ ap->link = phydev->link;
+ report_change = 1;
+ }
+
+ spin_unlock_irqrestore(&ap->lock, flags);
+
+ if (report_change)
+ phy_print_status(ap->phy_dev);
+}
+
+static int __devinit vmac_mii_probe(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ struct phy_device *phydev = NULL;
+ struct clk *vmac_clk;
+ unsigned long clock_rate;
+ int phy_addr, err;
+
+ /* find the first phy */
+ for (phy_addr = 0; phy_addr < PHY_MAX_ADDR; phy_addr++) {
+ if (ap->mii_bus->phy_map[phy_addr]) {
+ phydev = ap->mii_bus->phy_map[phy_addr];
+ break;
+ }
+ }
+
+ if (!phydev) {
+ dev_err(&ap->pdev->dev, "no PHY found\n");
+ return -ENODEV;
+ }
+
+ /* FIXME: add pin_irq, if avail */
+
+ phydev = phy_connect(dev, dev_name(&phydev->dev),
+ &vmac_handle_link_change, 0,
+ PHY_INTERFACE_MODE_MII);
+
+ if (IS_ERR(phydev)) {
+ err = PTR_ERR(phydev);
+ dev_err(&ap->pdev->dev, "could not attach to PHY %d\n", err);
+ goto err_out;
+ }
+
+ phydev->supported &= PHY_BASIC_FEATURES;
+ phydev->supported |= SUPPORTED_Asym_Pause | SUPPORTED_Pause;
+
+ vmac_clk = clk_get(&ap->pdev->dev, "arcvmac");
+ if (IS_ERR(vmac_clk)) {
+ err = PTR_ERR(vmac_clk);
+ goto err_disconnect;
+ }
+
+ clock_rate = clk_get_rate(vmac_clk);
+ clk_put(vmac_clk);
+
+ dev_dbg(&ap->pdev->dev, "vmac_clk: %lu Hz\n", clock_rate);
+
+ if (clock_rate < 25000000)
+ phydev->supported &= ~(SUPPORTED_100baseT_Half |
+ SUPPORTED_100baseT_Full);
+
+ phydev->advertising = phydev->supported;
+
+ ap->link = 0;
+ ap->speed = 0;
+ ap->duplex = -1;
+ ap->phy_dev = phydev;
+
+ return 0;
+
+err_disconnect:
+ phy_disconnect(phydev);
+err_out:
+ return err;
+}
+
+static int __devinit vmac_mii_init(struct vmac_priv *ap)
+{
+ unsigned long flags;
+ int err, i;
+
+ spin_lock_irqsave(&ap->lock, flags);
+
+ ap->mii_bus = mdiobus_alloc();
+ if (ap->mii_bus == NULL)
+ return -ENOMEM;
+
+ ap->mii_bus->name = "vmac_mii_bus";
+ ap->mii_bus->read = &vmac_mdio_read;
+ ap->mii_bus->write = &vmac_mdio_write;
+
+ snprintf(ap->mii_bus->id, MII_BUS_ID_SIZE, "%x", 0);
+
+ ap->mii_bus->priv = ap;
+
+ err = -ENOMEM;
+ ap->mii_bus->irq = kmalloc(sizeof(int) * PHY_MAX_ADDR, GFP_KERNEL);
+ if (!ap->mii_bus->irq)
+ goto err_out;
+
+ for (i = 0; i < PHY_MAX_ADDR; i++)
+ ap->mii_bus->irq[i] = PHY_POLL;
+
+ spin_unlock_irqrestore(&ap->lock, flags);
+
+ /* locking: mdio concurrency */
+
+ err = mdiobus_register(ap->mii_bus);
+ if (err)
+ goto err_out_free_mdio_irq;
+
+ err = vmac_mii_probe(ap->dev);
+ if (err)
+ goto err_out_unregister_bus;
+
+ return 0;
+
+err_out_unregister_bus:
+ mdiobus_unregister(ap->mii_bus);
+err_out_free_mdio_irq:
+ kfree(ap->mii_bus->irq);
+err_out:
+ mdiobus_free(ap->mii_bus);
+ return err;
+}
+
+static void vmac_mii_exit_unlocked(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+
+ if (ap->phy_dev)
+ phy_disconnect(ap->phy_dev);
+
+ mdiobus_unregister(ap->mii_bus);
+ kfree(ap->mii_bus->irq);
+ mdiobus_free(ap->mii_bus);
+}
+
+static int vmacether_get_settings(struct net_device *dev,
+ struct ethtool_cmd *cmd)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ struct phy_device *phydev = ap->phy_dev;
+
+ if (!phydev)
+ return -ENODEV;
+
+ return phy_ethtool_gset(phydev, cmd);
+}
+
+static int vmacether_set_settings(struct net_device *dev,
+ struct ethtool_cmd *cmd)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ struct phy_device *phydev = ap->phy_dev;
+
+ if (!phydev)
+ return -ENODEV;
+
+ return phy_ethtool_sset(phydev, cmd);
+}
+
+static int vmac_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ struct phy_device *phydev = ap->phy_dev;
+
+ if (!netif_running(dev))
+ return -EINVAL;
+
+ if (!phydev)
+ return -ENODEV;
+
+ return phy_mii_ioctl(phydev, rq, cmd);
+}
+
+static void vmacether_get_drvinfo(struct net_device *dev,
+ struct ethtool_drvinfo *info)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+
+ strlcpy(info->driver, DRV_NAME, sizeof(info->driver));
+ strlcpy(info->version, DRV_VERSION, sizeof(info->version));
+ snprintf(info->bus_info, sizeof(info->bus_info),
+ "platform 0x%pP", &ap->mem->start);
+}
+
+static int update_error_counters_unlocked(struct net_device *dev, int status)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ dev_dbg(&ap->pdev->dev, "rx error counter overrun. status = 0x%x\n",
+ status);
+
+ /* programming error */
+ WARN_ON(status & TXCH_MASK);
+ WARN_ON(!(status & (MSER_MASK | RXCR_MASK | RXFR_MASK | RXFL_MASK)));
+
+ if (status & MSER_MASK)
+ ap->stats.rx_over_errors += 256; /* ran out of BD */
+ if (status & RXCR_MASK)
+ ap->stats.rx_crc_errors += 256;
+ if (status & RXFR_MASK)
+ ap->stats.rx_frame_errors += 256;
+ if (status & RXFL_MASK)
+ ap->stats.rx_fifo_errors += 256;
+
+ return 0;
+}
+
+static void update_tx_errors_unlocked(struct net_device *dev, int status)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+
+ if (status & BD_UFLO)
+ ap->stats.tx_fifo_errors++;
+
+ if (ap->duplex)
+ return;
+
+ /* half duplex flags */
+ if (status & BD_LTCL)
+ ap->stats.tx_window_errors++;
+ if (status & BD_RETRY_CT)
+ ap->stats.collisions += (status & BD_RETRY_CT) >> 24;
+ if (status & BD_DROP) /* too many retries */
+ ap->stats.tx_aborted_errors++;
+ if (status & BD_DEFER)
+ dev_vdbg(&ap->pdev->dev, "\"defer to traffic\"\n");
+ if (status & BD_CARLOSS)
+ ap->stats.tx_carrier_errors++;
+}
+
+static int vmac_rx_reclaim_force_unlocked(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ int ct;
+
+ /* locking: no conurrency, runs only during shutdown */
+ WARN_ON(!ap->shutdown);
+
+ dev_dbg(&ap->pdev->dev, "need to release %d rx sk_buff\n",
+ fifo_used(&ap->rx_ring));
+
+ ct = 0;
+ while (!fifo_empty(&ap->rx_ring) && ct++ < ap->rx_ring.size) {
+ struct vmac_buffer_desc *desc;
+ struct sk_buff *skb;
+ int desc_idx;
+
+ desc_idx = ap->rx_ring.tail;
+ desc = &ap->rxbd[desc_idx];
+ fifo_inc_tail(&ap->rx_ring);
+
+ if (!ap->rx_skbuff[desc_idx]) {
+ dev_err(&ap->pdev->dev, "non-populated rx_skbuff found %d\n",
+ desc_idx);
+ continue;
+ }
+
+ skb = ap->rx_skbuff[desc_idx];
+ ap->rx_skbuff[desc_idx] = NULL;
+
+ dma_unmap_single(&ap->pdev->dev, desc->data, skb->len,
+ DMA_TO_DEVICE);
+
+ dev_kfree_skb(skb);
+ }
+
+ if (!fifo_empty(&ap->rx_ring)) {
+ dev_err(&ap->pdev->dev, "failed to reclaim %d rx sk_buff\n",
+ fifo_used(&ap->rx_ring));
+ }
+
+ return 0;
+}
+
+/* Function refills empty buffer descriptors and passes ownership to DMA */
+static int vmac_rx_refill_unlocked(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+
+ /* locking: protect from refill_timer */
+ /* locking: fct owns area outside rx_ring, head exclusive tail,
+ * modifies head */
+
+ spin_lock(&ap->refill_lock);
+
+ WARN_ON(fifo_full(&ap->rx_ring));
+
+ while (!fifo_full(&ap->rx_ring)) {
+ struct vmac_buffer_desc *desc;
+ struct sk_buff *skb;
+ dma_addr_t p;
+ int desc_idx;
+
+ desc_idx = ap->rx_ring.head;
+ desc = &ap->rxbd[desc_idx];
+
+ /* make sure we read the actual descriptor status */
+ rmb();
+
+ if (ap->rx_skbuff[desc_idx]) {
+ /* dropped packet / buffer chaining */
+ fifo_inc_head(&ap->rx_ring);
+
+ /* return to DMA */
+ wmb();
+ desc->info = cpu_to_le32(BD_DMA_OWN | ap->rx_skb_size);
+ continue;
+ }
+
+ skb = netdev_alloc_skb_ip_align(dev, ap->rx_skb_size);
+ if (!skb) {
+ dev_info(&ap->pdev->dev, "failed to allocate rx_skb, skb's left %d\n",
+ fifo_used(&ap->rx_ring));
+ break;
+ }
+
+ ap->rx_skbuff[desc_idx] = skb;
+
+ p = dma_map_single(&ap->pdev->dev, skb->data, ap->rx_skb_size,
+ DMA_FROM_DEVICE);
+
+ desc->data = p;
+
+ wmb();
+ desc->info = cpu_to_le32(BD_DMA_OWN | ap->rx_skb_size);
+
+ fifo_inc_head(&ap->rx_ring);
+ }
+
+ spin_unlock(&ap->refill_lock);
+
+ /* If rx ring is still empty, set a timer to try allocating
+ * again at a later time. */
+ if (fifo_empty(&ap->rx_ring) && netif_running(dev)) {
+ dev_warn(&ap->pdev->dev, "unable to refill rx ring\n");
+ ap->refill_timer.expires = jiffies + HZ;
+ add_timer(&ap->refill_timer);
+ }
+
+ return 0;
+}
+
+/*
+ * timer callback to defer refill rx queue in case we're OOM
+ */
+static void vmac_refill_rx_timer(unsigned long data)
+{
+ vmac_rx_refill_unlocked((struct net_device *)data);
+}
+
+/* merge buffer chaining */
+struct sk_buff *vmac_merge_rx_buffers_unlocked(struct net_device *dev,
+ struct vmac_buffer_desc *after,
+ int pkt_len) /* data */
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ struct sk_buff *merge_skb, *cur_skb;
+ struct dma_fifo *rx_ring;
+ struct vmac_buffer_desc *desc;
+
+ /* locking: same as vmac_rx_receive */
+
+ rx_ring = &ap->rx_ring;
+ desc = &ap->rxbd[rx_ring->tail];
+
+ WARN_ON(desc == after);
+
+ /* strip FCS */
+ pkt_len -= 4;
+
+ merge_skb = netdev_alloc_skb_ip_align(dev, pkt_len + NET_IP_ALIGN);
+ if (!merge_skb) {
+ dev_err(&ap->pdev->dev, "failed to allocate merged rx_skb, rx skb's left %d\n",
+ fifo_used(rx_ring));
+
+ return NULL;
+ }
+
+ while (desc != after && pkt_len) {
+ struct vmac_buffer_desc *desc;
+ int buf_len, valid;
+
+ /* desc needs wrapping */
+ desc = &ap->rxbd[rx_ring->tail];
+ cur_skb = ap->rx_skbuff[rx_ring->tail];
+ WARN_ON(!cur_skb);
+
+ dma_unmap_single(&ap->pdev->dev, desc->data, ap->rx_skb_size,
+ DMA_FROM_DEVICE);
+
+ /* do not copy FCS */
+ buf_len = le32_to_cpu(desc->info) & BD_LEN;
+ valid = min(pkt_len, buf_len);
+ pkt_len -= valid;
+
+ memcpy(skb_put(merge_skb, valid), cur_skb->data, valid);
+
+ fifo_inc_tail(rx_ring);
+ }
+
+ /* merging_pressure++ */
+
+ if (unlikely(pkt_len != 0))
+ dev_err(&ap->pdev->dev, "buffer chaining bytes missing %d\n",
+ pkt_len);
+
+ WARN_ON(desc != after);
+
+ return merge_skb;
+}
+
+int vmac_rx_receive(struct net_device *dev, int budget)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ struct vmac_buffer_desc *first;
+ int processed, pkt_len, pkt_err;
+ struct dma_fifo lookahead;
+
+ /* true concurrency -> DMA engine running in parallel */
+ /* locking: fct owns rx_ring tail to current DMA read position, alias
+ * 'received packets'. rx_refill owns area outside rx_ring, doesn't
+ * modify tail */
+
+ processed = 0;
+
+ first = NULL;
+ pkt_err = pkt_len = 0;
+
+ /* look ahead, till packet complete */
+ lookahead = ap->rx_ring;
+
+ do {
+ struct vmac_buffer_desc *desc; /* cur_ */
+ int desc_idx; /* cur_ */
+ struct sk_buff *skb; /* pkt_ */
+
+ desc_idx = lookahead.tail;
+ desc = &ap->rxbd[desc_idx];
+
+ /* make sure we read the actual descriptor status */
+ rmb();
+
+ /* break if dma ownership belongs to hw */
+ if (desc->info & cpu_to_le32(BD_DMA_OWN)) {
+ /* safe the dma position */
+ ap->dma_rx_head = vmac_readl(ap, MAC_RXRING_HEAD);
+ break;
+ }
+
+ if (desc->info & cpu_to_le32(BD_FRST)) {
+ pkt_len = 0;
+ pkt_err = 0;
+
+ /* don't free current */
+ ap->rx_ring.tail = lookahead.tail;
+ first = desc;
+ }
+
+ fifo_inc_tail(&lookahead);
+
+ /* check bd */
+
+ pkt_len += desc->info & cpu_to_le32(BD_LEN);
+ pkt_err |= desc->info & cpu_to_le32(BD_BUFF);
+
+ if (!(desc->info & cpu_to_le32(BD_LAST)))
+ continue;
+
+ /* received complete packet */
+
+ if (unlikely(pkt_err || !first)) {
+ /* recycle buffers */
+ ap->rx_ring.tail = lookahead.tail;
+ continue;
+ }
+
+#ifdef DEBUG
+ WARN_ON(!(first->info & cpu_to_le32(BD_FRST)) ||
+ !(desc->info & cpu_to_le32(BD_LAST)));
+ WARN_ON(pkt_err);
+#endif
+
+ /* -- valid packet -- */
+
+ if (first != desc) {
+ skb = vmac_merge_rx_buffers_unlocked(dev, desc,
+ pkt_len);
+
+ if (!skb) {
+ /* kill packet */
+ ap->rx_ring.tail = lookahead.tail;
+ ap->rx_merge_error++;
+ continue;
+ }
+ } else {
+ dma_unmap_single(&ap->pdev->dev, desc->data,
+ ap->rx_skb_size, DMA_FROM_DEVICE);
+
+ skb = ap->rx_skbuff[desc_idx];
+ ap->rx_skbuff[desc_idx] = NULL;
+ /* desc->data != skb->data => desc->data is DMA
+ * mapped */
+
+ /* strip FCS */
+ skb_put(skb, pkt_len - ETH_FCS_LEN);
+ }
+
+ /* free buffers */
+ ap->rx_ring.tail = lookahead.tail;
+
+#ifdef DEBUG
+ WARN_ON(skb->len != pkt_len - ETH_FCS_LEN);
+#endif
+ processed++;
+ skb->dev = dev;
+ skb->protocol = eth_type_trans(skb, dev);
+ ap->stats.rx_packets++;
+ ap->stats.rx_bytes += skb->len;
+ dev->last_rx = jiffies;
+ netif_rx(skb);
+
+ } while (!fifo_empty(&lookahead) && (processed < budget));
+
+ dev_vdbg(&ap->pdev->dev, "processed pkt %d, remaining rx buff %d\n",
+ processed,
+ fifo_used(&ap->rx_ring));
+
+ if (processed || fifo_empty(&ap->rx_ring))
+ vmac_rx_refill_unlocked(dev);
+
+ return processed;
+}
+
+static void vmac_toggle_irqmask_unlocked(struct net_device *dev, int enable,
+ int mask)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ unsigned long tmp;
+
+ tmp = vmac_readl(ap, ENABLE);
+ if (enable)
+ tmp |= mask;
+ else
+ tmp &= ~mask;
+ vmac_writel(ap, tmp, ENABLE);
+}
+
+static void vmac_toggle_txint(struct net_device *dev, int enable)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ unsigned long flags;
+
+ spin_lock_irqsave(&ap->lock, flags);
+ vmac_toggle_irqmask_unlocked(dev, enable, TXINT_MASK);
+ spin_unlock_irqrestore(&ap->lock, flags);
+}
+
+static void vmac_toggle_rxint_unlocked(struct net_device *dev, int enable)
+{
+ vmac_toggle_irqmask_unlocked(dev, enable, RXINT_MASK);
+}
+
+static int vmac_poll(struct napi_struct *napi, int budget)
+{
+ struct vmac_priv *ap;
+ struct net_device *dev;
+ int rx_work_done;
+ unsigned long flags;
+
+ ap = container_of(napi, struct vmac_priv, napi);
+ dev = ap->dev;
+
+ /* ack interrupt */
+ spin_lock_irqsave(&ap->lock, flags);
+ vmac_writel(ap, RXINT_MASK, STAT);
+ spin_unlock_irqrestore(&ap->lock, flags);
+
+ rx_work_done = vmac_rx_receive(dev, budget);
+
+ if (0 && printk_ratelimit()) {
+ dev_dbg(&ap->pdev->dev, "poll budget %d receive rx_work_done %d\n",
+ budget,
+ rx_work_done);
+ }
+
+ if (rx_work_done >= budget) {
+ /* rx queue is not yet empty/clean */
+ return rx_work_done;
+ }
+
+ /* no more packet in rx/tx queue, remove device from poll
+ * queue */
+ spin_lock_irqsave(&ap->lock, flags);
+ napi_complete(napi);
+ vmac_toggle_rxint_unlocked(dev, 1);
+ spin_unlock_irqrestore(&ap->lock, flags);
+
+ return rx_work_done;
+}
+
+static int vmac_tx_reclaim_unlocked(struct net_device *dev, int force);
+
+static irqreturn_t vmac_intr(int irq, void *dev_instance)
+{
+ struct net_device *dev = dev_instance;
+ struct vmac_priv *ap = netdev_priv(dev);
+ unsigned int status;
+
+ spin_lock(&ap->lock);
+
+ status = vmac_readl(ap, STAT);
+ vmac_writel(ap, status, STAT);
+
+#ifdef DEBUG
+ if (unlikely(ap->shutdown))
+ dev_err(&ap->pdev->dev, "ISR during close\n");
+
+ if (unlikely(!status & (RXINT_MASK|MDIO_MASK|ERR_MASK)))
+ dev_err(&ap->pdev->dev, "No source of IRQ found\n");
+#endif
+
+ if ((status & RXINT_MASK) &&
+ (ap->dma_rx_head !=
+ vmac_readl(ap, MAC_RXRING_HEAD))) {
+ vmac_toggle_rxint_unlocked(dev, 0);
+ napi_schedule(&ap->napi);
+ }
+
+ if (unlikely(netif_queue_stopped(dev) && (status & TXINT_MASK)))
+ vmac_tx_reclaim_unlocked(dev, 0);
+
+ if (status & MDIO_MASK)
+ complete(&ap->mdio_complete);
+
+ if (unlikely(status & ERR_MASK))
+ update_error_counters_unlocked(dev, status);
+
+ spin_unlock(&ap->lock);
+
+ return IRQ_HANDLED;
+}
+
+static int vmac_tx_reclaim_unlocked(struct net_device *dev, int force)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ int released = 0;
+
+ /* locking: modifies tx_ring tail, head only during shutdown */
+ /* locking: call with ap->lock held */
+ WARN_ON(force && !ap->shutdown);
+
+ /* buffer chaining not used, see vmac_start_xmit */
+
+ while (!fifo_empty(&ap->tx_ring)) {
+ struct vmac_buffer_desc *desc;
+ struct sk_buff *skb;
+ int desc_idx;
+
+ desc_idx = ap->tx_ring.tail;
+ desc = &ap->txbd[desc_idx];
+
+ /* ensure other field of the descriptor were not read
+ * before we checked ownership */
+ rmb();
+
+ if ((desc->info & cpu_to_le32(BD_DMA_OWN)) && !force)
+ break;
+
+ if (desc->info & cpu_to_le32(BD_TX_ERR)) {
+ update_tx_errors_unlocked(dev,
+ le32_to_cpu(desc->info));
+ /* recycle packet, let upper level deal with it */
+ }
+
+ skb = ap->tx_skbuff[desc_idx];
+ ap->tx_skbuff[desc_idx] = NULL;
+ WARN_ON(!skb);
+
+ dma_unmap_single(&ap->pdev->dev, desc->data, skb->len,
+ DMA_TO_DEVICE);
+
+ dev_kfree_skb_any(skb);
+
+ released++;
+ fifo_inc_tail(&ap->tx_ring);
+ }
+
+ if (netif_queue_stopped(dev) && released) {
+ netif_wake_queue(dev);
+ vmac_toggle_txint(dev, 0);
+ }
+
+ if (unlikely(force && !fifo_empty(&ap->tx_ring))) {
+ dev_err(&ap->pdev->dev, "failed to reclaim %d tx sk_buff\n",
+ fifo_used(&ap->tx_ring));
+ }
+
+ return released;
+}
+
+int vmac_start_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ struct vmac_buffer_desc *desc;
+ unsigned long flags;
+ unsigned int tmp;
+
+ /* running under xmit lock */
+ /* locking: modifies tx_ring head, tx_reclaim only tail */
+
+ /* no scatter/gatter see features below */
+ WARN_ON(skb_shinfo(skb)->nr_frags != 0);
+ WARN_ON(skb->len > MAX_TX_BUFFER_LEN);
+
+ if (unlikely(fifo_full(&ap->tx_ring))) {
+ netif_stop_queue(dev);
+ vmac_toggle_txint(dev, 1);
+ dev_err(&ap->pdev->dev, "xmit called with no tx desc available\n");
+ return NETDEV_TX_BUSY;
+ }
+
+ if (unlikely(skb->len < ETH_ZLEN)) {
+ struct sk_buff *short_skb;
+ short_skb = netdev_alloc_skb_ip_align(dev, ETH_ZLEN);
+ if (!short_skb)
+ return NETDEV_TX_LOCKED;
+
+ memset(short_skb->data, 0, ETH_ZLEN);
+ memcpy(skb_put(short_skb, ETH_ZLEN), skb->data, skb->len);
+ dev_kfree_skb(skb);
+ skb = short_skb;
+ }
+
+ /* fill descriptor */
+ ap->tx_skbuff[ap->tx_ring.head] = skb;
+ desc = &ap->txbd[ap->tx_ring.head];
+ WARN_ON(desc->info & cpu_to_le32(BD_DMA_OWN));
+
+ desc->data = dma_map_single(&ap->pdev->dev, skb->data, skb->len,
+ DMA_TO_DEVICE);
+
+ /* dma might already be polling */
+ wmb();
+ desc->info = cpu_to_le32(BD_DMA_OWN | BD_FRST | BD_LAST | skb->len);
+ wmb();
+
+ /* lock device data */
+ spin_lock_irqsave(&ap->lock, flags);
+
+ /* kick tx dma */
+ tmp = vmac_readl(ap, STAT);
+ vmac_writel(ap, tmp | TXPL_MASK, STAT);
+
+ ap->stats.tx_packets++;
+ ap->stats.tx_bytes += skb->len;
+ dev->trans_start = jiffies;
+ fifo_inc_head(&ap->tx_ring);
+
+ /* vmac_tx_reclaim outside of vmac_tx_timeout */
+ if (fifo_used(&ap->tx_ring) > 8)
+ vmac_tx_reclaim_unlocked(dev, 0);
+
+ /* unlock device data */
+ spin_unlock_irqrestore(&ap->lock, flags);
+
+ /* stop queue if no more desc available */
+ if (fifo_full(&ap->tx_ring)) {
+ netif_stop_queue(dev);
+ vmac_toggle_txint(dev, 1);
+ }
+
+ return NETDEV_TX_OK;
+}
+
+static int alloc_buffers_unlocked(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ int err = -ENOMEM;
+ int size;
+
+ fifo_init(&ap->rx_ring, RX_BDT_LEN);
+ fifo_init(&ap->tx_ring, TX_BDT_LEN);
+
+ /* initialize skb list */
+ memset(ap->rx_skbuff, 0, sizeof(ap->rx_skbuff));
+ memset(ap->tx_skbuff, 0, sizeof(ap->tx_skbuff));
+
+ /* allocate DMA received descriptors */
+ size = sizeof(*ap->rxbd) * ap->rx_ring.size;
+ ap->rxbd = dma_alloc_coherent(&ap->pdev->dev, size,
+ &ap->rxbd_dma,
+ GFP_KERNEL);
+ if (ap->rxbd == NULL)
+ goto err_out;
+
+ /* allocate DMA transmit descriptors */
+ size = sizeof(*ap->txbd) * ap->tx_ring.size;
+ ap->txbd = dma_alloc_coherent(&ap->pdev->dev, size,
+ &ap->txbd_dma,
+ GFP_KERNEL);
+ if (ap->txbd == NULL)
+ goto err_free_rxbd;
+
+ /* ensure 8-byte aligned */
+ WARN_ON(((uintptr_t)ap->txbd & 0x7) || ((uintptr_t)ap->rxbd & 0x7));
+
+ memset(ap->txbd, 0, sizeof(*ap->txbd) * ap->tx_ring.size);
+ memset(ap->rxbd, 0, sizeof(*ap->rxbd) * ap->rx_ring.size);
+
+ /* allocate rx skb */
+ err = vmac_rx_refill_unlocked(dev);
+ if (err)
+ goto err_free_txbd;
+
+ return 0;
+
+err_free_txbd:
+ dma_free_coherent(&ap->pdev->dev, sizeof(*ap->txbd) * ap->tx_ring.size,
+ ap->txbd, ap->txbd_dma);
+err_free_rxbd:
+ dma_free_coherent(&ap->pdev->dev, sizeof(*ap->rxbd) * ap->rx_ring.size,
+ ap->rxbd, ap->rxbd_dma);
+err_out:
+ return err;
+}
+
+static int free_buffers_unlocked(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+
+ /* free skbuff */
+ vmac_tx_reclaim_unlocked(dev, 1);
+ vmac_rx_reclaim_force_unlocked(dev);
+
+ /* free DMA ring */
+ dma_free_coherent(&ap->pdev->dev, sizeof(ap->txbd) * ap->tx_ring.size,
+ ap->txbd, ap->txbd_dma);
+ dma_free_coherent(&ap->pdev->dev, sizeof(ap->rxbd) * ap->rx_ring.size,
+ ap->rxbd, ap->rxbd_dma);
+
+ return 0;
+}
+
+static int vmac_hw_init(struct net_device *dev)
+{
+ struct vmac_priv *priv = netdev_priv(dev);
+
+ /* clear IRQ mask */
+ vmac_writel(priv, 0, ENABLE);
+
+ /* clear pending IRQ */
+ vmac_writel(priv, 0xffffffff, STAT);
+
+ /* Initialize logical address filter */
+ vmac_writel(priv, 0x0, LAFL);
+ vmac_writel(priv, 0x0, LAFH);
+
+ return 0;
+}
+
+int vmac_open(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ struct phy_device *phydev;
+ unsigned long flags;
+ unsigned int temp, ctrl;
+ int err = 0;
+
+ /* locking: no concurrency yet */
+
+ if (ap == NULL)
+ return -ENODEV;
+
+ spin_lock_irqsave(&ap->lock, flags);
+ ap->shutdown = 0;
+
+ err = get_register_map(ap);
+ if (err)
+ return err;
+
+ vmac_hw_init(dev);
+
+ /* mac address changed? */
+ write_mac_reg(dev, dev->dev_addr);
+
+ err = alloc_buffers_unlocked(dev);
+ if (err)
+ return err;
+
+ /* install DMA ring pointers */
+ vmac_writel(ap, ap->rxbd_dma, RXRINGPTR);
+ vmac_writel(ap, ap->txbd_dma, TXRINGPTR);
+
+ /* set poll rate to 1 ms */
+ vmac_writel(ap, POLLRATE_TIME, POLLRATE);
+
+ /* Set control */
+ ctrl = (RX_BDT_LEN << 24) | (TX_BDT_LEN << 16) | TXRN_MASK | RXRN_MASK;
+ vmac_writel(ap, ctrl, CONTROL);
+
+ /* make sure we enable napi before rx interrupt */
+ napi_enable(&ap->napi);
+
+ err = request_irq(dev->irq, &vmac_intr, 0, dev->name, dev);
+ if (err) {
+ dev_err(&ap->pdev->dev, "Unable to request IRQ %d (error %d)\n",
+ dev->irq, err);
+ goto err_free_buffers;
+ }
+
+ /* IRQ mask */
+ temp = RXINT_MASK | ERR_MASK | TXCH_MASK | MDIO_MASK;
+ vmac_writel(ap, temp, ENABLE);
+
+ /* enable, after all other bits are set */
+ vmac_writel(ap, ctrl | EN_MASK, CONTROL);
+ spin_unlock_irqrestore(&ap->lock, flags);
+
+ /* locking: concurrency */
+
+ netif_start_queue(dev);
+ netif_carrier_off(dev);
+
+ /* register the PHY board fixup, if needed */
+ err = vmac_mii_init(ap);
+ if (err)
+ goto err_free_irq;
+
+ /* schedule a link state check */
+ phy_start(ap->phy_dev);
+
+ phydev = ap->phy_dev;
+ dev_info(&ap->pdev->dev, "PHY driver [%s] (mii_bus:phy_addr=%s, irq=%d)\n",
+ phydev->drv->name, dev_name(&phydev->dev), phydev->irq);
+
+ return 0;
+
+err_free_irq:
+ free_irq(dev->irq, dev);
+err_free_buffers:
+ napi_disable(&ap->napi);
+ free_buffers_unlocked(dev);
+ return err;
+}
+
+int vmac_close(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ unsigned long flags;
+ unsigned int temp;
+
+ /* locking: protect everything, DMA / IRQ / timer */
+ spin_lock_irqsave(&ap->lock, flags);
+
+ /* complete running transfer, then stop */
+ temp = vmac_readl(ap, CONTROL);
+ temp &= ~(TXRN_MASK | RXRN_MASK);
+ vmac_writel(ap, temp, CONTROL);
+
+ /* reenable IRQ, process pending */
+ spin_unlock_irqrestore(&ap->lock, flags);
+
+ set_current_state(TASK_INTERRUPTIBLE);
+ schedule_timeout(msecs_to_jiffies(20));
+
+ /* shut it down now */
+ spin_lock_irqsave(&ap->lock, flags);
+ ap->shutdown = 1;
+
+ netif_stop_queue(dev);
+ napi_disable(&ap->napi);
+
+ /* disable phy */
+ phy_stop(ap->phy_dev);
+ vmac_mii_exit_unlocked(dev);
+ netif_carrier_off(dev);
+
+ /* disable interrupts */
+ vmac_writel(ap, 0, ENABLE);
+ free_irq(dev->irq, dev);
+
+ /* turn off vmac */
+ vmac_writel(ap, 0, CONTROL);
+ /* vmac_reset_hw(vmac) */
+
+ /* locking: concurrency off */
+ spin_unlock_irqrestore(&ap->lock, flags);
+
+ del_timer_sync(&ap->refill_timer);
+ free_buffers_unlocked(dev);
+
+ put_register_map(ap);
+
+ return 0;
+}
+
+void update_vmac_stats_unlocked(struct vmac_priv *ap)
+{
+ struct net_device_stats *_stats = &ap->stats;
+ unsigned long miss, rxerr;
+ unsigned long rxfram, rxcrc, rxoflow;
+
+ /* compare with /proc/net/dev,
+ * see net/core/dev.c:dev_seq_printf_stats */
+
+ /* rx stats */
+ rxerr = vmac_readl(ap, RXERR);
+ miss = vmac_readl(ap, MISS);
+
+ rxcrc = (rxerr & RXERR_CRC);
+ rxfram = (rxerr & RXERR_FRM) >> 8;
+ rxoflow = (rxerr & RXERR_OFLO) >> 16;
+
+ _stats->rx_length_errors = 0;
+ _stats->rx_over_errors += miss;
+ _stats->rx_crc_errors += rxcrc;
+ _stats->rx_frame_errors += rxfram;
+ _stats->rx_fifo_errors += rxoflow;
+ _stats->rx_missed_errors = 0;
+
+ /* TODO check rx_dropped/rx_errors/tx_dropped/tx_errors have not
+ * been updated elsewhere */
+ _stats->rx_dropped = _stats->rx_over_errors +
+ _stats->rx_fifo_errors +
+ ap->rx_merge_error;
+
+ _stats->rx_errors = _stats->rx_length_errors + _stats->rx_crc_errors +
+ _stats->rx_frame_errors +
+ _stats->rx_missed_errors +
+ _stats->rx_dropped;
+
+ /* tx stats */
+ _stats->tx_dropped = 0; /* otherwise queue stopped */
+
+ _stats->tx_errors = _stats->tx_aborted_errors +
+ _stats->tx_carrier_errors +
+ _stats->tx_fifo_errors +
+ _stats->tx_heartbeat_errors +
+ _stats->tx_window_errors +
+ _stats->tx_dropped +
+ ap->tx_timeout_error;
+}
+
+struct net_device_stats *vmac_stats(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ unsigned long flags;
+
+ spin_lock_irqsave(&ap->lock, flags);
+ update_vmac_stats_unlocked(ap);
+ spin_unlock_irqrestore(&ap->lock, flags);
+
+ return &ap->stats;
+}
+
+void vmac_tx_timeout(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ unsigned int status;
+ unsigned long flags;
+
+ spin_lock_irqsave(&ap->lock, flags);
+
+ /* queue did not progress for timeo jiffies */
+ WARN_ON(!netif_queue_stopped(dev));
+ WARN_ON(!fifo_full(&ap->tx_ring));
+
+ /* TX IRQ lost? */
+ status = vmac_readl(ap, STAT);
+ if (status & TXINT_MASK) {
+ dev_err(&ap->pdev->dev, "lost tx interrupt, IRQ mask %x\n",
+ vmac_readl(ap, ENABLE));
+ vmac_writel(ap, TXINT_MASK, STAT);
+ }
+
+ /* TODO RX/MDIO/ERR as well? */
+
+ vmac_tx_reclaim_unlocked(dev, 0);
+ if (fifo_full(&ap->tx_ring))
+ dev_err(&ap->pdev->dev, "DMA state machine not active\n");
+
+ /* We can accept TX packets again */
+ ap->tx_timeout_error++;
+ dev->trans_start = jiffies;
+ netif_wake_queue(dev);
+
+ spin_unlock_irqrestore(&ap->lock, flags);
+}
+
+static void create_multicast_filter(struct net_device *dev,
+ unsigned long *bitmask)
+{
+ unsigned long crc;
+ char *addrs;
+
+ /* locking: done by net_device */
+
+ WARN_ON(netdev_mc_count(dev) == 0);
+ WARN_ON(dev->flags & IFF_ALLMULTI);
+
+ bitmask[0] = bitmask[1] = 0;
+
+ {
+ struct netdev_hw_addr *ha;
+ netdev_for_each_mc_addr(ha, dev) {
+ addrs = ha->addr;
+
+ /* skip non-multicast addresses */
+ if (!(*addrs & 1))
+ continue;
+
+ crc = ether_crc_le(ETH_ALEN, addrs);
+ set_bit(crc >> 26, bitmask);
+ }
+ }
+}
+
+static void vmac_set_multicast_list(struct net_device *dev)
+{
+ struct vmac_priv *ap = netdev_priv(dev);
+ unsigned long flags, bitmask[2];
+ int promisc, reg;
+
+ spin_lock_irqsave(&ap->lock, flags);
+
+ promisc = !!(dev->flags & IFF_PROMISC);
+ reg = vmac_readl(ap, ENABLE);
+ if (promisc != !!(reg & PROM_MASK)) {
+ reg ^= PROM_MASK;
+ vmac_writel(ap, reg, ENABLE);
+ }
+
+ if (dev->flags & IFF_ALLMULTI)
+ memset(bitmask, 1, sizeof(bitmask));
+ else if (netdev_mc_count(dev) == 0)
+ memset(bitmask, 0, sizeof(bitmask));
+ else
+ create_multicast_filter(dev, bitmask);
+
+ vmac_writel(ap, bitmask[0], LAFL);
+ vmac_writel(ap, bitmask[1], LAFH);
+
+ spin_unlock_irqrestore(&ap->lock, flags);
+}
+
+static struct ethtool_ops vmac_ethtool_ops = {
+ .get_settings = vmacether_get_settings,
+ .set_settings = vmacether_set_settings,
+ .get_drvinfo = vmacether_get_drvinfo,
+ .get_link = ethtool_op_get_link,
+};
+
+static const struct net_device_ops vmac_netdev_ops = {
+ .ndo_open = vmac_open,
+ .ndo_stop = vmac_close,
+ .ndo_get_stats = vmac_stats,
+ .ndo_start_xmit = vmac_start_xmit,
+ .ndo_do_ioctl = vmac_ioctl,
+ .ndo_set_mac_address = eth_mac_addr,
+ .ndo_tx_timeout = vmac_tx_timeout,
+ .ndo_set_multicast_list = vmac_set_multicast_list,
+ .ndo_validate_addr = eth_validate_addr,
+ .ndo_change_mtu = eth_change_mtu,
+};
+
+static int get_register_map(struct vmac_priv *ap)
+{
+ int err;
+
+ err = -EBUSY;
+ if (!request_mem_region(ap->mem->start, resource_size(ap->mem),
+ DRV_NAME)) {
+ dev_err(&ap->pdev->dev, "no memory region available\n");
+ return err;
+ }
+
+ err = -ENOMEM;
+ ap->regs = ioremap(ap->mem->start, resource_size(ap->mem));
+ if (!ap->regs) {
+ dev_err(&ap->pdev->dev, "failed to map registers, aborting.\n");
+ goto err_out_release_mem;
+ }
+
+ return 0;
+
+err_out_release_mem:
+ release_mem_region(ap->mem->start, resource_size(ap->mem));
+ return err;
+}
+
+static int put_register_map(struct vmac_priv *ap)
+{
+ iounmap(ap->regs);
+ release_mem_region(ap->mem->start, resource_size(ap->mem));
+ return 0;
+}
+
+static int __devinit vmac_probe(struct platform_device *pdev)
+{
+ struct net_device *dev;
+ struct vmac_priv *ap;
+ struct resource *mem;
+ int err;
+
+ /* locking: no concurrency */
+
+ if (dma_get_mask(&pdev->dev) > DMA_BIT_MASK(32) ||
+ pdev->dev.coherent_dma_mask > DMA_BIT_MASK(32)) {
+ dev_err(&pdev->dev, "arcvmac supports only 32-bit DMA addresses\n");
+ return -ENODEV;
+ }
+
+ dev = alloc_etherdev(sizeof(*ap));
+ if (!dev) {
+ dev_err(&pdev->dev, "etherdev alloc failed, aborting.\n");
+ return -ENOMEM;
+ }
+
+ ap = netdev_priv(dev);
+
+ err = -ENODEV;
+ mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ if (!mem) {
+ dev_err(&pdev->dev, "no mmio resource defined\n");
+ goto err_out;
+ }
+ ap->mem = mem;
+
+ err = platform_get_irq(pdev, 0);
+ if (err < 0) {
+ dev_err(&pdev->dev, "no irq found\n");
+ goto err_out;
+ }
+ dev->irq = err;
+
+ spin_lock_init(&ap->lock);
+
+ SET_NETDEV_DEV(dev, &pdev->dev);
+ ap->dev = dev;
+ ap->pdev = pdev;
+
+ /* init rx timeout (used for oom) */
+ init_timer(&ap->refill_timer);
+ ap->refill_timer.function = vmac_refill_rx_timer;
+ ap->refill_timer.data = (unsigned long)dev;
+ spin_lock_init(&ap->refill_lock);
+
+ netif_napi_add(dev, &ap->napi, vmac_poll, 2);
+ dev->netdev_ops = &vmac_netdev_ops;
+ dev->ethtool_ops = &vmac_ethtool_ops;
+
+ dev->flags |= IFF_MULTICAST;
+
+ dev->base_addr = (unsigned long)ap->regs; /* TODO */
+
+ /* prevent buffer chaining, favor speed over space */
+ ap->rx_skb_size = ETH_FRAME_LEN + VMAC_BUFFER_PAD;
+
+ /* private struct functional */
+
+ /* temporarily map registers to fetch mac addr */
+ err = get_register_map(ap);
+ if (err)
+ goto err_out;
+
+ /* mac address intialize, set vmac_open */
+ read_mac_reg(dev, dev->dev_addr); /* TODO */
+
+ if (!is_valid_ether_addr(dev->dev_addr))
+ random_ether_addr(dev->dev_addr);
+
+ err = register_netdev(dev);
+ if (err) {
+ dev_err(&pdev->dev, "Cannot register net device, aborting.\n");
+ goto err_out;
+ }
+
+ /* release the memory region, till open is called */
+ put_register_map(ap);
+
+ dev_info(&pdev->dev, "ARC VMAC at 0x%pP irq %d %pM\n", &mem->start,
+ dev->irq, dev->dev_addr);
+ platform_set_drvdata(pdev, ap);
+
+ return 0;
+
+err_out:
+ free_netdev(dev);
+ return err;
+}
+
+static int __devexit vmac_remove(struct platform_device *pdev)
+{
+ struct vmac_priv *ap;
+
+ /* locking: no concurrency */
+
+ ap = platform_get_drvdata(pdev);
+ if (!ap) {
+ dev_err(&pdev->dev, "vmac_remove no valid dev found\n");
+ return 0;
+ }
+
+ /* MAC */
+ unregister_netdev(ap->dev);
+ netif_napi_del(&ap->napi);
+
+ platform_set_drvdata(pdev, NULL);
+ free_netdev(ap->dev);
+ return 0;
+}
+
+static struct platform_driver arcvmac_driver = {
+ .probe = vmac_probe,
+ .remove = __devexit_p(vmac_remove),
+ .driver = {
+ .name = "arcvmac",
+ },
+};
+
+static int __init vmac_init(void)
+{
+ return platform_driver_register(&arcvmac_driver);
+}
+
+static void __exit vmac_exit(void)
+{
+ platform_driver_unregister(&arcvmac_driver);
+}
+
+module_init(vmac_init);
+module_exit(vmac_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("ARC VMAC Ethernet driver");
+MODULE_AUTHOR("afenkart@gmail.com");
diff --git a/drivers/net/arcvmac.h b/drivers/net/arcvmac.h
new file mode 100644
index 0000000..e638b9b
--- /dev/null
+++ b/drivers/net/arcvmac.h
@@ -0,0 +1,266 @@
+/*
+ * linux/arch/arc/drivers/arcvmac.h
+ *
+ * Copyright (C) 2003-2006 Codito Technologies, for linux-2.4 port
+ * Copyright (C) 2006-2007 Celunite Inc, for linux-2.6 port
+ * Copyright (C) 2007-2008 Sagem Communications, Fehmi HAFSI
+ * Copyright (C) 2009 Sagem Communications, Andreas Fenkart
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ *
+ */
+
+#ifndef _ARCVMAC_H
+#define _ARCVMAC_H
+
+#define DRV_NAME "arcvmac"
+#define DRV_VERSION "1.0"
+
+/* Buffer descriptors */
+#define TX_BDT_LEN 16 /* Number of receive BD's */
+#define RX_BDT_LEN 256 /* Number of transmit BD's */
+
+/* BD poll rate, in 1024 cycles. @100Mhz: x * 1024 cy * 10ns = 1ms */
+#define POLLRATE_TIME 200
+
+/* next power of two, bigger than ETH_FRAME_LEN + VLAN */
+#define MAX_RX_BUFFER_LEN 0x800 /* 2^11 = 2048 = 0x800 */
+#define MAX_TX_BUFFER_LEN 0x800 /* 2^11 = 2048 = 0x800 */
+
+/* 14 bytes of ethernet header, 4 bytes VLAN, FCS,
+ * plus extra pad to prevent buffer chaining of
+ * maximum sized ethernet packets (1514 bytes) */
+#define VMAC_BUFFER_PAD (ETH_HLEN + 4 + ETH_FCS_LEN + 4)
+
+/* VMAC register definitions, offsets in bytes */
+#define VMAC_ID 0x00
+
+/* stat/enable use same bit mask */
+#define VMAC_STAT 0x04
+#define VMAC_ENABLE 0x08
+# define TXINT_MASK 0x00000001 /* Transmit interrupt */
+# define RXINT_MASK 0x00000002 /* Receive interrupt */
+# define ERR_MASK 0x00000004 /* Error interrupt */
+# define TXCH_MASK 0x00000008 /* Transmit chaining error */
+# define MSER_MASK 0x00000010 /* Missed packet counter error */
+# define RXCR_MASK 0x00000100 /* RXCRCERR counter rolled over */
+# define RXFR_MASK 0x00000200 /* RXFRAMEERR counter rolled over */
+# define RXFL_MASK 0x00000400 /* RXOFLOWERR counter rolled over */
+# define MDIO_MASK 0x00001000 /* MDIO complete */
+# define TXPL_MASK 0x80000000 /* TXPOLL */
+
+#define VMAC_CONTROL 0x0c
+# define EN_MASK 0x00000001 /* VMAC enable */
+# define TXRN_MASK 0x00000008 /* TX enable */
+# define RXRN_MASK 0x00000010 /* RX enable */
+# define DSBC_MASK 0x00000100 /* Disable receive broadcast */
+# define ENFL_MASK 0x00000400 /* Enable Full Duplex */
+# define PROM_MASK 0x00000800 /* Promiscuous mode */
+
+#define VMAC_POLLRATE 0x10
+
+#define VMAC_RXERR 0x14
+# define RXERR_CRC 0x000000ff
+# define RXERR_FRM 0x0000ff00
+# define RXERR_OFLO 0x00ff0000 /* fifo overflow */
+
+#define VMAC_MISS 0x18
+#define VMAC_TXRINGPTR 0x1c
+#define VMAC_RXRINGPTR 0x20
+#define VMAC_ADDRL 0x24
+#define VMAC_ADDRH 0x28
+#define VMAC_LAFL 0x2c
+#define VMAC_LAFH 0x30
+#define VMAC_MAC_TXRING_HEAD 0x38
+#define VMAC_MAC_RXRING_HEAD 0x3C
+
+#define VMAC_MDIO_DATA 0x34
+# define MDIO_SFD 0xC0000000
+# define MDIO_OP 0x30000000
+# define MDIO_ID_MASK 0x0F800000
+# define MDIO_REG_MASK 0x007C0000
+# define MDIO_TA 0x00030000
+# define MDIO_DATA_MASK 0x0000FFFF
+/* common combinations */
+# define MDIO_BASE 0x40020000
+# define MDIO_OP_READ 0x20000000
+# define MDIO_OP_WRITE 0x10000000
+
+/* Buffer descriptor INFO bit masks */
+#define BD_DMA_OWN 0x80000000 /* buffer ownership, 0 CPU, 1 DMA */
+#define BD_BUFF 0x40000000 /* buffer invalid, rx */
+#define BD_UFLO 0x20000000 /* underflow, tx */
+#define BD_LTCL 0x10000000 /* late collision, tx */
+#define BD_RETRY_CT 0x0f000000 /* tx */
+#define BD_DROP 0x00800000 /* drop, more than 16 retries, tx */
+#define BD_DEFER 0x00400000 /* traffic on the wire, tx */
+#define BD_CARLOSS 0x00200000 /* carrier loss while transmission, tx, rx? */
+/* 20:19 reserved */
+#define BD_ADCR 0x00040000 /* add crc, ignored if not disaddcrc */
+#define BD_LAST 0x00020000 /* Last buffer in chain */
+#define BD_FRST 0x00010000 /* First buffer in chain */
+/* 15:11 reserved */
+#define BD_LEN 0x000007FF
+
+/* common combinations */
+#define BD_TX_ERR (BD_UFLO | BD_LTCL | BD_RETRY_CT | BD_DROP | \
+ BD_DEFER | BD_CARLOSS)
+
+
+/* arcvmac private data structures */
+struct vmac_buffer_desc {
+ __le32 info;
+ __le32 data;
+};
+
+struct dma_fifo {
+ int head; /* head */
+ int tail; /* tail */
+ int size;
+};
+
+struct vmac_priv {
+ struct net_device *dev;
+ struct platform_device *pdev;
+ struct net_device_stats stats;
+
+ struct completion mdio_complete;
+ spinlock_t lock; /* protects structure plus hw regs of device */
+
+ /* base address of register set */
+ char *regs;
+ struct resource *mem;
+
+ /* DMA ring buffers */
+ struct vmac_buffer_desc *rxbd;
+ dma_addr_t rxbd_dma;
+
+ struct vmac_buffer_desc *txbd;
+ dma_addr_t txbd_dma;
+
+ /* socket buffers */
+ struct sk_buff *rx_skbuff[RX_BDT_LEN];
+ struct sk_buff *tx_skbuff[TX_BDT_LEN];
+ int rx_skb_size;
+
+ /* skb / dma desc managing */
+ struct dma_fifo rx_ring; /* valid rx buffers */
+ struct dma_fifo tx_ring;
+
+ /* descriptor last polled/processed by the VMAC */
+ unsigned long dma_rx_head;
+
+ /* timer to retry rx skb allocation, if failed during receive */
+ struct timer_list refill_timer;
+ spinlock_t refill_lock;
+
+ struct napi_struct napi;
+
+ /* rx buffer chaining */
+ int rx_merge_error;
+ int tx_timeout_error;
+
+ /* PHY stuff */
+ struct mii_bus *mii_bus;
+ struct phy_device *phy_dev;
+
+ int link;
+ int speed;
+ int duplex;
+
+ /* debug */
+ int shutdown;
+};
+
+/* DMA ring management */
+
+/* for a fifo with size n,
+ * - [0..n] fill levels are n + 1 states
+ * - there are only n different deltas (head - tail) values
+ * => not all fill levels can be represented with head, tail
+ * pointers only
+ * we give up the n fill level, aka fifo full */
+
+/* sacrifice one elt as a sentinel */
+static inline int fifo_used(struct dma_fifo *f);
+static inline int fifo_inc_ct(int ct, int size);
+static inline void fifo_dump(struct dma_fifo *fifo);
+
+static inline int fifo_empty(struct dma_fifo *f)
+{
+ return (f->head == f->tail);
+}
+
+static inline int fifo_free(struct dma_fifo *f)
+{
+ int free;
+
+ free = f->tail - f->head;
+ if (free <= 0)
+ free += f->size;
+
+ return free;
+}
+
+static inline int fifo_used(struct dma_fifo *f)
+{
+ int used;
+
+ used = f->head - f->tail;
+ if (used < 0)
+ used += f->size;
+
+ return used;
+}
+
+static inline int fifo_full(struct dma_fifo *f)
+{
+ return (fifo_used(f) + 1) == f->size;
+}
+
+/* manipulate */
+static inline void fifo_init(struct dma_fifo *fifo, int size)
+{
+ fifo->size = size;
+ fifo->head = fifo->tail = 0; /* empty */
+}
+
+static inline void fifo_inc_head(struct dma_fifo *fifo)
+{
+ BUG_ON(fifo_full(fifo));
+ fifo->head = fifo_inc_ct(fifo->head, fifo->size);
+}
+
+static inline void fifo_inc_tail(struct dma_fifo *fifo)
+{
+ BUG_ON(fifo_empty(fifo));
+ fifo->tail = fifo_inc_ct(fifo->tail, fifo->size);
+}
+
+/* internal funcs */
+static inline void fifo_dump(struct dma_fifo *fifo)
+{
+ printk(KERN_INFO "fifo: head %d, tail %d, size %d\n", fifo->head,
+ fifo->tail,
+ fifo->size);
+}
+
+static inline int fifo_inc_ct(int ct, int size)
+{
+ return (++ct == size) ? 0 : ct;
+}
+
+#endif /* _ARCVMAC_H */
--
1.7.2.3
^ permalink raw reply related
* Re: [PATCH] arp_notify: unconditionally send gratuitous ARP for NETDEV_NOTIFY_PEERS.
From: Ian Campbell @ 2011-02-15 9:13 UTC (permalink / raw)
To: David Miller; +Cc: netdev@vger.kernel.org
In-Reply-To: <20110214.174654.115944453.davem@davemloft.net>
On Tue, 2011-02-15 at 01:46 +0000, David Miller wrote:
> From: Ian Campbell <ian.campbell@citrix.com>
> Date: Fri, 11 Feb 2011 17:44:16 +0000
>
> > NETDEV_NOTIFY_PEER is an explicit request by the driver to send a link
> > notification while NETDEV_UP/NETDEV_CHANGEADDR generate link
> > notifications as a sort of side effect.
> >
> > In the later cases the sysctl option is present because link
> > notification events can have undesired effects e.g. if the link is
> > flapping. I don't think this applies in the case of an explicit
> > request from a driver.
> >
> > This patch makes NETDEV_NOTIFY_PEER unconditional, if preferred we
> > could add a new sysctl for this case which defaults to on.
> >
> > This change causes Xen post-migration ARP notifications (which cause
> > switches to relearn their MAC tables etc) to be sent by default.
> >
> > Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
>
> Ok, applied, thanks.
Thanks.
I nearly forgot -- the NETDEV_NOTIFY_PEER stuff was tagged for
stable/longterm backport (it appeared in 2.6.32.19 or so). I think this
change should likewise go back, what do you think?
Ian.
^ permalink raw reply
* (unknown),
From: Western Union Transfer @ 2011-02-15 1:24 UTC (permalink / raw)
We have been trying to reach you since the past few days,
as my associate has helped me to send your first payment
of $7,500 USD to you as instructed by Mr. David Cameron
the United Kingdom prime minister after the last G20
meeting that was held in United Kingdom, making you one
of the beneficiaries. Here is the information below.
MONEY TRANSFER CONTROL NUMBER: 3928738492
SENDER NAME:Solomon Daniel
AMOUNT: $7,500 USD
I told him to keep sending you $7,500 USD twice a week
until the FULL payment of ($360.000 United State Dollars)
is completed.
Note a certificate will be made to change the Receiver Name as
stated by the British prime minister, send your Full Names
and your direct phone contact via Email to: Mr Gary Moore
The money will not reflect until the clearance certificate is
issue to you by the G20 committee.
contact Mr. Gary Moore for your clearance certificate.
Mr Gary Moore
E-mail:western_uniontransfer09@zbavitu.net
D/L:+44 7024018331
^ permalink raw reply
* (unknown),
From: Western Union Transfer @ 2011-02-15 1:23 UTC (permalink / raw)
We have been trying to reach you since the past few days,
as my associate has helped me to send your first payment
of $7,500 USD to you as instructed by Mr. David Cameron
the United Kingdom prime minister after the last G20
meeting that was held in United Kingdom, making you one
of the beneficiaries. Here is the information below.
MONEY TRANSFER CONTROL NUMBER: 3928738492
SENDER NAME:Solomon Daniel
AMOUNT: $7,500 USD
I told him to keep sending you $7,500 USD twice a week
until the FULL payment of ($360.000 United State Dollars)
is completed.
Note a certificate will be made to change the Receiver Name as
stated by the British prime minister, send your Full Names
and your direct phone contact via Email to: Mr Gary Moore
The money will not reflect until the clearance certificate is
issue to you by the G20 committee.
contact Mr. Gary Moore for your clearance certificate.
Mr Gary Moore
E-mail:western_uniontransfer09@zbavitu.net
D/L:+44 7024018331
^ permalink raw reply
* Re: [PATCH 02/14] net/fec: release mem_region requested in probe in error path and remove
From: Uwe Kleine-König @ 2011-02-15 8:53 UTC (permalink / raw)
To: David Miller; +Cc: netdev, shawn.guo, kernel
In-Reply-To: <20110214.110549.193710827.davem@davemloft.net>
Hi David,
On Mon, Feb 14, 2011 at 11:05:49AM -0800, David Miller wrote:
> From: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> Date: Mon, 14 Feb 2011 09:25:25 +0100
> > On Sun, Feb 13, 2011 at 01:15:31PM -0800, David Miller wrote:
> >> From: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> >> Date: Sun, 13 Feb 2011 22:07:09 +0100
> >> > On Fri, Feb 11, 2011 at 09:25:32PM -0800, David Miller wrote:
> >> >> I can't pull from that tree because it is _NOT_ based upon net-next-2.6
> >> >> and therefore brings in all kinds of commits not related to your work.
> >> > Sorry, I'm not used to the customs on netdev. I can rebase, but still I
> >> > wonder about the reason you cannot pull for. The only reason I can
> >> > imagine is that you fear unrelated breakage when taking these patches
> >> > that are already in Linus' tree. But if it's that, wouldn't it be great
> >> > the realize this breakage already now and not only during the next merge
> >> > window?
> >>
> >> My trees only merge in Linus's tree when absolutely necessary,
> >> to resolve conflicts or similar.
> > You don't merge Linus' tree, you merge mine that just happen to be based
> > on a newer version of Linus' tree. I'm sure Linus won't yell on you for
> > that. He only objects to merge directly from his tree, because the
> > result for him is an "empty" merge.
>
> You don't get it.
>
> These merge commits look ugly and Linus wants them minimized.
Hmm, right, I don't get why this looks uglier for Linus than a merge of
a tree that bases on something you already have. I guess you're too
annoyed by now to explain why you think it does.
> Either you follow the rules and my expectations, which is that when you
> give me a GIT tree to pull from it's based upon one of my trees, or
> I don't pull from you.
So I rebased my tree on something older. It now starts at
c69b909 (pch_can: fix module reload issue with MSI)
which is already in net-next/master.
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-König |
Industrial Linux Solutions | http://www.pengutronix.de/ |
^ permalink raw reply
* [PATCH v2 5/5] OMAP: hsmmc: Enable MMC4 and MMC5 on OMAP4 platforms
From: Panduranga Mallireddy @ 2011-02-15 8:40 UTC (permalink / raw)
To: coelho, netdev, linux-omap, linux-mmc
Cc: ohad, benzyg, pradeepgurumath, vishalm, x-boudet, naveen_jain,
pavan_savoy, manjunatha_halli, tony, cjb, Kishore Kadiyala,
Panduranga Mallireddy
In-Reply-To: <1297759236-25323-5-git-send-email-panduranga_mallireddy@ti.com>
From: Kishore Kadiyala <kishore.kadiyala@ti.com>
OMAP4 supports up to 5 MMC controllers, but only 3 of these were
initialized. MMC5 is used by wl12xx chip. So initialize MMC4 and MMC5.
Signed-off-by: Kishore Kadiyala <kishore.kadiyala@ti.com>
Signed-off-by: Panduranga Mallireddy <panduranga_mallireddy@ti.com>
---
arch/arm/mach-omap2/hsmmc.c | 5 +++++
drivers/mmc/host/omap_hsmmc.c | 24 ++++++++++++++++++++----
2 files changed, 25 insertions(+), 4 deletions(-)
diff --git a/arch/arm/mach-omap2/hsmmc.c b/arch/arm/mach-omap2/hsmmc.c
index 8f1a484..3c0809f 100644
--- a/arch/arm/mach-omap2/hsmmc.c
+++ b/arch/arm/mach-omap2/hsmmc.c
@@ -348,6 +348,11 @@ void __init omap2_hsmmc_init(struct omap2_hsmmc_info *controllers)
mmc->slots[0].after_set_reg = NULL;
}
break;
+ case 4:
+ case 5:
+ mmc->slots[0].before_set_reg = NULL;
+ mmc->slots[0].after_set_reg = NULL;
+ break;
default:
pr_err("MMC%d configuration not supported!\n", c->mmc);
kfree(mmc);
diff --git a/drivers/mmc/host/omap_hsmmc.c b/drivers/mmc/host/omap_hsmmc.c
index f59f8da..2525071 100644
--- a/drivers/mmc/host/omap_hsmmc.c
+++ b/drivers/mmc/host/omap_hsmmc.c
@@ -260,7 +260,7 @@ static int omap_hsmmc_1_set_power(struct device *dev, int slot, int power_on,
return ret;
}
-static int omap_hsmmc_23_set_power(struct device *dev, int slot, int power_on,
+static int omap_hsmmc_235_set_power(struct device *dev, int slot, int power_on,
int vdd)
{
struct omap_hsmmc_host *host =
@@ -316,6 +316,12 @@ static int omap_hsmmc_23_set_power(struct device *dev, int slot, int power_on,
return ret;
}
+static int omap_hsmmc_4_set_power(struct device *dev, int slot, int power_on,
+ int vdd)
+{
+ return 0;
+}
+
static int omap_hsmmc_1_set_sleep(struct device *dev, int slot, int sleep,
int vdd, int cardsleep)
{
@@ -326,7 +332,7 @@ static int omap_hsmmc_1_set_sleep(struct device *dev, int slot, int sleep,
return regulator_set_mode(host->vcc, mode);
}
-static int omap_hsmmc_23_set_sleep(struct device *dev, int slot, int sleep,
+static int omap_hsmmc_235_set_sleep(struct device *dev, int slot, int sleep,
int vdd, int cardsleep)
{
struct omap_hsmmc_host *host =
@@ -365,6 +371,12 @@ static int omap_hsmmc_23_set_sleep(struct device *dev, int slot, int sleep,
return regulator_enable(host->vcc_aux);
}
+static int omap_hsmmc_4_set_sleep(struct device *dev, int slot, int sleep,
+ int vdd, int cardsleep)
+{
+ return 0;
+}
+
static int omap_hsmmc_reg_get(struct omap_hsmmc_host *host)
{
struct regulator *reg;
@@ -379,10 +391,14 @@ static int omap_hsmmc_reg_get(struct omap_hsmmc_host *host)
break;
case OMAP_MMC2_DEVID:
case OMAP_MMC3_DEVID:
+ case OMAP_MMC5_DEVID:
/* Off-chip level shifting, or none */
- mmc_slot(host).set_power = omap_hsmmc_23_set_power;
- mmc_slot(host).set_sleep = omap_hsmmc_23_set_sleep;
+ mmc_slot(host).set_power = omap_hsmmc_235_set_power;
+ mmc_slot(host).set_sleep = omap_hsmmc_235_set_sleep;
break;
+ case OMAP_MMC4_DEVID:
+ mmc_slot(host).set_power = omap_hsmmc_4_set_power;
+ mmc_slot(host).set_sleep = omap_hsmmc_4_set_sleep;
default:
pr_err("MMC%d configuration not supported!\n", host->id);
return -EINVAL;
--
1.5.6.3
^ permalink raw reply related
* [PATCH v2 4/5] omap: panda: add mmc5/wl1271 device support
From: Panduranga Mallireddy @ 2011-02-15 8:40 UTC (permalink / raw)
To: coelho, netdev, linux-omap, linux-mmc
Cc: ohad, benzyg, pradeepgurumath, vishalm, x-boudet, naveen_jain,
pavan_savoy, manjunatha_halli, tony, cjb, Panduranga Mallireddy
In-Reply-To: <1297759236-25323-4-git-send-email-panduranga_mallireddy@ti.com>
Add MMC5 support on PANDA, which has the wl1271 device hardwired to.
The wl1271 is a 4-wire, 1.8V, embedded SDIO WLAN device with an
external IRQ line, and power-controlled by a GPIO-based fixed regulator.
Based on the patch for mmc3/wl1271 device support for zoom by Ohad
Ben-Cohen <ohad@wizery.com>
Signed-off-by: Panduranga Mallireddy <panduranga_mallireddy@ti.com>
---
arch/arm/mach-omap2/board-omap4panda.c | 20 ++++++++++++++++++++
1 files changed, 20 insertions(+), 0 deletions(-)
diff --git a/arch/arm/mach-omap2/board-omap4panda.c b/arch/arm/mach-omap2/board-omap4panda.c
index cd25255..ef43010 100644
--- a/arch/arm/mach-omap2/board-omap4panda.c
+++ b/arch/arm/mach-omap2/board-omap4panda.c
@@ -27,6 +27,7 @@
#include <linux/i2c/twl.h>
#include <linux/regulator/machine.h>
#include <linux/regulator/fixed.h>
+#include <linux/wl12xx.h>
#include <mach/hardware.h>
#include <mach/omap4-common.h>
@@ -47,6 +48,7 @@
#define GPIO_HUB_POWER 1
#define GPIO_HUB_NRESET 62
#define GPIO_WIFI_PMENA 43
+#define GPIO_WIFI_IRQ 53
static struct gpio_led gpio_leds[] = {
{
@@ -164,6 +166,15 @@ static struct omap2_hsmmc_info mmc[] = {
.gpio_wp = -EINVAL,
.gpio_cd = -EINVAL,
},
+ {
+ .name = "wl1271",
+ .mmc = 5,
+ .caps = MMC_CAP_4_BIT_DATA | MMC_CAP_POWER_OFF_CARD,
+ .gpio_wp = -EINVAL,
+ .gpio_cd = -EINVAL,
+ .ocr_mask = MMC_VDD_165_195,
+ .nonremovable = true,
+ },
{} /* Terminator */
};
@@ -205,6 +216,12 @@ static struct platform_device omap_vwlan_device = {
},
};
+struct wl12xx_platform_data omap_panda_wlan_data __initdata = {
+ .irq = OMAP_GPIO_IRQ(GPIO_WIFI_IRQ),
+ /* PANDA ref clock is 38.4 MHz */
+ .board_ref_clock = 2,
+};
+
static int omap4_twl6030_hsmmc_late_init(struct device *dev)
{
int ret = 0;
@@ -461,6 +478,9 @@ static void __init omap4_panda_init(void)
package = OMAP_PACKAGE_CBL;
omap4_mux_init(board_mux, package);
+ if (wl12xx_set_platform_data(&omap_panda_wlan_data))
+ pr_err("error setting wl12xx data\n");
+
omap4_panda_i2c_init();
platform_add_devices(panda_devices, ARRAY_SIZE(panda_devices));
platform_device_register(&omap_vwlan_device);
--
1.5.6.3
^ permalink raw reply related
* [PATCH v2 2/5] omap: select REGULATOR_FIXED_VOLTAGE by default for panda and sdp4430
From: Panduranga Mallireddy @ 2011-02-15 8:40 UTC (permalink / raw)
To: coelho, netdev, linux-omap, linux-mmc
Cc: ohad, benzyg, pradeepgurumath, vishalm, x-boudet, naveen_jain,
pavan_savoy, manjunatha_halli, tony, cjb, Panduranga Mallireddy
In-Reply-To: <1297759236-25323-2-git-send-email-panduranga_mallireddy@ti.com>
Power to the wl12xx wlan device is controlled by a fixed regulator.
Boards that have the wl12xx should select REGULATOR_FIXED_VOLTAGE.
Signed-off-by: Panduranga Mallireddy <panduranga_mallireddy@ti.com>
---
arch/arm/mach-omap2/Kconfig | 2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/arch/arm/mach-omap2/Kconfig b/arch/arm/mach-omap2/Kconfig
index 1a2cf62..eeaeb3b 100644
--- a/arch/arm/mach-omap2/Kconfig
+++ b/arch/arm/mach-omap2/Kconfig
@@ -300,6 +300,7 @@ config MACH_OMAP_4430SDP
depends on ARCH_OMAP4
select OMAP_PACKAGE_CBL
select OMAP_PACKAGE_CBS
+ select REGULATOR_FIXED_VOLTAGE
config MACH_OMAP4_PANDA
bool "OMAP4 Panda Board"
@@ -307,6 +308,7 @@ config MACH_OMAP4_PANDA
depends on ARCH_OMAP4
select OMAP_PACKAGE_CBL
select OMAP_PACKAGE_CBS
+ select REGULATOR_FIXED_VOLTAGE
config OMAP3_EMU
bool "OMAP3 debugging peripherals"
--
1.5.6.3
^ permalink raw reply related
* [PATCH v2 1/5] omap: panda: wlan board muxing
From: Panduranga Mallireddy @ 2011-02-15 8:40 UTC (permalink / raw)
To: coelho, netdev, linux-omap, linux-mmc
Cc: ohad, benzyg, pradeepgurumath, vishalm, x-boudet, naveen_jain,
pavan_savoy, manjunatha_halli, tony, cjb, Panduranga Mallireddy
In-Reply-To: <1297759236-25323-1-git-send-email-panduranga_mallireddy@ti.com>
Add board muxing to support the wlan wl1271 chip that is
hardwired to mmc5 (fifth mmc controller) on the PANDA.
Based on the wlan board muxing for zoom3 by Ohad Ben-Cohen
<ohadb@ti.com>
Signed-off-by: Panduranga Mallireddy <panduranga_mallireddy@ti.com>
---
arch/arm/mach-omap2/board-omap4panda.c | 13 +++++++++++++
1 files changed, 13 insertions(+), 0 deletions(-)
diff --git a/arch/arm/mach-omap2/board-omap4panda.c b/arch/arm/mach-omap2/board-omap4panda.c
index 9218862..cc9df6c 100644
--- a/arch/arm/mach-omap2/board-omap4panda.c
+++ b/arch/arm/mach-omap2/board-omap4panda.c
@@ -401,6 +401,19 @@ static int __init omap4_panda_i2c_init(void)
#ifdef CONFIG_OMAP_MUX
static struct omap_board_mux board_mux[] __initdata = {
+ /* WLAN IRQ - GPIO 53 */
+ OMAP4_MUX(GPMC_NCS3, OMAP_MUX_MODE3 | OMAP_PIN_INPUT),
+ /* WLAN POWER ENABLE - GPIO 43 */
+ OMAP4_MUX(GPMC_A19, OMAP_MUX_MODE3 | OMAP_PIN_OUTPUT),
+ /* WLAN SDIO: MMC5 CMD */
+ OMAP4_MUX(SDMMC5_CMD, OMAP_MUX_MODE0 | OMAP_PIN_INPUT_PULLUP),
+ /* WLAN SDIO: MMC5 CLK */
+ OMAP4_MUX(SDMMC5_CLK, OMAP_MUX_MODE0 | OMAP_PIN_INPUT_PULLUP),
+ /* WLAN SDIO: MMC5 DAT[0-3] */
+ OMAP4_MUX(SDMMC5_DAT0, OMAP_MUX_MODE0 | OMAP_PIN_INPUT_PULLUP),
+ OMAP4_MUX(SDMMC5_DAT1, OMAP_MUX_MODE0 | OMAP_PIN_INPUT_PULLUP),
+ OMAP4_MUX(SDMMC5_DAT2, OMAP_MUX_MODE0 | OMAP_PIN_INPUT_PULLUP),
+ OMAP4_MUX(SDMMC5_DAT3, OMAP_MUX_MODE0 | OMAP_PIN_INPUT_PULLUP),
{ .reg_offset = OMAP_MUX_TERMINATOR },
};
#else
--
1.5.6.3
^ permalink raw reply related
* [PATCH v2 0/5] Panda: Support for WLAN on WL127x
From: Panduranga Mallireddy @ 2011-02-15 8:40 UTC (permalink / raw)
To: coelho, netdev, linux-omap, linux-mmc
Cc: ohad, benzyg, pradeepgurumath, vishalm, x-boudet, naveen_jain,
pavan_savoy, manjunatha_halli, tony, cjb, Panduranga Mallireddy
Fixes from V1:
1. Removing the pull up of WLAN IRQ line, since it is always held up by wl127x device.
Adding support for WLAN on Panda board using wl12xx and mac80211 drivers
Patches were tested on Panda board.
These patches add software control for emulating card detect events,
add board configurations to support the wl127x device.
These patches are dependent on the patches that enable clock for WLAN on panda.
Please refer the following links for these patches:
https://patchwork.kernel.org/patch/546381/
https://patchwork.kernel.org/patch/546401/
[PATCH 5/5] OMAP: hsmmc: Enable MMC4 and MMC5 on OMAP4 platforms
is dependent on hwmod adpotation for HSMMC patches series.
For original patches of adpotation for HSMMC refer:
http://www.mail-archive.com/linux-omap@vger.kernel.org/msg43464.html
Patches are based on mainline 2.6.38-rc3
Kishore Kadiyala (1):
OMAP: hsmmc: Enable MMC4 and MMC5 on OMAP4 platforms
Panduranga Mallireddy (4):
omap: panda: wlan board muxing
omap: select REGULATOR_FIXED_VOLTAGE by default for panda and sdp4430
omap: panda: add fixed regulator device for wlan
omap: panda: add mmc5/wl1271 device support
arch/arm/mach-omap2/Kconfig | 2 +
arch/arm/mach-omap2/board-omap4panda.c | 67 ++++++++++++++++++++++++++++++++
arch/arm/mach-omap2/hsmmc.c | 5 ++
drivers/mmc/host/omap_hsmmc.c | 24 +++++++++--
4 files changed, 94 insertions(+), 4 deletions(-)
^ permalink raw reply
* [PATCH v2 3/5] omap: panda: add fixed regulator device for wlan
From: Panduranga Mallireddy @ 2011-02-15 8:40 UTC (permalink / raw)
To: coelho, netdev, linux-omap, linux-mmc
Cc: ohad, benzyg, pradeepgurumath, vishalm, x-boudet, naveen_jain,
pavan_savoy, manjunatha_halli, tony, cjb, Panduranga Mallireddy
In-Reply-To: <1297759236-25323-3-git-send-email-panduranga_mallireddy@ti.com>
Add a fixed regulator vmmc device to enable power control
of the wl1271 wlan device.
Based on the patch for zoom by Ohad Ben-Cohen <ohad@wizery.com>
Signed-off-by: Panduranga Mallireddy <panduranga_mallireddy@ti.com>
---
arch/arm/mach-omap2/board-omap4panda.c | 34 ++++++++++++++++++++++++++++++++
1 files changed, 34 insertions(+), 0 deletions(-)
diff --git a/arch/arm/mach-omap2/board-omap4panda.c b/arch/arm/mach-omap2/board-omap4panda.c
index cc9df6c..cd25255 100644
--- a/arch/arm/mach-omap2/board-omap4panda.c
+++ b/arch/arm/mach-omap2/board-omap4panda.c
@@ -26,6 +26,7 @@
#include <linux/usb/otg.h>
#include <linux/i2c/twl.h>
#include <linux/regulator/machine.h>
+#include <linux/regulator/fixed.h>
#include <mach/hardware.h>
#include <mach/omap4-common.h>
@@ -45,6 +46,7 @@
#define GPIO_HUB_POWER 1
#define GPIO_HUB_NRESET 62
+#define GPIO_WIFI_PMENA 43
static struct gpio_led gpio_leds[] = {
{
@@ -172,6 +174,37 @@ static struct regulator_consumer_supply omap4_panda_vmmc_supply[] = {
},
};
+static struct regulator_consumer_supply omap4_panda_vmmc5_supply = {
+ .supply = "vmmc",
+ .dev_name = "mmci-omap-hs.4",
+};
+
+static struct regulator_init_data panda_vmmc5 = {
+ .constraints = {
+ .valid_ops_mask = REGULATOR_CHANGE_STATUS,
+ },
+ .num_consumer_supplies = 1,
+ .consumer_supplies = &omap4_panda_vmmc5_supply,
+};
+
+static struct fixed_voltage_config panda_vwlan = {
+ .supply_name = "vwl1271",
+ .microvolts = 1800000, /* 1.8V */
+ .gpio = GPIO_WIFI_PMENA,
+ .startup_delay = 70000, /* 70msec */
+ .enable_high = 1,
+ .enabled_at_boot = 0,
+ .init_data = &panda_vmmc5,
+};
+
+static struct platform_device omap_vwlan_device = {
+ .name = "reg-fixed-voltage",
+ .id = 1,
+ .dev = {
+ .platform_data = &panda_vwlan,
+ },
+};
+
static int omap4_twl6030_hsmmc_late_init(struct device *dev)
{
int ret = 0;
@@ -430,6 +463,7 @@ static void __init omap4_panda_init(void)
omap4_panda_i2c_init();
platform_add_devices(panda_devices, ARRAY_SIZE(panda_devices));
+ platform_device_register(&omap_vwlan_device);
omap_serial_init();
omap4_twl6030_hsmmc_init(mmc);
/* OMAP4 Panda uses internal transceiver so register nop transceiver */
--
1.5.6.3
^ permalink raw reply related
* FINANCIAL LENDER
From: ZENITH HOME LENDER INC. @ 2011-02-15 6:40 UTC (permalink / raw)
To: zenithhlinc
apply for a financial lending amount to start your business our interest is very affordable
our lending process is very fast as well as percentage rate of 2.5% yearly from $5 000.00
min To $100 000 000.00 max
get back to us on this information below
amount, duration time,address and phone number
regards,
zenith home lender inc
zenithhlinc@gmail.com
mr ronald.c. foster
^ permalink raw reply
* Re: [Bugme-new] [Bug 27212] New: Warning kmemcheck: Caught 64-bit read from uninitialized memory in netlink_broadcast_filtered
From: Pekka Enberg @ 2011-02-15 5:48 UTC (permalink / raw)
To: Eric Dumazet
Cc: Andrew Morton, netdev, bugzilla-daemon, bugme-daemon,
casteyde.christian, Changli Gao, Vegard Nossum, David Miller,
linux-kernel, Christoph Lameter, David Rientjes
In-Reply-To: <1297704922.2996.60.camel@edumazet-laptop>
On Mon, Feb 14, 2011 at 7:35 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> Le vendredi 21 janvier 2011 à 09:49 +0200, Pekka Enberg a écrit :
>
>> It actually looks like a bug in SLUB+kmemcheck. The
>> kmemcheck_slab_alloc() call in slab_post_alloc_hook() should use ksize()
>> instead of s->objsize. SLAB seems to do the right thing already. Anyone
>> care to send a patch my way?
>>
>
> Hmm, what do you think of following patch ?
>
> Thanks, and sorry for the delay.
Looks good to me. Christoph, David, any objections to the patch?
> [PATCH] slub: fix kmemcheck calls to match ksize() hints
>
> Recent use of ksize() in network stack (commit ca44ac38 : net: don't
> reallocate skb->head unless the current one hasn't the needed extra size
> or is shared) triggers kmemcheck warnings, because ksize() can return
> more space than kmemcheck is aware of.
>
> Pekka Enberg noticed SLAB+kmemcheck is doing the right thing, while SLUB
> +kmemcheck doesnt.
>
> Bugzilla reference #27212
>
> Reported-by: Christian Casteyde <casteyde.christian@free.fr>
> Suggested-by: Pekka Enberg <penberg@kernel.org>
> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
> CC: David Miller <davem@davemloft.net>
> CC: Changli Gao <xiaosuo@gmail.com>
> CC: Andrew Morton <akpm@linux-foundation.org>
> ---
> mm/slub.c | 49 ++++++++++++++++++++++++++-----------------------
> 1 file changed, 26 insertions(+), 23 deletions(-)
>
> diff --git a/mm/slub.c b/mm/slub.c
> index e15aa7f..ee0aeb8 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -797,10 +797,34 @@ static inline int slab_pre_alloc_hook(struct kmem_cache *s, gfp_t flags)
> return should_failslab(s->objsize, flags, s->flags);
> }
>
> +static inline size_t slab_ksize(const struct kmem_cache *s)
> +{
> +#ifdef CONFIG_SLUB_DEBUG
> + /*
> + * Debugging requires use of the padding between object
> + * and whatever may come after it.
> + */
> + if (s->flags & (SLAB_RED_ZONE | SLAB_POISON))
> + return s->objsize;
> +
> +#endif
> + /*
> + * If we have the need to store the freelist pointer
> + * back there or track user information then we can
> + * only use the space before that information.
> + */
> + if (s->flags & (SLAB_DESTROY_BY_RCU | SLAB_STORE_USER))
> + return s->inuse;
> + /*
> + * Else we can use all the padding etc for the allocation
> + */
> + return s->size;
> +}
> +
> static inline void slab_post_alloc_hook(struct kmem_cache *s, gfp_t flags, void *object)
> {
> flags &= gfp_allowed_mask;
> - kmemcheck_slab_alloc(s, flags, object, s->objsize);
> + kmemcheck_slab_alloc(s, flags, object, slab_ksize(s));
> kmemleak_alloc_recursive(object, s->objsize, 1, s->flags, flags);
> }
>
> @@ -2696,7 +2720,6 @@ EXPORT_SYMBOL(__kmalloc_node);
> size_t ksize(const void *object)
> {
> struct page *page;
> - struct kmem_cache *s;
>
> if (unlikely(object == ZERO_SIZE_PTR))
> return 0;
> @@ -2707,28 +2730,8 @@ size_t ksize(const void *object)
> WARN_ON(!PageCompound(page));
> return PAGE_SIZE << compound_order(page);
> }
> - s = page->slab;
>
> -#ifdef CONFIG_SLUB_DEBUG
> - /*
> - * Debugging requires use of the padding between object
> - * and whatever may come after it.
> - */
> - if (s->flags & (SLAB_RED_ZONE | SLAB_POISON))
> - return s->objsize;
> -
> -#endif
> - /*
> - * If we have the need to store the freelist pointer
> - * back there or track user information then we can
> - * only use the space before that information.
> - */
> - if (s->flags & (SLAB_DESTROY_BY_RCU | SLAB_STORE_USER))
> - return s->inuse;
> - /*
> - * Else we can use all the padding etc for the allocation
> - */
> - return s->size;
> + return slab_ksize(page->slab);
> }
> EXPORT_SYMBOL(ksize);
>
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
^ permalink raw reply
* Re: [RFC PATCH 0/5] Cache PMTU/redirects in inetpeer
From: David Miller @ 2011-02-15 5:46 UTC (permalink / raw)
To: netdev
In-Reply-To: <20110210.150957.115925131.davem@davemloft.net>
From: David Miller <davem@davemloft.net>
Date: Thu, 10 Feb 2011 15:09:57 -0800 (PST)
> From: David Miller <davem@davemloft.net>
> Date: Wed, 09 Feb 2011 22:12:49 -0800 (PST)
>
>> This is what I've been working on for the past several days.
>
> Just FYI I've pushed just the infrastructure bits (patches 1, 2,
> and 3) into net-next-2.6 since those have no side effects.
>
> I won't push the other two until I can do some good testing on
> them.
I've now pushed the final two patches from the original series
into net-next-2.6
Please if anyone runs into troubles, give a good bug report and I
will fix the problems as quickly as possible.
Thanks!
^ permalink raw reply
* Re: [PATCH] net: provide capability and group sets via SCM
From: David Miller @ 2011-02-15 1:49 UTC (permalink / raw)
To: casey
Cc: linux-kernel, netdev, ext-jarkko.2.sakkinen, Janne.Karhunen,
elena.reshetova
In-Reply-To: <4D51C38B.1060902@schaufler-ca.com>
From: Casey Schaufler <casey@schaufler-ca.com>
Date: Tue, 08 Feb 2011 14:28:27 -0800
>
> Subject: [PATCH] net: provide group lists and capability set via CMSG
>
> Provide the namespace converted group list of the peer
> process using the SCM mechanism. Provide the capability
> set of the peer process. The capability set is not
> namespace converted on the assumption that there is no
> such conversion available or required.
>
> Signed-off-by: Casey Schaufler <casey@schaufler-ca.com>
You can't just hit the asm-generic header, you have to also hit
all of the architectures that don't use the asm-generic header,
including sparc, powerpc, mips, s390, arm, alpha, cris, frv, h8300,
ia64, m32r, m68k, m68knommu, mn10300, parisc, and xtensa.
^ permalink raw reply
* Re: [PATCH] arp_notify: unconditionally send gratuitous ARP for NETDEV_NOTIFY_PEERS.
From: David Miller @ 2011-02-15 1:46 UTC (permalink / raw)
To: ian.campbell; +Cc: netdev
In-Reply-To: <1297446256-23917-1-git-send-email-ian.campbell@citrix.com>
From: Ian Campbell <ian.campbell@citrix.com>
Date: Fri, 11 Feb 2011 17:44:16 +0000
> NETDEV_NOTIFY_PEER is an explicit request by the driver to send a link
> notification while NETDEV_UP/NETDEV_CHANGEADDR generate link
> notifications as a sort of side effect.
>
> In the later cases the sysctl option is present because link
> notification events can have undesired effects e.g. if the link is
> flapping. I don't think this applies in the case of an explicit
> request from a driver.
>
> This patch makes NETDEV_NOTIFY_PEER unconditional, if preferred we
> could add a new sysctl for this case which defaults to on.
>
> This change causes Xen post-migration ARP notifications (which cause
> switches to relearn their MAC tables etc) to be sent by default.
>
> Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Ok, applied, thanks.
^ permalink raw reply
* Re: [PATCH] phy/micrel: add ability to support 50MHz RMII clock on KZS8051RNL
From: David Miller @ 2011-02-15 1:38 UTC (permalink / raw)
To: baruch; +Cc: netdev, david.choi
In-Reply-To: <e426eef4b15c4e1c77981094c3c8020012aaca42.1297685030.git.baruch@tkos.co.il>
From: Baruch Siach <baruch@tkos.co.il>
Date: Mon, 14 Feb 2011 14:05:33 +0200
> Platform code can now set the MICREL_PHY_50MHZ_CLK bit of dev_flags in a fixup
> routine (registered with phy_register_fixup_for_uid()), to make the KZS8051RNL
> PHY work with 50MHz RMII reference clock.
>
> Cc: David J. Choi <david.choi@micrel.com>
> Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Applied to net-next-2.6, thanks.
^ permalink raw reply
* Re: [PATCH net-26 0/4] cxgb4vf: minor bug fixes -- revised patch set
From: David Miller @ 2011-02-15 1:37 UTC (permalink / raw)
To: leedom; +Cc: netdev
In-Reply-To: <1297724185-27452-1-git-send-email-leedom@chelsio.com>
From: Casey Leedom <leedom@chelsio.com>
Date: Mon, 14 Feb 2011 14:56:21 -0800
> [01/04]: Check driver parameters in the right place ...
> [02/04]: Behave properly when CONFIG_DEBUG_FS isn't defined ...
> [03/04]: Quiesce Virtual Interfaces on shutdown ...
> [04/04]: Use defined Mailbox Timeout
All applied, thanks.
^ permalink raw reply
* (unknown),
From: Western Union Transfer @ 2011-02-14 23:21 UTC (permalink / raw)
We have been trying to reach you since the past few days,
as my associate has helped me to send your first payment
of $7,500 USD to you as instructed by Mr. David Cameron
the United Kingdom prime minister after the last G20
meeting that was held in United Kingdom, making you one
of the beneficiaries. Here is the information below.
MONEY TRANSFER CONTROL NUMBER: 3928738492
SENDER NAME:Solomon Daniel
AMOUNT: $7,500 USD
I told him to keep sending you $7,500 USD twice a week
until the FULL payment of ($360.000 United State Dollars)
is completed.
Note a certificate will be made to change the Receiver Name as
stated by the British prime minister, send your Full Names
and your direct phone contact via Email to: Mr Gary Moore
The money will not reflect until the clearance certificate is
issue to you by the G20 committee.
contact Mr. Gary Moore for your clearance certificate.
Mr Gary Moore
E-mail:western_uniontransfer09@zbavitu.net
D/L:+44 7024018331
^ permalink raw reply
* Re: Process for subsystem maintainers to get Hyper-V code out of staging. - CORRECTED RECIPIENTS
From: Robert Hancock @ 2011-02-15 0:45 UTC (permalink / raw)
To: Hank Janssen
Cc: shemminger@linux-foundation.org, netdev@vger.kernel.org,
davem@davemloft.net, ide, KY Srinivasan, Hashir Abdi,
Mike Sterling, Haiyang Zhang, gregkh@suse.de" "
In-Reply-To: <8AFC7968D54FB448A30D8F38F259C56233F957D3@TK5EX14MBXC114.redmond.corp.microsoft.com>
On 02/14/2011 05:42 PM, Hank Janssen wrote:
>
> MY APOLOGIES-I made a typo on James email address. I corrected it and resend.
> Sorry for the double email.
>
>
> Stephen/James/David,
>
> Greetings to you all. As you might be aware, we submitted Hyper-V drivers to the kernel 2009.
> We have been extending these drivers with additional functionality and our primary focus now
> is doing the work needed to exit the staging area.
>
> To give you some background, the following are Hyper-V specific Linux drivers:
>
> hv_vmbus The vmbus driver that is the bridge between guest and the
> host
> hv_storvsc The SCSI device driver
> hv_blkvsc The IDE driver
Given that the IDE subsystem (drivers/ide) is currently in
maintenance-only mode, and isn't used by modern distributions, you
likely want to make this a libata driver instead.
Though, from what's in current git, it's not clear to me what the HV IDE
(and SCSI) drivers are attempting to do. Is it really something that
looks like an IDE controller from the guest OS point of view? If not,
then having it as an IDE driver would be the wrong thing to do, it
should be more of a generic block driver. In that case, then, why are
there both SCSI and IDE drivers in the first place?
> hv_netvsc The network driver
>
> We think our drivers are pretty close to be reviewed by the subsystem maintainers.
>
> We have been working with Greg on hv_vmbus, and several other driver issues as it relates to
> exiting staging. And now we are looking for guidance for the other drivers.
>
> 1. Most important thing of course, did we contact the correct subsystem
> maintainers?
> i. IDE/Blkvsc David Miller
> ii. SCSI/Storvsc James Bottomley
> iii. Network/Netvsc Stephen Hemminger
> 2. What is the process to submit the code for review?
> 3. Which mailing list(s) do we need to use for the code reviews
> 4. I assume normal patch format is required?
> 5. What additional information is needed
> 6. Once they leave staging where do they need to go? Because they all
> pretty much come as a package we were thinking drivers/hyperv might
> be a good place.
>
> There are still a few outstanding items we are currently working on. But they should be
> wrapped up shortly. (There are a few remaining FIXME comments in the code we are
> cleaning up as I write this). But if possible we would like to get your feedback even
> as we are addressing the issues we currently know about.
>
> Thanks
>
> Hank.
>
>
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ide" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply
* Re: Process for subsystem maintainers to get Hyper-V code out of staging.
From: Greg KH @ 2011-02-14 23:45 UTC (permalink / raw)
To: Hank Janssen
Cc: shemminger@linux-foundation.org, netdev@vger.kernel.org,
davem@davemloft.net, linux-ide@vger.kernel.org,
Jame.Bottomley@HansenPartnership.com, linux-scsi@vger.kernel.org,
KY Srinivasan, Hashir Abdi, Mike Sterling, Haiyang Zhang
In-Reply-To: <8AFC7968D54FB448A30D8F38F259C56233F95737@TK5EX14MBXC114.redmond.corp.microsoft.com>
On Mon, Feb 14, 2011 at 11:30:07PM +0000, Hank Janssen wrote:
>
> Stephen/James/David,
>
> Greetings to you all. As you might be aware, we submitted Hyper-V drivers to the kernel 2009.
> We have been extending these drivers with additional functionality and our primary focus
> now is doing the work needed to exit the staging area.
>
> To give you some background, the following are Hyper-V specific Linux drivers:
>
> hv_vmbus The vmbus driver that is the bridge between guest and the
> host
> hv_storvsc The SCSI device driver
> hv_blkvsc The IDE driver
> hv_netvsc The network driver
>
> We think our drivers are pretty close to be reviewed by the subsystem maintainers.
>
> We have been working with Greg on hv_vmbus, and several other driver issues as it
> relates to exiting staging. And now we are looking for guidance for the other drivers.
>
> 1. Most important thing of course, did we contact the correct subsystem
> maintainers?
> i. IDE/Blkvsc David Miller
> ii. SCSI/Storvsc James Bottomley
> iii. Network/Netvsc Stephen Hemminger
That's what the MAINTAINERS file says, right?
> 2. What is the process to submit the code for review?
Like Documentation/SubmittingPatches shows, send patches.
> 3. Which mailing list(s) do we need to use for the code reviews
Again, MAINTAINERS shows this.
> 4. I assume normal patch format is required?
Yes.
> 5. What additional information is needed
What's normally needed.
> 6. Once they leave staging where do they need to go? Because they all
> pretty much come as a package we were thinking drivers/hyperv might
> be a good place.
That's up to the subsystem, if they want all network drivers in
drivers/net/ then they go there, same for scsi, it's up to the
maintainer.
> There are still a few outstanding items we are currently working on. But they should be
> wrapped up shortly. (There are a few remaining FIXME comments in the code we are
> cleaning up as I write this). But if possible we would like to get your feedback even as
> we are addressing the issues we currently know about.
I would wait and only ask for review _after_ you fix the things you know
about first. Otherwise it just wastes everyone's time.
thanks,
greg k-h
^ permalink raw reply
* RE: Process for subsystem maintainers to get Hyper-V code out of staging. - CORRECTED RECIPIENTS
From: Hank Janssen @ 2011-02-14 23:42 UTC (permalink / raw)
To: shemminger@linux-foundation.org, netdev@vger.kernel.org,
davem@davemloft.net, "linux-ide@vger.ke
Cc: KY Srinivasan, Hashir Abdi, Mike Sterling, Haiyang Zhang,
gregkh@suse.de
MY APOLOGIES-I made a typo on James email address. I corrected it and resend.
Sorry for the double email.
Stephen/James/David,
Greetings to you all. As you might be aware, we submitted Hyper-V drivers to the kernel 2009.
We have been extending these drivers with additional functionality and our primary focus now
is doing the work needed to exit the staging area.
To give you some background, the following are Hyper-V specific Linux drivers:
hv_vmbus The vmbus driver that is the bridge between guest and the
host
hv_storvsc The SCSI device driver
hv_blkvsc The IDE driver
hv_netvsc The network driver
We think our drivers are pretty close to be reviewed by the subsystem maintainers.
We have been working with Greg on hv_vmbus, and several other driver issues as it relates to
exiting staging. And now we are looking for guidance for the other drivers.
1. Most important thing of course, did we contact the correct subsystem
maintainers?
i. IDE/Blkvsc David Miller
ii. SCSI/Storvsc James Bottomley
iii. Network/Netvsc Stephen Hemminger
2. What is the process to submit the code for review?
3. Which mailing list(s) do we need to use for the code reviews
4. I assume normal patch format is required?
5. What additional information is needed
6. Once they leave staging where do they need to go? Because they all
pretty much come as a package we were thinking drivers/hyperv might
be a good place.
There are still a few outstanding items we are currently working on. But they should be
wrapped up shortly. (There are a few remaining FIXME comments in the code we are
cleaning up as I write this). But if possible we would like to get your feedback even
as we are addressing the issues we currently know about.
Thanks
Hank.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox