/*************************************************************************
* myri10ge.c: Myricom Myri-10G Ethernet driver.
*
- * Copyright (C) 2005, 2006 Myricom, Inc.
+ * Copyright (C) 2005 - 2007 Myricom, Inc.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
- * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
*
*
* If the eeprom on your board is not recent enough, you will need to get a
#include "myri10ge_mcp.h"
#include "myri10ge_mcp_gen_header.h"
-#define MYRI10GE_VERSION_STR "1.0.0"
+#define MYRI10GE_VERSION_STR "1.3.0-1.233"
MODULE_DESCRIPTION("Myricom 10G driver (10GbE)");
MODULE_AUTHOR("Maintainer: help@myri.com");
#define MYRI10GE_MAX_FRAGS_PER_FRAME (MYRI10GE_MAX_ETHER_MTU/MYRI10GE_ALLOC_SIZE + 1)
struct myri10ge_rx_buffer_state {
- struct sk_buff *skb;
struct page *page;
int page_offset;
DECLARE_PCI_UNMAP_ADDR(bus)
int intr_coal_delay;
__be32 __iomem *intr_coal_delay_ptr;
int mtrr;
+ int wc_enabled;
int wake_queue;
int stop_queue;
int down_cnt;
char *fw_name;
char eeprom_strings[MYRI10GE_EEPROM_STRINGS_SIZE];
char fw_version[128];
+ int fw_ver_major;
+ int fw_ver_minor;
+ int fw_ver_tiny;
+ int adopted_rx_filter_bug;
u8 mac_addr[6]; /* eeprom mac address */
unsigned long serial_number;
int vendor_specific_offset;
int fw_multicast_support;
- u32 devctl;
- u16 msi_flags;
u32 read_dma;
u32 write_dma;
u32 read_write_dma;
MODULE_PARM_DESC(myri10ge_small_bytes, "Threshold of small packets\n");
static int myri10ge_msi = 1; /* enable msi by default */
-module_param(myri10ge_msi, int, S_IRUGO);
+module_param(myri10ge_msi, int, S_IRUGO | S_IWUSR);
MODULE_PARM_DESC(myri10ge_msi, "Enable Message Signalled Interrupts\n");
-static int myri10ge_intr_coal_delay = 25;
+static int myri10ge_intr_coal_delay = 75;
module_param(myri10ge_intr_coal_delay, int, S_IRUGO);
MODULE_PARM_DESC(myri10ge_intr_coal_delay, "Interrupt coalescing delay\n");
MODULE_PARM_DESC(myri10ge_force_firmware,
"Force firmware to assume aligned completions\n");
-static int myri10ge_skb_cross_4k = 0;
-module_param(myri10ge_skb_cross_4k, int, S_IRUGO | S_IWUSR);
-MODULE_PARM_DESC(myri10ge_skb_cross_4k,
- "Can a small skb cross a 4KB boundary?\n");
-
static int myri10ge_initial_mtu = MYRI10GE_MAX_ETHER_MTU - ETH_HLEN;
module_param(myri10ge_initial_mtu, int, S_IRUGO);
MODULE_PARM_DESC(myri10ge_initial_mtu, "Initial MTU\n");
module_param(myri10ge_fill_thresh, int, S_IRUGO | S_IWUSR);
MODULE_PARM_DESC(myri10ge_fill_thresh, "Number of empty rx slots allowed\n");
+static int myri10ge_reset_recover = 1;
+
+static int myri10ge_wcfifo = 0;
+module_param(myri10ge_wcfifo, int, S_IRUGO);
+MODULE_PARM_DESC(myri10ge_wcfifo, "Enable WC Fifo when WC is enabled\n");
+
#define MYRI10GE_FW_OFFSET 1024*1024
#define MYRI10GE_HIGHPART_TO_U32(X) \
(sizeof (X) == 8) ? ((u32)((u64)(X) >> 32)) : (0)
#define myri10ge_pio_copy(to,from,size) __iowrite64_copy(to,from,size/8)
+static void myri10ge_set_multicast_list(struct net_device *dev);
+
static inline void put_be32(__be32 val, __be32 __iomem * p)
{
__raw_writel((__force __u32) val, (__force void __iomem *)p);
return 0;
} else if (result == MXGEFW_CMD_UNKNOWN) {
return -ENOSYS;
+ } else if (result == MXGEFW_CMD_ERROR_UNALIGNED) {
+ return -E2BIG;
} else {
dev_err(&mgp->pdev->dev,
"command %d failed, result = %d\n",
struct mcp_gen_header *hdr)
{
struct device *dev = &mgp->pdev->dev;
- int major, minor;
/* check firmware type */
if (ntohl(hdr->mcp_type) != MCP_TYPE_ETH) {
/* save firmware version for ethtool */
strncpy(mgp->fw_version, hdr->version, sizeof(mgp->fw_version));
- sscanf(mgp->fw_version, "%d.%d", &major, &minor);
+ sscanf(mgp->fw_version, "%d.%d.%d", &mgp->fw_ver_major,
+ &mgp->fw_ver_minor, &mgp->fw_ver_tiny);
- if (!(major == MXGEFW_VERSION_MAJOR && minor == MXGEFW_VERSION_MINOR)) {
+ if (!(mgp->fw_ver_major == MXGEFW_VERSION_MAJOR
+ && mgp->fw_ver_minor == MXGEFW_VERSION_MINOR)) {
dev_err(dev, "Found firmware version %s\n", mgp->fw_version);
dev_err(dev, "Driver needs %d.%d\n", MXGEFW_VERSION_MAJOR,
MXGEFW_VERSION_MINOR);
memcpy_fromio(hdr, mgp->sram + hdr_offset, bytes);
status = myri10ge_validate_firmware(mgp, hdr);
kfree(hdr);
+
+ /* check to see if adopted firmware has bug where adopting
+ * it will cause broadcasts to be filtered unless the NIC
+ * is kept in ALLMULTI mode */
+ if (mgp->fw_ver_major == 1 && mgp->fw_ver_minor == 4 &&
+ mgp->fw_ver_tiny >= 4 && mgp->fw_ver_tiny <= 11) {
+ mgp->adopted_rx_filter_bug = 1;
+ dev_warn(dev, "Adopting fw %d.%d.%d: "
+ "working around rx filter bug\n",
+ mgp->fw_ver_major, mgp->fw_ver_minor,
+ mgp->fw_ver_tiny);
+ }
return status;
}
mgp->dev->name);
}
+static int myri10ge_dma_test(struct myri10ge_priv *mgp, int test_type)
+{
+ struct myri10ge_cmd cmd;
+ int status;
+ u32 len;
+ struct page *dmatest_page;
+ dma_addr_t dmatest_bus;
+ char *test = " ";
+
+ dmatest_page = alloc_page(GFP_KERNEL);
+ if (!dmatest_page)
+ return -ENOMEM;
+ dmatest_bus = pci_map_page(mgp->pdev, dmatest_page, 0, PAGE_SIZE,
+ DMA_BIDIRECTIONAL);
+
+ /* Run a small DMA test.
+ * The magic multipliers to the length tell the firmware
+ * to do DMA read, write, or read+write tests. The
+ * results are returned in cmd.data0. The upper 16
+ * bits or the return is the number of transfers completed.
+ * The lower 16 bits is the time in 0.5us ticks that the
+ * transfers took to complete.
+ */
+
+ len = mgp->tx.boundary;
+
+ cmd.data0 = MYRI10GE_LOWPART_TO_U32(dmatest_bus);
+ cmd.data1 = MYRI10GE_HIGHPART_TO_U32(dmatest_bus);
+ cmd.data2 = len * 0x10000;
+ status = myri10ge_send_cmd(mgp, test_type, &cmd, 0);
+ if (status != 0) {
+ test = "read";
+ goto abort;
+ }
+ mgp->read_dma = ((cmd.data0 >> 16) * len * 2) / (cmd.data0 & 0xffff);
+ cmd.data0 = MYRI10GE_LOWPART_TO_U32(dmatest_bus);
+ cmd.data1 = MYRI10GE_HIGHPART_TO_U32(dmatest_bus);
+ cmd.data2 = len * 0x1;
+ status = myri10ge_send_cmd(mgp, test_type, &cmd, 0);
+ if (status != 0) {
+ test = "write";
+ goto abort;
+ }
+ mgp->write_dma = ((cmd.data0 >> 16) * len * 2) / (cmd.data0 & 0xffff);
+
+ cmd.data0 = MYRI10GE_LOWPART_TO_U32(dmatest_bus);
+ cmd.data1 = MYRI10GE_HIGHPART_TO_U32(dmatest_bus);
+ cmd.data2 = len * 0x10001;
+ status = myri10ge_send_cmd(mgp, test_type, &cmd, 0);
+ if (status != 0) {
+ test = "read/write";
+ goto abort;
+ }
+ mgp->read_write_dma = ((cmd.data0 >> 16) * len * 2 * 2) /
+ (cmd.data0 & 0xffff);
+
+abort:
+ pci_unmap_page(mgp->pdev, dmatest_bus, PAGE_SIZE, DMA_BIDIRECTIONAL);
+ put_page(dmatest_page);
+
+ if (status != 0 && test_type != MXGEFW_CMD_UNALIGNED_TEST)
+ dev_warn(&mgp->pdev->dev, "DMA %s benchmark failed: %d\n",
+ test, status);
+
+ return status;
+}
+
static int myri10ge_reset(struct myri10ge_priv *mgp)
{
struct myri10ge_cmd cmd;
int status;
size_t bytes;
- u32 len;
/* try to send a reset command to the card to see if it
* is alive */
return -ENXIO;
}
+ (void)myri10ge_dma_test(mgp, MXGEFW_DMA_TEST);
+
/* Now exchange information about interrupts */
bytes = myri10ge_max_intr_slots * sizeof(*mgp->rx_done.entry);
status |=
myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_IRQ_ACK_OFFSET, &cmd, 0);
mgp->irq_claim = (__iomem __be32 *) (mgp->sram + cmd.data0);
- if (!mgp->msi_enabled) {
- status |= myri10ge_send_cmd
- (mgp, MXGEFW_CMD_GET_IRQ_DEASSERT_OFFSET, &cmd, 0);
- mgp->irq_deassert = (__iomem __be32 *) (mgp->sram + cmd.data0);
+ status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_IRQ_DEASSERT_OFFSET,
+ &cmd, 0);
+ mgp->irq_deassert = (__iomem __be32 *) (mgp->sram + cmd.data0);
- }
status |= myri10ge_send_cmd
(mgp, MXGEFW_CMD_GET_INTR_COAL_DELAY_OFFSET, &cmd, 0);
mgp->intr_coal_delay_ptr = (__iomem __be32 *) (mgp->sram + cmd.data0);
}
put_be32(htonl(mgp->intr_coal_delay), mgp->intr_coal_delay_ptr);
- /* Run a small DMA test.
- * The magic multipliers to the length tell the firmware
- * to do DMA read, write, or read+write tests. The
- * results are returned in cmd.data0. The upper 16
- * bits or the return is the number of transfers completed.
- * The lower 16 bits is the time in 0.5us ticks that the
- * transfers took to complete.
- */
-
- len = mgp->tx.boundary;
-
- cmd.data0 = MYRI10GE_LOWPART_TO_U32(mgp->rx_done.bus);
- cmd.data1 = MYRI10GE_HIGHPART_TO_U32(mgp->rx_done.bus);
- cmd.data2 = len * 0x10000;
- status = myri10ge_send_cmd(mgp, MXGEFW_DMA_TEST, &cmd, 0);
- if (status == 0)
- mgp->read_dma = ((cmd.data0 >> 16) * len * 2) /
- (cmd.data0 & 0xffff);
- else
- dev_warn(&mgp->pdev->dev, "DMA read benchmark failed: %d\n",
- status);
- cmd.data0 = MYRI10GE_LOWPART_TO_U32(mgp->rx_done.bus);
- cmd.data1 = MYRI10GE_HIGHPART_TO_U32(mgp->rx_done.bus);
- cmd.data2 = len * 0x1;
- status = myri10ge_send_cmd(mgp, MXGEFW_DMA_TEST, &cmd, 0);
- if (status == 0)
- mgp->write_dma = ((cmd.data0 >> 16) * len * 2) /
- (cmd.data0 & 0xffff);
- else
- dev_warn(&mgp->pdev->dev, "DMA write benchmark failed: %d\n",
- status);
-
- cmd.data0 = MYRI10GE_LOWPART_TO_U32(mgp->rx_done.bus);
- cmd.data1 = MYRI10GE_HIGHPART_TO_U32(mgp->rx_done.bus);
- cmd.data2 = len * 0x10001;
- status = myri10ge_send_cmd(mgp, MXGEFW_DMA_TEST, &cmd, 0);
- if (status == 0)
- mgp->read_write_dma = ((cmd.data0 >> 16) * len * 2 * 2) /
- (cmd.data0 & 0xffff);
- else
- dev_warn(&mgp->pdev->dev,
- "DMA read/write benchmark failed: %d\n", status);
-
memset(mgp->rx_done.entry, 0, bytes);
/* reset mcp/driver shared state back to 0 */
mgp->rx_done.cnt = 0;
mgp->link_changes = 0;
status = myri10ge_update_mac_address(mgp, mgp->dev->dev_addr);
- myri10ge_change_promisc(mgp, 0, 0);
myri10ge_change_pause(mgp, mgp->pause);
+ myri10ge_set_multicast_list(mgp->dev);
return status;
}
mb();
}
-/*
- * Set of routines to get a new receive buffer. Any buffer which
- * crosses a 4KB boundary must start on a 4KB boundary due to PCIe
- * wdma restrictions. We also try to align any smaller allocation to
- * at least a 16 byte boundary for efficiency. We assume the linux
- * memory allocator works by powers of 2, and will not return memory
- * smaller than 2KB which crosses a 4KB boundary. If it does, we fall
- * back to allocating 2x as much space as required.
- *
- * We intend to replace large (>4KB) skb allocations by using
- * pages directly and building a fraglist in the near future.
- */
-
-static inline struct sk_buff *myri10ge_alloc_big(struct net_device *dev,
- int bytes)
-{
- struct sk_buff *skb;
- unsigned long data, roundup;
-
- skb = netdev_alloc_skb(dev, bytes + 4096 + MXGEFW_PAD);
- if (skb == NULL)
- return NULL;
-
- /* Correct skb->truesize so that socket buffer
- * accounting is not confused the rounding we must
- * do to satisfy alignment constraints.
- */
- skb->truesize -= 4096;
-
- data = (unsigned long)(skb->data);
- roundup = (-data) & (4095);
- skb_reserve(skb, roundup);
- return skb;
-}
-
-/* Allocate 2x as much space as required and use whichever portion
- * does not cross a 4KB boundary */
-static inline struct sk_buff *myri10ge_alloc_small_safe(struct net_device *dev,
- unsigned int bytes)
-{
- struct sk_buff *skb;
- unsigned long data, boundary;
-
- skb = netdev_alloc_skb(dev, 2 * (bytes + MXGEFW_PAD) - 1);
- if (unlikely(skb == NULL))
- return NULL;
-
- /* Correct skb->truesize so that socket buffer
- * accounting is not confused the rounding we must
- * do to satisfy alignment constraints.
- */
- skb->truesize -= bytes + MXGEFW_PAD;
-
- data = (unsigned long)(skb->data);
- boundary = (data + 4095UL) & ~4095UL;
- if ((boundary - data) >= (bytes + MXGEFW_PAD))
- return skb;
-
- skb_reserve(skb, boundary - data);
- return skb;
-}
-
-/* Allocate just enough space, and verify that the allocated
- * space does not cross a 4KB boundary */
-static inline struct sk_buff *myri10ge_alloc_small(struct net_device *dev,
- int bytes)
-{
- struct sk_buff *skb;
- unsigned long roundup, data, end;
-
- skb = netdev_alloc_skb(dev, bytes + 16 + MXGEFW_PAD);
- if (unlikely(skb == NULL))
- return NULL;
-
- /* Round allocated buffer to 16 byte boundary */
- data = (unsigned long)(skb->data);
- roundup = (-data) & 15UL;
- skb_reserve(skb, roundup);
- /* Verify that the data buffer does not cross a page boundary */
- data = (unsigned long)(skb->data);
- end = data + bytes + MXGEFW_PAD - 1;
- if (unlikely(((end >> 12) != (data >> 12)) && (data & 4095UL))) {
- printk(KERN_NOTICE
- "myri10ge_alloc_small: small skb crossed 4KB boundary\n");
- myri10ge_skb_cross_4k = 1;
- dev_kfree_skb_any(skb);
- skb = myri10ge_alloc_small_safe(dev, bytes);
- }
- return skb;
-}
-
-static inline int
-myri10ge_getbuf(struct myri10ge_rx_buf *rx, struct myri10ge_priv *mgp,
- int bytes, int idx)
-{
- struct net_device *dev = mgp->dev;
- struct pci_dev *pdev = mgp->pdev;
- struct sk_buff *skb;
- dma_addr_t bus;
- int len, retval = 0;
-
- bytes += VLAN_HLEN; /* account for 802.1q vlan tag */
-
- if ((bytes + MXGEFW_PAD) > (4096 - 16) /* linux overhead */ )
- skb = myri10ge_alloc_big(dev, bytes);
- else if (myri10ge_skb_cross_4k)
- skb = myri10ge_alloc_small_safe(dev, bytes);
- else
- skb = myri10ge_alloc_small(dev, bytes);
-
- if (unlikely(skb == NULL)) {
- rx->alloc_fail++;
- retval = -ENOBUFS;
- goto done;
- }
-
- /* set len so that it only covers the area we
- * need mapped for DMA */
- len = bytes + MXGEFW_PAD;
-
- bus = pci_map_single(pdev, skb->data, len, PCI_DMA_FROMDEVICE);
- rx->info[idx].skb = skb;
- pci_unmap_addr_set(&rx->info[idx], bus, bus);
- pci_unmap_len_set(&rx->info[idx], len, len);
- rx->shadow[idx].addr_low = htonl(MYRI10GE_LOWPART_TO_U32(bus));
- rx->shadow[idx].addr_high = htonl(MYRI10GE_HIGHPART_TO_U32(bus));
-
-done:
- /* copy 8 descriptors (64-bytes) to the mcp at a time */
- if ((idx & 7) == 7) {
- if (rx->wc_fifo == NULL)
- myri10ge_submit_8rx(&rx->lanai[idx - 7],
- &rx->shadow[idx - 7]);
- else {
- mb();
- myri10ge_pio_copy(rx->wc_fifo,
- &rx->shadow[idx - 7], 64);
- }
- }
- return retval;
-}
-
static inline void myri10ge_vlan_ip_csum(struct sk_buff *skb, __wsum hw_csum)
{
struct vlan_hdr *vh = (struct vlan_hdr *)(skb->data);
* skb_pull() (for ether_pad and eth_type_trans()) requires
* the beginning of the packet in skb_headlen(), move it
* manually */
- memcpy(skb->data, va, hlen);
+ skb_copy_to_linear_data(skb, va, hlen);
skb_shinfo(skb)->frags[0].page_offset += hlen;
skb_shinfo(skb)->frags[0].size -= hlen;
skb->data_len -= hlen;
/* try to refill entire ring */
while (rx->fill_cnt != (rx->cnt + rx->mask + 1)) {
idx = rx->fill_cnt & rx->mask;
-
- if ((bytes < MYRI10GE_ALLOC_SIZE / 2) &&
- (rx->page_offset + bytes <= MYRI10GE_ALLOC_SIZE)) {
+ if (rx->page_offset + bytes <= MYRI10GE_ALLOC_SIZE) {
/* we can use part of previous page */
get_page(rx->page);
} else {
/* start next packet on a cacheline boundary */
rx->page_offset += SKB_DATA_ALIGN(bytes);
+
+#if MYRI10GE_ALLOC_SIZE > 4096
+ /* don't cross a 4KB boundary */
+ if ((rx->page_offset >> 12) !=
+ ((rx->page_offset + bytes - 1) >> 12))
+ rx->page_offset = (rx->page_offset + 4096) & ~4095;
+#endif
rx->fill_cnt++;
/* copy 8 descriptors to the firmware at a time */
* page into an skb */
static inline int
-myri10ge_page_rx_done(struct myri10ge_priv *mgp, struct myri10ge_rx_buf *rx,
- int bytes, int len, __wsum csum)
+myri10ge_rx_done(struct myri10ge_priv *mgp, struct myri10ge_rx_buf *rx,
+ int bytes, int len, __wsum csum)
{
struct sk_buff *skb;
struct skb_frag_struct rx_frags[MYRI10GE_MAX_FRAGS_PER_FRAME];
skb_shinfo(skb)->nr_frags = 0;
}
skb->protocol = eth_type_trans(skb, dev);
- skb->dev = dev;
if (mgp->csum_flag) {
if ((skb->protocol == htons(ETH_P_IP)) ||
return 1;
}
-static inline unsigned long
-myri10ge_rx_done(struct myri10ge_priv *mgp, struct myri10ge_rx_buf *rx,
- int bytes, int len, __wsum csum)
-{
- dma_addr_t bus;
- struct sk_buff *skb;
- int idx, unmap_len;
-
- idx = rx->cnt & rx->mask;
- rx->cnt++;
-
- /* save a pointer to the received skb */
- skb = rx->info[idx].skb;
- bus = pci_unmap_addr(&rx->info[idx], bus);
- unmap_len = pci_unmap_len(&rx->info[idx], len);
-
- /* try to replace the received skb */
- if (myri10ge_getbuf(rx, mgp, bytes, idx)) {
- /* drop the frame -- the old skbuf is re-cycled */
- mgp->stats.rx_dropped += 1;
- return 0;
- }
-
- /* unmap the recvd skb */
- pci_unmap_single(mgp->pdev, bus, unmap_len, PCI_DMA_FROMDEVICE);
-
- /* mcp implicitly skips 1st bytes so that packet is properly
- * aligned */
- skb_reserve(skb, MXGEFW_PAD);
-
- /* set the length of the frame */
- skb_put(skb, len);
-
- skb->protocol = eth_type_trans(skb, mgp->dev);
- if (mgp->csum_flag) {
- if ((skb->protocol == htons(ETH_P_IP)) ||
- (skb->protocol == htons(ETH_P_IPV6))) {
- skb->csum = csum;
- skb->ip_summed = CHECKSUM_COMPLETE;
- } else
- myri10ge_vlan_ip_csum(skb, csum);
- }
-
- netif_receive_skb(skb);
- mgp->dev->last_rx = jiffies;
- return 1;
-}
-
static inline void myri10ge_tx_done(struct myri10ge_priv *mgp, int mcp_index)
{
struct pci_dev *pdev = mgp->pdev;
length, checksum);
else
rx_ok = myri10ge_rx_done(mgp, &mgp->rx_big,
- mgp->dev->mtu + ETH_HLEN,
+ mgp->big_bytes,
length, checksum);
rx_packets += rx_ok;
rx_bytes += rx_ok * (unsigned long)length;
rx_done->cnt = cnt;
mgp->stats.rx_packets += rx_packets;
mgp->stats.rx_bytes += rx_bytes;
+
+ /* restock receive rings if needed */
+ if (mgp->rx_small.fill_cnt - mgp->rx_small.cnt < myri10ge_fill_thresh)
+ myri10ge_alloc_rx_pages(mgp, &mgp->rx_small,
+ mgp->small_bytes + MXGEFW_PAD, 0);
+ if (mgp->rx_big.fill_cnt - mgp->rx_big.cnt < myri10ge_fill_thresh)
+ myri10ge_alloc_rx_pages(mgp, &mgp->rx_big, mgp->big_bytes, 0);
+
}
static inline void myri10ge_check_statblock(struct myri10ge_priv *mgp)
struct mcp_irq_data *stats = mgp->fw_stats;
if (unlikely(stats->stats_updated)) {
- if (mgp->link_state != stats->link_up) {
- mgp->link_state = stats->link_up;
- if (mgp->link_state) {
+ unsigned link_up = ntohl(stats->link_up);
+ if (mgp->link_state != link_up) {
+ mgp->link_state = link_up;
+
+ if (mgp->link_state == MXGEFW_LINK_UP) {
if (netif_msg_link(mgp))
printk(KERN_INFO
"myri10ge: %s: link up\n",
} else {
if (netif_msg_link(mgp))
printk(KERN_INFO
- "myri10ge: %s: link down\n",
- mgp->dev->name);
+ "myri10ge: %s: link %s\n",
+ mgp->dev->name,
+ (link_up == MXGEFW_LINK_MYRINET ?
+ "mismatch (Myrinet detected)" :
+ "down"));
netif_carrier_off(mgp->dev);
mgp->link_changes++;
}
"tx_req", "tx_done", "rx_small_cnt", "rx_big_cnt",
"wake_queue", "stop_queue", "watchdog_resets", "tx_linearized",
"link_changes", "link_up", "dropped_link_overflow",
- "dropped_link_error_or_filtered", "dropped_multicast_filtered",
+ "dropped_link_error_or_filtered",
+ "dropped_pause", "dropped_bad_phy", "dropped_bad_crc32",
+ "dropped_unicast_filtered", "dropped_multicast_filtered",
"dropped_runt", "dropped_overrun", "dropped_no_small_buffer",
"dropped_no_big_buffer"
};
data[i] = ((unsigned long *)&mgp->stats)[i];
data[i++] = (unsigned int)mgp->tx.boundary;
- data[i++] = (unsigned int)(mgp->mtrr >= 0);
+ data[i++] = (unsigned int)mgp->wc_enabled;
data[i++] = (unsigned int)mgp->pdev->irq;
data[i++] = (unsigned int)mgp->msi_enabled;
data[i++] = (unsigned int)mgp->read_dma;
data[i++] = (unsigned int)ntohl(mgp->fw_stats->dropped_link_overflow);
data[i++] =
(unsigned int)ntohl(mgp->fw_stats->dropped_link_error_or_filtered);
+ data[i++] = (unsigned int)ntohl(mgp->fw_stats->dropped_pause);
+ data[i++] = (unsigned int)ntohl(mgp->fw_stats->dropped_bad_phy);
+ data[i++] = (unsigned int)ntohl(mgp->fw_stats->dropped_bad_crc32);
+ data[i++] =
+ (unsigned int)ntohl(mgp->fw_stats->dropped_unicast_filtered);
data[i++] =
(unsigned int)ntohl(mgp->fw_stats->dropped_multicast_filtered);
data[i++] = (unsigned int)ntohl(mgp->fw_stats->dropped_runt);
.set_tx_csum = ethtool_op_set_tx_hw_csum,
.get_sg = ethtool_op_get_sg,
.set_sg = ethtool_op_set_sg,
-#ifdef NETIF_F_TSO
.get_tso = ethtool_op_get_tso,
.set_tso = ethtool_op_set_tso,
-#endif
+ .get_link = ethtool_op_get_link,
.get_strings = myri10ge_get_strings,
.get_stats_count = myri10ge_get_stats_count,
.get_ethtool_stats = myri10ge_get_ethtool_stats,
status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_SEND_RING_SIZE, &cmd, 0);
tx_ring_size = cmd.data0;
status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_RX_RING_SIZE, &cmd, 0);
+ if (status != 0)
+ return status;
rx_ring_size = cmd.data0;
tx_ring_entries = tx_ring_size / sizeof(struct mcp_kreq_ether_send);
mgp->tx.mask = tx_ring_entries - 1;
mgp->rx_small.mask = mgp->rx_big.mask = rx_ring_entries - 1;
+ status = -ENOMEM;
+
/* allocate the host shadow rings */
bytes = 8 + (MYRI10GE_MAX_SEND_DESC_TSO + 4)
goto abort_with_rx_small_info;
/* Fill the receive rings */
+ mgp->rx_big.cnt = 0;
+ mgp->rx_small.cnt = 0;
+ mgp->rx_big.fill_cnt = 0;
+ mgp->rx_small.fill_cnt = 0;
+ mgp->rx_small.page_offset = MYRI10GE_ALLOC_SIZE;
+ mgp->rx_big.page_offset = MYRI10GE_ALLOC_SIZE;
+ mgp->rx_small.watchdog_needed = 0;
+ mgp->rx_big.watchdog_needed = 0;
+ myri10ge_alloc_rx_pages(mgp, &mgp->rx_small,
+ mgp->small_bytes + MXGEFW_PAD, 0);
- for (i = 0; i <= mgp->rx_small.mask; i++) {
- status = myri10ge_getbuf(&mgp->rx_small, mgp,
- mgp->small_bytes, i);
- if (status) {
- printk(KERN_ERR
- "myri10ge: %s: alloced only %d small bufs\n",
- dev->name, i);
- goto abort_with_rx_small_ring;
- }
+ if (mgp->rx_small.fill_cnt < mgp->rx_small.mask + 1) {
+ printk(KERN_ERR "myri10ge: %s: alloced only %d small bufs\n",
+ dev->name, mgp->rx_small.fill_cnt);
+ goto abort_with_rx_small_ring;
}
- for (i = 0; i <= mgp->rx_big.mask; i++) {
- status =
- myri10ge_getbuf(&mgp->rx_big, mgp, dev->mtu + ETH_HLEN, i);
- if (status) {
- printk(KERN_ERR
- "myri10ge: %s: alloced only %d big bufs\n",
- dev->name, i);
- goto abort_with_rx_big_ring;
- }
+ myri10ge_alloc_rx_pages(mgp, &mgp->rx_big, mgp->big_bytes, 0);
+ if (mgp->rx_big.fill_cnt < mgp->rx_big.mask + 1) {
+ printk(KERN_ERR "myri10ge: %s: alloced only %d big bufs\n",
+ dev->name, mgp->rx_big.fill_cnt);
+ goto abort_with_rx_big_ring;
}
return 0;
abort_with_rx_big_ring:
- for (i = 0; i <= mgp->rx_big.mask; i++) {
- if (mgp->rx_big.info[i].skb != NULL)
- dev_kfree_skb_any(mgp->rx_big.info[i].skb);
- if (pci_unmap_len(&mgp->rx_big.info[i], len))
- pci_unmap_single(mgp->pdev,
- pci_unmap_addr(&mgp->rx_big.info[i],
- bus),
- pci_unmap_len(&mgp->rx_big.info[i],
- len),
- PCI_DMA_FROMDEVICE);
+ for (i = mgp->rx_big.cnt; i < mgp->rx_big.fill_cnt; i++) {
+ int idx = i & mgp->rx_big.mask;
+ myri10ge_unmap_rx_page(mgp->pdev, &mgp->rx_big.info[idx],
+ mgp->big_bytes);
+ put_page(mgp->rx_big.info[idx].page);
}
abort_with_rx_small_ring:
- for (i = 0; i <= mgp->rx_small.mask; i++) {
- if (mgp->rx_small.info[i].skb != NULL)
- dev_kfree_skb_any(mgp->rx_small.info[i].skb);
- if (pci_unmap_len(&mgp->rx_small.info[i], len))
- pci_unmap_single(mgp->pdev,
- pci_unmap_addr(&mgp->rx_small.info[i],
- bus),
- pci_unmap_len(&mgp->rx_small.info[i],
- len),
- PCI_DMA_FROMDEVICE);
+ for (i = mgp->rx_small.cnt; i < mgp->rx_small.fill_cnt; i++) {
+ int idx = i & mgp->rx_small.mask;
+ myri10ge_unmap_rx_page(mgp->pdev, &mgp->rx_small.info[idx],
+ mgp->small_bytes + MXGEFW_PAD);
+ put_page(mgp->rx_small.info[idx].page);
}
+
kfree(mgp->rx_big.info);
abort_with_rx_small_info:
mgp = netdev_priv(dev);
- for (i = 0; i <= mgp->rx_big.mask; i++) {
- if (mgp->rx_big.info[i].skb != NULL)
- dev_kfree_skb_any(mgp->rx_big.info[i].skb);
- if (pci_unmap_len(&mgp->rx_big.info[i], len))
- pci_unmap_single(mgp->pdev,
- pci_unmap_addr(&mgp->rx_big.info[i],
- bus),
- pci_unmap_len(&mgp->rx_big.info[i],
- len),
- PCI_DMA_FROMDEVICE);
- }
-
- for (i = 0; i <= mgp->rx_small.mask; i++) {
- if (mgp->rx_small.info[i].skb != NULL)
- dev_kfree_skb_any(mgp->rx_small.info[i].skb);
- if (pci_unmap_len(&mgp->rx_small.info[i], len))
- pci_unmap_single(mgp->pdev,
- pci_unmap_addr(&mgp->rx_small.info[i],
- bus),
- pci_unmap_len(&mgp->rx_small.info[i],
- len),
- PCI_DMA_FROMDEVICE);
+ for (i = mgp->rx_big.cnt; i < mgp->rx_big.fill_cnt; i++) {
+ idx = i & mgp->rx_big.mask;
+ if (i == mgp->rx_big.fill_cnt - 1)
+ mgp->rx_big.info[idx].page_offset = MYRI10GE_ALLOC_SIZE;
+ myri10ge_unmap_rx_page(mgp->pdev, &mgp->rx_big.info[idx],
+ mgp->big_bytes);
+ put_page(mgp->rx_big.info[idx].page);
}
+ for (i = mgp->rx_small.cnt; i < mgp->rx_small.fill_cnt; i++) {
+ idx = i & mgp->rx_small.mask;
+ if (i == mgp->rx_small.fill_cnt - 1)
+ mgp->rx_small.info[idx].page_offset =
+ MYRI10GE_ALLOC_SIZE;
+ myri10ge_unmap_rx_page(mgp->pdev, &mgp->rx_small.info[idx],
+ mgp->small_bytes + MXGEFW_PAD);
+ put_page(mgp->rx_small.info[idx].page);
+ }
tx = &mgp->tx;
while (tx->done != tx->req) {
idx = tx->done & tx->mask;
mgp->tx.req_list = NULL;
}
+static int myri10ge_request_irq(struct myri10ge_priv *mgp)
+{
+ struct pci_dev *pdev = mgp->pdev;
+ int status;
+
+ if (myri10ge_msi) {
+ status = pci_enable_msi(pdev);
+ if (status != 0)
+ dev_err(&pdev->dev,
+ "Error %d setting up MSI; falling back to xPIC\n",
+ status);
+ else
+ mgp->msi_enabled = 1;
+ } else {
+ mgp->msi_enabled = 0;
+ }
+ status = request_irq(pdev->irq, myri10ge_intr, IRQF_SHARED,
+ mgp->dev->name, mgp);
+ if (status != 0) {
+ dev_err(&pdev->dev, "failed to allocate IRQ\n");
+ if (mgp->msi_enabled)
+ pci_disable_msi(pdev);
+ }
+ return status;
+}
+
+static void myri10ge_free_irq(struct myri10ge_priv *mgp)
+{
+ struct pci_dev *pdev = mgp->pdev;
+
+ free_irq(pdev->irq, mgp);
+ if (mgp->msi_enabled)
+ pci_disable_msi(pdev);
+}
+
static int myri10ge_open(struct net_device *dev)
{
struct myri10ge_priv *mgp;
status = myri10ge_reset(mgp);
if (status != 0) {
printk(KERN_ERR "myri10ge: %s: failed reset\n", dev->name);
- mgp->running = MYRI10GE_ETH_STOPPED;
- return -ENXIO;
+ goto abort_with_nothing;
}
+ status = myri10ge_request_irq(mgp);
+ if (status != 0)
+ goto abort_with_nothing;
+
/* decide what small buffer size to use. For good TCP rx
* performance, it is important to not receive 1514 byte
* frames into jumbo buffers, as it confuses the socket buffer
*/
if (dev->mtu <= ETH_DATA_LEN)
- mgp->small_bytes = 128; /* enough for a TCP header */
+ /* enough for a TCP header */
+ mgp->small_bytes = (128 > SMP_CACHE_BYTES)
+ ? (128 - MXGEFW_PAD)
+ : (SMP_CACHE_BYTES - MXGEFW_PAD);
else
- mgp->small_bytes = ETH_FRAME_LEN; /* enough for an ETH_DATA_LEN frame */
+ /* enough for a vlan encapsulated ETH_DATA_LEN frame */
+ mgp->small_bytes = VLAN_ETH_FRAME_LEN;
/* Override the small buffer size? */
if (myri10ge_small_bytes > 0)
mgp->small_bytes = myri10ge_small_bytes;
- /* If the user sets an obscenely small MTU, adjust the small
- * bytes down to nearly nothing */
- if (mgp->small_bytes >= (dev->mtu + ETH_HLEN))
- mgp->small_bytes = 64;
-
/* get the lanai pointers to the send and receive rings */
status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_SEND_OFFSET, &cmd, 0);
"myri10ge: %s: failed to get ring sizes or locations\n",
dev->name);
mgp->running = MYRI10GE_ETH_STOPPED;
- return -ENXIO;
+ goto abort_with_irq;
}
- if (mgp->mtrr >= 0) {
+ if (myri10ge_wcfifo && mgp->wc_enabled) {
mgp->tx.wc_fifo = (u8 __iomem *) mgp->sram + MXGEFW_ETH_SEND_4;
mgp->rx_small.wc_fifo =
(u8 __iomem *) mgp->sram + MXGEFW_ETH_RECV_SMALL;
mgp->rx_big.wc_fifo = NULL;
}
- status = myri10ge_allocate_rings(dev);
- if (status != 0)
- goto abort_with_nothing;
-
/* Firmware needs the big buff size as a power of 2. Lie and
* tell him the buffer is larger, because we only use 1
* buffer/pkt, and the mtu will prevent overruns.
*/
- big_pow2 = dev->mtu + ETH_HLEN + MXGEFW_PAD;
- while ((big_pow2 & (big_pow2 - 1)) != 0)
- big_pow2++;
+ big_pow2 = dev->mtu + ETH_HLEN + VLAN_HLEN + MXGEFW_PAD;
+ if (big_pow2 < MYRI10GE_ALLOC_SIZE / 2) {
+ while ((big_pow2 & (big_pow2 - 1)) != 0)
+ big_pow2++;
+ mgp->big_bytes = dev->mtu + ETH_HLEN + VLAN_HLEN + MXGEFW_PAD;
+ } else {
+ big_pow2 = MYRI10GE_ALLOC_SIZE;
+ mgp->big_bytes = big_pow2;
+ }
+
+ status = myri10ge_allocate_rings(dev);
+ if (status != 0)
+ goto abort_with_irq;
/* now give firmware buffers sizes, and MTU */
cmd.data0 = dev->mtu + ETH_HLEN + VLAN_HLEN;
abort_with_rings:
myri10ge_free_rings(dev);
+abort_with_irq:
+ myri10ge_free_irq(mgp);
+
abort_with_nothing:
mgp->running = MYRI10GE_ETH_STOPPED;
return -ENOMEM;
printk(KERN_ERR "myri10ge: %s never got down irq\n", dev->name);
netif_tx_disable(dev);
-
+ myri10ge_free_irq(mgp);
myri10ge_free_rings(dev);
mgp->running = MYRI10GE_ETH_STOPPED;
mss = 0;
max_segments = MXGEFW_MAX_SEND_DESC;
-#ifdef NETIF_F_TSO
- if (skb->len > (dev->mtu + ETH_HLEN)) {
+ if (skb_is_gso(skb)) {
mss = skb_shinfo(skb)->gso_size;
- if (mss != 0)
- max_segments = MYRI10GE_MAX_SEND_DESC_TSO;
+ max_segments = MYRI10GE_MAX_SEND_DESC_TSO;
}
-#endif /*NETIF_F_TSO */
if ((unlikely(avail < max_segments))) {
/* we are out of transmit resources */
odd_flag = 0;
flags = (MXGEFW_FLAGS_NO_TSO | MXGEFW_FLAGS_FIRST);
if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
- cksum_offset = (skb->h.raw - skb->data);
+ cksum_offset = skb_transport_offset(skb);
pseudo_hdr_offset = cksum_offset + skb->csum_offset;
/* If the headers are excessively large, then we must
* fall back to a software checksum */
cum_len = 0;
-#ifdef NETIF_F_TSO
if (mss) { /* TSO */
/* this removes any CKSUM flag from before */
flags = (MXGEFW_FLAGS_TSO_HDR | MXGEFW_FLAGS_FIRST);
* send loop that we are still in the
* header portion of the TSO packet.
* TSO header must be at most 134 bytes long */
- cum_len = -((skb->h.raw - skb->data) + (skb->h.th->doff << 2));
+ cum_len = -(skb_transport_offset(skb) + tcp_hdrlen(skb));
/* for TSO, pseudo_hdr_offset holds mss.
* The firmware figures out where to put
* the checksum by parsing the header. */
pseudo_hdr_offset = mss;
} else
-#endif /*NETIF_F_TSO */
/* Mark small packets, and pad out tiny packets */
if (skb->len <= MXGEFW_SEND_SMALL_SIZE) {
flags |= MXGEFW_FLAGS_SMALL;
seglen = len;
flags_next = flags & ~MXGEFW_FLAGS_FIRST;
cum_len_next = cum_len + seglen;
-#ifdef NETIF_F_TSO
if (mss) { /* TSO */
(req - rdma_count)->rdma_count = rdma_count + 1;
(small * MXGEFW_FLAGS_SMALL);
}
}
-#endif /* NETIF_F_TSO */
req->addr_high = high_swapped;
req->addr_low = htonl(low);
req->pseudo_hdr_offset = htons(pseudo_hdr_offset);
}
(req - rdma_count)->rdma_count = rdma_count;
-#ifdef NETIF_F_TSO
if (mss)
do {
req--;
req->flags |= MXGEFW_FLAGS_TSO_LAST;
} while (!(req->flags & (MXGEFW_FLAGS_TSO_CHOP |
MXGEFW_FLAGS_FIRST)));
-#endif
idx = ((count - 1) + tx->req) & tx->mask;
tx->info[idx].last = 1;
if (tx->wc_fifo == NULL)
goto abort;
}
- if (dev->flags & IFF_ALLMULTI) {
+ if ((dev->flags & IFF_ALLMULTI) || mgp->adopted_rx_filter_bug) {
/* request to disable multicast filtering, so quit here */
return;
}
err_cap |= PCI_ERR_CAP_ECRC_GENE;
pci_write_config_dword(bridge, cap + PCI_ERR_CAP, err_cap);
dev_info(dev, "Enabled ECRC on upstream bridge %s\n", pci_name(bridge));
- mgp->tx.boundary = 4096;
- mgp->fw_name = myri10ge_fw_aligned;
}
/*
* firmware image, and set tx.boundary to 4KB.
*/
-#define PCI_DEVICE_ID_INTEL_E5000_PCIE23 0x25f7
-#define PCI_DEVICE_ID_INTEL_E5000_PCIE47 0x25fa
-
-static void myri10ge_select_firmware(struct myri10ge_priv *mgp)
+static void myri10ge_firmware_probe(struct myri10ge_priv *mgp)
{
- struct pci_dev *bridge = mgp->pdev->bus->self;
+ struct pci_dev *pdev = mgp->pdev;
+ struct device *dev = &pdev->dev;
+ int cap, status;
+ u16 val;
+ mgp->tx.boundary = 4096;
+ /*
+ * Verify the max read request size was set to 4KB
+ * before trying the test with 4KB.
+ */
+ cap = pci_find_capability(pdev, PCI_CAP_ID_EXP);
+ if (cap < 64) {
+ dev_err(dev, "Bad PCI_CAP_ID_EXP location %d\n", cap);
+ goto abort;
+ }
+ status = pci_read_config_word(pdev, cap + PCI_EXP_DEVCTL, &val);
+ if (status != 0) {
+ dev_err(dev, "Couldn't read max read req size: %d\n", status);
+ goto abort;
+ }
+ if ((val & (5 << 12)) != (5 << 12)) {
+ dev_warn(dev, "Max Read Request size != 4096 (0x%x)\n", val);
+ mgp->tx.boundary = 2048;
+ }
+ /*
+ * load the optimized firmware (which assumes aligned PCIe
+ * completions) in order to see if it works on this host.
+ */
+ mgp->fw_name = myri10ge_fw_aligned;
+ status = myri10ge_load_firmware(mgp);
+ if (status != 0) {
+ goto abort;
+ }
+
+ /*
+ * Enable ECRC if possible
+ */
+ myri10ge_enable_ecrc(mgp);
+
+ /*
+ * Run a DMA test which watches for unaligned completions and
+ * aborts on the first one seen.
+ */
+
+ status = myri10ge_dma_test(mgp, MXGEFW_CMD_UNALIGNED_TEST);
+ if (status == 0)
+ return; /* keep the aligned firmware */
+
+ if (status != -E2BIG)
+ dev_warn(dev, "DMA test failed: %d\n", status);
+ if (status == -ENOSYS)
+ dev_warn(dev, "Falling back to ethp! "
+ "Please install up to date fw\n");
+abort:
+ /* fall back to using the unaligned firmware */
mgp->tx.boundary = 2048;
mgp->fw_name = myri10ge_fw_unaligned;
+}
+
+static void myri10ge_select_firmware(struct myri10ge_priv *mgp)
+{
if (myri10ge_force_firmware == 0) {
int link_width, exp_cap;
u16 lnk;
pci_read_config_word(mgp->pdev, exp_cap + PCI_EXP_LNKSTA, &lnk);
link_width = (lnk >> 4) & 0x3f;
- myri10ge_enable_ecrc(mgp);
-
/* Check to see if Link is less than 8 or if the
* upstream bridge is known to provide aligned
* completions */
link_width);
mgp->tx.boundary = 4096;
mgp->fw_name = myri10ge_fw_aligned;
- } else if (bridge &&
- /* ServerWorks HT2000/HT1000 */
- ((bridge->vendor == PCI_VENDOR_ID_SERVERWORKS
- && bridge->device ==
- PCI_DEVICE_ID_SERVERWORKS_HT2000_PCIE)
- /* All Intel E5000 PCIE ports */
- || (bridge->vendor == PCI_VENDOR_ID_INTEL
- && bridge->device >=
- PCI_DEVICE_ID_INTEL_E5000_PCIE23
- && bridge->device <=
- PCI_DEVICE_ID_INTEL_E5000_PCIE47))) {
- dev_info(&mgp->pdev->dev,
- "Assuming aligned completions (0x%x:0x%x)\n",
- bridge->vendor, bridge->device);
- mgp->tx.boundary = 4096;
- mgp->fw_name = myri10ge_fw_aligned;
+ } else {
+ myri10ge_firmware_probe(mgp);
}
} else {
if (myri10ge_force_firmware == 1) {
}
}
-static void myri10ge_save_state(struct myri10ge_priv *mgp)
-{
- struct pci_dev *pdev = mgp->pdev;
- int cap;
-
- pci_save_state(pdev);
- /* now save PCIe and MSI state that Linux will not
- * save for us */
- cap = pci_find_capability(pdev, PCI_CAP_ID_EXP);
- pci_read_config_dword(pdev, cap + PCI_EXP_DEVCTL, &mgp->devctl);
- cap = pci_find_capability(pdev, PCI_CAP_ID_MSI);
- pci_read_config_word(pdev, cap + PCI_MSI_FLAGS, &mgp->msi_flags);
-}
-
-static void myri10ge_restore_state(struct myri10ge_priv *mgp)
-{
- struct pci_dev *pdev = mgp->pdev;
- int cap;
-
- /* restore PCIe and MSI state that linux will not */
- cap = pci_find_capability(pdev, PCI_CAP_ID_EXP);
- pci_write_config_dword(pdev, cap + PCI_CAP_ID_EXP, mgp->devctl);
- cap = pci_find_capability(pdev, PCI_CAP_ID_MSI);
- pci_write_config_word(pdev, cap + PCI_MSI_FLAGS, mgp->msi_flags);
-
- pci_restore_state(pdev);
-}
-
#ifdef CONFIG_PM
static int myri10ge_suspend(struct pci_dev *pdev, pm_message_t state)
rtnl_unlock();
}
myri10ge_dummy_rdma(mgp, 0);
- free_irq(pdev->irq, mgp);
- myri10ge_save_state(mgp);
+ pci_save_state(pdev);
pci_disable_device(pdev);
- pci_set_power_state(pdev, pci_choose_state(pdev, state));
- return 0;
+
+ return pci_set_power_state(pdev, pci_choose_state(pdev, state));
}
static int myri10ge_resume(struct pci_dev *pdev)
mgp->dev->name);
return -EIO;
}
- myri10ge_restore_state(mgp);
+
+ status = pci_restore_state(pdev);
+ if (status)
+ return status;
status = pci_enable_device(pdev);
- if (status < 0) {
+ if (status) {
dev_err(&pdev->dev, "failed to enable device\n");
- return -EIO;
+ return status;
}
pci_set_master(pdev);
- status = request_irq(pdev->irq, myri10ge_intr, IRQF_SHARED,
- netdev->name, mgp);
- if (status != 0) {
- dev_err(&pdev->dev, "failed to allocate IRQ\n");
- goto abort_with_enabled;
- }
-
myri10ge_reset(mgp);
myri10ge_dummy_rdma(mgp, 1);
/* Save configuration space to be restored if the
* nic resets due to a parity error */
- myri10ge_save_state(mgp);
+ pci_save_state(pdev);
if (netif_running(netdev)) {
rtnl_lock();
- myri10ge_open(netdev);
+ status = myri10ge_open(netdev);
rtnl_unlock();
+ if (status != 0)
+ goto abort_with_enabled;
+
}
netif_device_attach(netdev);
* For now, just report it */
reboot = myri10ge_read_reboot(mgp);
printk(KERN_ERR
- "myri10ge: %s: NIC rebooted (0x%x), resetting\n",
- mgp->dev->name, reboot);
+ "myri10ge: %s: NIC rebooted (0x%x),%s resetting\n",
+ mgp->dev->name, reboot,
+ myri10ge_reset_recover ? " " : " not");
+ if (myri10ge_reset_recover == 0)
+ return;
+
+ myri10ge_reset_recover--;
+
/*
* A rebooted nic will come back with config space as
* it was after power was applied to PCIe bus.
* when the driver was loaded, or the last time the
* nic was resumed from power saving mode.
*/
- myri10ge_restore_state(mgp);
+ pci_restore_state(mgp->pdev);
+
+ /* save state again for accounting reasons */
+ pci_save_state(mgp->pdev);
+
} else {
/* if we get back -1's from our slot, perhaps somebody
* powered off our card. Don't try to reset it in
struct myri10ge_priv *mgp;
mgp = (struct myri10ge_priv *)arg;
+
+ if (mgp->rx_small.watchdog_needed) {
+ myri10ge_alloc_rx_pages(mgp, &mgp->rx_small,
+ mgp->small_bytes + MXGEFW_PAD, 1);
+ if (mgp->rx_small.fill_cnt - mgp->rx_small.cnt >=
+ myri10ge_fill_thresh)
+ mgp->rx_small.watchdog_needed = 0;
+ }
+ if (mgp->rx_big.watchdog_needed) {
+ myri10ge_alloc_rx_pages(mgp, &mgp->rx_big, mgp->big_bytes, 1);
+ if (mgp->rx_big.fill_cnt - mgp->rx_big.cnt >=
+ myri10ge_fill_thresh)
+ mgp->rx_big.watchdog_needed = 0;
+ }
+
if (mgp->tx.req != mgp->tx.done &&
mgp->tx.done == mgp->watchdog_tx_done &&
mgp->watchdog_tx_req != mgp->watchdog_tx_done)
status = -ENODEV;
goto abort_with_netdev;
}
- myri10ge_select_firmware(mgp);
/* Find the vendor-specific cap so we can check
* the reboot register later on */
mgp->board_span = pci_resource_len(pdev, 0);
mgp->iomem_base = pci_resource_start(pdev, 0);
mgp->mtrr = -1;
+ mgp->wc_enabled = 0;
#ifdef CONFIG_MTRR
mgp->mtrr = mtrr_add(mgp->iomem_base, mgp->board_span,
MTRR_TYPE_WRCOMB, 1);
+ if (mgp->mtrr >= 0)
+ mgp->wc_enabled = 1;
#endif
/* Hack. need to get rid of these magic numbers */
mgp->sram_size =
goto abort_with_ioremap;
memset(mgp->rx_done.entry, 0, bytes);
+ myri10ge_select_firmware(mgp);
+
status = myri10ge_load_firmware(mgp);
if (status != 0) {
dev_err(&pdev->dev, "failed to load firmware\n");
goto abort_with_firmware;
}
- if (myri10ge_msi) {
- status = pci_enable_msi(pdev);
- if (status != 0)
- dev_err(&pdev->dev,
- "Error %d setting up MSI; falling back to xPIC\n",
- status);
- else
- mgp->msi_enabled = 1;
- }
-
- status = request_irq(pdev->irq, myri10ge_intr, IRQF_SHARED,
- netdev->name, mgp);
- if (status != 0) {
- dev_err(&pdev->dev, "failed to allocate IRQ\n");
- goto abort_with_firmware;
- }
-
pci_set_drvdata(pdev, mgp);
if ((myri10ge_initial_mtu + ETH_HLEN) > MYRI10GE_MAX_ETHER_MTU)
myri10ge_initial_mtu = MYRI10GE_MAX_ETHER_MTU - ETH_HLEN;
netdev->hard_start_xmit = myri10ge_xmit;
netdev->get_stats = myri10ge_get_stats;
netdev->base_addr = mgp->iomem_base;
- netdev->irq = pdev->irq;
netdev->change_mtu = myri10ge_change_mtu;
netdev->set_multicast_list = myri10ge_set_multicast_list;
netdev->set_mac_address = myri10ge_set_mac_address;
netdev->poll = myri10ge_poll;
netdev->weight = myri10ge_napi_weight;
+ /* make sure we can get an irq, and that MSI can be
+ * setup (if available). Also ensure netdev->irq
+ * is set to correct value if MSI is enabled */
+ status = myri10ge_request_irq(mgp);
+ if (status != 0)
+ goto abort_with_firmware;
+ netdev->irq = pdev->irq;
+ myri10ge_free_irq(mgp);
+
/* Save configuration space to be restored if the
* nic resets due to a parity error */
- myri10ge_save_state(mgp);
+ pci_save_state(pdev);
/* Setup the watchdog timer */
setup_timer(&mgp->watchdog_timer, myri10ge_watchdog_timer,
status = register_netdev(netdev);
if (status != 0) {
dev_err(&pdev->dev, "register_netdev failed: %d\n", status);
- goto abort_with_irq;
+ goto abort_with_state;
}
dev_info(dev, "%s IRQ %d, tx bndry %d, fw %s, WC %s\n",
(mgp->msi_enabled ? "MSI" : "xPIC"),
- pdev->irq, mgp->tx.boundary, mgp->fw_name,
- (mgp->mtrr >= 0 ? "Enabled" : "Disabled"));
+ netdev->irq, mgp->tx.boundary, mgp->fw_name,
+ (mgp->wc_enabled ? "Enabled" : "Disabled"));
return 0;
-abort_with_irq:
- free_irq(pdev->irq, mgp);
- if (mgp->msi_enabled)
- pci_disable_msi(pdev);
+abort_with_state:
+ pci_restore_state(pdev);
abort_with_firmware:
myri10ge_dummy_rdma(mgp, 0);
flush_scheduled_work();
netdev = mgp->dev;
unregister_netdev(netdev);
- free_irq(pdev->irq, mgp);
- if (mgp->msi_enabled)
- pci_disable_msi(pdev);
myri10ge_dummy_rdma(mgp, 0);
+ /* avoid a memory leak */
+ pci_restore_state(pdev);
+
bytes = myri10ge_max_intr_slots * sizeof(*mgp->rx_done.entry);
dma_free_coherent(&pdev->dev, bytes,
mgp->rx_done.entry, mgp->rx_done.bus);
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff