From: Flavio Leitner <fleitner@redhat.com> Date: Thu, 28 Aug 2008 19:42:50 -0300 Subject: [net] update myri10ge 10Gbs ethernet driver Message-id: 48B729EA.1090405@redhat.com O-Subject: [RHEL5.3 PATCH] update myri10ge 10Gbs ethernet driver Bugzilla: 357191 RH-Acked-by: David Miller <davem@redhat.com> RH-Acked-by: Andy Gospodarek <gospo@redhat.com> RH-Acked-by: John W. Linville <linville@redhat.com> resolves BZ 357191 - need updated myri10ge 10Gbs ethernet driver Description: This patch updates myri10ge driver to version 1.3.2-1.269 applying the upstream commits listed below. There is one extra modification adding myri10ge_compat.h needed to backport few defines. Brew build url: http://brewweb.devel.redhat.com/brew/taskinfo?taskID=1444314 Testing: There is a good feedback from one customer that needs multicast support fixed. The other one still needs more features not available in upstream yet. These tests were based on -65.el5 kernel and now we are waiting on feedback based on -106.el5 though it should work. Individual patches can be found at: http://people.redhat.com/fleitner/myri10ge/rhel5/106.el5/ Upstream commits: commit: e700f9f4a208bf6c5d2b4dd67816555930524476 [PATCH] myri10ge: define some previously hardwired firmware constants commit: fd6746daade513bf7f887625c107878b6aacccfd [PATCH] myri10ge: convert to netdev_alloc_skb commit: c58ac5caeb32ef17c2e4fc208f7dc93f6de32b7d [PATCH] myri10ge: use netif_msg_link commit: 85a7ea1b0a3263f3ad423b789a841d03c9acbb65 [PATCH] myri10ge: use multicast support in the firmware commit: ce7f93680aa1d37171c654536ae0ce9745d86a24 [PATCH] myri10ge: improve firmware selection commit: 84fa7933a33f806bbbaae6775e87459b1ec584c0 [NET]: Replace CHECKSUM_HW by CHECKSUM_PARTIAL/CHECKSUM_COMPLETE commit: 1e7bed081968c42469bd02842b4190a115008221 [PATCH] myri10ge: ServerWorks HT2000 PCI id is already defined in pci_ids.h commit: 40f6cff5c47efac2df361fbfc2eb2816729986c8 [PATCH] myri10ge annotations commit: e67bda55e27d3308ba0b4ce8cf2da51850ef1453 [PATCH] myri10ge: write as 2 32-byte blocks in myri10ge_submit_8rx commit: dd50f3361f9f0bb407658e9087947c9bdcdefffc [PATCH] myri10ge: add page-based skb routines commit: c7dab99b080accb2751c96bf66cd5ab12c78f8e4 [PATCH] myri10ge: switch to page-based skb commit: 52ea6fb39b6fd08ec8718b92cddb3fed2165a921 [PATCH] myri10ge: drop contiguous skb routines commit: de3c4507047f2457359551c49b093669acb4f190 [PATCH] myri10ge: Full vlan frame in small_bytes commit: 13348beee529cd1200deeac161e1b2de0705b495 [PATCH] myri10ge: fix big_bytes in case of vlan frames commit: 5796df19824bef34aabf5656f447b3b170d34c3b [PATCH] myri10ge: update driver version to 1.1.0 commit: 7adda30c82f403c948b016a40fc68a1ef427dbca [PATCH] myri10ge: match number of save_state and restore commit: df30a740e4647b035e44a94e6977ad42a94cb850 [PATCH] myri10ge: move request_irq to myri10ge_open commit: 3621cec5b5847348248e749c647219f409550f2f [PATCH] myri10ge: make msi configurable at runtime through sysfs commit: 83f6e1524586119df84f70b7127688c57349664a [PATCH] myri10ge: no need to save MSI and PCIe state in the driver commit: 1a63e846a4099e6fbff86a3b112064378515f254 [PATCH] myri10ge: handle failures in suspend and resume commit: 6ebc087a10c3953d59a61f362bd532f2d9468d1b [PATCH] myri10ge: make wc_fifo usage load-time tunable commit: 21d05db175202adb166b047f35cbb96dbf9b86a5 [PATCH] myri10ge: check that we can get an irq commit: b2db8dd4fdce7f31aa72a854e2433108d474ebb2 [PATCH] myri10ge: update driver version to 1.2.0 commit: 1d39ed565cfcc7c4fe586de621aef495c4f94ffb [PATCH] remove NETIF_F_TSO ifdefery commit: 9dc6f0e789ac8cdd4a7912a9c27027d937a6e784 [PATCH] myri10ge: workaround buggy adopted firmwares commit: 4a2e612a3b0cb6502d2317fa3e9d8f22be5db434 [PATCH] myri10ge: fix copyright and license commit: 355c7265f4745a2aa6f0482d466e2271e6b3c61a [PATCH] myri10ge: fix error checking and return value in myri10ge_allocate_rings commit: 34fdccea93e858aaece3f127628365b546f96857 [PATCH] myri10ge: use pci_map_page to prepare the dmatest buffer commit: 4c882dd808b907652c9a8fe4f7d345cb4dee6a92 [PATCH] myri10ge: prevent 4k rdma on SGI TIOCE chipset commit: 276e26c3160602efea1d0acc9fe7b45bb4c2c8e1 [PATCH] myri10ge: add a wc_enabled flag to myri10ge_priv commit: aafa70eb56edd1cd5332c978bf9b5e224373c980 [PATCH] myri10ge: Serverworks HT2100 provides aligned PCIe completion commit: f761fae1ae1e6e35ae15fce99d225d08d6cff1e7 [PATCH] myri10ge: update wcfifo and intr_coal_delay default values commit: b52a8b7f0aeff5b91921cd53728ac781cdb4cccf [PATCH] myri10ge: fix management of >4kB allocated pages commit: 2ea34672f8cec20b22ccb6ba78e3fe61b44e734a [PATCH] myri10ge: update driver version to 1.3.0-1.226 commit: 917690cd035b422b1ac933ac160d26016aa454ac [PATCH] myri10ge: correctly detect when TSO should be used commit: ae8509b1876e6e1074edc9846296e80983e30502 [PATCH] myri10ge: fix management of the firmware 4KB boundary crossing restriction commit: f19baaeaadf9d77bcc6c122500c70b27c5bbc7a4 [PATCH] myri10ge: more Intel chipsets providing aligned PCIe completions commit: 4b2281c7aa6a6e2615b35f8cebd93656dc6d714c [PATCH] myri10ge: update driver version to 1.3.0-1.233 commit: ea2ae17d6443abddc79480dc9f7af8feacabddc4 [PATCH] [SK_BUFF]: Introduce skb_transport_offset() commit: ab6a5bb6b28a970104a34f0f6959b73cf61bdc72 [PATCH] [TCP]: Introduce tcp_hdrlen() and tcp_optlen() commit: 27d7ff46a3498d3debc6ba68fb8014c702b81170 [PATCH] [SK_BUFF]: Introduce skb_copy_to_linear_data{_offset} commit: cee505db2459aa100a4c3619b8178ec323f1d11e [PATCH] myri10ge: support new firmware counters commit: 772a815804ae778aee7fcf937f3d29ab218ecdc7 [PATCH] myri10ge: update firmware headers commit: 2f76216fe071a67fc110a85a10b1eca038379e11 [PATCH] myri10ge: fix restoring of multicast list after reset commit: 0d6ac257ab556838c3c5b1437a36251c2802285e [PATCH] myri10ge: move the DMA test code into its own function commit: 5443e9ead4f53fd7a43e6846cf10fdc0c5366a93 [PATCH] myri10ge: replace the chipset whitelist with firmware autodetection commit: 6ffdd071e0386645d539a7386f77669777412fea [PATCH] myri10ge: report link up/down in standard ethtool way commit: f181137f9d70dc851dcb418cdad4df7888a0bfd3 [PATCH] myri10ge: limit the number of recoveries commit: 798a95dbd38b113a0f6f3bfe8a2769b261046d09 [PATCH] myri10ge: report when the link partner is running in Myrinet mode commit: 6f8a7c66e2dc8080950d28edc0259f37739aead9 [PATCH] myri10ge: update driver version commit: b245fb675bf03d716b84dcb40885aeb50041dcca [PATCH] myri10ge: SET_NETDEV_DEV() commit: 199126a23a091585670fba023529b73047bb2bad [PATCH] Use is_power_of_2() in myri10ge/myri10ge.c commit: c97884b5de95a11d1668f134a51c1d150d214391 [PATCH] myri10ge: Remove nonsensical limit in the tx done routine commit: 626fda948839fc2f488f527076b8f42b1b7a4687 [PATCH] myri10ge: Use the pause counter to avoid a needless device reset commit: 2972863768cc2ef94734abb22dec6a46b0891307 [PATCH] myri10ge: update driver version to 1.3.2-1.269 commit: a07bc1ffaeee9f05490193f66915ac086c6ea5c9 [PATCH] myri10ge: Add support for PCI device id 9 myri10ge.c | 1276 ++++++++++++++++++++++++++++------------------ myri10ge_compat.h | 14 myri10ge_mcp.h | 115 +++- myri10ge_mcp_gen_header.h | 2 4 files changed, 880 insertions(+), 527 deletions(-) diff --git a/drivers/net/myri10ge/myri10ge.c b/drivers/net/myri10ge/myri10ge.c index d0c4df1..8326926 100644 --- a/drivers/net/myri10ge/myri10ge.c +++ b/drivers/net/myri10ge/myri10ge.c @@ -1,7 +1,7 @@ /************************************************************************* * 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 @@ -16,17 +16,17 @@ * 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 @@ -60,6 +60,7 @@ #include <linux/crc32.h> #include <linux/moduleparam.h> #include <linux/io.h> +#include <linux/log2.h> #include <net/checksum.h> #include <asm/byteorder.h> #include <asm/io.h> @@ -68,10 +69,11 @@ #include <asm/mtrr.h> #endif +#include "myri10ge_compat.h" #include "myri10ge_mcp.h" #include "myri10ge_mcp_gen_header.h" -#define MYRI10GE_VERSION_STR "1.0.0" +#define MYRI10GE_VERSION_STR "1.3.2-1.269" MODULE_DESCRIPTION("Myricom 10G driver (10GbE)"); MODULE_AUTHOR("Maintainer: help@myri.com"); @@ -89,11 +91,16 @@ MODULE_LICENSE("Dual BSD/GPL"); #define MYRI10GE_EEPROM_STRINGS_SIZE 256 #define MYRI10GE_MAX_SEND_DESC_TSO ((65536 / 2048) * 2) -#define MYRI10GE_NO_CONFIRM_DATA 0xffffffff +#define MYRI10GE_NO_CONFIRM_DATA htonl(0xffffffff) #define MYRI10GE_NO_RESPONSE_RESULT 0xffffffff +#define MYRI10GE_ALLOC_ORDER 0 +#define MYRI10GE_ALLOC_SIZE ((1 << MYRI10GE_ALLOC_ORDER) * PAGE_SIZE) +#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) DECLARE_PCI_UNMAP_LEN(len) }; @@ -116,9 +123,14 @@ struct myri10ge_rx_buf { u8 __iomem *wc_fifo; /* w/c rx dma addr fifo address */ struct mcp_kreq_ether_recv *shadow; /* host shadow of recv ring */ struct myri10ge_rx_buffer_state *info; + struct page *page; + dma_addr_t bus; + int page_offset; int cnt; + int fill_cnt; int alloc_fail; int mask; /* number of rx slots -1 */ + int watchdog_needed; }; struct myri10ge_tx_buf { @@ -150,14 +162,15 @@ struct myri10ge_priv { struct myri10ge_rx_buf rx_big; struct myri10ge_rx_done rx_done; int small_bytes; + int big_bytes; struct net_device *dev; struct net_device_stats stats; u8 __iomem *sram; int sram_size; unsigned long board_span; unsigned long iomem_base; - u32 __iomem *irq_claim; - u32 __iomem *irq_deassert; + __be32 __iomem *irq_claim; + __be32 __iomem *irq_deassert; char *mac_addr_string; struct mcp_cmd_response *cmd; dma_addr_t cmd_bus; @@ -165,11 +178,12 @@ struct myri10ge_priv { dma_addr_t fw_stats_bus; struct pci_dev *pdev; int msi_enabled; - unsigned int link_state; + __be32 link_state; unsigned int rdma_tags_available; int intr_coal_delay; - u32 __iomem *intr_coal_delay_ptr; + __be32 __iomem *intr_coal_delay_ptr; int mtrr; + int wc_enabled; int wake_queue; int stop_queue; int down_cnt; @@ -178,20 +192,26 @@ struct myri10ge_priv { struct timer_list watchdog_timer; int watchdog_tx_done; int watchdog_tx_req; + int watchdog_pause; int watchdog_resets; int tx_linearized; int pause; 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; - u32 devctl; - u16 msi_flags; + int fw_multicast_support; u32 read_dma; u32 write_dma; u32 read_write_dma; + u32 link_changes; + u32 msg_enable; }; static char *myri10ge_fw_unaligned = "myri10ge_ethp_z8e.dat"; @@ -214,10 +234,10 @@ module_param(myri10ge_small_bytes, int, S_IRUGO | S_IWUSR); 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"); @@ -235,11 +255,6 @@ module_param(myri10ge_force_firmware, int, S_IRUGO); 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"); @@ -257,6 +272,22 @@ module_param(myri10ge_max_irq_loops, int, S_IRUGO); MODULE_PARM_DESC(myri10ge_max_irq_loops, "Set stuck legacy IRQ detection threshold\n"); +#define MYRI10GE_MSG_DEFAULT NETIF_MSG_LINK + +static int myri10ge_debug = -1; /* defaults above */ +module_param(myri10ge_debug, int, 0); +MODULE_PARM_DESC(myri10ge_debug, "Debug level (0=none,...,16=all)"); + +static int myri10ge_fill_thresh = 256; +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) @@ -264,6 +295,13 @@ MODULE_PARM_DESC(myri10ge_max_irq_loops, #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); +} + static int myri10ge_send_cmd(struct myri10ge_priv *mgp, u32 cmd, struct myri10ge_cmd *data, int atomic) @@ -271,7 +309,7 @@ myri10ge_send_cmd(struct myri10ge_priv *mgp, u32 cmd, struct mcp_cmd *buf; char buf_bytes[sizeof(*buf) + 8]; struct mcp_cmd_response *response = mgp->cmd; - char __iomem *cmd_addr = mgp->sram + MXGEFW_CMD_OFFSET; + char __iomem *cmd_addr = mgp->sram + MXGEFW_ETH_CMD; u32 dma_low, dma_high, result, value; int sleep_total = 0; @@ -287,7 +325,7 @@ myri10ge_send_cmd(struct myri10ge_priv *mgp, u32 cmd, buf->response_addr.low = htonl(dma_low); buf->response_addr.high = htonl(dma_high); - response->result = MYRI10GE_NO_RESPONSE_RESULT; + response->result = htonl(MYRI10GE_NO_RESPONSE_RESULT); mb(); myri10ge_pio_copy(cmd_addr, buf, sizeof(*buf)); @@ -302,14 +340,14 @@ myri10ge_send_cmd(struct myri10ge_priv *mgp, u32 cmd, * (1ms will be enough for those commands) */ for (sleep_total = 0; sleep_total < 1000 - && response->result == MYRI10GE_NO_RESPONSE_RESULT; + && response->result == htonl(MYRI10GE_NO_RESPONSE_RESULT); sleep_total += 10) udelay(10); } else { /* use msleep for most command */ for (sleep_total = 0; sleep_total < 15 - && response->result == MYRI10GE_NO_RESPONSE_RESULT; + && response->result == htonl(MYRI10GE_NO_RESPONSE_RESULT); sleep_total++) msleep(1); } @@ -320,6 +358,10 @@ myri10ge_send_cmd(struct myri10ge_priv *mgp, u32 cmd, if (result == 0) { data->data0 = value; 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", @@ -382,7 +424,7 @@ abort: static void myri10ge_dummy_rdma(struct myri10ge_priv *mgp, int enable) { char __iomem *submit; - u32 buf[16]; + __be32 buf[16]; u32 dma_low, dma_high; int i; @@ -399,12 +441,12 @@ static void myri10ge_dummy_rdma(struct myri10ge_priv *mgp, int enable) buf[0] = htonl(dma_high); /* confirm addr MSW */ buf[1] = htonl(dma_low); /* confirm addr LSW */ - buf[2] = htonl(MYRI10GE_NO_CONFIRM_DATA); /* confirm data */ + buf[2] = MYRI10GE_NO_CONFIRM_DATA; /* confirm data */ buf[3] = htonl(dma_high); /* dummy addr MSW */ buf[4] = htonl(dma_low); /* dummy addr LSW */ buf[5] = htonl(enable); /* enable? */ - submit = mgp->sram + 0xfc01c0; + submit = mgp->sram + MXGEFW_BOOT_DUMMY_RDMA; myri10ge_pio_copy(submit, &buf, sizeof(buf)); for (i = 0; mgp->cmd->data != MYRI10GE_NO_CONFIRM_DATA && i < 20; i++) @@ -419,7 +461,6 @@ myri10ge_validate_firmware(struct myri10ge_priv *mgp, 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) { @@ -430,9 +471,11 @@ myri10ge_validate_firmware(struct myri10ge_priv *mgp, /* 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); @@ -468,7 +511,7 @@ static int myri10ge_load_hotplug_firmware(struct myri10ge_priv *mgp, u32 * size) } /* check id */ - hdr_offset = ntohl(*(u32 *) (fw->data + MCP_HEADER_PTR_OFFSET)); + hdr_offset = ntohl(*(__be32 *) (fw->data + MCP_HEADER_PTR_OFFSET)); if ((hdr_offset & 3) || hdr_offset + sizeof(*hdr) > fw->size) { dev_err(dev, "Bad firmware file\n"); status = -EINVAL; @@ -533,13 +576,25 @@ static int myri10ge_adopt_running_firmware(struct myri10ge_priv *mgp) 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; } static int myri10ge_load_firmware(struct myri10ge_priv *mgp) { char __iomem *submit; - u32 buf[16]; + __be32 buf[16]; u32 dma_low, dma_high, size; int status, i; @@ -589,7 +644,7 @@ static int myri10ge_load_firmware(struct myri10ge_priv *mgp) buf[0] = htonl(dma_high); /* confirm addr MSW */ buf[1] = htonl(dma_low); /* confirm addr LSW */ - buf[2] = htonl(MYRI10GE_NO_CONFIRM_DATA); /* confirm data */ + buf[2] = MYRI10GE_NO_CONFIRM_DATA; /* confirm data */ /* FIX: All newest firmware should un-protect the bottom of * the sram before handoff. However, the very first interfaces @@ -600,7 +655,7 @@ static int myri10ge_load_firmware(struct myri10ge_priv *mgp) buf[5] = htonl(8); /* where to copy to */ buf[6] = htonl(0); /* where to jump to */ - submit = mgp->sram + 0xfc0000; + submit = mgp->sram + MXGEFW_BOOT_HANDOFF; myri10ge_pio_copy(submit, &buf, sizeof(buf)); mb(); @@ -666,12 +721,78 @@ myri10ge_change_promisc(struct myri10ge_priv *mgp, int promisc, int atomic) 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 */ @@ -682,6 +803,8 @@ static int myri10ge_reset(struct myri10ge_priv *mgp) 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); @@ -694,64 +817,19 @@ static int myri10ge_reset(struct myri10ge_priv *mgp) status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_IRQ_ACK_OFFSET, &cmd, 0); - mgp->irq_claim = (__iomem u32 *) (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 u32 *) (mgp->sram + cmd.data0); + mgp->irq_claim = (__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 u32 *) (mgp->sram + cmd.data0); + mgp->intr_coal_delay_ptr = (__iomem __be32 *) (mgp->sram + cmd.data0); if (status != 0) { dev_err(&mgp->pdev->dev, "failed set interrupt parameters\n"); return status; } - __raw_writel(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); + put_be32(htonl(mgp->intr_coal_delay), mgp->intr_coal_delay_ptr); memset(mgp->rx_done.entry, 0, bytes); @@ -764,9 +842,10 @@ static int myri10ge_reset(struct myri10ge_priv *mgp) mgp->rx_small.cnt = 0; mgp->rx_done.idx = 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; } @@ -774,212 +853,207 @@ static inline void myri10ge_submit_8rx(struct mcp_kreq_ether_recv __iomem * dst, struct mcp_kreq_ether_recv *src) { - u32 low; + __be32 low; low = src->addr_low; - src->addr_low = DMA_32BIT_MASK; - myri10ge_pio_copy(dst, src, 8 * sizeof(*src)); + src->addr_low = htonl(DMA_32BIT_MASK); + myri10ge_pio_copy(dst, src, 4 * sizeof(*src)); + mb(); + myri10ge_pio_copy(dst + 4, src + 4, 4 * sizeof(*src)); mb(); src->addr_low = low; - __raw_writel(low, &dst->addr_low); + put_be32(low, &dst->addr_low); 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(int bytes) -{ - struct sk_buff *skb; - unsigned long data, roundup; - - skb = dev_alloc_skb(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(unsigned int bytes) +static inline void myri10ge_vlan_ip_csum(struct sk_buff *skb, __wsum hw_csum) { - struct sk_buff *skb; - unsigned long data, boundary; - - skb = dev_alloc_skb(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; + struct vlan_hdr *vh = (struct vlan_hdr *)(skb->data); - skb_reserve(skb, boundary - data); - return skb; + if ((skb->protocol == htons(ETH_P_8021Q)) && + (vh->h_vlan_encapsulated_proto == htons(ETH_P_IP) || + vh->h_vlan_encapsulated_proto == htons(ETH_P_IPV6))) { + skb->csum = hw_csum; + skb->ip_summed = CHECKSUM_COMPLETE; + } } -/* Allocate just enough space, and verify that the allocated - * space does not cross a 4KB boundary */ -static inline struct sk_buff *myri10ge_alloc_small(int bytes) +static inline void +myri10ge_rx_skb_build(struct sk_buff *skb, u8 * va, + struct skb_frag_struct *rx_frags, int len, int hlen) { - struct sk_buff *skb; - unsigned long roundup, data, end; - - skb = dev_alloc_skb(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(bytes); - } - return skb; + struct skb_frag_struct *skb_frags; + + skb->len = skb->data_len = len; + skb->truesize = len + sizeof(struct sk_buff); + /* attach the page(s) */ + + skb_frags = skb_shinfo(skb)->frags; + while (len > 0) { + memcpy(skb_frags, rx_frags, sizeof(*skb_frags)); + len -= rx_frags->size; + skb_frags++; + rx_frags++; + skb_shinfo(skb)->nr_frags++; + } + + /* pskb_may_pull is not available in irq context, but + * skb_pull() (for ether_pad and eth_type_trans()) requires + * the beginning of the packet in skb_headlen(), move it + * manually */ + 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; + skb->tail += hlen; + skb_pull(skb, MXGEFW_PAD); } -static inline int -myri10ge_getbuf(struct myri10ge_rx_buf *rx, struct pci_dev *pdev, int bytes, - int idx) +static void +myri10ge_alloc_rx_pages(struct myri10ge_priv *mgp, struct myri10ge_rx_buf *rx, + int bytes, int watchdog) { - struct sk_buff *skb; - dma_addr_t bus; - int len, retval = 0; - - bytes += VLAN_HLEN; /* account for 802.1q vlan tag */ + struct page *page; + int idx; - if ((bytes + MXGEFW_PAD) > (4096 - 16) /* linux overhead */ ) - skb = myri10ge_alloc_big(bytes); - else if (myri10ge_skb_cross_4k) - skb = myri10ge_alloc_small_safe(bytes); - else - skb = myri10ge_alloc_small(bytes); + if (unlikely(rx->watchdog_needed && !watchdog)) + return; - 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); + /* try to refill entire ring */ + while (rx->fill_cnt != (rx->cnt + rx->mask + 1)) { + idx = rx->fill_cnt & rx->mask; + if (rx->page_offset + bytes <= MYRI10GE_ALLOC_SIZE) { + /* we can use part of previous page */ + get_page(rx->page); + } else { + /* we need a new page */ + page = + alloc_pages(GFP_ATOMIC | __GFP_COMP, + MYRI10GE_ALLOC_ORDER); + if (unlikely(page == NULL)) { + if (rx->fill_cnt - rx->cnt < 16) + rx->watchdog_needed = 1; + return; + } + rx->page = page; + rx->page_offset = 0; + rx->bus = pci_map_page(mgp->pdev, page, 0, + MYRI10GE_ALLOC_SIZE, + PCI_DMA_FROMDEVICE); + } + rx->info[idx].page = rx->page; + rx->info[idx].page_offset = rx->page_offset; + /* note that this is the address of the start of the + * page */ + pci_unmap_addr_set(&rx->info[idx], bus, rx->bus); + rx->shadow[idx].addr_low = + htonl(MYRI10GE_LOWPART_TO_U32(rx->bus) + rx->page_offset); + rx->shadow[idx].addr_high = + htonl(MYRI10GE_HIGHPART_TO_U32(rx->bus)); + + /* 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 */ + 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, u16 hw_csum) +static inline void +myri10ge_unmap_rx_page(struct pci_dev *pdev, + struct myri10ge_rx_buffer_state *info, int bytes) { - struct vlan_hdr *vh = (struct vlan_hdr *)(skb->data); - - if ((skb->protocol == ntohs(ETH_P_8021Q)) && - (vh->h_vlan_encapsulated_proto == htons(ETH_P_IP) || - vh->h_vlan_encapsulated_proto == htons(ETH_P_IPV6))) { - skb->csum = hw_csum; - skb->ip_summed = CHECKSUM_HW; + /* unmap the recvd page if we're the only or last user of it */ + if (bytes >= MYRI10GE_ALLOC_SIZE / 2 || + (info->page_offset + 2 * bytes) > MYRI10GE_ALLOC_SIZE) { + pci_unmap_page(pdev, (pci_unmap_addr(info, bus) + & ~(MYRI10GE_ALLOC_SIZE - 1)), + MYRI10GE_ALLOC_SIZE, PCI_DMA_FROMDEVICE); } } -static inline unsigned long +#define MYRI10GE_HLEN 64 /* The number of bytes to copy from a + * page into an skb */ + +static inline int myri10ge_rx_done(struct myri10ge_priv *mgp, struct myri10ge_rx_buf *rx, - int bytes, int len, int csum) + int bytes, int len, __wsum csum) { - dma_addr_t bus; struct sk_buff *skb; - int idx, unmap_len; + struct skb_frag_struct rx_frags[MYRI10GE_MAX_FRAGS_PER_FRAME]; + int i, idx, hlen, remainder; + struct pci_dev *pdev = mgp->pdev; + struct net_device *dev = mgp->dev; + u8 *va; + len += MXGEFW_PAD; idx = rx->cnt & rx->mask; - rx->cnt++; + va = page_address(rx->info[idx].page) + rx->info[idx].page_offset; + prefetch(va); + /* Fill skb_frag_struct(s) with data from our receive */ + for (i = 0, remainder = len; remainder > 0; i++) { + myri10ge_unmap_rx_page(pdev, &rx->info[idx], bytes); + rx_frags[i].page = rx->info[idx].page; + rx_frags[i].page_offset = rx->info[idx].page_offset; + if (remainder < MYRI10GE_ALLOC_SIZE) + rx_frags[i].size = remainder; + else + rx_frags[i].size = MYRI10GE_ALLOC_SIZE; + rx->cnt++; + idx = rx->cnt & rx->mask; + remainder -= MYRI10GE_ALLOC_SIZE; + } + + hlen = MYRI10GE_HLEN > len ? len : MYRI10GE_HLEN; - /* 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); + /* allocate an skb to attach the page(s) to. */ - /* try to replace the received skb */ - if (myri10ge_getbuf(rx, mgp->pdev, bytes, idx)) { - /* drop the frame -- the old skbuf is re-cycled */ - mgp->stats.rx_dropped += 1; + skb = netdev_alloc_skb(dev, MYRI10GE_HLEN + 16); + if (unlikely(skb == NULL)) { + mgp->stats.rx_dropped++; + do { + i--; + put_page(rx_frags[i].page); + } while (i != 0); 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); + /* Attach the pages to the skb, and trim off any padding */ + myri10ge_rx_skb_build(skb, va, rx_frags, len, hlen); + if (skb_shinfo(skb)->frags[0].size <= 0) { + put_page(skb_shinfo(skb)->frags[0].page); + skb_shinfo(skb)->nr_frags = 0; + } + skb->protocol = eth_type_trans(skb, dev); + skb->dev = dev; - skb->protocol = eth_type_trans(skb, mgp->dev); - skb->dev = mgp->dev; if (mgp->csum_flag) { - if ((skb->protocol == ntohs(ETH_P_IP)) || - (skb->protocol == ntohs(ETH_P_IPV6))) { - skb->csum = ntohs((u16) csum); - skb->ip_summed = CHECKSUM_HW; + 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, ntohs((u16) csum)); + myri10ge_vlan_ip_csum(skb, csum); } - netif_receive_skb(skb); - mgp->dev->last_rx = jiffies; + dev->last_rx = jiffies; return 1; } @@ -989,7 +1063,6 @@ static inline void myri10ge_tx_done(struct myri10ge_priv *mgp, int mcp_index) struct myri10ge_tx_buf *tx = &mgp->tx; struct sk_buff *skb; int idx, len; - int limit = 0; while (tx->pkt_done != mcp_index) { idx = tx->done & tx->mask; @@ -1020,11 +1093,6 @@ static inline void myri10ge_tx_done(struct myri10ge_priv *mgp, int mcp_index) bus), len, PCI_DMA_TODEVICE); } - - /* limit potential for livelock by only handling - * 2 full tx rings per call */ - if (unlikely(++limit > 2 * tx->mask)) - break; } /* start the queue if we've stopped it */ if (netif_queue_stopped(mgp->dev) @@ -1044,19 +1112,19 @@ static inline void myri10ge_clean_rx_done(struct myri10ge_priv *mgp, int *limit) int idx = rx_done->idx; int cnt = rx_done->cnt; u16 length; - u16 checksum; + __wsum checksum; while (rx_done->entry[idx].length != 0 && *limit != 0) { length = ntohs(rx_done->entry[idx].length); rx_done->entry[idx].length = 0; - checksum = ntohs(rx_done->entry[idx].checksum); + checksum = csum_unfold(rx_done->entry[idx].checksum); if (length <= mgp->small_bytes) rx_ok = myri10ge_rx_done(mgp, &mgp->rx_small, mgp->small_bytes, 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; @@ -1071,6 +1139,14 @@ static inline void myri10ge_clean_rx_done(struct myri10ge_priv *mgp, int *limit) 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) @@ -1078,16 +1154,27 @@ 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) { - printk(KERN_INFO "myri10ge: %s: link up\n", - mgp->dev->name); + 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", + mgp->dev->name); netif_carrier_on(mgp->dev); + mgp->link_changes++; } else { - printk(KERN_INFO "myri10ge: %s: link down\n", - mgp->dev->name); + if (netif_msg_link(mgp)) + printk(KERN_INFO + "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++; } } if (mgp->rdma_tags_available != @@ -1120,7 +1207,7 @@ static int myri10ge_poll(struct net_device *netdev, int *budget) if (rx_done->entry[rx_done->idx].length == 0 || !netif_running(netdev)) { netif_rx_complete(netdev); - __raw_writel(htonl(3), mgp->irq_claim); + put_be32(htonl(3), mgp->irq_claim); return 0; } return 1; @@ -1144,7 +1231,7 @@ static irqreturn_t myri10ge_intr(int irq, void *arg, struct pt_regs *regs) netif_rx_schedule(mgp->dev); if (!mgp->msi_enabled) { - __raw_writel(0, mgp->irq_deassert); + put_be32(0, mgp->irq_deassert); if (!myri10ge_deassert_wait) stats->valid = 0; mb(); @@ -1173,7 +1260,7 @@ static irqreturn_t myri10ge_intr(int irq, void *arg, struct pt_regs *regs) myri10ge_check_statblock(mgp); - __raw_writel(htonl(3), mgp->irq_claim + 1); + put_be32(htonl(3), mgp->irq_claim + 1); return (IRQ_HANDLED); } @@ -1211,7 +1298,7 @@ myri10ge_set_coalesce(struct net_device *netdev, struct ethtool_coalesce *coal) struct myri10ge_priv *mgp = netdev_priv(netdev); mgp->intr_coal_delay = coal->rx_coalesce_usecs; - __raw_writel(htonl(mgp->intr_coal_delay), mgp->intr_coal_delay_ptr); + put_be32(htonl(mgp->intr_coal_delay), mgp->intr_coal_delay_ptr); return 0; } @@ -1289,7 +1376,10 @@ static const char myri10ge_gstrings_stats[][ETH_GSTRING_LEN] = { "serial_number", "tx_pkt_start", "tx_pkt_done", "tx_req", "tx_done", "rx_small_cnt", "rx_big_cnt", "wake_queue", "stop_queue", "watchdog_resets", "tx_linearized", - "link_up", "dropped_link_overflow", "dropped_link_error_or_filtered", + "link_changes", "link_up", "dropped_link_overflow", + "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" }; @@ -1324,7 +1414,7 @@ myri10ge_get_ethtool_stats(struct net_device *netdev, 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; @@ -1341,16 +1431,36 @@ myri10ge_get_ethtool_stats(struct net_device *netdev, data[i++] = (unsigned int)mgp->stop_queue; data[i++] = (unsigned int)mgp->watchdog_resets; data[i++] = (unsigned int)mgp->tx_linearized; + data[i++] = (unsigned int)mgp->link_changes; data[i++] = (unsigned int)ntohl(mgp->fw_stats->link_up); 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); data[i++] = (unsigned int)ntohl(mgp->fw_stats->dropped_overrun); data[i++] = (unsigned int)ntohl(mgp->fw_stats->dropped_no_small_buffer); data[i++] = (unsigned int)ntohl(mgp->fw_stats->dropped_no_big_buffer); } +static void myri10ge_set_msglevel(struct net_device *netdev, u32 value) +{ + struct myri10ge_priv *mgp = netdev_priv(netdev); + mgp->msg_enable = value; +} + +static u32 myri10ge_get_msglevel(struct net_device *netdev) +{ + struct myri10ge_priv *mgp = netdev_priv(netdev); + return mgp->msg_enable; +} + static struct ethtool_ops myri10ge_ethtool_ops = { .get_settings = myri10ge_get_settings, .get_drvinfo = myri10ge_get_drvinfo, @@ -1365,13 +1475,14 @@ static struct ethtool_ops myri10ge_ethtool_ops = { .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 + .get_ethtool_stats = myri10ge_get_ethtool_stats, + .set_msglevel = myri10ge_set_msglevel, + .get_msglevel = myri10ge_get_msglevel }; static int myri10ge_allocate_rings(struct net_device *dev) @@ -1390,6 +1501,8 @@ static int myri10ge_allocate_rings(struct net_device *dev) 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); @@ -1397,6 +1510,8 @@ static int myri10ge_allocate_rings(struct net_device *dev) 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) @@ -1437,57 +1552,48 @@ static int myri10ge_allocate_rings(struct net_device *dev) 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->pdev, - 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->pdev, - 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: @@ -1520,30 +1626,24 @@ static void myri10ge_free_rings(struct net_device *dev) 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; @@ -1585,6 +1685,41 @@ static void myri10ge_free_rings(struct net_device *dev) 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; @@ -1600,10 +1735,13 @@ static int myri10ge_open(struct net_device *dev) 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 @@ -1611,19 +1749,18 @@ static int myri10ge_open(struct net_device *dev) */ 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); @@ -1644,30 +1781,38 @@ static int myri10ge_open(struct net_device *dev) "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) { - mgp->tx.wc_fifo = (u8 __iomem *) mgp->sram + 0x200000; - mgp->rx_small.wc_fifo = (u8 __iomem *) mgp->sram + 0x300000; - mgp->rx_big.wc_fifo = (u8 __iomem *) mgp->sram + 0x340000; + 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 = + (u8 __iomem *) mgp->sram + MXGEFW_ETH_RECV_BIG; } else { mgp->tx.wc_fifo = NULL; mgp->rx_small.wc_fifo = NULL; 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 (!is_power_of_2(big_pow2)) + 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; @@ -1686,14 +1831,28 @@ static int myri10ge_open(struct net_device *dev) cmd.data0 = MYRI10GE_LOWPART_TO_U32(mgp->fw_stats_bus); cmd.data1 = MYRI10GE_HIGHPART_TO_U32(mgp->fw_stats_bus); - status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_STATS_DMA, &cmd, 0); + cmd.data2 = sizeof(struct mcp_irq_data); + status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_STATS_DMA_V2, &cmd, 0); + if (status == -ENOSYS) { + dma_addr_t bus = mgp->fw_stats_bus; + bus += offsetof(struct mcp_irq_data, send_done_count); + cmd.data0 = MYRI10GE_LOWPART_TO_U32(bus); + cmd.data1 = MYRI10GE_HIGHPART_TO_U32(bus); + status = myri10ge_send_cmd(mgp, + MXGEFW_CMD_SET_STATS_DMA_OBSOLETE, + &cmd, 0); + /* Firmware cannot support multicast without STATS_DMA_V2 */ + mgp->fw_multicast_support = 0; + } else { + mgp->fw_multicast_support = 1; + } if (status) { printk(KERN_ERR "myri10ge: %s: Couldn't set stats DMA\n", dev->name); goto abort_with_rings; } - mgp->link_state = -1; + mgp->link_state = htonl(~0U); mgp->rdma_tags_available = 15; netif_poll_enable(mgp->dev); /* must happen prior to any irq */ @@ -1716,6 +1875,9 @@ static int myri10ge_open(struct net_device *dev) abort_with_rings: myri10ge_free_rings(dev); +abort_with_irq: + myri10ge_free_irq(mgp); + abort_with_nothing: mgp->running = MYRI10GE_ETH_STOPPED; return -ENOMEM; @@ -1752,7 +1914,7 @@ static int myri10ge_close(struct net_device *dev) 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; @@ -1821,7 +1983,7 @@ myri10ge_submit_req(struct myri10ge_tx_buf *tx, struct mcp_kreq_ether_send *src, /* re-write the last 32-bits with the valid flags */ src->flags = last_flags; - __raw_writel(*((u32 *) src + 3), (u32 __iomem *) dst + 3); + put_be32(*((__be32 *) src + 3), (__be32 __iomem *) dst + 3); tx->req += cnt; mb(); } @@ -1841,7 +2003,8 @@ myri10ge_submit_req_wc(struct myri10ge_tx_buf *tx, if (cnt > 0) { /* pad it to 64 bytes. The src is 64 bytes bigger than it * needs to be so that we don't overrun it */ - myri10ge_pio_copy(tx->wc_fifo + (cnt << 18), src, 64); + myri10ge_pio_copy(tx->wc_fifo + MXGEFW_ETH_SEND_OFFSET(cnt), + src, 64); mb(); } } @@ -1863,7 +2026,8 @@ static int myri10ge_xmit(struct sk_buff *skb, struct net_device *dev) struct myri10ge_tx_buf *tx = &mgp->tx; struct skb_frag_struct *frag; dma_addr_t bus; - u32 low, high_swapped; + u32 low; + __be32 high_swapped; unsigned int len; int idx, last_idx, avail, frag_cnt, frag_idx, count, mss, max_segments; u16 pseudo_hdr_offset, cksum_offset; @@ -1877,13 +2041,10 @@ again: 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 */ @@ -1897,8 +2058,8 @@ again: pseudo_hdr_offset = 0; odd_flag = 0; flags = (MXGEFW_FLAGS_NO_TSO | MXGEFW_FLAGS_FIRST); - if (likely(skb->ip_summed == CHECKSUM_HW)) { - cksum_offset = (skb->h.raw - skb->data); + if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) { + cksum_offset = skb_transport_offset(skb); pseudo_hdr_offset = (skb->h.raw + skb->csum) - skb->data; /* If the headers are excessively large, then we must * fall back to a software checksum */ @@ -1908,7 +2069,6 @@ again: cksum_offset = 0; pseudo_hdr_offset = 0; } else { - pseudo_hdr_offset = htons(pseudo_hdr_offset); odd_flag = MXGEFW_FLAGS_ALIGN_ODD; flags |= MXGEFW_FLAGS_CKSUM; } @@ -1916,7 +2076,6 @@ again: 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); @@ -1925,14 +2084,13 @@ again: * 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 = htons(mss); + 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; @@ -2000,7 +2158,6 @@ again: 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; @@ -2027,10 +2184,9 @@ again: (small * MXGEFW_FLAGS_SMALL); } } -#endif /* NETIF_F_TSO */ req->addr_high = high_swapped; req->addr_low = htonl(low); - req->pseudo_hdr_offset = pseudo_hdr_offset; + req->pseudo_hdr_offset = htons(pseudo_hdr_offset); req->pad = 0; /* complete solid 16-byte block; does this matter? */ req->rdma_count = 1; req->length = htons(seglen); @@ -2064,14 +2220,12 @@ again: } (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) @@ -2140,9 +2294,81 @@ static struct net_device_stats *myri10ge_get_stats(struct net_device *dev) static void myri10ge_set_multicast_list(struct net_device *dev) { + struct myri10ge_cmd cmd; + struct myri10ge_priv *mgp; + struct dev_mc_list *mc_list; + __be32 data[2] = {0, 0}; + int err; + + mgp = netdev_priv(dev); /* can be called from atomic contexts, * pass 1 to force atomicity in myri10ge_send_cmd() */ - myri10ge_change_promisc(netdev_priv(dev), dev->flags & IFF_PROMISC, 1); + myri10ge_change_promisc(mgp, dev->flags & IFF_PROMISC, 1); + + /* This firmware is known to not support multicast */ + if (!mgp->fw_multicast_support) + return; + + /* Disable multicast filtering */ + + err = myri10ge_send_cmd(mgp, MXGEFW_ENABLE_ALLMULTI, &cmd, 1); + if (err != 0) { + printk(KERN_ERR "myri10ge: %s: Failed MXGEFW_ENABLE_ALLMULTI," + " error status: %d\n", dev->name, err); + goto abort; + } + + if ((dev->flags & IFF_ALLMULTI) || mgp->adopted_rx_filter_bug) { + /* request to disable multicast filtering, so quit here */ + return; + } + + /* Flush the filters */ + + err = myri10ge_send_cmd(mgp, MXGEFW_LEAVE_ALL_MULTICAST_GROUPS, + &cmd, 1); + if (err != 0) { + printk(KERN_ERR + "myri10ge: %s: Failed MXGEFW_LEAVE_ALL_MULTICAST_GROUPS" + ", error status: %d\n", dev->name, err); + goto abort; + } + + /* Walk the multicast list, and add each address */ + for (mc_list = dev->mc_list; mc_list != NULL; mc_list = mc_list->next) { + memcpy(data, &mc_list->dmi_addr, 6); + cmd.data0 = ntohl(data[0]); + cmd.data1 = ntohl(data[1]); + err = myri10ge_send_cmd(mgp, MXGEFW_JOIN_MULTICAST_GROUP, + &cmd, 1); + + if (err != 0) { + printk(KERN_ERR "myri10ge: %s: Failed " + "MXGEFW_JOIN_MULTICAST_GROUP, error status:" + "%d\t", dev->name, err); + printk(KERN_ERR "MAC %02x:%02x:%02x:%02x:%02x:%02x\n", + ((unsigned char *)&mc_list->dmi_addr)[0], + ((unsigned char *)&mc_list->dmi_addr)[1], + ((unsigned char *)&mc_list->dmi_addr)[2], + ((unsigned char *)&mc_list->dmi_addr)[3], + ((unsigned char *)&mc_list->dmi_addr)[4], + ((unsigned char *)&mc_list->dmi_addr)[5] + ); + goto abort; + } + } + /* Enable multicast filtering */ + err = myri10ge_send_cmd(mgp, MXGEFW_DISABLE_ALLMULTI, &cmd, 1); + if (err != 0) { + printk(KERN_ERR "myri10ge: %s: Failed MXGEFW_DISABLE_ALLMULTI," + "error status: %d\n", dev->name, err); + goto abort; + } + + return; + +abort: + return; } static int myri10ge_set_mac_address(struct net_device *dev, void *addr) @@ -2265,8 +2491,6 @@ static void myri10ge_enable_ecrc(struct myri10ge_priv *mgp) 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; } /* @@ -2288,30 +2512,88 @@ static void myri10ge_enable_ecrc(struct myri10ge_priv *mgp) * firmware image, and set tx.boundary to 4KB. */ -#define PCI_DEVICE_ID_SERVERWORKS_HT2000_PCIE 0x0132 - -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) { - myri10ge_enable_ecrc(mgp); - - /* Check to see if the upstream bridge is known to - * provide aligned completions */ - if (bridge - /* ServerWorks HT2000/HT1000 */ - && bridge->vendor == PCI_VENDOR_ID_SERVERWORKS - && bridge->device == - PCI_DEVICE_ID_SERVERWORKS_HT2000_PCIE) { - dev_info(&mgp->pdev->dev, - "Assuming aligned completions (0x%x:0x%x)\n", - bridge->vendor, bridge->device); + int link_width, exp_cap; + u16 lnk; + + exp_cap = pci_find_capability(mgp->pdev, PCI_CAP_ID_EXP); + pci_read_config_word(mgp->pdev, exp_cap + PCI_EXP_LNKSTA, &lnk); + link_width = (lnk >> 4) & 0x3f; + + /* Check to see if Link is less than 8 or if the + * upstream bridge is known to provide aligned + * completions */ + if (link_width < 8) { + dev_info(&mgp->pdev->dev, "PCIE x%d Link\n", + link_width); mgp->tx.boundary = 4096; mgp->fw_name = myri10ge_fw_aligned; + } else { + myri10ge_firmware_probe(mgp); } } else { if (myri10ge_force_firmware == 1) { @@ -2333,34 +2615,6 @@ static void myri10ge_select_firmware(struct myri10ge_priv *mgp) } } -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) @@ -2381,11 +2635,10 @@ 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) @@ -2407,34 +2660,33 @@ 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); @@ -2482,8 +2734,14 @@ static void myri10ge_watchdog(void *arg) * 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. @@ -2491,7 +2749,11 @@ static void myri10ge_watchdog(void *arg) * 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 @@ -2540,20 +2802,45 @@ static void myri10ge_watchdog(void *arg) static void myri10ge_watchdog_timer(unsigned long arg) { struct myri10ge_priv *mgp; + u32 rx_pause_cnt; 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; + } + rx_pause_cnt = ntohl(mgp->fw_stats->dropped_pause); + if (mgp->tx.req != mgp->tx.done && mgp->tx.done == mgp->watchdog_tx_done && - mgp->watchdog_tx_req != mgp->watchdog_tx_done) + mgp->watchdog_tx_req != mgp->watchdog_tx_done) { /* nic seems like it might be stuck.. */ - schedule_work(&mgp->watchdog_work); - else - /* rearm timer */ - mod_timer(&mgp->watchdog_timer, - jiffies + myri10ge_watchdog_timeout * HZ); - + if (rx_pause_cnt != mgp->watchdog_pause) { + if (net_ratelimit()) + printk(KERN_WARNING "myri10ge %s:" + "TX paused, check link partner\n", + mgp->dev->name); + } else { + schedule_work(&mgp->watchdog_work); + return; + } + } + /* rearm timer */ + mod_timer(&mgp->watchdog_timer, + jiffies + myri10ge_watchdog_timeout * HZ); mgp->watchdog_tx_done = mgp->tx.done; mgp->watchdog_tx_req = mgp->tx.req; + mgp->watchdog_pause = rx_pause_cnt; } static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) @@ -2574,6 +2861,8 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) return -ENOMEM; } + SET_NETDEV_DEV(netdev, &pdev->dev); + mgp = netdev_priv(netdev); memset(mgp, 0, sizeof(*mgp)); mgp->dev = netdev; @@ -2581,6 +2870,7 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) mgp->csum_flag = MXGEFW_FLAGS_CKSUM; mgp->pause = myri10ge_flow_control; mgp->intr_coal_delay = myri10ge_intr_coal_delay; + mgp->msg_enable = netif_msg_init(myri10ge_debug, MYRI10GE_MSG_DEFAULT); init_waitqueue_head(&mgp->down_wq); if (pci_enable_device(pdev)) { @@ -2588,7 +2878,6 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) status = -ENODEV; goto abort_with_netdev; } - myri10ge_select_firmware(mgp); /* Find the vendor-specific cap so we can check * the reboot register later on */ @@ -2640,9 +2929,12 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) 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 = @@ -2678,6 +2970,8 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) 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"); @@ -2690,23 +2984,6 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) 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; @@ -2718,7 +2995,6 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) 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; @@ -2728,9 +3004,18 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) 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, @@ -2741,19 +3026,17 @@ static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent) 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); @@ -2804,12 +3087,12 @@ static void myri10ge_remove(struct pci_dev *pdev) 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); @@ -2831,9 +3114,12 @@ static void myri10ge_remove(struct pci_dev *pdev) } #define PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E 0x0008 +#define PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E_9 0x0009 static struct pci_device_id myri10ge_pci_tbl[] = { {PCI_DEVICE(PCI_VENDOR_ID_MYRICOM, PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E)}, + {PCI_DEVICE + (PCI_VENDOR_ID_MYRICOM, PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E_9)}, {0}, }; diff --git a/drivers/net/myri10ge/myri10ge_compat.h b/drivers/net/myri10ge/myri10ge_compat.h new file mode 100644 index 0000000..2c41db9 --- /dev/null +++ b/drivers/net/myri10ge/myri10ge_compat.h @@ -0,0 +1,14 @@ +#ifndef __MYRI10GE_COMPAT_H__ +#define __MYRI10GE_COMPAT_H__ + +/* RHEL5 compat */ +typedef __u16 __bitwise __sum16; +typedef __u32 __bitwise __wsum; + +static inline __wsum csum_unfold(__sum16 n) +{ + return (__force __wsum)n; +} + +#endif /* __MYRI10GE_COMPAT_H__ */ + diff --git a/drivers/net/myri10ge/myri10ge_mcp.h b/drivers/net/myri10ge/myri10ge_mcp.h index 0a6cae6..a1d2a22 100644 --- a/drivers/net/myri10ge/myri10ge_mcp.h +++ b/drivers/net/myri10ge/myri10ge_mcp.h @@ -6,23 +6,23 @@ /* 8 Bytes */ struct mcp_dma_addr { - u32 high; - u32 low; + __be32 high; + __be32 low; }; /* 4 Bytes */ struct mcp_slot { - u16 checksum; - u16 length; + __sum16 checksum; + __be16 length; }; /* 64 Bytes */ struct mcp_cmd { - u32 cmd; - u32 data0; /* will be low portion if data > 32 bits */ + __be32 cmd; + __be32 data0; /* will be low portion if data > 32 bits */ /* 8 */ - u32 data1; /* will be high portion if data > 32 bits */ - u32 data2; /* currently unused.. */ + __be32 data1; /* will be high portion if data > 32 bits */ + __be32 data2; /* currently unused.. */ /* 16 */ struct mcp_dma_addr response_addr; /* 24 */ @@ -31,8 +31,8 @@ struct mcp_cmd { /* 8 Bytes */ struct mcp_cmd_response { - u32 data; - u32 result; + __be32 data; + __be32 result; }; /* @@ -73,10 +73,10 @@ union mcp_pso_or_cumlen { /* 16 Bytes */ struct mcp_kreq_ether_send { - u32 addr_high; - u32 addr_low; - u16 pseudo_hdr_offset; - u16 length; + __be32 addr_high; + __be32 addr_low; + __be16 pseudo_hdr_offset; + __be16 length; u8 pad; u8 rdma_count; u8 cksum_offset; /* where to start computing cksum */ @@ -85,13 +85,25 @@ struct mcp_kreq_ether_send { /* 8 Bytes */ struct mcp_kreq_ether_recv { - u32 addr_high; - u32 addr_low; + __be32 addr_high; + __be32 addr_low; }; /* Commands */ -#define MXGEFW_CMD_OFFSET 0xf80000 +#define MXGEFW_BOOT_HANDOFF 0xfc0000 +#define MXGEFW_BOOT_DUMMY_RDMA 0xfc01c0 + +#define MXGEFW_ETH_CMD 0xf80000 +#define MXGEFW_ETH_SEND_4 0x200000 +#define MXGEFW_ETH_SEND_1 0x240000 +#define MXGEFW_ETH_SEND_2 0x280000 +#define MXGEFW_ETH_SEND_3 0x2c0000 +#define MXGEFW_ETH_RECV_SMALL 0x300000 +#define MXGEFW_ETH_RECV_BIG 0x340000 + +#define MXGEFW_ETH_SEND(n) (0x200000 + (((n) & 0x03) * 0x40000)) +#define MXGEFW_ETH_SEND_OFFSET(n) (MXGEFW_ETH_SEND(n) - MXGEFW_ETH_SEND_4) enum myri10ge_mcp_cmd_type { MXGEFW_CMD_NONE = 0, @@ -154,7 +166,7 @@ enum myri10ge_mcp_cmd_type { MXGEFW_CMD_SET_MTU, MXGEFW_CMD_GET_INTR_COAL_DELAY_OFFSET, /* in microseconds */ MXGEFW_CMD_SET_STATS_INTERVAL, /* in microseconds */ - MXGEFW_CMD_SET_STATS_DMA, + MXGEFW_CMD_SET_STATS_DMA_OBSOLETE, /* replaced by SET_STATS_DMA_V2 */ MXGEFW_ENABLE_PROMISC, MXGEFW_DISABLE_PROMISC, @@ -168,7 +180,33 @@ enum myri10ge_mcp_cmd_type { * data2 = RDMA length (MSH), WDMA length (LSH) * command return data = repetitions (MSH), 0.5-ms ticks (LSH) */ - MXGEFW_DMA_TEST + MXGEFW_DMA_TEST, + + MXGEFW_ENABLE_ALLMULTI, + MXGEFW_DISABLE_ALLMULTI, + + /* returns MXGEFW_CMD_ERROR_MULTICAST + * if there is no room in the cache + * data0,MSH(data1) = multicast group address */ + MXGEFW_JOIN_MULTICAST_GROUP, + /* returns MXGEFW_CMD_ERROR_MULTICAST + * if the address is not in the cache, + * or is equal to FF-FF-FF-FF-FF-FF + * data0,MSH(data1) = multicast group address */ + MXGEFW_LEAVE_MULTICAST_GROUP, + MXGEFW_LEAVE_ALL_MULTICAST_GROUPS, + + MXGEFW_CMD_SET_STATS_DMA_V2, + /* data0, data1 = bus addr, + * data2 = sizeof(struct mcp_irq_data) from driver point of view, allows + * adding new stuff to mcp_irq_data without changing the ABI */ + + MXGEFW_CMD_UNALIGNED_TEST, + /* same than DMA_TEST (same args) but abort with UNALIGNED on unaligned + * chipset */ + + MXGEFW_CMD_UNALIGNED_STATUS + /* return data = boolean, true if the chipset is known to be unaligned */ }; enum myri10ge_mcp_cmd_status { @@ -180,21 +218,36 @@ enum myri10ge_mcp_cmd_status { MXGEFW_CMD_ERROR_CLOSED, MXGEFW_CMD_ERROR_HASH_ERROR, MXGEFW_CMD_ERROR_BAD_PORT, - MXGEFW_CMD_ERROR_RESOURCES + MXGEFW_CMD_ERROR_RESOURCES, + MXGEFW_CMD_ERROR_MULTICAST, + MXGEFW_CMD_ERROR_UNALIGNED }; -/* 40 Bytes */ +#define MXGEFW_OLD_IRQ_DATA_LEN 40 + struct mcp_irq_data { - u32 send_done_count; - - u32 link_up; - u32 dropped_link_overflow; - u32 dropped_link_error_or_filtered; - u32 dropped_runt; - u32 dropped_overrun; - u32 dropped_no_small_buffer; - u32 dropped_no_big_buffer; - u32 rdma_tags_available; + /* add new counters at the beginning */ + __be32 future_use[1]; + __be32 dropped_pause; + __be32 dropped_unicast_filtered; + __be32 dropped_bad_crc32; + __be32 dropped_bad_phy; + __be32 dropped_multicast_filtered; + /* 40 Bytes */ + __be32 send_done_count; + +#define MXGEFW_LINK_DOWN 0 +#define MXGEFW_LINK_UP 1 +#define MXGEFW_LINK_MYRINET 2 +#define MXGEFW_LINK_UNKNOWN 3 + __be32 link_up; + __be32 dropped_link_overflow; + __be32 dropped_link_error_or_filtered; + __be32 dropped_runt; + __be32 dropped_overrun; + __be32 dropped_no_small_buffer; + __be32 dropped_no_big_buffer; + __be32 rdma_tags_available; u8 tx_stopped; u8 link_down; diff --git a/drivers/net/myri10ge/myri10ge_mcp_gen_header.h b/drivers/net/myri10ge/myri10ge_mcp_gen_header.h index 487f779..16a810d 100644 --- a/drivers/net/myri10ge/myri10ge_mcp_gen_header.h +++ b/drivers/net/myri10ge/myri10ge_mcp_gen_header.h @@ -36,7 +36,7 @@ struct mcp_gen_header { /* the first 4 fields are filled at compile time */ unsigned header_length; - unsigned mcp_type; + __be32 mcp_type; char version[128]; unsigned mcp_globals; /* pointer to mcp-type specific structure */