Skip to contentSkip to conversation

Broadcom 10G NIC Poll Mode Driver

PATCH In review Stephen Hemminger 4 participants 17 messages Feb 6, 2015 – Feb 13, 2015 12 messages not in the archive Library code

lore
Share & tools
lore
#0 lore raw

cover Broadcom 10G NIC Poll Mode Driver

From: Stephen Hemminger <[email protected]>

These are the patches to enable supporting the Broadcom NetExtreme II 10G devices (show up as bnx2x on Linux).

The driver has only been tested on Linux, there maybe issues with firmware loading and PCI config access on BSD.

Stephen Hemminger (4):
pci: allow access to PCI config space
bcm: add BCM pci device ids
bcm: new poll mode driver
bcm: enable BCM poll mode driver in config

 config/common_linuxapp                          |    10 +
 lib/Makefile                                    |     1 +
 lib/librte_eal/common/include/rte_pci.h         |    29 +
 lib/librte_eal/common/include/rte_pci_dev_ids.h |    30 +
 lib/librte_eal/linuxapp/eal/eal_pci.c           |    15 +
 lib/librte_eal/linuxapp/eal/eal_pci_uio.c       |    10 +
 lib/librte_eal/linuxapp/eal/rte_eal_version.map |     2 +
 lib/librte_pmd_bcm/Makefile                     |    28 +
 lib/librte_pmd_bcm/bcm.c                        | 11817 +++++++++++++++++++
 lib/librte_pmd_bcm/bcm.h                        |  1998 ++++
 lib/librte_pmd_bcm/bcm_ethdev.c                 |   544 +
 lib/librte_pmd_bcm/bcm_ethdev.h                 |    79 +
 lib/librte_pmd_bcm/bcm_logs.h                   |    51 +
 lib/librte_pmd_bcm/bcm_rxtx.c                   |   487 +
 lib/librte_pmd_bcm/bcm_rxtx.h                   |    85 +
 lib/librte_pmd_bcm/bcm_stats.c                  |  1619 +++
 lib/librte_pmd_bcm/bcm_stats.h                  |   633 +
 lib/librte_pmd_bcm/bcm_vfpf.c                   |   597 +
 lib/librte_pmd_bcm/bcm_vfpf.h                   |   315 +
 lib/librte_pmd_bcm/debug.c                      |   113 +
 lib/librte_pmd_bcm/ecore_fw_defs.h              |   423 +
 lib/librte_pmd_bcm/ecore_hsi.h                  |  6349 ++++++++++
 lib/librte_pmd_bcm/ecore_init.h                 |   842 ++
 lib/librte_pmd_bcm/ecore_init_ops.h             |   886 ++
 lib/librte_pmd_bcm/ecore_mfw_req.h              |   207 +
 lib/librte_pmd_bcm/ecore_reg.h                  |  3664 ++++++
 lib/librte_pmd_bcm/ecore_sp.c                   |  5455 +++++++++
 lib/librte_pmd_bcm/ecore_sp.h                   |  1796 +++
 lib/librte_pmd_bcm/elink.c                      | 13378 ++++++++++++++++++++++
 lib/librte_pmd_bcm/elink.h                      |   610 +
 30 files changed, 52073 insertions(+)
 create mode 100644 lib/librte_pmd_bcm/Makefile
 create mode 100644 lib/librte_pmd_bcm/bcm.c
 create mode 100644 lib/librte_pmd_bcm/bcm.h
 create mode 100644 lib/librte_pmd_bcm/bcm_ethdev.c
 create mode 100644 lib/librte_pmd_bcm/bcm_ethdev.h
 create mode 100644 lib/librte_pmd_bcm/bcm_logs.h
 create mode 100644 lib/librte_pmd_bcm/bcm_rxtx.c
 create mode 100644 lib/librte_pmd_bcm/bcm_rxtx.h
 create mode 100644 lib/librte_pmd_bcm/bcm_stats.c
 create mode 100644 lib/librte_pmd_bcm/bcm_stats.h
 create mode 100644 lib/librte_pmd_bcm/bcm_vfpf.c
 create mode 100644 lib/librte_pmd_bcm/bcm_vfpf.h
 create mode 100644 lib/librte_pmd_bcm/debug.c
 create mode 100644 lib/librte_pmd_bcm/ecore_fw_defs.h
 create mode 100644 lib/librte_pmd_bcm/ecore_hsi.h
 create mode 100644 lib/librte_pmd_bcm/ecore_init.h
 create mode 100644 lib/librte_pmd_bcm/ecore_init_ops.h
 create mode 100644 lib/librte_pmd_bcm/ecore_mfw_req.h
 create mode 100644 lib/librte_pmd_bcm/ecore_reg.h
 create mode 100644 lib/librte_pmd_bcm/ecore_sp.c
 create mode 100644 lib/librte_pmd_bcm/ecore_sp.h
 create mode 100644 lib/librte_pmd_bcm/elink.c
 create mode 100644 lib/librte_pmd_bcm/elink.h
Signature · 1 line
2.1.4

A message that is not in the archive (sent off-list or not yet received): 1423247795-22399-1-git-send-email-stephen-OTpzqLSitTUnbdJkjeBofR2eb7JE58TQ@public.gmane.org

#2 lore raw

PATCH 1/4 pci: allow access to PCI config space

From: Stephen Hemminger <[email protected]>

Some drivers need ability to access PCI config (for example for power management). This adds an abstraction to do this; only implemented on Linux, but should be possible on BSD.

---
 lib/librte_eal/common/include/rte_pci.h         | 29 +++++++++++++++++++++++++
 lib/librte_eal/linuxapp/eal/eal_pci.c           | 15 +++++++++++++
 lib/librte_eal/linuxapp/eal/eal_pci_uio.c       | 10 +++++++++
 lib/librte_eal/linuxapp/eal/rte_eal_version.map |  2 ++
 4 files changed, 56 insertions(+)
lib/librte_eal/common/include/rte_pci.h +29 -0
index 66ed793..a78081f 100644
@@ -152,6 +152,7 @@ struct rte_pci_device {
 	uint16_t max_vfs;                       /**< sriov enable if not zero */
 	int numa_node;                          /**< NUMA node connection */
 	struct rte_devargs *devargs;            /**< Device user arguments */
+	int config_fd;				/**< PCI config access */
 };
 
 /** Any PCI device identifier (vendor, device, ...) */
@@ -298,6 +299,34 @@ void rte_eal_pci_register(struct rte_pci_driver *driver);
  */
 void rte_eal_pci_unregister(struct rte_pci_driver *driver);
 
+/**
+ * Read PCI config space.
+ *
+ * @param device
+ *   A pointer to a rte_pci_device structure describing the device
+ *   to use
+ * @param buf
+ *   A data buffer where the bytes should be read into
+ * @param size
+ *   The length of the data buffer.
+ */
+int rte_eal_pci_read_config(const struct rte_pci_device *device,
+			    void *buf, size_t len, off_t offset);
+
+/**
+ * Write PCI config space.
+ *
+ * @param device
+ *   A pointer to a rte_pci_device structure describing the device
+ *   to use
+ * @param buf
+ *   A data buffer containing the bytes should be written
+ * @param size
+ *   The length of the data buffer.
+ */
+int rte_eal_pci_write_config(const struct rte_pci_device *device,
+			     const void *buf, size_t len, off_t offset);
+
 #ifdef __cplusplus
 }
 #endif
lib/librte_eal/linuxapp/eal/eal_pci.c +15 -0
index b5f5410..5bcfffa 100644
@@ -233,6 +233,7 @@ pci_scan_one(const char *dirname, uint16_t domain, uint8_t bus,
 	dev->addr.bus = bus;
 	dev->addr.devid = devid;
 	dev->addr.function = function;
+	dev->config_fd = -1;
 
 	/* get vendor id */
 	snprintf(filename, sizeof(filename), "%s/vendor", dirname);
@@ -592,6 +593,20 @@ rte_eal_pci_probe_one_driver(struct rte_pci_driver *dr, struct rte_pci_device *d
 	return 1;
 }
 
+/* Read PCI config space. */
+int rte_eal_pci_read_config(const struct rte_pci_device *device,
+			    void *buf, size_t len, off_t offset)
+{
+	return pread(device->config_fd, buf, len, offset);
+}
+
+/* Write PCI config space. */
+int rte_eal_pci_write_config(const struct rte_pci_device *device,
+			     const void *buf, size_t len, off_t offset)
+{
+	return pwrite(device->config_fd, buf, len, offset);
+}
+
 /* Init the PCI EAL subsystem */
 int
 rte_eal_pci_init(void)
lib/librte_eal/linuxapp/eal/eal_pci_uio.c +10 -0
index e53f06b..0396eaa 100644
@@ -213,6 +213,7 @@ pci_get_uio_dev(struct rte_pci_device *dev, char *dstbuf,
 	struct dirent *e;
 	DIR *dir;
 	char dirname[PATH_MAX];
+	char filename[PATH_MAX];
 
 	/* depending on kernel version, uio can be located in uio/uioX
 	 * or uio:uioX */
@@ -268,6 +269,15 @@ pci_get_uio_dev(struct rte_pci_device *dev, char *dstbuf,
 	if (e == NULL)
 		return -1;
 
+	/* Open fd for access to PCI config */
+	snprintf(filename, sizeof(filename), "%s/device/config", dirname);
+	dev->config_fd = open(filename, O_RDWR);
+	if (dev->config_fd < 0) {
+		RTE_LOG(ERR, EAL, "%s(): cannot open %s: %s\n",
+			__func__, filename, strerror(errno));
+		return -1;
+	}
+
 	/* create uio device if we've been asked to */
 	if (internal_config.create_uio_dev &&
 			pci_mknod_uio_dev(dstbuf, uio_num) < 0)
lib/librte_eal/linuxapp/eal/rte_eal_version.map +2 -0
index d36286e..626689c 100644
@@ -40,6 +40,8 @@ DPDK_2.0 {
 	rte_eal_pci_probe;
 	rte_eal_pci_register;
 	rte_eal_pci_unregister;
+	rte_eal_pci_read_config;
+	rte_eal_pci_write_config;
 	rte_eal_process_type;
 	rte_eal_remote_launch;
 	rte_eal_tailq_lookup;
Signature · 1 line
2.1.4

A message that is not in the archive (sent off-list or not yet received): 1423247795-22399-2-git-send-email-stephen-OTpzqLSitTUnbdJkjeBofR2eb7JE58TQ@public.gmane.org

#4 lore raw

Re: [PATCH 1/4] pci: allow access to PCI config space

Hello Stephen,

- It looks a bit odd to me, we end up with something asymetric between uio / vfio wrt pci config space.
Can we an api consistent between the two ?
Does this mean that your pmd cannot work / has not been used with vfio ?

- Anyway, I suppose we could reuse this api to remove the RTE_PCI_CONFIG #ifdef / hardcoded stuff from linux eal / igb_uio.
Opinion ?

Signature · 4 lines
David Marchand


On Fri, Feb 6, 2015 at 7:36 PM, Stephen Hemminger <
Quoted 152 lines · [email protected]> wrote:

From: Stephen Hemminger <[email protected]>

Some drivers need ability to access PCI config (for example for power management). This adds an abstraction to do this; only implemented on Linux, but should be possible on BSD.

---
 lib/librte_eal/common/include/rte_pci.h         | 29
+++++++++++++++++++++++++
 lib/librte_eal/linuxapp/eal/eal_pci.c           | 15 +++++++++++++
 lib/librte_eal/linuxapp/eal/eal_pci_uio.c       | 10 +++++++++
 lib/librte_eal/linuxapp/eal/rte_eal_version.map |  2 ++
 4 files changed, 56 insertions(+)
lib/librte_eal/common/include/rte_pci.h +0 -0
index 66ed793..a78081f 100644
@@ -152,6 +152,7 @@ struct rte_pci_device {
        uint16_t max_vfs;                       /**< sriov enable if not

zero */
int numa_node; /**< NUMA node connection
*/
struct rte_devargs *devargs; /**< Device user arguments
*/

lib/librte_eal/common/include/rte_pci.h +29 -0
+       int config_fd;                          /**< PCI config access */
 };
 
 /** Any PCI device identifier (vendor, device, ...) */
@@ -298,6 +299,34 @@ void rte_eal_pci_register(struct rte_pci_driver
 *driver);
  */
 void rte_eal_pci_unregister(struct rte_pci_driver *driver);
 
+/**
+ * Read PCI config space.
+ *
+ * @param device
+ *   A pointer to a rte_pci_device structure describing the device
+ *   to use
+ * @param buf
+ *   A data buffer where the bytes should be read into
+ * @param size
+ *   The length of the data buffer.
+ */
+int rte_eal_pci_read_config(const struct rte_pci_device *device,
+                           void *buf, size_t len, off_t offset);
+
+/**
+ * Write PCI config space.
+ *
+ * @param device
+ *   A pointer to a rte_pci_device structure describing the device
+ *   to use
+ * @param buf
+ *   A data buffer containing the bytes should be written
+ * @param size
+ *   The length of the data buffer.
+ */
+int rte_eal_pci_write_config(const struct rte_pci_device *device,
+                            const void *buf, size_t len, off_t offset);
+
 #ifdef __cplusplus
 }

#endif

lib/librte_eal/linuxapp/eal/eal_pci.c +1 -0
index b5f5410..5bcfffa 100644
@@ -233,6 +233,7 @@ pci_scan_one(const char *dirname, uint16_t domain,
 uint8_t bus,
        dev->addr.bus = bus;
        dev->addr.devid = devid;
        dev->addr.function = function;
+       dev->config_fd = -1;
 
        /* get vendor id */
        snprintf(filename, sizeof(filename), "%s/vendor", dirname);
lib/librte_eal/linuxapp/eal/eal_pci.c +14 -0
@@ -592,6 +593,20 @@ rte_eal_pci_probe_one_driver(struct rte_pci_driver
 *dr, struct rte_pci_device *d
        return 1;
 }
 
+/* Read PCI config space. */
+int rte_eal_pci_read_config(const struct rte_pci_device *device,
+                           void *buf, size_t len, off_t offset)
+{
+       return pread(device->config_fd, buf, len, offset);
+}
+
+/* Write PCI config space. */
+int rte_eal_pci_write_config(const struct rte_pci_device *device,
+                            const void *buf, size_t len, off_t offset)
+{
+       return pwrite(device->config_fd, buf, len, offset);
+}
+
 /* Init the PCI EAL subsystem */
 int

rte_eal_pci_init(void)

lib/librte_eal/linuxapp/eal/eal_pci_uio.c +1 -0
index e53f06b..0396eaa 100644
@@ -213,6 +213,7 @@ pci_get_uio_dev(struct rte_pci_device *dev, char
 *dstbuf,
        struct dirent *e;
        DIR *dir;
        char dirname[PATH_MAX];
+       char filename[PATH_MAX];
 
        /* depending on kernel version, uio can be located in uio/uioX

* or uio:uioX */

lib/librte_eal/linuxapp/eal/eal_pci_uio.c +9 -0
@@ -268,6 +269,15 @@ pci_get_uio_dev(struct rte_pci_device *dev, char
 *dstbuf,
        if (e == NULL)
                return -1;
 
+       /* Open fd for access to PCI config */
+       snprintf(filename, sizeof(filename), "%s/device/config", dirname);
+       dev->config_fd = open(filename, O_RDWR);
+       if (dev->config_fd < 0) {
+               RTE_LOG(ERR, EAL, "%s(): cannot open %s: %s\n",
+                       __func__, filename, strerror(errno));
+               return -1;
+       }
+
        /* create uio device if we've been asked to */
        if (internal_config.create_uio_dev &&
                        pci_mknod_uio_dev(dstbuf, uio_num) < 0)
lib/librte_eal/linuxapp/eal/rte_eal_version.map +2 -1
index d36286e..626689c 100644
@@ -40,6 +40,8 @@ DPDK_2.0 {
        rte_eal_pci_probe;
        rte_eal_pci_register;
        rte_eal_pci_unregister;
+       rte_eal_pci_read_config;
+       rte_eal_pci_write_config;
        rte_eal_process_type;
        rte_eal_remote_launch;
        rte_eal_tailq_lookup;
--

2.1.4

A message that is not in the archive (sent off-list or not yet received): CALwxeUu7Ur-yXMcKu7Zkq16zqR6f5vU8DEX4Z8CET9G2ttjKgA-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

#6 lore raw

Re: [PATCH 1/4] pci: allow access to PCI config space

Quoted 6 lines · On Mon, 9 Feb 2015 10:45:01 +0100 David Marchand…

Hello Stephen,

- It looks a bit odd to me, we end up with something asymetric between uio / vfio wrt pci config space. Can we an api consistent between the two ?
Does this mean that your pmd cannot work / has not been used with vfio ?

Quoted 6 lines

- Anyway, I suppose we could reuse this api to remove the RTE_PCI_CONFIG #ifdef / hardcoded stuff from linux eal / igb_uio. Opinion ?

--
David Marchand

Ok, please send patches to VFIO to use this API.
It is not possible to do VFIO style config access through older UIO. Therefore a wrapper like this is needed.

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): c058fdefd5564f3f96bd21593c9ea19a-sI283RzA5cMA50eAn2D9H4CDckiq/[email protected]

#9 lore raw

Re: [PATCH 1/4] pci: allow access to PCI config space

Here is a revised version that works for both UIO and VFIO. The config access is property of device not the I/O model.

From: Stephen Hemminger <[email protected]>
Subject: [PATCH 1/4] pci: allow access to PCI config space

Some drivers need ability to access PCI config (for example for power management). This adds an abstraction to do this; only implemented on Linux, but should be possible on BSD.

---
 lib/librte_eal/common/include/rte_pci.h   | 29 +++++++++++++++++++++++++++++
 lib/librte_eal/linuxapp/eal/eal_pci.c     | 15 +++++++++++++++
 lib/librte_eal/linuxapp/eal/eal_pci_uio.c | 10 ++++++++++
 3 files changed, 54 insertions(+)
lib/librte_eal/common/include/rte_pci.h +29 -0
Index: dpdk/lib/librte_eal/common/include/rte_pci.h
===================================================================
@@ -152,6 +152,7 @@
 	uint16_t max_vfs;                       /**< sriov enable if not zero */
 	int numa_node;                          /**< NUMA node connection */
 	struct rte_devargs *devargs;            /**< Device user arguments */
+	int config_fd;				/**< PCI config access */
 };
 
 /** Any PCI device identifier (vendor, device, ...) */
@@ -304,6 +305,34 @@
  */
 void rte_eal_pci_unregister(struct rte_pci_driver *driver);
 
+/**
+ * Read PCI config space.
+ *
+ * @param device
+ *   A pointer to a rte_pci_device structure describing the device
+ *   to use
+ * @param buf
+ *   A data buffer where the bytes should be read into
+ * @param size
+ *   The length of the data buffer.
+ */
+int rte_eal_pci_read_config(const struct rte_pci_device *device,
+			    void *buf, size_t len, off_t offset);
+
+/**
+ * Write PCI config space.
+ *
+ * @param device
+ *   A pointer to a rte_pci_device structure describing the device
+ *   to use
+ * @param buf
+ *   A data buffer containing the bytes should be written
+ * @param size
+ *   The length of the data buffer.
+ */
+int rte_eal_pci_write_config(const struct rte_pci_device *device,
+			     const void *buf, size_t len, off_t offset);
+
 #ifdef __cplusplus
 }
 #endif
lib/librte_eal/linuxapp/eal/eal_pci.c +38 -1
Index: dpdk/lib/librte_eal/linuxapp/eal/eal_pci.c
===================================================================
@@ -34,6 +34,7 @@
 #include <string.h>
 #include <dirent.h>
 #include <sys/mman.h>
+#include <sys/fcntl.h>
 
 #include <rte_log.h>
 #include <rte_pci.h>
@@ -233,6 +234,7 @@
 	dev->addr.bus = bus;
 	dev->addr.devid = devid;
 	dev->addr.function = function;
+	dev->config_fd = -1;
 
 	/* get vendor id */
 	snprintf(filename, sizeof(filename), "%s/vendor", dirname);
@@ -498,6 +500,26 @@
 #endif
 
 static int
+pci_config_open(struct rte_pci_device *dev)
+{
+	struct rte_pci_addr *loc = &dev->addr;
+	char filename[PATH_MAX];
+
+	/* Open fd for access to PCI config */
+	snprintf(filename, sizeof(filename),
+		 SYSFS_PCI_DEVICES "/" PCI_PRI_FMT "/config",
+		 loc->domain, loc->bus, loc->devid, loc->function);
+	dev->config_fd = open(filename, O_RDWR);
+	if (dev->config_fd < 0) {
+		RTE_LOG(ERR, EAL, "%s(): cannot open %s: %s\n",
+			__func__, filename, strerror(errno));
+		return -1;
+	}
+
+	return 0;
+}
+
+static int
 pci_map_device(struct rte_pci_device *dev)
 {
 	int ret, mapped = 0;
@@ -518,7 +540,8 @@
 		if (ret != 0)
 			return ret;
 	}
-	return 0;
+
+	return pci_config_open(dev);
 }
 
 /*
@@ -595,6 +618,20 @@
 	return 1;
 }
 
+/* Read PCI config space. */
+int rte_eal_pci_read_config(const struct rte_pci_device *device,
+			    void *buf, size_t len, off_t offset)
+{
+	return pread(device->config_fd, buf, len, offset);
+}
+
+/* Write PCI config space. */
+int rte_eal_pci_write_config(const struct rte_pci_device *device,
+			     const void *buf, size_t len, off_t offset)
+{
+	return pwrite(device->config_fd, buf, len, offset);
+}
+
 /* Init the PCI EAL subsystem */
 int
 rte_eal_pci_init(void)
#10 lore raw

PATCH 2/4 bcm: add BCM pci device ids

In reply to a message not in the archive

From: Stephen Hemminger <[email protected]>

Add PCI id's to enable BCM poll mode driver.

---
 lib/librte_eal/common/include/rte_pci_dev_ids.h | 30 +++++++++++++++++++++++++
 1 file changed, 30 insertions(+)
lib/librte_eal/common/include/rte_pci_dev_ids.h +30 -0
index c922de9..b53fd93 100644
@@ -132,6 +132,14 @@
 #define RTE_PCI_DEV_ID_DECL_VMXNET3(vend, dev)
 #endif
 
+#ifndef RTE_PCI_DEV_ID_DECL_BCM
+#define RTE_PCI_DEV_ID_DECL_BCM(vend, dev)
+#endif
+
+#ifndef RTE_PCI_DEV_ID_DECL_BCMVF
+#define RTE_PCI_DEV_ID_DECL_BCMVF(vend, dev)
+#endif
+
 #ifndef PCI_VENDOR_ID_INTEL
 /** Vendor ID used by Intel devices */
 #define PCI_VENDOR_ID_INTEL 0x8086
@@ -147,6 +155,11 @@
 #define PCI_VENDOR_ID_VMWARE 0x15AD
 #endif
 
+#ifndef PCI_VENDOR_ID_BROADCOM
+/** Vendor ID used by Broadcom devices */
+#define PCI_VENDOR_ID_BROADCOM 0x14E4
+#endif
+
 /******************** Physical EM devices from e1000_hw.h ********************/
 
 #define E1000_DEV_ID_82542                    0x1000
@@ -526,6 +539,21 @@ RTE_PCI_DEV_ID_DECL_VIRTIO(PCI_VENDOR_ID_QUMRANET, QUMRANET_DEV_ID_VIRTIO)
 
 RTE_PCI_DEV_ID_DECL_VMXNET3(PCI_VENDOR_ID_VMWARE, VMWARE_DEV_ID_VMXNET3)
 
+/****************** Broadcom devices ******************/
+
+/* Broadcom */
+#define BROADCOM_DEV_ID_57711		    0x164F
+#define BROADCOM_DEV_ID_57810		    0x168E
+#define BROADCOM_DEV_ID_57810_MF	    0x16AE
+#define BROADCOM_DEV_ID_57810_VF	    0x16AF
+
+RTE_PCI_DEV_ID_DECL_BCM(PCI_VENDOR_ID_BROADCOM, BROADCOM_DEV_ID_57711)
+RTE_PCI_DEV_ID_DECL_BCM(PCI_VENDOR_ID_BROADCOM, BROADCOM_DEV_ID_57810)
+#ifdef RTE_LIBRTE_BCM_MF_SUPPORT
+RTE_PCI_DEV_ID_DECL_BCM(PCI_VENDOR_ID_BROADCOM, BROADCOM_DEV_ID_57810_MF)
+#endif
+RTE_PCI_DEV_ID_DECL_BCMVF(PCI_VENDOR_ID_BROADCOM, BROADCOM_DEV_ID_57810_VF)
+
 /*
  * Undef all RTE_PCI_DEV_ID_DECL_* here.
  */
@@ -538,3 +566,5 @@ RTE_PCI_DEV_ID_DECL_VMXNET3(PCI_VENDOR_ID_VMWARE, VMWARE_DEV_ID_VMXNET3)
 #undef RTE_PCI_DEV_ID_DECL_I40EVF
 #undef RTE_PCI_DEV_ID_DECL_VIRTIO
 #undef RTE_PCI_DEV_ID_DECL_VMXNET3
+#undef RTE_PCI_DEV_ID_DECL_BCM
+#undef RTE_PCI_DEV_ID_DECL_BCMVF
Signature · 1 line
2.1.4
#11 lore raw

PATCH 3/4 bcm: new poll mode driver

In reply to a message not in the archive

From: Stephen Hemminger <[email protected]>

Add driver for the Broadcom NetXtremeII 10 gigabit devices.

---
 lib/Makefile                        |     1 +
 lib/librte_pmd_bcm/Makefile         |    28 +
 lib/librte_pmd_bcm/bcm.c            | 11817 ++++++++++++++++++++++++++++++
 lib/librte_pmd_bcm/bcm.h            |  1998 +++++
 lib/librte_pmd_bcm/bcm_ethdev.c     |   544 ++
 lib/librte_pmd_bcm/bcm_ethdev.h     |    79 +
 lib/librte_pmd_bcm/bcm_logs.h       |    51 +
 lib/librte_pmd_bcm/bcm_rxtx.c       |   487 ++
 lib/librte_pmd_bcm/bcm_rxtx.h       |    85 +
 lib/librte_pmd_bcm/bcm_stats.c      |  1619 ++++
 lib/librte_pmd_bcm/bcm_stats.h      |   633 ++
 lib/librte_pmd_bcm/bcm_vfpf.c       |   597 ++
 lib/librte_pmd_bcm/bcm_vfpf.h       |   315 +
 lib/librte_pmd_bcm/debug.c          |   113 +
 lib/librte_pmd_bcm/ecore_fw_defs.h  |   423 ++
 lib/librte_pmd_bcm/ecore_hsi.h      |  6349 ++++++++++++++++
 lib/librte_pmd_bcm/ecore_init.h     |   842 +++
 lib/librte_pmd_bcm/ecore_init_ops.h |   886 +++
 lib/librte_pmd_bcm/ecore_mfw_req.h  |   207 +
 lib/librte_pmd_bcm/ecore_reg.h      |  3664 ++++++++++
 lib/librte_pmd_bcm/ecore_sp.c       |  5455 ++++++++++++++
 lib/librte_pmd_bcm/ecore_sp.h       |  1796 +++++
 lib/librte_pmd_bcm/elink.c          | 13378 ++++++++++++++++++++++++++++++++++
 lib/librte_pmd_bcm/elink.h          |   610 ++
 24 files changed, 51977 insertions(+)
 create mode 100644 lib/librte_pmd_bcm/Makefile
 create mode 100644 lib/librte_pmd_bcm/bcm.c
 create mode 100644 lib/librte_pmd_bcm/bcm.h
 create mode 100644 lib/librte_pmd_bcm/bcm_ethdev.c
 create mode 100644 lib/librte_pmd_bcm/bcm_ethdev.h
 create mode 100644 lib/librte_pmd_bcm/bcm_logs.h
 create mode 100644 lib/librte_pmd_bcm/bcm_rxtx.c
 create mode 100644 lib/librte_pmd_bcm/bcm_rxtx.h
 create mode 100644 lib/librte_pmd_bcm/bcm_stats.c
 create mode 100644 lib/librte_pmd_bcm/bcm_stats.h
 create mode 100644 lib/librte_pmd_bcm/bcm_vfpf.c
 create mode 100644 lib/librte_pmd_bcm/bcm_vfpf.h
 create mode 100644 lib/librte_pmd_bcm/debug.c
 create mode 100644 lib/librte_pmd_bcm/ecore_fw_defs.h
 create mode 100644 lib/librte_pmd_bcm/ecore_hsi.h
 create mode 100644 lib/librte_pmd_bcm/ecore_init.h
 create mode 100644 lib/librte_pmd_bcm/ecore_init_ops.h
 create mode 100644 lib/librte_pmd_bcm/ecore_mfw_req.h
 create mode 100644 lib/librte_pmd_bcm/ecore_reg.h
 create mode 100644 lib/librte_pmd_bcm/ecore_sp.c
 create mode 100644 lib/librte_pmd_bcm/ecore_sp.h
 create mode 100644 lib/librte_pmd_bcm/elink.c
 create mode 100644 lib/librte_pmd_bcm/elink.h
lib/Makefile +1 -0
index d617d81..023fc9c 100644
@@ -45,6 +45,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_E1000_PMD) += librte_pmd_e1000
 DIRS-$(CONFIG_RTE_LIBRTE_IXGBE_PMD) += librte_pmd_ixgbe
 DIRS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += librte_pmd_i40e
 DIRS-$(CONFIG_RTE_LIBRTE_ENIC_PMD) += librte_pmd_enic
+DIRS-$(CONFIG_RTE_LIBRTE_BCM_PMD) += librte_pmd_bcm
 DIRS-$(CONFIG_RTE_LIBRTE_PMD_BOND) += librte_pmd_bond
 DIRS-$(CONFIG_RTE_LIBRTE_PMD_RING) += librte_pmd_ring
 DIRS-$(CONFIG_RTE_LIBRTE_PMD_PCAP) += librte_pmd_pcap
lib/librte_pmd_bcm/Makefile new file +28 -0
new file mode 100644
index 0000000..2f3df65
@@ -0,0 +1,28 @@
+include $(RTE_SDK)/mk/rte.vars.mk
+
+#
+# library name
+#
+LIB = librte_pmd_bcm.a
+
+CFLAGS += -O3 -g
+CFLAGS += $(WERROR_FLAGS)
+CFLAGS += -DZLIB_CONST
+
+#
+# all source are stored in SRCS-y
+#
+SRCS-$(CONFIG_RTE_LIBRTE_BCM_PMD) += bcm.c
+SRCS-$(CONFIG_RTE_LIBRTE_BCM_PMD) += bcm_rxtx.c
+SRCS-$(CONFIG_RTE_LIBRTE_BCM_PMD) += bcm_stats.c
+SRCS-$(CONFIG_RTE_LIBRTE_BCM_PMD) += bcm_ethdev.c
+SRCS-$(CONFIG_RTE_LIBRTE_BCM_PMD) += ecore_sp.c
+SRCS-$(CONFIG_RTE_LIBRTE_BCM_PMD) += elink.c
+SRCS-$(CONFIG_RTE_LIBRTE_BCM_PMD) += bcm_vfpf.c
+SRCS-$(CONFIG_RTE_LIBRTE_BCM_DEBUG) += debug.c
+
+# this lib depends upon:
+DEPDIRS-$(CONFIG_RTE_LIBRTE_BCM_PMD) += lib/librte_eal lib/librte_ether lib/librte_hash
+DEPDIRS-$(CONFIG_RTE_LIBRTE_BCM_PMD) += lib/librte_mempool lib/librte_mbuf
+
+include $(RTE_SDK)/mk/rte.lib.mk
lib/librte_pmd_bcm/bcm.c new file +11817 -0

2,965 lines not on this page · raw message

lib/librte_pmd_bcm/bcm.h new file +1998 -0

1,998 lines not on this page · raw message

lib/librte_pmd_bcm/bcm_ethdev.c new file +544 -0

544 lines not on this page · raw message

lib/librte_pmd_bcm/bcm_ethdev.h new file +79 -0

79 lines not on this page · raw message

lib/librte_pmd_bcm/bcm_logs.h new file +51 -0

51 lines not on this page · raw message

lib/librte_pmd_bcm/bcm_rxtx.c new file +487 -0

487 lines not on this page · raw message

lib/librte_pmd_bcm/bcm_rxtx.h new file +85 -0

85 lines not on this page · raw message

lib/librte_pmd_bcm/bcm_stats.c new file +1619 -0

1,619 lines not on this page · raw message

lib/librte_pmd_bcm/bcm_stats.h new file +633 -0

633 lines not on this page · raw message

lib/librte_pmd_bcm/bcm_vfpf.c new file +597 -0

597 lines not on this page · raw message

lib/librte_pmd_bcm/bcm_vfpf.h new file +315 -0

315 lines not on this page · raw message

lib/librte_pmd_bcm/debug.c new file +113 -0

113 lines not on this page · raw message

lib/librte_pmd_bcm/ecore_fw_defs.h new file +423 -0

423 lines not on this page · raw message

lib/librte_pmd_bcm/ecore_hsi.h new file +6349 -0

6,349 lines not on this page · raw message

lib/librte_pmd_bcm/ecore_init.h new file +842 -0

842 lines not on this page · raw message

lib/librte_pmd_bcm/ecore_init_ops.h new file +886 -0

886 lines not on this page · raw message

lib/librte_pmd_bcm/ecore_mfw_req.h new file +207 -0

207 lines not on this page · raw message

lib/librte_pmd_bcm/ecore_reg.h new file +3664 -0

3,664 lines not on this page · raw message

lib/librte_pmd_bcm/ecore_sp.c new file +5455 -0

5,455 lines not on this page · raw message

lib/librte_pmd_bcm/ecore_sp.h new file +1796 -0

1,796 lines not on this page · raw message

lib/librte_pmd_bcm/elink.c new file +13378 -0

13,378 lines not on this page · raw message

lib/librte_pmd_bcm/elink.h new file +610 -0

610 lines not on this page · raw message

The rest of this diff is not shown. View the raw message.

Signature · 1 line
2.1.4

Part of the diff is not shown here. View the raw message.

#12 lore raw

PATCH 4/4 bcm: enable BCM poll mode driver in config

In reply to a message not in the archive

From: Stephen Hemminger <[email protected]>

Enable BCM poll mode driver in default config.

---
 config/common_linuxapp | 10 ++++++++++
 1 file changed, 10 insertions(+)
config/common_linuxapp +10 -0
index d428f84..86b157d 100644
@@ -185,6 +185,16 @@ CONFIG_RTE_LIBRTE_I40E_ITR_INTERVAL=-1
 CONFIG_RTE_LIBRTE_ENIC_PMD=y
 
 #
+# Compile burst-oriented Broadcom PMD driver
+#
+CONFIG_RTE_LIBRTE_BCM_PMD=y
+CONFIG_RTE_LIBRTE_BCM_DEBUG_INIT=n
+CONFIG_RTE_LIBRTE_BCM_DEBUG=n
+CONFIG_RTE_LIBRTE_BCM_MF_SUPPORT=n
+CONFIG_RTE_LIBRTE_BCM_DEBUG_TX=n
+CONFIG_RTE_LIBRTE_BCM_DEBUG_RX=n
+
+#
 # Compile burst-oriented VIRTIO PMD driver
 #
 CONFIG_RTE_LIBRTE_VIRTIO_PMD=y
Signature · 1 line
2.1.4
#13 lore raw
In reply to a message not in the archive

Hi Stephen,

Thanks for your great work on this!

Could you elaborate on a few things:
- What's the methodology used in the driver porting?
e.g. what's the base source you ported from?
what's the major change against the base source?
- What kind of tests have been done on Linux? any known limitations?

Thanks,
Jun

Quoted 72 lines · Stephen Hemminger

From: Stephen Hemminger <[email protected]>

These are the patches to enable supporting the Broadcom NetExtreme II 10G devices (show up as bnx2x on Linux).

The driver has only been tested on Linux, there maybe issues with firmware loading and PCI config access on BSD.

Stephen Hemminger (4):
pci: allow access to PCI config space
bcm: add BCM pci device ids
bcm: new poll mode driver
bcm: enable BCM poll mode driver in config

 config/common_linuxapp                          |    10 +
 lib/Makefile                                    |     1 +
 lib/librte_eal/common/include/rte_pci.h         |    29 +
 lib/librte_eal/common/include/rte_pci_dev_ids.h |    30 +
 lib/librte_eal/linuxapp/eal/eal_pci.c           |    15 +
 lib/librte_eal/linuxapp/eal/eal_pci_uio.c       |    10 +
 lib/librte_eal/linuxapp/eal/rte_eal_version.map |     2 +
 lib/librte_pmd_bcm/Makefile                     |    28 +
 lib/librte_pmd_bcm/bcm.c                        | 11817 +++++++++++++++++++
 lib/librte_pmd_bcm/bcm.h                        |  1998 ++++
 lib/librte_pmd_bcm/bcm_ethdev.c                 |   544 +
 lib/librte_pmd_bcm/bcm_ethdev.h                 |    79 +
 lib/librte_pmd_bcm/bcm_logs.h                   |    51 +
 lib/librte_pmd_bcm/bcm_rxtx.c                   |   487 +
 lib/librte_pmd_bcm/bcm_rxtx.h                   |    85 +
 lib/librte_pmd_bcm/bcm_stats.c                  |  1619 +++
 lib/librte_pmd_bcm/bcm_stats.h                  |   633 +
 lib/librte_pmd_bcm/bcm_vfpf.c                   |   597 +
 lib/librte_pmd_bcm/bcm_vfpf.h                   |   315 +
 lib/librte_pmd_bcm/debug.c                      |   113 +
 lib/librte_pmd_bcm/ecore_fw_defs.h              |   423 +
 lib/librte_pmd_bcm/ecore_hsi.h                  |  6349 ++++++++++
 lib/librte_pmd_bcm/ecore_init.h                 |   842 ++
 lib/librte_pmd_bcm/ecore_init_ops.h             |   886 ++
 lib/librte_pmd_bcm/ecore_mfw_req.h              |   207 +
 lib/librte_pmd_bcm/ecore_reg.h                  |  3664 ++++++
 lib/librte_pmd_bcm/ecore_sp.c                   |  5455 +++++++++
 lib/librte_pmd_bcm/ecore_sp.h                   |  1796 +++
 lib/librte_pmd_bcm/elink.c                      | 13378 ++++++++++++++++++++++
 lib/librte_pmd_bcm/elink.h                      |   610 +
 30 files changed, 52073 insertions(+)
 create mode 100644 lib/librte_pmd_bcm/Makefile
 create mode 100644 lib/librte_pmd_bcm/bcm.c
 create mode 100644 lib/librte_pmd_bcm/bcm.h
 create mode 100644 lib/librte_pmd_bcm/bcm_ethdev.c
 create mode 100644 lib/librte_pmd_bcm/bcm_ethdev.h
 create mode 100644 lib/librte_pmd_bcm/bcm_logs.h
 create mode 100644 lib/librte_pmd_bcm/bcm_rxtx.c
 create mode 100644 lib/librte_pmd_bcm/bcm_rxtx.h
 create mode 100644 lib/librte_pmd_bcm/bcm_stats.c
 create mode 100644 lib/librte_pmd_bcm/bcm_stats.h
 create mode 100644 lib/librte_pmd_bcm/bcm_vfpf.c
 create mode 100644 lib/librte_pmd_bcm/bcm_vfpf.h
 create mode 100644 lib/librte_pmd_bcm/debug.c
 create mode 100644 lib/librte_pmd_bcm/ecore_fw_defs.h
 create mode 100644 lib/librte_pmd_bcm/ecore_hsi.h
 create mode 100644 lib/librte_pmd_bcm/ecore_init.h
 create mode 100644 lib/librte_pmd_bcm/ecore_init_ops.h
 create mode 100644 lib/librte_pmd_bcm/ecore_mfw_req.h
 create mode 100644 lib/librte_pmd_bcm/ecore_reg.h
 create mode 100644 lib/librte_pmd_bcm/ecore_sp.c
 create mode 100644 lib/librte_pmd_bcm/ecore_sp.h
 create mode 100644 lib/librte_pmd_bcm/elink.c
 create mode 100644 lib/librte_pmd_bcm/elink.h

--
2.1.4

A message that is not in the archive (sent off-list or not yet received): CACyt=TG8EgADhVpiWN4cu56z904hBFH4LyUJAPvgrWhddqemBQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

#15 lore raw

Source was from FreeBSD.

changes were to make it work and lots of de-uglification.

For example, the last change was to remove custom logging and debug macros.

There are still lots of style issues with the driver, because of the amount of
useless macro wrapping.

I don't believe in the vendor model of taking one driver with lots of macros and pretending
it is generic across OS's.

It has been tested by our QA group as part of our vRouter product release, which is
based on DPDK 1.6. There are no open problem reports. The performance is less
than IXGBE but that is probably because it a a port an not optimized.

Quoted 90 lines · Jun Xiao

Hi Stephen,

Thanks for your great work on this!

Could you elaborate on a few things:
- What's the methodology used in the driver porting?
e.g. what's the base source you ported from?
what's the major change against the base source?
- What kind of tests have been done on Linux? any known limitations?

Thanks,
Jun

Quoted 14 lines · Stephen Hemminger

From: Stephen Hemminger <[email protected]>

These are the patches to enable supporting the Broadcom NetExtreme II 10G devices (show up as bnx2x on Linux).

The driver has only been tested on Linux, there maybe issues with firmware loading and PCI config access on BSD.

Stephen Hemminger (4):
pci: allow access to PCI config space
bcm: add BCM pci device ids
bcm: new poll mode driver
bcm: enable BCM poll mode driver in config

Quoted 9 lines
 config/common_linuxapp                          |    10 +
 lib/Makefile                                    |     1 +
 lib/librte_eal/common/include/rte_pci.h         |    29 +
 lib/librte_eal/common/include/rte_pci_dev_ids.h |    30 +
 lib/librte_eal/linuxapp/eal/eal_pci.c           |    15 +
 lib/librte_eal/linuxapp/eal/eal_pci_uio.c       |    10 +
 lib/librte_eal/linuxapp/eal/rte_eal_version.map |     2 +
 lib/librte_pmd_bcm/Makefile                     |    28 +
 lib/librte_pmd_bcm/bcm.c                        | 11817

+++++++++++++++++++

Quoted 12 lines
 lib/librte_pmd_bcm/bcm.h                        |  1998 ++++
 lib/librte_pmd_bcm/bcm_ethdev.c                 |   544 +
 lib/librte_pmd_bcm/bcm_ethdev.h                 |    79 +
 lib/librte_pmd_bcm/bcm_logs.h                   |    51 +
 lib/librte_pmd_bcm/bcm_rxtx.c                   |   487 +
 lib/librte_pmd_bcm/bcm_rxtx.h                   |    85 +
 lib/librte_pmd_bcm/bcm_stats.c                  |  1619 +++
 lib/librte_pmd_bcm/bcm_stats.h                  |   633 +
 lib/librte_pmd_bcm/bcm_vfpf.c                   |   597 +
 lib/librte_pmd_bcm/bcm_vfpf.h                   |   315 +
 lib/librte_pmd_bcm/debug.c                      |   113 +
 lib/librte_pmd_bcm/ecore_fw_defs.h              |   423 +
Quoted 8 lines
 lib/librte_pmd_bcm/ecore_hsi.h                  |  6349 ++++++++++
 lib/librte_pmd_bcm/ecore_init.h                 |   842 ++
 lib/librte_pmd_bcm/ecore_init_ops.h             |   886 ++
 lib/librte_pmd_bcm/ecore_mfw_req.h              |   207 +
 lib/librte_pmd_bcm/ecore_reg.h                  |  3664 ++++++
 lib/librte_pmd_bcm/ecore_sp.c                   |  5455 +++++++++
 lib/librte_pmd_bcm/ecore_sp.h                   |  1796 +++
 lib/librte_pmd_bcm/elink.c                      | 13378

++++++++++++++++++++++

Quoted 29 lines
 lib/librte_pmd_bcm/elink.h                      |   610 +
 30 files changed, 52073 insertions(+)
 create mode 100644 lib/librte_pmd_bcm/Makefile
 create mode 100644 lib/librte_pmd_bcm/bcm.c
 create mode 100644 lib/librte_pmd_bcm/bcm.h
 create mode 100644 lib/librte_pmd_bcm/bcm_ethdev.c
 create mode 100644 lib/librte_pmd_bcm/bcm_ethdev.h
 create mode 100644 lib/librte_pmd_bcm/bcm_logs.h
 create mode 100644 lib/librte_pmd_bcm/bcm_rxtx.c
 create mode 100644 lib/librte_pmd_bcm/bcm_rxtx.h
 create mode 100644 lib/librte_pmd_bcm/bcm_stats.c
 create mode 100644 lib/librte_pmd_bcm/bcm_stats.h
 create mode 100644 lib/librte_pmd_bcm/bcm_vfpf.c
 create mode 100644 lib/librte_pmd_bcm/bcm_vfpf.h
 create mode 100644 lib/librte_pmd_bcm/debug.c
 create mode 100644 lib/librte_pmd_bcm/ecore_fw_defs.h
 create mode 100644 lib/librte_pmd_bcm/ecore_hsi.h
 create mode 100644 lib/librte_pmd_bcm/ecore_init.h
 create mode 100644 lib/librte_pmd_bcm/ecore_init_ops.h
 create mode 100644 lib/librte_pmd_bcm/ecore_mfw_req.h
 create mode 100644 lib/librte_pmd_bcm/ecore_reg.h
 create mode 100644 lib/librte_pmd_bcm/ecore_sp.c
 create mode 100644 lib/librte_pmd_bcm/ecore_sp.h
 create mode 100644 lib/librte_pmd_bcm/elink.c
 create mode 100644 lib/librte_pmd_bcm/elink.h

--
2.1.4

A message that is not in the archive (sent off-list or not yet received): CAOaVG17Yb0yYVxrpm_qo1w3mEyZXDzXN+LX2K8GAZ=7pc5Hv_w-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

#17 lore raw

Thanks Stephen for all the info!

Do you expect this pmd driver will be in sync with FreeBSD repo in the future, or it's going to be diverged? I am just asking from a maintenance stand point.

Quoted 115 lines · Stephen Hemminger

Source was from FreeBSD.

changes were to make it work and lots of de-uglification.

For example, the last change was to remove custom logging and debug macros.

There are still lots of style issues with the driver, because of the amount of
useless macro wrapping.

I don't believe in the vendor model of taking one driver with lots of macros and pretending
it is generic across OS's.

It has been tested by our QA group as part of our vRouter product release, which is
based on DPDK 1.6. There are no open problem reports. The performance is less
than IXGBE but that is probably because it a a port an not optimized.

Quoted 90 lines · Jun Xiao

Hi Stephen,

Thanks for your great work on this!

Could you elaborate on a few things:
- What's the methodology used in the driver porting?
e.g. what's the base source you ported from?
what's the major change against the base source?
- What kind of tests have been done on Linux? any known limitations?

Thanks,
Jun

Quoted 74 lines · Stephen Hemminger

From: Stephen Hemminger <[email protected]>

These are the patches to enable supporting the Broadcom NetExtreme II 10G devices (show up as bnx2x on Linux).

The driver has only been tested on Linux, there maybe issues with firmware loading and PCI config access on BSD.

Stephen Hemminger (4):
pci: allow access to PCI config space
bcm: add BCM pci device ids
bcm: new poll mode driver
bcm: enable BCM poll mode driver in config

 config/common_linuxapp                          |    10 +
 lib/Makefile                                    |     1 +
 lib/librte_eal/common/include/rte_pci.h         |    29 +
 lib/librte_eal/common/include/rte_pci_dev_ids.h |    30 +
 lib/librte_eal/linuxapp/eal/eal_pci.c           |    15 +
 lib/librte_eal/linuxapp/eal/eal_pci_uio.c       |    10 +
 lib/librte_eal/linuxapp/eal/rte_eal_version.map |     2 +
 lib/librte_pmd_bcm/Makefile                     |    28 +
 lib/librte_pmd_bcm/bcm.c                        | 11817
+++++++++++++++++++
 lib/librte_pmd_bcm/bcm.h                        |  1998 ++++
 lib/librte_pmd_bcm/bcm_ethdev.c                 |   544 +
 lib/librte_pmd_bcm/bcm_ethdev.h                 |    79 +
 lib/librte_pmd_bcm/bcm_logs.h                   |    51 +
 lib/librte_pmd_bcm/bcm_rxtx.c                   |   487 +
 lib/librte_pmd_bcm/bcm_rxtx.h                   |    85 +
 lib/librte_pmd_bcm/bcm_stats.c                  |  1619 +++
 lib/librte_pmd_bcm/bcm_stats.h                  |   633 +
 lib/librte_pmd_bcm/bcm_vfpf.c                   |   597 +
 lib/librte_pmd_bcm/bcm_vfpf.h                   |   315 +
 lib/librte_pmd_bcm/debug.c                      |   113 +
 lib/librte_pmd_bcm/ecore_fw_defs.h              |   423 +
 lib/librte_pmd_bcm/ecore_hsi.h                  |  6349 ++++++++++
 lib/librte_pmd_bcm/ecore_init.h                 |   842 ++
 lib/librte_pmd_bcm/ecore_init_ops.h             |   886 ++
 lib/librte_pmd_bcm/ecore_mfw_req.h              |   207 +
 lib/librte_pmd_bcm/ecore_reg.h                  |  3664 ++++++
 lib/librte_pmd_bcm/ecore_sp.c                   |  5455 +++++++++
 lib/librte_pmd_bcm/ecore_sp.h                   |  1796 +++
 lib/librte_pmd_bcm/elink.c                      | 13378
++++++++++++++++++++++
 lib/librte_pmd_bcm/elink.h                      |   610 +
 30 files changed, 52073 insertions(+)
 create mode 100644 lib/librte_pmd_bcm/Makefile
 create mode 100644 lib/librte_pmd_bcm/bcm.c
 create mode 100644 lib/librte_pmd_bcm/bcm.h
 create mode 100644 lib/librte_pmd_bcm/bcm_ethdev.c
 create mode 100644 lib/librte_pmd_bcm/bcm_ethdev.h
 create mode 100644 lib/librte_pmd_bcm/bcm_logs.h
 create mode 100644 lib/librte_pmd_bcm/bcm_rxtx.c
 create mode 100644 lib/librte_pmd_bcm/bcm_rxtx.h
 create mode 100644 lib/librte_pmd_bcm/bcm_stats.c
 create mode 100644 lib/librte_pmd_bcm/bcm_stats.h
 create mode 100644 lib/librte_pmd_bcm/bcm_vfpf.c
 create mode 100644 lib/librte_pmd_bcm/bcm_vfpf.h
 create mode 100644 lib/librte_pmd_bcm/debug.c
 create mode 100644 lib/librte_pmd_bcm/ecore_fw_defs.h
 create mode 100644 lib/librte_pmd_bcm/ecore_hsi.h
 create mode 100644 lib/librte_pmd_bcm/ecore_init.h
 create mode 100644 lib/librte_pmd_bcm/ecore_init_ops.h
 create mode 100644 lib/librte_pmd_bcm/ecore_mfw_req.h
 create mode 100644 lib/librte_pmd_bcm/ecore_reg.h
 create mode 100644 lib/librte_pmd_bcm/ecore_sp.c
 create mode 100644 lib/librte_pmd_bcm/ecore_sp.h
 create mode 100644 lib/librte_pmd_bcm/elink.c
 create mode 100644 lib/librte_pmd_bcm/elink.h

--
2.1.4

A message that is not in the archive (sent off-list or not yet received): CACyt=TFiM3m3GkDs8HhoYRat7GEf6TE4hJYn31rR-TU=+=N4EQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

#19 lore raw
Quoted 5 lines · On Sun, 8 Feb 2015 00:05:16 +0800 Jun Xiao…

Thanks Stephen for all the info!

Do you expect this pmd driver will be in sync with FreeBSD repo in the future, or it's going to be diverged? I am just asking from a maintenance stand point.

No. The driver is not going to be extended or updated by us. Bugfixes only.

#20 lore raw
In reply to a message not in the archive

Hi Stephen,

2015-02-07 07:15, Stephen Hemminger:

Quoted 3 lines

Source was from FreeBSD.

changes were to make it work and lots of de-uglification.

Maintenance would be eased by showing this work with separate commits. You should at least have a patch for importing FreeBSD source without modifications.

Other comment: there could be other drivers for Broadcom devices. Why not keeping the name bnx2x?

Thanks

#21 lore raw
Quoted 7 lines · On Thu, 12 Feb 2015 15:00:13 +0100 Thomas Monjalon…

Hi Stephen,

2015-02-07 07:15, Stephen Hemminger:

Quoted 3 lines

Source was from FreeBSD.

changes were to make it work and lots of de-uglification.

Quoted 8 lines

Maintenance would be eased by showing this work with separate commits. You should at least have a patch for importing FreeBSD source without modifications.

Other comment: there could be other drivers for Broadcom devices. Why not keeping the name bnx2x?

Thanks

That history is lost.

A message that is not in the archive (sent off-list or not yet received): 20150212063353.6fb8d6a6-CA4OZQ/[email protected]

#23 lore raw

2015-02-12 06:33, Stephen Hemminger:

Quoted 14 lines
Quoted 11 lines · On Thu, 12 Feb 2015 15:00:13 +0100 Thomas Monjalon…

Hi Stephen,

2015-02-07 07:15, Stephen Hemminger:

Quoted 3 lines

Source was from FreeBSD.

changes were to make it work and lots of de-uglification.

Maintenance would be eased by showing this work with separate commits. You should at least have a patch for importing FreeBSD source without modifications.

Quoted 6 lines
Quoted 4 lines

Other comment: there could be other drivers for Broadcom devices. Why not keeping the name bnx2x?

Thanks

That history is lost.

I don't understand your answer,
and you forgot one question.
Please reply below each question.

I said: "You should at least have a patch for importing FreeBSD source without modfications". Is the initial FreeBSD source lost?

Do you agree to rename to bnx2x?

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): 0a8c8ae93e354bf0a74dd8ade9d1cc70-sI283RzA5cMA50eAn2D9H4CDckiq/[email protected]

#26 lore raw
Quoted 22 lines · On Thu, 12 Feb 2015 14:56:06 +0000 Thomas Monjalon…

2015-02-12 06:33, Stephen Hemminger:

Quoted 14 lines
Quoted 11 lines · On Thu, 12 Feb 2015 15:00:13 +0100 Thomas Monjalon…

Hi Stephen,

2015-02-07 07:15, Stephen Hemminger:

Quoted 3 lines

Source was from FreeBSD.

changes were to make it work and lots of de-uglification.

Maintenance would be eased by showing this work with separate commits. You should at least have a patch for importing FreeBSD source without modifications.

Quoted 6 lines
Quoted 4 lines

Other comment: there could be other drivers for Broadcom devices. Why not keeping the name bnx2x?

Thanks

That history is lost.

Quoted 8 lines

I don't understand your answer,
and you forgot one question.
Please reply below each question.

I said: "You should at least have a patch for importing FreeBSD source without modfications". Is the initial FreeBSD source lost?

Do you agree to rename to bnx2x?

This was based of driver called bxe in FreeBsd but the changes are so massive that there is no point in creating such a patch except to keep some lawyer happy.

 37 files changed, 33668 insertions(+), 159047 deletions(-)

It might be okay to rename the visible config option and directory to bxe but not going to change internal things.

A message that is not in the archive (sent off-list or not yet received): 20150212164550.0a2b8e2a-CA4OZQ/[email protected]

#28 lore raw

2015-02-12 16:45, Stephen Hemminger:

Quoted 23 lines
Quoted 20 lines · On Thu, 12 Feb 2015 14:56:06 +0000 Thomas Monjalon…

2015-02-12 06:33, Stephen Hemminger:

Quoted 19 lines
Quoted 15 lines · On Thu, 12 Feb 2015 15:00:13 +0100 Thomas Monjalon…

Hi Stephen,

2015-02-07 07:15, Stephen Hemminger:

Quoted 3 lines

Source was from FreeBSD.

changes were to make it work and lots of de-uglification.

Maintenance would be eased by showing this work with separate commits. You should at least have a patch for importing FreeBSD source without modifications.

Other comment: there could be other drivers for Broadcom devices. Why not keeping the name bnx2x?

Thanks

Quoted 8 lines
Quoted 8 lines
Quoted 1 line

That history is lost.

I don't understand your answer,
and you forgot one question.
Please reply below each question.

I said: "You should at least have a patch for importing FreeBSD source without modfications". Is the initial FreeBSD source lost?

Without answer, I guess Brocade lost its initial import.

Quoted 10 lines
Quoted 1 line

Do you agree to rename to bnx2x?

This was based of driver called bxe in FreeBsd but the changes are so massive that there is no point in creating such a patch except to keep some lawyer happy.

 37 files changed, 33668 insertions(+), 159047 deletions(-)

It might be okay to rename the visible config option and directory to bxe but not going to change internal things.

Yes it's a minimal change.
By the way, you named it "bcm" but it has been acquired by QLogic:

	http://www.broadcom.com/press/release.php?id=s832628

So we should refer at it as "NetXtreme II".
I have no strong opinion about the name of the pmd (bnx2x, bxe, whatever), except it should not be bcm.

When making a v2, please don't forget to add a section in the MAINTAINERS file.

Thank you