[dpdk-dev] [PATCH v2 0/5] New library: rte_distributor
Tested-by: Waterman Cao This patch is composed of 5 files including cover letter. All files have been tested by Intel. We performed function and performance test on new Distributor patch. Total Cases Passed Failed 10 10 0 Please see example as the following: cd /root/dpdk/app/test make ./test -n 4 -c 0xC ./distributor_autotest Test Environment: Fedora 20 x86_64, Linux Kernel 3.13.6-200, GCC 4.8.2 Intel Xeon CPU E5-2680 v2 @ 2.80GHz,Intel Niantic 82599 Based off commit 41d9a8250dc8d35bcd4499ba8e59809be943e712.
[dpdk-dev] [PATCH v2 0/3] Support setting TX rate for queue and VF
Hi, Neil " but you're implementaiton requires that it be re-implemented for each PMD " [Changchun]: Different PMD(corresponding diff NIC) has different register to set. It makes sense to me has different implementation to support limit tx rate. " Why not just export max tx rates from the PMD and write a generic queuing libarary to do rate limitation for any PMD?" [Changchun]: Just export max tx rate is not enough for customer. I think if we can leverage HW feature to do it, why need a more complicated lib to do so? Thanks Changchun -Original Message- From: Neil Horman [mailto:nhor...@tuxdriver.com] Sent: Tuesday, May 27, 2014 6:53 AM To: Ouyang, Changchun Cc: dev at dpdk.org Subject: Re: [dpdk-dev] [PATCH v2 0/3] Support setting TX rate for queue and VF On Mon, May 26, 2014 at 03:45:28PM +0800, Ouyang Changchun wrote: > This patch v2 fixes some errors and warnings reported by checkpatch.pl. > > This patch series also contain the 3 items: > 1. Add API to support setting TX rate for a queue or a VF. > 2. Implement the functionality of setting TX rate for queue or VF in IXGBE > PMD. > 3. Add commands in testpmd to test the functionality of setting TX rate for > queue or VF. > > Ouyang Changchun (3): > Add API to support set TX rate for a queue and VF. > Implement the functionality of setting TX rate for queue or VF in > IXGBE PMD. > Add commands in testpmd to test the functionality of setting TX rate > for queue or VF. > > app/test-pmd/cmdline.c | 159 > +++- > app/test-pmd/config.c | 47 +++ > app/test-pmd/testpmd.h | 3 + > lib/librte_ether/rte_ethdev.c | 71 > lib/librte_ether/rte_ethdev.h | 51 > lib/librte_pmd_ixgbe/ixgbe_ethdev.c | 122 +++ > lib/librte_pmd_ixgbe/ixgbe_ethdev.h | 13 ++- > 7 files changed, 462 insertions(+), 4 deletions(-) > This seems a bit backwards. queue rate limiting is rather a generic function, that doesn't really need to know any details about the hardware, save for its maximum tx rate, but you're implementaiton requires that it be re-implemented for each PMD. Why not just export max tx rates from the PMD and write a generic queuing libarary to do rate limitation for any PMD? Neil > -- > 1.9.0 > >
[dpdk-dev] [PATCH v2 1/3] ether: Add API to support setting TX rate for queue and VF
Hi Thomas, As we can see below, There are already 4 existing functions for vf in the header file: rte_ethdev.h:int rte_eth_dev_set_vf_rxmode(uint8_t port, uint16_t vf, uint16_t rx_mode, rte_ethdev.h:rte_eth_dev_set_vf_tx(uint8_t port,uint16_t vf, uint8_t on); rte_ethdev.h:rte_eth_dev_set_vf_rx(uint8_t port,uint16_t vf, uint8_t on); rte_ethdev.h:rte_eth_dev_set_vf_vlan_filter(uint8_t port, uint16_t vlan_id, So do we have plan to move them or remove them as they are all for VF specifically? If no, why we can accept those functions, but not accept the rte_eth_set_vf_rate_limit? :-) I have 2 new api in this patch, the rte_eth_set_queue_rate_limit is more generic, whose Argument only have port and queue. but PRC customer has the requirement of API function to limit the vf tx rate, so personally I think rte_eth_set_vf_rate_limit is necessary for them. Thanks and regards, Changchun -Original Message- From: Thomas Monjalon [mailto:thomas.monja...@6wind.com] Sent: Wednesday, May 28, 2014 6:48 AM To: Ouyang, Changchun Cc: dev at dpdk.org Subject: Re: [dpdk-dev] [PATCH v2 1/3] ether: Add API to support setting TX rate for queue and VF Hi Changchun, 2014-05-26 15:45, Ouyang Changchun: > /** > + * Set the rate limitation for a queue on an Ethernet device. > + * > + * @param port_id > + * The port identifier of the Ethernet device. > + * @param queue_idx > + * The queue id. > + * @param tx_rate > + * The tx rate allocated from the total link speed for this queue. > + * @return > + * - (0) if successful. > + * - (-ENOTSUP) if hardware doesn't support this feature. > + * - (-ENODEV) if *port_id* invalid. > + * - (-EINVAL) if bad parameter. > + */ > +int rte_eth_set_queue_rate_limit(uint8_t port_id, uint16_t queue_idx, > + uint16_t tx_rate); > + > +/** > + * Set the rate limitation for a vf on an Ethernet device. > + * > + * @param port_id > + * The port identifier of the Ethernet device. > + * @param vf > + * VF id. > + * @param tx_rate > + * The tx rate allocated from the total link speed for this VF id. > + * @param q_msk > + * The queue mask which need to set the rate. > + * @return > + * - (0) if successful. > + * - (-ENOTSUP) if hardware doesn't support this feature. > + * - (-ENODEV) if *port_id* invalid. > + * - (-EINVAL) if bad parameter. > + */ > +int rte_eth_set_vf_rate_limit(uint8_t port_id, uint16_t vf, > + uint16_t tx_rate, uint64_t q_msk); You are defining an API function specifically for VF. It's not generic and shouldn't appear in the API. We now have to be careful about the API and try to build a robust generic API which could become stable. Is it possible to imagine another API where only port and queue parameters are required? Thanks -- Thomas
[dpdk-dev] [PATCH v2 03/27] pci: add macros and pci device IDs to support i40e
The pci device IDs of i40e have been added. The new macros have been added in rte_pci_dev_ids.h to support i40e. Signed-off-by: Helin Zhang Signed-off-by: Jing Chen Acked-by: Cunming Liang Acked-by: Jijiang Liu Acked-by: Jingjing Wu Tested-by: Waterman Cao --- lib/librte_eal/common/include/rte_pci_dev_ids.h | 44 + 1 file changed, 44 insertions(+) diff --git a/lib/librte_eal/common/include/rte_pci_dev_ids.h b/lib/librte_eal/common/include/rte_pci_dev_ids.h index a51c1ef..f89b0e8 100644 --- a/lib/librte_eal/common/include/rte_pci_dev_ids.h +++ b/lib/librte_eal/common/include/rte_pci_dev_ids.h @@ -73,6 +73,8 @@ * RTE_PCI_DEV_ID_DECL_IGBVF * RTE_PCI_DEV_ID_DECL_IXGBE * RTE_PCI_DEV_ID_DECL_IXGBEVF + * RTE_PCI_DEV_ID_DECL_I40E + * RTE_PCI_DEV_ID_DECL_I40EVF * RTE_PCI_DEV_ID_DECL_VIRTIO * at the time when this file is included. * @@ -114,6 +116,14 @@ #define RTE_PCI_DEV_ID_DECL_IXGBEVF(vend, dev) #endif +#ifndef RTE_PCI_DEV_ID_DECL_I40E +#define RTE_PCI_DEV_ID_DECL_I40E(vend, dev) +#endif + +#ifndef RTE_PCI_DEV_ID_DECL_I40EVF +#define RTE_PCI_DEV_ID_DECL_I40EVF(vend, dev) +#endif + #ifndef RTE_PCI_DEV_ID_DECL_VIRTIO #define RTE_PCI_DEV_ID_DECL_VIRTIO(vend, dev) #endif @@ -444,6 +454,30 @@ RTE_PCI_DEV_ID_DECL_IXGBE(PCI_VENDOR_ID_INTEL, IXGBE_DEV_ID_X540T1) RTE_PCI_DEV_ID_DECL_IXGBE(PCI_VENDOR_ID_INTEL, IXGBE_DEV_ID_82599_BYPASS) #endif +/*** Physical I40E devices from i40e_type.h */ + +#define I40E_DEV_ID_SFP_XL710 0x1572 +#define I40E_DEV_ID_SFP_X7100x1573 +#define I40E_DEV_ID_QEMU0x1574 +#define I40E_DEV_ID_KX_A0x157F +#define I40E_DEV_ID_KX_B0x1580 +#define I40E_DEV_ID_KX_C0x1581 +#define I40E_DEV_ID_KX_D0x1582 +#define I40E_DEV_ID_QSFP_A 0x1583 +#define I40E_DEV_ID_QSFP_B 0x1584 +#define I40E_DEV_ID_QSFP_C 0x1585 + +RTE_PCI_DEV_ID_DECL_I40E(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_SFP_XL710) +RTE_PCI_DEV_ID_DECL_I40E(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_SFP_X710) +RTE_PCI_DEV_ID_DECL_I40E(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_QEMU) +RTE_PCI_DEV_ID_DECL_I40E(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_KX_A) +RTE_PCI_DEV_ID_DECL_I40E(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_KX_B) +RTE_PCI_DEV_ID_DECL_I40E(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_KX_C) +RTE_PCI_DEV_ID_DECL_I40E(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_KX_D) +RTE_PCI_DEV_ID_DECL_I40E(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_QSFP_A) +RTE_PCI_DEV_ID_DECL_I40E(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_QSFP_B) +RTE_PCI_DEV_ID_DECL_I40E(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_QSFP_C) + /** Virtual IGB devices from e1000_hw.h **/ #define E1000_DEV_ID_82576_VF 0x10CA @@ -468,6 +502,14 @@ RTE_PCI_DEV_ID_DECL_IXGBEVF(PCI_VENDOR_ID_INTEL, IXGBE_DEV_ID_82599_VF_HV) RTE_PCI_DEV_ID_DECL_IXGBEVF(PCI_VENDOR_ID_INTEL, IXGBE_DEV_ID_X540_VF) RTE_PCI_DEV_ID_DECL_IXGBEVF(PCI_VENDOR_ID_INTEL, IXGBE_DEV_ID_X540_VF_HV) +/** Virtual I40E devices from i40e_type.h / + +#define I40E_DEV_ID_VF 0x154C +#define I40E_DEV_ID_VF_HV 0x1571 + +RTE_PCI_DEV_ID_DECL_I40EVF(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_VF) +RTE_PCI_DEV_ID_DECL_I40EVF(PCI_VENDOR_ID_INTEL, I40E_DEV_ID_VF_HV) + /** Virtio devices from virtio.h **/ #define QUMRANET_DEV_ID_VIRTIO 0x1000 @@ -488,5 +530,7 @@ RTE_PCI_DEV_ID_DECL_VMXNET3(PCI_VENDOR_ID_VMWARE, VMWARE_DEV_ID_VMXNET3) #undef RTE_PCI_DEV_ID_DECL_IGBVF #undef RTE_PCI_DEV_ID_DECL_IXGBE #undef RTE_PCI_DEV_ID_DECL_IXGBEVF +#undef RTE_PCI_DEV_ID_DECL_I40E +#undef RTE_PCI_DEV_ID_DECL_I40EVF #undef RTE_PCI_DEV_ID_DECL_VIRTIO #undef RTE_PCI_DEV_ID_DECL_VMXNET3 -- 1.8.1.4
[dpdk-dev] [PATCH v2 00/27] Add i40e PMD support
Summary: The 2nd version of series of patches are to add i40e PMD support. It contains the updated basic shared code, and some other enhancements. It adds the support of the latest version of firmware. * Add new PMD driver of i40e in the folder of librte_pmd_i40e * Add some neccessary definitions, changes in rte_mbuf.h and eth_dev * Add new configurations for i40e * Add or modifiy makefiles to support i40e compilation * Add neccessary changes in ixgbe, e1000 and vmxnet3 PMD, as hash flags has been enlarged from 16 bits to 64 bits to support i40e * Add neccessary changes in example applications and testpmd to use ETH_RSS_IP to replace all IP hash flags, as i40e introduced more hash flags. * Add command in testpmd for port based vlan insertion offload testing * Add neccessary changes in eth_dev to support configuring maximum packet length of less than 1518 * Add two sys files in igb_uio to support enabling/disabling 'Extended Tag' and resetting 'Max Read Request Size', as it has big impacts on i40e performance * Add neccessary changes in pci to read/write the above two sys files during probing PCI Features/enhancements to be implemented later: * Set link speed, and physically up/down * Double VLAN support, flow director, VMDq and DCB * VLAN insertion/stripping, RSS in VF Validation Summary: This patch includes 27 files, and has been tested by Intel(R). Please see test environment information as the following: * Fedora 20 x86_64 * Linux Kernel 3.11.10-301 * GCC 4.8.2 * Intel Xeon CPU E5-2680 v2 @ 2.80GHz We verified this patch with two NICs (i40e 4x10 SFP+, i40e 2x40G), and performed basic function test on PF and VF. All cases are passed. Total cases Passed Failed 138 138 0 Signed-off-by: Helin Zhang Signed-off-by: Jing Chen Acked-by: Cunming Liang Acked-by: Jijiang Liu Acked-by: Jingjing Wu Tested-by: Waterman Cao --- Helin Zhang (27): i40e: add basic shared code i40e: add PMD source files pci: add macros and pci device IDs to support i40e igb_uio: add i40e support mbuf: add new packet flags for i40e ethdev: add i40e support ethdev: support setting maximum packet length to less than 1518 vmxnet3: enlarge the hash flags of RSS to 64 bits ixgbe: enlarge the hash flags of RSS to 64 bits igb: enlarge the hash flags of RSS to 64 bits mk: add i40e support config: add configurations for i40e app/test-pmd: support displaying 32 bytes RX descriptors app/test-pmd: add command of 'tx_vlan set pvid port_id vlan_id (on|off)' app/testpmd: enlarge the hash flags of RSS to 64 bits app/test-pmd: add L3 packet type in offload flags examples/dpdk_qat: use ETH_RSS_IP to replace IP hash flags of RSS examples/ip_reassembly: use ETH_RSS_IP to replace IP hash flags of RSS examples/l3fwd-power: use ETH_RSS_IP to replace IP hash flags of RSS examples/l3fwd-vf: use ETH_RSS_IP to replace IP hash flags of RSS examples/l3fwd: use ETH_RSS_IP to replace IP hash flags of RSS examples/load_balancer: use ETH_RSS_IP to replace IP hash flags of RSS examples/multi_process: use ETH_RSS_IP to replace IP hash flags of RSS examples/qos_meter: use ETH_RSS_IP to replace IP hash flags of RSS igb_uio: add sys files to read/write specific bits in pci config space pci: support reading/writing sys files of 'extended_tag' and 'max_read_request_size' config: add configurations for enabling 'Extended Tag' or resetting 'Max Read Request Size' app/test-pmd/cmdline.c|65 +- app/test-pmd/config.c |45 +- app/test-pmd/csumonly.c | 2 + app/test-pmd/parameters.c | 5 +- app/test-pmd/testpmd.c| 2 +- app/test-pmd/testpmd.h| 4 +- config/common_bsdapp |20 + config/common_linuxapp|30 + examples/dpdk_qat/main.c | 2 +- examples/ip_reassembly/main.c | 2 +- examples/l3fwd-power/main.c | 2 +- examples/l3fwd-vf/main.c | 2 +- examples/l3fwd/main.c | 2 +- examples/load_balancer/init.c | 2 +- examples/multi_process/symmetric_mp/main.c| 2 +- examples/qos_meter/main.c | 2 +- lib/Makefile | 1 + lib/librte_eal/common/include/rte_pci_dev_ids.h |44 + lib/librte_eal/linuxapp/eal/eal_pci.c |99 + lib/librte_eal/linuxapp/igb_uio/igb_uio.c | 107 + lib/librte_ether/rte_ethdev.c |25 +- lib/librte_ether/rte_ethdev.h | 154 +- lib/librte_ether/rte_ether.h |24 + lib/librte_mbuf/rte_mbuf.h| 8 + lib/librte_pmd_e1000/igb_rxtx.c |
[dpdk-dev] [PATCH v2 05/27] mbuf: add new packet flags for i40e
New packet flags of both RX and TX have been added to support i40e.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
lib/librte_mbuf/rte_mbuf.h | 8
1 file changed, 8 insertions(+)
diff --git a/lib/librte_mbuf/rte_mbuf.h b/lib/librte_mbuf/rte_mbuf.h
index 4a9ab41..ded2ce6 100644
--- a/lib/librte_mbuf/rte_mbuf.h
+++ b/lib/librte_mbuf/rte_mbuf.h
@@ -85,6 +85,11 @@ struct rte_ctrlmbuf {
#define PKT_RX_FDIR 0x0004 /**< RX packet with FDIR infos. */
#define PKT_RX_L4_CKSUM_BAD 0x0008 /**< L4 cksum of RX pkt. is not OK. */
#define PKT_RX_IP_CKSUM_BAD 0x0010 /**< IP cksum of RX pkt. is not OK. */
+#define PKT_RX_EIP_CKSUM_BAD 0x /**< External IP header checksum error. */
+#define PKT_RX_OVERSIZE 0x /**< Num of desc of an RX pkt oversize. */
+#define PKT_RX_HBUF_OVERFLOW 0x /**< Header buffer overflow. */
+#define PKT_RX_RECIP_ERR 0x /**< Hardware processing error. */
+#define PKT_RX_MAC_ERR 0x /**< MAC error. */
#define PKT_RX_IPV4_HDR 0x0020 /**< RX packet with IPv4 header. */
#define PKT_RX_IPV4_HDR_EXT 0x0040 /**< RX packet with extended IPv4 header.
*/
#define PKT_RX_IPV6_HDR 0x0080 /**< RX packet with IPv6 header. */
@@ -94,6 +99,9 @@ struct rte_ctrlmbuf {
#define PKT_TX_VLAN_PKT 0x0800 /**< TX packet is a 802.1q VLAN packet. */
#define PKT_TX_IP_CKSUM 0x1000 /**< IP cksum of TX pkt. computed by NIC.
*/
+#define PKT_TX_IPV4_CSUM 0x1000 /**< Alias of PKT_TX_IP_CKSUM. */
+#define PKT_TX_IPV4 PKT_RX_IPV4_HDR /**< IPv4 with no IP checksum
offload. */
+#define PKT_TX_IPV6 PKT_RX_IPV6_HDR /**< IPv6 packet */
/*
* Bit 14~13 used for L4 packet type with checksum enabled.
* 00: Reserved
--
1.8.1.4
[dpdk-dev] [PATCH v2 06/27] ethdev: add i40e support
To support i40e,
- New link speeds have been defined.
- New ops has been added to support setting port based vlan insertion.
- i40e RSS flags have been added.
- A generic function of is_same_ether_addr() has been added to
compare two ethernet address.
- A new configuration of 'uint8_t rss_key_len' has been added in
'struct rte_eth_rss_conf' to support different length of RSS keys.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
lib/librte_ether/rte_ethdev.c | 15
lib/librte_ether/rte_ethdev.h | 154 +-
lib/librte_ether/rte_ether.h | 24 +++
3 files changed, 175 insertions(+), 18 deletions(-)
diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
index 11e877b..53cb7a9 100644
--- a/lib/librte_ether/rte_ethdev.c
+++ b/lib/librte_ether/rte_ethdev.c
@@ -1298,6 +1298,21 @@ rte_eth_dev_get_vlan_offload(uint8_t port_id)
return ret;
}
+int
+rte_eth_dev_set_vlan_pvid(uint8_t port_id, uint16_t pvid, int on)
+{
+ struct rte_eth_dev *dev;
+
+ if (port_id >= nb_ports) {
+ PMD_DEBUG_TRACE("Invalid port_id=%d\n", port_id);
+ return (-ENODEV);
+ }
+ dev = &rte_eth_devices[port_id];
+ FUNC_PTR_OR_ERR_RET(*dev->dev_ops->vlan_pvid_set, -ENOTSUP);
+ (*dev->dev_ops->vlan_pvid_set)(dev, pvid, on);
+
+ return 0;
+}
int
rte_eth_dev_fdir_add_signature_filter(uint8_t port_id,
diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h
index 948a2cc..593977e 100644
--- a/lib/librte_ether/rte_ethdev.h
+++ b/lib/librte_ether/rte_ethdev.h
@@ -230,6 +230,9 @@ struct rte_eth_link {
#define ETH_LINK_SPEED_100 100 /**< 100 megabits/second. */
#define ETH_LINK_SPEED_1000 1000/**< 1 gigabits/second. */
#define ETH_LINK_SPEED_11 /**< 10 gigabits/second. */
+#define ETH_LINK_SPEED_10G 1 /**< alias of 10 gigabits/second. */
+#define ETH_LINK_SPEED_20G 2 /**< 20 gigabits/second. */
+#define ETH_LINK_SPEED_40G 4 /**< 40 gigabits/second. */
#define ETH_LINK_AUTONEG_DUPLEX 0 /**< Auto-negotiate duplex. */
#define ETH_LINK_HALF_DUPLEX1 /**< Half-duplex connection. */
@@ -308,31 +311,116 @@ struct rte_eth_rxmode {
* A structure used to configure the Receive Side Scaling (RSS) feature
* of an Ethernet port.
* If not NULL, the *rss_key* pointer of the *rss_conf* structure points
- * to an array of 40 bytes holding the RSS key to use for hashing specific
- * header fields of received packets.
- * Otherwise, a default random hash key is used by the device driver.
+ * to an array holding the RSS key to use for hashing specific header
+ * fields of received packets. The length of this array should be indicated
+ * by *rss_key_len* below. Otherwise, a default random hash key is used by
+ * the device driver.
+ *
+ * The *rss_key_len* field of the *rss_conf* structure indicates the length
+ * in bytes of the array pointed by *rss_key*. To be compitable, this length
+ * will be checked in i40e only. Others assume 40 bytes to be used as before.
*
* The *rss_hf* field of the *rss_conf* structure indicates the different
* types of IPv4/IPv6 packets to which the RSS hashing must be applied.
* Supplying an *rss_hf* equal to zero disables the RSS feature.
*/
struct rte_eth_rss_conf {
- uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */
- uint16_t rss_hf; /**< Hash functions to apply - see below. */
+ uint8_t *rss_key;/**< If not NULL, 40-byte hash key. */
+ uint8_t rss_key_len; /**< hash key length in bytes. */
+ uint64_t rss_hf; /**< Hash functions to apply - see below. */
};
-#define ETH_RSS_IPV40x0001 /**< IPv4 packet. */
-#define ETH_RSS_IPV4_TCP0x0002 /**< IPv4/TCP packet. */
-#define ETH_RSS_IPV60x0004 /**< IPv6 packet. */
-#define ETH_RSS_IPV6_EX 0x0008 /**< IPv6 packet with extension headers.*/
-#define ETH_RSS_IPV6_TCP0x0010 /**< IPv6/TCP packet. */
-#define ETH_RSS_IPV6_TCP_EX 0x0020 /**< IPv6/TCP with extension headers. */
-/* Intel RSS extensions to UDP packets */
-#define ETH_RSS_IPV4_UDP0x0040 /**< IPv4/UDP packet. */
-#define ETH_RSS_IPV6_UDP0x0080 /**< IPv6/UDP packet. */
-#define ETH_RSS_IPV6_UDP_EX 0x0100 /**< IPv6/UDP with extension headers. */
-
-#define ETH_RSS_PROTO_MASK 0x01FF /**< Mask of valid RSS hash protocols */
+/* Supported RSS offloads */
+/* for 1G & 10G */
+#define ETH_RSS_IPV4_SHIFT0
+#define ETH_RSS_IPV4_TCP_SHIFT1
+#define ETH_RSS_IPV6_SHIFT2
+#define ETH_RSS_IPV6_EX_SHIFT 3
+#define ETH_RSS_IPV6_TCP_SHIFT4
+#define ETH_RSS_IPV6_TCP_EX_SHIFT 5
+#define ETH_RSS_IPV4_UDP_SHIFT6
+#define ETH_RSS_IPV6_UDP_SHIFT7
+#define ETH_RSS_I
[dpdk-dev] [PATCH v2 04/27] igb_uio: add i40e support
The definitions have been added to support i40e devices.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
lib/librte_eal/linuxapp/igb_uio/igb_uio.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c
b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c
index 09c40bf..79fe97d 100644
--- a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c
+++ b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c
@@ -79,6 +79,8 @@ static struct pci_device_id igbuio_pci_ids[] = {
#define RTE_PCI_DEV_ID_DECL_IGBVF(vend, dev) {PCI_DEVICE(vend, dev)},
#define RTE_PCI_DEV_ID_DECL_IXGBE(vend, dev) {PCI_DEVICE(vend, dev)},
#define RTE_PCI_DEV_ID_DECL_IXGBEVF(vend, dev) {PCI_DEVICE(vend, dev)},
+#define RTE_PCI_DEV_ID_DECL_I40E(vend, dev) {PCI_DEVICE(vend, dev)},
+#define RTE_PCI_DEV_ID_DECL_I40EVF(vend, dev) {PCI_DEVICE(vend, dev)},
#ifdef RTE_LIBRTE_VIRTIO_PMD
#define RTE_PCI_DEV_ID_DECL_VIRTIO(vend, dev) {PCI_DEVICE(vend, dev)},
#endif
--
1.8.1.4
[dpdk-dev] [PATCH v2 07/27] ethdev: support setting maximum packet length to less than 1518
In ethdev, it ignores setting maximum packet length to less than
1518. The changes is to remove this limitation and let less than
1518 can be set for 'maximum packet length'.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
lib/librte_ether/rte_ethdev.c | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
index 53cb7a9..1791ed7 100644
--- a/lib/librte_ether/rte_ethdev.c
+++ b/lib/librte_ether/rte_ethdev.c
@@ -657,9 +657,13 @@ rte_eth_dev_configure(uint8_t port_id, uint16_t nb_rx_q,
uint16_t nb_tx_q,
(unsigned)ETHER_MIN_LEN);
return (-EINVAL);
}
- } else
- /* Use default value */
- dev->data->dev_conf.rxmode.max_rx_pkt_len = ETHER_MAX_LEN;
+ } else {
+ if (dev_conf->rxmode.max_rx_pkt_len < ETHER_MIN_LEN ||
+ dev_conf->rxmode.max_rx_pkt_len > ETHER_MAX_LEN)
+ /* Use default value */
+ dev->data->dev_conf.rxmode.max_rx_pkt_len =
+ ETHER_MAX_LEN;
+ }
/* multipe queue mode checking */
diag = rte_eth_dev_check_mq_mode(port_id, nb_rx_q, nb_tx_q, dev_conf);
--
1.8.1.4
[dpdk-dev] [PATCH v2 02/27] i40e: add PMD source files
Add PMD source files to support Intel(R) 40G Ethernet Controllers. The new files are, i40e_osdep.h i40e_ethdev.c i40e_ethdev.h i40e_ethdev_vf.c i40e_logs.h i40e_pf.c i40e_pf.h i40e_rxtx.c i40e_rxtx.h Makefile Signed-off-by: Helin Zhang Signed-off-by: Jing Chen Acked-by: Cunming Liang Acked-by: Jijiang Liu Acked-by: Jingjing Wu Tested-by: Waterman Cao --- lib/librte_pmd_i40e/Makefile | 85 + lib/librte_pmd_i40e/i40e/i40e_osdep.h | 197 ++ lib/librte_pmd_i40e/i40e_ethdev.c | 4019 + lib/librte_pmd_i40e/i40e_ethdev.h | 356 +++ lib/librte_pmd_i40e/i40e_ethdev_vf.c | 1336 +++ lib/librte_pmd_i40e/i40e_logs.h | 74 + lib/librte_pmd_i40e/i40e_pf.c | 928 lib/librte_pmd_i40e/i40e_pf.h | 67 + lib/librte_pmd_i40e/i40e_rxtx.c | 2204 ++ lib/librte_pmd_i40e/i40e_rxtx.h | 189 ++ 10 files changed, 9455 insertions(+) create mode 100644 lib/librte_pmd_i40e/Makefile create mode 100644 lib/librte_pmd_i40e/i40e/i40e_osdep.h create mode 100644 lib/librte_pmd_i40e/i40e_ethdev.c create mode 100644 lib/librte_pmd_i40e/i40e_ethdev.h create mode 100644 lib/librte_pmd_i40e/i40e_ethdev_vf.c create mode 100644 lib/librte_pmd_i40e/i40e_logs.h create mode 100644 lib/librte_pmd_i40e/i40e_pf.c create mode 100644 lib/librte_pmd_i40e/i40e_pf.h create mode 100644 lib/librte_pmd_i40e/i40e_rxtx.c create mode 100644 lib/librte_pmd_i40e/i40e_rxtx.h diff --git a/lib/librte_pmd_i40e/Makefile b/lib/librte_pmd_i40e/Makefile new file mode 100644 index 000..09f2087 --- /dev/null +++ b/lib/librte_pmd_i40e/Makefile @@ -0,0 +1,85 @@ +# BSD LICENSE +# +# Copyright(c) 2010-2014 Intel Corporation. All rights reserved. +# All rights reserved. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in +# the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Intel Corporation nor the names of its +# contributors may be used to endorse or promote products derived +# from this software without specific prior written permission. +# +# 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 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. + +include $(RTE_SDK)/mk/rte.vars.mk + +# +# library name +# +LIB = librte_pmd_i40e.a + +CFLAGS += -O3 +CFLAGS += $(WERROR_FLAGS) + +ifeq ($(CC), icc) +CFLAGS_SHARED_DRIVERS = -wd593 +else +CFLAGS_SHARED_DRIVERS = -Wno-unused-but-set-variable +CFLAGS_SHARED_DRIVERS += -Wno-sign-compare +CFLAGS_SHARED_DRIVERS += -Wno-unused-value +CFLAGS_SHARED_DRIVERS += -Wno-unused-parameter +CFLAGS_SHARED_DRIVERS += -Wno-strict-aliasing +CFLAGS_SHARED_DRIVERS += -Wno-format +CFLAGS_SHARED_DRIVERS += -Wno-missing-field-initializers +CFLAGS_SHARED_DRIVERS += -Wno-pointer-to-int-cast +CFLAGS_SHARED_DRIVERS += -Wno-format-nonliteral +CFLAGS_SHARED_DRIVERS += -Wno-format-security +endif + +# +# Add extra flags for ND source files to disable warnings +# +SHARED_DRIVERS_OBJS=$(patsubst %.c,%.o,$(notdir $(wildcard $(RTE_SDK)/lib/librte_pmd_i40e/i40e/*.c))) +$(foreach obj, $(SHARED_DRIVERS_OBJS), $(eval CFLAGS_$(obj)+=$(CFLAGS_SHARED_DRIVERS))) + +VPATH += $(RTE_SDK)/lib/librte_pmd_i40e/i40e + +# +# all source are stored in SRCS-y +# +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_adminq.c +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_common.c +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_diag.c +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_hmc.c +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_lan_hmc.c +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_nvm.c +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_dcb.c + +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_ethdev.c +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_rxtx.c +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_ethdev_vf.c +SRCS-$(CONFIG_RTE_LIBRTE_I40E_PMD) += i40e_pf.c +# this lib depends upon: +DEPDIRS-$(CONFIG_RTE_L
[dpdk-dev] [PATCH v2 08/27] vmxnet3: enlarge the hash flags of RSS to 64 bits
As the hash flags of RSS has been enlarged from 16 bits to 64 bits
in 'struct rte_eth_rss_conf' in rte_ethdev.h, the size of it in
vmxnet3 has to be enlarged as well. In addition, the flags for
vmxnet3 only are masked, as there are flags for others in that
64 bits.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
lib/librte_pmd_vmxnet3/vmxnet3_rxtx.c | 16
1 file changed, 12 insertions(+), 4 deletions(-)
diff --git a/lib/librte_pmd_vmxnet3/vmxnet3_rxtx.c
b/lib/librte_pmd_vmxnet3/vmxnet3_rxtx.c
index 1072654..ad7cbb2 100644
--- a/lib/librte_pmd_vmxnet3/vmxnet3_rxtx.c
+++ b/lib/librte_pmd_vmxnet3/vmxnet3_rxtx.c
@@ -882,9 +882,16 @@ static uint8_t rss_intel_key[40] = {
int
vmxnet3_rss_configure(struct rte_eth_dev *dev)
{
+#define VMXNET3_RSS_OFFLOAD_ALL ( \
+ ETH_RSS_IPV4 | \
+ ETH_RSS_IPV4_TCP | \
+ ETH_RSS_IPV6 | \
+ ETH_RSS_IPV6_TCP)
+
struct vmxnet3_hw *hw;
struct VMXNET3_RSSConf *dev_rss_conf;
struct rte_eth_rss_conf *port_rss_conf;
+ uint64_t rss_hf;
uint8_t i, j;
PMD_INIT_FUNC_TRACE();
@@ -916,13 +923,14 @@ vmxnet3_rss_configure(struct rte_eth_dev *dev)
/* loading hashType */
dev_rss_conf->hashType = 0;
- if (port_rss_conf->rss_hf & ETH_RSS_IPV4)
+ rss_hf = port_rss_conf->rss_hf & VMXNET3_RSS_OFFLOAD_ALL;
+ if (rss_hf & ETH_RSS_IPV4)
dev_rss_conf->hashType |= VMXNET3_RSS_HASH_TYPE_IPV4;
- if (port_rss_conf->rss_hf & ETH_RSS_IPV4_TCP)
+ if (rss_hf & ETH_RSS_IPV4_TCP)
dev_rss_conf->hashType |= VMXNET3_RSS_HASH_TYPE_TCP_IPV4;
- if (port_rss_conf->rss_hf & ETH_RSS_IPV6)
+ if (rss_hf & ETH_RSS_IPV6)
dev_rss_conf->hashType |= VMXNET3_RSS_HASH_TYPE_IPV6;
- if (port_rss_conf->rss_hf & ETH_RSS_IPV6_TCP)
+ if (rss_hf & ETH_RSS_IPV6_TCP)
dev_rss_conf->hashType |= VMXNET3_RSS_HASH_TYPE_TCP_IPV6;
return VMXNET3_SUCCESS;
--
1.8.1.4
[dpdk-dev] [PATCH v2 11/27] mk: add i40e support
i40e support has been added in lib/Makefile and mk/rte.app.mk. Signed-off-by: Helin Zhang Signed-off-by: Jing Chen Acked-by: Cunming Liang Acked-by: Jijiang Liu Acked-by: Jingjing Wu Tested-by: Waterman Cao --- lib/Makefile | 1 + mk/rte.app.mk | 4 2 files changed, 5 insertions(+) diff --git a/lib/Makefile b/lib/Makefile index b92b392..e5b6635 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -42,6 +42,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_CMDLINE) += librte_cmdline DIRS-$(CONFIG_RTE_LIBRTE_ETHER) += librte_ether DIRS-$(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-$(CONFIG_RTE_LIBRTE_PMD_RING) += librte_pmd_ring DIRS-$(CONFIG_RTE_LIBRTE_PMD_PCAP) += librte_pmd_pcap DIRS-$(CONFIG_RTE_LIBRTE_VIRTIO_PMD) += librte_pmd_virtio diff --git a/mk/rte.app.mk b/mk/rte.app.mk index a836577..ceb7fd8 100644 --- a/mk/rte.app.mk +++ b/mk/rte.app.mk @@ -161,6 +161,10 @@ ifeq ($(CONFIG_RTE_LIBRTE_VIRTIO_PMD),y) LDLIBS += -lrte_pmd_virtio_uio endif +ifeq ($(CONFIG_RTE_LIBRTE_I40E_PMD),y) +LDLIBS += -lrte_pmd_i40e +endif + ifeq ($(CONFIG_RTE_LIBRTE_IXGBE_PMD),y) LDLIBS += -lrte_pmd_ixgbe endif -- 1.8.1.4
[dpdk-dev] [PATCH v2 13/27] app/test-pmd: support displaying 32 bytes RX descriptors
i40e supports both 16 and 32 bytes RX descriptors, while ixgbe and
igb support 16 bytes size only. Code changes have been made in
test-pmd to support both 16 and 32 bytes RX descriptors according
to the configuration in config files.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
app/test-pmd/config.c | 36 +---
1 file changed, 33 insertions(+), 3 deletions(-)
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index d6291e7..3dc7fba 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -558,13 +558,43 @@ union igb_ring_dword {
} words;
};
+#ifndef RTE_LIBRTE_I40E_16BYTE_RX_DESC
+struct igb_ring_desc_32B {
+ union igb_ring_dword lo_dword;
+ union igb_ring_dword hi_dword;
+ union igb_ring_dword resv1;
+ union igb_ring_dword resv2;
+};
+#endif
+
struct igb_ring_desc {
union igb_ring_dword lo_dword;
union igb_ring_dword hi_dword;
};
static void
-ring_descriptor_display(const struct rte_memzone *ring_mz, uint16_t desc_id)
+ring_rx_descriptor_display(const struct rte_memzone *ring_mz, uint16_t desc_id)
+{
+#ifdef RTE_LIBRTE_I40E_16BYTE_RX_DESC
+ struct igb_ring_desc *ring;
+ struct igb_ring_desc rd;
+
+ ring = (struct igb_ring_desc *) ring_mz->addr;
+#else
+ struct igb_ring_desc_32B *ring;
+ struct igb_ring_desc_32B rd;
+
+ ring = (struct igb_ring_desc_32B *) ring_mz->addr;
+#endif
+ rd.lo_dword = rte_le_to_cpu_64(ring[desc_id].lo_dword);
+ rd.hi_dword = rte_le_to_cpu_64(ring[desc_id].hi_dword);
+ printf("0x%08X - 0x%08X / 0x%08X - 0x%08X\n",
+ (unsigned)rd.lo_dword.words.lo, (unsigned)rd.lo_dword.words.hi,
+ (unsigned)rd.hi_dword.words.lo, (unsigned)rd.hi_dword.words.hi);
+}
+
+static void
+ring_tx_descriptor_display(const struct rte_memzone *ring_mz, uint16_t desc_id)
{
struct igb_ring_desc *ring;
struct igb_ring_desc rd;
@@ -591,7 +621,7 @@ rx_ring_desc_display(portid_t port_id, queueid_t rxq_id,
uint16_t rxd_id)
rx_mz = ring_dma_zone_lookup("rx_ring", port_id, rxq_id);
if (rx_mz == NULL)
return;
- ring_descriptor_display(rx_mz, rxd_id);
+ ring_rx_descriptor_display(rx_mz, rxd_id);
}
void
@@ -608,7 +638,7 @@ tx_ring_desc_display(portid_t port_id, queueid_t txq_id,
uint16_t txd_id)
tx_mz = ring_dma_zone_lookup("tx_ring", port_id, txq_id);
if (tx_mz == NULL)
return;
- ring_descriptor_display(tx_mz, txd_id);
+ ring_tx_descriptor_display(tx_mz, txd_id);
}
void
--
1.8.1.4
[dpdk-dev] [PATCH v2 10/27] igb: enlarge the hash flags of RSS to 64 bits
As the hash flags of RSS has been enlarged from 16 bits to 64 bits
in 'struct rte_eth_rss_conf' in rte_ethdev.h, the size of it in
igb has to be enlarged as well. In addition, the flags for igb
only are masked, as there are flags for others in those 64 bits.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
lib/librte_pmd_e1000/igb_rxtx.c | 21 -
1 file changed, 16 insertions(+), 5 deletions(-)
diff --git a/lib/librte_pmd_e1000/igb_rxtx.c b/lib/librte_pmd_e1000/igb_rxtx.c
index ae53428..cbb9588 100644
--- a/lib/librte_pmd_e1000/igb_rxtx.c
+++ b/lib/librte_pmd_e1000/igb_rxtx.c
@@ -73,6 +73,17 @@
#include "e1000/e1000_api.h"
#include "e1000_ethdev.h"
+#define IGB_RSS_OFFLOAD_ALL ( \
+ ETH_RSS_IPV4 | \
+ ETH_RSS_IPV4_TCP | \
+ ETH_RSS_IPV6 | \
+ ETH_RSS_IPV6_EX | \
+ ETH_RSS_IPV6_TCP | \
+ ETH_RSS_IPV6_TCP_EX | \
+ ETH_RSS_IPV4_UDP | \
+ ETH_RSS_IPV6_UDP | \
+ ETH_RSS_IPV6_UDP_EX)
+
static inline struct rte_mbuf *
rte_rxmbuf_alloc(struct rte_mempool *mp)
{
@@ -1524,7 +1535,7 @@ igb_hw_rss_hash_set(struct e1000_hw *hw, struct
rte_eth_rss_conf *rss_conf)
uint8_t *hash_key;
uint32_t rss_key;
uint32_t mrqc;
- uint16_t rss_hf;
+ uint64_t rss_hf;
uint16_t i;
hash_key = rss_conf->rss_key;
@@ -1569,7 +1580,7 @@ eth_igb_rss_hash_update(struct rte_eth_dev *dev,
{
struct e1000_hw *hw;
uint32_t mrqc;
- uint16_t rss_hf;
+ uint64_t rss_hf;
hw = E1000_DEV_PRIVATE_TO_HW(dev->data->dev_private);
@@ -1579,7 +1590,7 @@ eth_igb_rss_hash_update(struct rte_eth_dev *dev,
* initialization time, or does not attempt to enable RSS, if RSS was
* disabled at initialization time.
*/
- rss_hf = rss_conf->rss_hf;
+ rss_hf = rss_conf->rss_hf & IGB_RSS_OFFLOAD_ALL;
mrqc = E1000_READ_REG(hw, E1000_MRQC);
if (!(mrqc & E1000_MRQC_ENABLE_MASK)) { /* RSS disabled */
if (rss_hf != 0) /* Enable RSS */
@@ -1600,7 +1611,7 @@ int eth_igb_rss_hash_conf_get(struct rte_eth_dev *dev,
uint8_t *hash_key;
uint32_t rss_key;
uint32_t mrqc;
- uint16_t rss_hf;
+ uint64_t rss_hf;
uint16_t i;
hw = E1000_DEV_PRIVATE_TO_HW(dev->data->dev_private);
@@ -1676,7 +1687,7 @@ igb_rss_configure(struct rte_eth_dev *dev)
* the RSS hash of input packets.
*/
rss_conf = dev->data->dev_conf.rx_adv_conf.rss_conf;
- if (rss_conf.rss_hf == 0) {
+ if ((rss_conf.rss_hf & IGB_RSS_OFFLOAD_ALL) == 0) {
igb_rss_disable(dev);
return;
}
--
1.8.1.4
[dpdk-dev] [PATCH v2 09/27] ixgbe: enlarge the hash flags of RSS to 64 bits
As the hash flags of RSS has been enlarged from 16 bits to 64 bits
in 'struct rte_eth_rss_conf' in rte_ethdev.h, the size of it in
ixgbe has to be enlarged as well. In addition, the flags for ixgbe
only are masked, as there are flags for others in those 64 bits.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
lib/librte_pmd_ixgbe/ixgbe_rxtx.c | 22 +++---
1 file changed, 15 insertions(+), 7 deletions(-)
diff --git a/lib/librte_pmd_ixgbe/ixgbe_rxtx.c
b/lib/librte_pmd_ixgbe/ixgbe_rxtx.c
index dc79c4b..bdae98e 100644
--- a/lib/librte_pmd_ixgbe/ixgbe_rxtx.c
+++ b/lib/librte_pmd_ixgbe/ixgbe_rxtx.c
@@ -77,10 +77,18 @@
#include "ixgbe_ethdev.h"
#include "ixgbe/ixgbe_dcb.h"
#include "ixgbe/ixgbe_common.h"
-
-
#include "ixgbe_rxtx.h"
+#define IXGBE_RSS_OFFLOAD_ALL ( \
+ ETH_RSS_IPV4 | \
+ ETH_RSS_IPV4_TCP | \
+ ETH_RSS_IPV6 | \
+ ETH_RSS_IPV6_EX | \
+ ETH_RSS_IPV6_TCP | \
+ ETH_RSS_IPV6_TCP_EX | \
+ ETH_RSS_IPV4_UDP | \
+ ETH_RSS_IPV6_UDP | \
+ ETH_RSS_IPV6_UDP_EX)
static inline struct rte_mbuf *
rte_rxmbuf_alloc(struct rte_mempool *mp)
@@ -2299,7 +2307,7 @@ ixgbe_hw_rss_hash_set(struct ixgbe_hw *hw, struct
rte_eth_rss_conf *rss_conf)
uint8_t *hash_key;
uint32_t mrqc;
uint32_t rss_key;
- uint16_t rss_hf;
+ uint64_t rss_hf;
uint16_t i;
hash_key = rss_conf->rss_key;
@@ -2344,7 +2352,7 @@ ixgbe_dev_rss_hash_update(struct rte_eth_dev *dev,
{
struct ixgbe_hw *hw;
uint32_t mrqc;
- uint16_t rss_hf;
+ uint64_t rss_hf;
hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
@@ -2357,7 +2365,7 @@ ixgbe_dev_rss_hash_update(struct rte_eth_dev *dev,
* initialization time, or does not attempt to enable RSS, if RSS was
* disabled at initialization time.
*/
- rss_hf = rss_conf->rss_hf;
+ rss_hf = rss_conf->rss_hf & IXGBE_RSS_OFFLOAD_ALL;
mrqc = IXGBE_READ_REG(hw, IXGBE_MRQC);
if (!(mrqc & IXGBE_MRQC_RSSEN)) { /* RSS disabled */
if (rss_hf != 0) /* Enable RSS */
@@ -2379,7 +2387,7 @@ ixgbe_dev_rss_hash_conf_get(struct rte_eth_dev *dev,
uint8_t *hash_key;
uint32_t mrqc;
uint32_t rss_key;
- uint16_t rss_hf;
+ uint64_t rss_hf;
uint16_t i;
hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
@@ -2456,7 +2464,7 @@ ixgbe_rss_configure(struct rte_eth_dev *dev)
* the RSS hash of input packets.
*/
rss_conf = dev->data->dev_conf.rx_adv_conf.rss_conf;
- if (rss_conf.rss_hf == 0) {
+ if ((rss_conf.rss_hf & IXGBE_RSS_OFFLOAD_ALL) == 0) {
ixgbe_rss_disable(dev);
return;
}
--
1.8.1.4
[dpdk-dev] [PATCH v2 12/27] config: add configurations for i40e
configurations for i40e have been added in both common_bsdapp and common_linuxapp. Signed-off-by: Helin Zhang Signed-off-by: Jing Chen Acked-by: Cunming Liang Acked-by: Jijiang Liu Acked-by: Jingjing Wu Tested-by: Waterman Cao --- config/common_bsdapp | 20 config/common_linuxapp | 20 2 files changed, 40 insertions(+) diff --git a/config/common_bsdapp b/config/common_bsdapp index 2cc7b80..5e17d46 100644 --- a/config/common_bsdapp +++ b/config/common_bsdapp @@ -165,6 +165,26 @@ CONFIG_RTE_LIBRTE_IXGBE_RX_ALLOW_BULK_ALLOC=y CONFIG_RTE_LIBRTE_IXGBE_ALLOW_UNSUPPORTED_SFP=n # +# Compile burst-oriented I40E PMD driver +# +# CONFIG_RTE_LIBRTE_I40E_ITR_INTERVAL can be 0 to 8160 us, +# otherwise a default value will be used instead. +# The interval will be aligned to 2, due to the hardware restriction. +# +CONFIG_RTE_LIBRTE_I40E_PMD=y +CONFIG_RTE_LIBRTE_I40E_DEBUG_INIT=y +CONFIG_RTE_LIBRTE_I40E_DEBUG_RX=y +CONFIG_RTE_LIBRTE_I40E_DEBUG_TX=y +CONFIG_RTE_LIBRTE_I40E_DEBUG_TX_FREE=y +CONFIG_RTE_LIBRTE_I40E_DEBUG_DRIVER=y +CONFIG_RTE_LIBRTE_I40E_PF_DISABLE_STRIP_CRC=y +CONFIG_RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC=n +CONFIG_RTE_LIBRTE_I40E_ALLOW_UNSUPPORTED_SFP=y +CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC=n +CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_VF=4 +CONFIG_RTE_LIBRTE_I40E_ITR_INTERVAL=-1 + +# # Compile burst-oriented VIRTIO PMD driver # CONFIG_RTE_LIBRTE_VIRTIO_PMD=n diff --git a/config/common_linuxapp b/config/common_linuxapp index 62619c6..da20011 100644 --- a/config/common_linuxapp +++ b/config/common_linuxapp @@ -179,6 +179,26 @@ CONFIG_RTE_LIBRTE_IXGBE_RX_ALLOW_BULK_ALLOC=y CONFIG_RTE_LIBRTE_IXGBE_ALLOW_UNSUPPORTED_SFP=n # +# Compile burst-oriented I40E PMD driver +# +# CONFIG_RTE_LIBRTE_I40E_ITR_INTERVAL can be 0 to 8160 us, +# otherwise a default value will be used instead. +# The interval will be aligned to 2, due to the hardware restriction. +# +CONFIG_RTE_LIBRTE_I40E_PMD=y +CONFIG_RTE_LIBRTE_I40E_DEBUG_INIT=n +CONFIG_RTE_LIBRTE_I40E_DEBUG_RX=n +CONFIG_RTE_LIBRTE_I40E_DEBUG_TX=n +CONFIG_RTE_LIBRTE_I40E_DEBUG_TX_FREE=n +CONFIG_RTE_LIBRTE_I40E_DEBUG_DRIVER=n +CONFIG_RTE_LIBRTE_I40E_PF_DISABLE_STRIP_CRC=n +CONFIG_RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC=y +CONFIG_RTE_LIBRTE_I40E_ALLOW_UNSUPPORTED_SFP=n +CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC=n +CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_VF=4 +CONFIG_RTE_LIBRTE_I40E_ITR_INTERVAL=-1 + +# # Compile burst-oriented VIRTIO PMD driver # CONFIG_RTE_LIBRTE_VIRTIO_PMD=y -- 1.8.1.4
[dpdk-dev] [PATCH v2 18/27] examples/ip_reassembly: use ETH_RSS_IP to replace IP hash flags of RSS
As hash flags of RSS have been enlarged from 16 bits to 64 bits to
support more possible types in all PMDs, and new macro of
ETH_RSS_IP has been defined to cover all IP hash flags of RSS.
That macro should be used in ip_reassembly example application to
support all PMDs.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
examples/ip_reassembly/main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/examples/ip_reassembly/main.c b/examples/ip_reassembly/main.c
index bafa8d9..bb2f810 100644
--- a/examples/ip_reassembly/main.c
+++ b/examples/ip_reassembly/main.c
@@ -235,7 +235,7 @@ static struct rte_eth_conf port_conf = {
.rx_adv_conf = {
.rss_conf = {
.rss_key = NULL,
- .rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6,
+ .rss_hf = ETH_RSS_IP,
},
},
.txmode = {
--
1.8.1.4
[dpdk-dev] [PATCH v2 19/27] examples/l3fwd-power: use ETH_RSS_IP to replace IP hash flags of RSS
As hash flags of RSS have been enlarged from 16 bits to 64 bits to
support more possible types in all PMDs, and new macro of
ETH_RSS_IP has been defined to cover all IP hash flags of RSS.
That macro should be used in l3fwd-power example application to
support all PMDs.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
examples/l3fwd-power/main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/examples/l3fwd-power/main.c b/examples/l3fwd-power/main.c
index 598b7a2..ab5d2a1 100644
--- a/examples/l3fwd-power/main.c
+++ b/examples/l3fwd-power/main.c
@@ -247,7 +247,7 @@ static struct rte_eth_conf port_conf = {
.rx_adv_conf = {
.rss_conf = {
.rss_key = NULL,
- .rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6,
+ .rss_hf = ETH_RSS_IP,
},
},
.txmode = {
--
1.8.1.4
[dpdk-dev] [PATCH v2 20/27] examples/l3fwd-vf: use ETH_RSS_IP to replace IP hash flags of RSS
As hash flags of RSS have been enlarged from 16 bits to 64 bits to
support more possible types in all PMDs, and new macro of
ETH_RSS_IP has been defined to cover all IP hash flags of RSS.
That macro should be used in l3fwd-vf example application to
support all PMDs.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
examples/l3fwd-vf/main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/examples/l3fwd-vf/main.c b/examples/l3fwd-vf/main.c
index 793cacc..d37eb35 100644
--- a/examples/l3fwd-vf/main.c
+++ b/examples/l3fwd-vf/main.c
@@ -207,7 +207,7 @@ static struct rte_eth_conf port_conf = {
.rx_adv_conf = {
.rss_conf = {
.rss_key = NULL,
- .rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6,
+ .rss_hf = ETH_RSS_IP,
},
},
.txmode = {
--
1.8.1.4
[dpdk-dev] [PATCH v2 21/27] examples/l3fwd: use ETH_RSS_IP to replace IP hash flags of RSS
As hash flags of RSS have been enlarged from 16 bits to 64 bits to
support more possible types in all PMDs, and new macro of
ETH_RSS_IP has been defined to cover all IP hash flags of RSS.
That macro should be used in l3fwd example application to support
all PMDs.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
examples/l3fwd/main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/examples/l3fwd/main.c b/examples/l3fwd/main.c
index 8ee1af9..6588d3c 100755
--- a/examples/l3fwd/main.c
+++ b/examples/l3fwd/main.c
@@ -227,7 +227,7 @@ static struct rte_eth_conf port_conf = {
.rx_adv_conf = {
.rss_conf = {
.rss_key = NULL,
- .rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6,
+ .rss_hf = ETH_RSS_IP,
},
},
.txmode = {
--
1.8.1.4
[dpdk-dev] [PATCH v2 22/27] examples/load_balancer: use ETH_RSS_IP to replace IP hash flags of RSS
As hash flags of RSS have been enlarged from 16 bits to 64 bits to
support more possible types in all PMDs, and new macro of
ETH_RSS_IP has been defined to cover all IP hash flags of RSS.
That macro should be used in load_balancer example application to
support all PMDs.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
examples/load_balancer/init.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/examples/load_balancer/init.c b/examples/load_balancer/init.c
index e997238..160a938 100644
--- a/examples/load_balancer/init.c
+++ b/examples/load_balancer/init.c
@@ -87,7 +87,7 @@ static struct rte_eth_conf port_conf = {
.rx_adv_conf = {
.rss_conf = {
.rss_key = NULL,
- .rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6,
+ .rss_hf = ETH_RSS_IP,
},
},
.txmode = {
--
1.8.1.4
[dpdk-dev] [PATCH v2 15/27] app/testpmd: enlarge the hash flags of RSS to 64 bits
As the hash flags of RSS has been enlarged from 16 bits to 64 bits in
'struct rte_eth_rss_conf' in rte_ethdev.h, the size of it in testpmd
should be enlarged as well. In addition, macro of ETH_RSS_IP is used
to replace all IP hash flags of RSS.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
app/test-pmd/cmdline.c| 4 ++--
app/test-pmd/parameters.c | 5 ++---
app/test-pmd/testpmd.c| 2 +-
app/test-pmd/testpmd.h| 2 +-
4 files changed, 6 insertions(+), 7 deletions(-)
diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index ddb825d..fb52e89 100644
--- a/app/test-pmd/cmdline.c
+++ b/app/test-pmd/cmdline.c
@@ -1135,9 +1135,9 @@ cmd_config_rss_parsed(void *parsed_result,
uint8_t i;
if (!strcmp(res->value, "ip"))
- rss_conf.rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6;
+ rss_conf.rss_hf = ETH_RSS_IP;
else if (!strcmp(res->value, "udp"))
- rss_conf.rss_hf = ETH_RSS_IPV4_UDP | ETH_RSS_IPV6_UDP;
+ rss_conf.rss_hf = ETH_RSS_UDP;
else if (!strcmp(res->value, "none"))
rss_conf.rss_hf = 0;
else {
diff --git a/app/test-pmd/parameters.c b/app/test-pmd/parameters.c
index 7a60048..725ca6f 100644
--- a/app/test-pmd/parameters.c
+++ b/app/test-pmd/parameters.c
@@ -832,10 +832,9 @@ launch_args_parse(int argc, char** argv)
if (!strcmp(lgopts[opt_idx].name, "forward-mode"))
set_pkt_forwarding_mode(optarg);
if (!strcmp(lgopts[opt_idx].name, "rss-ip"))
- rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6;
+ rss_hf = ETH_RSS_IP;
if (!strcmp(lgopts[opt_idx].name, "rss-udp"))
- rss_hf = ETH_RSS_IPV4 |
- ETH_RSS_IPV6 | ETH_RSS_IPV4_UDP;
+ rss_hf = ETH_RSS_UDP;
if (!strcmp(lgopts[opt_idx].name, "rxq")) {
n = atoi(optarg);
if (n >= 1 && n <= (int) MAX_QUEUE_ID)
diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
index bc38305..eea2c6c 100644
--- a/app/test-pmd/testpmd.c
+++ b/app/test-pmd/testpmd.c
@@ -250,7 +250,7 @@ uint32_t txq_flags = 0; /* No flags set. */
/*
* Receive Side Scaling (RSS) configuration.
*/
-uint16_t rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6; /* RSS IP by default. */
+uint64_t rss_hf = ETH_RSS_IP; /* RSS IP by default. */
/*
* Port topology configuration
diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h
index c2b970e..f91f9cd 100644
--- a/app/test-pmd/testpmd.h
+++ b/app/test-pmd/testpmd.h
@@ -326,7 +326,7 @@ extern portid_t fwd_ports_ids[RTE_MAX_ETHPORTS];
extern struct rte_port *ports;
extern struct rte_eth_rxmode rx_mode;
-extern uint16_t rss_hf;
+extern uint64_t rss_hf;
extern queueid_t nb_rxq;
extern queueid_t nb_txq;
--
1.8.1.4
[dpdk-dev] [PATCH v2 14/27] app/test-pmd: add command of 'tx_vlan set pvid port_id vlan_id (on|off)'
New command of 'tx_vlan set pvid port_id vlan_id (on|off)' has been
added in test-pmd to configure port based vlan insertion.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
app/test-pmd/cmdline.c | 61 ++
app/test-pmd/config.c | 9
app/test-pmd/testpmd.h | 2 +-
3 files changed, 71 insertions(+), 1 deletion(-)
diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index 0be28f6..ddb825d 100644
--- a/app/test-pmd/cmdline.c
+++ b/app/test-pmd/cmdline.c
@@ -285,6 +285,9 @@ static void cmd_help_long_parsed(void *parsed_result,
"Set hardware insertion of VLAN ID in packets sent"
" on a port.\n\n"
+ "tx_vlan set pvid port_id vlan_id (on|off)\n"
+ "Set port based TX VLAN insertion.\n\n"
+
"tx_vlan reset (port_id)\n"
"Disable hardware insertion of a VLAN header in"
" packets sent on a port.\n\n"
@@ -2352,6 +2355,63 @@ cmdline_parse_inst_t cmd_tx_vlan_set = {
},
};
+/* *** ENABLE/DISABLE PORT BASED TX VLAN INSERTION *** */
+struct cmd_tx_vlan_set_pvid_result {
+ cmdline_fixed_string_t tx_vlan;
+ cmdline_fixed_string_t set;
+ cmdline_fixed_string_t pvid;
+ uint8_t port_id;
+ uint16_t vlan_id;
+ cmdline_fixed_string_t mode;
+};
+
+static void
+cmd_tx_vlan_set_pvid_parsed(void *parsed_result,
+ __attribute__((unused)) struct cmdline *cl,
+ __attribute__((unused)) void *data)
+{
+ struct cmd_tx_vlan_set_pvid_result *res = parsed_result;
+
+ if (strcmp(res->mode, "on") == 0)
+ tx_vlan_pvid_set(res->port_id, res->vlan_id, 1);
+ else
+ tx_vlan_pvid_set(res->port_id, res->vlan_id, 0);
+}
+
+cmdline_parse_token_string_t cmd_tx_vlan_set_pvid_tx_vlan =
+ TOKEN_STRING_INITIALIZER(struct cmd_tx_vlan_set_pvid_result,
+tx_vlan, "tx_vlan");
+cmdline_parse_token_string_t cmd_tx_vlan_set_pvid_set =
+ TOKEN_STRING_INITIALIZER(struct cmd_tx_vlan_set_pvid_result,
+set, "set");
+cmdline_parse_token_string_t cmd_tx_vlan_set_pvid_pvid =
+ TOKEN_STRING_INITIALIZER(struct cmd_tx_vlan_set_pvid_result,
+pvid, "pvid");
+cmdline_parse_token_num_t cmd_tx_vlan_set_pvid_port_id =
+ TOKEN_NUM_INITIALIZER(struct cmd_tx_vlan_set_pvid_result,
+port_id, UINT8);
+cmdline_parse_token_num_t cmd_tx_vlan_set_pvid_vlan_id =
+ TOKEN_NUM_INITIALIZER(struct cmd_tx_vlan_set_pvid_result,
+ vlan_id, UINT16);
+cmdline_parse_token_string_t cmd_tx_vlan_set_pvid_mode =
+ TOKEN_STRING_INITIALIZER(struct cmd_tx_vlan_set_pvid_result,
+mode, "on#off");
+
+cmdline_parse_inst_t cmd_tx_vlan_set_pvid = {
+ .f = cmd_tx_vlan_set_pvid_parsed,
+ .data = NULL,
+ .help_str = "tx_vlan set pvid port_id vlan_id (on|off)",
+ .tokens = {
+ (void *)&cmd_tx_vlan_set_pvid_tx_vlan,
+ (void *)&cmd_tx_vlan_set_pvid_set,
+ (void *)&cmd_tx_vlan_set_pvid_pvid,
+ (void *)&cmd_tx_vlan_set_pvid_port_id,
+ (void *)&cmd_tx_vlan_set_pvid_vlan_id,
+ (void *)&cmd_tx_vlan_set_pvid_mode,
+ NULL,
+ },
+};
+
/* *** DISABLE HARDWARE INSERTION OF VLAN HEADER IN TX PACKETS *** */
struct cmd_tx_vlan_reset_result {
cmdline_fixed_string_t tx_vlan;
@@ -5339,6 +5399,7 @@ cmdline_parse_ctx_t main_ctx[] = {
(cmdline_parse_inst_t *)&cmd_rx_vlan_filter,
(cmdline_parse_inst_t *)&cmd_tx_vlan_set,
(cmdline_parse_inst_t *)&cmd_tx_vlan_reset,
+ (cmdline_parse_inst_t *)&cmd_tx_vlan_set_pvid,
(cmdline_parse_inst_t *)&cmd_tx_cksum_set,
(cmdline_parse_inst_t *)&cmd_link_flow_control_set,
(cmdline_parse_inst_t *)&cmd_priority_flow_control_set,
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index 3dc7fba..a5814ca 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -1603,6 +1603,15 @@ tx_vlan_reset(portid_t port_id)
}
void
+tx_vlan_pvid_set(portid_t port_id, uint16_t vlan_id, int on)
+{
+ if (port_id_is_invalid(port_id))
+ return;
+
+ rte_eth_dev_set_vlan_pvid(port_id, vlan_id, on);
+}
+
+void
set_qmap(portid_t port_id, uint8_t is_rx, uint16_t queue_id, uint8_t map_value)
{
uint16_t i;
diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h
index 0cf5a92..c2b970e 100644
--- a/app/test-pmd/testpmd.h
+++ b/app/test-pmd/testpmd.h
@@ -486,7 +486,7 @@ void vlan_extend_set(portid_t port_id, int on);
void vlan_tpid_set(portid_t port_id, uint16_t tp_id);
void tx_vlan_set(porti
[dpdk-dev] [PATCH v2 25/27] igb_uio: add sys files to read/write specific bits in pci config space
Enabling 'Extended Tag' and resetting 'Max Read Request Size' in PCI
config space have big impacts to i40e performance. They cannot be
changed on some BIOS implementations, though can on others. Two sys
files of 'extended_tag' and 'max_read_request_size' are added to
support changing them by 'echo' in user space.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
lib/librte_eal/linuxapp/igb_uio/igb_uio.c | 105 ++
1 file changed, 105 insertions(+)
diff --git a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c
b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c
index 79fe97d..55e801d 100644
--- a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c
+++ b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c
@@ -47,6 +47,15 @@
#define PCI_MSIX_ENTRY_CTRL_MASKBIT 1
#endif
+#ifdef RTE_PCI_CONFIG
+#define PCI_SYS_FILE_BUF_SIZE 10
+#define PCI_DEV_CAP_REG0xA4
+#define PCI_DEV_CTRL_REG 0xA8
+#define PCI_DEV_CAP_EXT_TAG_MASK 0x20
+#define PCI_DEV_CTRL_EXT_TAG_SHIFT 8
+#define PCI_DEV_CTRL_EXT_TAG_MASK (1 << PCI_DEV_CTRL_EXT_TAG_SHIFT)
+#endif
+
#define IGBUIO_NUM_MSI_VECTORS 1
/* interrupt mode */
@@ -151,9 +160,105 @@ store_max_vfs(struct device *dev, struct device_attribute
*attr,
return err ? err : count;
}
+#ifdef RTE_PCI_CONFIG
+static ssize_t
+show_extended_tag(struct device *dev, struct device_attribute *attr, char *buf)
+{
+ struct pci_dev *pci_dev = container_of(dev, struct pci_dev, dev);
+ uint32_t val = 0;
+
+ pci_read_config_dword(pci_dev, PCI_DEV_CAP_REG, &val);
+ if (!(val & PCI_DEV_CAP_EXT_TAG_MASK)) /* Not supported */
+ return snprintf(buf, PCI_SYS_FILE_BUF_SIZE, "%s\n", "invalid");
+
+ val = 0;
+ pci_bus_read_config_dword(pci_dev->bus, pci_dev->devfn,
+ PCI_DEV_CTRL_REG, &val);
+
+ return snprintf(buf, PCI_SYS_FILE_BUF_SIZE, "%s\n",
+ (val & PCI_DEV_CTRL_EXT_TAG_MASK) ? "on" : "off");
+}
+
+static ssize_t
+store_extended_tag(struct device *dev,
+ struct device_attribute *attr,
+ const char *buf,
+ size_t count)
+{
+ struct pci_dev *pci_dev = container_of(dev, struct pci_dev, dev);
+ uint32_t val = 0, enable;
+
+ if (strncmp(buf, "on", 2) == 0)
+ enable = 1;
+ else if (strncmp(buf, "off", 3) == 0)
+ enable = 0;
+ else
+ return -EINVAL;
+
+ pci_bus_read_config_dword(pci_dev->bus, pci_dev->devfn,
+ PCI_DEV_CAP_REG, &val);
+ if (!(val & PCI_DEV_CAP_EXT_TAG_MASK)) /* Not supported */
+ return -EPERM;
+
+ val = 0;
+ pci_bus_read_config_dword(pci_dev->bus, pci_dev->devfn,
+ PCI_DEV_CTRL_REG, &val);
+ if (enable)
+ val |= PCI_DEV_CTRL_EXT_TAG_MASK;
+ else
+ val &= ~PCI_DEV_CTRL_EXT_TAG_MASK;
+ pci_bus_write_config_dword(pci_dev->bus, pci_dev->devfn,
+ PCI_DEV_CTRL_REG, val);
+
+ return count;
+}
+
+static ssize_t
+show_max_read_request_size(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct pci_dev *pci_dev = container_of(dev, struct pci_dev, dev);
+ int val = pcie_get_readrq(pci_dev);
+
+ return snprintf(buf, PCI_SYS_FILE_BUF_SIZE, "%d\n", val);
+}
+
+static ssize_t
+store_max_read_request_size(struct device *dev,
+ struct device_attribute *attr,
+ const char *buf,
+ size_t count)
+{
+ struct pci_dev *pci_dev = container_of(dev, struct pci_dev, dev);
+ unsigned long size = 0;
+ int ret;
+
+ if (strict_strtoul(buf, 0, &size) != 0)
+ return -EINVAL;
+
+ ret = pcie_set_readrq(pci_dev, (int)size);
+ if (ret < 0)
+ return ret;
+
+ return count;
+}
+#endif
+
static DEVICE_ATTR(max_vfs, S_IRUGO | S_IWUSR, show_max_vfs, store_max_vfs);
+#ifdef RTE_PCI_CONFIG
+static DEVICE_ATTR(extended_tag, S_IRUGO | S_IWUSR, show_extended_tag, \
+ store_extended_tag);
+static DEVICE_ATTR(max_read_request_size, S_IRUGO | S_IWUSR, \
+ show_max_read_request_size, store_max_read_request_size);
+#endif
+
static struct attribute *dev_attrs[] = {
&dev_attr_max_vfs.attr,
+#ifdef RTE_PCI_CONFIG
+ &dev_attr_extended_tag.attr,
+ &dev_attr_max_read_request_size.attr,
+#endif
NULL,
};
--
1.8.1.4
[dpdk-dev] [PATCH v2 16/27] app/test-pmd: add L3 packet type in offload flags
As i40e PMD need to know the L3 packet type for TX checksum offloading,
the packet type has been added in rte_mbuf offload flags.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
app/test-pmd/csumonly.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/app/test-pmd/csumonly.c b/app/test-pmd/csumonly.c
index 3568ba0..0e6441a 100644
--- a/app/test-pmd/csumonly.c
+++ b/app/test-pmd/csumonly.c
@@ -313,6 +313,7 @@ pkt_burst_checksum_forward(struct fwd_stream *fs)
ol_flags |= PKT_TX_IP_CKSUM;
}
else {
+ ol_flags |= PKT_TX_IPV4;
/* SW checksum calculation */
ipv4_hdr->src_addr++;
ipv4_hdr->hdr_checksum =
get_ipv4_cksum(ipv4_hdr);
@@ -373,6 +374,7 @@ pkt_burst_checksum_forward(struct fwd_stream *fs)
unsigned char *) + l2_len);
l3_len = sizeof(struct ipv6_hdr) ;
l4_proto = ipv6_hdr->proto;
+ ol_flags |= PKT_TX_IPV6;
if (l4_proto == IPPROTO_UDP) {
udp_hdr = (struct udp_hdr*)
(rte_pktmbuf_mtod(mb,
--
1.8.1.4
[dpdk-dev] [PATCH v2 23/27] examples/multi_process: use ETH_RSS_IP to replace IP hash flags of RSS
As hash flags of RSS have been enlarged from 16 bits to 64 bits to
support more possible types in all PMDs, and new macro of
ETH_RSS_IP has been defined to cover all IP hash flags of RSS.
That macro should be used in multi_process example application to
support all PMDs.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
examples/multi_process/symmetric_mp/main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/examples/multi_process/symmetric_mp/main.c
b/examples/multi_process/symmetric_mp/main.c
index 028f98d..7e0a357 100644
--- a/examples/multi_process/symmetric_mp/main.c
+++ b/examples/multi_process/symmetric_mp/main.c
@@ -251,7 +251,7 @@ smp_port_init(uint8_t port, struct rte_mempool *mbuf_pool,
uint16_t num_queues)
.rx_adv_conf = {
.rss_conf = {
.rss_key = NULL,
- .rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6,
+ .rss_hf = ETH_RSS_IP,
},
},
.txmode = {
--
1.8.1.4
[dpdk-dev] [PATCH v2 17/27] examples/dpdk_qat: use ETH_RSS_IP to replace IP hash flags of RSS
As hash flags of RSS have been enlarged from 16 bits to 64 bits to
support more possible types in all PMDs, and new macro of
ETH_RSS_IP has been defined to cover all IP hash flags of RSS.
That macro should be used in dpdk_qat example application to
support all PMDs.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
examples/dpdk_qat/main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/examples/dpdk_qat/main.c b/examples/dpdk_qat/main.c
index 741d3f3..76a1ac5 100644
--- a/examples/dpdk_qat/main.c
+++ b/examples/dpdk_qat/main.c
@@ -170,7 +170,7 @@ static struct rte_eth_conf port_conf = {
.rx_adv_conf = {
.rss_conf = {
.rss_key = NULL,
- .rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6,
+ .rss_hf = ETH_RSS_IP,
},
},
.txmode = {
--
1.8.1.4
[dpdk-dev] [PATCH v2 24/27] examples/qos_meter: use ETH_RSS_IP to replace IP hash flags of RSS
As hash flags of RSS have been enlarged from 16 bits to 64 bits to
support more possible types in all PMDs, and new macro of
ETH_RSS_IP has been defined to cover all IP hash flags of RSS.
That macro should be used in qos_meter example application to
support all PMDs.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
examples/qos_meter/main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/examples/qos_meter/main.c b/examples/qos_meter/main.c
index e1698cc..55a9e98 100755
--- a/examples/qos_meter/main.c
+++ b/examples/qos_meter/main.c
@@ -94,7 +94,7 @@ static struct rte_eth_conf port_conf = {
.rx_adv_conf = {
.rss_conf = {
.rss_key = NULL,
- .rss_hf = ETH_RSS_IPV4 | ETH_RSS_IPV6,
+ .rss_hf = ETH_RSS_IP,
},
},
.txmode = {
--
1.8.1.4
[dpdk-dev] [PATCH v2 26/27] pci: support reading/writing sys files of 'extended_tag' and 'max_read_request_size'
Sys files of 'extended_tag' and 'max_read_request_size' are
supported by igb_uio. Reading or writing them to enable/disable
'Extended Tag' or reset 'Max Read Request Size' automatically
according to the configurations are added.
Signed-off-by: Helin Zhang
Signed-off-by: Jing Chen
Acked-by: Cunming Liang
Acked-by: Jijiang Liu
Acked-by: Jingjing Wu
Tested-by: Waterman Cao
---
lib/librte_eal/linuxapp/eal/eal_pci.c | 99 +++
1 file changed, 99 insertions(+)
diff --git a/lib/librte_eal/linuxapp/eal/eal_pci.c
b/lib/librte_eal/linuxapp/eal/eal_pci.c
index ac2c1fe..17f5675 100644
--- a/lib/librte_eal/linuxapp/eal/eal_pci.c
+++ b/lib/librte_eal/linuxapp/eal/eal_pci.c
@@ -106,6 +106,9 @@ TAILQ_HEAD(uio_res_list, uio_resource);
static struct uio_res_list *uio_res_list = NULL;
static int pci_parse_sysfs_value(const char *filename, uint64_t *val);
+#ifdef RTE_PCI_CONFIG
+static void pci_config_space_set(struct rte_pci_device *dev);
+#endif
/* unbind kernel driver for this device */
static int
@@ -840,6 +843,13 @@ rte_eal_pci_probe_one_driver(struct rte_pci_driver *dr,
struct rte_pci_device *d
}
if (dr->drv_flags & RTE_PCI_DRV_NEED_IGB_UIO) {
+#ifdef RTE_PCI_CONFIG
+ /*
+* Set PCIe config space for high performance.
+* Return value can be ignored.
+*/
+ pci_config_space_set(dev);
+#endif
/* map resources for devices that use igb_uio */
if (pci_uio_map_resource(dev) < 0)
return -1;
@@ -878,3 +888,92 @@ rte_eal_pci_init(void)
}
return 0;
}
+
+#ifdef RTE_PCI_CONFIG
+static int
+pci_config_extended_tag(struct rte_pci_device *dev)
+{
+ struct rte_pci_addr *loc = &dev->addr;
+ char filename[PATH_MAX];
+ char buf[BUFSIZ];
+ FILE *f;
+
+ /* not configured, let it as is */
+ if (strncmp(RTE_PCI_EXTENDED_TAG, "on", 2) != 0 &&
+ strncmp(RTE_PCI_EXTENDED_TAG, "off", 3) != 0)
+ return 0;
+
+ rte_snprintf(filename, sizeof(filename),
+ SYSFS_PCI_DEVICES "/" PCI_PRI_FMT "/" "extended_tag",
+ loc->domain, loc->bus, loc->devid, loc->function);
+ f = fopen(filename, "rw+");
+ if (!f)
+ return -1;
+
+ fgets(buf, sizeof(buf), f);
+ if (strncmp(RTE_PCI_EXTENDED_TAG, "on", 2) == 0) {
+ /* enable Extended Tag*/
+ if (strncmp(buf, "on", 2) != 0) {
+ fseek(f, 0, SEEK_SET);
+ fputs("on", f);
+ }
+ } else {
+ /* disable Extended Tag */
+ if (strncmp(buf, "off", 3) != 0) {
+ fseek(f, 0, SEEK_SET);
+ fputs("off", f);
+ }
+ }
+ fclose(f);
+
+ return 0;
+}
+
+static int
+pci_config_max_read_request_size(struct rte_pci_device *dev)
+{
+ struct rte_pci_addr *loc = &dev->addr;
+ char filename[PATH_MAX];
+ char buf[BUFSIZ], param[BUFSIZ];
+ FILE *f;
+ uint32_t max_size = RTE_PCI_MAX_READ_REQUEST_SIZE;
+
+ /* not configured, let it as is */
+ if (!max_size)
+ return 0;
+
+ rte_snprintf(filename, sizeof(filename),
+ SYSFS_PCI_DEVICES "/" PCI_PRI_FMT "/" "max_read_request_size",
+ loc->domain, loc->bus, loc->devid, loc->function);
+ f = fopen(filename, "rw+");
+ if (!f)
+ return -1;
+
+ fgets(buf, sizeof(buf), f);
+ rte_snprintf(param, sizeof(param), "%d", max_size);
+
+ /* check if the size to be set is the same as current */
+ if (strcmp(buf, param) == 0) {
+ fclose(f);
+ return 0;
+ }
+ fseek(f, 0, SEEK_SET);
+ fputs(param, f);
+ fclose(f);
+
+ return 0;
+}
+
+static void
+pci_config_space_set(struct rte_pci_device *dev)
+{
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return;
+
+ /* configure extended tag */
+ pci_config_extended_tag(dev);
+
+ /* configure max read request size */
+ pci_config_max_read_request_size(dev);
+}
+#endif
--
1.8.1.4
[dpdk-dev] [PATCH v2 27/27] config: add configurations for enabling 'Extended Tag' or resetting 'Max Read Request Size'
Sys files of 'extended_tag' and 'max_read_request_size' have been supported in igb_uio, and can be changed during probing PCI. Three items in configuration files are needed to support them at compile time. Those three items are, - CONFIG_RTE_PCI_CONFIG - CONFIG_RTE_PCI_EXTENDED_TAG - CONFIG_RTE_PCI_MAX_READ_REQUEST_SIZE Signed-off-by: Helin Zhang Signed-off-by: Jing Chen Acked-by: Cunming Liang Acked-by: Jijiang Liu Acked-by: Jingjing Wu Tested-by: Waterman Cao --- config/common_linuxapp | 10 ++ 1 file changed, 10 insertions(+) diff --git a/config/common_linuxapp b/config/common_linuxapp index da20011..a020451 100644 --- a/config/common_linuxapp +++ b/config/common_linuxapp @@ -125,6 +125,16 @@ CONFIG_RTE_EAL_ALLOW_INV_SOCKET_ID=n CONFIG_RTE_EAL_ALWAYS_PANIC_ON_ERROR=n # +# Special configurations in PCI Config Space for high performance +# CONFIG_RTE_PCI_CONFIG is the compile switch for two features below +# CONFIG_RTE_PCI_EXTENDED_TAG can be "on", "off" +# CONFIG_RTE_PCI_MAX_READ_REQUEST_SIZE can be 128, 256, 512, 1024, 2048, 4096 +# +CONFIG_RTE_PCI_CONFIG=n +CONFIG_RTE_PCI_EXTENDED_TAG="" +CONFIG_RTE_PCI_MAX_READ_REQUEST_SIZE=0 + +# # Compile Environment Abstraction Layer for linux # CONFIG_RTE_LIBRTE_EAL_LINUXAPP=y -- 1.8.1.4
[dpdk-dev] [PATCH 0/3] *** Upgrade NIC share codes ***
Acked-by: Helin Zhang -Original Message- From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Jijiang Liu Sent: Thursday, May 15, 2014 10:18 AM To: dev at dpdk.org Subject: [dpdk-dev] [PATCH 0/3] *** Upgrade NIC share codes *** This patchset upgrades NIC share code in ixgbe & e1000 directories and fixs an issue of hash calculation of flow director introduced by upgrading this. Short summary: * Upgrading NIC share code in ixgbe & e1000 directories * Changing README in ixgbe & e1000 directories * Changing ixgbe_osdep.h file * Fixng an issue of hash calculation of flow director * Changing Copyright date of e1000_osdep.c and e1000_osdep.h files Signed-off-by: Jijiang Liu jijiangl (3): Upgrade NIC shared code in ixgbe & e1000 directories Fix an issue of hash calculation of flow director introduced by ixgbe_type.h change Change Copyright date of e1000_osdep.c and e1000_osdep.h files lib/librte_eal/common/include/rte_pci_dev_ids.h |8 - lib/librte_pmd_e1000/e1000/README |2 +- lib/librte_pmd_e1000/e1000/e1000_80003es2lan.c | 44 +- lib/librte_pmd_e1000/e1000/e1000_80003es2lan.h |2 +- lib/librte_pmd_e1000/e1000/e1000_82540.c| 24 +- lib/librte_pmd_e1000/e1000/e1000_82541.c| 14 +- lib/librte_pmd_e1000/e1000/e1000_82541.h|2 +- lib/librte_pmd_e1000/e1000/e1000_82542.c|4 +- lib/librte_pmd_e1000/e1000/e1000_82543.c| 38 +- lib/librte_pmd_e1000/e1000/e1000_82543.h|2 +- lib/librte_pmd_e1000/e1000/e1000_82571.c| 40 +- lib/librte_pmd_e1000/e1000/e1000_82571.h|2 +- lib/librte_pmd_e1000/e1000/e1000_82575.c| 134 +- lib/librte_pmd_e1000/e1000/e1000_82575.h|4 +- lib/librte_pmd_e1000/e1000/e1000_api.c | 10 +- lib/librte_pmd_e1000/e1000/e1000_api.h |2 +- lib/librte_pmd_e1000/e1000/e1000_defines.h |7 +- lib/librte_pmd_e1000/e1000/e1000_hw.h | 13 +- lib/librte_pmd_e1000/e1000/e1000_i210.c | 94 +++- lib/librte_pmd_e1000/e1000/e1000_i210.h | 15 +- lib/librte_pmd_e1000/e1000/e1000_ich8lan.c | 250 +++--- lib/librte_pmd_e1000/e1000/e1000_ich8lan.h | 11 +- lib/librte_pmd_e1000/e1000/e1000_mac.c |5 +- lib/librte_pmd_e1000/e1000/e1000_mac.h |5 +- lib/librte_pmd_e1000/e1000/e1000_manage.c |2 +- lib/librte_pmd_e1000/e1000/e1000_manage.h |2 +- lib/librte_pmd_e1000/e1000/e1000_mbx.c |2 +- lib/librte_pmd_e1000/e1000/e1000_mbx.h |2 +- lib/librte_pmd_e1000/e1000/e1000_nvm.c | 16 +- lib/librte_pmd_e1000/e1000/e1000_nvm.h |2 +- lib/librte_pmd_e1000/e1000/e1000_osdep.c|2 +- lib/librte_pmd_e1000/e1000/e1000_osdep.h|2 +- lib/librte_pmd_e1000/e1000/e1000_phy.c | 38 +- lib/librte_pmd_e1000/e1000/e1000_phy.h |5 +- lib/librte_pmd_e1000/e1000/e1000_regs.h |2 +- lib/librte_pmd_e1000/e1000/e1000_vf.c |4 +- lib/librte_pmd_e1000/e1000/e1000_vf.h |2 +- lib/librte_pmd_ixgbe/ixgbe/README |2 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_82598.c| 136 -- lib/librte_pmd_ixgbe/ixgbe/ixgbe_82598.h|3 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_82599.c| 506 +++--- lib/librte_pmd_ixgbe/ixgbe/ixgbe_82599.h| 12 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_api.c | 54 ++- lib/librte_pmd_ixgbe/ixgbe/ixgbe_api.h | 22 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_common.c | 652 +++ lib/librte_pmd_ixgbe/ixgbe/ixgbe_common.h | 22 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_dcb.c |4 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_dcb.h |2 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_dcb_82598.c|2 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_dcb_82598.h|2 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_dcb_82599.c| 11 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_dcb_82599.h|5 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_mbx.c | 26 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_mbx.h |2 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_osdep.h|7 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_phy.c | 570 +--- lib/librte_pmd_ixgbe/ixgbe/ixgbe_phy.h | 23 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_type.h | 230 +++-- lib/librte_pmd_ixgbe/ixgbe/ixgbe_vf.c | 133 +++-- lib/librte_pmd_ixgbe/ixgbe/ixgbe_vf.h | 10 +- lib/librte_pmd_ixgbe/ixgbe/ixgbe_x540.c | 131 +++-- lib/librte_pmd_ixgbe/ixgbe/ixgbe_x540.h |8 +- lib/librte_pmd_ixgbe/ixgbe_ethdev.c |2 +- lib/librte_pmd_ixgbe/ixgbe_fdir.c |5 +- 64 files changed, 2178 insertions(+), 1217 deletions(-)
[dpdk-dev] [PATCH v2 0/3] Support setting TX rate for queue and VF
Acked-by: Jijiang Liu -Original Message- From: Ouyang, Changchun Sent: Monday, May 26, 2014 3:45 PM To: dev at dpdk.org Cc: Ouyang, Changchun Subject: [PATCH v2 0/3] Support setting TX rate for queue and VF This patch v2 fixes some errors and warnings reported by checkpatch.pl. This patch series also contain the 3 items: 1. Add API to support setting TX rate for a queue or a VF. 2. Implement the functionality of setting TX rate for queue or VF in IXGBE PMD. 3. Add commands in testpmd to test the functionality of setting TX rate for queue or VF. Ouyang Changchun (3): Add API to support set TX rate for a queue and VF. Implement the functionality of setting TX rate for queue or VF in IXGBE PMD. Add commands in testpmd to test the functionality of setting TX rate for queue or VF. app/test-pmd/cmdline.c | 159 +++- app/test-pmd/config.c | 47 +++ app/test-pmd/testpmd.h | 3 + lib/librte_ether/rte_ethdev.c | 71 lib/librte_ether/rte_ethdev.h | 51 lib/librte_pmd_ixgbe/ixgbe_ethdev.c | 122 +++ lib/librte_pmd_ixgbe/ixgbe_ethdev.h | 13 ++- 7 files changed, 462 insertions(+), 4 deletions(-) -- 1.9.0
[dpdk-dev] [PATCH v2 0/4] Link Bonding Library
Acked-by: Pablo de Lara Guarch > -Original Message- > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of > declan.doherty at intel.com > Sent: Wednesday, June 04, 2014 4:18 PM > To: dev at dpdk.org; dev at dpdk.org > Subject: [dpdk-dev] [PATCH v2 0/4] Link Bonding Library > > From: Declan Doherty > > v2 patch additions, > fix for tx burst broadcast, incrementing the reference count on each mbuf by > the number of slaves - 1 add/remove slave behavior chnange to fix primary > slave port assignment patchcheck code fixes > > Initial release of Link Bonding Library (lib/librte_bond) with support for > bonding modes : > 0 - Round Robin > 1 - Active Backup > 2 - Balance l2 / l23 / l34 > 3 - Broadcast > > patches split: > 1 - library + makefile changes > 2 - Unit test suite, including code to generate packet bursts for > testing rx and tx functionality of bonded device and a > virtual/stubbed out ethdev for use as slave ethdev in testing > 3 - Link bonding integration into testpmd, including : > - Includes the ability to create new bonded devices. > - Add /remove bonding slave devices. > - Interogate bonded device stats/configuration > - Change bonding modes and select balance transmit polices > 4 - Add Link Bonding Library to Doxygen > > > app/test-pmd/cmdline.c| 570 ++ > app/test-pmd/config.c |4 +- > app/test-pmd/parameters.c |4 +- > app/test-pmd/testpmd.c| 37 +- > app/test-pmd/testpmd.h|2 + > app/test/Makefile |3 + > app/test/commands.c |3 + > app/test/packet_burst_generator.c | 289 +++ > app/test/packet_burst_generator.h | 78 + > app/test/test.h |1 + > app/test/test_link_bonding.c | 3943 > + > app/test/virtual_pmd.c| 574 ++ > app/test/virtual_pmd.h| 74 + > config/common_bsdapp |5 + > config/common_linuxapp|5 + > doc/doxy-api-index.md |1 + > doc/doxy-api.conf |1 + > lib/Makefile |1 + > lib/librte_bond/Makefile | 28 + > lib/librte_bond/rte_bond.c| 1682 > lib/librte_bond/rte_bond.h| 228 +++ > mk/rte.app.mk |5 + > 22 files changed, 7531 insertions(+), 7 deletions(-) create mode 100644 > app/test/packet_burst_generator.c create mode 100644 > app/test/packet_burst_generator.h create mode 100644 > app/test/test_link_bonding.c create mode 100644 app/test/virtual_pmd.c > create mode 100644 app/test/virtual_pmd.h create mode 100644 > lib/librte_bond/Makefile create mode 100644 lib/librte_bond/rte_bond.c > create mode 100644 lib/librte_bond/rte_bond.h > > -- > 1.8.5.3
[dpdk-dev] [PATCH v2 00/27] Add i40e PMD support
Acked-by: Heqing Zhu -Original Message- From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Helin Zhang Sent: Thursday, June 05, 2014 1:09 PM To: dev at dpdk.org Subject: [dpdk-dev] [PATCH v2 00/27] Add i40e PMD support Summary: The 2nd version of series of patches are to add i40e PMD support. It contains the updated basic shared code, and some other enhancements. It adds the support of the latest version of firmware. * Add new PMD driver of i40e in the folder of librte_pmd_i40e * Add some neccessary definitions, changes in rte_mbuf.h and eth_dev * Add new configurations for i40e * Add or modifiy makefiles to support i40e compilation * Add neccessary changes in ixgbe, e1000 and vmxnet3 PMD, as hash flags has been enlarged from 16 bits to 64 bits to support i40e * Add neccessary changes in example applications and testpmd to use ETH_RSS_IP to replace all IP hash flags, as i40e introduced more hash flags. * Add command in testpmd for port based vlan insertion offload testing * Add neccessary changes in eth_dev to support configuring maximum packet length of less than 1518 * Add two sys files in igb_uio to support enabling/disabling 'Extended Tag' and resetting 'Max Read Request Size', as it has big impacts on i40e performance * Add neccessary changes in pci to read/write the above two sys files during probing PCI Features/enhancements to be implemented later: * Set link speed, and physically up/down * Double VLAN support, flow director, VMDq and DCB * VLAN insertion/stripping, RSS in VF Validation Summary: This patch includes 27 files, and has been tested by Intel(R). Please see test environment information as the following: * Fedora 20 x86_64 * Linux Kernel 3.11.10-301 * GCC 4.8.2 * Intel Xeon CPU E5-2680 v2 @ 2.80GHz We verified this patch with two NICs (i40e 4x10 SFP+, i40e 2x40G), and performed basic function test on PF and VF. All cases are passed. Total cases Passed Failed 138 138 0 Signed-off-by: Helin Zhang Signed-off-by: Jing Chen Acked-by: Cunming Liang Acked-by: Jijiang Liu Acked-by: Jingjing Wu Tested-by: Waterman Cao --- Helin Zhang (27): i40e: add basic shared code i40e: add PMD source files pci: add macros and pci device IDs to support i40e igb_uio: add i40e support mbuf: add new packet flags for i40e ethdev: add i40e support ethdev: support setting maximum packet length to less than 1518 vmxnet3: enlarge the hash flags of RSS to 64 bits ixgbe: enlarge the hash flags of RSS to 64 bits igb: enlarge the hash flags of RSS to 64 bits mk: add i40e support config: add configurations for i40e app/test-pmd: support displaying 32 bytes RX descriptors app/test-pmd: add command of 'tx_vlan set pvid port_id vlan_id (on|off)' app/testpmd: enlarge the hash flags of RSS to 64 bits app/test-pmd: add L3 packet type in offload flags examples/dpdk_qat: use ETH_RSS_IP to replace IP hash flags of RSS examples/ip_reassembly: use ETH_RSS_IP to replace IP hash flags of RSS examples/l3fwd-power: use ETH_RSS_IP to replace IP hash flags of RSS examples/l3fwd-vf: use ETH_RSS_IP to replace IP hash flags of RSS examples/l3fwd: use ETH_RSS_IP to replace IP hash flags of RSS examples/load_balancer: use ETH_RSS_IP to replace IP hash flags of RSS examples/multi_process: use ETH_RSS_IP to replace IP hash flags of RSS examples/qos_meter: use ETH_RSS_IP to replace IP hash flags of RSS igb_uio: add sys files to read/write specific bits in pci config space pci: support reading/writing sys files of 'extended_tag' and 'max_read_request_size' config: add configurations for enabling 'Extended Tag' or resetting 'Max Read Request Size' app/test-pmd/cmdline.c|65 +- app/test-pmd/config.c |45 +- app/test-pmd/csumonly.c | 2 + app/test-pmd/parameters.c | 5 +- app/test-pmd/testpmd.c| 2 +- app/test-pmd/testpmd.h| 4 +- config/common_bsdapp |20 + config/common_linuxapp|30 + examples/dpdk_qat/main.c | 2 +- examples/ip_reassembly/main.c | 2 +- examples/l3fwd-power/main.c | 2 +- examples/l3fwd-vf/main.c | 2 +- examples/l3fwd/main.c | 2 +- examples/load_balancer/init.c | 2 +- examples/multi_process/symmetric_mp/main.c| 2 +- examples/qos_meter/main.c | 2 +- lib/Makefile | 1 + lib/librte_eal/common/include/rte_pci_dev_ids.h |44 + lib/librte_eal/linuxapp/eal/eal_pci.c |99 + lib/librte_eal/linuxapp/igb_uio/igb_uio.c | 107 + lib/librte_ether/rte_ethdev.c |25 +- lib/li
[dpdk-dev] [v2 00/23] Packet Framework
Acked-by: Pablo de Lara Guarch > -Original Message- > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Cristian Dumitrescu > Sent: Wednesday, June 04, 2014 7:08 PM > To: dev at dpdk.org > Subject: [dpdk-dev] [v2 00/23] Packet Framework > > (Version 2 changes are exclusively style changes (checkpatch.pl) and patch > consolidation, no functional change) > > Intel DPDK Packet Framework provides a standard methodology (logically > similar to OpenFlow) for rapid development of complex packet processing > pipelines out of ports, tables and actions. > > A pipeline is constructed by connecting its input ports to its output ports > through a chain of lookup tables. As result of lookup operation into the > current table, one of the table entries (or the default table entry, in case > of > lookup miss) is identified to provide the actions to be executed on the > current packet and the associated action meta-data. The behavior of user > actions is defined through the configurable table action handler, while the > reserved actions define the next hop for the current packet (either another > table, an output port or packet drop) and are handled transparently by the > framework. > > Three new Intel DPDK libraries are introduced for Packet Framework: > librte_port, librte_table, librte_pipeline. Please check the Intel DPDK > Programmer's Guide for full description of the Packet Framework design. > > Two sample applications are provided for Packet Framework: app/test- > pipeline and examples/ip_pipeline. Please check the Intel Sample Apps Guide > for a detailed description of how these sample apps. > > Cristian Dumitrescu (23): > librte_lpm: rule_is_present > mbuf: meta-data > Packet Framework librte_port: Port API > Packet Framework librte_port: ethdev ports > Packet Framework librte_port: ring ports > Packet Framework librte_port: IPv4 frag port > Packet Framework librte_port: IPv4 reassembly > Packet Framework librte_port: hierarchical scheduler port > Packet Framework librte_port: Source/Sink ports > Packet Framework librte_port: Build infrastructure > Packet Framework librte_table: Table API > Packet Framework librte_table: LPM IPv4 table > Packet Framework librte_table: LPM IPv6 table > Packet Framework librte_table: ACL table > Packet Framework librte_table: Hash tables > Packet Framework librte_table: array table > Packet Framework librte_table: Stub table > Packet Framework librte_table: Build infrastructure > Packet Framework librte_pipeline: Pipeline > librte_cfgfile: interpret config files > Packet Framework performance application > Packet Framework IPv4 pipeline sample app > Packet Framework unit tests > > app/Makefile |1 + > app/test-pipeline/Makefile | 66 + > app/test-pipeline/config.c | 248 +++ > app/test-pipeline/init.c | 295 +++ > app/test-pipeline/main.c | 180 ++ > app/test-pipeline/main.h | 148 ++ > app/test-pipeline/pipeline_acl.c | 278 +++ > app/test-pipeline/pipeline_hash.c | 487 + > app/test-pipeline/pipeline_lpm.c | 196 ++ > app/test-pipeline/pipeline_lpm_ipv6.c | 200 ++ > app/test-pipeline/pipeline_stub.c | 165 ++ > app/test-pipeline/runtime.c| 185 ++ > app/test/Makefile |6 + > app/test/commands.c|4 +- > app/test/test.h|1 + > app/test/test_table.c | 220 +++ > app/test/test_table.h | 204 ++ > app/test/test_table_acl.c | 593 ++ > app/test/test_table_acl.h | 35 + > app/test/test_table_combined.c | 784 > app/test/test_table_combined.h | 55 + > app/test/test_table_pipeline.c | 603 ++ > app/test/test_table_pipeline.h | 35 + > app/test/test_table_ports.c| 224 +++ > app/test/test_table_ports.h| 42 + > app/test/test_table_tables.c | 907 + > app/test/test_table_tables.h | 50 + > config/common_bsdapp | 25 + > config/common_linuxapp | 24 + > doc/doxy-api-index.md | 17 + > doc/doxy-api.conf |3 + > examples/ip_pipeline/Makefile | 67 + > examples/ip_pipeline/cmdline.c | 1976 > > examples/ip_pipeline/config.c | 420 + > examples/ip_pipeline/init.c| 614 +++
[dpdk-dev] [PATCH v2 0/3] Support setting TX rate for queue and VF
On Thu, Jun 05, 2014 at 03:11:46AM +, Ouyang, Changchun wrote: > Hi, Neil > > " but you're implementaiton requires that it be re-implemented for each PMD " > > [Changchun]: Different PMD(corresponding diff NIC) has different register to > set. > It makes sense to me has different implementation to support limit tx rate. > > " Why not just export max tx rates from the PMD and write a generic queuing > libarary to do rate limitation for any PMD?" > > [Changchun]: Just export max tx rate is not enough for customer. > I think if we can leverage HW feature to do it, why need a more complicated > lib to do so? > > Thanks > Changchun > sorry, I reread my initial comment and see I wasn't being at all clear. You actually do have a nice generic layer for applications to call through in rte_ethdev, which is good. What you don't have is a backoff mechanism. That is to say the return of ENOSUP in rte_eth_dev_set_queue_rate_limit means that your queuing library doesnt' work for any device that doesn't support hardware rate limiting (implying none of the virtual devs get to implement this). Don't you think you should implement a generic software queue rate limit layer for the myrriad of devices that cant' do what ixgbe can? Regards Neil > -Original Message- > From: Neil Horman [mailto:nhorman at tuxdriver.com] > Sent: Tuesday, May 27, 2014 6:53 AM > To: Ouyang, Changchun > Cc: dev at dpdk.org > Subject: Re: [dpdk-dev] [PATCH v2 0/3] Support setting TX rate for queue and > VF > > On Mon, May 26, 2014 at 03:45:28PM +0800, Ouyang Changchun wrote: > > This patch v2 fixes some errors and warnings reported by checkpatch.pl. > > > > This patch series also contain the 3 items: > > 1. Add API to support setting TX rate for a queue or a VF. > > 2. Implement the functionality of setting TX rate for queue or VF in IXGBE > > PMD. > > 3. Add commands in testpmd to test the functionality of setting TX rate for > > queue or VF. > > > > Ouyang Changchun (3): > > Add API to support set TX rate for a queue and VF. > > Implement the functionality of setting TX rate for queue or VF in > > IXGBE PMD. > > Add commands in testpmd to test the functionality of setting TX rate > > for queue or VF. > > > > app/test-pmd/cmdline.c | 159 > > +++- > > app/test-pmd/config.c | 47 +++ > > app/test-pmd/testpmd.h | 3 + > > lib/librte_ether/rte_ethdev.c | 71 > > lib/librte_ether/rte_ethdev.h | 51 > > lib/librte_pmd_ixgbe/ixgbe_ethdev.c | 122 +++ > > lib/librte_pmd_ixgbe/ixgbe_ethdev.h | 13 ++- > > 7 files changed, 462 insertions(+), 4 deletions(-) > > > This seems a bit backwards. queue rate limiting is rather a generic > function, that doesn't really need to know any details about the hardware, > save for its maximum tx rate, but you're implementaiton requires that it be > re-implemented for each PMD. Why not just export max tx rates from the PMD > and write a generic queuing libarary to do rate limitation for any PMD? > > Neil > > > -- > > 1.9.0 > > > > >
[dpdk-dev] [PATCH v2 0/4] Link Bonding Library
On Wed, Jun 04, 2014 at 04:18:01PM +0100, declan.doherty at intel.com wrote: > From: Declan Doherty > > v2 patch additions, > fix for tx burst broadcast, incrementing the reference count on each mbuf by > the number of slaves - 1 > add/remove slave behavior chnange to fix primary slave port assignment > patchcheck code fixes > > Initial release of Link Bonding Library (lib/librte_bond) with support for > bonding modes : > 0 - Round Robin > 1 - Active Backup > 2 - Balance l2 / l23 / l34 > 3 - Broadcast > > patches split: > 1 - library + makefile changes > 2 - Unit test suite, including code to generate packet bursts for > testing rx and tx functionality of bonded device and a > virtual/stubbed out ethdev for use as slave ethdev in testing > 3 - Link bonding integration into testpmd, including : > - Includes the ability to create new bonded devices. > - Add /remove bonding slave devices. > - Interogate bonded device stats/configuration > - Change bonding modes and select balance transmit polices > 4 - Add Link Bonding Library to Doxygen > > > app/test-pmd/cmdline.c| 570 ++ > app/test-pmd/config.c |4 +- > app/test-pmd/parameters.c |4 +- > app/test-pmd/testpmd.c| 37 +- > app/test-pmd/testpmd.h|2 + > app/test/Makefile |3 + > app/test/commands.c |3 + > app/test/packet_burst_generator.c | 289 +++ > app/test/packet_burst_generator.h | 78 + > app/test/test.h |1 + > app/test/test_link_bonding.c | 3943 > + > app/test/virtual_pmd.c| 574 ++ > app/test/virtual_pmd.h| 74 + > config/common_bsdapp |5 + > config/common_linuxapp|5 + > doc/doxy-api-index.md |1 + > doc/doxy-api.conf |1 + > lib/Makefile |1 + > lib/librte_bond/Makefile | 28 + > lib/librte_bond/rte_bond.c| 1682 > lib/librte_bond/rte_bond.h| 228 +++ > mk/rte.app.mk |5 + > 22 files changed, 7531 insertions(+), 7 deletions(-) > create mode 100644 app/test/packet_burst_generator.c > create mode 100644 app/test/packet_burst_generator.h > create mode 100644 app/test/test_link_bonding.c > create mode 100644 app/test/virtual_pmd.c > create mode 100644 app/test/virtual_pmd.h > create mode 100644 lib/librte_bond/Makefile > create mode 100644 lib/librte_bond/rte_bond.c > create mode 100644 lib/librte_bond/rte_bond.h > > -- > 1.8.5.3 > > This doesn't address any of the comments I made previously. Neil
[dpdk-dev] vETH is DOWN by default
Hi, We are running the sample DPDK KNI application on a single physical port. We are able to see the vETH interface getting created via IFCONFIG. However the vETH interface is DOWN by default despite the physical interface being UP and RUNNING. Only when we configured IP address via IFCONFIG, the vETH came UP. Any reason for this ? Regards, Ajith
[dpdk-dev] [v2 00/23] Packet Framework
Tested-by: Waterman Cao Totally this patch is composed of 24 files including cover letter, and has been tested by Intel. We verified packet framework patch with ip pipeline example and unit test, all cases passed. Please see test result as the following: test_flow_management Passed test_frame_sizes Passed test_incremental_ipPassed test_route_management Passed test_hash_tables Passed test_lpm_table Passed test_none_tablePassed Test environment: Fedora 20, Linux Kernel 3.13.6-200, GCC 4.8.2, Intel Xeon processor E5-2680 v2, with Intel Niantic 82599.
[dpdk-dev] [PATCH v2]xen:reserve memory at installing dom0_mm kernel module
Tested-by: Waterman Cao This patch has been tested by Intel. Test Environment: Fedora 16 with kernel 3.10.0 and Red Hat 5.0 with kernel 2.6.32 pvops. Linux kernel has the following features enabled: Huge page support, UIO, HPET ,Xen support. Enable PCI passthrough by add intel_iommu=on iommu=pt in kernel grub XEN should be version 4.0.1 on Red Hat and 4.2.3 on Fedora 16.
[dpdk-dev] [PATCH v2 0/3] Support setting TX rate for queue and VF
Tested-by: Waterman Cao This patch bug fix has been tested by Intel. Please see information as the following: Fedora 20 x86_64, Linux Kernel 3.13.9-200, GCC 4.8.2 Intel Xeon CPU E5-2680 v2 @ 2.80GHz NIC: Intel Niantic 82599, Intel i350, Intel 82580 and Intel 82576
[dpdk-dev] [PATCH 04/29] mbuf: added offset of packet meta-data in the packet buffer just after mbuf
Hi,
In order to minimize the number of iterations, Ivan and I had an offline
discussion on this.
Ivan is concerned with the way the mbuf cloning/scatter-gather feature is
currently implemented in DPDK, and not with this particular patch.
We agreed to take the discussion on cloning/scatter-gather implementation
during the DPDK 1.8 time-frame, at this belongs to the mbuf refresh discussion.
The mbuf library is not just the format of the mbuf data structure, it also
includes all the features associated with the mbuf, as: accessing mbuf fields
through get/set methods, buffer chaining, cloning/scatter-gather, enabling HW
offloads through mbuf flags and fields, etc.
Regards,
Cristian
-Original Message-
From: Ivan Boule [mailto:ivan.bo...@6wind.com]
Sent: Monday, June 2, 2014 1:24 PM
To: Dumitrescu, Cristian; dev at dpdk.org
Subject: Re: [dpdk-dev] [PATCH 04/29] mbuf: added offset of packet meta-data in
the packet buffer just after mbuf
Hi Cristian,
I agree with you, the natural way to store application metadata into
mbufs consists in putting it right after the rte_mbuf data structure.
This can be simply implemented this way:
struct metadata {
...
};
struct app_mbuf {
struct rte_mbuf mbuf;
struct metadata mtdt;
};
With such a representation, the application initializes the "buf_addr"
field of each mbuf pointed to by a (struct app_mbuf *)amb pointer as:
amb->mbuf.buf_addr = ((char *amb) + sizeof(struct app_mbuf));
But such a natural programming approach breaks the assumptions of the
macros RTE_MBUF_FROM_BADDR, RTE_MBUF_TO_BADDR, RTE_MBUF_INDIRECT, and
RTE_MBUF_DIRECT.
For instance, in the function __rte_pktmbuf_prefree_seg(m), after the line
struct rte_mbuf *md = RTE_MBUF_FROM_BADDR(m->buf_addr);
"md" is always different from "m", and thus systematically (and most of
the time wrongly) considered as an indirect mbuf.
In the same way, the operations done by the function rte_pktmbuf_detach(m)
void *buf = RTE_MBUF_TO_BADDR(m);
m->buf_addr = buf;
does not set buf_addr to its [dafault] correct value.
To summarize, adding metadata after the rte_mbuf data structure is
incompatible with the packet cloning feature behind the [wrongly named]
RTE_MBUF_SCATTER_GATHER configuration option.
By the way, you suggest to use the headroom to also store packet metadata.
But, then, how does an application can store both types of information
in a given mbuf at the same time, when the actual length of network
headers in a mbuf is variable, as it depends on the protocol path
followed by the packet in a networking stack (tunnels, etc)?
Regards,
Ivan
On 05/30/2014 12:28 AM, Dumitrescu, Cristian wrote:
> Hi Ivan,
>
> I hate to disagree with you :), but no, we do not break the scatter-gather
> feature. We actually implemented the Packet Framework IPv4 fragmentation and
> reassembly ports to validate exactly this.
>
> Maybe the confusion comes from the RTE_MBUF_TO_BADDR macro. This macro only
> works (and it is only used) for direct mbufs, so probably the correct name
> for this macro should be RTE_MBUF_DIRECT_TO_BADDR, as it cannot work (and it
> is not used) for indirect mbufs.
>
> I am describing the rationale behind meta-data design below, sorry for the
> long description that looks like a design document ...
>
> Quick recap:
> - mbuf->buf_addr points to the first address where a byte of data for the
> current segment _can_ be placed
> - direct mbuf: the packet descriptor (mbuf) sits in the same mempool buffer
> with the packet itself, so mbuf->buf_addr = mbuf + 1;
> - indirect mbuf: the packet descriptor is located in a different mempool
> buffer than the packet itself, so mbuf->buf_addr != mbuf + 1;
> - Regardless of the mbuf type, it is not necessarily true that the first byte
> of data is located at mbuf->buf_addr, as we save a headroom (configurable, by
> default initially of CONFIG_RTE_PKTMBUF_HEADROOM = 128 bytes) at the start of
> the data buffer (mbuf->buf_addr) for prepending packet headers ( ... or, why
> not, meta-data!). The location of the first real data byte is
> mbuf->pkt.data, which is variable, as opposed to mbuf->buf_addr, which does
> not change.
>
> Packet meta-data rationale:
> - I think we both agree that we need to be able to store packet meta-data in
> the packet buffer. The packet buffer regions where meta-data could be stored
> can only be the in the headroom or in the tailroom, which are both managed by
> the application and both go up and down during the packet lifetime.
>
> - To me, packet-metadata is part of the packet descriptor: mbuf is the
> mandatory & standard part of the packet descriptor, while meta-data is the
> optional & non-standard (per application) part of the packet descriptor.
> Therefore, it makes sense to put the meta-data immediately after mbuf, but
> this is not mandatory.
>
> - The zero-size field called meta-data is really just an offset: it points to
> the first buffer location where
[dpdk-dev] [PATCH 0/4] librte_pmd_virtio :Fix: virtio_pci.h non-existent virtio feature bit-flag tested
This series addresses an issue with librte_pmd_virtio where the offset to the virtio device specific header may be incorrect depending on whether MSI-X has been enabled or not. If MSI-X is configured the device specific header is placed at byte offset 24 relative to the IO base address. If MSI-X is not configured the device specific header is placed at byte offset 20. The following macro defined in virtio_pci.h is used to test the presence of the MSI-X header and determine the correct offset: #define VIRTIO_PCI_CONFIG(hw) (((hw)->guest_features & VIRTIO_PCI_FLAG_MSIX) ? 24 : 20) However, VIRTIO_PCI_FLAG_MSIX is not a guest_feature nor is it part of the Virtio Specification and resolves to the VIRTIO_NET_F_MAC feature as both are #defined as 0x20. VIRTIO_PCI_FLAG_MSIX or similar flag should instead be set by the kernel driver allocating resources and passed to user space for testing. i.e. #define VIRTIO_PCI_CONFIG(hw) (((hw)->intr_mode & IGBUIO_MSIX_INTR_MODE) ? 24 : 20) To enable this testing of interrupt mode, this series allows for the kernel driver(igb_uio) to place the configured interrupt mode into a sysfs entry. sysfs is then parsed by eal_pci to determine the configured mode, which allows all user space devices to correctly determine the interrupt mode, including virtio_ethdev. This series should be applied prior to Anatoly Burakov's [VFIO] Add VFIO support to DPDK series Alan Carew (4): igb_uio: Add interrupt_mode sysfs entry for igb_uio devices eal_pci: Add interrupt mode to rte_pci_device and parsing to eal_pci FreeBSD: Adds the equivalent interrupt mode setting and parsing virtio: Fixes the VIRTIO_PCI_CONFIG macro to use the correct offset to the Virtio header lib/librte_eal/bsdapp/eal/eal_pci.c| 44 ++ lib/librte_eal/bsdapp/nic_uio/nic_uio.c| 14 +++ lib/librte_eal/common/Makefile |1 + lib/librte_eal/common/include/rte_pci.h|2 + .../common/include/rte_pci_dev_feature_defs.h | 85 .../common/include/rte_pci_dev_features.h | 70 lib/librte_eal/linuxapp/eal/eal_pci.c | 78 ++ lib/librte_eal/linuxapp/igb_uio/igb_uio.c | 48 +--- lib/librte_pmd_virtio/virtio_ethdev.c |1 + lib/librte_pmd_virtio/virtio_pci.h |4 +- 10 files changed, 336 insertions(+), 11 deletions(-) create mode 100755 lib/librte_eal/common/include/rte_pci_dev_feature_defs.h create mode 100755 lib/librte_eal/common/include/rte_pci_dev_features.h
[dpdk-dev] [PATCH 1/4] [PATCH 1/4] igb_uio: Add interrupt_mode sysfs entry for igb_uio devices
This patch adds an "interrupt_mode" sysfs entry for igb_uio devices,
allowing userspace eal_pci to track which interrupt mode has been enabled in
kernel space.
The sysfs entry can be inspected via
/sys/bus/pci/devices/
---
.../common/include/rte_pci_dev_feature_defs.h | 85
.../common/include/rte_pci_dev_features.h | 70
lib/librte_eal/linuxapp/igb_uio/igb_uio.c | 48 +---
3 files changed, 193 insertions(+), 10 deletions(-)
create mode 100755 lib/librte_eal/common/include/rte_pci_dev_feature_defs.h
create mode 100755 lib/librte_eal/common/include/rte_pci_dev_features.h
diff --git a/lib/librte_eal/common/include/rte_pci_dev_feature_defs.h
b/lib/librte_eal/common/include/rte_pci_dev_feature_defs.h
new file mode 100755
index 000..d23ed7d
--- /dev/null
+++ b/lib/librte_eal/common/include/rte_pci_dev_feature_defs.h
@@ -0,0 +1,85 @@
+/*-
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ * The full GNU General Public License is included in this distribution
+ * in the file called LICENSE.GPL.
+ *
+ * Contact Information:
+ * Intel Corporation
+ *
+ * BSD LICENSE
+ *
+ * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ * * Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * 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 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.
+ *
+ */
+
+#ifndef _RTE_PCI_DEV_DEFS_H_
+#define _RTE_PCI_DEV_DEFS_H_
+
+#define RTE_PCI_DEV_FEATURE_INTR_MODE "interrupt_mode"
+
+#define INTR_NAME_LEN 10
+
+#define IGBUIO_NONE_INTR_NAME "none"
+#define IGBUIO_LEGACY_INTR_NAME "legacy"
+#define IGBUIO_MSI_INTR_NAME "msi"
+#define IGBUIO_MSIX_INTR_NAME "msix"
+
+
+#define INTR_MODE(id, mode_name) \
+.mode = (id), .name = (mode_name)
+
+/* interrupt mode */
+enum igbuio_intr_mode {
+ IGBUIO_NONE_INTR_MODE = 0,
+ IGBUIO_LEGACY_INTR_MODE,
+ IGBUIO_MSI_INTR_MODE,
+ IGBUIO_MSIX_INTR_MODE,
+ IGBUIO_INTR_MODE_MAX
+};
+
+#endif /* _RTE_PCI_DEV_DEFS_H_ */
diff --git a/lib/librte_eal/common/include/rte_pci_dev_features.h
b/lib/librte_eal/common/include/rte_pci_dev_features.h
new file mode 100755
index 000..a45c056
--- /dev/null
+++ b/lib/librte_eal/common/include/rte_pci_dev_features.h
@@ -0,0 +1,70 @@
+/*-
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ *
[dpdk-dev] [PATCH 2/4] [PATCH 2/4] eal_pci: Add interrupt mode to rte_pci_device and parsing to eal_pci
This patch adds a shared enum between user and kernel space to rte_pci_device.
The value of intr_mode is parsed by eal_pci during pci_uio_map_resource
Signed-off-by: Alan Carew
---
lib/librte_eal/common/Makefile |1 +
lib/librte_eal/common/include/rte_pci.h |2 +
lib/librte_eal/linuxapp/eal/eal_pci.c | 78 +++
3 files changed, 81 insertions(+), 0 deletions(-)
diff --git a/lib/librte_eal/common/Makefile b/lib/librte_eal/common/Makefile
index 0016fc5..52d46f3 100644
--- a/lib/librte_eal/common/Makefile
+++ b/lib/librte_eal/common/Makefile
@@ -35,6 +35,7 @@ INC := rte_atomic.h rte_branch_prediction.h rte_byteorder.h
rte_common.h
INC += rte_cycles.h rte_debug.h rte_eal.h rte_errno.h rte_launch.h rte_lcore.h
INC += rte_log.h rte_memcpy.h rte_memory.h rte_memzone.h rte_pci.h
INC += rte_pci_dev_ids.h rte_per_lcore.h rte_prefetch.h rte_random.h
+INC += rte_pci_dev_feature_defs.h rte_pci_dev_features.h
INC += rte_rwlock.h rte_spinlock.h rte_tailq.h rte_interrupts.h rte_alarm.h
INC += rte_string_fns.h rte_cpuflags.h rte_version.h rte_tailq_elem.h
INC += rte_eal_memconfig.h rte_malloc_heap.h
diff --git a/lib/librte_eal/common/include/rte_pci.h
b/lib/librte_eal/common/include/rte_pci.h
index c793773..5883cd9 100644
--- a/lib/librte_eal/common/include/rte_pci.h
+++ b/lib/librte_eal/common/include/rte_pci.h
@@ -81,6 +81,7 @@ extern "C" {
#include
#include
#include
+#include
TAILQ_HEAD(pci_device_list, rte_pci_device); /**< PCI devices in D-linked Q. */
TAILQ_HEAD(pci_driver_list, rte_pci_driver); /**< PCI drivers in D-linked Q. */
@@ -150,6 +151,7 @@ struct rte_pci_device {
const struct rte_pci_driver *driver;/**< Associated driver */
uint16_t max_vfs; /**< sriov enable if not zero */
int numa_node; /**< NUMA node connection */
+ enum igbuio_intr_mode intr_mode;/**< Interrupt mode */
struct rte_devargs *devargs;/**< Device user arguments */
};
diff --git a/lib/librte_eal/linuxapp/eal/eal_pci.c
b/lib/librte_eal/linuxapp/eal/eal_pci.c
index ac2c1fe..7dba446 100644
--- a/lib/librte_eal/linuxapp/eal/eal_pci.c
+++ b/lib/librte_eal/linuxapp/eal/eal_pci.c
@@ -67,6 +67,7 @@
#include
#include "rte_pci_dev_ids.h"
+#include "rte_pci_dev_feature_defs.h"
#include "eal_filesystem.h"
#include "eal_private.h"
@@ -89,6 +90,17 @@ struct uio_map {
uint64_t phaddr;
};
+struct rte_pci_dev_intr_mode {
+ enum igbuio_intr_mode mode;
+ const char *name;
+};
+
+/* Table of interrupt modes */
+const struct rte_pci_dev_intr_mode interrupt_modes[] = {
+#define RTE_PCI_DEV_INTR_MODE(id, mode_name) {INTR_MODE(id, mode_name)},
+#include
+};
+
/*
* For multi-process we need to reproduce all PCI mappings in secondary
* processes, so save them in a tailq.
@@ -106,6 +118,7 @@ TAILQ_HEAD(uio_res_list, uio_resource);
static struct uio_res_list *uio_res_list = NULL;
static int pci_parse_sysfs_value(const char *filename, uint64_t *val);
+static int pci_parse_sysfs_intr_mode(const char *filename, struct
rte_pci_device *dev);
/* unbind kernel driver for this device */
static int
@@ -400,6 +413,7 @@ pci_uio_map_resource(struct rte_pci_device *dev)
int i, j;
char dirname[PATH_MAX];
char devname[PATH_MAX]; /* contains the /dev/uioX */
+char filename[PATH_MAX];
void *mapaddr;
int uio_num;
uint64_t phaddr;
@@ -435,6 +449,18 @@ pci_uio_map_resource(struct rte_pci_device *dev)
}
dev->intr_handle.type = RTE_INTR_HANDLE_UIO;
+ rte_snprintf(filename, sizeof(filename),
+ SYSFS_PCI_DEVICES "/" PCI_PRI_FMT
+ "/"RTE_PCI_DEV_FEATURE_INTR_MODE,
+ loc->domain, loc->bus, loc->devid, loc->function);
+
+/* Get the kernel configured interrupt mode */
+ if (pci_parse_sysfs_intr_mode(filename, dev) < 0) {
+ RTE_LOG(ERR, EAL, "%s(): cannot determine interrupt_mode\n",
+ __func__);
+ return -1;
+ }
+
/* allocate the mapping details for secondary processes*/
if ((uio_res = rte_zmalloc("UIO_RES", sizeof (*uio_res), 0)) == NULL) {
RTE_LOG(ERR, EAL,
@@ -591,6 +617,58 @@ pci_parse_sysfs_value(const char *filename, uint64_t *val)
return 0;
}
+/*
+ * Parse a sysfs file containing a string
+ */
+static int
+pci_parse_sysfs_string(const char *filename, char *result_str, int max_len)
+{
+ FILE *f;
+ size_t len;
+
+ f = fopen(filename, "r");
+ if (f == NULL) {
+ RTE_LOG(ERR, EAL, "%s(): cannot open sysfs value %s\n",
+ __func__, filename);
+ return -1;
+ }
+ if (fgets(result_str, max_len, f) == NULL) {
+ RTE_LOG(ERR, EAL, "%s(): cannot read sysfs value %s\n",
+ __func__, filename);
+
[dpdk-dev] [PATCH 3/4] [PATCH 3/4] FreeBSD: Adds the equivalent interrupt mode setting and parsing
This patch adds the equivalent functionality to FreeBSD as with patches 1 and 2
Signed-off-by: Alan Carew
---
lib/librte_eal/bsdapp/eal/eal_pci.c | 45 +++
lib/librte_eal/bsdapp/nic_uio/nic_uio.c | 14 +
2 files changed, 59 insertions(+), 0 deletions(-)
diff --git a/lib/librte_eal/bsdapp/eal/eal_pci.c
b/lib/librte_eal/bsdapp/eal/eal_pci.c
index 94ae461..7c270bb 100644
--- a/lib/librte_eal/bsdapp/eal/eal_pci.c
+++ b/lib/librte_eal/bsdapp/eal/eal_pci.c
@@ -50,6 +50,7 @@
#include
#include
#include
+#include
#include
#include
@@ -108,6 +109,44 @@ TAILQ_HEAD(uio_res_list, uio_resource);
static struct uio_res_list *uio_res_list = NULL;
+struct rte_pci_dev_intr_mode {
+ enum igbuio_intr_mode mode;
+ const char *name;
+};
+
+/* Table of interrupt modes */
+const struct rte_pci_dev_intr_mode interrupt_modes[] = {
+#define RTE_PCI_DEV_INTR_MODE(id, mode_name) {INTR_MODE (id, mode_name)},
+#include
+};
+
+/*
+ * Determine the kernel configured interrupt mode
+ */
+static int
+pci_parse_intr_mode(struct rte_pci_device *dev)
+{
+ char intr_mode[INTR_NAME_LEN];
+ unsigned int i, num_intr_modes = RTE_DIM(interrupt_modes);
+ size_t sysctl_size = sizeof(intr_mode);
+
+ if (sysctlbyname("hw.nic_uio."RTE_PCI_DEV_FEATURE_INTR_MODE, &intr_mode,
+ &sysctl_size, NULL, 0) < 0) {
+ RTE_LOG(ERR, EAL,
+ "%s(): cannot get sysctlbyname: hw.nic_uio.intr_mode\n",
+__func__);
+ return (-1);
+ }
+ for (i = 0; i < num_intr_modes; i++) {
+ if (!strncmp(intr_mode, interrupt_modes[i].name,
INTR_NAME_LEN)) {
+ dev->intr_mode = interrupt_modes[i].mode;
+ return 0;
+ }
+ }
+ return -1;
+
+}
+
/* unbind kernel driver for this device */
static int
pci_unbind_kernel_driver(struct rte_pci_device *dev)
@@ -220,6 +259,12 @@ pci_uio_map_resource(struct rte_pci_device *dev)
return -1;
}
+ if (pci_parse_intr_mode(dev) < 0) {
+ RTE_LOG(ERR, EAL,
+ "%s(): unable to determine interrupt mode\n",
__func__);
+ return (-1);
+ }
+
/* save fd if in primary process */
dev->intr_handle.fd = open(devname, O_RDWR);
if (dev->intr_handle.fd < 0) {
diff --git a/lib/librte_eal/bsdapp/nic_uio/nic_uio.c
b/lib/librte_eal/bsdapp/nic_uio/nic_uio.c
index c10e9aa..0e17d63 100644
--- a/lib/librte_eal/bsdapp/nic_uio/nic_uio.c
+++ b/lib/librte_eal/bsdapp/nic_uio/nic_uio.c
@@ -51,6 +51,10 @@ __FBSDID("$FreeBSD$");
#include
#include
#include
+#include
+
+#include
+#include
#define MAX_BARS (PCIR_MAX_BAR_0 + 1)
@@ -116,6 +120,16 @@ const struct device devices[] = {
};
#define NUM_DEVICES (sizeof(devices)/sizeof(devices[0]))
+static char nic_uio_intr_mode[] = {IGBUIO_NONE_INTR_NAME};
+
+TUNABLE_STR("hw.nic_uio."RTE_PCI_DEV_FEATURE_INTR_MODE, nic_uio_intr_mode,
sizeof(nic_uio_intr_mode));
+
+static SYSCTL_NODE(_hw, OID_AUTO, nic_uio, CTLFLAG_RD, 0, "nic_uio");
+
+SYSCTL_STRING(_hw_nic_uio, OID_AUTO, interrupt_mode, CTLFLAG_RW,
+ &nic_uio_intr_mode, sizeof(nic_uio_intr_mode),
+ "Configured interrupt mode");
+
static devclass_t nic_uio_devclass;
--
1.7.0.7
[dpdk-dev] [PATCH 4/4] [PATCH 4/4] virtio: Fixes the VIRTIO_PCI_CONFIG macro to use the correct offset to the Virtio header
This final patch address the issue of not being able to determine the correct
offet when MSI-X is disabled.
Signed-off-by: Alan Carew
---
lib/librte_pmd_virtio/virtio_ethdev.c |1 +
lib/librte_pmd_virtio/virtio_pci.h|4 +++-
2 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/lib/librte_pmd_virtio/virtio_ethdev.c
b/lib/librte_pmd_virtio/virtio_ethdev.c
index c2b4dfb..819e7d7 100644
--- a/lib/librte_pmd_virtio/virtio_ethdev.c
+++ b/lib/librte_pmd_virtio/virtio_ethdev.c
@@ -719,6 +719,7 @@ eth_virtio_dev_init(__rte_unused struct eth_driver *eth_drv,
}
#endif
hw->io_base = (uint32_t)(uintptr_t)pci_dev->mem_resource[0].addr;
+ hw->intr_mode = pci_dev->intr_mode;
/* Reset the device although not necessary at startup */
vtpci_reset(hw);
diff --git a/lib/librte_pmd_virtio/virtio_pci.h
b/lib/librte_pmd_virtio/virtio_pci.h
index 96443c7..7b3d0ef 100644
--- a/lib/librte_pmd_virtio/virtio_pci.h
+++ b/lib/librte_pmd_virtio/virtio_pci.h
@@ -44,6 +44,7 @@
#endif
#include
+#include
struct virtqueue;
@@ -177,6 +178,7 @@ struct virtio_hw {
uint16_tsubsystem_device_id;
uint16_tsubsystem_vendor_id;
uint8_t revision_id;
+ enumigbuio_intr_mode intr_mode;
uint8_t mac_addr[ETHER_ADDR_LEN];
int adapter_stopped;
struct rte_eth_stats eth_stats;
@@ -201,7 +203,7 @@ struct virtio_net_config {
* The remaining space is defined by each driver as the per-driver
* configuration space.
*/
-#define VIRTIO_PCI_CONFIG(hw) (((hw)->guest_features & VIRTIO_PCI_FLAG_MSIX) ?
24 : 20)
+#define VIRTIO_PCI_CONFIG(hw) (((hw)->intr_mode == IGBUIO_MSIX_INTR_MODE) ? 24
: 20)
/*
* How many bits to shift physical queue address written to QUEUE_PFN.
--
1.7.0.7
[dpdk-dev] [v2 00/23] Packet Framework
On 06/04/2014 08:08 PM, Cristian Dumitrescu wrote: > (Version 2 changes are exclusively style changes (checkpatch.pl) and patch > consolidation, no functional change) > > Intel DPDK Packet Framework provides a standard methodology (logically > similar to OpenFlow) for rapid development of complex packet processing > pipelines out of ports, tables and actions. > > A pipeline is constructed by connecting its input ports to its output ports > through a chain of lookup tables. As result of lookup operation into the > current table, one of the table entries (or the default table entry, in case > of lookup miss) is identified to provide the actions to be executed on the > current packet and the associated action meta-data. The behavior of user > actions is defined through the configurable table action handler, while the > reserved actions define the next hop for the current packet (either another > table, an output port or packet drop) and are handled transparently by the > framework. > > Three new Intel DPDK libraries are introduced for Packet Framework: > librte_port, librte_table, librte_pipeline. Please check the Intel DPDK > Programmer's Guide for full description of the Packet Framework design. > > Two sample applications are provided for Packet Framework: app/test-pipeline > and examples/ip_pipeline. Please check the Intel Sample Apps Guide for a > detailed description of how these sample apps. > > Cristian Dumitrescu (23): >librte_lpm: rule_is_present >mbuf: meta-data >Packet Framework librte_port: Port API >Packet Framework librte_port: ethdev ports >Packet Framework librte_port: ring ports >Packet Framework librte_port: IPv4 frag port >Packet Framework librte_port: IPv4 reassembly >Packet Framework librte_port: hierarchical scheduler port >Packet Framework librte_port: Source/Sink ports >Packet Framework librte_port: Build infrastructure >Packet Framework librte_table: Table API >Packet Framework librte_table: LPM IPv4 table >Packet Framework librte_table: LPM IPv6 table >Packet Framework librte_table: ACL table >Packet Framework librte_table: Hash tables >Packet Framework librte_table: array table >Packet Framework librte_table: Stub table >Packet Framework librte_table: Build infrastructure >Packet Framework librte_pipeline: Pipeline >librte_cfgfile: interpret config files >Packet Framework performance application >Packet Framework IPv4 pipeline sample app >Packet Framework unit tests > Acked by: Ivan Boule -- Ivan Boule 6WIND Development Engineer
[dpdk-dev] [PATCH v2 1/4] Link Bonding Library
On Wed, 4 Jun 2014 16:18:02 +0100
declan.doherty at intel.com wrote:
> From: Declan Doherty
>
> - Broadcast TX burst broadcast bug fix
> - Add/remove slave behavior fix
> - Checkpatch fixes
>
> Signed-off-by: Declan Doherty
There are some pretty weak hash functions in there.
What about using:
--- a/lib/librte_eal/common/include/rte_common.h2014-02-27
09:02:23.424698279 -0800
+++ b/lib/librte_eal/common/include/rte_common.h2014-06-05
08:03:51.615839548 -0700
@@ -365,6 +365,31 @@ rte_str_to_size(const char *str)
}
/**
+ * Multiplicative hash functions useful to distrbute
+ * a uniform set of consective keys
+ *
+ * @param val
+ * The input uniform key
+ * @param bits
+ * Number of bits desired
+ */
+#define GOLDEN_RATIO_32 2654404609U
+static inline uint32_t
+rte_fib_hash32(uint32_t val, unsigned int bits)
+{
+ val *= GOLDEN_RATIO_32;
+ return val >> (32 - bits);
+}
+
+#defineGOLDEN_RATIO_64 0x9e37fffc0001UL
+static inline uint64_t
+rte_fib_hash64(uint64_t val, unsigned bits)
+{
+ val *= GOLDEN_RATIO_64;
+ return val >> (64 - bits);
+}
+
+/**
* Function to terminate the application immediately, printing an error
* message and returning the exit_code back to the shell.
*
--- a/lib/librte_ether/rte_ether.h 2014-02-27 09:02:23.438697987 -0800
+++ b/lib/librte_ether/rte_ether.h 2014-06-05 08:14:15.438080058 -0700
@@ -48,6 +48,8 @@ extern "C" {
#include
#include
+#include
+#include
#define ETHER_ADDR_LEN 6 /**< Length of Ethernet address. */
#define ETHER_TYPE_LEN 2 /**< Length of Ethernet type field. */
@@ -250,6 +252,80 @@ struct ether_hdr {
uint16_t ether_type; /**< Frame type. */
} __attribute__((__packed__));
+
+/*
+ * These functions work by aliasing the 6 byte Ethernet address
+ * into a 64 bit value. The macro shift16, removes the extra
+ * bytes with the correct shift depending on byte order
+ */
+#ifdef __BYTE_ORDER
+ #if __BYTE_ORDER == __BIG_ENDIAN
+#define shift16(s) (s >>= 16)
+ #else
+#define shift16(s) (s <<= 16)
+ #endif
+#endif
+
+/**
+ * Fast compare of Ethernet address.
+ *
+ * @param e1
+ * A pointer to a ether_addr structure one address
+ * @param e2
+ * A pointer to a ether_addr structure holding other address
+ * @return
+ * True (1) if addresses are the same
+ * false (0) otherwise.
+ */
+static inline int
+ether_addr_equal(const struct ether_addr *e1, const struct ether_addr *e2)
+{
+ uint64_t e1_addr = *(const uint64_t *) e1;
+ shift16(e1_addr);
+ uint64_t e2_addr = *(const uint64_t *) e2;
+ shift16(e2_addr);
+
+ return (e1_addr == e2_addr);
+}
+
+/**
+ * Fast hash of ethernet address
+ *
+ * @param ea
+ * A pointer to a ether_addr structure
+ * @param bits
+ * Number of bits desired
+ * @return
+ * Calculated hash value.
+ */
+static inline uint32_t
+ether_addr_hash(const struct ether_addr *ea, unsigned bits)
+{
+ uint64_t val = *(const uint64_t *) ea;
+
+ shift16(val);
+ return rte_fib_hash64(val, bits);
+}
+#undef shift16
+
+/**
+ * Hash of ethernet source and destination
+ *
+ * @param ea
+ * A pointer to a ether_hdr structure
+ * @param initval
+ * Initialising value of hash.
+ * @return
+ * Calculated hash value.
+ */
+static inline uint32_t
+ether_header_hash(const struct ether_hdr *hdr, unsigned seed)
+{
+ const uint32_t *key = (const uint32_t *)hdr;
+
+ return rte_jhash(key, 2*ETHER_ADDR_LEN, seed);
+}
+
/**
* Ethernet VLAN Header.
* Contains the 16-bit VLAN Tag Control Identifier and the Ethernet type
[dpdk-dev] [PATCH 0/4] librte_pmd_virtio :Fix: virtio_pci.h non-existent virtio feature bit-flag tested
On Thu, 5 Jun 2014 15:39:15 +0100 Alan Carew wrote: > This series addresses an issue with librte_pmd_virtio where the offset to the > virtio device specific header may be incorrect depending on whether MSI-X has > been enabled or not. > > If MSI-X is configured the device specific header is placed at byte offset 24 > relative to the IO base address. > If MSI-X is not configured the device specific header is placed at > byte offset 20. > > The following macro defined in virtio_pci.h is used to test the > presence of the MSI-X header and determine the correct offset: > #define VIRTIO_PCI_CONFIG(hw) (((hw)->guest_features & VIRTIO_PCI_FLAG_MSIX) > ? 24 : 20) > > However, VIRTIO_PCI_FLAG_MSIX is not a guest_feature nor is it part of the > Virtio Specification and resolves to the VIRTIO_NET_F_MAC feature as both > are #defined as 0x20. > > VIRTIO_PCI_FLAG_MSIX or similar flag should instead be set by the kernel > driver allocating resources and passed to user space for testing. > i.e. > #define VIRTIO_PCI_CONFIG(hw) (((hw)->intr_mode & IGBUIO_MSIX_INTR_MODE) ? 24 > : 20) > > To enable this testing of interrupt mode, this series allows for the kernel > driver(igb_uio) to place the configured interrupt mode into a sysfs entry. > sysfs is then parsed by eal_pci to determine the configured mode, which > allows all user space devices to correctly determine the interrupt mode, > including virtio_ethdev. > > This series should be applied prior to Anatoly Burakov's > [VFIO] Add VFIO support to DPDK series > > Alan Carew (4): > igb_uio: Add interrupt_mode sysfs entry for igb_uio devices > eal_pci: Add interrupt mode to rte_pci_device and parsing to eal_pci > FreeBSD: Adds the equivalent interrupt mode setting and parsing > virtio: Fixes the VIRTIO_PCI_CONFIG macro to use the correct offset > to the Virtio header > > lib/librte_eal/bsdapp/eal/eal_pci.c| 44 ++ > lib/librte_eal/bsdapp/nic_uio/nic_uio.c| 14 +++ > lib/librte_eal/common/Makefile |1 + > lib/librte_eal/common/include/rte_pci.h|2 + > .../common/include/rte_pci_dev_feature_defs.h | 85 > > .../common/include/rte_pci_dev_features.h | 70 > lib/librte_eal/linuxapp/eal/eal_pci.c | 78 ++ > lib/librte_eal/linuxapp/igb_uio/igb_uio.c | 48 +--- > lib/librte_pmd_virtio/virtio_ethdev.c |1 + > lib/librte_pmd_virtio/virtio_pci.h |4 +- > 10 files changed, 336 insertions(+), 11 deletions(-) > create mode 100755 lib/librte_eal/common/include/rte_pci_dev_feature_defs.h > create mode 100755 lib/librte_eal/common/include/rte_pci_dev_features.h > I took the simpler approach. Since DPDK devices doesn't really need the multiple vectors of MSI-X at all, I just converted igb_uio to use MSI instead.
[dpdk-dev] [PATCH v2 05/27] mbuf: add new packet flags for i40e
On Thu, 5 Jun 2014 13:08:49 +0800 Helin Zhang wrote: > +#define PKT_RX_EIP_CKSUM_BAD 0x /**< External IP header checksum error. > */ > +#define PKT_RX_OVERSIZE 0x /**< Num of desc of an RX pkt oversize. > */ > +#define PKT_RX_HBUF_OVERFLOW 0x /**< Header buffer overflow. */ > +#define PKT_RX_RECIP_ERR 0x /**< Hardware processing error. */ > +#define PKT_RX_MAC_ERR 0x /**< MAC error. */ Flags are useless if all zero!
[dpdk-dev] [PATCH] fix for eth_pcap_tx() can cause mbuf corruption
Acked-by: Pablo de Lara Guarch > -Original Message- > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Konstantin Ananyev > Sent: Wednesday, May 28, 2014 3:47 PM > To: dev at dpdk.org; dev at dpdk.org > Subject: [dpdk-dev] [PATCH] fix for eth_pcap_tx() can cause mbuf corruption > > If pcap_sendpacket() fails, then eth_pcap_tx shouldn't silently free that > mbuf and continue. > > Signed-off-by: Konstantin Ananyev > --- > lib/librte_pmd_pcap/rte_eth_pcap.c |5 +++-- > 1 files changed, 3 insertions(+), 2 deletions(-) > > diff --git a/lib/librte_pmd_pcap/rte_eth_pcap.c > b/lib/librte_pmd_pcap/rte_eth_pcap.c > index dc4670c..6f026ab 100644 > --- a/lib/librte_pmd_pcap/rte_eth_pcap.c > +++ b/lib/librte_pmd_pcap/rte_eth_pcap.c > @@ -239,8 +239,9 @@ eth_pcap_tx(void *queue, > mbuf = bufs[i]; > ret = pcap_sendpacket(tx_queue->pcap, (u_char*) mbuf- > >pkt.data, > mbuf->pkt.data_len); > - if(likely(!ret)) > - num_tx++; > + if (unlikely(ret != 0)) > + break; > + num_tx++; > rte_pktmbuf_free(mbuf); > } > > -- > 1.7.7.6
[dpdk-dev] [PATCH 1/2] rte_tailq.h: Fix compilation under FreeBSD
Recent change to rte_dump_tailq, which now uses a FILE parameter
causes compilation to fail under FreeBSD and sourced to a
missing include of stdio.h
This and next patch(both small) allows to compile without error.
Signed-off-by: Alan Carew
---
lib/librte_eal/common/include/rte_tailq.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/lib/librte_eal/common/include/rte_tailq.h
b/lib/librte_eal/common/include/rte_tailq.h
index 42df7d2..0ddcc11 100644
--- a/lib/librte_eal/common/include/rte_tailq.h
+++ b/lib/librte_eal/common/include/rte_tailq.h
@@ -45,6 +45,7 @@ extern "C" {
#endif
#include
+#include
/** dummy structure type used by the rte_tailq APIs */
struct rte_dummy {
--
1.9.3
[dpdk-dev] [PATCH 2/2] eal_pci: Fix compilation under FreeBSD
I'm not sure why this has not caused an issue before, perhaps
missing -Werror=unused-parameter or super-set
and subsequently fixed.
This patch adds __rte_unused to
pci_unbind_kernel_driver(struct rte_pci_device *dev)
Signed-off-by: Alan Carew
---
lib/librte_eal/bsdapp/eal/eal_pci.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/librte_eal/bsdapp/eal/eal_pci.c
b/lib/librte_eal/bsdapp/eal/eal_pci.c
index 94ae461..5c4d81b 100644
--- a/lib/librte_eal/bsdapp/eal/eal_pci.c
+++ b/lib/librte_eal/bsdapp/eal/eal_pci.c
@@ -110,7 +110,7 @@ static struct uio_res_list *uio_res_list = NULL;
/* unbind kernel driver for this device */
static int
-pci_unbind_kernel_driver(struct rte_pci_device *dev)
+pci_unbind_kernel_driver(struct rte_pci_device *dev __rte_unused)
{
RTE_LOG(ERR, EAL, "RTE_PCI_DRV_FORCE_UNBIND flag is not implemented "
"for BSD\n");
--
1.9.3
[dpdk-dev] [PATCH 0/4] librte_pmd_virtio :Fix: virtio_pci.h non-existent virtio feature bit-flag tested
On Thu, 5 Jun 2014 15:39:15 +0100 Alan Carew wrote: > This series addresses an issue with librte_pmd_virtio where the offset to the > virtio device specific header may be incorrect depending on whether MSI-X has > been enabled or not. > > If MSI-X is configured the device specific header is placed at byte offset 24 > relative to the IO base address. > If MSI-X is not configured the device specific header is placed at > byte offset 20. > > The following macro defined in virtio_pci.h is used to test the > presence of the MSI-X header and determine the correct offset: > #define VIRTIO_PCI_CONFIG(hw) (((hw)->guest_features & VIRTIO_PCI_FLAG_MSIX) > ? 24 : 20) > > However, VIRTIO_PCI_FLAG_MSIX is not a guest_feature nor is it part of the > Virtio Specification and resolves to the VIRTIO_NET_F_MAC feature as both > are #defined as 0x20. > > VIRTIO_PCI_FLAG_MSIX or similar flag should instead be set by the kernel > driver allocating resources and passed to user space for testing. > i.e. > #define VIRTIO_PCI_CONFIG(hw) (((hw)->intr_mode & IGBUIO_MSIX_INTR_MODE) ? 24 > : 20) > > To enable this testing of interrupt mode, this series allows for the kernel > driver(igb_uio) to place the configured interrupt mode into a sysfs entry. > sysfs is then parsed by eal_pci to determine the configured mode, which > allows all user space devices to correctly determine the interrupt mode, > including virtio_ethdev. Why not scan for the IRQ in /proc/interrupts rather than changing the kernel driver. I ask because my goal is to get rid of igb_uio and it's quirks and use uio_pci_generic instead.
[dpdk-dev] [PATCH 1/4] [PATCH 1/4] igb_uio: Add interrupt_mode sysfs entry for igb_uio devices
On Thu, 5 Jun 2014 15:39:16 +0100
Alan Carew wrote:
> This patch adds an "interrupt_mode" sysfs entry for igb_uio devices,
> allowing userspace eal_pci to track which interrupt mode has been enabled in
> kernel space.
> The sysfs entry can be inspected via
> /sys/bus/pci/devices/
> Signed-off-by: Alan Carew
> ---
> .../common/include/rte_pci_dev_feature_defs.h | 85
>
> .../common/include/rte_pci_dev_features.h | 70
> lib/librte_eal/linuxapp/igb_uio/igb_uio.c | 48 +---
> 3 files changed, 193 insertions(+), 10 deletions(-)
> create mode 100755 lib/librte_eal/common/include/rte_pci_dev_feature_defs.h
> create mode 100755 lib/librte_eal/common/include/rte_pci_dev_features.h
>
> diff --git a/lib/librte_eal/common/include/rte_pci_dev_feature_defs.h
> b/lib/librte_eal/common/include/rte_pci_dev_feature_defs.h
> new file mode 100755
> index 000..d23ed7d
> --- /dev/null
> +++ b/lib/librte_eal/common/include/rte_pci_dev_feature_defs.h
> @@ -0,0 +1,85 @@
> +/*-
> + * This file is provided under a dual BSD/GPLv2 license. When using or
> + * redistributing this file, you may do so under either license.
> + *
> + * GPL LICENSE SUMMARY
> + *
> + * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of version 2 of the GNU General Public License as
> + * published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful, but
> + * WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + * General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program; if not, write to the Free Software
> + * Foundation, Inc., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301
> USA.
> + * The full GNU General Public License is included in this distribution
> + * in the file called LICENSE.GPL.
> + *
> + * Contact Information:
> + * Intel Corporation
> + *
> + * BSD LICENSE
> + *
> + * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
> + * All rights reserved.
> + *
> + * Redistribution and use in source and binary forms, with or without
> + * modification, are permitted provided that the following conditions
> + * are met:
> + *
> + * * Redistributions of source code must retain the above copyright
> + * notice, this list of conditions and the following disclaimer.
> + * * Redistributions in binary form must reproduce the above copyright
> + * notice, this list of conditions and the following disclaimer in
> + * the documentation and/or other materials provided with the
> + * distribution.
> + * * Neither the name of Intel Corporation nor the names of its
> + * contributors may be used to endorse or promote products derived
> + * from this software without specific prior written permission.
> + *
> + * 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 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.
> + *
> + */
> +
> +#ifndef _RTE_PCI_DEV_DEFS_H_
> +#define _RTE_PCI_DEV_DEFS_H_
> +
> +#define RTE_PCI_DEV_FEATURE_INTR_MODE "interrupt_mode"
> +
> +#define INTR_NAME_LEN 10
> +
> +#define IGBUIO_NONE_INTR_NAME "none"
> +#define IGBUIO_LEGACY_INTR_NAME "legacy"
> +#define IGBUIO_MSI_INTR_NAME "msi"
> +#define IGBUIO_MSIX_INTR_NAME "msix"
> +
> +
> +#define INTR_MODE(id, mode_name) \
> +.mode = (id), .name = (mode_name)
> +
> +/* interrupt mode */
> +enum igbuio_intr_mode {
> + IGBUIO_NONE_INTR_MODE = 0,
> + IGBUIO_LEGACY_INTR_MODE,
> + IGBUIO_MSI_INTR_MODE,
> + IGBUIO_MSIX_INTR_MODE,
> + IGBUIO_INTR_MODE_MAX
> +};
> +
> +#endif /* _RTE_PCI_DEV_DEFS_H_ */
> diff --git a/lib/librte_eal/common/include/rte_pci_dev_features.h
> b/lib/librte_eal/common/include/rte_pci_dev_features.h
> new file mode 100755
> index 000..a45c056
> --- /dev/null
> +++ b/lib/librte_eal/common/include/rte_pci_dev_features.h
> @@ -0,0 +1,70 @@
> +/*-
> + * This file is provid
[dpdk-dev] Licensing consistency
Hey all- One of the things that came up during the dpdk package review for Fedora was the inconsistency of License reporting in the upstream project. DPDK is triple licensed, whcih isn't in and of itself a big deal, but indications of which file(s) are under which license is fairly scattered. For instance: 1) The kni module has a GPLv2 license at the top of each file 2) The kni MODULE_LICENSE macro indicates the license is dual BSD/GPLv2 3) The rte_kni_common.h file is licensed dual BSD/LGPL v2 4) The linux kernel modules for hardware pmds have no license file in them at all, but do have a README which contains a BSD license (though no clear indicator that this license applies to the files in this directory). Theres several more examples of this, but the point is, its often not clear what bits fall under what license. Has any effort been made to consolodate licensing here, or at least to make it consistent and clear where to find license information for a file? If not I would propose that all files in the DPDK be required to carry the license that they are distributed under in the top of said file, and that we add a LICENSE file to the tree root indicating that each file contains its own licensing terms. Thoughts? Neil
[dpdk-dev] [PATCH v2 01/27] i40e: add basic shared code
Add shared code source files to support basic operations to be called in poll mode driver. The new files are, - i40e_adminq.c - i40e_adminq.h - i40e_adminq_cmd.h - i40e_alloc.h - i40e_common.c - i40e_dcb.c - i40e_dcb.h - i40e_diag.c - i40e_diag.h - i40e_hmc.c - i40e_hmc.h - i40e_lan_hmc.c - i40e_lan_hmc.h - i40e_nvm.c - i40e_prototype.h - i40e_register.h - i40e_register_x710_int.h - i40e_status.h - i40e_type.h - i40e_virtchnl.h Signed-off-by: Helin Zhang Signed-off-by: Jing Chen Acked-by: Cunming Liang Acked-by: Jijiang Liu Acked-by: Jingjing Wu Tested-by: Waterman Cao --- lib/librte_pmd_i40e/i40e/i40e_adminq.c| 1103 ++ lib/librte_pmd_i40e/i40e/i40e_adminq.h| 158 + lib/librte_pmd_i40e/i40e/i40e_adminq_cmd.h| 2179 lib/librte_pmd_i40e/i40e/i40e_alloc.h |65 + lib/librte_pmd_i40e/i40e/i40e_common.c| 4846 + lib/librte_pmd_i40e/i40e/i40e_dcb.c | 1104 ++ lib/librte_pmd_i40e/i40e/i40e_dcb.h | 264 + lib/librte_pmd_i40e/i40e/i40e_diag.c | 188 + lib/librte_pmd_i40e/i40e/i40e_diag.h |61 + lib/librte_pmd_i40e/i40e/i40e_hmc.c | 373 + lib/librte_pmd_i40e/i40e/i40e_hmc.h | 244 + lib/librte_pmd_i40e/i40e/i40e_lan_hmc.c | 1616 +++ lib/librte_pmd_i40e/i40e/i40e_lan_hmc.h | 214 + lib/librte_pmd_i40e/i40e/i40e_nvm.c | 942 ++ lib/librte_pmd_i40e/i40e/i40e_prototype.h | 435 + lib/librte_pmd_i40e/i40e/i40e_register.h | 3377 ++ lib/librte_pmd_i40e/i40e/i40e_register_x710_int.h | 10712 lib/librte_pmd_i40e/i40e/i40e_status.h| 107 + lib/librte_pmd_i40e/i40e/i40e_type.h | 1462 +++ lib/librte_pmd_i40e/i40e/i40e_virtchnl.h | 373 + 20 files changed, 29823 insertions(+) create mode 100644 lib/librte_pmd_i40e/i40e/i40e_adminq.c create mode 100644 lib/librte_pmd_i40e/i40e/i40e_adminq.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_adminq_cmd.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_alloc.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_common.c create mode 100644 lib/librte_pmd_i40e/i40e/i40e_dcb.c create mode 100644 lib/librte_pmd_i40e/i40e/i40e_dcb.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_diag.c create mode 100644 lib/librte_pmd_i40e/i40e/i40e_diag.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_hmc.c create mode 100644 lib/librte_pmd_i40e/i40e/i40e_hmc.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_lan_hmc.c create mode 100644 lib/librte_pmd_i40e/i40e/i40e_lan_hmc.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_nvm.c create mode 100644 lib/librte_pmd_i40e/i40e/i40e_prototype.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_register.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_register_x710_int.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_status.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_type.h create mode 100644 lib/librte_pmd_i40e/i40e/i40e_virtchnl.h diff --git a/lib/librte_pmd_i40e/i40e/i40e_adminq.c b/lib/librte_pmd_i40e/i40e/i40e_adminq.c new file mode 100644 index 000..d078cea --- /dev/null +++ b/lib/librte_pmd_i40e/i40e/i40e_adminq.c @@ -0,0 +1,1103 @@ +/*** + +Copyright (c) 2013 - 2014, Intel Corporation +All rights reserved. + +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + 1. Redistributions of source code must retain the above copyright notice, +this list of conditions and the following disclaimer. + + 2. Redistributions in binary form must reproduce the above copyright +notice, this list of conditions and the following disclaimer in the +documentation and/or other materials provided with the distribution. + + 3. Neither the name of the Intel Corporation nor the names of its +contributors may be used to endorse or promote products derived from +this software without specific prior written permission. + +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 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. + +***
