Merge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into staging

# gpg: Signature made Thu 02 Jun 2016 07:23:18 BST using RSA key ID 398D6211
# gpg: Good signature from "Jason Wang (Jason Wang on RedHat) <jasowang@redhat.com>"
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg:          It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 215D 46F4 8246 689E C77F  3562 EF04 965B 398D 6211

* remotes/jasowang/tags/net-pull-request: (31 commits)
  Add ENET device to i.MX6 SOC.
  Add ENET/Gbps Ethernet support to FEC device
  i.MX: move FEC device to a register array structure.
  i.MX: Rename i.MX FEC defines to ENET_XXX
  i.MX: reset TX/RX descriptors when FEC is disabled.
  i.MX: Fix FEC code for ECR register reset value.
  i.MX: Fix FEC code for MDIO address selection
  i.MX: Fix FEC code for MDIO operation selection
  net: handle optional VLAN header in checksum computation.
  net: improve UDP/TCP checksum computation.
  e1000e: Introduce qtest for e1000e device
  net: Introduce e1000e device emulation
  e1000: Move out code that will be reused in e1000e
  e1000_regs: Add definitions for Intel 82574-specific bits
  vmxnet3: Use pci_dma_* API instead of cpu_physical_memory_*
  net_pkt: Extend packet abstraction as required by e1000e functionality
  rtl8139: Move more TCP definitions to common header
  net_pkt: Name vmxnet3 packet abstractions more generic
  vmxnet3: Use common MAC address tracing macros
  net: Add macros for MAC address tracing
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

10 files changed, 402 insertions, 104 deletions

diff --git a/include/hw/arm/fsl-imx6.h b/include/hw/arm/fsl-imx6.h
index e9157ea4b3..ec6c509d74 100644
--- a/include/hw/arm/fsl-imx6.h
+++ b/include/hw/arm/fsl-imx6.h
@@ -28,6 +28,7 @@
 #include "hw/gpio/imx_gpio.h"
 #include "hw/sd/sdhci.h"
 #include "hw/ssi/imx_spi.h"
+#include "hw/net/imx_fec.h"
 #include "exec/memory.h"
 #include "cpu.h"
 
@@ -58,6 +59,7 @@ typedef struct FslIMX6State {
     IMXGPIOState   gpio[FSL_IMX6_NUM_GPIOS];
     SDHCIState     esdhc[FSL_IMX6_NUM_ESDHCS];
     IMXSPIState    spi[FSL_IMX6_NUM_ECSPIS];
+    IMXFECState    eth;
     MemoryRegion   rom;
     MemoryRegion   caam;
     MemoryRegion   ocram;
@@ -436,8 +438,8 @@ typedef struct FslIMX6State {
 #define FSL_IMX6_HDMI_MASTER_IRQ 115
 #define FSL_IMX6_HDMI_CEC_IRQ 116
 #define FSL_IMX6_MLB150_LOW_IRQ 117
-#define FSL_IMX6_ENET_MAC_IRQ 118
-#define FSL_IMX6_ENET_MAC_1588_IRQ 119
+#define FSL_IMX6_ENET_MAC_1588_IRQ 118
+#define FSL_IMX6_ENET_MAC_IRQ 119
 #define FSL_IMX6_PCIE1_IRQ 120
 #define FSL_IMX6_PCIE2_IRQ 121
 #define FSL_IMX6_PCIE3_IRQ 122
diff --git a/include/hw/net/imx_fec.h b/include/hw/net/imx_fec.h
index cbf86509e8..62ad473b05 100644
--- a/include/hw/net/imx_fec.h
+++ b/include/hw/net/imx_fec.h
@@ -1,5 +1,5 @@
 /*
- * i.MX Fast Ethernet Controller emulation.
+ * i.MX FEC/ENET Ethernet Controller emulation.
  *
  * Copyright (c) 2013 Jean-Christophe Dubois. <jcd@tribudubois.net>
  *
@@ -27,27 +27,147 @@
 #define TYPE_IMX_FEC "imx.fec"
 #define IMX_FEC(obj) OBJECT_CHECK(IMXFECState, (obj), TYPE_IMX_FEC)
 
+#define TYPE_IMX_ENET "imx.enet"
+
 #include "hw/sysbus.h"
 #include "net/net.h"
 
-#define FEC_MAX_FRAME_SIZE 2032
-
-#define FEC_INT_HB      (1 << 31)
-#define FEC_INT_BABR    (1 << 30)
-#define FEC_INT_BABT    (1 << 29)
-#define FEC_INT_GRA     (1 << 28)
-#define FEC_INT_TXF     (1 << 27)
-#define FEC_INT_TXB     (1 << 26)
-#define FEC_INT_RXF     (1 << 25)
-#define FEC_INT_RXB     (1 << 24)
-#define FEC_INT_MII     (1 << 23)
-#define FEC_INT_EBERR   (1 << 22)
-#define FEC_INT_LC      (1 << 21)
-#define FEC_INT_RL      (1 << 20)
-#define FEC_INT_UN      (1 << 19)
-
-#define FEC_EN      2
-#define FEC_RESET   1
+#define ENET_EIR               1
+#define ENET_EIMR              2
+#define ENET_RDAR              4
+#define ENET_TDAR              5
+#define ENET_ECR               9
+#define ENET_MMFR              16
+#define ENET_MSCR              17
+#define ENET_MIBC              25
+#define ENET_RCR               33
+#define ENET_TCR               49
+#define ENET_PALR              57
+#define ENET_PAUR              58
+#define ENET_OPD               59
+#define ENET_IAUR              70
+#define ENET_IALR              71
+#define ENET_GAUR              72
+#define ENET_GALR              73
+#define ENET_TFWR              81
+#define ENET_FRBR              83
+#define ENET_FRSR              84
+#define ENET_RDSR              96
+#define ENET_TDSR              97
+#define ENET_MRBR              98
+#define ENET_RSFL              100
+#define ENET_RSEM              101
+#define ENET_RAEM              102
+#define ENET_RAFL              103
+#define ENET_TSEM              104
+#define ENET_TAEM              105
+#define ENET_TAFL              106
+#define ENET_TIPG              107
+#define ENET_FTRL              108
+#define ENET_TACC              112
+#define ENET_RACC              113
+#define ENET_MIIGSK_CFGR       192
+#define ENET_MIIGSK_ENR        194
+#define ENET_ATCR              256
+#define ENET_ATVR              257
+#define ENET_ATOFF             258
+#define ENET_ATPER             259
+#define ENET_ATCOR             260
+#define ENET_ATINC             261
+#define ENET_ATSTMP            262
+#define ENET_TGSR              385
+#define ENET_TCSR0             386
+#define ENET_TCCR0             387
+#define ENET_TCSR1             388
+#define ENET_TCCR1             389
+#define ENET_TCSR2             390
+#define ENET_TCCR2             391
+#define ENET_TCSR3             392
+#define ENET_TCCR3             393
+#define ENET_MAX               400
+
+#define ENET_MAX_FRAME_SIZE    2032
+
+/* EIR and EIMR */
+#define ENET_INT_HB            (1 << 31)
+#define ENET_INT_BABR          (1 << 30)
+#define ENET_INT_BABT          (1 << 29)
+#define ENET_INT_GRA           (1 << 28)
+#define ENET_INT_TXF           (1 << 27)
+#define ENET_INT_TXB           (1 << 26)
+#define ENET_INT_RXF           (1 << 25)
+#define ENET_INT_RXB           (1 << 24)
+#define ENET_INT_MII           (1 << 23)
+#define ENET_INT_EBERR         (1 << 22)
+#define ENET_INT_LC            (1 << 21)
+#define ENET_INT_RL            (1 << 20)
+#define ENET_INT_UN            (1 << 19)
+#define ENET_INT_PLR           (1 << 18)
+#define ENET_INT_WAKEUP        (1 << 17)
+#define ENET_INT_TS_AVAIL      (1 << 16)
+#define ENET_INT_TS_TIMER      (1 << 15)
+
+#define ENET_INT_MAC           (ENET_INT_HB | ENET_INT_BABR | ENET_INT_BABT | \
+                                ENET_INT_GRA | ENET_INT_TXF | ENET_INT_TXB | \
+                                ENET_INT_RXF | ENET_INT_RXB | ENET_INT_MII | \
+                                ENET_INT_EBERR | ENET_INT_LC | ENET_INT_RL | \
+                                ENET_INT_UN | ENET_INT_PLR | ENET_INT_WAKEUP | \
+                                ENET_INT_TS_AVAIL)
+
+/* RDAR */
+#define ENET_RDAR_RDAR         (1 << 24)
+
+/* TDAR */
+#define ENET_TDAR_TDAR         (1 << 24)
+
+/* ECR */
+#define ENET_ECR_RESET         (1 << 0)
+#define ENET_ECR_ETHEREN       (1 << 1)
+#define ENET_ECR_MAGICEN       (1 << 2)
+#define ENET_ECR_SLEEP         (1 << 3)
+#define ENET_ECR_EN1588        (1 << 4)
+#define ENET_ECR_SPEED         (1 << 5)
+#define ENET_ECR_DBGEN         (1 << 6)
+#define ENET_ECR_STOPEN        (1 << 7)
+#define ENET_ECR_DSBWP         (1 << 8)
+
+/* MIBC */
+#define ENET_MIBC_MIB_DIS      (1 << 31)
+#define ENET_MIBC_MIB_IDLE     (1 << 30)
+#define ENET_MIBC_MIB_CLEAR    (1 << 29)
+
+/* RCR */
+#define ENET_RCR_LOOP          (1 << 0)
+#define ENET_RCR_DRT           (1 << 1)
+#define ENET_RCR_MII_MODE      (1 << 2)
+#define ENET_RCR_PROM          (1 << 3)
+#define ENET_RCR_BC_REJ        (1 << 4)
+#define ENET_RCR_FCE           (1 << 5)
+#define ENET_RCR_RGMII_EN      (1 << 6)
+#define ENET_RCR_RMII_MODE     (1 << 8)
+#define ENET_RCR_RMII_10T      (1 << 9)
+#define ENET_RCR_PADEN         (1 << 12)
+#define ENET_RCR_PAUFWD        (1 << 13)
+#define ENET_RCR_CRCFWD        (1 << 14)
+#define ENET_RCR_CFEN          (1 << 15)
+#define ENET_RCR_MAX_FL_SHIFT  (16)
+#define ENET_RCR_MAX_FL_LENGTH (14)
+#define ENET_RCR_NLC           (1 << 30)
+#define ENET_RCR_GRS           (1 << 31)
+
+/* TCR */
+#define ENET_TCR_GTS           (1 << 0)
+#define ENET_TCR_FDEN          (1 << 2)
+#define ENET_TCR_TFC_PAUSE     (1 << 3)
+#define ENET_TCR_RFC_PAUSE     (1 << 4)
+#define ENET_TCR_ADDSEL_SHIFT  (5)
+#define ENET_TCR_ADDSEL_LENGTH (3)
+#define ENET_TCR_CRCFWD        (1 << 9)
+
+/* RDSR */
+#define ENET_TWFR_TFWR_SHIFT   (0)
+#define ENET_TWFR_TFWR_LENGTH  (6)
+#define ENET_TWFR_STRFWD       (1 << 8)
 
 /* Buffer Descriptor.  */
 typedef struct {
@@ -56,22 +176,60 @@ typedef struct {
     uint32_t data;
 } IMXFECBufDesc;
 
-#define FEC_BD_R    (1 << 15)
-#define FEC_BD_E    (1 << 15)
-#define FEC_BD_O1   (1 << 14)
-#define FEC_BD_W    (1 << 13)
-#define FEC_BD_O2   (1 << 12)
-#define FEC_BD_L    (1 << 11)
-#define FEC_BD_TC   (1 << 10)
-#define FEC_BD_ABC  (1 << 9)
-#define FEC_BD_M    (1 << 8)
-#define FEC_BD_BC   (1 << 7)
-#define FEC_BD_MC   (1 << 6)
-#define FEC_BD_LG   (1 << 5)
-#define FEC_BD_NO   (1 << 4)
-#define FEC_BD_CR   (1 << 2)
-#define FEC_BD_OV   (1 << 1)
-#define FEC_BD_TR   (1 << 0)
+#define ENET_BD_R              (1 << 15)
+#define ENET_BD_E              (1 << 15)
+#define ENET_BD_O1             (1 << 14)
+#define ENET_BD_W              (1 << 13)
+#define ENET_BD_O2             (1 << 12)
+#define ENET_BD_L              (1 << 11)
+#define ENET_BD_TC             (1 << 10)
+#define ENET_BD_ABC            (1 << 9)
+#define ENET_BD_M              (1 << 8)
+#define ENET_BD_BC             (1 << 7)
+#define ENET_BD_MC             (1 << 6)
+#define ENET_BD_LG             (1 << 5)
+#define ENET_BD_NO             (1 << 4)
+#define ENET_BD_CR             (1 << 2)
+#define ENET_BD_OV             (1 << 1)
+#define ENET_BD_TR             (1 << 0)
+
+typedef struct {
+    uint16_t length;
+    uint16_t flags;
+    uint32_t data;
+    uint16_t status;
+    uint16_t option;
+    uint16_t checksum;
+    uint16_t head_proto;
+    uint32_t last_buffer;
+    uint32_t timestamp;
+    uint32_t reserved[2];
+} IMXENETBufDesc;
+
+#define ENET_BD_ME             (1 << 15)
+#define ENET_BD_TX_INT         (1 << 14)
+#define ENET_BD_TS             (1 << 13)
+#define ENET_BD_PINS           (1 << 12)
+#define ENET_BD_IINS           (1 << 11)
+#define ENET_BD_PE             (1 << 10)
+#define ENET_BD_CE             (1 << 9)
+#define ENET_BD_UC             (1 << 8)
+#define ENET_BD_RX_INT         (1 << 7)
+
+#define ENET_BD_TXE            (1 << 15)
+#define ENET_BD_UE             (1 << 13)
+#define ENET_BD_EE             (1 << 12)
+#define ENET_BD_FE             (1 << 11)
+#define ENET_BD_LCE            (1 << 10)
+#define ENET_BD_OE             (1 << 9)
+#define ENET_BD_TSE            (1 << 8)
+#define ENET_BD_ICE            (1 << 5)
+#define ENET_BD_PCR            (1 << 4)
+#define ENET_BD_VLAN           (1 << 2)
+#define ENET_BD_IPV6           (1 << 1)
+#define ENET_BD_FRAG           (1 << 0)
+
+#define ENET_BD_BDU            (1 << 31)
 
 typedef struct IMXFECState {
     /*< private >*/
@@ -80,34 +238,20 @@ typedef struct IMXFECState {
     /*< public >*/
     NICState *nic;
     NICConf conf;
-    qemu_irq irq;
+    qemu_irq irq[2];
     MemoryRegion iomem;
 
-    uint32_t irq_state;
-    uint32_t eir;
-    uint32_t eimr;
-    uint32_t rx_enabled;
+    uint32_t regs[ENET_MAX];
     uint32_t rx_descriptor;
     uint32_t tx_descriptor;
-    uint32_t ecr;
-    uint32_t mmfr;
-    uint32_t mscr;
-    uint32_t mibc;
-    uint32_t rcr;
-    uint32_t tcr;
-    uint32_t tfwr;
-    uint32_t frsr;
-    uint32_t erdsr;
-    uint32_t etdsr;
-    uint32_t emrbr;
-    uint32_t miigsk_cfgr;
-    uint32_t miigsk_enr;
 
     uint32_t phy_status;
     uint32_t phy_control;
     uint32_t phy_advertise;
     uint32_t phy_int;
     uint32_t phy_int_mask;
+
+    bool is_fec;
 } IMXFECState;
 
 #endif
diff --git a/include/hw/pci/msix.h b/include/hw/pci/msix.h
index 72e5f931c5..048a29dd2f 100644
--- a/include/hw/pci/msix.h
+++ b/include/hw/pci/msix.h
@@ -29,6 +29,7 @@ int msix_present(PCIDevice *dev);
 
 bool msix_is_masked(PCIDevice *dev, unsigned vector);
 void msix_set_pending(PCIDevice *dev, unsigned vector);
+void msix_clr_pending(PCIDevice *dev, int vector);
 
 int msix_vector_use(PCIDevice *dev, unsigned vector);
 void msix_vector_unuse(PCIDevice *dev, unsigned vector);
diff --git a/include/hw/pci/pci.h b/include/hw/pci/pci.h
index ef6ba51f6c..4420f47598 100644
--- a/include/hw/pci/pci.h
+++ b/include/hw/pci/pci.h
@@ -465,16 +465,23 @@ pci_get_long(const uint8_t *config)
     return ldl_le_p(config);
 }
 
+/*
+ * PCI capabilities and/or their fields
+ * are generally DWORD aligned only so
+ * mechanism used by pci_set/get_quad()
+ * must be tolerant to unaligned pointers
+ *
+ */
 static inline void
 pci_set_quad(uint8_t *config, uint64_t val)
 {
-    cpu_to_le64w((uint64_t *)config, val);
+    stq_le_p(config, val);
 }
 
 static inline uint64_t
 pci_get_quad(const uint8_t *config)
 {
-    return le64_to_cpup((const uint64_t *)config);
+    return ldq_le_p(config);
 }
 
 static inline void
diff --git a/include/hw/pci/pci_regs.h b/include/hw/pci/pci_regs.h
index ba8cbe9278..7a83142578 100644
--- a/include/hw/pci/pci_regs.h
+++ b/include/hw/pci/pci_regs.h
@@ -1 +1,3 @@
 #include "standard-headers/linux/pci_regs.h"
+
+#define  PCI_PM_CAP_VER_1_1     0x0002  /* PCI PM spec ver. 1.1 */
diff --git a/include/hw/pci/pcie.h b/include/hw/pci/pcie.h
index b48a7a2c5a..056d25e53c 100644
--- a/include/hw/pci/pcie.h
+++ b/include/hw/pci/pcie.h
@@ -80,8 +80,12 @@ struct PCIExpressDevice {
 
 /* PCI express capability helper functions */
 int pcie_cap_init(PCIDevice *dev, uint8_t offset, uint8_t type, uint8_t port);
+int pcie_cap_v1_init(PCIDevice *dev, uint8_t offset,
+                     uint8_t type, uint8_t port);
 int pcie_endpoint_cap_init(PCIDevice *dev, uint8_t offset);
 void pcie_cap_exit(PCIDevice *dev);
+int pcie_endpoint_cap_v1_init(PCIDevice *dev, uint8_t offset);
+void pcie_cap_v1_exit(PCIDevice *dev);
 uint8_t pcie_cap_get_type(const PCIDevice *dev);
 void pcie_cap_flags_set_vector(PCIDevice *dev, uint8_t vector);
 uint8_t pcie_cap_flags_get_vector(PCIDevice *dev);
@@ -115,6 +119,7 @@ void pcie_add_capability(PCIDevice *dev,
                          uint16_t offset, uint16_t size);
 
 void pcie_ari_init(PCIDevice *dev, uint16_t offset, uint16_t nextfn);
+void pcie_dev_ser_num_init(PCIDevice *dev, uint16_t offset, uint64_t ser_num);
 
 extern const VMStateDescription vmstate_pcie_device;
 
diff --git a/include/hw/pci/pcie_regs.h b/include/hw/pci/pcie_regs.h
index 6a28b33e69..a95522a13b 100644
--- a/include/hw/pci/pcie_regs.h
+++ b/include/hw/pci/pcie_regs.h
@@ -11,6 +11,7 @@
 
 /* express capability */
 
+#define PCI_EXP_VER1_SIZEOF             0x14 /* express capability of ver. 1 */
 #define PCI_EXP_VER2_SIZEOF             0x3c /* express capability of ver. 2 */
 #define PCI_EXT_CAP_VER_SHIFT           16
 #define PCI_EXT_CAP_NEXT_SHIFT          20
@@ -26,11 +27,11 @@
     (((x) + PCI_EXT_CAP_ALIGN - 1) & ~(PCI_EXT_CAP_ALIGN - 1))
 
 /* PCI_EXP_FLAGS */
-#define PCI_EXP_FLAGS_VER2              2 /* for now, supports only ver. 2 */
+#define PCI_EXP_FLAGS_VER1              1
+#define PCI_EXP_FLAGS_VER2              2
 #define PCI_EXP_FLAGS_IRQ_SHIFT         ctz32(PCI_EXP_FLAGS_IRQ)
 #define PCI_EXP_FLAGS_TYPE_SHIFT        ctz32(PCI_EXP_FLAGS_TYPE)
 
-
 /* PCI_EXP_LINK{CAP, STA} */
 /* link speed */
 #define PCI_EXP_LNK_LS_25               1
diff --git a/include/net/checksum.h b/include/net/checksum.h
index 7de1acb79a..7df472c2fe 100644
--- a/include/net/checksum.h
+++ b/include/net/checksum.h
@@ -18,6 +18,7 @@
 #ifndef QEMU_NET_CHECKSUM_H
 #define QEMU_NET_CHECKSUM_H
 
+#include "qemu/bswap.h"
 struct iovec;
 
 uint32_t net_checksum_add_cont(int len, uint8_t *buf, int seq);
@@ -45,9 +46,55 @@ net_raw_checksum(uint8_t *data, int length)
  * @iov_cnt: number of array elements
  * @iov_off: starting iov offset for checksumming
  * @size: length of data to be checksummed
+ * @csum_offset: offset of the checksum chunk
  */
 uint32_t net_checksum_add_iov(const struct iovec *iov,
                               const unsigned int iov_cnt,
-                              uint32_t iov_off, uint32_t size);
+                              uint32_t iov_off, uint32_t size,
+                              uint32_t csum_offset);
+
+typedef struct toeplitz_key_st {
+    uint32_t leftmost_32_bits;
+    uint8_t *next_byte;
+} net_toeplitz_key;
+
+static inline
+void net_toeplitz_key_init(net_toeplitz_key *key, uint8_t *key_bytes)
+{
+    key->leftmost_32_bits = be32_to_cpu(*(uint32_t *)key_bytes);
+    key->next_byte = key_bytes + sizeof(uint32_t);
+}
+
+static inline
+void net_toeplitz_add(uint32_t *result,
+                      uint8_t *input,
+                      uint32_t len,
+                      net_toeplitz_key *key)
+{
+    register uint32_t accumulator = *result;
+    register uint32_t leftmost_32_bits = key->leftmost_32_bits;
+    register uint32_t byte;
+
+    for (byte = 0; byte < len; byte++) {
+        register uint8_t input_byte = input[byte];
+        register uint8_t key_byte = *(key->next_byte++);
+        register uint8_t bit;
+
+        for (bit = 0; bit < 8; bit++) {
+            if (input_byte & (1 << 7)) {
+                accumulator ^= leftmost_32_bits;
+            }
+
+            leftmost_32_bits =
+                (leftmost_32_bits << 1) | ((key_byte & (1 << 7)) >> 7);
+
+            input_byte <<= 1;
+            key_byte <<= 1;
+        }
+    }
+
+    key->leftmost_32_bits = leftmost_32_bits;
+    *result = accumulator;
+}
 
 #endif /* QEMU_NET_CHECKSUM_H */
diff --git a/include/net/eth.h b/include/net/eth.h
index 18d0be3b16..2013175857 100644
--- a/include/net/eth.h
+++ b/include/net/eth.h
@@ -67,6 +67,16 @@ typedef struct tcp_header {
     uint16_t th_urp;            /* urgent pointer */
 } tcp_header;
 
+#define TCP_FLAGS_ONLY(flags) ((flags) & 0x3f)
+
+#define TCP_HEADER_FLAGS(tcp) \
+    TCP_FLAGS_ONLY(be16_to_cpu((tcp)->th_offset_flags))
+
+#define TCP_FLAG_ACK  0x10
+
+#define TCP_HEADER_DATA_OFFSET(tcp) \
+    (((be16_to_cpu((tcp)->th_offset_flags) >> 12) & 0xf) << 2)
+
 typedef struct udp_header {
     uint16_t uh_sport; /* source port */
     uint16_t uh_dport; /* destination port */
@@ -108,11 +118,34 @@ struct ip6_header {
     struct in6_address ip6_dst;    /* destination address */
 };
 
+typedef struct ip6_pseudo_header {
+    struct in6_address ip6_src;
+    struct in6_address ip6_dst;
+    uint32_t           len;
+    uint8_t            zero[3];
+    uint8_t            next_hdr;
+} ip6_pseudo_header;
+
 struct ip6_ext_hdr {
     uint8_t        ip6r_nxt;   /* next header */
     uint8_t        ip6r_len;   /* length in units of 8 octets */
 };
 
+struct ip6_ext_hdr_routing {
+    uint8_t     nxt;
+    uint8_t     len;
+    uint8_t     rtype;
+    uint8_t     segleft;
+    uint8_t     rsvd[4];
+};
+
+struct ip6_option_hdr {
+#define IP6_OPT_PAD1   (0x00)
+#define IP6_OPT_HOME   (0xC9)
+    uint8_t type;
+    uint8_t len;
+};
+
 struct udp_hdr {
   uint16_t uh_sport;           /* source port */
   uint16_t uh_dport;           /* destination port */
@@ -161,19 +194,22 @@ struct tcp_hdr {
 #define PKT_GET_IP_HDR(p)         \
     ((struct ip_header *)(((uint8_t *)(p)) + eth_get_l2_hdr_length(p)))
 #define IP_HDR_GET_LEN(p)         \
-    ((((struct ip_header *)p)->ip_ver_len & 0x0F) << 2)
+    ((((struct ip_header *)(p))->ip_ver_len & 0x0F) << 2)
 #define PKT_GET_IP_HDR_LEN(p)     \
     (IP_HDR_GET_LEN(PKT_GET_IP_HDR(p)))
 #define PKT_GET_IP6_HDR(p)        \
     ((struct ip6_header *) (((uint8_t *)(p)) + eth_get_l2_hdr_length(p)))
 #define IP_HEADER_VERSION(ip)     \
-    ((ip->ip_ver_len >> 4)&0xf)
+    (((ip)->ip_ver_len >> 4) & 0xf)
+#define IP4_IS_FRAGMENT(ip) \
+    ((be16_to_cpu((ip)->ip_off) & (IP_OFFMASK | IP_MF)) != 0)
 
 #define ETH_P_IP                  (0x0800)      /* Internet Protocol packet  */
 #define ETH_P_ARP                 (0x0806)      /* Address Resolution packet */
 #define ETH_P_IPV6                (0x86dd)
 #define ETH_P_VLAN                (0x8100)
 #define ETH_P_DVLAN               (0x88a8)
+#define ETH_P_UNKNOWN             (0xffff)
 #define VLAN_VID_MASK             0x0fff
 #define IP_HEADER_VERSION_4       (4)
 #define IP_HEADER_VERSION_6       (6)
@@ -258,7 +294,7 @@ eth_get_l2_hdr_length(const void *p)
     case ETH_P_VLAN:
         return sizeof(struct eth_header) + sizeof(struct vlan_header);
     case ETH_P_DVLAN:
-        if (hvlan->h_proto == ETH_P_VLAN) {
+        if (be16_to_cpu(hvlan->h_proto) == ETH_P_VLAN) {
             return sizeof(struct eth_header) + 2 * sizeof(struct vlan_header);
         } else {
             return sizeof(struct eth_header) + sizeof(struct vlan_header);
@@ -268,6 +304,19 @@ eth_get_l2_hdr_length(const void *p)
     }
 }
 
+static inline uint32_t
+eth_get_l2_hdr_length_iov(const struct iovec *iov, int iovcnt)
+{
+    uint8_t p[sizeof(struct eth_header) + sizeof(struct vlan_header)];
+    size_t copied = iov_to_buf(iov, iovcnt, 0, p, ARRAY_SIZE(p));
+
+    if (copied < ARRAY_SIZE(p)) {
+        return copied;
+    }
+
+    return eth_get_l2_hdr_length(p);
+}
+
 static inline uint16_t
 eth_get_pkt_tci(const void *p)
 {
@@ -282,51 +331,67 @@ eth_get_pkt_tci(const void *p)
     }
 }
 
-static inline bool
-eth_strip_vlan(const void *p, uint8_t *new_ehdr_buf,
-               uint16_t *payload_offset, uint16_t *tci)
-{
-    uint16_t proto = be16_to_cpu(PKT_GET_ETH_HDR(p)->h_proto);
-    struct vlan_header *hvlan = PKT_GET_VLAN_HDR(p);
-    struct eth_header *new_ehdr = (struct eth_header *) new_ehdr_buf;
+bool
+eth_strip_vlan(const struct iovec *iov, int iovcnt, size_t iovoff,
+               uint8_t *new_ehdr_buf,
+               uint16_t *payload_offset, uint16_t *tci);
 
-    switch (proto) {
-    case ETH_P_VLAN:
-    case ETH_P_DVLAN:
-        memcpy(new_ehdr->h_source, PKT_GET_ETH_HDR(p)->h_source, ETH_ALEN);
-        memcpy(new_ehdr->h_dest, PKT_GET_ETH_HDR(p)->h_dest, ETH_ALEN);
-        new_ehdr->h_proto = hvlan->h_proto;
-        *tci = be16_to_cpu(hvlan->h_tci);
-        *payload_offset =
-            sizeof(struct eth_header) + sizeof(struct vlan_header);
-        if (be16_to_cpu(new_ehdr->h_proto) == ETH_P_VLAN) {
-            memcpy(PKT_GET_VLAN_HDR(new_ehdr),
-                   PKT_GET_DVLAN_HDR(p),
-                   sizeof(struct vlan_header));
-            *payload_offset += sizeof(struct vlan_header);
-        }
-        return true;
-    default:
-        return false;
-    }
-}
+bool
+eth_strip_vlan_ex(const struct iovec *iov, int iovcnt, size_t iovoff,
+                  uint16_t vet, uint8_t *new_ehdr_buf,
+                  uint16_t *payload_offset, uint16_t *tci);
 
-static inline uint16_t
-eth_get_l3_proto(const void *l2hdr, size_t l2hdr_len)
+uint16_t
+eth_get_l3_proto(const struct iovec *l2hdr_iov, int iovcnt, size_t l2hdr_len);
+
+void eth_setup_vlan_headers_ex(struct eth_header *ehdr, uint16_t vlan_tag,
+    uint16_t vlan_ethtype, bool *is_new);
+
+static inline void
+eth_setup_vlan_headers(struct eth_header *ehdr, uint16_t vlan_tag,
+    bool *is_new)
 {
-    uint8_t *proto_ptr = (uint8_t *) l2hdr + l2hdr_len - sizeof(uint16_t);
-    return be16_to_cpup((uint16_t *)proto_ptr);
+    eth_setup_vlan_headers_ex(ehdr, vlan_tag, ETH_P_VLAN, is_new);
 }
 
-void eth_setup_vlan_headers(struct eth_header *ehdr, uint16_t vlan_tag,
-    bool *is_new);
 
 uint8_t eth_get_gso_type(uint16_t l3_proto, uint8_t *l3_hdr, uint8_t l4proto);
 
-void eth_get_protocols(const uint8_t *headers,
-                       uint32_t hdr_length,
+typedef struct eth_ip6_hdr_info_st {
+    uint8_t l4proto;
+    size_t  full_hdr_len;
+    struct  ip6_header ip6_hdr;
+    bool    has_ext_hdrs;
+    bool    rss_ex_src_valid;
+    struct  in6_address rss_ex_src;
+    bool    rss_ex_dst_valid;
+    struct  in6_address rss_ex_dst;
+    bool    fragment;
+} eth_ip6_hdr_info;
+
+typedef struct eth_ip4_hdr_info_st {
+    struct ip_header ip4_hdr;
+    bool   fragment;
+} eth_ip4_hdr_info;
+
+typedef struct eth_l4_hdr_info_st {
+    union {
+        struct tcp_header tcp;
+        struct udp_header udp;
+    } hdr;
+
+    bool has_tcp_data;
+} eth_l4_hdr_info;
+
+void eth_get_protocols(const struct iovec *iov, int iovcnt,
                        bool *isip4, bool *isip6,
-                       bool *isudp, bool *istcp);
+                       bool *isudp, bool *istcp,
+                       size_t *l3hdr_off,
+                       size_t *l4hdr_off,
+                       size_t *l5hdr_off,
+                       eth_ip6_hdr_info *ip6hdr_info,
+                       eth_ip4_hdr_info *ip4hdr_info,
+                       eth_l4_hdr_info  *l4hdr_info);
 
 void eth_setup_ip4_fragmentation(const void *l2hdr, size_t l2hdr_len,
                                  void *l3hdr, size_t l3hdr_len,
@@ -337,11 +402,18 @@ void
 eth_fix_ip4_checksum(void *l3hdr, size_t l3hdr_len);
 
 uint32_t
-eth_calc_pseudo_hdr_csum(struct ip_header *iphdr, uint16_t csl);
+eth_calc_ip4_pseudo_hdr_csum(struct ip_header *iphdr,
+                             uint16_t csl,
+                             uint32_t *cso);
+
+uint32_t
+eth_calc_ip6_pseudo_hdr_csum(struct ip6_header *iphdr,
+                             uint16_t csl,
+                             uint8_t l4_proto,
+                             uint32_t *cso);
 
 bool
-eth_parse_ipv6_hdr(struct iovec *pkt, int pkt_frags,
-                   size_t ip6hdr_off, uint8_t *l4proto,
-                   size_t *full_hdr_len);
+eth_parse_ipv6_hdr(const struct iovec *pkt, int pkt_frags,
+                   size_t ip6hdr_off, eth_ip6_hdr_info *info);
 
 #endif
diff --git a/include/net/net.h b/include/net/net.h
index 73e4c466e2..a69e382ba7 100644
--- a/include/net/net.h
+++ b/include/net/net.h
@@ -9,6 +9,11 @@
 #include "migration/vmstate.h"
 #include "qapi-types.h"
 
+#define MAC_FMT "%02X:%02X:%02X:%02X:%02X:%02X"
+#define MAC_ARG(x) ((uint8_t *)(x))[0], ((uint8_t *)(x))[1], \
+                   ((uint8_t *)(x))[2], ((uint8_t *)(x))[3], \
+                   ((uint8_t *)(x))[4], ((uint8_t *)(x))[5]
+
 #define MAX_QUEUE_NUM 1024
 
 /* Maximum GSO packet size (64k) plus plenty of room for
@@ -57,6 +62,8 @@ typedef void (SetOffload)(NetClientState *, int, int, int, int, int);
 typedef void (SetVnetHdrLen)(NetClientState *, int);
 typedef int (SetVnetLE)(NetClientState *, bool);
 typedef int (SetVnetBE)(NetClientState *, bool);
+typedef struct SocketReadState SocketReadState;
+typedef void (SocketReadStateFinalize)(SocketReadState *rs);
 
 typedef struct NetClientInfo {
     NetClientOptionsKind type;
@@ -102,6 +109,15 @@ typedef struct NICState {
     bool peer_deleted;
 } NICState;
 
+struct SocketReadState {
+    int state; /* 0 = getting length, 1 = getting data */
+    uint32_t index;
+    uint32_t packet_len;
+    uint8_t buf[NET_BUFSIZE];
+    SocketReadStateFinalize *finalize;
+};
+
+int net_fill_rstate(SocketReadState *rs, const uint8_t *buf, int size);
 char *qemu_mac_strdup_printf(const uint8_t *macaddr);
 NetClientState *qemu_find_netdev(const char *id);
 int qemu_find_net_clients_except(const char *id, NetClientState **ncs,
@@ -160,6 +176,8 @@ ssize_t qemu_deliver_packet_iov(NetClientState *sender,
 
 void print_net_client(Monitor *mon, NetClientState *nc);
 void hmp_info_network(Monitor *mon, const QDict *qdict);
+void net_socket_rs_init(SocketReadState *rs,
+                        SocketReadStateFinalize *finalize);
 
 /* NIC info */
 
@@ -178,7 +196,6 @@ struct NICInfo {
 
 extern int nb_nics;
 extern NICInfo nd_table[MAX_NICS];
-extern int default_net;
 extern const char *host_net_devices[];
 
 /* from net.c */