Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next
Jeff Kirsher says:
====================
Intel Wired LAN Driver Updates 2014-06-26
This series contains updates to i40e and i40evf.
Kamil provides a cleanup patch to i40e where we do not need to acquire the
NVM for shadow RAM checksum calculation, since we only read the shadow RAM
through SRCTL register.
Paul provides a fix for handling HMC for big endian architectures for i40e
and i40evf.
Mitch provides four cleanup and fixes for i40evf. Fix an issue where if
the VF driver fails to complete early init, then rmmod can cause a softlock
when the driver tries to stop a watchdog timer that never got initialized.
So add a check to see if the timer is actually initialized before stopping
it. Make the function i40evf_send_api_ver() return more useful information,
instead of just returning -EIO by propagating firmware errors back to the
caller and log a message if the PF sends an invalid reply. Fix up a log
message that was missing a word, which makes the log message more readable.
Fix an initialization failure if many VFs are instantiated at the same time
and the VF module is autoloaded by simply resending firmware request if
there is no response the first time.
Jacob does a rename of the function i40e_ptp_enable() to
i40e_ptp_feature_enable(), like he did for ixgbe, to reduce possible
confusion and ambugity in the purpose of the function. Does follow on
PTP work on i40e, like he did for ixgbe, by breaking the PTP hardware
control from the ioctl command for timestamping mode. By doing this,
we can maintain state about the 1588 timestamping mode and properly
re-enable to the last known mode during a re-initialization of 1588 bits.
Anjali cleans up the i40e driver where TCP-IPv4 filters were being added
twice, which seems to be left over from when we had to add two PTYPEs for
one filter. Fixes the flow director sideband logic to detect when there
is a full flow director table. Also fixes the programming of FDIR where
a couple of fields in the descriptor setup that were not being
programmed, which left the opportunity for stale data to be pushed as
part of the descriptor next time it was used.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/drivers/net/ethernet/amd/Kconfig b/drivers/net/ethernet/amd/Kconfig
index bbaf36d..6e314db 100644
--- a/drivers/net/ethernet/amd/Kconfig
+++ b/drivers/net/ethernet/amd/Kconfig
@@ -182,6 +182,8 @@
depends on OF_NET
select PHYLIB
select AMD_XGBE_PHY
+ select BITREVERSE
+ select CRC32
---help---
This driver supports the AMD 10GbE Ethernet device found on an
AMD SoC.
diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-common.h b/drivers/net/ethernet/amd/xgbe/xgbe-common.h
index bf462ee8..ccbceba 100644
--- a/drivers/net/ethernet/amd/xgbe/xgbe-common.h
+++ b/drivers/net/ethernet/amd/xgbe/xgbe-common.h
@@ -276,13 +276,6 @@
#define MAC_PFR 0x0008
#define MAC_WTR 0x000c
#define MAC_HTR0 0x0010
-#define MAC_HTR1 0x0014
-#define MAC_HTR2 0x0018
-#define MAC_HTR3 0x001c
-#define MAC_HTR4 0x0020
-#define MAC_HTR5 0x0024
-#define MAC_HTR6 0x0028
-#define MAC_HTR7 0x002c
#define MAC_VLANTR 0x0050
#define MAC_VLANHTR 0x0058
#define MAC_VLANIR 0x0060
@@ -315,6 +308,7 @@
#define MAC_QTFCR_INC 4
#define MAC_MACA_INC 4
+#define MAC_HTR_INC 4
/* MAC register entry bit positions and sizes */
#define MAC_HWF0R_ADDMACADRSEL_INDEX 18
@@ -387,12 +381,16 @@
#define MAC_MACA1HR_AE_WIDTH 1
#define MAC_PFR_HMC_INDEX 2
#define MAC_PFR_HMC_WIDTH 1
+#define MAC_PFR_HPF_INDEX 10
+#define MAC_PFR_HPF_WIDTH 1
#define MAC_PFR_HUC_INDEX 1
#define MAC_PFR_HUC_WIDTH 1
#define MAC_PFR_PM_INDEX 4
#define MAC_PFR_PM_WIDTH 1
#define MAC_PFR_PR_INDEX 0
#define MAC_PFR_PR_WIDTH 1
+#define MAC_PFR_VTFE_INDEX 16
+#define MAC_PFR_VTFE_WIDTH 1
#define MAC_PMTCSR_MGKPKTEN_INDEX 1
#define MAC_PMTCSR_MGKPKTEN_WIDTH 1
#define MAC_PMTCSR_PWRDWN_INDEX 0
@@ -427,16 +425,30 @@
#define MAC_TCR_SS_WIDTH 2
#define MAC_TCR_TE_INDEX 0
#define MAC_TCR_TE_WIDTH 1
+#define MAC_VLANHTR_VLHT_INDEX 0
+#define MAC_VLANHTR_VLHT_WIDTH 16
+#define MAC_VLANIR_VLTI_INDEX 20
+#define MAC_VLANIR_VLTI_WIDTH 1
+#define MAC_VLANIR_CSVL_INDEX 19
+#define MAC_VLANIR_CSVL_WIDTH 1
#define MAC_VLANTR_DOVLTC_INDEX 20
#define MAC_VLANTR_DOVLTC_WIDTH 1
#define MAC_VLANTR_ERSVLM_INDEX 19
#define MAC_VLANTR_ERSVLM_WIDTH 1
#define MAC_VLANTR_ESVL_INDEX 18
#define MAC_VLANTR_ESVL_WIDTH 1
+#define MAC_VLANTR_ETV_INDEX 16
+#define MAC_VLANTR_ETV_WIDTH 1
#define MAC_VLANTR_EVLS_INDEX 21
#define MAC_VLANTR_EVLS_WIDTH 2
#define MAC_VLANTR_EVLRXS_INDEX 24
#define MAC_VLANTR_EVLRXS_WIDTH 1
+#define MAC_VLANTR_VL_INDEX 0
+#define MAC_VLANTR_VL_WIDTH 16
+#define MAC_VLANTR_VTHM_INDEX 25
+#define MAC_VLANTR_VTHM_WIDTH 1
+#define MAC_VLANTR_VTIM_INDEX 17
+#define MAC_VLANTR_VTIM_WIDTH 1
#define MAC_VR_DEVID_INDEX 8
#define MAC_VR_DEVID_WIDTH 8
#define MAC_VR_SNPSVER_INDEX 0
diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-debugfs.c b/drivers/net/ethernet/amd/xgbe/xgbe-debugfs.c
index 6bb76d5..8119858 100644
--- a/drivers/net/ethernet/amd/xgbe/xgbe-debugfs.c
+++ b/drivers/net/ethernet/amd/xgbe/xgbe-debugfs.c
@@ -151,7 +151,7 @@
{
char workarea[32];
ssize_t len;
- unsigned int scan_value;
+ int ret;
if (*ppos != 0)
return 0;
@@ -165,10 +165,9 @@
return len;
workarea[len] = '\0';
- if (sscanf(workarea, "%x", &scan_value) == 1)
- *value = scan_value;
- else
- return -EIO;
+ ret = kstrtouint(workarea, 0, value);
+ if (ret)
+ return ret;
return len;
}
diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-desc.c b/drivers/net/ethernet/amd/xgbe/xgbe-desc.c
index 6f1c859..a9ce56d 100644
--- a/drivers/net/ethernet/amd/xgbe/xgbe-desc.c
+++ b/drivers/net/ethernet/amd/xgbe/xgbe-desc.c
@@ -131,7 +131,7 @@
if (ring->rdata) {
for (i = 0; i < ring->rdesc_count; i++) {
- rdata = GET_DESC_DATA(ring, i);
+ rdata = XGBE_GET_DESC_DATA(ring, i);
xgbe_unmap_skb(pdata, rdata);
}
@@ -256,7 +256,7 @@
rdesc_dma = ring->rdesc_dma;
for (j = 0; j < ring->rdesc_count; j++) {
- rdata = GET_DESC_DATA(ring, j);
+ rdata = XGBE_GET_DESC_DATA(ring, j);
rdata->rdesc = rdesc;
rdata->rdesc_dma = rdesc_dma;
@@ -298,7 +298,7 @@
rdesc_dma = ring->rdesc_dma;
for (j = 0; j < ring->rdesc_count; j++) {
- rdata = GET_DESC_DATA(ring, j);
+ rdata = XGBE_GET_DESC_DATA(ring, j);
rdata->rdesc = rdesc;
rdata->rdesc_dma = rdesc_dma;
@@ -392,7 +392,7 @@
if ((tso && (packet->mss != ring->tx.cur_mss)) ||
(vlan && (packet->vlan_ctag != ring->tx.cur_vlan_ctag)))
cur_index++;
- rdata = GET_DESC_DATA(ring, cur_index);
+ rdata = XGBE_GET_DESC_DATA(ring, cur_index);
if (tso) {
DBGPR(" TSO packet\n");
@@ -413,12 +413,12 @@
packet->length += packet->header_len;
cur_index++;
- rdata = GET_DESC_DATA(ring, cur_index);
+ rdata = XGBE_GET_DESC_DATA(ring, cur_index);
}
/* Map the (remainder of the) packet */
for (datalen = skb_headlen(skb) - offset; datalen; ) {
- len = min_t(unsigned int, datalen, TX_MAX_BUF_SIZE);
+ len = min_t(unsigned int, datalen, XGBE_TX_MAX_BUF_SIZE);
skb_dma = dma_map_single(pdata->dev, skb->data + offset, len,
DMA_TO_DEVICE);
@@ -437,7 +437,7 @@
packet->length += len;
cur_index++;
- rdata = GET_DESC_DATA(ring, cur_index);
+ rdata = XGBE_GET_DESC_DATA(ring, cur_index);
}
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
@@ -447,7 +447,8 @@
offset = 0;
for (datalen = skb_frag_size(frag); datalen; ) {
- len = min_t(unsigned int, datalen, TX_MAX_BUF_SIZE);
+ len = min_t(unsigned int, datalen,
+ XGBE_TX_MAX_BUF_SIZE);
skb_dma = skb_frag_dma_map(pdata->dev, frag, offset,
len, DMA_TO_DEVICE);
@@ -468,7 +469,7 @@
packet->length += len;
cur_index++;
- rdata = GET_DESC_DATA(ring, cur_index);
+ rdata = XGBE_GET_DESC_DATA(ring, cur_index);
}
}
@@ -484,7 +485,7 @@
err_out:
while (start_index < cur_index) {
- rdata = GET_DESC_DATA(ring, start_index++);
+ rdata = XGBE_GET_DESC_DATA(ring, start_index++);
xgbe_unmap_skb(pdata, rdata);
}
@@ -507,7 +508,7 @@
ring->rx.realloc_index);
for (i = 0; i < ring->dirty; i++) {
- rdata = GET_DESC_DATA(ring, ring->rx.realloc_index);
+ rdata = XGBE_GET_DESC_DATA(ring, ring->rx.realloc_index);
/* Reset rdata values */
xgbe_unmap_skb(pdata, rdata);
diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-dev.c b/drivers/net/ethernet/amd/xgbe/xgbe-dev.c
index 002293b..a56069c9 100644
--- a/drivers/net/ethernet/amd/xgbe/xgbe-dev.c
+++ b/drivers/net/ethernet/amd/xgbe/xgbe-dev.c
@@ -116,6 +116,8 @@
#include <linux/phy.h>
#include <linux/clk.h>
+#include <linux/bitrev.h>
+#include <linux/crc32.h>
#include "xgbe.h"
#include "xgbe-common.h"
@@ -547,24 +549,16 @@
return 0;
}
-static int xgbe_set_addn_mac_addrs(struct xgbe_prv_data *pdata,
- unsigned int am_mode)
+static void xgbe_set_mac_reg(struct xgbe_prv_data *pdata,
+ struct netdev_hw_addr *ha, unsigned int *mac_reg)
{
- struct netdev_hw_addr *ha;
- unsigned int mac_reg;
unsigned int mac_addr_hi, mac_addr_lo;
u8 *mac_addr;
- unsigned int i;
- XGMAC_IOWRITE_BITS(pdata, MAC_PFR, HUC, 0);
- XGMAC_IOWRITE_BITS(pdata, MAC_PFR, HMC, 0);
+ mac_addr_lo = 0;
+ mac_addr_hi = 0;
- i = 0;
- mac_reg = MAC_MACA1HR;
-
- netdev_for_each_uc_addr(ha, pdata->netdev) {
- mac_addr_lo = 0;
- mac_addr_hi = 0;
+ if (ha) {
mac_addr = (u8 *)&mac_addr_lo;
mac_addr[0] = ha->addr[0];
mac_addr[1] = ha->addr[1];
@@ -574,54 +568,93 @@
mac_addr[0] = ha->addr[4];
mac_addr[1] = ha->addr[5];
- DBGPR(" adding unicast address %pM at 0x%04x\n",
- ha->addr, mac_reg);
+ DBGPR(" adding mac address %pM at 0x%04x\n", ha->addr,
+ *mac_reg);
XGMAC_SET_BITS(mac_addr_hi, MAC_MACA1HR, AE, 1);
-
- XGMAC_IOWRITE(pdata, mac_reg, mac_addr_hi);
- mac_reg += MAC_MACA_INC;
- XGMAC_IOWRITE(pdata, mac_reg, mac_addr_lo);
- mac_reg += MAC_MACA_INC;
-
- i++;
}
- if (!am_mode) {
- netdev_for_each_mc_addr(ha, pdata->netdev) {
- mac_addr_lo = 0;
- mac_addr_hi = 0;
- mac_addr = (u8 *)&mac_addr_lo;
- mac_addr[0] = ha->addr[0];
- mac_addr[1] = ha->addr[1];
- mac_addr[2] = ha->addr[2];
- mac_addr[3] = ha->addr[3];
- mac_addr = (u8 *)&mac_addr_hi;
- mac_addr[0] = ha->addr[4];
- mac_addr[1] = ha->addr[5];
+ XGMAC_IOWRITE(pdata, *mac_reg, mac_addr_hi);
+ *mac_reg += MAC_MACA_INC;
+ XGMAC_IOWRITE(pdata, *mac_reg, mac_addr_lo);
+ *mac_reg += MAC_MACA_INC;
+}
- DBGPR(" adding multicast address %pM at 0x%04x\n",
- ha->addr, mac_reg);
+static void xgbe_set_mac_addn_addrs(struct xgbe_prv_data *pdata)
+{
+ struct net_device *netdev = pdata->netdev;
+ struct netdev_hw_addr *ha;
+ unsigned int mac_reg;
+ unsigned int addn_macs;
- XGMAC_SET_BITS(mac_addr_hi, MAC_MACA1HR, AE, 1);
+ mac_reg = MAC_MACA1HR;
+ addn_macs = pdata->hw_feat.addn_mac;
- XGMAC_IOWRITE(pdata, mac_reg, mac_addr_hi);
- mac_reg += MAC_MACA_INC;
- XGMAC_IOWRITE(pdata, mac_reg, mac_addr_lo);
- mac_reg += MAC_MACA_INC;
+ if (netdev_uc_count(netdev) > addn_macs) {
+ xgbe_set_promiscuous_mode(pdata, 1);
+ } else {
+ netdev_for_each_uc_addr(ha, netdev) {
+ xgbe_set_mac_reg(pdata, ha, &mac_reg);
+ addn_macs--;
+ }
- i++;
+ if (netdev_mc_count(netdev) > addn_macs) {
+ xgbe_set_all_multicast_mode(pdata, 1);
+ } else {
+ netdev_for_each_mc_addr(ha, netdev) {
+ xgbe_set_mac_reg(pdata, ha, &mac_reg);
+ addn_macs--;
+ }
}
}
/* Clear remaining additional MAC address entries */
- for (; i < pdata->hw_feat.addn_mac; i++) {
- XGMAC_IOWRITE(pdata, mac_reg, 0);
- mac_reg += MAC_MACA_INC;
- XGMAC_IOWRITE(pdata, mac_reg, 0);
- mac_reg += MAC_MACA_INC;
+ while (addn_macs--)
+ xgbe_set_mac_reg(pdata, NULL, &mac_reg);
+}
+
+static void xgbe_set_mac_hash_table(struct xgbe_prv_data *pdata)
+{
+ struct net_device *netdev = pdata->netdev;
+ struct netdev_hw_addr *ha;
+ unsigned int hash_reg;
+ unsigned int hash_table_shift, hash_table_count;
+ u32 hash_table[XGBE_MAC_HASH_TABLE_SIZE];
+ u32 crc;
+ unsigned int i;
+
+ hash_table_shift = 26 - (pdata->hw_feat.hash_table_size >> 7);
+ hash_table_count = pdata->hw_feat.hash_table_size / 32;
+ memset(hash_table, 0, sizeof(hash_table));
+
+ /* Build the MAC Hash Table register values */
+ netdev_for_each_uc_addr(ha, netdev) {
+ crc = bitrev32(~crc32_le(~0, ha->addr, ETH_ALEN));
+ crc >>= hash_table_shift;
+ hash_table[crc >> 5] |= (1 << (crc & 0x1f));
}
+ netdev_for_each_mc_addr(ha, netdev) {
+ crc = bitrev32(~crc32_le(~0, ha->addr, ETH_ALEN));
+ crc >>= hash_table_shift;
+ hash_table[crc >> 5] |= (1 << (crc & 0x1f));
+ }
+
+ /* Set the MAC Hash Table registers */
+ hash_reg = MAC_HTR0;
+ for (i = 0; i < hash_table_count; i++) {
+ XGMAC_IOWRITE(pdata, hash_reg, hash_table[i]);
+ hash_reg += MAC_HTR_INC;
+ }
+}
+
+static int xgbe_add_mac_addresses(struct xgbe_prv_data *pdata)
+{
+ if (pdata->hw_feat.hash_table_size)
+ xgbe_set_mac_hash_table(pdata);
+ else
+ xgbe_set_mac_addn_addrs(pdata);
+
return 0;
}
@@ -738,6 +771,89 @@
return 0;
}
+static int xgbe_enable_rx_vlan_filtering(struct xgbe_prv_data *pdata)
+{
+ /* Enable VLAN filtering */
+ XGMAC_IOWRITE_BITS(pdata, MAC_PFR, VTFE, 1);
+
+ /* Enable VLAN Hash Table filtering */
+ XGMAC_IOWRITE_BITS(pdata, MAC_VLANTR, VTHM, 1);
+
+ /* Disable VLAN tag inverse matching */
+ XGMAC_IOWRITE_BITS(pdata, MAC_VLANTR, VTIM, 0);
+
+ /* Only filter on the lower 12-bits of the VLAN tag */
+ XGMAC_IOWRITE_BITS(pdata, MAC_VLANTR, ETV, 1);
+
+ /* In order for the VLAN Hash Table filtering to be effective,
+ * the VLAN tag identifier in the VLAN Tag Register must not
+ * be zero. Set the VLAN tag identifier to "1" to enable the
+ * VLAN Hash Table filtering. This implies that a VLAN tag of
+ * 1 will always pass filtering.
+ */
+ XGMAC_IOWRITE_BITS(pdata, MAC_VLANTR, VL, 1);
+
+ return 0;
+}
+
+static int xgbe_disable_rx_vlan_filtering(struct xgbe_prv_data *pdata)
+{
+ /* Disable VLAN filtering */
+ XGMAC_IOWRITE_BITS(pdata, MAC_PFR, VTFE, 0);
+
+ return 0;
+}
+
+#ifndef CRCPOLY_LE
+#define CRCPOLY_LE 0xedb88320
+#endif
+static u32 xgbe_vid_crc32_le(__le16 vid_le)
+{
+ u32 poly = CRCPOLY_LE;
+ u32 crc = ~0;
+ u32 temp = 0;
+ unsigned char *data = (unsigned char *)&vid_le;
+ unsigned char data_byte = 0;
+ int i, bits;
+
+ bits = get_bitmask_order(VLAN_VID_MASK);
+ for (i = 0; i < bits; i++) {
+ if ((i % 8) == 0)
+ data_byte = data[i / 8];
+
+ temp = ((crc & 1) ^ data_byte) & 1;
+ crc >>= 1;
+ data_byte >>= 1;
+
+ if (temp)
+ crc ^= poly;
+ }
+
+ return crc;
+}
+
+static int xgbe_update_vlan_hash_table(struct xgbe_prv_data *pdata)
+{
+ u32 crc;
+ u16 vid;
+ __le16 vid_le;
+ u16 vlan_hash_table = 0;
+
+ /* Generate the VLAN Hash Table value */
+ for_each_set_bit(vid, pdata->active_vlans, VLAN_N_VID) {
+ /* Get the CRC32 value of the VLAN ID */
+ vid_le = cpu_to_le16(vid);
+ crc = bitrev32(~xgbe_vid_crc32_le(vid_le)) >> 28;
+
+ vlan_hash_table |= (1 << crc);
+ }
+
+ /* Set the VLAN Hash Table filtering register */
+ XGMAC_IOWRITE_BITS(pdata, MAC_VLANHTR, VLHT, vlan_hash_table);
+
+ return 0;
+}
+
static void xgbe_tx_desc_reset(struct xgbe_ring_data *rdata)
{
struct xgbe_ring_desc *rdesc = rdata->rdesc;
@@ -766,7 +882,7 @@
/* Initialze all descriptors */
for (i = 0; i < ring->rdesc_count; i++) {
- rdata = GET_DESC_DATA(ring, i);
+ rdata = XGBE_GET_DESC_DATA(ring, i);
rdesc = rdata->rdesc;
/* Initialize Tx descriptor
@@ -791,7 +907,7 @@
XGMAC_DMA_IOWRITE(channel, DMA_CH_TDRLR, ring->rdesc_count - 1);
/* Update the starting address of descriptor ring */
- rdata = GET_DESC_DATA(ring, start_index);
+ rdata = XGBE_GET_DESC_DATA(ring, start_index);
XGMAC_DMA_IOWRITE(channel, DMA_CH_TDLR_HI,
upper_32_bits(rdata->rdesc_dma));
XGMAC_DMA_IOWRITE(channel, DMA_CH_TDLR_LO,
@@ -848,7 +964,7 @@
/* Initialize all descriptors */
for (i = 0; i < ring->rdesc_count; i++) {
- rdata = GET_DESC_DATA(ring, i);
+ rdata = XGBE_GET_DESC_DATA(ring, i);
rdesc = rdata->rdesc;
/* Initialize Rx descriptor
@@ -882,14 +998,14 @@
XGMAC_DMA_IOWRITE(channel, DMA_CH_RDRLR, ring->rdesc_count - 1);
/* Update the starting address of descriptor ring */
- rdata = GET_DESC_DATA(ring, start_index);
+ rdata = XGBE_GET_DESC_DATA(ring, start_index);
XGMAC_DMA_IOWRITE(channel, DMA_CH_RDLR_HI,
upper_32_bits(rdata->rdesc_dma));
XGMAC_DMA_IOWRITE(channel, DMA_CH_RDLR_LO,
lower_32_bits(rdata->rdesc_dma));
/* Update the Rx Descriptor Tail Pointer */
- rdata = GET_DESC_DATA(ring, start_index + ring->rdesc_count - 1);
+ rdata = XGBE_GET_DESC_DATA(ring, start_index + ring->rdesc_count - 1);
XGMAC_DMA_IOWRITE(channel, DMA_CH_RDTR_LO,
lower_32_bits(rdata->rdesc_dma));
@@ -933,7 +1049,7 @@
if (tx_coalesce && !channel->tx_timer_active)
ring->coalesce_count = 0;
- rdata = GET_DESC_DATA(ring, ring->cur);
+ rdata = XGBE_GET_DESC_DATA(ring, ring->cur);
rdesc = rdata->rdesc;
/* Create a context descriptor if this is a TSO packet */
@@ -977,7 +1093,7 @@
}
ring->cur++;
- rdata = GET_DESC_DATA(ring, ring->cur);
+ rdata = XGBE_GET_DESC_DATA(ring, ring->cur);
rdesc = rdata->rdesc;
}
@@ -1034,7 +1150,7 @@
for (i = ring->cur - start_index + 1; i < packet->rdesc_count; i++) {
ring->cur++;
- rdata = GET_DESC_DATA(ring, ring->cur);
+ rdata = XGBE_GET_DESC_DATA(ring, ring->cur);
rdesc = rdata->rdesc;
/* Update buffer address */
@@ -1074,7 +1190,7 @@
wmb();
/* Set OWN bit for the first descriptor */
- rdata = GET_DESC_DATA(ring, start_index);
+ rdata = XGBE_GET_DESC_DATA(ring, start_index);
rdesc = rdata->rdesc;
XGMAC_SET_BITS_LE(rdesc->desc3, TX_NORMAL_DESC3, OWN, 1);
@@ -1088,7 +1204,7 @@
/* Issue a poll command to Tx DMA by writing address
* of next immediate free descriptor */
ring->cur++;
- rdata = GET_DESC_DATA(ring, ring->cur);
+ rdata = XGBE_GET_DESC_DATA(ring, ring->cur);
XGMAC_DMA_IOWRITE(channel, DMA_CH_TDTR_LO,
lower_32_bits(rdata->rdesc_dma));
@@ -1113,11 +1229,12 @@
struct xgbe_ring_data *rdata;
struct xgbe_ring_desc *rdesc;
struct xgbe_packet_data *packet = &ring->packet_data;
+ struct net_device *netdev = channel->pdata->netdev;
unsigned int err, etlt;
DBGPR("-->xgbe_dev_read: cur = %d\n", ring->cur);
- rdata = GET_DESC_DATA(ring, ring->cur);
+ rdata = XGBE_GET_DESC_DATA(ring, ring->cur);
rdesc = rdata->rdesc;
/* Check for data availability */
@@ -1153,7 +1270,8 @@
DBGPR(" err=%u, etlt=%#x\n", err, etlt);
if (!err || (err && !etlt)) {
- if (etlt == 0x09) {
+ if ((etlt == 0x09) &&
+ (netdev->features & NETIF_F_HW_VLAN_CTAG_RX)) {
XGMAC_SET_BITS(packet->attributes, RX_PACKET_ATTRIBUTES,
VLAN_CTAG, 1);
packet->vlan_ctag = XGMAC_GET_BITS_LE(rdesc->desc0,
@@ -1195,7 +1313,7 @@
if (int_state == XGMAC_INT_STATE_SAVE) {
channel->saved_ier = XGMAC_DMA_IOREAD(channel, DMA_CH_IER);
- channel->saved_ier &= DMA_INTERRUPT_MASK;
+ channel->saved_ier &= XGBE_DMA_INTERRUPT_MASK;
} else {
dma_ch_ier = XGMAC_DMA_IOREAD(channel, DMA_CH_IER);
dma_ch_ier |= channel->saved_ier;
@@ -1275,7 +1393,7 @@
xgbe_save_interrupt_status(channel, XGMAC_INT_STATE_SAVE);
dma_ch_ier = XGMAC_DMA_IOREAD(channel, DMA_CH_IER);
- dma_ch_ier &= ~DMA_INTERRUPT_MASK;
+ dma_ch_ier &= ~XGBE_DMA_INTERRUPT_MASK;
XGMAC_DMA_IOWRITE(channel, DMA_CH_IER, dma_ch_ier);
break;
default:
@@ -1342,23 +1460,23 @@
unsigned int arcache, awcache;
arcache = 0;
- XGMAC_SET_BITS(arcache, DMA_AXIARCR, DRC, DMA_ARCACHE_SETTING);
- XGMAC_SET_BITS(arcache, DMA_AXIARCR, DRD, DMA_ARDOMAIN_SETTING);
- XGMAC_SET_BITS(arcache, DMA_AXIARCR, TEC, DMA_ARCACHE_SETTING);
- XGMAC_SET_BITS(arcache, DMA_AXIARCR, TED, DMA_ARDOMAIN_SETTING);
- XGMAC_SET_BITS(arcache, DMA_AXIARCR, THC, DMA_ARCACHE_SETTING);
- XGMAC_SET_BITS(arcache, DMA_AXIARCR, THD, DMA_ARDOMAIN_SETTING);
+ XGMAC_SET_BITS(arcache, DMA_AXIARCR, DRC, XGBE_DMA_ARCACHE);
+ XGMAC_SET_BITS(arcache, DMA_AXIARCR, DRD, XGBE_DMA_ARDOMAIN);
+ XGMAC_SET_BITS(arcache, DMA_AXIARCR, TEC, XGBE_DMA_ARCACHE);
+ XGMAC_SET_BITS(arcache, DMA_AXIARCR, TED, XGBE_DMA_ARDOMAIN);
+ XGMAC_SET_BITS(arcache, DMA_AXIARCR, THC, XGBE_DMA_ARCACHE);
+ XGMAC_SET_BITS(arcache, DMA_AXIARCR, THD, XGBE_DMA_ARDOMAIN);
XGMAC_IOWRITE(pdata, DMA_AXIARCR, arcache);
awcache = 0;
- XGMAC_SET_BITS(awcache, DMA_AXIAWCR, DWC, DMA_AWCACHE_SETTING);
- XGMAC_SET_BITS(awcache, DMA_AXIAWCR, DWD, DMA_AWDOMAIN_SETTING);
- XGMAC_SET_BITS(awcache, DMA_AXIAWCR, RPC, DMA_AWCACHE_SETTING);
- XGMAC_SET_BITS(awcache, DMA_AXIAWCR, RPD, DMA_AWDOMAIN_SETTING);
- XGMAC_SET_BITS(awcache, DMA_AXIAWCR, RHC, DMA_AWCACHE_SETTING);
- XGMAC_SET_BITS(awcache, DMA_AXIAWCR, RHD, DMA_AWDOMAIN_SETTING);
- XGMAC_SET_BITS(awcache, DMA_AXIAWCR, TDC, DMA_AWCACHE_SETTING);
- XGMAC_SET_BITS(awcache, DMA_AXIAWCR, TDD, DMA_AWDOMAIN_SETTING);
+ XGMAC_SET_BITS(awcache, DMA_AXIAWCR, DWC, XGBE_DMA_AWCACHE);
+ XGMAC_SET_BITS(awcache, DMA_AXIAWCR, DWD, XGBE_DMA_AWDOMAIN);
+ XGMAC_SET_BITS(awcache, DMA_AXIAWCR, RPC, XGBE_DMA_AWCACHE);
+ XGMAC_SET_BITS(awcache, DMA_AXIAWCR, RPD, XGBE_DMA_AWDOMAIN);
+ XGMAC_SET_BITS(awcache, DMA_AXIAWCR, RHC, XGBE_DMA_AWCACHE);
+ XGMAC_SET_BITS(awcache, DMA_AXIAWCR, RHD, XGBE_DMA_AWDOMAIN);
+ XGMAC_SET_BITS(awcache, DMA_AXIAWCR, TDC, XGBE_DMA_AWCACHE);
+ XGMAC_SET_BITS(awcache, DMA_AXIAWCR, TDD, XGBE_DMA_AWDOMAIN);
XGMAC_IOWRITE(pdata, DMA_AXIAWCR, awcache);
}
@@ -1388,66 +1506,66 @@
/* Calculate Tx/Rx fifo share per queue */
switch (fifo_size) {
case 0:
- q_fifo_size = FIFO_SIZE_B(128);
+ q_fifo_size = XGBE_FIFO_SIZE_B(128);
break;
case 1:
- q_fifo_size = FIFO_SIZE_B(256);
+ q_fifo_size = XGBE_FIFO_SIZE_B(256);
break;
case 2:
- q_fifo_size = FIFO_SIZE_B(512);
+ q_fifo_size = XGBE_FIFO_SIZE_B(512);
break;
case 3:
- q_fifo_size = FIFO_SIZE_KB(1);
+ q_fifo_size = XGBE_FIFO_SIZE_KB(1);
break;
case 4:
- q_fifo_size = FIFO_SIZE_KB(2);
+ q_fifo_size = XGBE_FIFO_SIZE_KB(2);
break;
case 5:
- q_fifo_size = FIFO_SIZE_KB(4);
+ q_fifo_size = XGBE_FIFO_SIZE_KB(4);
break;
case 6:
- q_fifo_size = FIFO_SIZE_KB(8);
+ q_fifo_size = XGBE_FIFO_SIZE_KB(8);
break;
case 7:
- q_fifo_size = FIFO_SIZE_KB(16);
+ q_fifo_size = XGBE_FIFO_SIZE_KB(16);
break;
case 8:
- q_fifo_size = FIFO_SIZE_KB(32);
+ q_fifo_size = XGBE_FIFO_SIZE_KB(32);
break;
case 9:
- q_fifo_size = FIFO_SIZE_KB(64);
+ q_fifo_size = XGBE_FIFO_SIZE_KB(64);
break;
case 10:
- q_fifo_size = FIFO_SIZE_KB(128);
+ q_fifo_size = XGBE_FIFO_SIZE_KB(128);
break;
case 11:
- q_fifo_size = FIFO_SIZE_KB(256);
+ q_fifo_size = XGBE_FIFO_SIZE_KB(256);
break;
}
q_fifo_size = q_fifo_size / queue_count;
/* Set the queue fifo size programmable value */
- if (q_fifo_size >= FIFO_SIZE_KB(256))
+ if (q_fifo_size >= XGBE_FIFO_SIZE_KB(256))
p_fifo = XGMAC_MTL_FIFO_SIZE_256K;
- else if (q_fifo_size >= FIFO_SIZE_KB(128))
+ else if (q_fifo_size >= XGBE_FIFO_SIZE_KB(128))
p_fifo = XGMAC_MTL_FIFO_SIZE_128K;
- else if (q_fifo_size >= FIFO_SIZE_KB(64))
+ else if (q_fifo_size >= XGBE_FIFO_SIZE_KB(64))
p_fifo = XGMAC_MTL_FIFO_SIZE_64K;
- else if (q_fifo_size >= FIFO_SIZE_KB(32))
+ else if (q_fifo_size >= XGBE_FIFO_SIZE_KB(32))
p_fifo = XGMAC_MTL_FIFO_SIZE_32K;
- else if (q_fifo_size >= FIFO_SIZE_KB(16))
+ else if (q_fifo_size >= XGBE_FIFO_SIZE_KB(16))
p_fifo = XGMAC_MTL_FIFO_SIZE_16K;
- else if (q_fifo_size >= FIFO_SIZE_KB(8))
+ else if (q_fifo_size >= XGBE_FIFO_SIZE_KB(8))
p_fifo = XGMAC_MTL_FIFO_SIZE_8K;
- else if (q_fifo_size >= FIFO_SIZE_KB(4))
+ else if (q_fifo_size >= XGBE_FIFO_SIZE_KB(4))
p_fifo = XGMAC_MTL_FIFO_SIZE_4K;
- else if (q_fifo_size >= FIFO_SIZE_KB(2))
+ else if (q_fifo_size >= XGBE_FIFO_SIZE_KB(2))
p_fifo = XGMAC_MTL_FIFO_SIZE_2K;
- else if (q_fifo_size >= FIFO_SIZE_KB(1))
+ else if (q_fifo_size >= XGBE_FIFO_SIZE_KB(1))
p_fifo = XGMAC_MTL_FIFO_SIZE_1K;
- else if (q_fifo_size >= FIFO_SIZE_B(512))
+ else if (q_fifo_size >= XGBE_FIFO_SIZE_B(512))
p_fifo = XGMAC_MTL_FIFO_SIZE_512;
- else if (q_fifo_size >= FIFO_SIZE_B(256))
+ else if (q_fifo_size >= XGBE_FIFO_SIZE_B(256))
p_fifo = XGMAC_MTL_FIFO_SIZE_256;
return p_fifo;
@@ -1520,6 +1638,13 @@
static void xgbe_config_mac_address(struct xgbe_prv_data *pdata)
{
xgbe_set_mac_address(pdata, pdata->netdev->dev_addr);
+
+ /* Filtering is done using perfect filtering and hash filtering */
+ if (pdata->hw_feat.hash_table_size) {
+ XGMAC_IOWRITE_BITS(pdata, MAC_PFR, HPF, 1);
+ XGMAC_IOWRITE_BITS(pdata, MAC_PFR, HUC, 1);
+ XGMAC_IOWRITE_BITS(pdata, MAC_PFR, HMC, 1);
+ }
}
static void xgbe_config_jumbo_enable(struct xgbe_prv_data *pdata)
@@ -1541,6 +1666,18 @@
static void xgbe_config_vlan_support(struct xgbe_prv_data *pdata)
{
+ /* Indicate that VLAN Tx CTAGs come from context descriptors */
+ XGMAC_IOWRITE_BITS(pdata, MAC_VLANIR, CSVL, 0);
+ XGMAC_IOWRITE_BITS(pdata, MAC_VLANIR, VLTI, 1);
+
+ /* Set the current VLAN Hash Table register value */
+ xgbe_update_vlan_hash_table(pdata);
+
+ if (pdata->netdev->features & NETIF_F_HW_VLAN_CTAG_FILTER)
+ xgbe_enable_rx_vlan_filtering(pdata);
+ else
+ xgbe_disable_rx_vlan_filtering(pdata);
+
if (pdata->netdev->features & NETIF_F_HW_VLAN_CTAG_RX)
xgbe_enable_rx_vlan_stripping(pdata);
else
@@ -2104,7 +2241,7 @@
hw_if->set_promiscuous_mode = xgbe_set_promiscuous_mode;
hw_if->set_all_multicast_mode = xgbe_set_all_multicast_mode;
- hw_if->set_addn_mac_addrs = xgbe_set_addn_mac_addrs;
+ hw_if->add_mac_addresses = xgbe_add_mac_addresses;
hw_if->set_mac_address = xgbe_set_mac_address;
hw_if->enable_rx_csum = xgbe_enable_rx_csum;
@@ -2112,6 +2249,9 @@
hw_if->enable_rx_vlan_stripping = xgbe_enable_rx_vlan_stripping;
hw_if->disable_rx_vlan_stripping = xgbe_disable_rx_vlan_stripping;
+ hw_if->enable_rx_vlan_filtering = xgbe_enable_rx_vlan_filtering;
+ hw_if->disable_rx_vlan_filtering = xgbe_disable_rx_vlan_filtering;
+ hw_if->update_vlan_hash_table = xgbe_update_vlan_hash_table;
hw_if->read_mmd_regs = xgbe_read_mmd_regs;
hw_if->write_mmd_regs = xgbe_write_mmd_regs;
diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-drv.c b/drivers/net/ethernet/amd/xgbe/xgbe-drv.c
index cfe3d93..72dd611 100644
--- a/drivers/net/ethernet/amd/xgbe/xgbe-drv.c
+++ b/drivers/net/ethernet/amd/xgbe/xgbe-drv.c
@@ -144,9 +144,10 @@
}
rx_buf_size = mtu + ETH_HLEN + ETH_FCS_LEN + VLAN_HLEN;
- if (rx_buf_size < RX_MIN_BUF_SIZE)
- rx_buf_size = RX_MIN_BUF_SIZE;
- rx_buf_size = (rx_buf_size + RX_BUF_ALIGN - 1) & ~(RX_BUF_ALIGN - 1);
+ if (rx_buf_size < XGBE_RX_MIN_BUF_SIZE)
+ rx_buf_size = XGBE_RX_MIN_BUF_SIZE;
+ rx_buf_size = (rx_buf_size + XGBE_RX_BUF_ALIGN - 1) &
+ ~(XGBE_RX_BUF_ALIGN - 1);
return rx_buf_size;
}
@@ -377,6 +378,21 @@
hw_feat->pps_out_num = XGMAC_GET_BITS(mac_hfr2, MAC_HWF2R, PPSOUTNUM);
hw_feat->aux_snap_num = XGMAC_GET_BITS(mac_hfr2, MAC_HWF2R, AUXSNAPNUM);
+ /* Translate the Hash Table size into actual number */
+ switch (hw_feat->hash_table_size) {
+ case 0:
+ break;
+ case 1:
+ hw_feat->hash_table_size = 64;
+ break;
+ case 2:
+ hw_feat->hash_table_size = 128;
+ break;
+ case 3:
+ hw_feat->hash_table_size = 256;
+ break;
+ }
+
/* The Queue and Channel counts are zero based so increment them
* to get the actual number
*/
@@ -446,7 +462,7 @@
break;
for (j = 0; j < ring->rdesc_count; j++) {
- rdata = GET_DESC_DATA(ring, j);
+ rdata = XGBE_GET_DESC_DATA(ring, j);
desc_if->unmap_skb(pdata, rdata);
}
}
@@ -471,7 +487,7 @@
break;
for (j = 0; j < ring->rdesc_count; j++) {
- rdata = GET_DESC_DATA(ring, j);
+ rdata = XGBE_GET_DESC_DATA(ring, j);
desc_if->unmap_skb(pdata, rdata);
}
}
@@ -726,14 +742,14 @@
for (len = skb_headlen(skb); len;) {
packet->rdesc_count++;
- len -= min_t(unsigned int, len, TX_MAX_BUF_SIZE);
+ len -= min_t(unsigned int, len, XGBE_TX_MAX_BUF_SIZE);
}
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
frag = &skb_shinfo(skb)->frags[i];
for (len = skb_frag_size(frag); len; ) {
packet->rdesc_count++;
- len -= min_t(unsigned int, len, TX_MAX_BUF_SIZE);
+ len -= min_t(unsigned int, len, XGBE_TX_MAX_BUF_SIZE);
}
}
}
@@ -911,18 +927,10 @@
pr_mode = ((netdev->flags & IFF_PROMISC) != 0);
am_mode = ((netdev->flags & IFF_ALLMULTI) != 0);
- if (netdev_uc_count(netdev) > pdata->hw_feat.addn_mac)
- pr_mode = 1;
- if (netdev_mc_count(netdev) > pdata->hw_feat.addn_mac)
- am_mode = 1;
- if ((netdev_uc_count(netdev) + netdev_mc_count(netdev)) >
- pdata->hw_feat.addn_mac)
- pr_mode = 1;
-
hw_if->set_promiscuous_mode(pdata, pr_mode);
hw_if->set_all_multicast_mode(pdata, am_mode);
- if (!pr_mode)
- hw_if->set_addn_mac_addrs(pdata, am_mode);
+
+ hw_if->add_mac_addresses(pdata);
DBGPR("<--xgbe_set_rx_mode\n");
}
@@ -999,6 +1007,38 @@
return s;
}
+static int xgbe_vlan_rx_add_vid(struct net_device *netdev, __be16 proto,
+ u16 vid)
+{
+ struct xgbe_prv_data *pdata = netdev_priv(netdev);
+ struct xgbe_hw_if *hw_if = &pdata->hw_if;
+
+ DBGPR("-->%s\n", __func__);
+
+ set_bit(vid, pdata->active_vlans);
+ hw_if->update_vlan_hash_table(pdata);
+
+ DBGPR("<--%s\n", __func__);
+
+ return 0;
+}
+
+static int xgbe_vlan_rx_kill_vid(struct net_device *netdev, __be16 proto,
+ u16 vid)
+{
+ struct xgbe_prv_data *pdata = netdev_priv(netdev);
+ struct xgbe_hw_if *hw_if = &pdata->hw_if;
+
+ DBGPR("-->%s\n", __func__);
+
+ clear_bit(vid, pdata->active_vlans);
+ hw_if->update_vlan_hash_table(pdata);
+
+ DBGPR("<--%s\n", __func__);
+
+ return 0;
+}
+
#ifdef CONFIG_NET_POLL_CONTROLLER
static void xgbe_poll_controller(struct net_device *netdev)
{
@@ -1021,26 +1061,26 @@
{
struct xgbe_prv_data *pdata = netdev_priv(netdev);
struct xgbe_hw_if *hw_if = &pdata->hw_if;
- unsigned int rxcsum_enabled, rxvlan_enabled;
+ unsigned int rxcsum, rxvlan, rxvlan_filter;
- rxcsum_enabled = !!(pdata->netdev_features & NETIF_F_RXCSUM);
- rxvlan_enabled = !!(pdata->netdev_features & NETIF_F_HW_VLAN_CTAG_RX);
+ rxcsum = pdata->netdev_features & NETIF_F_RXCSUM;
+ rxvlan = pdata->netdev_features & NETIF_F_HW_VLAN_CTAG_RX;
+ rxvlan_filter = pdata->netdev_features & NETIF_F_HW_VLAN_CTAG_FILTER;
- if ((features & NETIF_F_RXCSUM) && !rxcsum_enabled) {
+ if ((features & NETIF_F_RXCSUM) && !rxcsum)
hw_if->enable_rx_csum(pdata);
- netdev_alert(netdev, "state change - rxcsum enabled\n");
- } else if (!(features & NETIF_F_RXCSUM) && rxcsum_enabled) {
+ else if (!(features & NETIF_F_RXCSUM) && rxcsum)
hw_if->disable_rx_csum(pdata);
- netdev_alert(netdev, "state change - rxcsum disabled\n");
- }
- if ((features & NETIF_F_HW_VLAN_CTAG_RX) && !rxvlan_enabled) {
+ if ((features & NETIF_F_HW_VLAN_CTAG_RX) && !rxvlan)
hw_if->enable_rx_vlan_stripping(pdata);
- netdev_alert(netdev, "state change - rxvlan enabled\n");
- } else if (!(features & NETIF_F_HW_VLAN_CTAG_RX) && rxvlan_enabled) {
+ else if (!(features & NETIF_F_HW_VLAN_CTAG_RX) && rxvlan)
hw_if->disable_rx_vlan_stripping(pdata);
- netdev_alert(netdev, "state change - rxvlan disabled\n");
- }
+
+ if ((features & NETIF_F_HW_VLAN_CTAG_FILTER) && !rxvlan_filter)
+ hw_if->enable_rx_vlan_filtering(pdata);
+ else if (!(features & NETIF_F_HW_VLAN_CTAG_FILTER) && rxvlan_filter)
+ hw_if->disable_rx_vlan_filtering(pdata);
pdata->netdev_features = features;
@@ -1058,6 +1098,8 @@
.ndo_validate_addr = eth_validate_addr,
.ndo_change_mtu = xgbe_change_mtu,
.ndo_get_stats64 = xgbe_get_stats64,
+ .ndo_vlan_rx_add_vid = xgbe_vlan_rx_add_vid,
+ .ndo_vlan_rx_kill_vid = xgbe_vlan_rx_kill_vid,
#ifdef CONFIG_NET_POLL_CONTROLLER
.ndo_poll_controller = xgbe_poll_controller,
#endif
@@ -1089,8 +1131,9 @@
spin_lock_irqsave(&ring->lock, flags);
- while ((processed < TX_DESC_MAX_PROC) && (ring->dirty < ring->cur)) {
- rdata = GET_DESC_DATA(ring, ring->dirty);
+ while ((processed < XGBE_TX_DESC_MAX_PROC) &&
+ (ring->dirty < ring->cur)) {
+ rdata = XGBE_GET_DESC_DATA(ring, ring->dirty);
rdesc = rdata->rdesc;
if (!hw_if->tx_complete(rdesc))
@@ -1109,7 +1152,7 @@
}
if ((ring->tx.queue_stopped == 1) &&
- (xgbe_tx_avail_desc(ring) > TX_DESC_MIN_FREE)) {
+ (xgbe_tx_avail_desc(ring) > XGBE_TX_DESC_MIN_FREE)) {
ring->tx.queue_stopped = 0;
netif_wake_subqueue(netdev, channel->queue_index);
}
@@ -1152,7 +1195,7 @@
cur_len = 0;
read_again:
- rdata = GET_DESC_DATA(ring, ring->cur);
+ rdata = XGBE_GET_DESC_DATA(ring, ring->cur);
if (hw_if->dev_read(channel))
break;
@@ -1244,7 +1287,7 @@
/* Update the Rx Tail Pointer Register with address of
* the last cleaned entry */
- rdata = GET_DESC_DATA(ring, ring->rx.realloc_index - 1);
+ rdata = XGBE_GET_DESC_DATA(ring, ring->rx.realloc_index - 1);
XGMAC_DMA_IOWRITE(channel, DMA_CH_RDTR_LO,
lower_32_bits(rdata->rdesc_dma));
}
@@ -1296,7 +1339,7 @@
struct xgbe_ring_desc *rdesc;
while (count--) {
- rdata = GET_DESC_DATA(ring, idx);
+ rdata = XGBE_GET_DESC_DATA(ring, idx);
rdesc = rdata->rdesc;
DBGPR("TX_NORMAL_DESC[%d %s] = %08x:%08x:%08x:%08x\n", idx,
(flag == 1) ? "QUEUED FOR TX" : "TX BY DEVICE",
diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-main.c b/drivers/net/ethernet/amd/xgbe/xgbe-main.c
index c83584a..b411ac5 100644
--- a/drivers/net/ethernet/amd/xgbe/xgbe-main.c
+++ b/drivers/net/ethernet/amd/xgbe/xgbe-main.c
@@ -247,16 +247,16 @@
mutex_init(&pdata->xpcs_mutex);
/* Set and validate the number of descriptors for a ring */
- BUILD_BUG_ON_NOT_POWER_OF_2(TX_DESC_CNT);
- pdata->tx_desc_count = TX_DESC_CNT;
+ BUILD_BUG_ON_NOT_POWER_OF_2(XGBE_TX_DESC_CNT);
+ pdata->tx_desc_count = XGBE_TX_DESC_CNT;
if (pdata->tx_desc_count & (pdata->tx_desc_count - 1)) {
dev_err(dev, "tx descriptor count (%d) is not valid\n",
pdata->tx_desc_count);
ret = -EINVAL;
goto err_io;
}
- BUILD_BUG_ON_NOT_POWER_OF_2(RX_DESC_CNT);
- pdata->rx_desc_count = RX_DESC_CNT;
+ BUILD_BUG_ON_NOT_POWER_OF_2(XGBE_RX_DESC_CNT);
+ pdata->rx_desc_count = XGBE_RX_DESC_CNT;
if (pdata->rx_desc_count & (pdata->rx_desc_count - 1)) {
dev_err(dev, "rx descriptor count (%d) is not valid\n",
pdata->rx_desc_count);
@@ -385,7 +385,8 @@
NETIF_F_TSO6 |
NETIF_F_GRO |
NETIF_F_HW_VLAN_CTAG_RX |
- NETIF_F_HW_VLAN_CTAG_TX;
+ NETIF_F_HW_VLAN_CTAG_TX |
+ NETIF_F_HW_VLAN_CTAG_FILTER;
netdev->vlan_features |= NETIF_F_SG |
NETIF_F_IP_CSUM |
@@ -396,6 +397,8 @@
netdev->features |= netdev->hw_features;
pdata->netdev_features = netdev->features;
+ netdev->priv_flags |= IFF_UNICAST_FLT;
+
xgbe_init_rx_coalesce(pdata);
xgbe_init_tx_coalesce(pdata);
diff --git a/drivers/net/ethernet/amd/xgbe/xgbe.h b/drivers/net/ethernet/amd/xgbe/xgbe.h
index ab06271..a2d5f5f 100644
--- a/drivers/net/ethernet/amd/xgbe/xgbe.h
+++ b/drivers/net/ethernet/amd/xgbe/xgbe.h
@@ -121,6 +121,8 @@
#include <linux/netdevice.h>
#include <linux/workqueue.h>
#include <linux/phy.h>
+#include <linux/if_vlan.h>
+#include <linux/bitops.h>
#define XGBE_DRV_NAME "amd-xgbe"
@@ -128,22 +130,25 @@
#define XGBE_DRV_DESC "AMD 10 Gigabit Ethernet Driver"
/* Descriptor related defines */
-#define TX_DESC_CNT 512
-#define TX_DESC_MIN_FREE (TX_DESC_CNT >> 3)
-#define TX_DESC_MAX_PROC (TX_DESC_CNT >> 1)
-#define RX_DESC_CNT 512
+#define XGBE_TX_DESC_CNT 512
+#define XGBE_TX_DESC_MIN_FREE (XGBE_TX_DESC_CNT >> 3)
+#define XGBE_TX_DESC_MAX_PROC (XGBE_TX_DESC_CNT >> 1)
+#define XGBE_RX_DESC_CNT 512
-#define TX_MAX_BUF_SIZE (0x3fff & ~(64 - 1))
+#define XGBE_TX_MAX_BUF_SIZE (0x3fff & ~(64 - 1))
-#define RX_MIN_BUF_SIZE (ETH_FRAME_LEN + ETH_FCS_LEN + VLAN_HLEN)
-#define RX_BUF_ALIGN 64
+#define XGBE_RX_MIN_BUF_SIZE (ETH_FRAME_LEN + ETH_FCS_LEN + VLAN_HLEN)
+#define XGBE_RX_BUF_ALIGN 64
#define XGBE_MAX_DMA_CHANNELS 16
-#define DMA_ARDOMAIN_SETTING 0x2
-#define DMA_ARCACHE_SETTING 0xb
-#define DMA_AWDOMAIN_SETTING 0x2
-#define DMA_AWCACHE_SETTING 0x7
-#define DMA_INTERRUPT_MASK 0x31c7
+
+/* DMA cache settings - Outer sharable, write-back, write-allocate */
+#define XGBE_DMA_ARDOMAIN 0x2
+#define XGBE_DMA_ARCACHE 0xb
+#define XGBE_DMA_AWDOMAIN 0x2
+#define XGBE_DMA_AWCACHE 0x7
+
+#define XGBE_DMA_INTERRUPT_MASK 0x31c7
#define XGMAC_MIN_PACKET 60
#define XGMAC_STD_PACKET_MTU 1500
@@ -151,10 +156,6 @@
#define XGMAC_JUMBO_PACKET_MTU 9000
#define XGMAC_MAX_JUMBO_PACKET 9018
-#define MAX_MULTICAST_LIST 14
-#define TX_FLAGS_IP_PKT 0x00000001
-#define TX_FLAGS_TCP_PKT 0x00000002
-
/* MDIO bus phy name */
#define XGBE_PHY_NAME "amd_xgbe_phy"
#define XGBE_PRTAD 0
@@ -163,18 +164,18 @@
#define XGMAC_DRIVER_CONTEXT 1
#define XGMAC_IOCTL_CONTEXT 2
-#define FIFO_SIZE_B(x) (x)
-#define FIFO_SIZE_KB(x) (x * 1024)
+#define XGBE_FIFO_SIZE_B(x) (x)
+#define XGBE_FIFO_SIZE_KB(x) (x * 1024)
#define XGBE_TC_CNT 2
/* Helper macro for descriptor handling
- * Always use GET_DESC_DATA to access the descriptor data
+ * Always use XGBE_GET_DESC_DATA to access the descriptor data
* since the index is free-running and needs to be and-ed
* with the descriptor count value of the ring to index to
* the proper descriptor data.
*/
-#define GET_DESC_DATA(_ring, _idx) \
+#define XGBE_GET_DESC_DATA(_ring, _idx) \
((_ring)->rdata + \
((_idx) & ((_ring)->rdesc_count - 1)))
@@ -190,6 +191,8 @@
/* Flow control queue count */
#define XGMAC_MAX_FLOW_CONTROL_QUEUES 8
+/* Maximum MAC address hash table size (256 bits = 8 bytes) */
+#define XGBE_MAC_HASH_TABLE_SIZE 8
struct xgbe_prv_data;
@@ -219,7 +222,7 @@
/* Structure used to hold information related to the descriptor
* and the packet associated with the descriptor (always use
- * use the GET_DESC_DATA macro to access this data from the ring)
+ * use the XGBE_GET_DESC_DATA macro to access this data from the ring)
*/
struct xgbe_ring_data {
struct xgbe_ring_desc *rdesc; /* Virtual address of descriptor */
@@ -250,7 +253,7 @@
unsigned int rdesc_count;
/* Array of descriptor data corresponding the descriptor memory
- * (always use the GET_DESC_DATA macro to access this data)
+ * (always use the XGBE_GET_DESC_DATA macro to access this data)
*/
struct xgbe_ring_data *rdata;
@@ -386,7 +389,7 @@
int (*set_promiscuous_mode)(struct xgbe_prv_data *, unsigned int);
int (*set_all_multicast_mode)(struct xgbe_prv_data *, unsigned int);
- int (*set_addn_mac_addrs)(struct xgbe_prv_data *, unsigned int);
+ int (*add_mac_addresses)(struct xgbe_prv_data *);
int (*set_mac_address)(struct xgbe_prv_data *, u8 *addr);
int (*enable_rx_csum)(struct xgbe_prv_data *);
@@ -394,6 +397,9 @@
int (*enable_rx_vlan_stripping)(struct xgbe_prv_data *);
int (*disable_rx_vlan_stripping)(struct xgbe_prv_data *);
+ int (*enable_rx_vlan_filtering)(struct xgbe_prv_data *);
+ int (*disable_rx_vlan_filtering)(struct xgbe_prv_data *);
+ int (*update_vlan_hash_table)(struct xgbe_prv_data *);
int (*read_mmd_regs)(struct xgbe_prv_data *, int, int);
void (*write_mmd_regs)(struct xgbe_prv_data *, int, int, int);
@@ -589,6 +595,9 @@
struct napi_struct napi;
struct xgbe_mmc_stats mmc_stats;
+ /* Filtering support */
+ unsigned long active_vlans[BITS_TO_LONGS(VLAN_N_VID)];
+
/* System clock value used for Rx watchdog */
struct clk *sysclock;
diff --git a/net/tipc/link.c b/net/tipc/link.c
index ad2c57f..a081e7d 100644
--- a/net/tipc/link.c
+++ b/net/tipc/link.c
@@ -82,9 +82,6 @@
static int tipc_link_tunnel_rcv(struct tipc_node *n_ptr,
struct sk_buff **buf);
static void link_set_supervision_props(struct tipc_link *l_ptr, u32 tolerance);
-static int tipc_link_iovec_long_xmit(struct tipc_port *sender,
- struct iovec const *msg_sect,
- unsigned int len, u32 destnode);
static void link_state_event(struct tipc_link *l_ptr, u32 event);
static void link_reset_statistics(struct tipc_link *l_ptr);
static void link_print(struct tipc_link *l_ptr, const char *str);
@@ -335,13 +332,15 @@
static int link_schedule_port(struct tipc_link *l_ptr, u32 origport, u32 sz)
{
struct tipc_port *p_ptr;
+ struct tipc_sock *tsk;
spin_lock_bh(&tipc_port_list_lock);
p_ptr = tipc_port_lock(origport);
if (p_ptr) {
if (!list_empty(&p_ptr->wait_list))
goto exit;
- p_ptr->congested = 1;
+ tsk = tipc_port_to_sock(p_ptr);
+ tsk->link_cong = 1;
p_ptr->waiting_pkts = 1 + ((sz - 1) / l_ptr->max_pkt);
list_add_tail(&p_ptr->wait_list, &l_ptr->waiting_ports);
l_ptr->stats.link_congs++;
@@ -355,6 +354,7 @@
void tipc_link_wakeup_ports(struct tipc_link *l_ptr, int all)
{
struct tipc_port *p_ptr;
+ struct tipc_sock *tsk;
struct tipc_port *temp_p_ptr;
int win = l_ptr->queue_limit[0] - l_ptr->out_queue_size;
@@ -370,10 +370,11 @@
wait_list) {
if (win <= 0)
break;
+ tsk = tipc_port_to_sock(p_ptr);
list_del_init(&p_ptr->wait_list);
spin_lock_bh(p_ptr->lock);
- p_ptr->congested = 0;
- tipc_port_wakeup(p_ptr);
+ tsk->link_cong = 0;
+ tipc_sock_wakeup(tsk);
win -= p_ptr->waiting_pkts;
spin_unlock_bh(p_ptr->lock);
}
@@ -850,6 +851,144 @@
return res;
}
+/* tipc_link_cong: determine return value and how to treat the
+ * sent buffer during link congestion.
+ * - For plain, errorless user data messages we keep the buffer and
+ * return -ELINKONG.
+ * - For all other messages we discard the buffer and return -EHOSTUNREACH
+ * - For TIPC internal messages we also reset the link
+ */
+static int tipc_link_cong(struct tipc_link *link, struct sk_buff *buf)
+{
+ struct tipc_msg *msg = buf_msg(buf);
+ uint psz = msg_size(msg);
+ uint imp = tipc_msg_tot_importance(msg);
+ u32 oport = msg_tot_origport(msg);
+
+ if (likely(imp <= TIPC_CRITICAL_IMPORTANCE)) {
+ if (!msg_errcode(msg) && !msg_reroute_cnt(msg)) {
+ link_schedule_port(link, oport, psz);
+ return -ELINKCONG;
+ }
+ } else {
+ pr_warn("%s<%s>, send queue full", link_rst_msg, link->name);
+ tipc_link_reset(link);
+ }
+ kfree_skb_list(buf);
+ return -EHOSTUNREACH;
+}
+
+/**
+ * __tipc_link_xmit2(): same as tipc_link_xmit2, but destlink is known & locked
+ * @link: link to use
+ * @buf: chain of buffers containing message
+ * Consumes the buffer chain, except when returning -ELINKCONG
+ * Returns 0 if success, otherwise errno: -ELINKCONG, -EMSGSIZE (plain socket
+ * user data messages) or -EHOSTUNREACH (all other messages/senders)
+ * Only the socket functions tipc_send_stream() and tipc_send_packet() need
+ * to act on the return value, since they may need to do more send attempts.
+ */
+int __tipc_link_xmit2(struct tipc_link *link, struct sk_buff *buf)
+{
+ struct tipc_msg *msg = buf_msg(buf);
+ uint psz = msg_size(msg);
+ uint qsz = link->out_queue_size;
+ uint sndlim = link->queue_limit[0];
+ uint imp = tipc_msg_tot_importance(msg);
+ uint mtu = link->max_pkt;
+ uint ack = mod(link->next_in_no - 1);
+ uint seqno = link->next_out_no;
+ uint bc_last_in = link->owner->bclink.last_in;
+ struct tipc_media_addr *addr = &link->media_addr;
+ struct sk_buff *next = buf->next;
+
+ /* Match queue limits against msg importance: */
+ if (unlikely(qsz >= link->queue_limit[imp]))
+ return tipc_link_cong(link, buf);
+
+ /* Has valid packet limit been used ? */
+ if (unlikely(psz > mtu)) {
+ kfree_skb_list(buf);
+ return -EMSGSIZE;
+ }
+
+ /* Prepare each packet for sending, and add to outqueue: */
+ while (buf) {
+ next = buf->next;
+ msg = buf_msg(buf);
+ msg_set_word(msg, 2, ((ack << 16) | mod(seqno)));
+ msg_set_bcast_ack(msg, bc_last_in);
+
+ if (!link->first_out) {
+ link->first_out = buf;
+ } else if (qsz < sndlim) {
+ link->last_out->next = buf;
+ } else if (tipc_msg_bundle(link->last_out, buf, mtu)) {
+ link->stats.sent_bundled++;
+ buf = next;
+ next = buf->next;
+ continue;
+ } else if (tipc_msg_make_bundle(&buf, mtu, link->addr)) {
+ link->stats.sent_bundled++;
+ link->stats.sent_bundles++;
+ link->last_out->next = buf;
+ if (!link->next_out)
+ link->next_out = buf;
+ } else {
+ link->last_out->next = buf;
+ if (!link->next_out)
+ link->next_out = buf;
+ }
+
+ /* Send packet if possible: */
+ if (likely(++qsz <= sndlim)) {
+ tipc_bearer_send(link->bearer_id, buf, addr);
+ link->next_out = next;
+ link->unacked_window = 0;
+ }
+ seqno++;
+ link->last_out = buf;
+ buf = next;
+ }
+ link->next_out_no = seqno;
+ link->out_queue_size = qsz;
+ return 0;
+}
+
+/**
+ * tipc_link_xmit2() is the general link level function for message sending
+ * @buf: chain of buffers containing message
+ * @dsz: amount of user data to be sent
+ * @dnode: address of destination node
+ * @selector: a number used for deterministic link selection
+ * Consumes the buffer chain, except when returning -ELINKCONG
+ * Returns 0 if success, otherwise errno: -ELINKCONG,-EHOSTUNREACH,-EMSGSIZE
+ */
+int tipc_link_xmit2(struct sk_buff *buf, u32 dnode, u32 selector)
+{
+ struct tipc_link *link = NULL;
+ struct tipc_node *node;
+ int rc = -EHOSTUNREACH;
+
+ node = tipc_node_find(dnode);
+ if (node) {
+ tipc_node_lock(node);
+ link = node->active_links[selector & 1];
+ if (link)
+ rc = __tipc_link_xmit2(link, buf);
+ tipc_node_unlock(node);
+ }
+
+ if (link)
+ return rc;
+
+ if (likely(in_own_node(dnode)))
+ return tipc_sk_rcv(buf);
+
+ kfree_skb_list(buf);
+ return rc;
+}
+
/*
* tipc_link_sync_xmit - synchronize broadcast link endpoints.
*
@@ -933,252 +1072,6 @@
}
/*
- * tipc_link_xmit_fast: Entry for data messages where the
- * destination link is known and the header is complete,
- * inclusive total message length. Very time critical.
- * Link is locked. Returns user data length.
- */
-static int tipc_link_xmit_fast(struct tipc_link *l_ptr, struct sk_buff *buf,
- u32 *used_max_pkt)
-{
- struct tipc_msg *msg = buf_msg(buf);
- int res = msg_data_sz(msg);
-
- if (likely(!link_congested(l_ptr))) {
- if (likely(msg_size(msg) <= l_ptr->max_pkt)) {
- link_add_to_outqueue(l_ptr, buf, msg);
- tipc_bearer_send(l_ptr->bearer_id, buf,
- &l_ptr->media_addr);
- l_ptr->unacked_window = 0;
- return res;
- }
- else
- *used_max_pkt = l_ptr->max_pkt;
- }
- return __tipc_link_xmit(l_ptr, buf); /* All other cases */
-}
-
-/*
- * tipc_link_iovec_xmit_fast: Entry for messages where the
- * destination processor is known and the header is complete,
- * except for total message length.
- * Returns user data length or errno.
- */
-int tipc_link_iovec_xmit_fast(struct tipc_port *sender,
- struct iovec const *msg_sect,
- unsigned int len, u32 destaddr)
-{
- struct tipc_msg *hdr = &sender->phdr;
- struct tipc_link *l_ptr;
- struct sk_buff *buf;
- struct tipc_node *node;
- int res;
- u32 selector = msg_origport(hdr) & 1;
-
-again:
- /*
- * Try building message using port's max_pkt hint.
- * (Must not hold any locks while building message.)
- */
- res = tipc_msg_build(hdr, msg_sect, len, sender->max_pkt, &buf);
- /* Exit if build request was invalid */
- if (unlikely(res < 0))
- return res;
-
- node = tipc_node_find(destaddr);
- if (likely(node)) {
- tipc_node_lock(node);
- l_ptr = node->active_links[selector];
- if (likely(l_ptr)) {
- if (likely(buf)) {
- res = tipc_link_xmit_fast(l_ptr, buf,
- &sender->max_pkt);
-exit:
- tipc_node_unlock(node);
- return res;
- }
-
- /* Exit if link (or bearer) is congested */
- if (link_congested(l_ptr)) {
- res = link_schedule_port(l_ptr,
- sender->ref, res);
- goto exit;
- }
-
- /*
- * Message size exceeds max_pkt hint; update hint,
- * then re-try fast path or fragment the message
- */
- sender->max_pkt = l_ptr->max_pkt;
- tipc_node_unlock(node);
-
-
- if ((msg_hdr_sz(hdr) + res) <= sender->max_pkt)
- goto again;
-
- return tipc_link_iovec_long_xmit(sender, msg_sect,
- len, destaddr);
- }
- tipc_node_unlock(node);
- }
-
- /* Couldn't find a link to the destination node */
- kfree_skb(buf);
- tipc_port_iovec_reject(sender, hdr, msg_sect, len, TIPC_ERR_NO_NODE);
- return -ENETUNREACH;
-}
-
-/*
- * tipc_link_iovec_long_xmit(): Entry for long messages where the
- * destination node is known and the header is complete,
- * inclusive total message length.
- * Link and bearer congestion status have been checked to be ok,
- * and are ignored if they change.
- *
- * Note that fragments do not use the full link MTU so that they won't have
- * to undergo refragmentation if link changeover causes them to be sent
- * over another link with an additional tunnel header added as prefix.
- * (Refragmentation will still occur if the other link has a smaller MTU.)
- *
- * Returns user data length or errno.
- */
-static int tipc_link_iovec_long_xmit(struct tipc_port *sender,
- struct iovec const *msg_sect,
- unsigned int len, u32 destaddr)
-{
- struct tipc_link *l_ptr;
- struct tipc_node *node;
- struct tipc_msg *hdr = &sender->phdr;
- u32 dsz = len;
- u32 max_pkt, fragm_sz, rest;
- struct tipc_msg fragm_hdr;
- struct sk_buff *buf, *buf_chain, *prev;
- u32 fragm_crs, fragm_rest, hsz, sect_rest;
- const unchar __user *sect_crs;
- int curr_sect;
- u32 fragm_no;
- int res = 0;
-
-again:
- fragm_no = 1;
- max_pkt = sender->max_pkt - INT_H_SIZE;
- /* leave room for tunnel header in case of link changeover */
- fragm_sz = max_pkt - INT_H_SIZE;
- /* leave room for fragmentation header in each fragment */
- rest = dsz;
- fragm_crs = 0;
- fragm_rest = 0;
- sect_rest = 0;
- sect_crs = NULL;
- curr_sect = -1;
-
- /* Prepare reusable fragment header */
- tipc_msg_init(&fragm_hdr, MSG_FRAGMENTER, FIRST_FRAGMENT,
- INT_H_SIZE, msg_destnode(hdr));
- msg_set_size(&fragm_hdr, max_pkt);
- msg_set_fragm_no(&fragm_hdr, 1);
-
- /* Prepare header of first fragment */
- buf_chain = buf = tipc_buf_acquire(max_pkt);
- if (!buf)
- return -ENOMEM;
- buf->next = NULL;
- skb_copy_to_linear_data(buf, &fragm_hdr, INT_H_SIZE);
- hsz = msg_hdr_sz(hdr);
- skb_copy_to_linear_data_offset(buf, INT_H_SIZE, hdr, hsz);
-
- /* Chop up message */
- fragm_crs = INT_H_SIZE + hsz;
- fragm_rest = fragm_sz - hsz;
-
- do { /* For all sections */
- u32 sz;
-
- if (!sect_rest) {
- sect_rest = msg_sect[++curr_sect].iov_len;
- sect_crs = msg_sect[curr_sect].iov_base;
- }
-
- if (sect_rest < fragm_rest)
- sz = sect_rest;
- else
- sz = fragm_rest;
-
- if (copy_from_user(buf->data + fragm_crs, sect_crs, sz)) {
- res = -EFAULT;
-error:
- kfree_skb_list(buf_chain);
- return res;
- }
- sect_crs += sz;
- sect_rest -= sz;
- fragm_crs += sz;
- fragm_rest -= sz;
- rest -= sz;
-
- if (!fragm_rest && rest) {
-
- /* Initiate new fragment: */
- if (rest <= fragm_sz) {
- fragm_sz = rest;
- msg_set_type(&fragm_hdr, LAST_FRAGMENT);
- } else {
- msg_set_type(&fragm_hdr, FRAGMENT);
- }
- msg_set_size(&fragm_hdr, fragm_sz + INT_H_SIZE);
- msg_set_fragm_no(&fragm_hdr, ++fragm_no);
- prev = buf;
- buf = tipc_buf_acquire(fragm_sz + INT_H_SIZE);
- if (!buf) {
- res = -ENOMEM;
- goto error;
- }
-
- buf->next = NULL;
- prev->next = buf;
- skb_copy_to_linear_data(buf, &fragm_hdr, INT_H_SIZE);
- fragm_crs = INT_H_SIZE;
- fragm_rest = fragm_sz;
- }
- } while (rest > 0);
-
- /*
- * Now we have a buffer chain. Select a link and check
- * that packet size is still OK
- */
- node = tipc_node_find(destaddr);
- if (likely(node)) {
- tipc_node_lock(node);
- l_ptr = node->active_links[sender->ref & 1];
- if (!l_ptr) {
- tipc_node_unlock(node);
- goto reject;
- }
- if (l_ptr->max_pkt < max_pkt) {
- sender->max_pkt = l_ptr->max_pkt;
- tipc_node_unlock(node);
- kfree_skb_list(buf_chain);
- goto again;
- }
- } else {
-reject:
- kfree_skb_list(buf_chain);
- tipc_port_iovec_reject(sender, hdr, msg_sect, len,
- TIPC_ERR_NO_NODE);
- return -ENETUNREACH;
- }
-
- /* Append chain of fragments to send queue & send them */
- l_ptr->long_msg_seq_no++;
- link_add_chain_to_outqueue(l_ptr, buf_chain, l_ptr->long_msg_seq_no);
- l_ptr->stats.sent_fragments += fragm_no;
- l_ptr->stats.sent_fragmented++;
- tipc_link_push_queue(l_ptr);
- tipc_node_unlock(node);
- return dsz;
-}
-
-/*
* tipc_link_push_packet: Push one unsent packet to the media
*/
static u32 tipc_link_push_packet(struct tipc_link *l_ptr)
@@ -1238,7 +1131,7 @@
tipc_bearer_send(l_ptr->bearer_id, buf,
&l_ptr->media_addr);
if (msg_user(msg) == MSG_BUNDLER)
- msg_set_type(msg, CLOSED_MSG);
+ msg_set_type(msg, BUNDLE_CLOSED);
l_ptr->next_out = buf->next;
return 0;
}
@@ -1590,6 +1483,7 @@
case TIPC_MEDIUM_IMPORTANCE:
case TIPC_HIGH_IMPORTANCE:
case TIPC_CRITICAL_IMPORTANCE:
+ case CONN_MANAGER:
tipc_node_unlock(n_ptr);
tipc_sk_rcv(buf);
continue;
@@ -1604,10 +1498,6 @@
tipc_node_unlock(n_ptr);
tipc_named_rcv(buf);
continue;
- case CONN_MANAGER:
- tipc_node_unlock(n_ptr);
- tipc_port_proto_rcv(buf);
- continue;
case BCAST_PROTOCOL:
tipc_link_sync_rcv(n_ptr, buf);
break;
@@ -2217,6 +2107,7 @@
u32 msgcount = msg_msgcnt(buf_msg(buf));
u32 pos = INT_H_SIZE;
struct sk_buff *obuf;
+ struct tipc_msg *omsg;
while (msgcount--) {
obuf = buf_extract(buf, pos);
@@ -2224,8 +2115,16 @@
pr_warn("Link unable to unbundle message(s)\n");
break;
}
- pos += align(msg_size(buf_msg(obuf)));
- tipc_net_route_msg(obuf);
+ omsg = buf_msg(obuf);
+ pos += align(msg_size(omsg));
+ if (msg_isdata(omsg) || (msg_user(omsg) == CONN_MANAGER)) {
+ tipc_sk_rcv(obuf);
+ } else if (msg_user(omsg) == NAME_DISTRIBUTOR) {
+ tipc_named_rcv(obuf);
+ } else {
+ pr_warn("Illegal bundled msg: %u\n", msg_user(omsg));
+ kfree_skb(obuf);
+ }
}
kfree_skb(buf);
}
diff --git a/net/tipc/link.h b/net/tipc/link.h
index 200d518..227ff81 100644
--- a/net/tipc/link.h
+++ b/net/tipc/link.h
@@ -227,8 +227,10 @@
void tipc_link_reset(struct tipc_link *l_ptr);
void tipc_link_reset_list(unsigned int bearer_id);
int tipc_link_xmit(struct sk_buff *buf, u32 dest, u32 selector);
+int tipc_link_xmit2(struct sk_buff *buf, u32 dest, u32 selector);
void tipc_link_names_xmit(struct list_head *message_list, u32 dest);
int __tipc_link_xmit(struct tipc_link *l_ptr, struct sk_buff *buf);
+int __tipc_link_xmit2(struct tipc_link *link, struct sk_buff *buf);
int tipc_link_send_buf(struct tipc_link *l_ptr, struct sk_buff *buf);
u32 tipc_link_get_max_pkt(u32 dest, u32 selector);
int tipc_link_iovec_xmit_fast(struct tipc_port *sender,
diff --git a/net/tipc/msg.c b/net/tipc/msg.c
index 8be6e94..6ec9584 100644
--- a/net/tipc/msg.c
+++ b/net/tipc/msg.c
@@ -36,21 +36,16 @@
#include "core.h"
#include "msg.h"
+#include "addr.h"
+#include "name_table.h"
-u32 tipc_msg_tot_importance(struct tipc_msg *m)
+#define MAX_FORWARD_SIZE 1024
+
+static unsigned int align(unsigned int i)
{
- if (likely(msg_isdata(m))) {
- if (likely(msg_orignode(m) == tipc_own_addr))
- return msg_importance(m);
- return msg_importance(m) + 4;
- }
- if ((msg_user(m) == MSG_FRAGMENTER) &&
- (msg_type(m) == FIRST_FRAGMENT))
- return msg_importance(msg_get_wrapped(m));
- return msg_importance(m);
+ return (i + 3) & ~3u;
}
-
void tipc_msg_init(struct tipc_msg *m, u32 user, u32 type, u32 hsize,
u32 destnode)
{
@@ -152,3 +147,268 @@
kfree_skb(*buf);
return 0;
}
+
+
+/**
+ * tipc_msg_build2 - create buffer chain containing specified header and data
+ * @mhdr: Message header, to be prepended to data
+ * @iov: User data
+ * @offset: Posision in iov to start copying from
+ * @dsz: Total length of user data
+ * @pktmax: Max packet size that can be used
+ * @chain: Buffer or chain of buffers to be returned to caller
+ * Returns message data size or errno: -ENOMEM, -EFAULT
+ */
+int tipc_msg_build2(struct tipc_msg *mhdr, struct iovec const *iov,
+ int offset, int dsz, int pktmax , struct sk_buff **chain)
+{
+ int mhsz = msg_hdr_sz(mhdr);
+ int msz = mhsz + dsz;
+ int pktno = 1;
+ int pktsz;
+ int pktrem = pktmax;
+ int drem = dsz;
+ struct tipc_msg pkthdr;
+ struct sk_buff *buf, *prev;
+ char *pktpos;
+ int rc;
+
+ msg_set_size(mhdr, msz);
+
+ /* No fragmentation needed? */
+ if (likely(msz <= pktmax)) {
+ buf = tipc_buf_acquire(msz);
+ *chain = buf;
+ if (unlikely(!buf))
+ return -ENOMEM;
+ skb_copy_to_linear_data(buf, mhdr, mhsz);
+ pktpos = buf->data + mhsz;
+ if (!dsz || !memcpy_fromiovecend(pktpos, iov, offset, dsz))
+ return dsz;
+ rc = -EFAULT;
+ goto error;
+ }
+
+ /* Prepare reusable fragment header */
+ tipc_msg_init(&pkthdr, MSG_FRAGMENTER, FIRST_FRAGMENT,
+ INT_H_SIZE, msg_destnode(mhdr));
+ msg_set_size(&pkthdr, pktmax);
+ msg_set_fragm_no(&pkthdr, pktno);
+
+ /* Prepare first fragment */
+ *chain = buf = tipc_buf_acquire(pktmax);
+ if (!buf)
+ return -ENOMEM;
+ pktpos = buf->data;
+ skb_copy_to_linear_data(buf, &pkthdr, INT_H_SIZE);
+ pktpos += INT_H_SIZE;
+ pktrem -= INT_H_SIZE;
+ skb_copy_to_linear_data_offset(buf, INT_H_SIZE, mhdr, mhsz);
+ pktpos += mhsz;
+ pktrem -= mhsz;
+
+ do {
+ if (drem < pktrem)
+ pktrem = drem;
+
+ if (memcpy_fromiovecend(pktpos, iov, offset, pktrem)) {
+ rc = -EFAULT;
+ goto error;
+ }
+ drem -= pktrem;
+ offset += pktrem;
+
+ if (!drem)
+ break;
+
+ /* Prepare new fragment: */
+ if (drem < (pktmax - INT_H_SIZE))
+ pktsz = drem + INT_H_SIZE;
+ else
+ pktsz = pktmax;
+ prev = buf;
+ buf = tipc_buf_acquire(pktsz);
+ if (!buf) {
+ rc = -ENOMEM;
+ goto error;
+ }
+ prev->next = buf;
+ msg_set_type(&pkthdr, FRAGMENT);
+ msg_set_size(&pkthdr, pktsz);
+ msg_set_fragm_no(&pkthdr, ++pktno);
+ skb_copy_to_linear_data(buf, &pkthdr, INT_H_SIZE);
+ pktpos = buf->data + INT_H_SIZE;
+ pktrem = pktsz - INT_H_SIZE;
+
+ } while (1);
+
+ msg_set_type(buf_msg(buf), LAST_FRAGMENT);
+ return dsz;
+error:
+ kfree_skb_list(*chain);
+ *chain = NULL;
+ return rc;
+}
+
+/**
+ * tipc_msg_bundle(): Append contents of a buffer to tail of an existing one
+ * @bbuf: the existing buffer ("bundle")
+ * @buf: buffer to be appended
+ * @mtu: max allowable size for the bundle buffer
+ * Consumes buffer if successful
+ * Returns true if bundling could be performed, otherwise false
+ */
+bool tipc_msg_bundle(struct sk_buff *bbuf, struct sk_buff *buf, u32 mtu)
+{
+ struct tipc_msg *bmsg = buf_msg(bbuf);
+ struct tipc_msg *msg = buf_msg(buf);
+ unsigned int bsz = msg_size(bmsg);
+ unsigned int msz = msg_size(msg);
+ u32 start = align(bsz);
+ u32 max = mtu - INT_H_SIZE;
+ u32 pad = start - bsz;
+
+ if (likely(msg_user(msg) == MSG_FRAGMENTER))
+ return false;
+ if (unlikely(msg_user(msg) == CHANGEOVER_PROTOCOL))
+ return false;
+ if (unlikely(msg_user(msg) == BCAST_PROTOCOL))
+ return false;
+ if (likely(msg_user(bmsg) != MSG_BUNDLER))
+ return false;
+ if (likely(msg_type(bmsg) != BUNDLE_OPEN))
+ return false;
+ if (unlikely(skb_tailroom(bbuf) < (pad + msz)))
+ return false;
+ if (unlikely(max < (start + msz)))
+ return false;
+
+ skb_put(bbuf, pad + msz);
+ skb_copy_to_linear_data_offset(bbuf, start, buf->data, msz);
+ msg_set_size(bmsg, start + msz);
+ msg_set_msgcnt(bmsg, msg_msgcnt(bmsg) + 1);
+ bbuf->next = buf->next;
+ kfree_skb(buf);
+ return true;
+}
+
+/**
+ * tipc_msg_make_bundle(): Create bundle buf and append message to its tail
+ * @buf: buffer to be appended and replaced
+ * @mtu: max allowable size for the bundle buffer, inclusive header
+ * @dnode: destination node for message. (Not always present in header)
+ * Replaces buffer if successful
+ * Returns true if sucess, otherwise false
+ */
+bool tipc_msg_make_bundle(struct sk_buff **buf, u32 mtu, u32 dnode)
+{
+ struct sk_buff *bbuf;
+ struct tipc_msg *bmsg;
+ struct tipc_msg *msg = buf_msg(*buf);
+ u32 msz = msg_size(msg);
+ u32 max = mtu - INT_H_SIZE;
+
+ if (msg_user(msg) == MSG_FRAGMENTER)
+ return false;
+ if (msg_user(msg) == CHANGEOVER_PROTOCOL)
+ return false;
+ if (msg_user(msg) == BCAST_PROTOCOL)
+ return false;
+ if (msz > (max / 2))
+ return false;
+
+ bbuf = tipc_buf_acquire(max);
+ if (!bbuf)
+ return false;
+
+ skb_trim(bbuf, INT_H_SIZE);
+ bmsg = buf_msg(bbuf);
+ tipc_msg_init(bmsg, MSG_BUNDLER, BUNDLE_OPEN, INT_H_SIZE, dnode);
+ msg_set_seqno(bmsg, msg_seqno(msg));
+ msg_set_ack(bmsg, msg_ack(msg));
+ msg_set_bcast_ack(bmsg, msg_bcast_ack(msg));
+ bbuf->next = (*buf)->next;
+ tipc_msg_bundle(bbuf, *buf, mtu);
+ *buf = bbuf;
+ return true;
+}
+
+/**
+ * tipc_msg_reverse(): swap source and destination addresses and add error code
+ * @buf: buffer containing message to be reversed
+ * @dnode: return value: node where to send message after reversal
+ * @err: error code to be set in message
+ * Consumes buffer if failure
+ * Returns true if success, otherwise false
+ */
+bool tipc_msg_reverse(struct sk_buff *buf, u32 *dnode, int err)
+{
+ struct tipc_msg *msg = buf_msg(buf);
+ uint imp = msg_importance(msg);
+ struct tipc_msg ohdr;
+ uint rdsz = min_t(uint, msg_data_sz(msg), MAX_FORWARD_SIZE);
+
+ if (skb_linearize(buf))
+ goto exit;
+ if (msg_dest_droppable(msg))
+ goto exit;
+ if (msg_errcode(msg))
+ goto exit;
+
+ memcpy(&ohdr, msg, msg_hdr_sz(msg));
+ imp = min_t(uint, imp + 1, TIPC_CRITICAL_IMPORTANCE);
+ if (msg_isdata(msg))
+ msg_set_importance(msg, imp);
+ msg_set_errcode(msg, err);
+ msg_set_origport(msg, msg_destport(&ohdr));
+ msg_set_destport(msg, msg_origport(&ohdr));
+ msg_set_prevnode(msg, tipc_own_addr);
+ if (!msg_short(msg)) {
+ msg_set_orignode(msg, msg_destnode(&ohdr));
+ msg_set_destnode(msg, msg_orignode(&ohdr));
+ }
+ msg_set_size(msg, msg_hdr_sz(msg) + rdsz);
+ skb_trim(buf, msg_size(msg));
+ skb_orphan(buf);
+ *dnode = msg_orignode(&ohdr);
+ return true;
+exit:
+ kfree_skb(buf);
+ return false;
+}
+
+/**
+ * tipc_msg_eval: determine fate of message that found no destination
+ * @buf: the buffer containing the message.
+ * @dnode: return value: next-hop node, if message to be forwarded
+ * @err: error code to use, if message to be rejected
+ *
+ * Does not consume buffer
+ * Returns 0 (TIPC_OK) if message ok and we can try again, -TIPC error
+ * code if message to be rejected
+ */
+int tipc_msg_eval(struct sk_buff *buf, u32 *dnode)
+{
+ struct tipc_msg *msg = buf_msg(buf);
+ u32 dport;
+
+ if (msg_type(msg) != TIPC_NAMED_MSG)
+ return -TIPC_ERR_NO_PORT;
+ if (skb_linearize(buf))
+ return -TIPC_ERR_NO_NAME;
+ if (msg_data_sz(msg) > MAX_FORWARD_SIZE)
+ return -TIPC_ERR_NO_NAME;
+ if (msg_reroute_cnt(msg) > 0)
+ return -TIPC_ERR_NO_NAME;
+
+ *dnode = addr_domain(msg_lookup_scope(msg));
+ dport = tipc_nametbl_translate(msg_nametype(msg),
+ msg_nameinst(msg),
+ dnode);
+ if (!dport)
+ return -TIPC_ERR_NO_NAME;
+ msg_incr_reroute_cnt(msg);
+ msg_set_destnode(msg, *dnode);
+ msg_set_destport(msg, dport);
+ return TIPC_OK;
+}
diff --git a/net/tipc/msg.h b/net/tipc/msg.h
index 5035119..7d57434 100644
--- a/net/tipc/msg.h
+++ b/net/tipc/msg.h
@@ -463,6 +463,11 @@
#define FRAGMENT 1
#define LAST_FRAGMENT 2
+/* Bundling protocol message types
+ */
+#define BUNDLE_OPEN 0
+#define BUNDLE_CLOSED 1
+
/*
* Link management protocol message types
*/
@@ -706,12 +711,37 @@
msg_set_bits(m, 9, 0, 0xffff, n);
}
-u32 tipc_msg_tot_importance(struct tipc_msg *m);
+static inline u32 tipc_msg_tot_importance(struct tipc_msg *m)
+{
+ if ((msg_user(m) == MSG_FRAGMENTER) && (msg_type(m) == FIRST_FRAGMENT))
+ return msg_importance(msg_get_wrapped(m));
+ return msg_importance(m);
+}
+
+static inline u32 msg_tot_origport(struct tipc_msg *m)
+{
+ if ((msg_user(m) == MSG_FRAGMENTER) && (msg_type(m) == FIRST_FRAGMENT))
+ return msg_origport(msg_get_wrapped(m));
+ return msg_origport(m);
+}
+
+bool tipc_msg_reverse(struct sk_buff *buf, u32 *dnode, int err);
+
+int tipc_msg_eval(struct sk_buff *buf, u32 *dnode);
+
void tipc_msg_init(struct tipc_msg *m, u32 user, u32 type, u32 hsize,
u32 destnode);
+
int tipc_msg_build(struct tipc_msg *hdr, struct iovec const *msg_sect,
unsigned int len, int max_size, struct sk_buff **buf);
int tipc_buf_append(struct sk_buff **headbuf, struct sk_buff **buf);
+bool tipc_msg_bundle(struct sk_buff *bbuf, struct sk_buff *buf, u32 mtu);
+
+bool tipc_msg_make_bundle(struct sk_buff **buf, u32 mtu, u32 dnode);
+
+int tipc_msg_build2(struct tipc_msg *mhdr, struct iovec const *iov,
+ int offset, int dsz, int mtu , struct sk_buff **chain);
+
#endif
diff --git a/net/tipc/net.c b/net/tipc/net.c
index f64375e..7fcc949 100644
--- a/net/tipc/net.c
+++ b/net/tipc/net.c
@@ -1,7 +1,7 @@
/*
* net/tipc/net.c: TIPC network routing code
*
- * Copyright (c) 1995-2006, Ericsson AB
+ * Copyright (c) 1995-2006, 2014, Ericsson AB
* Copyright (c) 2005, 2010-2011, Wind River Systems
* All rights reserved.
*
@@ -104,67 +104,6 @@
* - A local spin_lock protecting the queue of subscriber events.
*/
-static void net_route_named_msg(struct sk_buff *buf)
-{
- struct tipc_msg *msg = buf_msg(buf);
- u32 dnode;
- u32 dport;
-
- if (!msg_named(msg)) {
- kfree_skb(buf);
- return;
- }
-
- dnode = addr_domain(msg_lookup_scope(msg));
- dport = tipc_nametbl_translate(msg_nametype(msg), msg_nameinst(msg), &dnode);
- if (dport) {
- msg_set_destnode(msg, dnode);
- msg_set_destport(msg, dport);
- tipc_net_route_msg(buf);
- return;
- }
- tipc_reject_msg(buf, TIPC_ERR_NO_NAME);
-}
-
-void tipc_net_route_msg(struct sk_buff *buf)
-{
- struct tipc_msg *msg;
- u32 dnode;
-
- if (!buf)
- return;
- msg = buf_msg(buf);
-
- /* Handle message for this node */
- dnode = msg_short(msg) ? tipc_own_addr : msg_destnode(msg);
- if (tipc_in_scope(dnode, tipc_own_addr)) {
- if (msg_isdata(msg)) {
- if (msg_mcast(msg))
- tipc_port_mcast_rcv(buf, NULL);
- else if (msg_destport(msg))
- tipc_sk_rcv(buf);
- else
- net_route_named_msg(buf);
- return;
- }
- switch (msg_user(msg)) {
- case NAME_DISTRIBUTOR:
- tipc_named_rcv(buf);
- break;
- case CONN_MANAGER:
- tipc_port_proto_rcv(buf);
- break;
- default:
- kfree_skb(buf);
- }
- return;
- }
-
- /* Handle message for another node */
- skb_trim(buf, msg_size(msg));
- tipc_link_xmit(buf, dnode, msg_link_selector(msg));
-}
-
int tipc_net_start(u32 addr)
{
char addr_string[16];
diff --git a/net/tipc/net.h b/net/tipc/net.h
index c6c2b46..59ef338 100644
--- a/net/tipc/net.h
+++ b/net/tipc/net.h
@@ -37,8 +37,6 @@
#ifndef _TIPC_NET_H
#define _TIPC_NET_H
-void tipc_net_route_msg(struct sk_buff *buf);
-
int tipc_net_start(u32 addr);
void tipc_net_stop(void);
diff --git a/net/tipc/node.c b/net/tipc/node.c
index 5b44c30..d959343 100644
--- a/net/tipc/node.c
+++ b/net/tipc/node.c
@@ -1,7 +1,7 @@
/*
* net/tipc/node.c: TIPC node management routines
*
- * Copyright (c) 2000-2006, 2012 Ericsson AB
+ * Copyright (c) 2000-2006, 2012-2014, Ericsson AB
* Copyright (c) 2005-2006, 2010-2014, Wind River Systems
* All rights reserved.
*
@@ -155,21 +155,25 @@
if (!active[0]) {
active[0] = active[1] = l_ptr;
node_established_contact(n_ptr);
- return;
+ goto exit;
}
if (l_ptr->priority < active[0]->priority) {
pr_info("New link <%s> becomes standby\n", l_ptr->name);
- return;
+ goto exit;
}
tipc_link_dup_queue_xmit(active[0], l_ptr);
if (l_ptr->priority == active[0]->priority) {
active[0] = l_ptr;
- return;
+ goto exit;
}
pr_info("Old link <%s> becomes standby\n", active[0]->name);
if (active[1] != active[0])
pr_info("Old link <%s> becomes standby\n", active[1]->name);
active[0] = active[1] = l_ptr;
+exit:
+ /* Leave room for changeover header when returning 'mtu' to users: */
+ n_ptr->act_mtus[0] = active[0]->max_pkt - INT_H_SIZE;
+ n_ptr->act_mtus[1] = active[1]->max_pkt - INT_H_SIZE;
}
/**
@@ -229,6 +233,19 @@
tipc_link_failover_send_queue(l_ptr);
else
node_lost_contact(n_ptr);
+
+ /* Leave room for changeover header when returning 'mtu' to users: */
+ if (active[0]) {
+ n_ptr->act_mtus[0] = active[0]->max_pkt - INT_H_SIZE;
+ n_ptr->act_mtus[1] = active[1]->max_pkt - INT_H_SIZE;
+ return;
+ }
+
+ /* Loopback link went down? No fragmentation needed from now on. */
+ if (n_ptr->addr == tipc_own_addr) {
+ n_ptr->act_mtus[0] = MAX_MSG_SIZE;
+ n_ptr->act_mtus[1] = MAX_MSG_SIZE;
+ }
}
int tipc_node_active_links(struct tipc_node *n_ptr)
diff --git a/net/tipc/node.h b/net/tipc/node.h
index 9087063..b61716a 100644
--- a/net/tipc/node.h
+++ b/net/tipc/node.h
@@ -41,6 +41,7 @@
#include "addr.h"
#include "net.h"
#include "bearer.h"
+#include "msg.h"
/*
* Out-of-range value for node signature
@@ -105,6 +106,7 @@
spinlock_t lock;
struct hlist_node hash;
struct tipc_link *active_links[2];
+ u32 act_mtus[2];
struct tipc_link *links[MAX_BEARERS];
unsigned int action_flags;
struct tipc_node_bclink bclink;
@@ -143,4 +145,19 @@
TIPC_NOTIFY_NODE_DOWN | TIPC_WAIT_OWN_LINKS_DOWN));
}
+static inline uint tipc_node_get_mtu(u32 addr, u32 selector)
+{
+ struct tipc_node *node;
+ u32 mtu;
+
+ node = tipc_node_find(addr);
+
+ if (likely(node))
+ mtu = node->act_mtus[selector & 1];
+ else
+ mtu = MAX_MSG_SIZE;
+
+ return mtu;
+}
+
#endif
diff --git a/net/tipc/node_subscr.c b/net/tipc/node_subscr.c
index 7c59ab1..2d13eea 100644
--- a/net/tipc/node_subscr.c
+++ b/net/tipc/node_subscr.c
@@ -84,11 +84,13 @@
void tipc_nodesub_notify(struct list_head *nsub_list)
{
struct tipc_node_subscr *ns, *safe;
+ net_ev_handler handle_node_down;
list_for_each_entry_safe(ns, safe, nsub_list, nodesub_list) {
- if (ns->handle_node_down) {
- ns->handle_node_down(ns->usr_handle);
+ handle_node_down = ns->handle_node_down;
+ if (handle_node_down) {
ns->handle_node_down = NULL;
+ handle_node_down(ns->usr_handle);
}
}
}
diff --git a/net/tipc/port.c b/net/tipc/port.c
index 5fd7acc..0d09dcb 100644
--- a/net/tipc/port.c
+++ b/net/tipc/port.c
@@ -42,8 +42,6 @@
/* Connection management: */
#define PROBING_INTERVAL 3600000 /* [ms] => 1 h */
-#define CONFIRMED 0
-#define PROBING 1
#define MAX_REJECT_SIZE 1024
@@ -188,12 +186,6 @@
tipc_port_list_free(dp);
}
-
-void tipc_port_wakeup(struct tipc_port *port)
-{
- tipc_sock_wakeup(tipc_port_to_sock(port));
-}
-
/* tipc_port_init - intiate TIPC port and lock it
*
* Returns obtained reference if initialization is successful, zero otherwise
@@ -235,6 +227,8 @@
void tipc_port_destroy(struct tipc_port *p_ptr)
{
struct sk_buff *buf = NULL;
+ struct tipc_msg *msg = NULL;
+ u32 peer;
tipc_withdraw(p_ptr, 0, NULL);
@@ -246,14 +240,15 @@
if (p_ptr->connected) {
buf = port_build_peer_abort_msg(p_ptr, TIPC_ERR_NO_PORT);
tipc_nodesub_unsubscribe(&p_ptr->subscription);
+ msg = buf_msg(buf);
+ peer = msg_destnode(msg);
+ tipc_link_xmit2(buf, peer, msg_link_selector(msg));
}
-
spin_lock_bh(&tipc_port_list_lock);
list_del(&p_ptr->port_list);
list_del(&p_ptr->wait_list);
spin_unlock_bh(&tipc_port_list_lock);
k_term_timer(&p_ptr->timer);
- tipc_net_route_msg(buf);
}
/*
@@ -275,100 +270,16 @@
msg_set_destport(msg, tipc_port_peerport(p_ptr));
msg_set_origport(msg, p_ptr->ref);
msg_set_msgcnt(msg, ack);
+ buf->next = NULL;
}
return buf;
}
-int tipc_reject_msg(struct sk_buff *buf, u32 err)
-{
- struct tipc_msg *msg = buf_msg(buf);
- struct sk_buff *rbuf;
- struct tipc_msg *rmsg;
- int hdr_sz;
- u32 imp;
- u32 data_sz = msg_data_sz(msg);
- u32 src_node;
- u32 rmsg_sz;
-
- /* discard rejected message if it shouldn't be returned to sender */
- if (WARN(!msg_isdata(msg),
- "attempt to reject message with user=%u", msg_user(msg))) {
- dump_stack();
- goto exit;
- }
- if (msg_errcode(msg) || msg_dest_droppable(msg))
- goto exit;
-
- /*
- * construct returned message by copying rejected message header and
- * data (or subset), then updating header fields that need adjusting
- */
- hdr_sz = msg_hdr_sz(msg);
- rmsg_sz = hdr_sz + min_t(u32, data_sz, MAX_REJECT_SIZE);
-
- rbuf = tipc_buf_acquire(rmsg_sz);
- if (rbuf == NULL)
- goto exit;
-
- rmsg = buf_msg(rbuf);
- skb_copy_to_linear_data(rbuf, msg, rmsg_sz);
-
- if (msg_connected(rmsg)) {
- imp = msg_importance(rmsg);
- if (imp < TIPC_CRITICAL_IMPORTANCE)
- msg_set_importance(rmsg, ++imp);
- }
- msg_set_non_seq(rmsg, 0);
- msg_set_size(rmsg, rmsg_sz);
- msg_set_errcode(rmsg, err);
- msg_set_prevnode(rmsg, tipc_own_addr);
- msg_swap_words(rmsg, 4, 5);
- if (!msg_short(rmsg))
- msg_swap_words(rmsg, 6, 7);
-
- /* send self-abort message when rejecting on a connected port */
- if (msg_connected(msg)) {
- struct tipc_port *p_ptr = tipc_port_lock(msg_destport(msg));
-
- if (p_ptr) {
- struct sk_buff *abuf = NULL;
-
- if (p_ptr->connected)
- abuf = port_build_self_abort_msg(p_ptr, err);
- tipc_port_unlock(p_ptr);
- tipc_net_route_msg(abuf);
- }
- }
-
- /* send returned message & dispose of rejected message */
- src_node = msg_prevnode(msg);
- if (in_own_node(src_node))
- tipc_sk_rcv(rbuf);
- else
- tipc_link_xmit(rbuf, src_node, msg_link_selector(rmsg));
-exit:
- kfree_skb(buf);
- return data_sz;
-}
-
-int tipc_port_iovec_reject(struct tipc_port *p_ptr, struct tipc_msg *hdr,
- struct iovec const *msg_sect, unsigned int len,
- int err)
-{
- struct sk_buff *buf;
- int res;
-
- res = tipc_msg_build(hdr, msg_sect, len, MAX_MSG_SIZE, &buf);
- if (!buf)
- return res;
-
- return tipc_reject_msg(buf, err);
-}
-
static void port_timeout(unsigned long ref)
{
struct tipc_port *p_ptr = tipc_port_lock(ref);
struct sk_buff *buf = NULL;
+ struct tipc_msg *msg = NULL;
if (!p_ptr)
return;
@@ -379,15 +290,16 @@
}
/* Last probe answered ? */
- if (p_ptr->probing_state == PROBING) {
+ if (p_ptr->probing_state == TIPC_CONN_PROBING) {
buf = port_build_self_abort_msg(p_ptr, TIPC_ERR_NO_PORT);
} else {
buf = port_build_proto_msg(p_ptr, CONN_PROBE, 0);
- p_ptr->probing_state = PROBING;
+ p_ptr->probing_state = TIPC_CONN_PROBING;
k_start_timer(&p_ptr->timer, p_ptr->probing_interval);
}
tipc_port_unlock(p_ptr);
- tipc_net_route_msg(buf);
+ msg = buf_msg(buf);
+ tipc_link_xmit2(buf, msg_destnode(msg), msg_link_selector(msg));
}
@@ -395,12 +307,14 @@
{
struct tipc_port *p_ptr = tipc_port_lock(ref);
struct sk_buff *buf = NULL;
+ struct tipc_msg *msg = NULL;
if (!p_ptr)
return;
buf = port_build_self_abort_msg(p_ptr, TIPC_ERR_NO_NODE);
tipc_port_unlock(p_ptr);
- tipc_net_route_msg(buf);
+ msg = buf_msg(buf);
+ tipc_link_xmit2(buf, msg_destnode(msg), msg_link_selector(msg));
}
@@ -412,6 +326,7 @@
struct tipc_msg *msg = buf_msg(buf);
msg_swap_words(msg, 4, 5);
msg_swap_words(msg, 6, 7);
+ buf->next = NULL;
}
return buf;
}
@@ -436,60 +351,11 @@
if (imp < TIPC_CRITICAL_IMPORTANCE)
msg_set_importance(msg, ++imp);
msg_set_errcode(msg, err);
+ buf->next = NULL;
}
return buf;
}
-void tipc_port_proto_rcv(struct sk_buff *buf)
-{
- struct tipc_msg *msg = buf_msg(buf);
- struct tipc_port *p_ptr;
- struct sk_buff *r_buf = NULL;
- u32 destport = msg_destport(msg);
- int wakeable;
-
- /* Validate connection */
- p_ptr = tipc_port_lock(destport);
- if (!p_ptr || !p_ptr->connected || !tipc_port_peer_msg(p_ptr, msg)) {
- r_buf = tipc_buf_acquire(BASIC_H_SIZE);
- if (r_buf) {
- msg = buf_msg(r_buf);
- tipc_msg_init(msg, TIPC_HIGH_IMPORTANCE, TIPC_CONN_MSG,
- BASIC_H_SIZE, msg_orignode(msg));
- msg_set_errcode(msg, TIPC_ERR_NO_PORT);
- msg_set_origport(msg, destport);
- msg_set_destport(msg, msg_origport(msg));
- }
- if (p_ptr)
- tipc_port_unlock(p_ptr);
- goto exit;
- }
-
- /* Process protocol message sent by peer */
- switch (msg_type(msg)) {
- case CONN_ACK:
- wakeable = tipc_port_congested(p_ptr) && p_ptr->congested;
- p_ptr->acked += msg_msgcnt(msg);
- if (!tipc_port_congested(p_ptr)) {
- p_ptr->congested = 0;
- if (wakeable)
- tipc_port_wakeup(p_ptr);
- }
- break;
- case CONN_PROBE:
- r_buf = port_build_proto_msg(p_ptr, CONN_PROBE_REPLY, 0);
- break;
- default:
- /* CONN_PROBE_REPLY or unrecognized - no action required */
- break;
- }
- p_ptr->probing_state = CONFIRMED;
- tipc_port_unlock(p_ptr);
-exit:
- tipc_net_route_msg(r_buf);
- kfree_skb(buf);
-}
-
static int port_print(struct tipc_port *p_ptr, char *buf, int len, int full_id)
{
struct publication *publ;
@@ -581,16 +447,19 @@
{
struct tipc_port *p_ptr;
struct sk_buff *buf = NULL;
+ struct tipc_msg *msg;
p_ptr = tipc_port_lock(ref);
if (!p_ptr)
return;
- if (p_ptr->connected) {
- p_ptr->conn_unacked -= ack;
+ if (p_ptr->connected)
buf = port_build_proto_msg(p_ptr, CONN_ACK, ack);
- }
+
tipc_port_unlock(p_ptr);
- tipc_net_route_msg(buf);
+ if (!buf)
+ return;
+ msg = buf_msg(buf);
+ tipc_link_xmit2(buf, msg_destnode(msg), msg_link_selector(msg));
}
int tipc_publish(struct tipc_port *p_ptr, unsigned int scope,
@@ -689,7 +558,7 @@
msg_set_hdr_sz(msg, SHORT_H_SIZE);
p_ptr->probing_interval = PROBING_INTERVAL;
- p_ptr->probing_state = CONFIRMED;
+ p_ptr->probing_state = TIPC_CONN_OK;
p_ptr->connected = 1;
k_start_timer(&p_ptr->timer, p_ptr->probing_interval);
@@ -698,7 +567,7 @@
(net_ev_handler)port_handle_node_down);
res = 0;
exit:
- p_ptr->max_pkt = tipc_link_get_max_pkt(peer->node, ref);
+ p_ptr->max_pkt = tipc_node_get_mtu(peer->node, ref);
return res;
}
@@ -741,6 +610,7 @@
*/
int tipc_port_shutdown(u32 ref)
{
+ struct tipc_msg *msg;
struct tipc_port *p_ptr;
struct sk_buff *buf = NULL;
@@ -750,149 +620,7 @@
buf = port_build_peer_abort_msg(p_ptr, TIPC_CONN_SHUTDOWN);
tipc_port_unlock(p_ptr);
- tipc_net_route_msg(buf);
+ msg = buf_msg(buf);
+ tipc_link_xmit2(buf, msg_destnode(msg), msg_link_selector(msg));
return tipc_port_disconnect(ref);
}
-
-/*
- * tipc_port_iovec_rcv: Concatenate and deliver sectioned
- * message for this node.
- */
-static int tipc_port_iovec_rcv(struct tipc_port *sender,
- struct iovec const *msg_sect,
- unsigned int len)
-{
- struct sk_buff *buf;
- int res;
-
- res = tipc_msg_build(&sender->phdr, msg_sect, len, MAX_MSG_SIZE, &buf);
- if (likely(buf))
- tipc_sk_rcv(buf);
- return res;
-}
-
-/**
- * tipc_send - send message sections on connection
- */
-int tipc_send(struct tipc_port *p_ptr,
- struct iovec const *msg_sect,
- unsigned int len)
-{
- u32 destnode;
- int res;
-
- if (!p_ptr->connected)
- return -EINVAL;
-
- p_ptr->congested = 1;
- if (!tipc_port_congested(p_ptr)) {
- destnode = tipc_port_peernode(p_ptr);
- if (likely(!in_own_node(destnode)))
- res = tipc_link_iovec_xmit_fast(p_ptr, msg_sect, len,
- destnode);
- else
- res = tipc_port_iovec_rcv(p_ptr, msg_sect, len);
-
- if (likely(res != -ELINKCONG)) {
- p_ptr->congested = 0;
- if (res > 0)
- p_ptr->sent++;
- return res;
- }
- }
- if (tipc_port_unreliable(p_ptr)) {
- p_ptr->congested = 0;
- return len;
- }
- return -ELINKCONG;
-}
-
-/**
- * tipc_send2name - send message sections to port name
- */
-int tipc_send2name(struct tipc_port *p_ptr,
- struct tipc_name const *name,
- unsigned int domain,
- struct iovec const *msg_sect,
- unsigned int len)
-{
- struct tipc_msg *msg;
- u32 destnode = domain;
- u32 destport;
- int res;
-
- if (p_ptr->connected)
- return -EINVAL;
-
- msg = &p_ptr->phdr;
- msg_set_type(msg, TIPC_NAMED_MSG);
- msg_set_hdr_sz(msg, NAMED_H_SIZE);
- msg_set_nametype(msg, name->type);
- msg_set_nameinst(msg, name->instance);
- msg_set_lookup_scope(msg, tipc_addr_scope(domain));
- destport = tipc_nametbl_translate(name->type, name->instance, &destnode);
- msg_set_destnode(msg, destnode);
- msg_set_destport(msg, destport);
-
- if (likely(destport || destnode)) {
- if (likely(in_own_node(destnode)))
- res = tipc_port_iovec_rcv(p_ptr, msg_sect, len);
- else if (tipc_own_addr)
- res = tipc_link_iovec_xmit_fast(p_ptr, msg_sect, len,
- destnode);
- else
- res = tipc_port_iovec_reject(p_ptr, msg, msg_sect,
- len, TIPC_ERR_NO_NODE);
- if (likely(res != -ELINKCONG)) {
- if (res > 0)
- p_ptr->sent++;
- return res;
- }
- if (tipc_port_unreliable(p_ptr))
- return len;
-
- return -ELINKCONG;
- }
- return tipc_port_iovec_reject(p_ptr, msg, msg_sect, len,
- TIPC_ERR_NO_NAME);
-}
-
-/**
- * tipc_send2port - send message sections to port identity
- */
-int tipc_send2port(struct tipc_port *p_ptr,
- struct tipc_portid const *dest,
- struct iovec const *msg_sect,
- unsigned int len)
-{
- struct tipc_msg *msg;
- int res;
-
- if (p_ptr->connected)
- return -EINVAL;
-
- msg = &p_ptr->phdr;
- msg_set_type(msg, TIPC_DIRECT_MSG);
- msg_set_lookup_scope(msg, 0);
- msg_set_destnode(msg, dest->node);
- msg_set_destport(msg, dest->ref);
- msg_set_hdr_sz(msg, BASIC_H_SIZE);
-
- if (in_own_node(dest->node))
- res = tipc_port_iovec_rcv(p_ptr, msg_sect, len);
- else if (tipc_own_addr)
- res = tipc_link_iovec_xmit_fast(p_ptr, msg_sect, len,
- dest->node);
- else
- res = tipc_port_iovec_reject(p_ptr, msg, msg_sect, len,
- TIPC_ERR_NO_NODE);
- if (likely(res != -ELINKCONG)) {
- if (res > 0)
- p_ptr->sent++;
- return res;
- }
- if (tipc_port_unreliable(p_ptr))
- return len;
-
- return -ELINKCONG;
-}
diff --git a/net/tipc/port.h b/net/tipc/port.h
index cf4ca5b..0e47052 100644
--- a/net/tipc/port.h
+++ b/net/tipc/port.h
@@ -53,17 +53,13 @@
* @connected: non-zero if port is currently connected to a peer port
* @conn_type: TIPC type used when connection was established
* @conn_instance: TIPC instance used when connection was established
- * @conn_unacked: number of unacknowledged messages received from peer port
* @published: non-zero if port has one or more associated names
- * @congested: non-zero if cannot send because of link or port congestion
* @max_pkt: maximum packet size "hint" used when building messages sent by port
* @ref: unique reference to port in TIPC object registry
* @phdr: preformatted message header used when sending messages
* @port_list: adjacent ports in TIPC's global list of ports
* @wait_list: adjacent ports in list of ports waiting on link congestion
* @waiting_pkts:
- * @sent: # of non-empty messages sent by port
- * @acked: # of non-empty message acknowledgements from connected port's peer
* @publications: list of publications for port
* @pub_count: total # of publications port has made during its lifetime
* @probing_state:
@@ -76,17 +72,13 @@
int connected;
u32 conn_type;
u32 conn_instance;
- u32 conn_unacked;
int published;
- u32 congested;
u32 max_pkt;
u32 ref;
struct tipc_msg phdr;
struct list_head port_list;
struct list_head wait_list;
u32 waiting_pkts;
- u32 sent;
- u32 acked;
struct list_head publications;
u32 pub_count;
u32 probing_state;
@@ -104,8 +96,6 @@
u32 tipc_port_init(struct tipc_port *p_ptr,
const unsigned int importance);
-int tipc_reject_msg(struct sk_buff *buf, u32 err);
-
void tipc_acknowledge(u32 port_ref, u32 ack);
void tipc_port_destroy(struct tipc_port *p_ptr);
@@ -122,8 +112,6 @@
int tipc_port_shutdown(u32 ref);
-void tipc_port_wakeup(struct tipc_port *port);
-
/*
* The following routines require that the port be locked on entry
*/
@@ -136,34 +124,12 @@
* TIPC messaging routines
*/
-int tipc_send(struct tipc_port *port,
- struct iovec const *msg_sect,
- unsigned int len);
-
-int tipc_send2name(struct tipc_port *port,
- struct tipc_name const *name,
- u32 domain,
- struct iovec const *msg_sect,
- unsigned int len);
-
-int tipc_send2port(struct tipc_port *port,
- struct tipc_portid const *dest,
- struct iovec const *msg_sect,
- unsigned int len);
-
int tipc_port_mcast_xmit(struct tipc_port *port,
struct tipc_name_seq const *seq,
struct iovec const *msg,
unsigned int len);
-int tipc_port_iovec_reject(struct tipc_port *p_ptr,
- struct tipc_msg *hdr,
- struct iovec const *msg_sect,
- unsigned int len,
- int err);
-
struct sk_buff *tipc_port_get_ports(void);
-void tipc_port_proto_rcv(struct sk_buff *buf);
void tipc_port_mcast_rcv(struct sk_buff *buf, struct tipc_port_list *dp);
void tipc_port_reinit(void);
@@ -185,12 +151,6 @@
spin_unlock_bh(p_ptr->lock);
}
-static inline int tipc_port_congested(struct tipc_port *p_ptr)
-{
- return ((p_ptr->sent - p_ptr->acked) >= TIPC_FLOWCTRL_WIN);
-}
-
-
static inline u32 tipc_port_peernode(struct tipc_port *p_ptr)
{
return msg_destnode(&p_ptr->phdr);
diff --git a/net/tipc/socket.c b/net/tipc/socket.c
index ef04755..ede78b1 100644
--- a/net/tipc/socket.c
+++ b/net/tipc/socket.c
@@ -36,14 +36,17 @@
#include "core.h"
#include "port.h"
+#include "name_table.h"
#include "node.h"
-
+#include "link.h"
#include <linux/export.h>
+#include "link.h"
#define SS_LISTENING -1 /* socket is listening */
#define SS_READY -2 /* socket is connectionless */
#define CONN_TIMEOUT_DEFAULT 8000 /* default connect timeout = 8s */
+#define TIPC_FWD_MSG 1
static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *skb);
static void tipc_data_ready(struct sock *sk);
@@ -123,9 +126,12 @@
static void reject_rx_queue(struct sock *sk)
{
struct sk_buff *buf;
+ u32 dnode;
- while ((buf = __skb_dequeue(&sk->sk_receive_queue)))
- tipc_reject_msg(buf, TIPC_ERR_NO_PORT);
+ while ((buf = __skb_dequeue(&sk->sk_receive_queue))) {
+ if (tipc_msg_reverse(buf, &dnode, TIPC_ERR_NO_PORT))
+ tipc_link_xmit2(buf, dnode, 0);
+ }
}
/**
@@ -201,6 +207,7 @@
sk->sk_data_ready = tipc_data_ready;
sk->sk_write_space = tipc_write_space;
tsk->conn_timeout = CONN_TIMEOUT_DEFAULT;
+ tsk->sent_unacked = 0;
atomic_set(&tsk->dupl_rcvcnt, 0);
tipc_port_unlock(port);
@@ -303,6 +310,7 @@
struct tipc_sock *tsk;
struct tipc_port *port;
struct sk_buff *buf;
+ u32 dnode;
/*
* Exit if socket isn't fully initialized (occurs when a failed accept()
@@ -331,7 +339,8 @@
sock->state = SS_DISCONNECTING;
tipc_port_disconnect(port->ref);
}
- tipc_reject_msg(buf, TIPC_ERR_NO_PORT);
+ if (tipc_msg_reverse(buf, &dnode, TIPC_ERR_NO_PORT))
+ tipc_link_xmit2(buf, dnode, 0);
}
}
@@ -504,12 +513,12 @@
switch ((int)sock->state) {
case SS_UNCONNECTED:
- if (!tsk->port.congested)
+ if (!tsk->link_cong)
mask |= POLLOUT;
break;
case SS_READY:
case SS_CONNECTED:
- if (!tsk->port.congested)
+ if (!tsk->link_cong && !tipc_sk_conn_cong(tsk))
mask |= POLLOUT;
/* fall thru' */
case SS_CONNECTING:
@@ -526,6 +535,43 @@
}
/**
+ * tipc_sk_proto_rcv - receive a connection mng protocol message
+ * @tsk: receiving socket
+ * @dnode: node to send response message to, if any
+ * @buf: buffer containing protocol message
+ * Returns 0 (TIPC_OK) if message was consumed, 1 (TIPC_FWD_MSG) if
+ * (CONN_PROBE_REPLY) message should be forwarded.
+ */
+int tipc_sk_proto_rcv(struct tipc_sock *tsk, u32 *dnode, struct sk_buff *buf)
+{
+ struct tipc_msg *msg = buf_msg(buf);
+ struct tipc_port *port = &tsk->port;
+ int conn_cong;
+
+ /* Ignore if connection cannot be validated: */
+ if (!port->connected || !tipc_port_peer_msg(port, msg))
+ goto exit;
+
+ port->probing_state = TIPC_CONN_OK;
+
+ if (msg_type(msg) == CONN_ACK) {
+ conn_cong = tipc_sk_conn_cong(tsk);
+ tsk->sent_unacked -= msg_msgcnt(msg);
+ if (conn_cong)
+ tipc_sock_wakeup(tsk);
+ } else if (msg_type(msg) == CONN_PROBE) {
+ if (!tipc_msg_reverse(buf, dnode, TIPC_OK))
+ return TIPC_OK;
+ msg_set_type(msg, CONN_PROBE_REPLY);
+ return TIPC_FWD_MSG;
+ }
+ /* Do nothing if msg_type() == CONN_PROBE_REPLY */
+exit:
+ kfree_skb(buf);
+ return TIPC_OK;
+}
+
+/**
* dest_name_check - verify user is permitted to send to specified port name
* @dest: destination address
* @m: descriptor for message to be sent
@@ -539,6 +585,8 @@
{
struct tipc_cfg_msg_hdr hdr;
+ if (unlikely(dest->addrtype == TIPC_ADDR_ID))
+ return 0;
if (likely(dest->addr.name.name.type >= TIPC_RESERVED_TYPES))
return 0;
if (likely(dest->addr.name.name.type == TIPC_TOP_SRV))
@@ -575,19 +623,55 @@
return sock_intr_errno(*timeo_p);
prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
- done = sk_wait_event(sk, timeo_p, !tsk->port.congested);
+ done = sk_wait_event(sk, timeo_p, !tsk->link_cong);
finish_wait(sk_sleep(sk), &wait);
} while (!done);
return 0;
}
+/**
+ * tipc_sendmcast - send multicast message
+ * @sock: socket structure
+ * @seq: destination address
+ * @iov: message data to send
+ * @dsz: total length of message data
+ * @timeo: timeout to wait for wakeup
+ *
+ * Called from function tipc_sendmsg(), which has done all sanity checks
+ * Returns the number of bytes sent on success, or errno
+ */
+static int tipc_sendmcast(struct socket *sock, struct tipc_name_seq *seq,
+ struct iovec *iov, size_t dsz, long timeo)
+{
+ struct sock *sk = sock->sk;
+ struct tipc_sock *tsk = tipc_sk(sk);
+ int rc;
+
+ do {
+ if (sock->state != SS_READY) {
+ rc = -EOPNOTSUPP;
+ break;
+ }
+ rc = tipc_port_mcast_xmit(&tsk->port, seq, iov, dsz);
+ if (likely(rc >= 0)) {
+ if (sock->state != SS_READY)
+ sock->state = SS_CONNECTING;
+ break;
+ }
+ if (rc != -ELINKCONG)
+ break;
+ rc = tipc_wait_for_sndmsg(sock, &timeo);
+ } while (!rc);
+
+ return rc;
+}
/**
* tipc_sendmsg - send message in connectionless manner
* @iocb: if NULL, indicates that socket lock is already held
* @sock: socket structure
* @m: message to send
- * @total_len: length of message
+ * @dsz: amount of user data to be sent
*
* Message must have an destination specified explicitly.
* Used for SOCK_RDM and SOCK_DGRAM messages,
@@ -597,100 +681,122 @@
* Returns the number of bytes sent on success, or errno otherwise
*/
static int tipc_sendmsg(struct kiocb *iocb, struct socket *sock,
- struct msghdr *m, size_t total_len)
+ struct msghdr *m, size_t dsz)
{
+ DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
struct sock *sk = sock->sk;
struct tipc_sock *tsk = tipc_sk(sk);
struct tipc_port *port = &tsk->port;
- DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
- int needs_conn;
+ struct tipc_msg *mhdr = &port->phdr;
+ struct iovec *iov = m->msg_iov;
+ u32 dnode, dport;
+ struct sk_buff *buf;
+ struct tipc_name_seq *seq = &dest->addr.nameseq;
+ u32 mtu;
long timeo;
- int res = -EINVAL;
+ int rc = -EINVAL;
if (unlikely(!dest))
return -EDESTADDRREQ;
+
if (unlikely((m->msg_namelen < sizeof(*dest)) ||
(dest->family != AF_TIPC)))
return -EINVAL;
- if (total_len > TIPC_MAX_USER_MSG_SIZE)
+
+ if (dsz > TIPC_MAX_USER_MSG_SIZE)
return -EMSGSIZE;
if (iocb)
lock_sock(sk);
- needs_conn = (sock->state != SS_READY);
- if (unlikely(needs_conn)) {
+ if (unlikely(sock->state != SS_READY)) {
if (sock->state == SS_LISTENING) {
- res = -EPIPE;
+ rc = -EPIPE;
goto exit;
}
if (sock->state != SS_UNCONNECTED) {
- res = -EISCONN;
+ rc = -EISCONN;
goto exit;
}
if (tsk->port.published) {
- res = -EOPNOTSUPP;
+ rc = -EOPNOTSUPP;
goto exit;
}
if (dest->addrtype == TIPC_ADDR_NAME) {
tsk->port.conn_type = dest->addr.name.name.type;
tsk->port.conn_instance = dest->addr.name.name.instance;
}
-
- /* Abort any pending connection attempts (very unlikely) */
- reject_rx_queue(sk);
}
+ rc = dest_name_check(dest, m);
+ if (rc)
+ goto exit;
timeo = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT);
+
+ if (dest->addrtype == TIPC_ADDR_MCAST) {
+ rc = tipc_sendmcast(sock, seq, iov, dsz, timeo);
+ goto exit;
+ } else if (dest->addrtype == TIPC_ADDR_NAME) {
+ u32 type = dest->addr.name.name.type;
+ u32 inst = dest->addr.name.name.instance;
+ u32 domain = dest->addr.name.domain;
+
+ dnode = domain;
+ msg_set_type(mhdr, TIPC_NAMED_MSG);
+ msg_set_hdr_sz(mhdr, NAMED_H_SIZE);
+ msg_set_nametype(mhdr, type);
+ msg_set_nameinst(mhdr, inst);
+ msg_set_lookup_scope(mhdr, tipc_addr_scope(domain));
+ dport = tipc_nametbl_translate(type, inst, &dnode);
+ msg_set_destnode(mhdr, dnode);
+ msg_set_destport(mhdr, dport);
+ if (unlikely(!dport && !dnode)) {
+ rc = -EHOSTUNREACH;
+ goto exit;
+ }
+ } else if (dest->addrtype == TIPC_ADDR_ID) {
+ dnode = dest->addr.id.node;
+ msg_set_type(mhdr, TIPC_DIRECT_MSG);
+ msg_set_lookup_scope(mhdr, 0);
+ msg_set_destnode(mhdr, dnode);
+ msg_set_destport(mhdr, dest->addr.id.ref);
+ msg_set_hdr_sz(mhdr, BASIC_H_SIZE);
+ }
+
+new_mtu:
+ mtu = tipc_node_get_mtu(dnode, tsk->port.ref);
+ rc = tipc_msg_build2(mhdr, iov, 0, dsz, mtu, &buf);
+ if (rc < 0)
+ goto exit;
+
do {
- if (dest->addrtype == TIPC_ADDR_NAME) {
- res = dest_name_check(dest, m);
- if (res)
- break;
- res = tipc_send2name(port,
- &dest->addr.name.name,
- dest->addr.name.domain,
- m->msg_iov,
- total_len);
- } else if (dest->addrtype == TIPC_ADDR_ID) {
- res = tipc_send2port(port,
- &dest->addr.id,
- m->msg_iov,
- total_len);
- } else if (dest->addrtype == TIPC_ADDR_MCAST) {
- if (needs_conn) {
- res = -EOPNOTSUPP;
- break;
- }
- res = dest_name_check(dest, m);
- if (res)
- break;
- res = tipc_port_mcast_xmit(port,
- &dest->addr.nameseq,
- m->msg_iov,
- total_len);
- }
- if (likely(res != -ELINKCONG)) {
- if (needs_conn && (res >= 0))
+ rc = tipc_link_xmit2(buf, dnode, tsk->port.ref);
+ if (likely(rc >= 0)) {
+ if (sock->state != SS_READY)
sock->state = SS_CONNECTING;
+ rc = dsz;
break;
}
- res = tipc_wait_for_sndmsg(sock, &timeo);
- if (res)
+ if (rc == -EMSGSIZE)
+ goto new_mtu;
+
+ if (rc != -ELINKCONG)
break;
- } while (1);
+
+ rc = tipc_wait_for_sndmsg(sock, &timeo);
+ } while (!rc);
exit:
if (iocb)
release_sock(sk);
- return res;
+
+ return rc;
}
static int tipc_wait_for_sndpkt(struct socket *sock, long *timeo_p)
{
struct sock *sk = sock->sk;
struct tipc_sock *tsk = tipc_sk(sk);
- struct tipc_port *port = &tsk->port;
DEFINE_WAIT(wait);
int done;
@@ -709,71 +815,20 @@
prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
done = sk_wait_event(sk, timeo_p,
- (!port->congested || !port->connected));
+ (!tsk->link_cong &&
+ !tipc_sk_conn_cong(tsk)) ||
+ !tsk->port.connected);
finish_wait(sk_sleep(sk), &wait);
} while (!done);
return 0;
}
/**
- * tipc_send_packet - send a connection-oriented message
- * @iocb: if NULL, indicates that socket lock is already held
- * @sock: socket structure
- * @m: message to send
- * @total_len: length of message
- *
- * Used for SOCK_SEQPACKET messages and SOCK_STREAM data.
- *
- * Returns the number of bytes sent on success, or errno otherwise
- */
-static int tipc_send_packet(struct kiocb *iocb, struct socket *sock,
- struct msghdr *m, size_t total_len)
-{
- struct sock *sk = sock->sk;
- struct tipc_sock *tsk = tipc_sk(sk);
- DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
- int res = -EINVAL;
- long timeo;
-
- /* Handle implied connection establishment */
- if (unlikely(dest))
- return tipc_sendmsg(iocb, sock, m, total_len);
-
- if (total_len > TIPC_MAX_USER_MSG_SIZE)
- return -EMSGSIZE;
-
- if (iocb)
- lock_sock(sk);
-
- if (unlikely(sock->state != SS_CONNECTED)) {
- if (sock->state == SS_DISCONNECTING)
- res = -EPIPE;
- else
- res = -ENOTCONN;
- goto exit;
- }
-
- timeo = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT);
- do {
- res = tipc_send(&tsk->port, m->msg_iov, total_len);
- if (likely(res != -ELINKCONG))
- break;
- res = tipc_wait_for_sndpkt(sock, &timeo);
- if (res)
- break;
- } while (1);
-exit:
- if (iocb)
- release_sock(sk);
- return res;
-}
-
-/**
* tipc_send_stream - send stream-oriented data
* @iocb: (unused)
* @sock: socket structure
* @m: data to send
- * @total_len: total length of data to be sent
+ * @dsz: total length of data to be transmitted
*
* Used for SOCK_STREAM data.
*
@@ -781,89 +836,93 @@
* or errno if no data sent
*/
static int tipc_send_stream(struct kiocb *iocb, struct socket *sock,
- struct msghdr *m, size_t total_len)
+ struct msghdr *m, size_t dsz)
{
struct sock *sk = sock->sk;
struct tipc_sock *tsk = tipc_sk(sk);
- struct msghdr my_msg;
- struct iovec my_iov;
- struct iovec *curr_iov;
- int curr_iovlen;
- char __user *curr_start;
- u32 hdr_size;
- int curr_left;
- int bytes_to_send;
- int bytes_sent;
- int res;
+ struct tipc_port *port = &tsk->port;
+ struct tipc_msg *mhdr = &port->phdr;
+ struct sk_buff *buf;
+ DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
+ u32 ref = port->ref;
+ int rc = -EINVAL;
+ long timeo;
+ u32 dnode;
+ uint mtu, send, sent = 0;
- lock_sock(sk);
+ /* Handle implied connection establishment */
+ if (unlikely(dest)) {
+ rc = tipc_sendmsg(iocb, sock, m, dsz);
+ if (dsz && (dsz == rc))
+ tsk->sent_unacked = 1;
+ return rc;
+ }
+ if (dsz > (uint)INT_MAX)
+ return -EMSGSIZE;
- /* Handle special cases where there is no connection */
+ if (iocb)
+ lock_sock(sk);
+
if (unlikely(sock->state != SS_CONNECTED)) {
- if (sock->state == SS_UNCONNECTED)
- res = tipc_send_packet(NULL, sock, m, total_len);
+ if (sock->state == SS_DISCONNECTING)
+ rc = -EPIPE;
else
- res = sock->state == SS_DISCONNECTING ? -EPIPE : -ENOTCONN;
+ rc = -ENOTCONN;
goto exit;
}
- if (unlikely(m->msg_name)) {
- res = -EISCONN;
+ timeo = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT);
+ dnode = tipc_port_peernode(port);
+
+next:
+ mtu = port->max_pkt;
+ send = min_t(uint, dsz - sent, TIPC_MAX_USER_MSG_SIZE);
+ rc = tipc_msg_build2(mhdr, m->msg_iov, sent, send, mtu, &buf);
+ if (unlikely(rc < 0))
goto exit;
- }
-
- if (total_len > (unsigned int)INT_MAX) {
- res = -EMSGSIZE;
- goto exit;
- }
-
- /*
- * Send each iovec entry using one or more messages
- *
- * Note: This algorithm is good for the most likely case
- * (i.e. one large iovec entry), but could be improved to pass sets
- * of small iovec entries into send_packet().
- */
- curr_iov = m->msg_iov;
- curr_iovlen = m->msg_iovlen;
- my_msg.msg_iov = &my_iov;
- my_msg.msg_iovlen = 1;
- my_msg.msg_flags = m->msg_flags;
- my_msg.msg_name = NULL;
- bytes_sent = 0;
-
- hdr_size = msg_hdr_sz(&tsk->port.phdr);
-
- while (curr_iovlen--) {
- curr_start = curr_iov->iov_base;
- curr_left = curr_iov->iov_len;
-
- while (curr_left) {
- bytes_to_send = tsk->port.max_pkt - hdr_size;
- if (bytes_to_send > TIPC_MAX_USER_MSG_SIZE)
- bytes_to_send = TIPC_MAX_USER_MSG_SIZE;
- if (curr_left < bytes_to_send)
- bytes_to_send = curr_left;
- my_iov.iov_base = curr_start;
- my_iov.iov_len = bytes_to_send;
- res = tipc_send_packet(NULL, sock, &my_msg,
- bytes_to_send);
- if (res < 0) {
- if (bytes_sent)
- res = bytes_sent;
- goto exit;
+ do {
+ if (likely(!tipc_sk_conn_cong(tsk))) {
+ rc = tipc_link_xmit2(buf, dnode, ref);
+ if (likely(!rc)) {
+ tsk->sent_unacked++;
+ sent += send;
+ if (sent == dsz)
+ break;
+ goto next;
}
- curr_left -= bytes_to_send;
- curr_start += bytes_to_send;
- bytes_sent += bytes_to_send;
+ if (rc == -EMSGSIZE) {
+ port->max_pkt = tipc_node_get_mtu(dnode, ref);
+ goto next;
+ }
+ if (rc != -ELINKCONG)
+ break;
}
-
- curr_iov++;
- }
- res = bytes_sent;
+ rc = tipc_wait_for_sndpkt(sock, &timeo);
+ } while (!rc);
exit:
- release_sock(sk);
- return res;
+ if (iocb)
+ release_sock(sk);
+ return sent ? sent : rc;
+}
+
+/**
+ * tipc_send_packet - send a connection-oriented message
+ * @iocb: if NULL, indicates that socket lock is already held
+ * @sock: socket structure
+ * @m: message to send
+ * @dsz: length of data to be transmitted
+ *
+ * Used for SOCK_SEQPACKET messages.
+ *
+ * Returns the number of bytes sent on success, or errno otherwise
+ */
+static int tipc_send_packet(struct kiocb *iocb, struct socket *sock,
+ struct msghdr *m, size_t dsz)
+{
+ if (dsz > TIPC_MAX_USER_MSG_SIZE)
+ return -EMSGSIZE;
+
+ return tipc_send_stream(iocb, sock, m, dsz);
}
/**
@@ -1104,8 +1163,10 @@
/* Consume received message (optional) */
if (likely(!(flags & MSG_PEEK))) {
if ((sock->state != SS_READY) &&
- (++port->conn_unacked >= TIPC_CONNACK_INTV))
- tipc_acknowledge(port->ref, port->conn_unacked);
+ (++tsk->rcv_unacked >= TIPC_CONNACK_INTV)) {
+ tipc_acknowledge(port->ref, tsk->rcv_unacked);
+ tsk->rcv_unacked = 0;
+ }
advance_rx_queue(sk);
}
exit:
@@ -1213,8 +1274,10 @@
/* Consume received message (optional) */
if (likely(!(flags & MSG_PEEK))) {
- if (unlikely(++port->conn_unacked >= TIPC_CONNACK_INTV))
- tipc_acknowledge(port->ref, port->conn_unacked);
+ if (unlikely(++tsk->rcv_unacked >= TIPC_CONNACK_INTV)) {
+ tipc_acknowledge(port->ref, tsk->rcv_unacked);
+ tsk->rcv_unacked = 0;
+ }
advance_rx_queue(sk);
}
@@ -1269,17 +1332,16 @@
* @tsk: TIPC socket
* @msg: message
*
- * Returns TIPC error status code and socket error status code
- * once it encounters some errors
+ * Returns 0 (TIPC_OK) if everyting ok, -TIPC_ERR_NO_PORT otherwise
*/
-static u32 filter_connect(struct tipc_sock *tsk, struct sk_buff **buf)
+static int filter_connect(struct tipc_sock *tsk, struct sk_buff **buf)
{
struct sock *sk = &tsk->sk;
struct tipc_port *port = &tsk->port;
struct socket *sock = sk->sk_socket;
struct tipc_msg *msg = buf_msg(*buf);
- u32 retval = TIPC_ERR_NO_PORT;
+ int retval = -TIPC_ERR_NO_PORT;
int res;
if (msg_mcast(msg))
@@ -1382,32 +1444,37 @@
*
* Called with socket lock already taken; port lock may also be taken.
*
- * Returns TIPC error status code (TIPC_OK if message is not to be rejected)
+ * Returns 0 (TIPC_OK) if message was consumed, -TIPC error code if message
+ * to be rejected, 1 (TIPC_FWD_MSG) if (CONN_MANAGER) message to be forwarded
*/
-static u32 filter_rcv(struct sock *sk, struct sk_buff *buf)
+static int filter_rcv(struct sock *sk, struct sk_buff *buf)
{
struct socket *sock = sk->sk_socket;
struct tipc_sock *tsk = tipc_sk(sk);
struct tipc_msg *msg = buf_msg(buf);
unsigned int limit = rcvbuf_limit(sk, buf);
- u32 res = TIPC_OK;
+ u32 onode;
+ int rc = TIPC_OK;
+
+ if (unlikely(msg_user(msg) == CONN_MANAGER))
+ return tipc_sk_proto_rcv(tsk, &onode, buf);
/* Reject message if it is wrong sort of message for socket */
if (msg_type(msg) > TIPC_DIRECT_MSG)
- return TIPC_ERR_NO_PORT;
+ return -TIPC_ERR_NO_PORT;
if (sock->state == SS_READY) {
if (msg_connected(msg))
- return TIPC_ERR_NO_PORT;
+ return -TIPC_ERR_NO_PORT;
} else {
- res = filter_connect(tsk, &buf);
- if (res != TIPC_OK || buf == NULL)
- return res;
+ rc = filter_connect(tsk, &buf);
+ if (rc != TIPC_OK || buf == NULL)
+ return rc;
}
/* Reject message if there isn't room to queue it */
if (sk_rmem_alloc_get(sk) + buf->truesize >= limit)
- return TIPC_ERR_OVERLOAD;
+ return -TIPC_ERR_OVERLOAD;
/* Enqueue message */
TIPC_SKB_CB(buf)->handle = NULL;
@@ -1429,16 +1496,23 @@
*/
static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *buf)
{
- u32 res;
+ int rc;
+ u32 onode;
struct tipc_sock *tsk = tipc_sk(sk);
uint truesize = buf->truesize;
- res = filter_rcv(sk, buf);
- if (unlikely(res))
- tipc_reject_msg(buf, res);
+ rc = filter_rcv(sk, buf);
- if (atomic_read(&tsk->dupl_rcvcnt) < TIPC_CONN_OVERLOAD_LIMIT)
- atomic_add(truesize, &tsk->dupl_rcvcnt);
+ if (likely(!rc)) {
+ if (atomic_read(&tsk->dupl_rcvcnt) < TIPC_CONN_OVERLOAD_LIMIT)
+ atomic_add(truesize, &tsk->dupl_rcvcnt);
+ return 0;
+ }
+
+ if ((rc < 0) && !tipc_msg_reverse(buf, &onode, -rc))
+ return 0;
+
+ tipc_link_xmit2(buf, onode, 0);
return 0;
}
@@ -1455,19 +1529,14 @@
struct tipc_port *port;
struct sock *sk;
u32 dport = msg_destport(buf_msg(buf));
- int err = TIPC_OK;
+ int rc = TIPC_OK;
uint limit;
+ u32 dnode;
- /* Forward unresolved named message */
- if (unlikely(!dport)) {
- tipc_net_route_msg(buf);
- return 0;
- }
-
- /* Validate destination */
+ /* Validate destination and message */
port = tipc_port_lock(dport);
if (unlikely(!port)) {
- err = TIPC_ERR_NO_PORT;
+ rc = tipc_msg_eval(buf, &dnode);
goto exit;
}
@@ -1478,23 +1547,25 @@
bh_lock_sock(sk);
if (!sock_owned_by_user(sk)) {
- err = filter_rcv(sk, buf);
+ rc = filter_rcv(sk, buf);
} else {
if (sk->sk_backlog.len == 0)
atomic_set(&tsk->dupl_rcvcnt, 0);
limit = rcvbuf_limit(sk, buf) + atomic_read(&tsk->dupl_rcvcnt);
if (sk_add_backlog(sk, buf, limit))
- err = TIPC_ERR_OVERLOAD;
+ rc = -TIPC_ERR_OVERLOAD;
}
-
bh_unlock_sock(sk);
tipc_port_unlock(port);
- if (likely(!err))
+ if (likely(!rc))
return 0;
exit:
- tipc_reject_msg(buf, err);
- return -EHOSTUNREACH;
+ if ((rc < 0) && !tipc_msg_reverse(buf, &dnode, -rc))
+ return -EHOSTUNREACH;
+
+ tipc_link_xmit2(buf, dnode, 0);
+ return (rc < 0) ? -EHOSTUNREACH : 0;
}
static int tipc_wait_for_connect(struct socket *sock, long *timeo_p)
@@ -1758,6 +1829,7 @@
struct tipc_sock *tsk = tipc_sk(sk);
struct tipc_port *port = &tsk->port;
struct sk_buff *buf;
+ u32 peer;
int res;
if (how != SHUT_RDWR)
@@ -1778,7 +1850,8 @@
goto restart;
}
tipc_port_disconnect(port->ref);
- tipc_reject_msg(buf, TIPC_CONN_SHUTDOWN);
+ if (tipc_msg_reverse(buf, &peer, TIPC_CONN_SHUTDOWN))
+ tipc_link_xmit2(buf, peer, 0);
} else {
tipc_port_shutdown(port->ref);
}
diff --git a/net/tipc/socket.h b/net/tipc/socket.h
index 3afcd2a..2cdede9 100644
--- a/net/tipc/socket.h
+++ b/net/tipc/socket.h
@@ -38,6 +38,9 @@
#include "port.h"
#include <net/sock.h>
+#define TIPC_CONN_OK 0
+#define TIPC_CONN_PROBING 1
+
/**
* struct tipc_sock - TIPC socket structure
* @sk: socket - interacts with 'port' and with user via the socket API
@@ -45,6 +48,9 @@
* @peer_name: the peer of the connection, if any
* @conn_timeout: the time we can wait for an unresponded setup request
* @dupl_rcvcnt: number of bytes counted twice, in both backlog and rcv queue
+ * @link_cong: non-zero if owner must sleep because of link congestion
+ * @sent_unacked: # messages sent by socket, and not yet acked by peer
+ * @rcv_unacked: # messages read by user, but not yet acked back to peer
*/
struct tipc_sock {
@@ -52,6 +58,9 @@
struct tipc_port port;
unsigned int conn_timeout;
atomic_t dupl_rcvcnt;
+ int link_cong;
+ uint sent_unacked;
+ uint rcv_unacked;
};
static inline struct tipc_sock *tipc_sk(const struct sock *sk)
@@ -69,6 +78,11 @@
tsk->sk.sk_write_space(&tsk->sk);
}
+static inline int tipc_sk_conn_cong(struct tipc_sock *tsk)
+{
+ return tsk->sent_unacked >= TIPC_FLOWCTRL_WIN;
+}
+
int tipc_sk_rcv(struct sk_buff *buf);
#endif