| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <03f0279a-8ed8-4068-80a7-3cdf7740c6b8@intel.com>
Date: Wed, 5 Jun 2024 10:54:38 +0200
From: Wojciech Drewek <wojciech.drewek@...el.com>
To: Mengyuan Lou <mengyuanlou@...-swift.com>, <netdev@...r.kernel.org>
CC: <jiawenwu@...stnetic.com>, <duanqiangwen@...-swift.com>
Subject: Re: [PATCH net-next v4 3/6] net: libwx: Redesign flow when sriov is
enabled
On 04.06.2024 17:57, Mengyuan Lou wrote:
> Reallocate queue and int resources when sriov is enabled.
> Redefine macro VMDQ to make it work in VT mode.
>
> Signed-off-by: Mengyuan Lou <mengyuanlou@...-swift.com>
> ---
Reviewed-by: Wojciech Drewek <wojciech.drewek@...el.com>
> drivers/net/ethernet/wangxun/libwx/wx_hw.c | 293 ++++++++++++++++++-
> drivers/net/ethernet/wangxun/libwx/wx_lib.c | 129 +++++++-
> drivers/net/ethernet/wangxun/libwx/wx_type.h | 37 ++-
> 3 files changed, 442 insertions(+), 17 deletions(-)
>
> diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.c b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
> index 7c4b6881a93f..8affcb9f7dbb 100644
> --- a/drivers/net/ethernet/wangxun/libwx/wx_hw.c
> +++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
> @@ -10,6 +10,7 @@
>
> #include "wx_type.h"
> #include "wx_lib.h"
> +#include "wx_sriov.h"
> #include "wx_hw.h"
>
> static int wx_phy_read_reg_mdi(struct mii_bus *bus, int phy_addr, int devnum, int regnum)
> @@ -804,11 +805,28 @@ static void wx_sync_mac_table(struct wx *wx)
> }
> }
>
> +static void wx_full_sync_mac_table(struct wx *wx)
> +{
> + int i;
> +
> + for (i = 0; i < wx->mac.num_rar_entries; i++) {
> + if (wx->mac_table[i].state & WX_MAC_STATE_IN_USE) {
> + wx_set_rar(wx, i,
> + wx->mac_table[i].addr,
> + wx->mac_table[i].pools,
> + WX_PSR_MAC_SWC_AD_H_AV);
> + } else {
> + wx_clear_rar(wx, i);
> + }
> + wx->mac_table[i].state &= ~(WX_MAC_STATE_MODIFIED);
> + }
> +}
> +
> /* this function destroys the first RAR entry */
> void wx_mac_set_default_filter(struct wx *wx, u8 *addr)
> {
> memcpy(&wx->mac_table[0].addr, addr, ETH_ALEN);
> - wx->mac_table[0].pools = 1ULL;
> + wx->mac_table[0].pools = BIT(VMDQ_P(0));
> wx->mac_table[0].state = (WX_MAC_STATE_DEFAULT | WX_MAC_STATE_IN_USE);
> wx_set_rar(wx, 0, wx->mac_table[0].addr,
> wx->mac_table[0].pools,
> @@ -1046,6 +1064,35 @@ static void wx_update_mc_addr_list(struct wx *wx, struct net_device *netdev)
> wx_dbg(wx, "Update mc addr list Complete\n");
> }
>
> +static void wx_restore_vf_multicasts(struct wx *wx)
> +{
> + u32 i, j, vector_bit, vector_reg;
> + struct vf_data_storage *vfinfo;
> +
> + for (i = 0; i < wx->num_vfs; i++) {
> + u32 vmolr = rd32(wx, WX_PSR_VM_L2CTL(i));
> +
> + vfinfo = &wx->vfinfo[i];
> + for (j = 0; j < vfinfo->num_vf_mc_hashes; j++) {
> + wx->addr_ctrl.mta_in_use++;
> + vector_reg = (vfinfo->vf_mc_hashes[j] >> 5) & GENMASK(6, 0);
> + vector_bit = vfinfo->vf_mc_hashes[j] & GENMASK(4, 0);
> + wr32m(wx, WX_PSR_MC_TBL(vector_reg),
> + BIT(vector_bit), BIT(vector_bit));
> + /* errata 5: maintain a copy of the reg table conf */
> + wx->mac.mta_shadow[vector_reg] |= BIT(vector_bit);
> + }
> + if (vfinfo->num_vf_mc_hashes)
> + vmolr |= WX_PSR_VM_L2CTL_ROMPE;
> + else
> + vmolr &= ~WX_PSR_VM_L2CTL_ROMPE;
> + wr32(wx, WX_PSR_VM_L2CTL(i), vmolr);
> + }
> +
> + /* Restore any VF macvlans */
> + wx_full_sync_mac_table(wx);
> +}
> +
> /**
> * wx_write_mc_addr_list - write multicast addresses to MTA
> * @netdev: network interface device structure
> @@ -1063,6 +1110,9 @@ static int wx_write_mc_addr_list(struct net_device *netdev)
>
> wx_update_mc_addr_list(wx, netdev);
>
> + if (test_bit(WX_FLAG_SRIOV_ENABLED, wx->flags))
> + wx_restore_vf_multicasts(wx);
> +
> return netdev_mc_count(netdev);
> }
>
> @@ -1083,7 +1133,7 @@ int wx_set_mac(struct net_device *netdev, void *p)
> if (retval)
> return retval;
>
> - wx_del_mac_filter(wx, wx->mac.addr, 0);
> + wx_del_mac_filter(wx, wx->mac.addr, VMDQ_P(0));
> eth_hw_addr_set(netdev, addr->sa_data);
> memcpy(wx->mac.addr, addr->sa_data, netdev->addr_len);
>
> @@ -1178,6 +1228,10 @@ static int wx_hpbthresh(struct wx *wx)
> /* Calculate delay value for device */
> dv_id = WX_DV(link, tc);
>
> + /* Loopback switch introduces additional latency */
> + if (test_bit(WX_FLAG_SRIOV_ENABLED, wx->flags))
> + dv_id += WX_B2BT(tc);
> +
> /* Delay value is calculated in bit times convert to KB */
> kb = WX_BT2KB(dv_id);
> rx_pba = rd32(wx, WX_RDB_PB_SZ(0)) >> WX_RDB_PB_SZ_SHIFT;
> @@ -1233,12 +1287,106 @@ static void wx_pbthresh_setup(struct wx *wx)
> wx->fc.low_water = 0;
> }
>
> +static void wx_set_ethertype_anti_spoofing(struct wx *wx, bool enable, int vf)
> +{
> + u32 pfvfspoof, reg_offset, vf_shift;
> +
> + vf_shift = vf % 32;
> + reg_offset = vf / 32;
> +
> + pfvfspoof = rd32(wx, WX_TDM_ETYPE_AS(reg_offset));
> + if (enable)
> + pfvfspoof |= BIT(vf_shift);
> + else
> + pfvfspoof &= ~BIT(vf_shift);
> + wr32(wx, WX_TDM_ETYPE_AS(reg_offset), pfvfspoof);
> +}
> +
> +static int wx_set_vf_spoofchk(struct net_device *netdev, int vf, bool setting)
> +{
> + u32 index = vf / 32, vf_bit = vf % 32;
> + struct wx *wx = netdev_priv(netdev);
> + u32 regval;
> +
> + if (vf >= wx->num_vfs)
> + return -EINVAL;
> +
> + wx->vfinfo[vf].spoofchk_enabled = setting;
> +
> + regval = (setting << vf_bit);
> + wr32m(wx, WX_TDM_MAC_AS(index), regval | BIT(vf_bit), regval);
> +
> + if (wx->vfinfo[vf].vlan_count)
> + wr32m(wx, WX_TDM_VLAN_AS(index), regval | BIT(vf_bit), regval);
> +
> + return 0;
> +}
> +
> +static void wx_configure_virtualization(struct wx *wx)
> +{
> + u16 pool = wx->num_rx_pools;
> + u32 reg_offset, vf_shift;
> + u32 i;
> +
> + if (!test_bit(WX_FLAG_SRIOV_ENABLED, wx->flags))
> + return;
> +
> + wr32m(wx, WX_PSR_VM_CTL,
> + WX_PSR_VM_CTL_POOL_MASK | WX_PSR_VM_CTL_REPLEN,
> + FIELD_PREP(WX_PSR_VM_CTL_POOL_MASK, VMDQ_P(0)) |
> + WX_PSR_VM_CTL_REPLEN);
> + while (pool--)
> + wr32m(wx, WX_PSR_VM_L2CTL(pool), WX_PSR_VM_L2CTL_AUPE, WX_PSR_VM_L2CTL_AUPE);
> +
> + if (wx->mac.type == wx_mac_sp) {
> + vf_shift = VMDQ_P(0) % 32;
> + reg_offset = VMDQ_P(0) / 32;
> +
> + /* Enable only the PF pools for Tx/Rx */
> + wr32(wx, WX_RDM_VF_RE(reg_offset), GENMASK(31, vf_shift));
> + wr32(wx, WX_RDM_VF_RE(reg_offset ^ 1), reg_offset - 1);
> + wr32(wx, WX_TDM_VF_TE(reg_offset), GENMASK(31, vf_shift));
> + wr32(wx, WX_TDM_VF_TE(reg_offset ^ 1), reg_offset - 1);
> + } else {
> + vf_shift = BIT(VMDQ_P(0));
> + /* Enable only the PF pools for Tx/Rx */
> + wr32(wx, WX_RDM_VF_RE(0), vf_shift);
> + wr32(wx, WX_TDM_VF_TE(0), vf_shift);
> + }
> +
> + /* clear VLAN promisc flag so VFTA will be updated if necessary */
> + clear_bit(WX_FLAG2_VLAN_PROMISC, wx->flags);
> +
> + for (i = 0; i < wx->num_vfs; i++) {
> + if (!wx->vfinfo[i].spoofchk_enabled)
> + wx_set_vf_spoofchk(wx->netdev, i, false);
> + /* enable ethertype anti spoofing if hw supports it */
> + wx_set_ethertype_anti_spoofing(wx, true, i);
> + }
> +}
> +
> static void wx_configure_port(struct wx *wx)
> {
> u32 value, i;
>
> - value = WX_CFG_PORT_CTL_D_VLAN | WX_CFG_PORT_CTL_QINQ;
> + if (wx->mac.type == wx_mac_em) {
> + value = (wx->num_vfs == 0) ?
> + WX_CFG_PORT_CTL_NUM_VT_NONE :
> + WX_CFG_PORT_CTL_NUM_VT_8;
> + } else {
> + if (test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags)) {
> + if (wx->ring_feature[RING_F_RSS].indices == 4)
> + value = WX_CFG_PORT_CTL_NUM_VT_32;
> + else
> + value = WX_CFG_PORT_CTL_NUM_VT_64;
> + } else {
> + value = 0;
> + }
> + }
> +
> + value |= WX_CFG_PORT_CTL_D_VLAN | WX_CFG_PORT_CTL_QINQ;
> wr32m(wx, WX_CFG_PORT_CTL,
> + WX_CFG_PORT_CTL_NUM_VT_MASK |
> WX_CFG_PORT_CTL_D_VLAN |
> WX_CFG_PORT_CTL_QINQ,
> value);
> @@ -1297,6 +1445,83 @@ static void wx_vlan_strip_control(struct wx *wx, bool enable)
> }
> }
>
> +static void wx_vlan_promisc_enable(struct wx *wx)
> +{
> + u32 vlnctrl, i, vind, bits, reg_idx;
> +
> + vlnctrl = rd32(wx, WX_PSR_VLAN_CTL);
> + if (test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags)) {
> + /* we need to keep the VLAN filter on in SRIOV */
> + vlnctrl |= WX_PSR_VLAN_CTL_VFE;
> + wr32(wx, WX_PSR_VLAN_CTL, vlnctrl);
> + } else {
> + vlnctrl &= ~WX_PSR_VLAN_CTL_VFE;
> + wr32(wx, WX_PSR_VLAN_CTL, vlnctrl);
> + return;
> + }
> + /* We are already in VLAN promisc, nothing to do */
> + if (test_bit(WX_FLAG2_VLAN_PROMISC, wx->flags))
> + return;
> + /* Set flag so we don't redo unnecessary work */
> + set_bit(WX_FLAG2_VLAN_PROMISC, wx->flags);
> + /* Add PF to all active pools */
> + for (i = WX_PSR_VLAN_SWC_ENTRIES; --i;) {
> + wr32(wx, WX_PSR_VLAN_SWC_IDX, i);
> + reg_idx = VMDQ_P(0) / 32;
> + vind = VMDQ_P(0) % 32;
> + bits = rd32(wx, WX_PSR_VLAN_SWC_VM(reg_idx));
> + bits |= BIT(vind);
> + wr32(wx, WX_PSR_VLAN_SWC_VM(reg_idx), bits);
> + }
> + /* Set all bits in the VLAN filter table array */
> + for (i = 0; i < wx->mac.vft_size; i++)
> + wr32(wx, WX_PSR_VLAN_TBL(i), U32_MAX);
> +}
> +
> +static void wx_scrub_vfta(struct wx *wx)
> +{
> + u32 i, vid, bits, vfta, vind, vlvf, reg_idx;
> +
> + for (i = WX_PSR_VLAN_SWC_ENTRIES; --i;) {
> + wr32(wx, WX_PSR_VLAN_SWC_IDX, i);
> + vlvf = rd32(wx, WX_PSR_VLAN_SWC_IDX);
> + /* pull VLAN ID from VLVF */
> + vid = vlvf & ~WX_PSR_VLAN_SWC_VIEN;
> + if (vlvf & WX_PSR_VLAN_SWC_VIEN) {
> + /* if PF is part of this then continue */
> + if (test_bit(vid, wx->active_vlans))
> + continue;
> + }
> + /* remove PF from the pool */
> + reg_idx = VMDQ_P(0) / 32;
> + vind = VMDQ_P(0) % 32;
> + bits = rd32(wx, WX_PSR_VLAN_SWC_VM(reg_idx));
> + bits &= ~BIT(vind);
> + wr32(wx, WX_PSR_VLAN_SWC_VM(reg_idx), bits);
> + }
> + /* extract values from vft_shadow and write back to VFTA */
> + for (i = 0; i < wx->mac.vft_size; i++) {
> + vfta = wx->mac.vft_shadow[i];
> + wr32(wx, WX_PSR_VLAN_TBL(i), vfta);
> + }
> +}
> +
> +static void wx_vlan_promisc_disable(struct wx *wx)
> +{
> + u32 vlnctrl;
> +
> + /* configure vlan filtering */
> + vlnctrl = rd32(wx, WX_PSR_VLAN_CTL);
> + vlnctrl |= WX_PSR_VLAN_CTL_VFE;
> + wr32(wx, WX_PSR_VLAN_CTL, vlnctrl);
> + /* We are not in VLAN promisc, nothing to do */
> + if (!test_bit(WX_FLAG2_VLAN_PROMISC, wx->flags))
> + return;
> + /* Set flag so we don't redo unnecessary work */
> + clear_bit(WX_FLAG2_VLAN_PROMISC, wx->flags);
> + wx_scrub_vfta(wx);
> +}
> +
> void wx_set_rx_mode(struct net_device *netdev)
> {
> struct wx *wx = netdev_priv(netdev);
> @@ -1309,7 +1534,7 @@ void wx_set_rx_mode(struct net_device *netdev)
> /* Check for Promiscuous and All Multicast modes */
> fctrl = rd32(wx, WX_PSR_CTL);
> fctrl &= ~(WX_PSR_CTL_UPE | WX_PSR_CTL_MPE);
> - vmolr = rd32(wx, WX_PSR_VM_L2CTL(0));
> + vmolr = rd32(wx, WX_PSR_VM_L2CTL(VMDQ_P(0)));
> vmolr &= ~(WX_PSR_VM_L2CTL_UPE |
> WX_PSR_VM_L2CTL_MPE |
> WX_PSR_VM_L2CTL_ROPE |
> @@ -1330,7 +1555,10 @@ void wx_set_rx_mode(struct net_device *netdev)
> fctrl |= WX_PSR_CTL_UPE | WX_PSR_CTL_MPE;
> /* pf don't want packets routing to vf, so clear UPE */
> vmolr |= WX_PSR_VM_L2CTL_MPE;
> - vlnctrl &= ~WX_PSR_VLAN_CTL_VFE;
> + if (test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags) &&
> + test_bit(WX_FLAG_SRIOV_ENABLED, wx->flags))
> + vlnctrl |= WX_PSR_VLAN_CTL_VFE;
> + features &= ~NETIF_F_HW_VLAN_CTAG_FILTER;
> }
>
> if (netdev->flags & IFF_ALLMULTI) {
> @@ -1353,7 +1581,7 @@ void wx_set_rx_mode(struct net_device *netdev)
> * sufficient space to store all the addresses then enable
> * unicast promiscuous mode
> */
> - count = wx_write_uc_addr_list(netdev, 0);
> + count = wx_write_uc_addr_list(netdev, VMDQ_P(0));
> if (count < 0) {
> vmolr &= ~WX_PSR_VM_L2CTL_ROPE;
> vmolr |= WX_PSR_VM_L2CTL_UPE;
> @@ -1371,7 +1599,7 @@ void wx_set_rx_mode(struct net_device *netdev)
>
> wr32(wx, WX_PSR_VLAN_CTL, vlnctrl);
> wr32(wx, WX_PSR_CTL, fctrl);
> - wr32(wx, WX_PSR_VM_L2CTL(0), vmolr);
> + wr32(wx, WX_PSR_VM_L2CTL(VMDQ_P(0)), vmolr);
>
> if ((features & NETIF_F_HW_VLAN_CTAG_RX) &&
> (features & NETIF_F_HW_VLAN_STAG_RX))
> @@ -1379,6 +1607,10 @@ void wx_set_rx_mode(struct net_device *netdev)
> else
> wx_vlan_strip_control(wx, false);
>
> + if (features & NETIF_F_HW_VLAN_CTAG_FILTER)
> + wx_vlan_promisc_disable(wx);
> + else
> + wx_vlan_promisc_enable(wx);
> }
> EXPORT_SYMBOL(wx_set_rx_mode);
>
> @@ -1621,6 +1853,13 @@ static void wx_setup_reta(struct wx *wx)
> u32 random_key_size = WX_RSS_KEY_SIZE / 4;
> u32 i, j;
>
> + if (test_bit(WX_FLAG_SRIOV_ENABLED, wx->flags)) {
> + if (wx->mac.type == wx_mac_sp)
> + rss_i = rss_i < 4 ? 4 : rss_i;
> + else if (wx->mac.type == wx_mac_em)
> + rss_i = 1;
> + }
> +
> /* Fill out hash function seeds */
> for (i = 0; i < random_key_size; i++)
> wr32(wx, WX_RDB_RSSRK(i), wx->rss_key[i]);
> @@ -1638,10 +1877,40 @@ static void wx_setup_reta(struct wx *wx)
> wx_store_reta(wx);
> }
>
> +static void wx_setup_psrtype(struct wx *wx)
> +{
> + int rss_i = wx->ring_feature[RING_F_RSS].indices;
> + u32 psrtype;
> + int pool;
> +
> + psrtype = WX_RDB_PL_CFG_L4HDR |
> + WX_RDB_PL_CFG_L3HDR |
> + WX_RDB_PL_CFG_L2HDR |
> + WX_RDB_PL_CFG_TUN_OUTL2HDR |
> + WX_RDB_PL_CFG_TUN_TUNHDR;
> +
> + if (wx->mac.type == wx_mac_sp) {
> + if (rss_i > 3)
> + psrtype |= FIELD_PREP(GENMASK(31, 29), 2);
> + else if (rss_i > 1)
> + psrtype |= FIELD_PREP(GENMASK(31, 29), 1);
> +
> + for_each_set_bit(pool, &wx->fwd_bitmask, 32)
> + wr32(wx, WX_RDB_PL_CFG(VMDQ_P(pool)), psrtype);
> + } else {
> + for_each_set_bit(pool, &wx->fwd_bitmask, 8)
> + wr32(wx, WX_RDB_PL_CFG(VMDQ_P(pool)), psrtype);
> + }
> +}
> +
> static void wx_setup_mrqc(struct wx *wx)
> {
> u32 rss_field = 0;
>
> + /* VT, and RSS do not coexist at the same time */
> + if (test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags))
> + return;
> +
> /* Disable indicating checksum in descriptor, enables RSS hash */
> wr32m(wx, WX_PSR_CTL, WX_PSR_CTL_PCSD, WX_PSR_CTL_PCSD);
>
> @@ -1671,16 +1940,11 @@ static void wx_setup_mrqc(struct wx *wx)
> **/
> void wx_configure_rx(struct wx *wx)
> {
> - u32 psrtype, i;
> int ret;
> + u32 i;
>
> wx_disable_rx(wx);
> -
> - psrtype = WX_RDB_PL_CFG_L4HDR |
> - WX_RDB_PL_CFG_L3HDR |
> - WX_RDB_PL_CFG_L2HDR |
> - WX_RDB_PL_CFG_TUN_TUNHDR;
> - wr32(wx, WX_RDB_PL_CFG(0), psrtype);
> + wx_setup_psrtype(wx);
>
> /* enable hw crc stripping */
> wr32m(wx, WX_RSC_CTL, WX_RSC_CTL_CRC_STRIP, WX_RSC_CTL_CRC_STRIP);
> @@ -1728,6 +1992,7 @@ void wx_configure(struct wx *wx)
> {
> wx_set_rxpba(wx);
> wx_pbthresh_setup(wx);
> + wx_configure_virtualization(wx);
> wx_configure_port(wx);
>
> wx_set_rx_mode(wx->netdev);
> diff --git a/drivers/net/ethernet/wangxun/libwx/wx_lib.c b/drivers/net/ethernet/wangxun/libwx/wx_lib.c
> index 68bde91b67a0..8e4c0e24a4a3 100644
> --- a/drivers/net/ethernet/wangxun/libwx/wx_lib.c
> +++ b/drivers/net/ethernet/wangxun/libwx/wx_lib.c
> @@ -1558,6 +1558,65 @@ void wx_napi_disable_all(struct wx *wx)
> }
> EXPORT_SYMBOL(wx_napi_disable_all);
>
> +static bool wx_set_vmdq_queues(struct wx *wx)
> +{
> + u16 vmdq_i = wx->ring_feature[RING_F_VMDQ].limit;
> + u16 rss_i = wx->ring_feature[RING_F_RSS].limit;
> + u16 rss_m = WX_RSS_DISABLED_MASK;
> + u16 vmdq_m = 0;
> +
> + /* only proceed if VMDq is enabled */
> + if (!test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags))
> + return false;
> + /* Add starting offset to total pool count */
> + vmdq_i += wx->ring_feature[RING_F_VMDQ].offset;
> +
> + if (wx->mac.type == wx_mac_sp) {
> + /* double check we are limited to maximum pools */
> + vmdq_i = min_t(u16, 64, vmdq_i);
> +
> + /* 64 pool mode with 2 queues per pool, or
> + * 16/32/64 pool mode with 1 queue per pool
> + */
> + if (vmdq_i > 32 || rss_i < 4) {
> + vmdq_m = WX_VMDQ_2Q_MASK;
> + rss_m = WX_RSS_2Q_MASK;
> + rss_i = min_t(u16, rss_i, 2);
> + /* 32 pool mode with 4 queues per pool */
> + } else {
> + vmdq_m = WX_VMDQ_4Q_MASK;
> + rss_m = WX_RSS_4Q_MASK;
> + rss_i = 4;
> + }
> + } else {
> + /* double check we are limited to maximum pools */
> + vmdq_i = min_t(u16, 8, vmdq_i);
> +
> + /* when VMDQ on, disable RSS */
> + rss_i = 1;
> + }
> +
> + /* remove the starting offset from the pool count */
> + vmdq_i -= wx->ring_feature[RING_F_VMDQ].offset;
> +
> + /* save features for later use */
> + wx->ring_feature[RING_F_VMDQ].indices = vmdq_i;
> + wx->ring_feature[RING_F_VMDQ].mask = vmdq_m;
> +
> + /* limit RSS based on user input and save for later use */
> + wx->ring_feature[RING_F_RSS].indices = rss_i;
> + wx->ring_feature[RING_F_RSS].mask = rss_m;
> +
> + wx->queues_per_pool = rss_i;/*maybe same to num_rx_queues_per_pool*/
> + wx->num_rx_pools = vmdq_i;
> + wx->num_rx_queues_per_pool = rss_i;
> +
> + wx->num_rx_queues = vmdq_i * rss_i;
> + wx->num_tx_queues = vmdq_i * rss_i;
> +
> + return true;
> +}
> +
> /**
> * wx_set_rss_queues: Allocate queues for RSS
> * @wx: board private structure to initialize
> @@ -1574,6 +1633,11 @@ static void wx_set_rss_queues(struct wx *wx)
> f = &wx->ring_feature[RING_F_RSS];
> f->indices = f->limit;
>
> + if (wx->mac.type == wx_mac_sp)
> + f->mask = WX_RSS_64Q_MASK;
> + else
> + f->mask = WX_RSS_8Q_MASK;
> +
> wx->num_rx_queues = f->limit;
> wx->num_tx_queues = f->limit;
> }
> @@ -1585,6 +1649,9 @@ static void wx_set_num_queues(struct wx *wx)
> wx->num_tx_queues = 1;
> wx->queues_per_pool = 1;
>
> + if (wx_set_vmdq_queues(wx))
> + return;
> +
> wx_set_rss_queues(wx);
> }
>
> @@ -1665,6 +1732,10 @@ static int wx_set_interrupt_capability(struct wx *wx)
> if (ret == 0 || (ret == -ENOMEM))
> return ret;
>
> + /* Disable VMDq support */
> + dev_warn(&wx->pdev->dev, "Disabling VMQQ support\n");
> + clear_bit(WX_FLAG_VMDQ_ENABLED, wx->flags);
> +
> /* Disable RSS */
> dev_warn(&wx->pdev->dev, "Disabling RSS support\n");
> wx->ring_feature[RING_F_RSS].limit = 1;
> @@ -1690,6 +1761,49 @@ static int wx_set_interrupt_capability(struct wx *wx)
> return 0;
> }
>
> +static bool wx_cache_ring_vmdq(struct wx *wx)
> +{
> + struct wx_ring_feature *vmdq = &wx->ring_feature[RING_F_VMDQ];
> + struct wx_ring_feature *rss = &wx->ring_feature[RING_F_RSS];
> + u16 reg_idx;
> + int i;
> +
> + /* only proceed if VMDq is enabled */
> + if (!test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags))
> + return false;
> +
> + if (wx->mac.type == wx_mac_sp) {
> + /* start at VMDq register offset for SR-IOV enabled setups */
> + reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask);
> + for (i = 0; i < wx->num_rx_queues; i++, reg_idx++) {
> + /* If we are greater than indices move to next pool */
> + if ((reg_idx & ~vmdq->mask) >= rss->indices)
> + reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask);
> + wx->rx_ring[i]->reg_idx = reg_idx;
> + }
> + reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask);
> + for (i = 0; i < wx->num_tx_queues; i++, reg_idx++) {
> + /* If we are greater than indices move to next pool */
> + if ((reg_idx & rss->mask) >= rss->indices)
> + reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask);
> + wx->tx_ring[i]->reg_idx = reg_idx;
> + }
> + } else {
> + /* start at VMDq register offset for SR-IOV enabled setups */
> + reg_idx = vmdq->offset;
> + for (i = 0; i < wx->num_rx_queues; i++)
> + /* If we are greater than indices move to next pool */
> + wx->rx_ring[i]->reg_idx = reg_idx + i;
> +
> + reg_idx = vmdq->offset;
> + for (i = 0; i < wx->num_tx_queues; i++)
> + /* If we are greater than indices move to next pool */
> + wx->tx_ring[i]->reg_idx = reg_idx + i;
> + }
> +
> + return true;
> +}
> +
> /**
> * wx_cache_ring_rss - Descriptor ring to register mapping for RSS
> * @wx: board private structure to initialize
> @@ -1701,6 +1815,9 @@ static void wx_cache_ring_rss(struct wx *wx)
> {
> u16 i;
>
> + if (wx_cache_ring_vmdq(wx))
> + return;
> +
> for (i = 0; i < wx->num_rx_queues; i++)
> wx->rx_ring[i]->reg_idx = i;
>
> @@ -2089,7 +2206,8 @@ static void wx_set_ivar(struct wx *wx, s8 direction,
> wr32(wx, WX_PX_MISC_IVAR, ivar);
> } else {
> /* tx or rx causes */
> - msix_vector += 1; /* offset for queue vectors */
> + if (!(wx->mac.type == wx_mac_em && wx->num_vfs == 7))
> + msix_vector += 1; /* offset for queue vectors */
> msix_vector |= WX_PX_IVAR_ALLOC_VAL;
> index = ((16 * (queue & 1)) + (8 * direction));
> ivar = rd32(wx, WX_PX_IVAR(queue >> 1));
> @@ -2134,10 +2252,17 @@ void wx_configure_vectors(struct wx *wx)
> {
> struct pci_dev *pdev = wx->pdev;
> u32 eitrsel = 0;
> - u16 v_idx;
> + u16 v_idx, i;
>
> if (pdev->msix_enabled) {
> /* Populate MSIX to EITR Select */
> + if (wx->mac.type == wx_mac_sp) {
> + if (wx->num_vfs >= 32)
> + eitrsel = BIT(wx->num_vfs % 32) - 1;
> + } else if (wx->mac.type == wx_mac_em) {
> + for (i = 0; i < wx->num_vfs; i++)
> + eitrsel |= BIT(i);
> + }
> wr32(wx, WX_PX_ITRSEL, eitrsel);
> /* use EIAM to auto-mask when MSI-X interrupt is asserted
> * this saves a register write for every interrupt
> diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h
> index 7dad022e01e9..126416534181 100644
> --- a/drivers/net/ethernet/wangxun/libwx/wx_type.h
> +++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h
> @@ -19,6 +19,7 @@
> #define WX_PCIE_MSIX_TBL_SZ_MASK 0x7FF
> #define WX_PCI_LINK_STATUS 0xB2
> #define WX_MAX_PF_MACVLANS 15
> +#define WX_MAX_VF_MC_ENTRIES 30
>
> /**************** Global Registers ****************************/
> /* chip control Registers */
> @@ -75,6 +76,7 @@
> #define WX_MAC_LXONOFFRXC 0x11E0C
>
> /*********************** Receive DMA registers **************************/
> +#define WX_RDM_VF_RE(_i) (0x12004 + ((_i) * 4))
> #define WX_RDM_DRP_PKT 0x12500
> #define WX_RDM_PKT_CNT 0x12504
> #define WX_RDM_BYTE_CNT_LSB 0x12508
> @@ -89,6 +91,7 @@
> #define WX_CFG_TAG_TPID(_i) (0x14430 + ((_i) * 4))
> #define WX_CFG_PORT_CTL_NUM_VT_MASK GENMASK(13, 12) /* number of TVs */
>
> +#define WX_CFG_PORT_CTL_NUM_VT_NONE 0
> #define WX_CFG_PORT_CTL_NUM_VT_8 FIELD_PREP(GENMASK(13, 12), 1)
> #define WX_CFG_PORT_CTL_NUM_VT_32 FIELD_PREP(GENMASK(13, 12), 2)
> #define WX_CFG_PORT_CTL_NUM_VT_64 FIELD_PREP(GENMASK(13, 12), 3)
> @@ -114,6 +117,10 @@
> /*********************** Transmit DMA registers **************************/
> /* transmit global control */
> #define WX_TDM_CTL 0x18000
> +#define WX_TDM_VF_TE(_i) (0x18004 + ((_i) * 4))
> +#define WX_TDM_MAC_AS(_i) (0x18060 + ((_i) * 4))
> +#define WX_TDM_VLAN_AS(_i) (0x18070 + ((_i) * 4))
> +
> /* TDM CTL BIT */
> #define WX_TDM_CTL_TE BIT(0) /* Transmit Enable */
> #define WX_TDM_PB_THRE(_i) (0x18020 + ((_i) * 4))
> @@ -186,6 +193,7 @@
> /* mcasst/ucast overflow tbl */
> #define WX_PSR_MC_TBL(_i) (0x15200 + ((_i) * 4))
> #define WX_PSR_UC_TBL(_i) (0x15400 + ((_i) * 4))
> +#define WX_PSR_VM_CTL_REPLEN BIT(30) /* replication enabled */
> #define WX_PSR_VM_CTL_POOL_MASK GENMASK(12, 7)
>
> /* VM L2 contorl */
> @@ -230,6 +238,7 @@
> #define WX_PSR_VLAN_SWC 0x16220
> #define WX_PSR_VLAN_SWC_VM_L 0x16224
> #define WX_PSR_VLAN_SWC_VM_H 0x16228
> +#define WX_PSR_VLAN_SWC_VM(_i) (0x16224 + ((_i) * 4))
> #define WX_PSR_VLAN_SWC_IDX 0x16230 /* 64 vlan entries */
> /* VLAN pool filtering masks */
> #define WX_PSR_VLAN_SWC_VIEN BIT(31) /* filter is valid */
> @@ -244,6 +253,10 @@
> #define WX_RSC_ST 0x17004
> #define WX_RSC_ST_RSEC_RDY BIT(0)
>
> +/*********************** Transmit DMA registers **************************/
> +/* transmit global control */
> +#define WX_TDM_ETYPE_AS(_i) (0x18058 + ((_i) * 4))
> +
> /****************************** TDB ******************************************/
> #define WX_TDB_PB_SZ(_i) (0x1CC00 + ((_i) * 4))
> #define WX_TXPKT_SIZE_MAX 0xA /* Max Tx Packet size */
> @@ -371,6 +384,15 @@ enum WX_MSCA_CMD_value {
> /* Number of 80 microseconds we wait for PCI Express master disable */
> #define WX_PCI_MASTER_DISABLE_TIMEOUT 80000
>
> +#define WX_RSS_64Q_MASK 0x3F
> +#define WX_RSS_8Q_MASK 0x7
> +#define WX_RSS_4Q_MASK 0x3
> +#define WX_RSS_2Q_MASK 0x1
> +#define WX_RSS_DISABLED_MASK 0x0
> +
> +#define WX_VMDQ_4Q_MASK 0x7C
> +#define WX_VMDQ_2Q_MASK 0x7E
> +
> /****************** Manageablility Host Interface defines ********************/
> #define WX_HI_MAX_BLOCK_BYTE_LENGTH 256 /* Num of bytes in range */
> #define WX_HI_COMMAND_TIMEOUT 1000 /* Process HI command limit */
> @@ -435,7 +457,12 @@ enum WX_MSCA_CMD_value {
> #define WX_REQ_TX_DESCRIPTOR_MULTIPLE 8
>
> #define WX_MAX_JUMBO_FRAME_SIZE 9432 /* max payload 9414 */
> -#define VMDQ_P(p) p
> +/* must account for pools assigned to VFs. */
> +#ifdef CONFIG_PCI_IOV
> +#define VMDQ_P(p) ((p) + wx->ring_feature[RING_F_VMDQ].offset)
> +#else
> +#define VMDQ_P(p) (p)
> +#endif
>
> /* Supported Rx Buffer Sizes */
> #define WX_RXBUFFER_256 256 /* Used for skb receive header */
> @@ -1005,6 +1032,10 @@ struct vf_data_storage {
> bool link_enable;
> bool trusted;
> int xcast_mode;
> +
> + u16 vf_mc_hashes[WX_MAX_VF_MC_ENTRIES];
> + u16 num_vf_mc_hashes;
> + u16 vlan_count;
> };
>
> struct vf_macvlans {
> @@ -1017,6 +1048,7 @@ struct vf_macvlans {
>
> enum wx_pf_flags {
> WX_FLAG_VMDQ_ENABLED,
> + WX_FLAG2_VLAN_PROMISC,
> WX_FLAG_SRIOV_ENABLED,
> WX_PF_FLAGS_NBITS /* must be last */
> };
> @@ -1085,6 +1117,8 @@ struct wx {
> struct wx_ring *tx_ring[64] ____cacheline_aligned_in_smp;
> struct wx_ring *rx_ring[64];
> struct wx_q_vector *q_vector[64];
> + int num_rx_pools; /* does not include pools assigned to VFs */
> + int num_rx_queues_per_pool;
>
> unsigned int queues_per_pool;
> struct msix_entry *msix_q_entries;
> @@ -1118,6 +1152,7 @@ struct wx {
> struct vf_data_storage *vfinfo;
> struct vf_macvlans vf_mvs;
> struct vf_macvlans *mv_list;
> + unsigned long fwd_bitmask; /* bitmask indicating in use pools */
>
> int (*setup_tc)(struct net_device *netdev, u8 tc);
> void (*do_reset)(struct net_device *netdev);
Powered by blists - more mailing lists