diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2012-07-24 21:01:50 +0400 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2012-07-24 21:01:50 +0400 |
commit | 3c4cfadef6a1665d9cd02a543782d03d3e6740c6 (patch) | |
tree | 3df72faaacd494d5ac8c9668df4f529b1b5e4457 /drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c | |
parent | e017507f37d5cb8b541df165a824958bc333bec3 (diff) | |
parent | 320f5ea0cedc08ef65d67e056bcb9d181386ef2c (diff) | |
download | linux-3c4cfadef6a1665d9cd02a543782d03d3e6740c6.tar.xz |
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next
Pull networking changes from David S Miller:
1) Remove the ipv4 routing cache. Now lookups go directly into the FIB
trie and use prebuilt routes cached there.
No more garbage collection, no more rDOS attacks on the routing
cache. Instead we now get predictable and consistent performance,
no matter what the pattern of traffic we service.
This has been almost 2 years in the making. Special thanks to
Julian Anastasov, Eric Dumazet, Steffen Klassert, and others who
have helped along the way.
I'm sure that with a change of this magnitude there will be some
kind of fallout, but such things ought the be simple to fix at this
point. Luckily I'm not European so I'll be around all of August to
fix things :-)
The major stages of this work here are each fronted by a forced
merge commit whose commit message contains a top-level description
of the motivations and implementation issues.
2) Pre-demux of established ipv4 TCP sockets, saves a route demux on
input.
3) TCP SYN/ACK performance tweaks from Eric Dumazet.
4) Add namespace support for netfilter L4 conntrack helpers, from Gao
Feng.
5) Add config mechanism for Energy Efficient Ethernet to ethtool, from
Yuval Mintz.
6) Remove quadratic behavior from /proc/net/unix, from Eric Dumazet.
7) Support for connection tracker helpers in userspace, from Pablo
Neira Ayuso.
8) Allow userspace driven TX load balancing functions in TEAM driver,
from Jiri Pirko.
9) Kill off NLMSG_PUT and RTA_PUT macros, more gross stuff with
embedded gotos.
10) TCP Small Queues, essentially minimize the amount of TCP data queued
up in the packet scheduler layer. Whereas the existing BQL (Byte
Queue Limits) limits the pkt_sched --> netdevice queuing levels,
this controls the TCP --> pkt_sched queueing levels.
From Eric Dumazet.
11) Reduce the number of get_page/put_page ops done on SKB fragments,
from Alexander Duyck.
12) Implement protection against blind resets in TCP (RFC 5961), from
Eric Dumazet.
13) Support the client side of TCP Fast Open, basically the ability to
send data in the SYN exchange, from Yuchung Cheng.
Basically, the sender queues up data with a sendmsg() call using
MSG_FASTOPEN, then they do the connect() which emits the queued up
fastopen data.
14) Avoid all the problems we get into in TCP when timers or PMTU events
hit a locked socket. The TCP Small Queues changes added a
tcp_release_cb() that allows us to queue work up to the
release_sock() caller, and that's what we use here too. From Eric
Dumazet.
15) Zero copy on TX support for TUN driver, from Michael S. Tsirkin.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1870 commits)
genetlink: define lockdep_genl_is_held() when CONFIG_LOCKDEP
r8169: revert "add byte queue limit support".
ipv4: Change rt->rt_iif encoding.
net: Make skb->skb_iif always track skb->dev
ipv4: Prepare for change of rt->rt_iif encoding.
ipv4: Remove all RTCF_DIRECTSRC handliing.
ipv4: Really ignore ICMP address requests/replies.
decnet: Don't set RTCF_DIRECTSRC.
net/ipv4/ip_vti.c: Fix __rcu warnings detected by sparse.
ipv4: Remove redundant assignment
rds: set correct msg_namelen
openvswitch: potential NULL deref in sample()
tcp: dont drop MTU reduction indications
bnx2x: Add new 57840 device IDs
tcp: avoid oops in tcp_metrics and reset tcpm_stamp
niu: Change niu_rbr_fill() to use unlikely() to check niu_rbr_add_page() return value
niu: Fix to check for dma mapping errors.
net: Fix references to out-of-scope variables in put_cmsg_compat()
net: ethernet: davinci_emac: add pm_runtime support
net: ethernet: davinci_emac: Remove unnecessary #include
...
Diffstat (limited to 'drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c')
-rw-r--r-- | drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c | 838 |
1 files changed, 536 insertions, 302 deletions
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c index c377706e81a8..17ecbcedd548 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c @@ -28,28 +28,83 @@ #include "ixgbe.h" #include "ixgbe_sriov.h" +#ifdef CONFIG_IXGBE_DCB /** - * ixgbe_cache_ring_rss - Descriptor ring to register mapping for RSS + * ixgbe_cache_ring_dcb_sriov - Descriptor ring to register mapping for SR-IOV * @adapter: board private structure to initialize * - * Cache the descriptor ring offsets for RSS to the assigned rings. + * Cache the descriptor ring offsets for SR-IOV to the assigned rings. It + * will also try to cache the proper offsets if RSS/FCoE are enabled along + * with VMDq. * **/ -static inline bool ixgbe_cache_ring_rss(struct ixgbe_adapter *adapter) +static bool ixgbe_cache_ring_dcb_sriov(struct ixgbe_adapter *adapter) { +#ifdef IXGBE_FCOE + struct ixgbe_ring_feature *fcoe = &adapter->ring_feature[RING_F_FCOE]; +#endif /* IXGBE_FCOE */ + struct ixgbe_ring_feature *vmdq = &adapter->ring_feature[RING_F_VMDQ]; int i; + u16 reg_idx; + u8 tcs = netdev_get_num_tc(adapter->netdev); - if (!(adapter->flags & IXGBE_FLAG_RSS_ENABLED)) + /* verify we have DCB queueing enabled before proceeding */ + if (tcs <= 1) return false; - for (i = 0; i < adapter->num_rx_queues; i++) - adapter->rx_ring[i]->reg_idx = i; - for (i = 0; i < adapter->num_tx_queues; i++) - adapter->tx_ring[i]->reg_idx = i; + /* verify we have VMDq enabled before proceeding */ + if (!(adapter->flags & IXGBE_FLAG_SRIOV_ENABLED)) + return false; + + /* start at VMDq register offset for SR-IOV enabled setups */ + reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask); + for (i = 0; i < adapter->num_rx_queues; i++, reg_idx++) { + /* If we are greater than indices move to next pool */ + if ((reg_idx & ~vmdq->mask) >= tcs) + reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask); + adapter->rx_ring[i]->reg_idx = reg_idx; + } + + reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask); + for (i = 0; i < adapter->num_tx_queues; i++, reg_idx++) { + /* If we are greater than indices move to next pool */ + if ((reg_idx & ~vmdq->mask) >= tcs) + reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask); + adapter->tx_ring[i]->reg_idx = reg_idx; + } + +#ifdef IXGBE_FCOE + /* nothing to do if FCoE is disabled */ + if (!(adapter->flags & IXGBE_FLAG_FCOE_ENABLED)) + return true; + + /* The work is already done if the FCoE ring is shared */ + if (fcoe->offset < tcs) + return true; + + /* The FCoE rings exist separately, we need to move their reg_idx */ + if (fcoe->indices) { + u16 queues_per_pool = __ALIGN_MASK(1, ~vmdq->mask); + u8 fcoe_tc = ixgbe_fcoe_get_tc(adapter); + + reg_idx = (vmdq->offset + vmdq->indices) * queues_per_pool; + for (i = fcoe->offset; i < adapter->num_rx_queues; i++) { + reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask) + fcoe_tc; + adapter->rx_ring[i]->reg_idx = reg_idx; + reg_idx++; + } + + reg_idx = (vmdq->offset + vmdq->indices) * queues_per_pool; + for (i = fcoe->offset; i < adapter->num_tx_queues; i++) { + reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask) + fcoe_tc; + adapter->tx_ring[i]->reg_idx = reg_idx; + reg_idx++; + } + } +#endif /* IXGBE_FCOE */ return true; } -#ifdef CONFIG_IXGBE_DCB /* ixgbe_get_first_reg_idx - Return first register index associated with ring */ static void ixgbe_get_first_reg_idx(struct ixgbe_adapter *adapter, u8 tc, @@ -64,42 +119,37 @@ static void ixgbe_get_first_reg_idx(struct ixgbe_adapter *adapter, u8 tc, switch (hw->mac.type) { case ixgbe_mac_82598EB: - *tx = tc << 2; - *rx = tc << 3; + /* TxQs/TC: 4 RxQs/TC: 8 */ + *tx = tc << 2; /* 0, 4, 8, 12, 16, 20, 24, 28 */ + *rx = tc << 3; /* 0, 8, 16, 24, 32, 40, 48, 56 */ break; case ixgbe_mac_82599EB: case ixgbe_mac_X540: if (num_tcs > 4) { - if (tc < 3) { - *tx = tc << 5; - *rx = tc << 4; - } else if (tc < 5) { - *tx = ((tc + 2) << 4); - *rx = tc << 4; - } else if (tc < num_tcs) { - *tx = ((tc + 8) << 3); - *rx = tc << 4; - } + /* + * TCs : TC0/1 TC2/3 TC4-7 + * TxQs/TC: 32 16 8 + * RxQs/TC: 16 16 16 + */ + *rx = tc << 4; + if (tc < 3) + *tx = tc << 5; /* 0, 32, 64 */ + else if (tc < 5) + *tx = (tc + 2) << 4; /* 80, 96 */ + else + *tx = (tc + 8) << 3; /* 104, 112, 120 */ } else { - *rx = tc << 5; - switch (tc) { - case 0: - *tx = 0; - break; - case 1: - *tx = 64; - break; - case 2: - *tx = 96; - break; - case 3: - *tx = 112; - break; - default: - break; - } + /* + * TCs : TC0 TC1 TC2/3 + * TxQs/TC: 64 32 16 + * RxQs/TC: 32 32 32 + */ + *rx = tc << 5; + if (tc < 2) + *tx = tc << 6; /* 0, 64 */ + else + *tx = (tc + 4) << 4; /* 96, 112 */ } - break; default: break; } @@ -112,106 +162,115 @@ static void ixgbe_get_first_reg_idx(struct ixgbe_adapter *adapter, u8 tc, * Cache the descriptor ring offsets for DCB to the assigned rings. * **/ -static inline bool ixgbe_cache_ring_dcb(struct ixgbe_adapter *adapter) +static bool ixgbe_cache_ring_dcb(struct ixgbe_adapter *adapter) { struct net_device *dev = adapter->netdev; - int i, j, k; + unsigned int tx_idx, rx_idx; + int tc, offset, rss_i, i; u8 num_tcs = netdev_get_num_tc(dev); - if (!num_tcs) + /* verify we have DCB queueing enabled before proceeding */ + if (num_tcs <= 1) return false; - for (i = 0, k = 0; i < num_tcs; i++) { - unsigned int tx_s, rx_s; - u16 count = dev->tc_to_txq[i].count; + rss_i = adapter->ring_feature[RING_F_RSS].indices; - ixgbe_get_first_reg_idx(adapter, i, &tx_s, &rx_s); - for (j = 0; j < count; j++, k++) { - adapter->tx_ring[k]->reg_idx = tx_s + j; - adapter->rx_ring[k]->reg_idx = rx_s + j; - adapter->tx_ring[k]->dcb_tc = i; - adapter->rx_ring[k]->dcb_tc = i; + for (tc = 0, offset = 0; tc < num_tcs; tc++, offset += rss_i) { + ixgbe_get_first_reg_idx(adapter, tc, &tx_idx, &rx_idx); + for (i = 0; i < rss_i; i++, tx_idx++, rx_idx++) { + adapter->tx_ring[offset + i]->reg_idx = tx_idx; + adapter->rx_ring[offset + i]->reg_idx = rx_idx; + adapter->tx_ring[offset + i]->dcb_tc = tc; + adapter->rx_ring[offset + i]->dcb_tc = tc; } } return true; } -#endif +#endif /** - * ixgbe_cache_ring_fdir - Descriptor ring to register mapping for Flow Director + * ixgbe_cache_ring_sriov - Descriptor ring to register mapping for sriov * @adapter: board private structure to initialize * - * Cache the descriptor ring offsets for Flow Director to the assigned rings. + * SR-IOV doesn't use any descriptor rings but changes the default if + * no other mapping is used. * - **/ -static inline bool ixgbe_cache_ring_fdir(struct ixgbe_adapter *adapter) + */ +static bool ixgbe_cache_ring_sriov(struct ixgbe_adapter *adapter) { +#ifdef IXGBE_FCOE + struct ixgbe_ring_feature *fcoe = &adapter->ring_feature[RING_F_FCOE]; +#endif /* IXGBE_FCOE */ + struct ixgbe_ring_feature *vmdq = &adapter->ring_feature[RING_F_VMDQ]; + struct ixgbe_ring_feature *rss = &adapter->ring_feature[RING_F_RSS]; int i; - bool ret = false; - - if ((adapter->flags & IXGBE_FLAG_RSS_ENABLED) && - (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE)) { - for (i = 0; i < adapter->num_rx_queues; i++) - adapter->rx_ring[i]->reg_idx = i; - for (i = 0; i < adapter->num_tx_queues; i++) - adapter->tx_ring[i]->reg_idx = i; - ret = true; + u16 reg_idx; + + /* only proceed if VMDq is enabled */ + if (!(adapter->flags & IXGBE_FLAG_VMDQ_ENABLED)) + return false; + + /* start at VMDq register offset for SR-IOV enabled setups */ + reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask); + for (i = 0; i < adapter->num_rx_queues; i++, reg_idx++) { +#ifdef IXGBE_FCOE + /* Allow first FCoE queue to be mapped as RSS */ + if (fcoe->offset && (i > fcoe->offset)) + break; +#endif + /* 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); + adapter->rx_ring[i]->reg_idx = reg_idx; } - return ret; -} +#ifdef IXGBE_FCOE + /* FCoE uses a linear block of queues so just assigning 1:1 */ + for (; i < adapter->num_rx_queues; i++, reg_idx++) + adapter->rx_ring[i]->reg_idx = reg_idx; +#endif + reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask); + for (i = 0; i < adapter->num_tx_queues; i++, reg_idx++) { #ifdef IXGBE_FCOE -/** - * ixgbe_cache_ring_fcoe - Descriptor ring to register mapping for the FCoE - * @adapter: board private structure to initialize - * - * Cache the descriptor ring offsets for FCoE mode to the assigned rings. - * - */ -static inline bool ixgbe_cache_ring_fcoe(struct ixgbe_adapter *adapter) -{ - struct ixgbe_ring_feature *f = &adapter->ring_feature[RING_F_FCOE]; - int i; - u8 fcoe_rx_i = 0, fcoe_tx_i = 0; + /* Allow first FCoE queue to be mapped as RSS */ + if (fcoe->offset && (i > fcoe->offset)) + break; +#endif + /* 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); + adapter->tx_ring[i]->reg_idx = reg_idx; + } - if (!(adapter->flags & IXGBE_FLAG_FCOE_ENABLED)) - return false; +#ifdef IXGBE_FCOE + /* FCoE uses a linear block of queues so just assigning 1:1 */ + for (; i < adapter->num_tx_queues; i++, reg_idx++) + adapter->tx_ring[i]->reg_idx = reg_idx; - if (adapter->flags & IXGBE_FLAG_RSS_ENABLED) { - if (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE) - ixgbe_cache_ring_fdir(adapter); - else - ixgbe_cache_ring_rss(adapter); +#endif - fcoe_rx_i = f->mask; - fcoe_tx_i = f->mask; - } - for (i = 0; i < f->indices; i++, fcoe_rx_i++, fcoe_tx_i++) { - adapter->rx_ring[f->mask + i]->reg_idx = fcoe_rx_i; - adapter->tx_ring[f->mask + i]->reg_idx = fcoe_tx_i; - } return true; } -#endif /* IXGBE_FCOE */ /** - * ixgbe_cache_ring_sriov - Descriptor ring to register mapping for sriov + * ixgbe_cache_ring_rss - Descriptor ring to register mapping for RSS * @adapter: board private structure to initialize * - * SR-IOV doesn't use any descriptor rings but changes the default if - * no other mapping is used. + * Cache the descriptor ring offsets for RSS to the assigned rings. * - */ -static inline bool ixgbe_cache_ring_sriov(struct ixgbe_adapter *adapter) + **/ +static bool ixgbe_cache_ring_rss(struct ixgbe_adapter *adapter) { - adapter->rx_ring[0]->reg_idx = adapter->num_vfs * 2; - adapter->tx_ring[0]->reg_idx = adapter->num_vfs * 2; - if (adapter->num_vfs) - return true; - else - return false; + int i; + + for (i = 0; i < adapter->num_rx_queues; i++) + adapter->rx_ring[i]->reg_idx = i; + for (i = 0; i < adapter->num_tx_queues; i++) + adapter->tx_ring[i]->reg_idx = i; + + return true; } /** @@ -231,186 +290,384 @@ static void ixgbe_cache_ring_register(struct ixgbe_adapter *adapter) adapter->rx_ring[0]->reg_idx = 0; adapter->tx_ring[0]->reg_idx = 0; - if (ixgbe_cache_ring_sriov(adapter)) - return; - #ifdef CONFIG_IXGBE_DCB - if (ixgbe_cache_ring_dcb(adapter)) + if (ixgbe_cache_ring_dcb_sriov(adapter)) return; -#endif -#ifdef IXGBE_FCOE - if (ixgbe_cache_ring_fcoe(adapter)) + if (ixgbe_cache_ring_dcb(adapter)) return; -#endif /* IXGBE_FCOE */ - if (ixgbe_cache_ring_fdir(adapter)) +#endif + if (ixgbe_cache_ring_sriov(adapter)) return; - if (ixgbe_cache_ring_rss(adapter)) - return; + ixgbe_cache_ring_rss(adapter); } -/** - * ixgbe_set_sriov_queues: Allocate queues for IOV use - * @adapter: board private structure to initialize - * - * IOV doesn't actually use anything, so just NAK the - * request for now and let the other queue routines - * figure out what to do. - */ -static inline bool ixgbe_set_sriov_queues(struct ixgbe_adapter *adapter) -{ - return false; -} +#define IXGBE_RSS_16Q_MASK 0xF +#define IXGBE_RSS_8Q_MASK 0x7 +#define IXGBE_RSS_4Q_MASK 0x3 +#define IXGBE_RSS_2Q_MASK 0x1 +#define IXGBE_RSS_DISABLED_MASK 0x0 +#ifdef CONFIG_IXGBE_DCB /** - * ixgbe_set_rss_queues: Allocate queues for RSS + * ixgbe_set_dcb_sriov_queues: Allocate queues for SR-IOV devices w/ DCB * @adapter: board private structure to initialize * - * This is our "base" multiqueue mode. RSS (Receive Side Scaling) will try - * to allocate one Rx queue per CPU, and if available, one Tx queue per CPU. + * When SR-IOV (Single Root IO Virtualiztion) is enabled, allocate queues + * and VM pools where appropriate. Also assign queues based on DCB + * priorities and map accordingly.. * **/ -static inline bool ixgbe_set_rss_queues(struct ixgbe_adapter *adapter) +static bool ixgbe_set_dcb_sriov_queues(struct ixgbe_adapter *adapter) { - bool ret = false; - struct ixgbe_ring_feature *f = &adapter->ring_feature[RING_F_RSS]; - - if (adapter->flags & IXGBE_FLAG_RSS_ENABLED) { - f->mask = 0xF; - adapter->num_rx_queues = f->indices; - adapter->num_tx_queues = f->indices; - ret = true; + int i; + u16 vmdq_i = adapter->ring_feature[RING_F_VMDQ].limit; + u16 vmdq_m = 0; +#ifdef IXGBE_FCOE + u16 fcoe_i = 0; +#endif + u8 tcs = netdev_get_num_tc(adapter->netdev); + + /* verify we have DCB queueing enabled before proceeding */ + if (tcs <= 1) + return false; + + /* verify we have VMDq enabled before proceeding */ + if (!(adapter->flags & IXGBE_FLAG_SRIOV_ENABLED)) + return false; + + /* Add starting offset to total pool count */ + vmdq_i += adapter->ring_feature[RING_F_VMDQ].offset; + + /* 16 pools w/ 8 TC per pool */ + if (tcs > 4) { + vmdq_i = min_t(u16, vmdq_i, 16); + vmdq_m = IXGBE_82599_VMDQ_8Q_MASK; + /* 32 pools w/ 4 TC per pool */ + } else { + vmdq_i = min_t(u16, vmdq_i, 32); + vmdq_m = IXGBE_82599_VMDQ_4Q_MASK; } - return ret; -} +#ifdef IXGBE_FCOE + /* queues in the remaining pools are available for FCoE */ + fcoe_i = (128 / __ALIGN_MASK(1, ~vmdq_m)) - vmdq_i; -/** - * ixgbe_set_fdir_queues: Allocate queues for Flow Director - * @adapter: board private structure to initialize - * - * Flow Director is an advanced Rx filter, attempting to get Rx flows back - * to the original CPU that initiated the Tx session. This runs in addition - * to RSS, so if a packet doesn't match an FDIR filter, we can still spread the - * Rx load across CPUs using RSS. - * - **/ -static inline bool ixgbe_set_fdir_queues(struct ixgbe_adapter *adapter) -{ - bool ret = false; - struct ixgbe_ring_feature *f_fdir = &adapter->ring_feature[RING_F_FDIR]; +#endif + /* remove the starting offset from the pool count */ + vmdq_i -= adapter->ring_feature[RING_F_VMDQ].offset; - f_fdir->indices = min_t(int, num_online_cpus(), f_fdir->indices); - f_fdir->mask = 0; + /* save features for later use */ + adapter->ring_feature[RING_F_VMDQ].indices = vmdq_i; + adapter->ring_feature[RING_F_VMDQ].mask = vmdq_m; /* - * Use RSS in addition to Flow Director to ensure the best - * distribution of flows across cores, even when an FDIR flow - * isn't matched. + * We do not support DCB, VMDq, and RSS all simultaneously + * so we will disable RSS since it is the lowest priority */ - if ((adapter->flags & IXGBE_FLAG_RSS_ENABLED) && - (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE)) { - adapter->num_tx_queues = f_fdir->indices; - adapter->num_rx_queues = f_fdir->indices; - ret = true; - } else { - adapter->flags &= ~IXGBE_FLAG_FDIR_HASH_CAPABLE; + adapter->ring_feature[RING_F_RSS].indices = 1; + adapter->ring_feature[RING_F_RSS].mask = IXGBE_RSS_DISABLED_MASK; + + /* disable ATR as it is not supported when VMDq is enabled */ + adapter->flags &= ~IXGBE_FLAG_FDIR_HASH_CAPABLE; + + adapter->num_rx_pools = vmdq_i; + adapter->num_rx_queues_per_pool = tcs; + + adapter->num_tx_queues = vmdq_i * tcs; + adapter->num_rx_queues = vmdq_i * tcs; + +#ifdef IXGBE_FCOE + if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED) { + struct ixgbe_ring_feature *fcoe; + + fcoe = &adapter->ring_feature[RING_F_FCOE]; + + /* limit ourselves based on feature limits */ + fcoe_i = min_t(u16, fcoe_i, num_online_cpus()); + fcoe_i = min_t(u16, fcoe_i, fcoe->limit); + + if (fcoe_i) { + /* alloc queues for FCoE separately */ + fcoe->indices = fcoe_i; + fcoe->offset = vmdq_i * tcs; + + /* add queues to adapter */ + adapter->num_tx_queues += fcoe_i; + adapter->num_rx_queues += fcoe_i; + } else if (tcs > 1) { + /* use queue belonging to FcoE TC */ + fcoe->indices = 1; + fcoe->offset = ixgbe_fcoe_get_tc(adapter); + } else { + adapter->flags &= ~IXGBE_FLAG_FCOE_ENABLED; + + fcoe->indices = 0; + fcoe->offset = 0; + } } - return ret; + +#endif /* IXGBE_FCOE */ + /* configure TC to queue mapping */ + for (i = 0; i < tcs; i++) + netdev_set_tc_queue(adapter->netdev, i, 1, i); + + return true; } +static bool ixgbe_set_dcb_queues(struct ixgbe_adapter *adapter) +{ + struct net_device *dev = adapter->netdev; + struct ixgbe_ring_feature *f; + int rss_i, rss_m, i; + int tcs; + + /* Map queue offset and counts onto allocated tx queues */ + tcs = netdev_get_num_tc(dev); + + /* verify we have DCB queueing enabled before proceeding */ + if (tcs <= 1) + return false; + + /* determine the upper limit for our current DCB mode */ + rss_i = dev->num_tx_queues / tcs; + if (adapter->hw.mac.type == ixgbe_mac_82598EB) { + /* 8 TC w/ 4 queues per TC */ + rss_i = min_t(u16, rss_i, 4); + rss_m = IXGBE_RSS_4Q_MASK; + } else if (tcs > 4) { + /* 8 TC w/ 8 queues per TC */ + rss_i = min_t(u16, rss_i, 8); + rss_m = IXGBE_RSS_8Q_MASK; + } else { + /* 4 TC w/ 16 queues per TC */ + rss_i = min_t(u16, rss_i, 16); + rss_m = IXGBE_RSS_16Q_MASK; + } + + /* set RSS mask and indices */ + f = &adapter->ring_feature[RING_F_RSS]; + rss_i = min_t(int, rss_i, f->limit); + f->indices = rss_i; + f->mask = rss_m; + + /* disable ATR as it is not supported when multiple TCs are enabled */ + adapter->flags &= ~IXGBE_FLAG_FDIR_HASH_CAPABLE; + #ifdef IXGBE_FCOE + /* FCoE enabled queues require special configuration indexed + * by feature specific indices and offset. Here we map FCoE + * indices onto the DCB queue pairs allowing FCoE to own + * configuration later. + */ + if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED) { + u8 tc = ixgbe_fcoe_get_tc(adapter); + + f = &adapter->ring_feature[RING_F_FCOE]; + f->indices = min_t(u16, rss_i, f->limit); + f->offset = rss_i * tc; + } + +#endif /* IXGBE_FCOE */ + for (i = 0; i < tcs; i++) + netdev_set_tc_queue(dev, i, rss_i, rss_i * i); + + adapter->num_tx_queues = rss_i * tcs; + adapter->num_rx_queues = rss_i * tcs; + + return true; +} + +#endif /** - * ixgbe_set_fcoe_queues: Allocate queues for Fiber Channel over Ethernet (FCoE) + * ixgbe_set_sriov_queues - Allocate queues for SR-IOV devices * @adapter: board private structure to initialize * - * FCoE RX FCRETA can use up to 8 rx queues for up to 8 different exchanges. - * The ring feature mask is not used as a mask for FCoE, as it can take any 8 - * rx queues out of the max number of rx queues, instead, it is used as the - * index of the first rx queue used by FCoE. + * When SR-IOV (Single Root IO Virtualiztion) is enabled, allocate queues + * and VM pools where appropriate. If RSS is available, then also try and + * enable RSS and map accordingly. * **/ -static inline bool ixgbe_set_fcoe_queues(struct ixgbe_adapter *adapter) +static bool ixgbe_set_sriov_queues(struct ixgbe_adapter *adapter) { - struct ixgbe_ring_feature *f = &adapter->ring_feature[RING_F_FCOE]; + u16 vmdq_i = adapter->ring_feature[RING_F_VMDQ].limit; + u16 vmdq_m = 0; + u16 rss_i = adapter->ring_feature[RING_F_RSS].limit; + u16 rss_m = IXGBE_RSS_DISABLED_MASK; +#ifdef IXGBE_FCOE + u16 fcoe_i = 0; +#endif - if (!(adapter->flags & IXGBE_FLAG_FCOE_ENABLED)) + /* only proceed if SR-IOV is enabled */ + if (!(adapter->flags & IXGBE_FLAG_SRIOV_ENABLED)) return false; - f->indices = min_t(int, num_online_cpus(), f->indices); + /* Add starting offset to total pool count */ + vmdq_i += adapter->ring_feature[RING_F_VMDQ].offset; - adapter->num_rx_queues = 1; - adapter->num_tx_queues = 1; + /* double check we are limited to maximum pools */ + vmdq_i = min_t(u16, IXGBE_MAX_VMDQ_INDICES, vmdq_i); - if (adapter->flags & IXGBE_FLAG_RSS_ENABLED) { - e_info(probe, "FCoE enabled with RSS\n"); - if (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE) - ixgbe_set_fdir_queues(adapter); - else - ixgbe_set_rss_queues(adapter); + /* 64 pool mode with 2 queues per pool */ + if ((vmdq_i > 32) || (rss_i < 4)) { + vmdq_m = IXGBE_82599_VMDQ_2Q_MASK; + rss_m = IXGBE_RSS_2Q_MASK; + rss_i = min_t(u16, rss_i, 2); + /* 32 pool mode with 4 queues per pool */ + } else { + vmdq_m = IXGBE_82599_VMDQ_4Q_MASK; + rss_m = IXGBE_RSS_4Q_MASK; + rss_i = 4; } - /* adding FCoE rx rings to the end */ - f->mask = adapter->num_rx_queues; - adapter->num_rx_queues += f->indices; - adapter->num_tx_queues += f->indices; +#ifdef IXGBE_FCOE + /* queues in the remaining pools are available for FCoE */ + fcoe_i = 128 - (vmdq_i * __ALIGN_MASK(1, ~vmdq_m)); + +#endif + /* remove the starting offset from the pool count */ + vmdq_i -= adapter->ring_feature[RING_F_VMDQ].offset; + + /* save features for later use */ + adapter->ring_feature[RING_F_VMDQ].indices = vmdq_i; + adapter->ring_feature[RING_F_VMDQ].mask = vmdq_m; + + /* limit RSS based on user input and save for later use */ + adapter->ring_feature[RING_F_RSS].indices = rss_i; + adapter->ring_feature[RING_F_RSS].mask = rss_m; + adapter->num_rx_pools = vmdq_i; + adapter->num_rx_queues_per_pool = rss_i; + + adapter->num_rx_queues = vmdq_i * rss_i; + adapter->num_tx_queues = vmdq_i * rss_i; + + /* disable ATR as it is not supported when VMDq is enabled */ + adapter->flags &= ~IXGBE_FLAG_FDIR_HASH_CAPABLE; + +#ifdef IXGBE_FCOE + /* + * FCoE can use rings from adjacent buffers to allow RSS + * like behavior. To account for this we need to add the + * FCoE indices to the total ring count. + */ + if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED) { + struct ixgbe_ring_feature *fcoe; + + fcoe = &adapter->ring_feature[RING_F_FCOE]; + + /* limit ourselves based on feature limits */ + fcoe_i = min_t(u16, fcoe_i, fcoe->limit); + + if (vmdq_i > 1 && fcoe_i) { + /* reserve no more than number of CPUs */ + fcoe_i = min_t(u16, fcoe_i, num_online_cpus()); + + /* alloc queues for FCoE separately */ + fcoe->indices = fcoe_i; + fcoe->offset = vmdq_i * rss_i; + } else { + /* merge FCoE queues with RSS queues */ + fcoe_i = min_t(u16, fcoe_i + rss_i, num_online_cpus()); + + /* limit indices to rss_i if MSI-X is disabled */ + if (!(adapter->flags & IXGBE_FLAG_MSIX_ENABLED)) + fcoe_i = rss_i; + + /* attempt to reserve some queues for just FCoE */ + fcoe->indices = min_t(u16, fcoe_i, fcoe->limit); + fcoe->offset = fcoe_i - fcoe->indices; + + fcoe_i -= rss_i; + } + + /* add queues to adapter */ + adapter->num_tx_queues += fcoe_i; + adapter->num_rx_queues += fcoe_i; + } + +#endif return true; } -#endif /* IXGBE_FCOE */ -/* Artificial max queue cap per traffic class in DCB mode */ -#define DCB_QUEUE_CAP 8 - -#ifdef CONFIG_IXGBE_DCB -static inline bool ixgbe_set_dcb_queues(struct ixgbe_adapter *adapter) +/** + * ixgbe_set_rss_queues - Allocate queues for RSS + * @adapter: board private structure to initialize + * + * This is our "base" multiqueue mode. RSS (Receive Side Scaling) will try + * to allocate one Rx queue per CPU, and if available, one Tx queue per CPU. + * + **/ +static bool ixgbe_set_rss_queues(struct ixgbe_adapter *adapter) { - int per_tc_q, q, i, offset = 0; - struct net_device *dev = adapter->netdev; - int tcs = netdev_get_num_tc(dev); + struct ixgbe_ring_feature *f; + u16 rss_i; - if (!tcs) - return false; + /* set mask for 16 queue limit of RSS */ + f = &adapter->ring_feature[RING_F_RSS]; + rss_i = f->limit; - /* Map queue offset and counts onto allocated tx queues */ - per_tc_q = min_t(unsigned int, dev->num_tx_queues / tcs, DCB_QUEUE_CAP); - q = min_t(int, num_online_cpus(), per_tc_q); + f->indices = rss_i; + f->mask = IXGBE_RSS_16Q_MASK; - for (i = 0; i < tcs; i++) { - netdev_set_tc_queue(dev, i, q, offset); - offset += q; - } + /* disable ATR by default, it will be configured below */ + adapter->flags &= ~IXGBE_FLAG_FDIR_HASH_CAPABLE; + + /* + * Use Flow Director in addition to RSS to ensure the best + * distribution of flows across cores, even when an FDIR flow + * isn't matched. + */ + if (rss_i > 1 && adapter->atr_sample_rate) { + f = &adapter->ring_feature[RING_F_FDIR]; - adapter->num_tx_queues = q * tcs; - adapter->num_rx_queues = q * tcs; + f->indices = min_t(u16, num_online_cpus(), f->limit); + rss_i = max_t(u16, rss_i, f->indices); + + if (!(adapter->flags & IXGBE_FLAG_FDIR_PERFECT_CAPABLE)) + adapter->flags |= IXGBE_FLAG_FDIR_HASH_CAPABLE; + } #ifdef IXGBE_FCOE - /* FCoE enabled queues require special configuration indexed - * by feature specific indices and mask. Here we map FCoE - * indices onto the DCB queue pairs allowing FCoE to own - * configuration later. + /* + * FCoE can exist on the same rings as standard network traffic + * however it is preferred to avoid that if possible. In order + * to get the best performance we allocate as many FCoE queues + * as we can and we place them at the end of the ring array to + * avoid sharing queues with standard RSS on systems with 24 or + * more CPUs. */ if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED) { - u8 prio_tc[MAX_USER_PRIORITY] = {0}; - int tc; - struct ixgbe_ring_feature *f = - &adapter->ring_feature[RING_F_FCOE]; - - ixgbe_dcb_unpack_map(&adapter->dcb_cfg, DCB_TX_CONFIG, prio_tc); - tc = prio_tc[adapter->fcoe.up]; - f->indices = dev->tc_to_txq[tc].count; - f->mask = dev->tc_to_txq[tc].offset; + struct net_device *dev = adapter->netdev; + u16 fcoe_i; + + f = &adapter->ring_feature[RING_F_FCOE]; + + /* merge FCoE queues with RSS queues */ + fcoe_i = min_t(u16, f->limit + rss_i, num_online_cpus()); + fcoe_i = min_t(u16, fcoe_i, dev->num_tx_queues); + + /* limit indices to rss_i if MSI-X is disabled */ + if (!(adapter->flags & IXGBE_FLAG_MSIX_ENABLED)) + fcoe_i = rss_i; + + /* attempt to reserve some queues for just FCoE */ + f->indices = min_t(u16, fcoe_i, f->limit); + f->offset = fcoe_i - f->indices; + rss_i = max_t(u16, fcoe_i, rss_i); } -#endif + +#endif /* IXGBE_FCOE */ + adapter->num_rx_queues = rss_i; + adapter->num_tx_queues = rss_i; return true; } -#endif /** - * ixgbe_set_num_queues: Allocate queues for device, feature dependent + * ixgbe_set_num_queues - Allocate queues for device, feature dependent * @adapter: board private structure to initialize * * This is the top level queue allocation routine. The order here is very @@ -420,7 +677,7 @@ static inline bool ixgbe_set_dcb_queues(struct ixgbe_adapter *adapter) * fallthrough conditions. * **/ -static int ixgbe_set_num_queues(struct ixgbe_adapter *adapter) +static void ixgbe_set_num_queues(struct ixgbe_adapter *adapter) { /* Start with base case */ adapter->num_rx_queues = 1; @@ -428,38 +685,18 @@ static int ixgbe_set_num_queues(struct ixgbe_adapter *adapter) adapter->num_rx_pools = adapter->num_rx_queues; adapter->num_rx_queues_per_pool = 1; - if (ixgbe_set_sriov_queues(adapter)) - goto done; - #ifdef CONFIG_IXGBE_DCB + if (ixgbe_set_dcb_sriov_queues(adapter)) + return; + if (ixgbe_set_dcb_queues(adapter)) - goto done; + return; #endif -#ifdef IXGBE_FCOE - if (ixgbe_set_fcoe_queues(adapter)) - goto done; - -#endif /* IXGBE_FCOE */ - if (ixgbe_set_fdir_queues(adapter)) - goto done; - - if (ixgbe_set_rss_queues(adapter)) - goto done; - - /* fallback to base case */ - adapter->num_rx_queues = 1; - adapter->num_tx_queues = 1; - -done: - if ((adapter->netdev->reg_state == NETREG_UNREGISTERED) || - (adapter->netdev->reg_state == NETREG_UNREGISTERING)) - return 0; + if (ixgbe_set_sriov_queues(adapter)) + return; - /* Notify the stack of the (possibly) reduced queue counts. */ - netif_set_real_num_tx_queues(adapter->netdev, adapter->num_tx_queues); - return netif_set_real_num_rx_queues(adapter->netdev, - adapter->num_rx_queues); + ixgbe_set_rss_queues(adapter); } static void ixgbe_acquire_msix_vectors(struct ixgbe_adapter *adapter, @@ -507,8 +744,8 @@ static void ixgbe_acquire_msix_vectors(struct ixgbe_adapter *adapter, * of max_msix_q_vectors + NON_Q_VECTORS, or the number of * vectors we were allocated. */ - adapter->num_msix_vectors = min(vectors, - adapter->max_msix_q_vectors + NON_Q_VECTORS); + vectors -= NON_Q_VECTORS; + adapter->num_q_vectors = min(vectors, adapter->max_q_vectors); } } @@ -632,8 +869,8 @@ static int ixgbe_alloc_q_vector(struct ixgbe_adapter *adapter, if (adapter->netdev->features & NETIF_F_FCOE_MTU) { struct ixgbe_ring_feature *f; f = &adapter->ring_feature[RING_F_FCOE]; - if ((rxr_idx >= f->mask) && - (rxr_idx < f->mask + f->indices)) + if ((rxr_idx >= f->offset) && + (rxr_idx < f->offset + f->indices)) set_bit(__IXGBE_RX_FCOE, &ring->state); } @@ -695,7 +932,7 @@ static void ixgbe_free_q_vector(struct ixgbe_adapter *adapter, int v_idx) **/ static int ixgbe_alloc_q_vectors(struct ixgbe_adapter *adapter) { - int q_vectors = adapter->num_msix_vectors - NON_Q_VECTORS; + int q_vectors = adapter->num_q_vectors; int rxr_remaining = adapter->num_rx_queues; int txr_remaining = adapter->num_tx_queues; int rxr_idx = 0, txr_idx = 0, v_idx = 0; @@ -739,10 +976,12 @@ static int ixgbe_alloc_q_vectors(struct ixgbe_adapter *adapter) return 0; err_out: - while (v_idx) { - v_idx--; + adapter->num_tx_queues = 0; + adapter->num_rx_queues = 0; + adapter->num_q_vectors = 0; + + while (v_idx--) ixgbe_free_q_vector(adapter, v_idx); - } return -ENOMEM; } @@ -757,14 +996,13 @@ err_out: **/ static void ixgbe_free_q_vectors(struct ixgbe_adapter *adapter) { - int v_idx, q_vectors; + int v_idx = adapter->num_q_vectors; - if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) - q_vectors = adapter->num_msix_vectors - NON_Q_VECTORS; - else - q_vectors = 1; + adapter->num_tx_queues = 0; + adapter->num_rx_queues = 0; + adapter->num_q_vectors = 0; - for (v_idx = 0; v_idx < q_vectors; v_idx++) + while (v_idx--) ixgbe_free_q_vector(adapter, v_idx); } @@ -788,11 +1026,10 @@ static void ixgbe_reset_interrupt_capability(struct ixgbe_adapter *adapter) * Attempt to configure the interrupts using the best available * capabilities of the hardware and the kernel. **/ -static int ixgbe_set_interrupt_capability(struct ixgbe_adapter *adapter) +static void ixgbe_set_interrupt_capability(struct ixgbe_adapter *adapter) { struct ixgbe_hw *hw = &adapter->hw; - int err = 0; - int vector, v_budget; + int vector, v_budget, err; /* * It's easy to be greedy for MSI-X vectors, but it really @@ -825,38 +1062,41 @@ static int ixgbe_set_interrupt_capability(struct ixgbe_adapter *adapter) ixgbe_acquire_msix_vectors(adapter, v_budget); if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) - goto out; + return; } - adapter->flags &= ~IXGBE_FLAG_DCB_ENABLED; - adapter->flags &= ~IXGBE_FLAG_RSS_ENABLED; - if (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE) { - e_err(probe, - "ATR is not supported while multiple " - "queues are disabled. Disabling Flow Director\n"); + /* disable DCB if number of TCs exceeds 1 */ + if (netdev_get_num_tc(adapter->netdev) > 1) { + e_err(probe, "num TCs exceeds number of queues - disabling DCB\n"); + netdev_reset_tc(adapter->netdev); + + if (adapter->hw.mac.type == ixgbe_mac_82598EB) + adapter->hw.fc.requested_mode = adapter->last_lfc_mode; + + adapter->flags &= ~IXGBE_FLAG_DCB_ENABLED; + adapter->temp_dcb_cfg.pfc_mode_enable = false; + adapter->dcb_cfg.pfc_mode_enable = false; } - adapter->flags &= ~IXGBE_FLAG_FDIR_HASH_CAPABLE; - adapter->atr_sample_rate = 0; - if (adapter->flags & IXGBE_FLAG_SRIOV_ENABLED) - ixgbe_disable_sriov(adapter); + adapter->dcb_cfg.num_tcs.pg_tcs = 1; + adapter->dcb_cfg.num_tcs.pfc_tcs = 1; + + /* disable SR-IOV */ + ixgbe_disable_sriov(adapter); - err = ixgbe_set_num_queues(adapter); - if (err) - return err; + /* disable RSS */ + adapter->ring_feature[RING_F_RSS].limit = 1; + + ixgbe_set_num_queues(adapter); + adapter->num_q_vectors = 1; err = pci_enable_msi(adapter->pdev); - if (!err) { - adapter->flags |= IXGBE_FLAG_MSI_ENABLED; - } else { + if (err) { netif_printk(adapter, hw, KERN_DEBUG, adapter->netdev, "Unable to allocate MSI interrupt, " "falling back to legacy. Error: %d\n", err); - /* reset err */ - err = 0; + return; } - -out: - return err; + adapter->flags |= IXGBE_FLAG_MSI_ENABLED; } /** @@ -874,15 +1114,10 @@ int ixgbe_init_interrupt_scheme(struct ixgbe_adapter *adapter) int err; /* Number of supported queues */ - err = ixgbe_set_num_queues(adapter); - if (err) - return err; + ixgbe_set_num_queues(adapter); - err = ixgbe_set_interrupt_capability(adapter); - if (err) { - e_dev_err("Unable to setup interrupt capabilities\n"); - goto err_set_interrupt; - } + /* Set interrupt mode */ + ixgbe_set_interrupt_capability(adapter); err = ixgbe_alloc_q_vectors(adapter); if (err) { @@ -902,7 +1137,6 @@ int ixgbe_init_interrupt_scheme(struct ixgbe_adapter *adapter) err_alloc_q_vectors: ixgbe_reset_interrupt_capability(adapter); -err_set_interrupt: return err; } |