Merge branch 'cxgb4-tc-offload'

Rahul Lakkireddy says:

====================
cxgb4: add support for offloading TC u32 filters

This series of patches add support to offload TC u32 filters onto
Chelsio NICs.

Patch 1 moves current common filter code to separate files
in order to provide a common api for performing packet classification
and filtering in Chelsio NICs.

Patch 2 enables filters for normal NIC configuration and implements
common api for setting and deleting filters.

Patches 3-5 add support for TC u32 offload via ndo_setup_tc.

---
v3:

Based on review and suggestion from David Miller <davem@davemloft.net>
- Fixed all local variable declarations by placing them in longest line
  first and shortest line last order.

v2:

Based on review and suggestions from Jiri Pirko <jiri@resnulli.us>:
- Replaced macros S and U with appropriate static helper functions.
- Moved completion code for set and delete filters to respective
  functions cxgb4_set_filter() and cxgb4_del_filter().  Renamed the
  original functions to __cxgb4_set_filter() and __cxgb4_del_filter()
  in case synchronization is not required.
- Dropped debugfs patch.
- Merged code for inserting and deleting u32 filters into a single
  patch.
- Reworked and fixed bugs with traversing the actions list.
- Removed all unnecessary extra ().
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/chelsio/cxgb4/Makefile

@@ -4,7 +4,7 @@
 
 obj-$(CONFIG_CHELSIO_T4) += cxgb4.o
 
-cxgb4-objs := cxgb4_main.o l2t.o t4_hw.o sge.o clip_tbl.o cxgb4_ethtool.o cxgb4_uld.o sched.o
+cxgb4-objs := cxgb4_main.o l2t.o t4_hw.o sge.o clip_tbl.o cxgb4_ethtool.o cxgb4_uld.o sched.o cxgb4_filter.o cxgb4_tc_u32.o
 cxgb4-$(CONFIG_CHELSIO_T4_DCB) +=  cxgb4_dcb.o
 cxgb4-$(CONFIG_CHELSIO_T4_FCOE) +=  cxgb4_fcoe.o
 cxgb4-$(CONFIG_DEBUG_FS) += cxgb4_debugfs.o

drivers/net/ethernet/chelsio/cxgb4/cxgb4.h

@@ -851,6 +851,9 @@ struct adapter {
 
 	spinlock_t stats_lock;
 	spinlock_t win0_lock ____cacheline_aligned_in_smp;
+
+	/* TC u32 offload */
+	struct cxgb4_tc_u32_table *tc_u32;
 };
 
 /* Support for "sched-class" command to allow a TX Scheduling Class to be
@@ -1025,6 +1028,32 @@ enum {
 	VLAN_REWRITE
 };
 
+/* Host shadow copy of ingress filter entry.  This is in host native format
+ * and doesn't match the ordering or bit order, etc. of the hardware of the
+ * firmware command.  The use of bit-field structure elements is purely to
+ * remind ourselves of the field size limitations and save memory in the case
+ * where the filter table is large.
+ */
+struct filter_entry {
+	/* Administrative fields for filter. */
+	u32 valid:1;            /* filter allocated and valid */
+	u32 locked:1;           /* filter is administratively locked */
+
+	u32 pending:1;          /* filter action is pending firmware reply */
+	u32 smtidx:8;           /* Source MAC Table index for smac */
+	struct filter_ctx *ctx; /* Caller's completion hook */
+	struct l2t_entry *l2t;  /* Layer Two Table entry for dmac */
+	struct net_device *dev; /* Associated net device */
+	u32 tid;                /* This will store the actual tid */
+
+	/* The filter itself.  Most of this is a straight copy of information
+	 * provided by the extended ioctl().  Some fields are translated to
+	 * internal forms -- for instance the Ingress Queue ID passed in from
+	 * the ioctl() is translated into the Absolute Ingress Queue ID.
+	 */
+	struct ch_filter_specification fs;
+};
+
 static inline int is_offload(const struct adapter *adap)
 {
 	return adap->params.offload;

drivers/net/ethernet/chelsio/cxgb4/cxgb4_filter.c

@@ -0,0 +1,721 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2016 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include "cxgb4.h"
+#include "t4_regs.h"
+#include "l2t.h"
+#include "t4fw_api.h"
+#include "cxgb4_filter.h"
+
+static inline bool is_field_set(u32 val, u32 mask)
+{
+	return val || mask;
+}
+
+static inline bool unsupported(u32 conf, u32 conf_mask, u32 val, u32 mask)
+{
+	return !(conf & conf_mask) && is_field_set(val, mask);
+}
+
+/* Validate filter spec against configuration done on the card. */
+static int validate_filter(struct net_device *dev,
+			   struct ch_filter_specification *fs)
+{
+	struct adapter *adapter = netdev2adap(dev);
+	u32 fconf, iconf;
+
+	/* Check for unconfigured fields being used. */
+	fconf = adapter->params.tp.vlan_pri_map;
+	iconf = adapter->params.tp.ingress_config;
+
+	if (unsupported(fconf, FCOE_F, fs->val.fcoe, fs->mask.fcoe) ||
+	    unsupported(fconf, PORT_F, fs->val.iport, fs->mask.iport) ||
+	    unsupported(fconf, TOS_F, fs->val.tos, fs->mask.tos) ||
+	    unsupported(fconf, ETHERTYPE_F, fs->val.ethtype,
+			fs->mask.ethtype) ||
+	    unsupported(fconf, MACMATCH_F, fs->val.macidx, fs->mask.macidx) ||
+	    unsupported(fconf, MPSHITTYPE_F, fs->val.matchtype,
+			fs->mask.matchtype) ||
+	    unsupported(fconf, FRAGMENTATION_F, fs->val.frag, fs->mask.frag) ||
+	    unsupported(fconf, PROTOCOL_F, fs->val.proto, fs->mask.proto) ||
+	    unsupported(fconf, VNIC_ID_F, fs->val.pfvf_vld,
+			fs->mask.pfvf_vld) ||
+	    unsupported(fconf, VNIC_ID_F, fs->val.ovlan_vld,
+			fs->mask.ovlan_vld) ||
+	    unsupported(fconf, VLAN_F, fs->val.ivlan_vld, fs->mask.ivlan_vld))
+		return -EOPNOTSUPP;
+
+	/* T4 inconveniently uses the same FT_VNIC_ID_W bits for both the Outer
+	 * VLAN Tag and PF/VF/VFvld fields based on VNIC_F being set
+	 * in TP_INGRESS_CONFIG.  Hense the somewhat crazy checks
+	 * below.  Additionally, since the T4 firmware interface also
+	 * carries that overlap, we need to translate any PF/VF
+	 * specification into that internal format below.
+	 */
+	if (is_field_set(fs->val.pfvf_vld, fs->mask.pfvf_vld) &&
+	    is_field_set(fs->val.ovlan_vld, fs->mask.ovlan_vld))
+		return -EOPNOTSUPP;
+	if (unsupported(iconf, VNIC_F, fs->val.pfvf_vld, fs->mask.pfvf_vld) ||
+	    (is_field_set(fs->val.ovlan_vld, fs->mask.ovlan_vld) &&
+	     (iconf & VNIC_F)))
+		return -EOPNOTSUPP;
+	if (fs->val.pf > 0x7 || fs->val.vf > 0x7f)
+		return -ERANGE;
+	fs->mask.pf &= 0x7;
+	fs->mask.vf &= 0x7f;
+
+	/* If the user is requesting that the filter action loop
+	 * matching packets back out one of our ports, make sure that
+	 * the egress port is in range.
+	 */
+	if (fs->action == FILTER_SWITCH &&
+	    fs->eport >= adapter->params.nports)
+		return -ERANGE;
+
+	/* Don't allow various trivially obvious bogus out-of-range values... */
+	if (fs->val.iport >= adapter->params.nports)
+		return -ERANGE;
+
+	/* T4 doesn't support removing VLAN Tags for loop back filters. */
+	if (is_t4(adapter->params.chip) &&
+	    fs->action == FILTER_SWITCH &&
+	    (fs->newvlan == VLAN_REMOVE ||
+	     fs->newvlan == VLAN_REWRITE))
+		return -EOPNOTSUPP;
+
+	return 0;
+}
+
+static unsigned int get_filter_steerq(struct net_device *dev,
+				      struct ch_filter_specification *fs)
+{
+	struct adapter *adapter = netdev2adap(dev);
+	unsigned int iq;
+
+	/* If the user has requested steering matching Ingress Packets
+	 * to a specific Queue Set, we need to make sure it's in range
+	 * for the port and map that into the Absolute Queue ID of the
+	 * Queue Set's Response Queue.
+	 */
+	if (!fs->dirsteer) {
+		if (fs->iq)
+			return -EINVAL;
+		iq = 0;
+	} else {
+		struct port_info *pi = netdev_priv(dev);
+
+		/* If the iq id is greater than the number of qsets,
+		 * then assume it is an absolute qid.
+		 */
+		if (fs->iq < pi->nqsets)
+			iq = adapter->sge.ethrxq[pi->first_qset +
+						 fs->iq].rspq.abs_id;
+		else
+			iq = fs->iq;
+	}
+
+	return iq;
+}
+
+static int cxgb4_set_ftid(struct tid_info *t, int fidx, int family)
+{
+	spin_lock_bh(&t->ftid_lock);
+
+	if (test_bit(fidx, t->ftid_bmap)) {
+		spin_unlock_bh(&t->ftid_lock);
+		return -EBUSY;
+	}
+
+	if (family == PF_INET)
+		__set_bit(fidx, t->ftid_bmap);
+	else
+		bitmap_allocate_region(t->ftid_bmap, fidx, 2);
+
+	spin_unlock_bh(&t->ftid_lock);
+	return 0;
+}
+
+static void cxgb4_clear_ftid(struct tid_info *t, int fidx, int family)
+{
+	spin_lock_bh(&t->ftid_lock);
+	if (family == PF_INET)
+		__clear_bit(fidx, t->ftid_bmap);
+	else
+		bitmap_release_region(t->ftid_bmap, fidx, 2);
+	spin_unlock_bh(&t->ftid_lock);
+}
+
+/* Delete the filter at a specified index. */
+static int del_filter_wr(struct adapter *adapter, int fidx)
+{
+	struct filter_entry *f = &adapter->tids.ftid_tab[fidx];
+	struct fw_filter_wr *fwr;
+	struct sk_buff *skb;
+	unsigned int len;
+
+	len = sizeof(*fwr);
+
+	skb = alloc_skb(len, GFP_KERNEL);
+	if (!skb)
+		return -ENOMEM;
+
+	fwr = (struct fw_filter_wr *)__skb_put(skb, len);
+	t4_mk_filtdelwr(f->tid, fwr, adapter->sge.fw_evtq.abs_id);
+
+	/* Mark the filter as "pending" and ship off the Filter Work Request.
+	 * When we get the Work Request Reply we'll clear the pending status.
+	 */
+	f->pending = 1;
+	t4_mgmt_tx(adapter, skb);
+	return 0;
+}
+
+/* Send a Work Request to write the filter at a specified index.  We construct
+ * a Firmware Filter Work Request to have the work done and put the indicated
+ * filter into "pending" mode which will prevent any further actions against
+ * it till we get a reply from the firmware on the completion status of the
+ * request.
+ */
+int set_filter_wr(struct adapter *adapter, int fidx)
+{
+	struct filter_entry *f = &adapter->tids.ftid_tab[fidx];
+	struct fw_filter_wr *fwr;
+	struct sk_buff *skb;
+
+	skb = alloc_skb(sizeof(*fwr), GFP_KERNEL);
+	if (!skb)
+		return -ENOMEM;
+
+	/* If the new filter requires loopback Destination MAC and/or VLAN
+	 * rewriting then we need to allocate a Layer 2 Table (L2T) entry for
+	 * the filter.
+	 */
+	if (f->fs.newdmac || f->fs.newvlan) {
+		/* allocate L2T entry for new filter */
+		f->l2t = t4_l2t_alloc_switching(adapter, f->fs.vlan,
+						f->fs.eport, f->fs.dmac);
+		if (!f->l2t) {
+			kfree_skb(skb);
+			return -ENOMEM;
+		}
+	}
+
+	fwr = (struct fw_filter_wr *)__skb_put(skb, sizeof(*fwr));
+	memset(fwr, 0, sizeof(*fwr));
+
+	/* It would be nice to put most of the following in t4_hw.c but most
+	 * of the work is translating the cxgbtool ch_filter_specification
+	 * into the Work Request and the definition of that structure is
+	 * currently in cxgbtool.h which isn't appropriate to pull into the
+	 * common code.  We may eventually try to come up with a more neutral
+	 * filter specification structure but for now it's easiest to simply
+	 * put this fairly direct code in line ...
+	 */
+	fwr->op_pkd = htonl(FW_WR_OP_V(FW_FILTER_WR));
+	fwr->len16_pkd = htonl(FW_WR_LEN16_V(sizeof(*fwr) / 16));
+	fwr->tid_to_iq =
+		htonl(FW_FILTER_WR_TID_V(f->tid) |
+		      FW_FILTER_WR_RQTYPE_V(f->fs.type) |
+		      FW_FILTER_WR_NOREPLY_V(0) |
+		      FW_FILTER_WR_IQ_V(f->fs.iq));
+	fwr->del_filter_to_l2tix =
+		htonl(FW_FILTER_WR_RPTTID_V(f->fs.rpttid) |
+		      FW_FILTER_WR_DROP_V(f->fs.action == FILTER_DROP) |
+		      FW_FILTER_WR_DIRSTEER_V(f->fs.dirsteer) |
+		      FW_FILTER_WR_MASKHASH_V(f->fs.maskhash) |
+		      FW_FILTER_WR_DIRSTEERHASH_V(f->fs.dirsteerhash) |
+		      FW_FILTER_WR_LPBK_V(f->fs.action == FILTER_SWITCH) |
+		      FW_FILTER_WR_DMAC_V(f->fs.newdmac) |
+		      FW_FILTER_WR_SMAC_V(f->fs.newsmac) |
+		      FW_FILTER_WR_INSVLAN_V(f->fs.newvlan == VLAN_INSERT ||
+					     f->fs.newvlan == VLAN_REWRITE) |
+		      FW_FILTER_WR_RMVLAN_V(f->fs.newvlan == VLAN_REMOVE ||
+					    f->fs.newvlan == VLAN_REWRITE) |
+		      FW_FILTER_WR_HITCNTS_V(f->fs.hitcnts) |
+		      FW_FILTER_WR_TXCHAN_V(f->fs.eport) |
+		      FW_FILTER_WR_PRIO_V(f->fs.prio) |
+		      FW_FILTER_WR_L2TIX_V(f->l2t ? f->l2t->idx : 0));
+	fwr->ethtype = htons(f->fs.val.ethtype);
+	fwr->ethtypem = htons(f->fs.mask.ethtype);
+	fwr->frag_to_ovlan_vldm =
+		(FW_FILTER_WR_FRAG_V(f->fs.val.frag) |
+		 FW_FILTER_WR_FRAGM_V(f->fs.mask.frag) |
+		 FW_FILTER_WR_IVLAN_VLD_V(f->fs.val.ivlan_vld) |
+		 FW_FILTER_WR_OVLAN_VLD_V(f->fs.val.ovlan_vld) |
+		 FW_FILTER_WR_IVLAN_VLDM_V(f->fs.mask.ivlan_vld) |
+		 FW_FILTER_WR_OVLAN_VLDM_V(f->fs.mask.ovlan_vld));
+	fwr->smac_sel = 0;
+	fwr->rx_chan_rx_rpl_iq =
+		htons(FW_FILTER_WR_RX_CHAN_V(0) |
+		      FW_FILTER_WR_RX_RPL_IQ_V(adapter->sge.fw_evtq.abs_id));
+	fwr->maci_to_matchtypem =
+		htonl(FW_FILTER_WR_MACI_V(f->fs.val.macidx) |
+		      FW_FILTER_WR_MACIM_V(f->fs.mask.macidx) |
+		      FW_FILTER_WR_FCOE_V(f->fs.val.fcoe) |
+		      FW_FILTER_WR_FCOEM_V(f->fs.mask.fcoe) |
+		      FW_FILTER_WR_PORT_V(f->fs.val.iport) |
+		      FW_FILTER_WR_PORTM_V(f->fs.mask.iport) |
+		      FW_FILTER_WR_MATCHTYPE_V(f->fs.val.matchtype) |
+		      FW_FILTER_WR_MATCHTYPEM_V(f->fs.mask.matchtype));
+	fwr->ptcl = f->fs.val.proto;
+	fwr->ptclm = f->fs.mask.proto;
+	fwr->ttyp = f->fs.val.tos;
+	fwr->ttypm = f->fs.mask.tos;
+	fwr->ivlan = htons(f->fs.val.ivlan);
+	fwr->ivlanm = htons(f->fs.mask.ivlan);
+	fwr->ovlan = htons(f->fs.val.ovlan);
+	fwr->ovlanm = htons(f->fs.mask.ovlan);
+	memcpy(fwr->lip, f->fs.val.lip, sizeof(fwr->lip));
+	memcpy(fwr->lipm, f->fs.mask.lip, sizeof(fwr->lipm));
+	memcpy(fwr->fip, f->fs.val.fip, sizeof(fwr->fip));
+	memcpy(fwr->fipm, f->fs.mask.fip, sizeof(fwr->fipm));
+	fwr->lp = htons(f->fs.val.lport);
+	fwr->lpm = htons(f->fs.mask.lport);
+	fwr->fp = htons(f->fs.val.fport);
+	fwr->fpm = htons(f->fs.mask.fport);
+	if (f->fs.newsmac)
+		memcpy(fwr->sma, f->fs.smac, sizeof(fwr->sma));
+
+	/* Mark the filter as "pending" and ship off the Filter Work Request.
+	 * When we get the Work Request Reply we'll clear the pending status.
+	 */
+	f->pending = 1;
+	set_wr_txq(skb, CPL_PRIORITY_CONTROL, f->fs.val.iport & 0x3);
+	t4_ofld_send(adapter, skb);
+	return 0;
+}
+
+/* Return an error number if the indicated filter isn't writable ... */
+int writable_filter(struct filter_entry *f)
+{
+	if (f->locked)
+		return -EPERM;
+	if (f->pending)
+		return -EBUSY;
+
+	return 0;
+}
+
+/* Delete the filter at the specified index (if valid).  The checks for all
+ * the common problems with doing this like the filter being locked, currently
+ * pending in another operation, etc.
+ */
+int delete_filter(struct adapter *adapter, unsigned int fidx)
+{
+	struct filter_entry *f;
+	int ret;
+
+	if (fidx >= adapter->tids.nftids + adapter->tids.nsftids)
+		return -EINVAL;
+
+	f = &adapter->tids.ftid_tab[fidx];
+	ret = writable_filter(f);
+	if (ret)
+		return ret;
+	if (f->valid)
+		return del_filter_wr(adapter, fidx);
+
+	return 0;
+}
+
+/* Clear a filter and release any of its resources that we own.  This also
+ * clears the filter's "pending" status.
+ */
+void clear_filter(struct adapter *adap, struct filter_entry *f)
+{
+	/* If the new or old filter have loopback rewriteing rules then we'll
+	 * need to free any existing Layer Two Table (L2T) entries of the old
+	 * filter rule.  The firmware will handle freeing up any Source MAC
+	 * Table (SMT) entries used for rewriting Source MAC Addresses in
+	 * loopback rules.
+	 */
+	if (f->l2t)
+		cxgb4_l2t_release(f->l2t);
+
+	/* The zeroing of the filter rule below clears the filter valid,
+	 * pending, locked flags, l2t pointer, etc. so it's all we need for
+	 * this operation.
+	 */
+	memset(f, 0, sizeof(*f));
+}
+
+void clear_all_filters(struct adapter *adapter)
+{
+	unsigned int i;
+
+	if (adapter->tids.ftid_tab) {
+		struct filter_entry *f = &adapter->tids.ftid_tab[0];
+		unsigned int max_ftid = adapter->tids.nftids +
+					adapter->tids.nsftids;
+
+		for (i = 0; i < max_ftid; i++, f++)
+			if (f->valid || f->pending)
+				clear_filter(adapter, f);
+	}
+}
+
+/* Fill up default masks for set match fields. */
+static void fill_default_mask(struct ch_filter_specification *fs)
+{
+	unsigned int lip = 0, lip_mask = 0;
+	unsigned int fip = 0, fip_mask = 0;
+	unsigned int i;
+
+	if (fs->val.iport && !fs->mask.iport)
+		fs->mask.iport |= ~0;
+	if (fs->val.fcoe && !fs->mask.fcoe)
+		fs->mask.fcoe |= ~0;
+	if (fs->val.matchtype && !fs->mask.matchtype)
+		fs->mask.matchtype |= ~0;
+	if (fs->val.macidx && !fs->mask.macidx)
+		fs->mask.macidx |= ~0;
+	if (fs->val.ethtype && !fs->mask.ethtype)
+		fs->mask.ethtype |= ~0;
+	if (fs->val.ivlan && !fs->mask.ivlan)
+		fs->mask.ivlan |= ~0;
+	if (fs->val.ovlan && !fs->mask.ovlan)
+		fs->mask.ovlan |= ~0;
+	if (fs->val.frag && !fs->mask.frag)
+		fs->mask.frag |= ~0;
+	if (fs->val.tos && !fs->mask.tos)
+		fs->mask.tos |= ~0;
+	if (fs->val.proto && !fs->mask.proto)
+		fs->mask.proto |= ~0;
+
+	for (i = 0; i < ARRAY_SIZE(fs->val.lip); i++) {
+		lip |= fs->val.lip[i];
+		lip_mask |= fs->mask.lip[i];
+		fip |= fs->val.fip[i];
+		fip_mask |= fs->mask.fip[i];
+	}
+
+	if (lip && !lip_mask)
+		memset(fs->mask.lip, ~0, sizeof(fs->mask.lip));
+
+	if (fip && !fip_mask)
+		memset(fs->mask.fip, ~0, sizeof(fs->mask.lip));
+
+	if (fs->val.lport && !fs->mask.lport)
+		fs->mask.lport = ~0;
+	if (fs->val.fport && !fs->mask.fport)
+		fs->mask.fport = ~0;
+}
+
+/* Check a Chelsio Filter Request for validity, convert it into our internal
+ * format and send it to the hardware.  Return 0 on success, an error number
+ * otherwise.  We attach any provided filter operation context to the internal
+ * filter specification in order to facilitate signaling completion of the
+ * operation.
+ */
+int __cxgb4_set_filter(struct net_device *dev, int filter_id,
+		       struct ch_filter_specification *fs,
+		       struct filter_ctx *ctx)
+{
+	struct adapter *adapter = netdev2adap(dev);
+	unsigned int max_fidx, fidx, iq;
+	struct filter_entry *f;
+	u32 iconf;
+	int ret;
+
+	max_fidx = adapter->tids.nftids;
+	if (filter_id != (max_fidx + adapter->tids.nsftids - 1) &&
+	    filter_id >= max_fidx)
+		return -E2BIG;
+
+	fill_default_mask(fs);
+
+	ret = validate_filter(dev, fs);
+	if (ret)
+		return ret;
+
+	iq = get_filter_steerq(dev, fs);
+	if (iq < 0)
+		return iq;
+
+	/* IPv6 filters occupy four slots and must be aligned on
+	 * four-slot boundaries.  IPv4 filters only occupy a single
+	 * slot and have no alignment requirements but writing a new
+	 * IPv4 filter into the middle of an existing IPv6 filter
+	 * requires clearing the old IPv6 filter and hence we prevent
+	 * insertion.
+	 */
+	if (fs->type == 0) { /* IPv4 */
+		/* If our IPv4 filter isn't being written to a
+		 * multiple of four filter index and there's an IPv6
+		 * filter at the multiple of 4 base slot, then we
+		 * prevent insertion.
+		 */
+		fidx = filter_id & ~0x3;
+		if (fidx != filter_id &&
+		    adapter->tids.ftid_tab[fidx].fs.type) {
+			f = &adapter->tids.ftid_tab[fidx];
+			if (f->valid) {
+				dev_err(adapter->pdev_dev,
+					"Invalid location. IPv6 requires 4 slots and is occupying slots %u to %u\n",
+					fidx, fidx + 3);
+				return -EINVAL;
+			}
+		}
+	} else { /* IPv6 */
+		/* Ensure that the IPv6 filter is aligned on a
+		 * multiple of 4 boundary.
+		 */
+		if (filter_id & 0x3) {
+			dev_err(adapter->pdev_dev,
+				"Invalid location. IPv6 must be aligned on a 4-slot boundary\n");
+			return -EINVAL;
+		}
+
+		/* Check all except the base overlapping IPv4 filter slots. */
+		for (fidx = filter_id + 1; fidx < filter_id + 4; fidx++) {
+			f = &adapter->tids.ftid_tab[fidx];
+			if (f->valid) {
+				dev_err(adapter->pdev_dev,
+					"Invalid location.  IPv6 requires 4 slots and an IPv4 filter exists at %u\n",
+					fidx);
+				return -EINVAL;
+			}
+		}
+	}
+
+	/* Check to make sure that provided filter index is not
+	 * already in use by someone else
+	 */
+	f = &adapter->tids.ftid_tab[filter_id];
+	if (f->valid)
+		return -EBUSY;
+
+	fidx = filter_id + adapter->tids.ftid_base;
+	ret = cxgb4_set_ftid(&adapter->tids, filter_id,
+			     fs->type ? PF_INET6 : PF_INET);
+	if (ret)
+		return ret;
+
+	/* Check to make sure the filter requested is writable ... */
+	ret = writable_filter(f);
+	if (ret) {
+		/* Clear the bits we have set above */
+		cxgb4_clear_ftid(&adapter->tids, filter_id,
+				 fs->type ? PF_INET6 : PF_INET);
+		return ret;
+	}
+
+	/* Clear out any old resources being used by the filter before
+	 * we start constructing the new filter.
+	 */
+	if (f->valid)
+		clear_filter(adapter, f);
+
+	/* Convert the filter specification into our internal format.
+	 * We copy the PF/VF specification into the Outer VLAN field
+	 * here so the rest of the code -- including the interface to
+	 * the firmware -- doesn't have to constantly do these checks.
+	 */
+	f->fs = *fs;
+	f->fs.iq = iq;
+	f->dev = dev;
+
+	iconf = adapter->params.tp.ingress_config;
+	if (iconf & VNIC_F) {
+		f->fs.val.ovlan = (fs->val.pf << 13) | fs->val.vf;
+		f->fs.mask.ovlan = (fs->mask.pf << 13) | fs->mask.vf;
+		f->fs.val.ovlan_vld = fs->val.pfvf_vld;
+		f->fs.mask.ovlan_vld = fs->mask.pfvf_vld;
+	}
+
+	/* Attempt to set the filter.  If we don't succeed, we clear
+	 * it and return the failure.
+	 */
+	f->ctx = ctx;
+	f->tid = fidx; /* Save the actual tid */
+	ret = set_filter_wr(adapter, filter_id);
+	if (ret) {
+		cxgb4_clear_ftid(&adapter->tids, filter_id,
+				 fs->type ? PF_INET6 : PF_INET);
+		clear_filter(adapter, f);
+	}
+
+	return ret;
+}
+
+/* Check a delete filter request for validity and send it to the hardware.
+ * Return 0 on success, an error number otherwise.  We attach any provided
+ * filter operation context to the internal filter specification in order to
+ * facilitate signaling completion of the operation.
+ */
+int __cxgb4_del_filter(struct net_device *dev, int filter_id,
+		       struct filter_ctx *ctx)
+{
+	struct adapter *adapter = netdev2adap(dev);
+	struct filter_entry *f;
+	unsigned int max_fidx;
+	int ret;
+
+	max_fidx = adapter->tids.nftids;
+	if (filter_id != (max_fidx + adapter->tids.nsftids - 1) &&
+	    filter_id >= max_fidx)
+		return -E2BIG;
+
+	f = &adapter->tids.ftid_tab[filter_id];
+	ret = writable_filter(f);
+	if (ret)
+		return ret;
+
+	if (f->valid) {
+		f->ctx = ctx;
+		cxgb4_clear_ftid(&adapter->tids, filter_id,
+				 f->fs.type ? PF_INET6 : PF_INET);
+		return del_filter_wr(adapter, filter_id);
+	}
+
+	/* If the caller has passed in a Completion Context then we need to
+	 * mark it as a successful completion so they don't stall waiting
+	 * for it.
+	 */
+	if (ctx) {
+		ctx->result = 0;
+		complete(&ctx->completion);
+	}
+	return ret;
+}
+
+int cxgb4_set_filter(struct net_device *dev, int filter_id,
+		     struct ch_filter_specification *fs)
+{
+	struct filter_ctx ctx;
+	int ret;
+
+	init_completion(&ctx.completion);
+
+	ret = __cxgb4_set_filter(dev, filter_id, fs, &ctx);
+	if (ret)
+		goto out;
+
+	/* Wait for reply */
+	ret = wait_for_completion_timeout(&ctx.completion, 10 * HZ);
+	if (!ret)
+		return -ETIMEDOUT;
+
+	ret = ctx.result;
+out:
+	return ret;
+}
+
+int cxgb4_del_filter(struct net_device *dev, int filter_id)
+{
+	struct filter_ctx ctx;
+	int ret;
+
+	init_completion(&ctx.completion);
+
+	ret = __cxgb4_del_filter(dev, filter_id, &ctx);
+	if (ret)
+		goto out;
+
+	/* Wait for reply */
+	ret = wait_for_completion_timeout(&ctx.completion, 10 * HZ);
+	if (!ret)
+		return -ETIMEDOUT;
+
+	ret = ctx.result;
+out:
+	return ret;
+}
+
+/* Handle a filter write/deletion reply. */
+void filter_rpl(struct adapter *adap, const struct cpl_set_tcb_rpl *rpl)
+{
+	unsigned int tid = GET_TID(rpl);
+	struct filter_entry *f = NULL;
+	unsigned int max_fidx;
+	int idx;
+
+	max_fidx = adap->tids.nftids + adap->tids.nsftids;
+	/* Get the corresponding filter entry for this tid */
+	if (adap->tids.ftid_tab) {
+		/* Check this in normal filter region */
+		idx = tid - adap->tids.ftid_base;
+		if (idx >= max_fidx)
+			return;
+		f = &adap->tids.ftid_tab[idx];
+		if (f->tid != tid)
+			return;
+	}
+
+	/* We found the filter entry for this tid */
+	if (f) {
+		unsigned int ret = TCB_COOKIE_G(rpl->cookie);
+		struct filter_ctx *ctx;
+
+		/* Pull off any filter operation context attached to the
+		 * filter.
+		 */
+		ctx = f->ctx;
+		f->ctx = NULL;
+
+		if (ret == FW_FILTER_WR_FLT_DELETED) {
+			/* Clear the filter when we get confirmation from the
+			 * hardware that the filter has been deleted.
+			 */
+			clear_filter(adap, f);
+			if (ctx)
+				ctx->result = 0;
+		} else if (ret == FW_FILTER_WR_SMT_TBL_FULL) {
+			dev_err(adap->pdev_dev, "filter %u setup failed due to full SMT\n",
+				idx);
+			clear_filter(adap, f);
+			if (ctx)
+				ctx->result = -ENOMEM;
+		} else if (ret == FW_FILTER_WR_FLT_ADDED) {
+			f->smtidx = (be64_to_cpu(rpl->oldval) >> 24) & 0xff;
+			f->pending = 0;  /* asynchronous setup completed */
+			f->valid = 1;
+			if (ctx) {
+				ctx->result = 0;
+				ctx->tid = idx;
+			}
+		} else {
+			/* Something went wrong.  Issue a warning about the
+			 * problem and clear everything out.
+			 */
+			dev_err(adap->pdev_dev, "filter %u setup failed with error %u\n",
+				idx, ret);
+			clear_filter(adap, f);
+			if (ctx)
+				ctx->result = -EINVAL;
+		}
+		if (ctx)
+			complete(&ctx->completion);
+	}
+}

drivers/net/ethernet/chelsio/cxgb4/cxgb4_filter.h

@@ -0,0 +1,48 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2016 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __CXGB4_FILTER_H
+#define __CXGB4_FILTER_H
+
+#include "t4_msg.h"
+
+void filter_rpl(struct adapter *adap, const struct cpl_set_tcb_rpl *rpl);
+void clear_filter(struct adapter *adap, struct filter_entry *f);
+
+int set_filter_wr(struct adapter *adapter, int fidx);
+int delete_filter(struct adapter *adapter, unsigned int fidx);
+
+int writable_filter(struct filter_entry *f);
+void clear_all_filters(struct adapter *adapter);
+#endif /* __CXGB4_FILTER_H */

drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c

@@ -67,6 +67,7 @@
 #include <linux/crash_dump.h>
 
 #include "cxgb4.h"
+#include "cxgb4_filter.h"
 #include "t4_regs.h"
 #include "t4_values.h"
 #include "t4_msg.h"
@@ -77,6 +78,7 @@
 #include "clip_tbl.h"
 #include "l2t.h"
 #include "sched.h"
+#include "cxgb4_tc_u32.h"
 
 char cxgb4_driver_name[] = KBUILD_MODNAME;
 
@@ -87,30 +89,6 @@ char cxgb4_driver_name[] = KBUILD_MODNAME;
 const char cxgb4_driver_version[] = DRV_VERSION;
 #define DRV_DESC "Chelsio T4/T5/T6 Network Driver"
 
-/* Host shadow copy of ingress filter entry.  This is in host native format
- * and doesn't match the ordering or bit order, etc. of the hardware of the
- * firmware command.  The use of bit-field structure elements is purely to
- * remind ourselves of the field size limitations and save memory in the case
- * where the filter table is large.
- */
-struct filter_entry {
-	/* Administrative fields for filter.
-	 */
-	u32 valid:1;            /* filter allocated and valid */
-	u32 locked:1;           /* filter is administratively locked */
-
-	u32 pending:1;          /* filter action is pending firmware reply */
-	u32 smtidx:8;           /* Source MAC Table index for smac */
-	struct l2t_entry *l2t;  /* Layer Two Table entry for dmac */
-
-	/* The filter itself.  Most of this is a straight copy of information
-	 * provided by the extended ioctl().  Some fields are translated to
-	 * internal forms -- for instance the Ingress Queue ID passed in from
-	 * the ioctl() is translated into the Absolute Ingress Queue ID.
-	 */
-	struct ch_filter_specification fs;
-};
-
 #define DFLT_MSG_ENABLE (NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_LINK | \
 			 NETIF_MSG_TIMER | NETIF_MSG_IFDOWN | NETIF_MSG_IFUP |\
 			 NETIF_MSG_RX_ERR | NETIF_MSG_TX_ERR)
@@ -527,66 +505,6 @@ static void dcb_rpl(struct adapter *adap, const struct fw_port_cmd *pcmd)
 }
 #endif /* CONFIG_CHELSIO_T4_DCB */
 
-/* Clear a filter and release any of its resources that we own.  This also
- * clears the filter's "pending" status.
- */
-static void clear_filter(struct adapter *adap, struct filter_entry *f)
-{
-	/* If the new or old filter have loopback rewriteing rules then we'll
-	 * need to free any existing Layer Two Table (L2T) entries of the old
-	 * filter rule.  The firmware will handle freeing up any Source MAC
-	 * Table (SMT) entries used for rewriting Source MAC Addresses in
-	 * loopback rules.
-	 */
-	if (f->l2t)
-		cxgb4_l2t_release(f->l2t);
-
-	/* The zeroing of the filter rule below clears the filter valid,
-	 * pending, locked flags, l2t pointer, etc. so it's all we need for
-	 * this operation.
-	 */
-	memset(f, 0, sizeof(*f));
-}
-
-/* Handle a filter write/deletion reply.
- */
-static void filter_rpl(struct adapter *adap, const struct cpl_set_tcb_rpl *rpl)
-{
-	unsigned int idx = GET_TID(rpl);
-	unsigned int nidx = idx - adap->tids.ftid_base;
-	unsigned int ret;
-	struct filter_entry *f;
-
-	if (idx >= adap->tids.ftid_base && nidx <
-	   (adap->tids.nftids + adap->tids.nsftids)) {
-		idx = nidx;
-		ret = TCB_COOKIE_G(rpl->cookie);
-		f = &adap->tids.ftid_tab[idx];
-
-		if (ret == FW_FILTER_WR_FLT_DELETED) {
-			/* Clear the filter when we get confirmation from the
-			 * hardware that the filter has been deleted.
-			 */
-			clear_filter(adap, f);
-		} else if (ret == FW_FILTER_WR_SMT_TBL_FULL) {
-			dev_err(adap->pdev_dev, "filter %u setup failed due to full SMT\n",
-				idx);
-			clear_filter(adap, f);
-		} else if (ret == FW_FILTER_WR_FLT_ADDED) {
-			f->smtidx = (be64_to_cpu(rpl->oldval) >> 24) & 0xff;
-			f->pending = 0;  /* asynchronous setup completed */
-			f->valid = 1;
-		} else {
-			/* Something went wrong.  Issue a warning about the
-			 * problem and clear everything out.
-			 */
-			dev_err(adap->pdev_dev, "filter %u setup failed with error %u\n",
-				idx, ret);
-			clear_filter(adap, f);
-		}
-	}
-}
-
 /* Response queue handler for the FW event queue.
  */
 static int fwevtq_handler(struct sge_rspq *q, const __be64 *rsp,
@@ -1026,151 +944,6 @@ void t4_free_mem(void *addr)
 	kvfree(addr);
 }
 
-/* Send a Work Request to write the filter at a specified index.  We construct
- * a Firmware Filter Work Request to have the work done and put the indicated
- * filter into "pending" mode which will prevent any further actions against
- * it till we get a reply from the firmware on the completion status of the
- * request.
- */
-static int set_filter_wr(struct adapter *adapter, int fidx)
-{
-	struct filter_entry *f = &adapter->tids.ftid_tab[fidx];
-	struct sk_buff *skb;
-	struct fw_filter_wr *fwr;
-	unsigned int ftid;
-
-	skb = alloc_skb(sizeof(*fwr), GFP_KERNEL);
-	if (!skb)
-		return -ENOMEM;
-
-	/* If the new filter requires loopback Destination MAC and/or VLAN
-	 * rewriting then we need to allocate a Layer 2 Table (L2T) entry for
-	 * the filter.
-	 */
-	if (f->fs.newdmac || f->fs.newvlan) {
-		/* allocate L2T entry for new filter */
-		f->l2t = t4_l2t_alloc_switching(adapter, f->fs.vlan,
-						f->fs.eport, f->fs.dmac);
-		if (f->l2t == NULL) {
-			kfree_skb(skb);
-			return -ENOMEM;
-		}
-	}
-
-	ftid = adapter->tids.ftid_base + fidx;
-
-	fwr = (struct fw_filter_wr *)__skb_put(skb, sizeof(*fwr));
-	memset(fwr, 0, sizeof(*fwr));
-
-	/* It would be nice to put most of the following in t4_hw.c but most
-	 * of the work is translating the cxgbtool ch_filter_specification
-	 * into the Work Request and the definition of that structure is
-	 * currently in cxgbtool.h which isn't appropriate to pull into the
-	 * common code.  We may eventually try to come up with a more neutral
-	 * filter specification structure but for now it's easiest to simply
-	 * put this fairly direct code in line ...
-	 */
-	fwr->op_pkd = htonl(FW_WR_OP_V(FW_FILTER_WR));
-	fwr->len16_pkd = htonl(FW_WR_LEN16_V(sizeof(*fwr)/16));
-	fwr->tid_to_iq =
-		htonl(FW_FILTER_WR_TID_V(ftid) |
-		      FW_FILTER_WR_RQTYPE_V(f->fs.type) |
-		      FW_FILTER_WR_NOREPLY_V(0) |
-		      FW_FILTER_WR_IQ_V(f->fs.iq));
-	fwr->del_filter_to_l2tix =
-		htonl(FW_FILTER_WR_RPTTID_V(f->fs.rpttid) |
-		      FW_FILTER_WR_DROP_V(f->fs.action == FILTER_DROP) |
-		      FW_FILTER_WR_DIRSTEER_V(f->fs.dirsteer) |
-		      FW_FILTER_WR_MASKHASH_V(f->fs.maskhash) |
-		      FW_FILTER_WR_DIRSTEERHASH_V(f->fs.dirsteerhash) |
-		      FW_FILTER_WR_LPBK_V(f->fs.action == FILTER_SWITCH) |
-		      FW_FILTER_WR_DMAC_V(f->fs.newdmac) |
-		      FW_FILTER_WR_SMAC_V(f->fs.newsmac) |
-		      FW_FILTER_WR_INSVLAN_V(f->fs.newvlan == VLAN_INSERT ||
-					     f->fs.newvlan == VLAN_REWRITE) |
-		      FW_FILTER_WR_RMVLAN_V(f->fs.newvlan == VLAN_REMOVE ||
-					    f->fs.newvlan == VLAN_REWRITE) |
-		      FW_FILTER_WR_HITCNTS_V(f->fs.hitcnts) |
-		      FW_FILTER_WR_TXCHAN_V(f->fs.eport) |
-		      FW_FILTER_WR_PRIO_V(f->fs.prio) |
-		      FW_FILTER_WR_L2TIX_V(f->l2t ? f->l2t->idx : 0));
-	fwr->ethtype = htons(f->fs.val.ethtype);
-	fwr->ethtypem = htons(f->fs.mask.ethtype);
-	fwr->frag_to_ovlan_vldm =
-		(FW_FILTER_WR_FRAG_V(f->fs.val.frag) |
-		 FW_FILTER_WR_FRAGM_V(f->fs.mask.frag) |
-		 FW_FILTER_WR_IVLAN_VLD_V(f->fs.val.ivlan_vld) |
-		 FW_FILTER_WR_OVLAN_VLD_V(f->fs.val.ovlan_vld) |
-		 FW_FILTER_WR_IVLAN_VLDM_V(f->fs.mask.ivlan_vld) |
-		 FW_FILTER_WR_OVLAN_VLDM_V(f->fs.mask.ovlan_vld));
-	fwr->smac_sel = 0;
-	fwr->rx_chan_rx_rpl_iq =
-		htons(FW_FILTER_WR_RX_CHAN_V(0) |
-		      FW_FILTER_WR_RX_RPL_IQ_V(adapter->sge.fw_evtq.abs_id));
-	fwr->maci_to_matchtypem =
-		htonl(FW_FILTER_WR_MACI_V(f->fs.val.macidx) |
-		      FW_FILTER_WR_MACIM_V(f->fs.mask.macidx) |
-		      FW_FILTER_WR_FCOE_V(f->fs.val.fcoe) |
-		      FW_FILTER_WR_FCOEM_V(f->fs.mask.fcoe) |
-		      FW_FILTER_WR_PORT_V(f->fs.val.iport) |
-		      FW_FILTER_WR_PORTM_V(f->fs.mask.iport) |
-		      FW_FILTER_WR_MATCHTYPE_V(f->fs.val.matchtype) |
-		      FW_FILTER_WR_MATCHTYPEM_V(f->fs.mask.matchtype));
-	fwr->ptcl = f->fs.val.proto;
-	fwr->ptclm = f->fs.mask.proto;
-	fwr->ttyp = f->fs.val.tos;
-	fwr->ttypm = f->fs.mask.tos;
-	fwr->ivlan = htons(f->fs.val.ivlan);
-	fwr->ivlanm = htons(f->fs.mask.ivlan);
-	fwr->ovlan = htons(f->fs.val.ovlan);
-	fwr->ovlanm = htons(f->fs.mask.ovlan);
-	memcpy(fwr->lip, f->fs.val.lip, sizeof(fwr->lip));
-	memcpy(fwr->lipm, f->fs.mask.lip, sizeof(fwr->lipm));
-	memcpy(fwr->fip, f->fs.val.fip, sizeof(fwr->fip));
-	memcpy(fwr->fipm, f->fs.mask.fip, sizeof(fwr->fipm));
-	fwr->lp = htons(f->fs.val.lport);
-	fwr->lpm = htons(f->fs.mask.lport);
-	fwr->fp = htons(f->fs.val.fport);
-	fwr->fpm = htons(f->fs.mask.fport);
-	if (f->fs.newsmac)
-		memcpy(fwr->sma, f->fs.smac, sizeof(fwr->sma));
-
-	/* Mark the filter as "pending" and ship off the Filter Work Request.
-	 * When we get the Work Request Reply we'll clear the pending status.
-	 */
-	f->pending = 1;
-	set_wr_txq(skb, CPL_PRIORITY_CONTROL, f->fs.val.iport & 0x3);
-	t4_ofld_send(adapter, skb);
-	return 0;
-}
-
-/* Delete the filter at a specified index.
- */
-static int del_filter_wr(struct adapter *adapter, int fidx)
-{
-	struct filter_entry *f = &adapter->tids.ftid_tab[fidx];
-	struct sk_buff *skb;
-	struct fw_filter_wr *fwr;
-	unsigned int len, ftid;
-
-	len = sizeof(*fwr);
-	ftid = adapter->tids.ftid_base + fidx;
-
-	skb = alloc_skb(len, GFP_KERNEL);
-	if (!skb)
-		return -ENOMEM;
-
-	fwr = (struct fw_filter_wr *)__skb_put(skb, len);
-	t4_mk_filtdelwr(ftid, fwr, adapter->sge.fw_evtq.abs_id);
-
-	/* Mark the filter as "pending" and ship off the Filter Work Request.
-	 * When we get the Work Request Reply we'll clear the pending status.
-	 */
-	f->pending = 1;
-	t4_mgmt_tx(adapter, skb);
-	return 0;
-}
-
 static u16 cxgb_select_queue(struct net_device *dev, struct sk_buff *skb,
 			     void *accel_priv, select_queue_fallback_t fallback)
 {
@@ -1552,19 +1325,22 @@ EXPORT_SYMBOL(cxgb4_remove_tid);
  */
 static int tid_init(struct tid_info *t)
 {
-	size_t size;
-	unsigned int stid_bmap_size;
-	unsigned int natids = t->natids;
 	struct adapter *adap = container_of(t, struct adapter, tids);
+	unsigned int max_ftids = t->nftids + t->nsftids;
+	unsigned int natids = t->natids;
+	unsigned int stid_bmap_size;
+	unsigned int ftid_bmap_size;
+	size_t size;
 
 	stid_bmap_size = BITS_TO_LONGS(t->nstids + t->nsftids);
+	ftid_bmap_size = BITS_TO_LONGS(t->nftids);
 	size = t->ntids * sizeof(*t->tid_tab) +
 	       natids * sizeof(*t->atid_tab) +
 	       t->nstids * sizeof(*t->stid_tab) +
 	       t->nsftids * sizeof(*t->stid_tab) +
 	       stid_bmap_size * sizeof(long) +
-	       t->nftids * sizeof(*t->ftid_tab) +
-	       t->nsftids * sizeof(*t->ftid_tab);
+	       max_ftids * sizeof(*t->ftid_tab) +
+	       ftid_bmap_size * sizeof(long);
 
 	t->tid_tab = t4_alloc_mem(size);
 	if (!t->tid_tab)
@@ -1574,8 +1350,10 @@ static int tid_init(struct tid_info *t)
 	t->stid_tab = (struct serv_entry *)&t->atid_tab[natids];
 	t->stid_bmap = (unsigned long *)&t->stid_tab[t->nstids + t->nsftids];
 	t->ftid_tab = (struct filter_entry *)&t->stid_bmap[stid_bmap_size];
+	t->ftid_bmap = (unsigned long *)&t->ftid_tab[max_ftids];
 	spin_lock_init(&t->stid_lock);
 	spin_lock_init(&t->atid_lock);
+	spin_lock_init(&t->ftid_lock);
 
 	t->stids_in_use = 0;
 	t->sftids_in_use = 0;
@@ -1590,12 +1368,16 @@ static int tid_init(struct tid_info *t)
 			t->atid_tab[natids - 1].next = &t->atid_tab[natids];
 		t->afree = t->atid_tab;
 	}
-	bitmap_zero(t->stid_bmap, t->nstids + t->nsftids);
-	/* Reserve stid 0 for T4/T5 adapters */
-	if (!t->stid_base &&
-	    (CHELSIO_CHIP_VERSION(adap->params.chip) <= CHELSIO_T5))
-		__set_bit(0, t->stid_bmap);
 
+	if (is_offload(adap)) {
+		bitmap_zero(t->stid_bmap, t->nstids + t->nsftids);
+		/* Reserve stid 0 for T4/T5 adapters */
+		if (!t->stid_base &&
+		    CHELSIO_CHIP_VERSION(adap->params.chip) <= CHELSIO_T5)
+			__set_bit(0, t->stid_bmap);
+	}
+
+	bitmap_zero(t->ftid_bmap, t->nftids);
 	return 0;
 }
 
@@ -2514,40 +2296,6 @@ static int cxgb_close(struct net_device *dev)
 	return t4_enable_vi(adapter, adapter->pf, pi->viid, false, false);
 }
 
-/* Return an error number if the indicated filter isn't writable ...
- */
-static int writable_filter(struct filter_entry *f)
-{
-	if (f->locked)
-		return -EPERM;
-	if (f->pending)
-		return -EBUSY;
-
-	return 0;
-}
-
-/* Delete the filter at the specified index (if valid).  The checks for all
- * the common problems with doing this like the filter being locked, currently
- * pending in another operation, etc.
- */
-static int delete_filter(struct adapter *adapter, unsigned int fidx)
-{
-	struct filter_entry *f;
-	int ret;
-
-	if (fidx >= adapter->tids.nftids + adapter->tids.nsftids)
-		return -EINVAL;
-
-	f = &adapter->tids.ftid_tab[fidx];
-	ret = writable_filter(f);
-	if (ret)
-		return ret;
-	if (f->valid)
-		return del_filter_wr(adapter, fidx);
-
-	return 0;
-}
-
 int cxgb4_create_server_filter(const struct net_device *dev, unsigned int stid,
 		__be32 sip, __be16 sport, __be16 vlan,
 		unsigned int queue, unsigned char port, unsigned char mask)
@@ -2964,6 +2712,35 @@ static int cxgb_set_tx_maxrate(struct net_device *dev, int index, u32 rate)
 	return err;
 }
 
+int cxgb_setup_tc(struct net_device *dev, u32 handle, __be16 proto,
+		  struct tc_to_netdev *tc)
+{
+	struct port_info *pi = netdev2pinfo(dev);
+	struct adapter *adap = netdev2adap(dev);
+
+	if (!(adap->flags & FULL_INIT_DONE)) {
+		dev_err(adap->pdev_dev,
+			"Failed to setup tc on port %d. Link Down?\n",
+			pi->port_id);
+		return -EINVAL;
+	}
+
+	if (TC_H_MAJ(handle) == TC_H_MAJ(TC_H_INGRESS) &&
+	    tc->type == TC_SETUP_CLSU32) {
+		switch (tc->cls_u32->command) {
+		case TC_CLSU32_NEW_KNODE:
+		case TC_CLSU32_REPLACE_KNODE:
+			return cxgb4_config_knode(dev, proto, tc->cls_u32);
+		case TC_CLSU32_DELETE_KNODE:
+			return cxgb4_delete_knode(dev, proto, tc->cls_u32);
+		default:
+			return -EOPNOTSUPP;
+		}
+	}
+
+	return -EOPNOTSUPP;
+}
+
 static const struct net_device_ops cxgb4_netdev_ops = {
 	.ndo_open             = cxgb_open,
 	.ndo_stop             = cxgb_close,
@@ -2987,6 +2764,7 @@ static const struct net_device_ops cxgb4_netdev_ops = {
 	.ndo_busy_poll        = cxgb_busy_poll,
 #endif
 	.ndo_set_tx_maxrate   = cxgb_set_tx_maxrate,
+	.ndo_setup_tc         = cxgb_setup_tc,
 };
 
 #ifdef CONFIG_PCI_IOV
@@ -4659,6 +4437,7 @@ static void free_some_resources(struct adapter *adapter)
 	t4_free_mem(adapter->l2t);
 	t4_cleanup_sched(adapter);
 	t4_free_mem(adapter->tids.tid_tab);
+	cxgb4_cleanup_tc_u32(adapter);
 	kfree(adapter->sge.egr_map);
 	kfree(adapter->sge.ingr_map);
 	kfree(adapter->sge.starving_fl);
@@ -5003,7 +4782,8 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 		netdev->hw_features = NETIF_F_SG | TSO_FLAGS |
 			NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM |
 			NETIF_F_RXCSUM | NETIF_F_RXHASH |
-			NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX;
+			NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX |
+			NETIF_F_HW_TC;
 		if (highdma)
 			netdev->hw_features |= NETIF_F_HIGHDMA;
 		netdev->features |= netdev->hw_features;
@@ -5087,10 +4867,16 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 				 i);
 	}
 
-	if (is_offload(adapter) && tid_init(&adapter->tids) < 0) {
+	if (tid_init(&adapter->tids) < 0) {
 		dev_warn(&pdev->dev, "could not allocate TID table, "
 			 "continuing\n");
 		adapter->params.offload = 0;
+	} else {
+		adapter->tc_u32 = cxgb4_init_tc_u32(adapter,
+						    CXGB4_MAX_LINK_HANDLE);
+		if (!adapter->tc_u32)
+			dev_warn(&pdev->dev,
+				 "could not offload tc u32, continuing\n");
 	}
 
 	if (is_offload(adapter)) {
@@ -5274,13 +5060,7 @@ static void remove_one(struct pci_dev *pdev)
 		/* If we allocated filters, free up state associated with any
 		 * valid filters ...
 		 */
-		if (adapter->tids.ftid_tab) {
-			struct filter_entry *f = &adapter->tids.ftid_tab[0];
-			for (i = 0; i < (adapter->tids.nftids +
-					adapter->tids.nsftids); i++, f++)
-				if (f->valid)
-					clear_filter(adapter, f);
-		}
+		clear_all_filters(adapter);
 
 		if (adapter->flags & FULL_INIT_DONE)
 			cxgb_down(adapter);

drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_u32.c

@@ -0,0 +1,483 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2016 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <net/tc_act/tc_gact.h>
+#include <net/tc_act/tc_mirred.h>
+
+#include "cxgb4.h"
+#include "cxgb4_tc_u32_parse.h"
+#include "cxgb4_tc_u32.h"
+
+/* Fill ch_filter_specification with parsed match value/mask pair. */
+static int fill_match_fields(struct adapter *adap,
+			     struct ch_filter_specification *fs,
+			     struct tc_cls_u32_offload *cls,
+			     const struct cxgb4_match_field *entry,
+			     bool next_header)
+{
+	unsigned int i, j;
+	u32 val, mask;
+	int off, err;
+	bool found;
+
+	for (i = 0; i < cls->knode.sel->nkeys; i++) {
+		off = cls->knode.sel->keys[i].off;
+		val = cls->knode.sel->keys[i].val;
+		mask = cls->knode.sel->keys[i].mask;
+
+		if (next_header) {
+			/* For next headers, parse only keys with offmask */
+			if (!cls->knode.sel->keys[i].offmask)
+				continue;
+		} else {
+			/* For the remaining, parse only keys without offmask */
+			if (cls->knode.sel->keys[i].offmask)
+				continue;
+		}
+
+		found = false;
+
+		for (j = 0; entry[j].val; j++) {
+			if (off == entry[j].off) {
+				found = true;
+				err = entry[j].val(fs, val, mask);
+				if (err)
+					return err;
+				break;
+			}
+		}
+
+		if (!found)
+			return -EINVAL;
+	}
+
+	return 0;
+}
+
+/* Fill ch_filter_specification with parsed action. */
+static int fill_action_fields(struct adapter *adap,
+			      struct ch_filter_specification *fs,
+			      struct tc_cls_u32_offload *cls)
+{
+	unsigned int num_actions = 0;
+	const struct tc_action *a;
+	struct tcf_exts *exts;
+	LIST_HEAD(actions);
+
+	exts = cls->knode.exts;
+	if (tc_no_actions(exts))
+		return -EINVAL;
+
+	tcf_exts_to_list(exts, &actions);
+	list_for_each_entry(a, &actions, list) {
+		/* Don't allow more than one action per rule. */
+		if (num_actions)
+			return -EINVAL;
+
+		/* Drop in hardware. */
+		if (is_tcf_gact_shot(a)) {
+			fs->action = FILTER_DROP;
+			num_actions++;
+			continue;
+		}
+
+		/* Re-direct to specified port in hardware. */
+		if (is_tcf_mirred_redirect(a)) {
+			struct net_device *n_dev;
+			unsigned int i, index;
+			bool found = false;
+
+			index = tcf_mirred_ifindex(a);
+			for_each_port(adap, i) {
+				n_dev = adap->port[i];
+				if (index == n_dev->ifindex) {
+					fs->action = FILTER_SWITCH;
+					fs->eport = i;
+					found = true;
+					break;
+				}
+			}
+
+			/* Interface doesn't belong to any port of
+			 * the underlying hardware.
+			 */
+			if (!found)
+				return -EINVAL;
+
+			num_actions++;
+			continue;
+		}
+
+		/* Un-supported action. */
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int cxgb4_config_knode(struct net_device *dev, __be16 protocol,
+		       struct tc_cls_u32_offload *cls)
+{
+	const struct cxgb4_match_field *start, *link_start = NULL;
+	struct adapter *adapter = netdev2adap(dev);
+	struct ch_filter_specification fs;
+	struct cxgb4_tc_u32_table *t;
+	struct cxgb4_link *link;
+	unsigned int filter_id;
+	u32 uhtid, link_uhtid;
+	bool is_ipv6 = false;
+	int ret;
+
+	if (!can_tc_u32_offload(dev))
+		return -EOPNOTSUPP;
+
+	if (protocol != htons(ETH_P_IP) && protocol != htons(ETH_P_IPV6))
+		return -EOPNOTSUPP;
+
+	/* Fetch the location to insert the filter. */
+	filter_id = cls->knode.handle & 0xFFFFF;
+
+	if (filter_id > adapter->tids.nftids) {
+		dev_err(adapter->pdev_dev,
+			"Location %d out of range for insertion. Max: %d\n",
+			filter_id, adapter->tids.nftids);
+		return -ERANGE;
+	}
+
+	t = adapter->tc_u32;
+	uhtid = TC_U32_USERHTID(cls->knode.handle);
+	link_uhtid = TC_U32_USERHTID(cls->knode.link_handle);
+
+	/* Ensure that uhtid is either root u32 (i.e. 0x800)
+	 * or a a valid linked bucket.
+	 */
+	if (uhtid != 0x800 && uhtid >= t->size)
+		return -EINVAL;
+
+	/* Ensure link handle uhtid is sane, if specified. */
+	if (link_uhtid >= t->size)
+		return -EINVAL;
+
+	memset(&fs, 0, sizeof(fs));
+
+	if (protocol == htons(ETH_P_IPV6)) {
+		start = cxgb4_ipv6_fields;
+		is_ipv6 = true;
+	} else {
+		start = cxgb4_ipv4_fields;
+		is_ipv6 = false;
+	}
+
+	if (uhtid != 0x800) {
+		/* Link must exist from root node before insertion. */
+		if (!t->table[uhtid - 1].link_handle)
+			return -EINVAL;
+
+		/* Link must have a valid supported next header. */
+		link_start = t->table[uhtid - 1].match_field;
+		if (!link_start)
+			return -EINVAL;
+	}
+
+	/* Parse links and record them for subsequent jumps to valid
+	 * next headers.
+	 */
+	if (link_uhtid) {
+		const struct cxgb4_next_header *next;
+		bool found = false;
+		unsigned int i, j;
+		u32 val, mask;
+		int off;
+
+		if (t->table[link_uhtid - 1].link_handle) {
+			dev_err(adapter->pdev_dev,
+				"Link handle exists for: 0x%x\n",
+				link_uhtid);
+			return -EINVAL;
+		}
+
+		next = is_ipv6 ? cxgb4_ipv6_jumps : cxgb4_ipv4_jumps;
+
+		/* Try to find matches that allow jumps to next header. */
+		for (i = 0; next[i].jump; i++) {
+			if (next[i].offoff != cls->knode.sel->offoff ||
+			    next[i].shift != cls->knode.sel->offshift ||
+			    next[i].mask != cls->knode.sel->offmask ||
+			    next[i].offset != cls->knode.sel->off)
+				continue;
+
+			/* Found a possible candidate.  Find a key that
+			 * matches the corresponding offset, value, and
+			 * mask to jump to next header.
+			 */
+			for (j = 0; j < cls->knode.sel->nkeys; j++) {
+				off = cls->knode.sel->keys[j].off;
+				val = cls->knode.sel->keys[j].val;
+				mask = cls->knode.sel->keys[j].mask;
+
+				if (next[i].match_off == off &&
+				    next[i].match_val == val &&
+				    next[i].match_mask == mask) {
+					found = true;
+					break;
+				}
+			}
+
+			if (!found)
+				continue; /* Try next candidate. */
+
+			/* Candidate to jump to next header found.
+			 * Translate all keys to internal specification
+			 * and store them in jump table. This spec is copied
+			 * later to set the actual filters.
+			 */
+			ret = fill_match_fields(adapter, &fs, cls,
+						start, false);
+			if (ret)
+				goto out;
+
+			link = &t->table[link_uhtid - 1];
+			link->match_field = next[i].jump;
+			link->link_handle = cls->knode.handle;
+			memcpy(&link->fs, &fs, sizeof(fs));
+			break;
+		}
+
+		/* No candidate found to jump to next header. */
+		if (!found)
+			return -EINVAL;
+
+		return 0;
+	}
+
+	/* Fill ch_filter_specification match fields to be shipped to hardware.
+	 * Copy the linked spec (if any) first.  And then update the spec as
+	 * needed.
+	 */
+	if (uhtid != 0x800 && t->table[uhtid - 1].link_handle) {
+		/* Copy linked ch_filter_specification */
+		memcpy(&fs, &t->table[uhtid - 1].fs, sizeof(fs));
+		ret = fill_match_fields(adapter, &fs, cls,
+					link_start, true);
+		if (ret)
+			goto out;
+	}
+
+	ret = fill_match_fields(adapter, &fs, cls, start, false);
+	if (ret)
+		goto out;
+
+	/* Fill ch_filter_specification action fields to be shipped to
+	 * hardware.
+	 */
+	ret = fill_action_fields(adapter, &fs, cls);
+	if (ret)
+		goto out;
+
+	/* The filter spec has been completely built from the info
+	 * provided from u32.  We now set some default fields in the
+	 * spec for sanity.
+	 */
+
+	/* Match only packets coming from the ingress port where this
+	 * filter will be created.
+	 */
+	fs.val.iport = netdev2pinfo(dev)->port_id;
+	fs.mask.iport = ~0;
+
+	/* Enable filter hit counts. */
+	fs.hitcnts = 1;
+
+	/* Set type of filter - IPv6 or IPv4 */
+	fs.type = is_ipv6 ? 1 : 0;
+
+	/* Set the filter */
+	ret = cxgb4_set_filter(dev, filter_id, &fs);
+	if (ret)
+		goto out;
+
+	/* If this is a linked bucket, then set the corresponding
+	 * entry in the bitmap to mark it as belonging to this linked
+	 * bucket.
+	 */
+	if (uhtid != 0x800 && t->table[uhtid - 1].link_handle)
+		set_bit(filter_id, t->table[uhtid - 1].tid_map);
+
+out:
+	return ret;
+}
+
+int cxgb4_delete_knode(struct net_device *dev, __be16 protocol,
+		       struct tc_cls_u32_offload *cls)
+{
+	struct adapter *adapter = netdev2adap(dev);
+	unsigned int filter_id, max_tids, i, j;
+	struct cxgb4_link *link = NULL;
+	struct cxgb4_tc_u32_table *t;
+	u32 handle, uhtid;
+	int ret;
+
+	if (!can_tc_u32_offload(dev))
+		return -EOPNOTSUPP;
+
+	/* Fetch the location to delete the filter. */
+	filter_id = cls->knode.handle & 0xFFFFF;
+
+	if (filter_id > adapter->tids.nftids) {
+		dev_err(adapter->pdev_dev,
+			"Location %d out of range for deletion. Max: %d\n",
+			filter_id, adapter->tids.nftids);
+		return -ERANGE;
+	}
+
+	t = adapter->tc_u32;
+	handle = cls->knode.handle;
+	uhtid = TC_U32_USERHTID(cls->knode.handle);
+
+	/* Ensure that uhtid is either root u32 (i.e. 0x800)
+	 * or a a valid linked bucket.
+	 */
+	if (uhtid != 0x800 && uhtid >= t->size)
+		return -EINVAL;
+
+	/* Delete the specified filter */
+	if (uhtid != 0x800) {
+		link = &t->table[uhtid - 1];
+		if (!link->link_handle)
+			return -EINVAL;
+
+		if (!test_bit(filter_id, link->tid_map))
+			return -EINVAL;
+	}
+
+	ret = cxgb4_del_filter(dev, filter_id);
+	if (ret)
+		goto out;
+
+	if (link)
+		clear_bit(filter_id, link->tid_map);
+
+	/* If a link is being deleted, then delete all filters
+	 * associated with the link.
+	 */
+	max_tids = adapter->tids.nftids;
+	for (i = 0; i < t->size; i++) {
+		link = &t->table[i];
+
+		if (link->link_handle == handle) {
+			for (j = 0; j < max_tids; j++) {
+				if (!test_bit(j, link->tid_map))
+					continue;
+
+				ret = __cxgb4_del_filter(dev, j, NULL);
+				if (ret)
+					goto out;
+
+				clear_bit(j, link->tid_map);
+			}
+
+			/* Clear the link state */
+			link->match_field = NULL;
+			link->link_handle = 0;
+			memset(&link->fs, 0, sizeof(link->fs));
+			break;
+		}
+	}
+
+out:
+	return ret;
+}
+
+void cxgb4_cleanup_tc_u32(struct adapter *adap)
+{
+	struct cxgb4_tc_u32_table *t;
+	unsigned int i;
+
+	if (!adap->tc_u32)
+		return;
+
+	/* Free up all allocated memory. */
+	t = adap->tc_u32;
+	for (i = 0; i < t->size; i++) {
+		struct cxgb4_link *link = &t->table[i];
+
+		t4_free_mem(link->tid_map);
+	}
+	t4_free_mem(adap->tc_u32);
+}
+
+struct cxgb4_tc_u32_table *cxgb4_init_tc_u32(struct adapter *adap,
+					     unsigned int size)
+{
+	struct cxgb4_tc_u32_table *t;
+	unsigned int i;
+
+	if (!size)
+		return NULL;
+
+	t = t4_alloc_mem(sizeof(*t) +
+			 (size * sizeof(struct cxgb4_link)));
+	if (!t)
+		return NULL;
+
+	t->size = size;
+
+	for (i = 0; i < t->size; i++) {
+		struct cxgb4_link *link = &t->table[i];
+		unsigned int bmap_size;
+		unsigned int max_tids;
+
+		max_tids = adap->tids.nftids;
+		bmap_size = BITS_TO_LONGS(max_tids);
+		link->tid_map = t4_alloc_mem(sizeof(unsigned long) * bmap_size);
+		if (!link->tid_map)
+			goto out_no_mem;
+		bitmap_zero(link->tid_map, max_tids);
+	}
+
+	return t;
+
+out_no_mem:
+	for (i = 0; i < t->size; i++) {
+		struct cxgb4_link *link = &t->table[i];
+
+		if (link->tid_map)
+			t4_free_mem(link->tid_map);
+	}
+
+	if (t)
+		t4_free_mem(t);
+
+	return NULL;
+}

drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_u32.h

@@ -0,0 +1,57 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2016 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __CXGB4_TC_U32_H
+#define __CXGB4_TC_U32_H
+
+#include <net/pkt_cls.h>
+
+#define CXGB4_MAX_LINK_HANDLE 32
+
+static inline bool can_tc_u32_offload(struct net_device *dev)
+{
+	struct adapter *adap = netdev2adap(dev);
+
+	return (dev->features & NETIF_F_HW_TC) && adap->tc_u32 ? true : false;
+}
+
+int cxgb4_config_knode(struct net_device *dev, __be16 protocol,
+		       struct tc_cls_u32_offload *cls);
+int cxgb4_delete_knode(struct net_device *dev, __be16 protocol,
+		       struct tc_cls_u32_offload *cls);
+
+void cxgb4_cleanup_tc_u32(struct adapter *adapter);
+struct cxgb4_tc_u32_table *cxgb4_init_tc_u32(struct adapter *adap,
+					     unsigned int size);
+#endif /* __CXGB4_TC_U32_H */

drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_u32_parse.h

@@ -0,0 +1,294 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2016 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __CXGB4_TC_U32_PARSE_H
+#define __CXGB4_TC_U32_PARSE_H
+
+struct cxgb4_match_field {
+	int off; /* Offset from the beginning of the header to match */
+	/* Fill the value/mask pair in the spec if matched */
+	int (*val)(struct ch_filter_specification *f, u32 val, u32 mask);
+};
+
+/* IPv4 match fields */
+static inline int cxgb4_fill_ipv4_tos(struct ch_filter_specification *f,
+				      u32 val, u32 mask)
+{
+	f->val.tos  = (ntohl(val)  >> 16) & 0x000000FF;
+	f->mask.tos = (ntohl(mask) >> 16) & 0x000000FF;
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv4_frag(struct ch_filter_specification *f,
+				       u32 val, u32 mask)
+{
+	u32 mask_val;
+	u8 frag_val;
+
+	frag_val = (ntohl(val) >> 13) & 0x00000007;
+	mask_val = ntohl(mask) & 0x0000FFFF;
+
+	if (frag_val == 0x1 && mask_val != 0x3FFF) { /* MF set */
+		f->val.frag = 1;
+		f->mask.frag = 1;
+	} else if (frag_val == 0x2 && mask_val != 0x3FFF) { /* DF set */
+		f->val.frag = 0;
+		f->mask.frag = 1;
+	} else {
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv4_proto(struct ch_filter_specification *f,
+					u32 val, u32 mask)
+{
+	f->val.proto  = (ntohl(val)  >> 16) & 0x000000FF;
+	f->mask.proto = (ntohl(mask) >> 16) & 0x000000FF;
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv4_src_ip(struct ch_filter_specification *f,
+					 u32 val, u32 mask)
+{
+	memcpy(&f->val.fip[0],  &val,  sizeof(u32));
+	memcpy(&f->mask.fip[0], &mask, sizeof(u32));
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv4_dst_ip(struct ch_filter_specification *f,
+					 u32 val, u32 mask)
+{
+	memcpy(&f->val.lip[0],  &val,  sizeof(u32));
+	memcpy(&f->mask.lip[0], &mask, sizeof(u32));
+
+	return 0;
+}
+
+static const struct cxgb4_match_field cxgb4_ipv4_fields[] = {
+	{ .off = 0,  .val = cxgb4_fill_ipv4_tos },
+	{ .off = 4,  .val = cxgb4_fill_ipv4_frag },
+	{ .off = 8,  .val = cxgb4_fill_ipv4_proto },
+	{ .off = 12, .val = cxgb4_fill_ipv4_src_ip },
+	{ .off = 16, .val = cxgb4_fill_ipv4_dst_ip },
+	{ .val = NULL }
+};
+
+/* IPv6 match fields */
+static inline int cxgb4_fill_ipv6_tos(struct ch_filter_specification *f,
+				      u32 val, u32 mask)
+{
+	f->val.tos  = (ntohl(val)  >> 20) & 0x000000FF;
+	f->mask.tos = (ntohl(mask) >> 20) & 0x000000FF;
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv6_proto(struct ch_filter_specification *f,
+					u32 val, u32 mask)
+{
+	f->val.proto  = (ntohl(val)  >> 8) & 0x000000FF;
+	f->mask.proto = (ntohl(mask) >> 8) & 0x000000FF;
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv6_src_ip0(struct ch_filter_specification *f,
+					  u32 val, u32 mask)
+{
+	memcpy(&f->val.fip[0],  &val,  sizeof(u32));
+	memcpy(&f->mask.fip[0], &mask, sizeof(u32));
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv6_src_ip1(struct ch_filter_specification *f,
+					  u32 val, u32 mask)
+{
+	memcpy(&f->val.fip[4],  &val,  sizeof(u32));
+	memcpy(&f->mask.fip[4], &mask, sizeof(u32));
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv6_src_ip2(struct ch_filter_specification *f,
+					  u32 val, u32 mask)
+{
+	memcpy(&f->val.fip[8],  &val,  sizeof(u32));
+	memcpy(&f->mask.fip[8], &mask, sizeof(u32));
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv6_src_ip3(struct ch_filter_specification *f,
+					  u32 val, u32 mask)
+{
+	memcpy(&f->val.fip[12],  &val,  sizeof(u32));
+	memcpy(&f->mask.fip[12], &mask, sizeof(u32));
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv6_dst_ip0(struct ch_filter_specification *f,
+					  u32 val, u32 mask)
+{
+	memcpy(&f->val.lip[0],  &val,  sizeof(u32));
+	memcpy(&f->mask.lip[0], &mask, sizeof(u32));
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv6_dst_ip1(struct ch_filter_specification *f,
+					  u32 val, u32 mask)
+{
+	memcpy(&f->val.lip[4],  &val,  sizeof(u32));
+	memcpy(&f->mask.lip[4], &mask, sizeof(u32));
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv6_dst_ip2(struct ch_filter_specification *f,
+					  u32 val, u32 mask)
+{
+	memcpy(&f->val.lip[8],  &val,  sizeof(u32));
+	memcpy(&f->mask.lip[8], &mask, sizeof(u32));
+
+	return 0;
+}
+
+static inline int cxgb4_fill_ipv6_dst_ip3(struct ch_filter_specification *f,
+					  u32 val, u32 mask)
+{
+	memcpy(&f->val.lip[12],  &val,  sizeof(u32));
+	memcpy(&f->mask.lip[12], &mask, sizeof(u32));
+
+	return 0;
+}
+
+static const struct cxgb4_match_field cxgb4_ipv6_fields[] = {
+	{ .off = 0,  .val = cxgb4_fill_ipv6_tos },
+	{ .off = 4,  .val = cxgb4_fill_ipv6_proto },
+	{ .off = 8,  .val = cxgb4_fill_ipv6_src_ip0 },
+	{ .off = 12, .val = cxgb4_fill_ipv6_src_ip1 },
+	{ .off = 16, .val = cxgb4_fill_ipv6_src_ip2 },
+	{ .off = 20, .val = cxgb4_fill_ipv6_src_ip3 },
+	{ .off = 24, .val = cxgb4_fill_ipv6_dst_ip0 },
+	{ .off = 28, .val = cxgb4_fill_ipv6_dst_ip1 },
+	{ .off = 32, .val = cxgb4_fill_ipv6_dst_ip2 },
+	{ .off = 36, .val = cxgb4_fill_ipv6_dst_ip3 },
+	{ .val = NULL }
+};
+
+/* TCP/UDP match */
+static inline int cxgb4_fill_l4_ports(struct ch_filter_specification *f,
+				      u32 val, u32 mask)
+{
+	f->val.fport  = ntohl(val)  >> 16;
+	f->mask.fport = ntohl(mask) >> 16;
+	f->val.lport  = ntohl(val)  & 0x0000FFFF;
+	f->mask.lport = ntohl(mask) & 0x0000FFFF;
+
+	return 0;
+};
+
+static const struct cxgb4_match_field cxgb4_tcp_fields[] = {
+	{ .off = 0, .val = cxgb4_fill_l4_ports },
+	{ .val = NULL }
+};
+
+static const struct cxgb4_match_field cxgb4_udp_fields[] = {
+	{ .off = 0, .val = cxgb4_fill_l4_ports },
+	{ .val = NULL }
+};
+
+struct cxgb4_next_header {
+	unsigned int offset; /* Offset to next header */
+	/* offset, shift, and mask added to offset above
+	 * to get to next header.  Useful when using a header
+	 * field's value to jump to next header such as IHL field
+	 * in IPv4 header.
+	 */
+	unsigned int offoff;
+	u32 shift;
+	u32 mask;
+	/* match criteria to make this jump */
+	unsigned int match_off;
+	u32 match_val;
+	u32 match_mask;
+	/* location of jump to make */
+	const struct cxgb4_match_field *jump;
+};
+
+/* Accept a rule with a jump to transport layer header based on IHL field in
+ * IPv4 header.
+ */
+static const struct cxgb4_next_header cxgb4_ipv4_jumps[] = {
+	{ .offset = 0, .offoff = 0, .shift = 6, .mask = 0xF,
+	  .match_off = 8, .match_val = 0x600, .match_mask = 0xFF00,
+	  .jump = cxgb4_tcp_fields },
+	{ .offset = 0, .offoff = 0, .shift = 6, .mask = 0xF,
+	  .match_off = 8, .match_val = 0x1100, .match_mask = 0xFF00,
+	  .jump = cxgb4_udp_fields },
+	{ .jump = NULL }
+};
+
+/* Accept a rule with a jump directly past the 40 Bytes of IPv6 fixed header
+ * to get to transport layer header.
+ */
+static const struct cxgb4_next_header cxgb4_ipv6_jumps[] = {
+	{ .offset = 0x28, .offoff = 0, .shift = 0, .mask = 0,
+	  .match_off = 4, .match_val = 0x60000, .match_mask = 0xFF0000,
+	  .jump = cxgb4_tcp_fields },
+	{ .offset = 0x28, .offoff = 0, .shift = 0, .mask = 0,
+	  .match_off = 4, .match_val = 0x110000, .match_mask = 0xFF0000,
+	  .jump = cxgb4_udp_fields },
+	{ .jump = NULL }
+};
+
+struct cxgb4_link {
+	const struct cxgb4_match_field *match_field;  /* Next header */
+	struct ch_filter_specification fs; /* Match spec associated with link */
+	u32 link_handle;         /* Knode handle associated with the link */
+	unsigned long *tid_map;  /* Bitmap for filter tids */
+};
+
+struct cxgb4_tc_u32_table {
+	unsigned int size;          /* number of entries in table */
+	struct cxgb4_link table[0]; /* Jump table */
+};
+#endif /* __CXGB4_TC_U32_PARSE_H */

drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h

@@ -1,7 +1,7 @@
 /*
  * This file is part of the Chelsio T4 Ethernet driver for Linux.
  *
- * Copyright (c) 2003-2014 Chelsio Communications, Inc. All rights reserved.
+ * Copyright (c) 2003-2016 Chelsio Communications, Inc. All rights reserved.
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU
@@ -106,6 +106,7 @@ struct tid_info {
 	unsigned int atid_base;
 
 	struct filter_entry *ftid_tab;
+	unsigned long *ftid_bmap;
 	unsigned int nftids;
 	unsigned int ftid_base;
 	unsigned int aftid_base;
@@ -126,6 +127,8 @@ struct tid_info {
 	atomic_t tids_in_use;
 	/* TIDs in the HASH */
 	atomic_t hash_tids_in_use;
+	/* lock for setting/clearing filter bitmap */
+	spinlock_t ftid_lock;
 };
 
 static inline void *lookup_tid(const struct tid_info *t, unsigned int tid)
@@ -185,6 +188,27 @@ int cxgb4_create_server_filter(const struct net_device *dev, unsigned int stid,
 int cxgb4_remove_server_filter(const struct net_device *dev, unsigned int stid,
 			       unsigned int queue, bool ipv6);
 
+/* Filter operation context to allow callers of cxgb4_set_filter() and
+ * cxgb4_del_filter() to wait for an asynchronous completion.
+ */
+struct filter_ctx {
+	struct completion completion;	/* completion rendezvous */
+	void *closure;			/* caller's opaque information */
+	int result;			/* result of operation */
+	u32 tid;			/* to store tid */
+};
+
+struct ch_filter_specification;
+
+int __cxgb4_set_filter(struct net_device *dev, int filter_id,
+		       struct ch_filter_specification *fs,
+		       struct filter_ctx *ctx);
+int __cxgb4_del_filter(struct net_device *dev, int filter_id,
+		       struct filter_ctx *ctx);
+int cxgb4_set_filter(struct net_device *dev, int filter_id,
+		     struct ch_filter_specification *fs);
+int cxgb4_del_filter(struct net_device *dev, int filter_id);
+
 static inline void set_wr_txq(struct sk_buff *skb, int prio, int queue)
 {
 	skb_set_queue_mapping(skb, (queue << 1) | prio);