freebsd-dev/sys/dev/ice/ice_flow.c
Eric Joyner 56429daea2
ice(4): Update to 1.34.2-k
- Adds FW logging support
  - Once enabled, this lets the firmware print event and error messages
    to the log, increasing the visibility into what the hardware is
    doing; this is useful for debugging
- General bug fixes
- Adds inital DCB support to the driver
  - Notably, this adds support for DCBX to the driver; now with the
    fw_lldp sysctl set to 1, the driver and adapter will adopt a DCBX
    configuration sent from a link partner
  - Adds statistcs sysctls for priority flow control frames
  - Adds new configuration sysctls for DCB-related features: (VLAN) user
    priority to TC mapping; ETS bandwidth allocation; priority flow
    control
- Remove unused SR-IOV files (until support gets added)

Signed-off-by: Eric Joyner <erj@FreeBSD.org>

Tested by:	jeffrey.e.pieper@intel.com
MFC after:	3 days
MFC with:	213e91399b, e438f0a975
Relnotes:	yes
Sponsored by:	Intel Corporation
Differential Revision:	https://reviews.freebsd.org/D34024
2022-03-04 10:51:41 -08:00

1951 lines
63 KiB
C

/* SPDX-License-Identifier: BSD-3-Clause */
/* Copyright (c) 2021, Intel Corporation
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
*
* 2. 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.
*
* 3. Neither the name of the Intel Corporation nor the names of its
* contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
/*$FreeBSD$*/
#include "ice_common.h"
#include "ice_flow.h"
/* Size of known protocol header fields */
#define ICE_FLOW_FLD_SZ_ETH_TYPE 2
#define ICE_FLOW_FLD_SZ_VLAN 2
#define ICE_FLOW_FLD_SZ_IPV4_ADDR 4
#define ICE_FLOW_FLD_SZ_IPV6_ADDR 16
#define ICE_FLOW_FLD_SZ_IP_DSCP 1
#define ICE_FLOW_FLD_SZ_IP_TTL 1
#define ICE_FLOW_FLD_SZ_IP_PROT 1
#define ICE_FLOW_FLD_SZ_PORT 2
#define ICE_FLOW_FLD_SZ_TCP_FLAGS 1
#define ICE_FLOW_FLD_SZ_ICMP_TYPE 1
#define ICE_FLOW_FLD_SZ_ICMP_CODE 1
#define ICE_FLOW_FLD_SZ_ARP_OPER 2
#define ICE_FLOW_FLD_SZ_GRE_KEYID 4
/* Describe properties of a protocol header field */
struct ice_flow_field_info {
enum ice_flow_seg_hdr hdr;
s16 off; /* Offset from start of a protocol header, in bits */
u16 size; /* Size of fields in bits */
};
#define ICE_FLOW_FLD_INFO(_hdr, _offset_bytes, _size_bytes) { \
.hdr = _hdr, \
.off = (_offset_bytes) * BITS_PER_BYTE, \
.size = (_size_bytes) * BITS_PER_BYTE, \
}
/* Table containing properties of supported protocol header fields */
static const
struct ice_flow_field_info ice_flds_info[ICE_FLOW_FIELD_IDX_MAX] = {
/* Ether */
/* ICE_FLOW_FIELD_IDX_ETH_DA */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_ETH, 0, ETH_ALEN),
/* ICE_FLOW_FIELD_IDX_ETH_SA */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_ETH, ETH_ALEN, ETH_ALEN),
/* ICE_FLOW_FIELD_IDX_S_VLAN */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_VLAN, 12, ICE_FLOW_FLD_SZ_VLAN),
/* ICE_FLOW_FIELD_IDX_C_VLAN */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_VLAN, 14, ICE_FLOW_FLD_SZ_VLAN),
/* ICE_FLOW_FIELD_IDX_ETH_TYPE */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_ETH, 0, ICE_FLOW_FLD_SZ_ETH_TYPE),
/* IPv4 / IPv6 */
/* ICE_FLOW_FIELD_IDX_IPV4_DSCP */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_IPV4, 1, ICE_FLOW_FLD_SZ_IP_DSCP),
/* ICE_FLOW_FIELD_IDX_IPV6_DSCP */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_IPV6, 0, ICE_FLOW_FLD_SZ_IP_DSCP),
/* ICE_FLOW_FIELD_IDX_IPV4_TTL */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_NONE, 8, ICE_FLOW_FLD_SZ_IP_TTL),
/* ICE_FLOW_FIELD_IDX_IPV4_PROT */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_NONE, 9, ICE_FLOW_FLD_SZ_IP_PROT),
/* ICE_FLOW_FIELD_IDX_IPV6_TTL */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_NONE, 7, ICE_FLOW_FLD_SZ_IP_TTL),
/* ICE_FLOW_FIELD_IDX_IPV4_PROT */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_NONE, 6, ICE_FLOW_FLD_SZ_IP_PROT),
/* ICE_FLOW_FIELD_IDX_IPV4_SA */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_IPV4, 12, ICE_FLOW_FLD_SZ_IPV4_ADDR),
/* ICE_FLOW_FIELD_IDX_IPV4_DA */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_IPV4, 16, ICE_FLOW_FLD_SZ_IPV4_ADDR),
/* ICE_FLOW_FIELD_IDX_IPV6_SA */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_IPV6, 8, ICE_FLOW_FLD_SZ_IPV6_ADDR),
/* ICE_FLOW_FIELD_IDX_IPV6_DA */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_IPV6, 24, ICE_FLOW_FLD_SZ_IPV6_ADDR),
/* Transport */
/* ICE_FLOW_FIELD_IDX_TCP_SRC_PORT */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_TCP, 0, ICE_FLOW_FLD_SZ_PORT),
/* ICE_FLOW_FIELD_IDX_TCP_DST_PORT */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_TCP, 2, ICE_FLOW_FLD_SZ_PORT),
/* ICE_FLOW_FIELD_IDX_UDP_SRC_PORT */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_UDP, 0, ICE_FLOW_FLD_SZ_PORT),
/* ICE_FLOW_FIELD_IDX_UDP_DST_PORT */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_UDP, 2, ICE_FLOW_FLD_SZ_PORT),
/* ICE_FLOW_FIELD_IDX_SCTP_SRC_PORT */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_SCTP, 0, ICE_FLOW_FLD_SZ_PORT),
/* ICE_FLOW_FIELD_IDX_SCTP_DST_PORT */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_SCTP, 2, ICE_FLOW_FLD_SZ_PORT),
/* ICE_FLOW_FIELD_IDX_TCP_FLAGS */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_TCP, 13, ICE_FLOW_FLD_SZ_TCP_FLAGS),
/* ARP */
/* ICE_FLOW_FIELD_IDX_ARP_SIP */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_ARP, 14, ICE_FLOW_FLD_SZ_IPV4_ADDR),
/* ICE_FLOW_FIELD_IDX_ARP_DIP */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_ARP, 24, ICE_FLOW_FLD_SZ_IPV4_ADDR),
/* ICE_FLOW_FIELD_IDX_ARP_SHA */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_ARP, 8, ETH_ALEN),
/* ICE_FLOW_FIELD_IDX_ARP_DHA */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_ARP, 18, ETH_ALEN),
/* ICE_FLOW_FIELD_IDX_ARP_OP */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_ARP, 6, ICE_FLOW_FLD_SZ_ARP_OPER),
/* ICMP */
/* ICE_FLOW_FIELD_IDX_ICMP_TYPE */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_ICMP, 0, ICE_FLOW_FLD_SZ_ICMP_TYPE),
/* ICE_FLOW_FIELD_IDX_ICMP_CODE */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_ICMP, 1, ICE_FLOW_FLD_SZ_ICMP_CODE),
/* GRE */
/* ICE_FLOW_FIELD_IDX_GRE_KEYID */
ICE_FLOW_FLD_INFO(ICE_FLOW_SEG_HDR_GRE, 12, ICE_FLOW_FLD_SZ_GRE_KEYID),
};
/* Bitmaps indicating relevant packet types for a particular protocol header
*
* Packet types for packets with an Outer/First/Single MAC header
*/
static const u32 ice_ptypes_mac_ofos[] = {
0xFDC00846, 0xBFBF7F7E, 0xF70001DF, 0xFEFDFDFB,
0x0000077E, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Innermost/Last MAC VLAN header */
static const u32 ice_ptypes_macvlan_il[] = {
0x00000000, 0xBC000000, 0x000001DF, 0xF0000000,
0x0000077E, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Outer/First/Single non-frag IPv4 header,
* does NOT include IPV4 other PTYPEs
*/
static const u32 ice_ptypes_ipv4_ofos[] = {
0x1D800000, 0x04000800, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Outer/First/Single non-frag IPv4 header,
* includes IPV4 other PTYPEs
*/
static const u32 ice_ptypes_ipv4_ofos_all[] = {
0x1D800000, 0x04000800, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Innermost/Last IPv4 header */
static const u32 ice_ptypes_ipv4_il[] = {
0xE0000000, 0xB807700E, 0x80000003, 0xE01DC03B,
0x0000000E, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Outer/First/Single non-frag IPv6 header,
* does NOT include IVP6 other PTYPEs
*/
static const u32 ice_ptypes_ipv6_ofos[] = {
0x00000000, 0x00000000, 0x76000000, 0x10002000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Outer/First/Single non-frag IPv6 header,
* includes IPV6 other PTYPEs
*/
static const u32 ice_ptypes_ipv6_ofos_all[] = {
0x00000000, 0x00000000, 0x76000000, 0x10002000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Innermost/Last IPv6 header */
static const u32 ice_ptypes_ipv6_il[] = {
0x00000000, 0x03B80770, 0x000001DC, 0x0EE00000,
0x00000770, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Outer/First/Single
* non-frag IPv4 header - no L4
*/
static const u32 ice_ptypes_ipv4_ofos_no_l4[] = {
0x10800000, 0x04000800, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Innermost/Last IPv4 header - no L4 */
static const u32 ice_ptypes_ipv4_il_no_l4[] = {
0x60000000, 0x18043008, 0x80000002, 0x6010c021,
0x00000008, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Outer/First/Single
* non-frag IPv6 header - no L4
*/
static const u32 ice_ptypes_ipv6_ofos_no_l4[] = {
0x00000000, 0x00000000, 0x42000000, 0x10002000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Innermost/Last IPv6 header - no L4 */
static const u32 ice_ptypes_ipv6_il_no_l4[] = {
0x00000000, 0x02180430, 0x0000010c, 0x086010c0,
0x00000430, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Outermost/First ARP header */
static const u32 ice_ptypes_arp_of[] = {
0x00000800, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* UDP Packet types for non-tunneled packets or tunneled
* packets with inner UDP.
*/
static const u32 ice_ptypes_udp_il[] = {
0x81000000, 0x20204040, 0x04000010, 0x80810102,
0x00000040, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Innermost/Last TCP header */
static const u32 ice_ptypes_tcp_il[] = {
0x04000000, 0x80810102, 0x10000040, 0x02040408,
0x00000102, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Innermost/Last SCTP header */
static const u32 ice_ptypes_sctp_il[] = {
0x08000000, 0x01020204, 0x20000081, 0x04080810,
0x00000204, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Outermost/First ICMP header */
static const u32 ice_ptypes_icmp_of[] = {
0x10000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Innermost/Last ICMP header */
static const u32 ice_ptypes_icmp_il[] = {
0x00000000, 0x02040408, 0x40000102, 0x08101020,
0x00000408, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Outermost/First GRE header */
static const u32 ice_ptypes_gre_of[] = {
0x00000000, 0xBFBF7800, 0x000001DF, 0xFEFDE000,
0x0000017E, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Packet types for packets with an Innermost/Last MAC header */
static const u32 ice_ptypes_mac_il[] = {
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
0x00000000, 0x00000000, 0x00000000, 0x00000000,
};
/* Manage parameters and info. used during the creation of a flow profile */
struct ice_flow_prof_params {
enum ice_block blk;
u16 entry_length; /* # of bytes formatted entry will require */
u8 es_cnt;
struct ice_flow_prof *prof;
/* For ACL, the es[0] will have the data of ICE_RX_MDID_PKT_FLAGS_15_0
* This will give us the direction flags.
*/
struct ice_fv_word es[ICE_MAX_FV_WORDS];
ice_declare_bitmap(ptypes, ICE_FLOW_PTYPE_MAX);
};
#define ICE_FLOW_SEG_HDRS_L3_MASK \
(ICE_FLOW_SEG_HDR_IPV4 | ICE_FLOW_SEG_HDR_IPV6 | \
ICE_FLOW_SEG_HDR_ARP)
#define ICE_FLOW_SEG_HDRS_L4_MASK \
(ICE_FLOW_SEG_HDR_ICMP | ICE_FLOW_SEG_HDR_TCP | ICE_FLOW_SEG_HDR_UDP | \
ICE_FLOW_SEG_HDR_SCTP)
/* mask for L4 protocols that are NOT part of IPV4/6 OTHER PTYPE groups */
#define ICE_FLOW_SEG_HDRS_L4_MASK_NO_OTHER \
(ICE_FLOW_SEG_HDR_TCP | ICE_FLOW_SEG_HDR_UDP | ICE_FLOW_SEG_HDR_SCTP)
/**
* ice_flow_val_hdrs - validates packet segments for valid protocol headers
* @segs: array of one or more packet segments that describe the flow
* @segs_cnt: number of packet segments provided
*/
static enum ice_status
ice_flow_val_hdrs(struct ice_flow_seg_info *segs, u8 segs_cnt)
{
u8 i;
for (i = 0; i < segs_cnt; i++) {
/* Multiple L3 headers */
if (segs[i].hdrs & ICE_FLOW_SEG_HDRS_L3_MASK &&
!ice_is_pow2(segs[i].hdrs & ICE_FLOW_SEG_HDRS_L3_MASK))
return ICE_ERR_PARAM;
/* Multiple L4 headers */
if (segs[i].hdrs & ICE_FLOW_SEG_HDRS_L4_MASK &&
!ice_is_pow2(segs[i].hdrs & ICE_FLOW_SEG_HDRS_L4_MASK))
return ICE_ERR_PARAM;
}
return ICE_SUCCESS;
}
/**
* ice_flow_proc_seg_hdrs - process protocol headers present in pkt segments
* @params: information about the flow to be processed
*
* This function identifies the packet types associated with the protocol
* headers being present in packet segments of the specified flow profile.
*/
static enum ice_status
ice_flow_proc_seg_hdrs(struct ice_flow_prof_params *params)
{
struct ice_flow_prof *prof;
u8 i;
ice_memset(params->ptypes, 0xff, sizeof(params->ptypes),
ICE_NONDMA_MEM);
prof = params->prof;
for (i = 0; i < params->prof->segs_cnt; i++) {
const ice_bitmap_t *src;
u32 hdrs;
hdrs = prof->segs[i].hdrs;
if (hdrs & ICE_FLOW_SEG_HDR_ETH) {
src = !i ? (const ice_bitmap_t *)ice_ptypes_mac_ofos :
(const ice_bitmap_t *)ice_ptypes_mac_il;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
}
if (i && hdrs & ICE_FLOW_SEG_HDR_VLAN) {
src = (const ice_bitmap_t *)ice_ptypes_macvlan_il;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
}
if (!i && hdrs & ICE_FLOW_SEG_HDR_ARP) {
ice_and_bitmap(params->ptypes, params->ptypes,
(const ice_bitmap_t *)ice_ptypes_arp_of,
ICE_FLOW_PTYPE_MAX);
}
if ((hdrs & ICE_FLOW_SEG_HDR_IPV4) &&
(hdrs & ICE_FLOW_SEG_HDR_IPV_OTHER)) {
src = i ?
(const ice_bitmap_t *)ice_ptypes_ipv4_il :
(const ice_bitmap_t *)ice_ptypes_ipv4_ofos_all;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
} else if ((hdrs & ICE_FLOW_SEG_HDR_IPV6) &&
(hdrs & ICE_FLOW_SEG_HDR_IPV_OTHER)) {
src = i ?
(const ice_bitmap_t *)ice_ptypes_ipv6_il :
(const ice_bitmap_t *)ice_ptypes_ipv6_ofos_all;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
} else if ((hdrs & ICE_FLOW_SEG_HDR_IPV4) &&
!(hdrs & ICE_FLOW_SEG_HDRS_L4_MASK_NO_OTHER)) {
src = !i ? (const ice_bitmap_t *)ice_ptypes_ipv4_ofos_no_l4 :
(const ice_bitmap_t *)ice_ptypes_ipv4_il_no_l4;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
} else if (hdrs & ICE_FLOW_SEG_HDR_IPV4) {
src = !i ? (const ice_bitmap_t *)ice_ptypes_ipv4_ofos :
(const ice_bitmap_t *)ice_ptypes_ipv4_il;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
} else if ((hdrs & ICE_FLOW_SEG_HDR_IPV6) &&
!(hdrs & ICE_FLOW_SEG_HDRS_L4_MASK_NO_OTHER)) {
src = !i ? (const ice_bitmap_t *)ice_ptypes_ipv6_ofos_no_l4 :
(const ice_bitmap_t *)ice_ptypes_ipv6_il_no_l4;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
} else if (hdrs & ICE_FLOW_SEG_HDR_IPV6) {
src = !i ? (const ice_bitmap_t *)ice_ptypes_ipv6_ofos :
(const ice_bitmap_t *)ice_ptypes_ipv6_il;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
}
if (hdrs & ICE_FLOW_SEG_HDR_UDP) {
src = (const ice_bitmap_t *)ice_ptypes_udp_il;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
} else if (hdrs & ICE_FLOW_SEG_HDR_TCP) {
ice_and_bitmap(params->ptypes, params->ptypes,
(const ice_bitmap_t *)ice_ptypes_tcp_il,
ICE_FLOW_PTYPE_MAX);
} else if (hdrs & ICE_FLOW_SEG_HDR_SCTP) {
src = (const ice_bitmap_t *)ice_ptypes_sctp_il;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
}
if (hdrs & ICE_FLOW_SEG_HDR_ICMP) {
src = !i ? (const ice_bitmap_t *)ice_ptypes_icmp_of :
(const ice_bitmap_t *)ice_ptypes_icmp_il;
ice_and_bitmap(params->ptypes, params->ptypes, src,
ICE_FLOW_PTYPE_MAX);
} else if (hdrs & ICE_FLOW_SEG_HDR_GRE) {
if (!i) {
src = (const ice_bitmap_t *)ice_ptypes_gre_of;
ice_and_bitmap(params->ptypes, params->ptypes,
src, ICE_FLOW_PTYPE_MAX);
}
}
}
return ICE_SUCCESS;
}
/**
* ice_flow_xtract_fld - Create an extraction sequence entry for the given field
* @hw: pointer to the HW struct
* @params: information about the flow to be processed
* @seg: packet segment index of the field to be extracted
* @fld: ID of field to be extracted
*
* This function determines the protocol ID, offset, and size of the given
* field. It then allocates one or more extraction sequence entries for the
* given field, and fill the entries with protocol ID and offset information.
*/
static enum ice_status
ice_flow_xtract_fld(struct ice_hw *hw, struct ice_flow_prof_params *params,
u8 seg, enum ice_flow_field fld)
{
enum ice_flow_field sib = ICE_FLOW_FIELD_IDX_MAX;
enum ice_prot_id prot_id = ICE_PROT_ID_INVAL;
u8 fv_words = hw->blk[params->blk].es.fvw;
struct ice_flow_fld_info *flds;
u16 cnt, ese_bits, i;
u16 off;
flds = params->prof->segs[seg].fields;
switch (fld) {
case ICE_FLOW_FIELD_IDX_ETH_DA:
case ICE_FLOW_FIELD_IDX_ETH_SA:
case ICE_FLOW_FIELD_IDX_S_VLAN:
case ICE_FLOW_FIELD_IDX_C_VLAN:
prot_id = seg == 0 ? ICE_PROT_MAC_OF_OR_S : ICE_PROT_MAC_IL;
break;
case ICE_FLOW_FIELD_IDX_ETH_TYPE:
prot_id = seg == 0 ? ICE_PROT_ETYPE_OL : ICE_PROT_ETYPE_IL;
break;
case ICE_FLOW_FIELD_IDX_IPV4_DSCP:
prot_id = seg == 0 ? ICE_PROT_IPV4_OF_OR_S : ICE_PROT_IPV4_IL;
break;
case ICE_FLOW_FIELD_IDX_IPV6_DSCP:
prot_id = seg == 0 ? ICE_PROT_IPV6_OF_OR_S : ICE_PROT_IPV6_IL;
break;
case ICE_FLOW_FIELD_IDX_IPV4_TTL:
case ICE_FLOW_FIELD_IDX_IPV4_PROT:
prot_id = seg == 0 ? ICE_PROT_IPV4_OF_OR_S : ICE_PROT_IPV4_IL;
/* TTL and PROT share the same extraction seq. entry.
* Each is considered a sibling to the other in terms of sharing
* the same extraction sequence entry.
*/
if (fld == ICE_FLOW_FIELD_IDX_IPV4_TTL)
sib = ICE_FLOW_FIELD_IDX_IPV4_PROT;
else
sib = ICE_FLOW_FIELD_IDX_IPV4_TTL;
break;
case ICE_FLOW_FIELD_IDX_IPV6_TTL:
case ICE_FLOW_FIELD_IDX_IPV6_PROT:
prot_id = seg == 0 ? ICE_PROT_IPV6_OF_OR_S : ICE_PROT_IPV6_IL;
/* TTL and PROT share the same extraction seq. entry.
* Each is considered a sibling to the other in terms of sharing
* the same extraction sequence entry.
*/
if (fld == ICE_FLOW_FIELD_IDX_IPV6_TTL)
sib = ICE_FLOW_FIELD_IDX_IPV6_PROT;
else
sib = ICE_FLOW_FIELD_IDX_IPV6_TTL;
break;
case ICE_FLOW_FIELD_IDX_IPV4_SA:
case ICE_FLOW_FIELD_IDX_IPV4_DA:
prot_id = seg == 0 ? ICE_PROT_IPV4_OF_OR_S : ICE_PROT_IPV4_IL;
break;
case ICE_FLOW_FIELD_IDX_IPV6_SA:
case ICE_FLOW_FIELD_IDX_IPV6_DA:
prot_id = seg == 0 ? ICE_PROT_IPV6_OF_OR_S : ICE_PROT_IPV6_IL;
break;
case ICE_FLOW_FIELD_IDX_TCP_SRC_PORT:
case ICE_FLOW_FIELD_IDX_TCP_DST_PORT:
case ICE_FLOW_FIELD_IDX_TCP_FLAGS:
prot_id = ICE_PROT_TCP_IL;
break;
case ICE_FLOW_FIELD_IDX_UDP_SRC_PORT:
case ICE_FLOW_FIELD_IDX_UDP_DST_PORT:
prot_id = ICE_PROT_UDP_IL_OR_S;
break;
case ICE_FLOW_FIELD_IDX_SCTP_SRC_PORT:
case ICE_FLOW_FIELD_IDX_SCTP_DST_PORT:
prot_id = ICE_PROT_SCTP_IL;
break;
case ICE_FLOW_FIELD_IDX_ARP_SIP:
case ICE_FLOW_FIELD_IDX_ARP_DIP:
case ICE_FLOW_FIELD_IDX_ARP_SHA:
case ICE_FLOW_FIELD_IDX_ARP_DHA:
case ICE_FLOW_FIELD_IDX_ARP_OP:
prot_id = ICE_PROT_ARP_OF;
break;
case ICE_FLOW_FIELD_IDX_ICMP_TYPE:
case ICE_FLOW_FIELD_IDX_ICMP_CODE:
/* ICMP type and code share the same extraction seq. entry */
prot_id = (params->prof->segs[seg].hdrs &
ICE_FLOW_SEG_HDR_IPV4) ?
ICE_PROT_ICMP_IL : ICE_PROT_ICMPV6_IL;
sib = fld == ICE_FLOW_FIELD_IDX_ICMP_TYPE ?
ICE_FLOW_FIELD_IDX_ICMP_CODE :
ICE_FLOW_FIELD_IDX_ICMP_TYPE;
break;
case ICE_FLOW_FIELD_IDX_GRE_KEYID:
prot_id = ICE_PROT_GRE_OF;
break;
default:
return ICE_ERR_NOT_IMPL;
}
/* Each extraction sequence entry is a word in size, and extracts a
* word-aligned offset from a protocol header.
*/
ese_bits = ICE_FLOW_FV_EXTRACT_SZ * BITS_PER_BYTE;
flds[fld].xtrct.prot_id = prot_id;
flds[fld].xtrct.off = (ice_flds_info[fld].off / ese_bits) *
ICE_FLOW_FV_EXTRACT_SZ;
flds[fld].xtrct.disp = (u8)(ice_flds_info[fld].off % ese_bits);
flds[fld].xtrct.idx = params->es_cnt;
/* Adjust the next field-entry index after accommodating the number of
* entries this field consumes
*/
cnt = DIVIDE_AND_ROUND_UP(flds[fld].xtrct.disp +
ice_flds_info[fld].size, ese_bits);
/* Fill in the extraction sequence entries needed for this field */
off = flds[fld].xtrct.off;
for (i = 0; i < cnt; i++) {
/* Only consume an extraction sequence entry if there is no
* sibling field associated with this field or the sibling entry
* already extracts the word shared with this field.
*/
if (sib == ICE_FLOW_FIELD_IDX_MAX ||
flds[sib].xtrct.prot_id == ICE_PROT_ID_INVAL ||
flds[sib].xtrct.off != off) {
u8 idx;
/* Make sure the number of extraction sequence required
* does not exceed the block's capability
*/
if (params->es_cnt >= fv_words)
return ICE_ERR_MAX_LIMIT;
/* some blocks require a reversed field vector layout */
if (hw->blk[params->blk].es.reverse)
idx = fv_words - params->es_cnt - 1;
else
idx = params->es_cnt;
params->es[idx].prot_id = prot_id;
params->es[idx].off = off;
params->es_cnt++;
}
off += ICE_FLOW_FV_EXTRACT_SZ;
}
return ICE_SUCCESS;
}
/**
* ice_flow_create_xtrct_seq - Create an extraction sequence for given segments
* @hw: pointer to the HW struct
* @params: information about the flow to be processed
*
* This function iterates through all matched fields in the given segments, and
* creates an extraction sequence for the fields.
*/
static enum ice_status
ice_flow_create_xtrct_seq(struct ice_hw *hw,
struct ice_flow_prof_params *params)
{
enum ice_status status = ICE_SUCCESS;
u8 i;
for (i = 0; i < params->prof->segs_cnt; i++) {
u64 match = params->prof->segs[i].match;
enum ice_flow_field j;
ice_for_each_set_bit(j, (ice_bitmap_t *)&match,
ICE_FLOW_FIELD_IDX_MAX) {
status = ice_flow_xtract_fld(hw, params, i, j);
if (status)
return status;
ice_clear_bit(j, (ice_bitmap_t *)&match);
}
}
return status;
}
/**
* ice_flow_proc_segs - process all packet segments associated with a profile
* @hw: pointer to the HW struct
* @params: information about the flow to be processed
*/
static enum ice_status
ice_flow_proc_segs(struct ice_hw *hw, struct ice_flow_prof_params *params)
{
enum ice_status status;
status = ice_flow_proc_seg_hdrs(params);
if (status)
return status;
status = ice_flow_create_xtrct_seq(hw, params);
if (status)
return status;
switch (params->blk) {
case ICE_BLK_RSS:
status = ICE_SUCCESS;
break;
default:
return ICE_ERR_NOT_IMPL;
}
return status;
}
#define ICE_FLOW_FIND_PROF_CHK_FLDS 0x00000001
#define ICE_FLOW_FIND_PROF_CHK_VSI 0x00000002
#define ICE_FLOW_FIND_PROF_NOT_CHK_DIR 0x00000004
/**
* ice_flow_find_prof_conds - Find a profile matching headers and conditions
* @hw: pointer to the HW struct
* @blk: classification stage
* @dir: flow direction
* @segs: array of one or more packet segments that describe the flow
* @segs_cnt: number of packet segments provided
* @vsi_handle: software VSI handle to check VSI (ICE_FLOW_FIND_PROF_CHK_VSI)
* @conds: additional conditions to be checked (ICE_FLOW_FIND_PROF_CHK_*)
*/
static struct ice_flow_prof *
ice_flow_find_prof_conds(struct ice_hw *hw, enum ice_block blk,
enum ice_flow_dir dir, struct ice_flow_seg_info *segs,
u8 segs_cnt, u16 vsi_handle, u32 conds)
{
struct ice_flow_prof *p, *prof = NULL;
ice_acquire_lock(&hw->fl_profs_locks[blk]);
LIST_FOR_EACH_ENTRY(p, &hw->fl_profs[blk], ice_flow_prof, l_entry)
if ((p->dir == dir || conds & ICE_FLOW_FIND_PROF_NOT_CHK_DIR) &&
segs_cnt && segs_cnt == p->segs_cnt) {
u8 i;
/* Check for profile-VSI association if specified */
if ((conds & ICE_FLOW_FIND_PROF_CHK_VSI) &&
ice_is_vsi_valid(hw, vsi_handle) &&
!ice_is_bit_set(p->vsis, vsi_handle))
continue;
/* Protocol headers must be checked. Matched fields are
* checked if specified.
*/
for (i = 0; i < segs_cnt; i++)
if (segs[i].hdrs != p->segs[i].hdrs ||
((conds & ICE_FLOW_FIND_PROF_CHK_FLDS) &&
segs[i].match != p->segs[i].match))
break;
/* A match is found if all segments are matched */
if (i == segs_cnt) {
prof = p;
break;
}
}
ice_release_lock(&hw->fl_profs_locks[blk]);
return prof;
}
/**
* ice_flow_find_prof - Look up a profile matching headers and matched fields
* @hw: pointer to the HW struct
* @blk: classification stage
* @dir: flow direction
* @segs: array of one or more packet segments that describe the flow
* @segs_cnt: number of packet segments provided
*/
u64
ice_flow_find_prof(struct ice_hw *hw, enum ice_block blk, enum ice_flow_dir dir,
struct ice_flow_seg_info *segs, u8 segs_cnt)
{
struct ice_flow_prof *p;
p = ice_flow_find_prof_conds(hw, blk, dir, segs, segs_cnt,
ICE_MAX_VSI, ICE_FLOW_FIND_PROF_CHK_FLDS);
return p ? p->id : ICE_FLOW_PROF_ID_INVAL;
}
/**
* ice_flow_find_prof_id - Look up a profile with given profile ID
* @hw: pointer to the HW struct
* @blk: classification stage
* @prof_id: unique ID to identify this flow profile
*/
static struct ice_flow_prof *
ice_flow_find_prof_id(struct ice_hw *hw, enum ice_block blk, u64 prof_id)
{
struct ice_flow_prof *p;
LIST_FOR_EACH_ENTRY(p, &hw->fl_profs[blk], ice_flow_prof, l_entry)
if (p->id == prof_id)
return p;
return NULL;
}
/**
* ice_flow_get_hw_prof - return the HW profile for a specific profile ID handle
* @hw: pointer to the HW struct
* @blk: classification stage
* @prof_id: the profile ID handle
* @hw_prof_id: pointer to variable to receive the HW profile ID
*/
enum ice_status
ice_flow_get_hw_prof(struct ice_hw *hw, enum ice_block blk, u64 prof_id,
u8 *hw_prof_id)
{
enum ice_status status = ICE_ERR_DOES_NOT_EXIST;
struct ice_prof_map *map;
ice_acquire_lock(&hw->blk[blk].es.prof_map_lock);
map = ice_search_prof_id(hw, blk, prof_id);
if (map) {
*hw_prof_id = map->prof_id;
status = ICE_SUCCESS;
}
ice_release_lock(&hw->blk[blk].es.prof_map_lock);
return status;
}
/**
* ice_flow_add_prof_sync - Add a flow profile for packet segments and fields
* @hw: pointer to the HW struct
* @blk: classification stage
* @dir: flow direction
* @prof_id: unique ID to identify this flow profile
* @segs: array of one or more packet segments that describe the flow
* @segs_cnt: number of packet segments provided
* @acts: array of default actions
* @acts_cnt: number of default actions
* @prof: stores the returned flow profile added
*
* Assumption: the caller has acquired the lock to the profile list
*/
static enum ice_status
ice_flow_add_prof_sync(struct ice_hw *hw, enum ice_block blk,
enum ice_flow_dir dir, u64 prof_id,
struct ice_flow_seg_info *segs, u8 segs_cnt,
struct ice_flow_action *acts, u8 acts_cnt,
struct ice_flow_prof **prof)
{
struct ice_flow_prof_params *params;
enum ice_status status;
u8 i;
if (!prof || (acts_cnt && !acts))
return ICE_ERR_BAD_PTR;
params = (struct ice_flow_prof_params *)ice_malloc(hw, sizeof(*params));
if (!params)
return ICE_ERR_NO_MEMORY;
params->prof = (struct ice_flow_prof *)
ice_malloc(hw, sizeof(*params->prof));
if (!params->prof) {
status = ICE_ERR_NO_MEMORY;
goto free_params;
}
/* initialize extraction sequence to all invalid (0xff) */
for (i = 0; i < ICE_MAX_FV_WORDS; i++) {
params->es[i].prot_id = ICE_PROT_INVALID;
params->es[i].off = ICE_FV_OFFSET_INVAL;
}
params->blk = blk;
params->prof->id = prof_id;
params->prof->dir = dir;
params->prof->segs_cnt = segs_cnt;
/* Make a copy of the segments that need to be persistent in the flow
* profile instance
*/
for (i = 0; i < segs_cnt; i++)
ice_memcpy(&params->prof->segs[i], &segs[i], sizeof(*segs),
ICE_NONDMA_TO_NONDMA);
status = ice_flow_proc_segs(hw, params);
if (status) {
ice_debug(hw, ICE_DBG_FLOW, "Error processing a flow's packet segments\n");
goto out;
}
/* Add a HW profile for this flow profile */
status = ice_add_prof(hw, blk, prof_id, (u8 *)params->ptypes,
params->es);
if (status) {
ice_debug(hw, ICE_DBG_FLOW, "Error adding a HW flow profile\n");
goto out;
}
*prof = params->prof;
out:
if (status) {
ice_free(hw, params->prof);
}
free_params:
ice_free(hw, params);
return status;
}
/**
* ice_flow_rem_prof_sync - remove a flow profile
* @hw: pointer to the hardware structure
* @blk: classification stage
* @prof: pointer to flow profile to remove
*
* Assumption: the caller has acquired the lock to the profile list
*/
static enum ice_status
ice_flow_rem_prof_sync(struct ice_hw *hw, enum ice_block blk,
struct ice_flow_prof *prof)
{
enum ice_status status;
/* Remove all hardware profiles associated with this flow profile */
status = ice_rem_prof(hw, blk, prof->id);
if (!status) {
LIST_DEL(&prof->l_entry);
ice_free(hw, prof);
}
return status;
}
/**
* ice_flow_assoc_vsig_vsi - associate a VSI with VSIG
* @hw: pointer to the hardware structure
* @blk: classification stage
* @vsi_handle: software VSI handle
* @vsig: target VSI group
*
* Assumption: the caller has already verified that the VSI to
* be added has the same characteristics as the VSIG and will
* thereby have access to all resources added to that VSIG.
*/
enum ice_status
ice_flow_assoc_vsig_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi_handle,
u16 vsig)
{
enum ice_status status;
if (!ice_is_vsi_valid(hw, vsi_handle) || blk >= ICE_BLK_COUNT)
return ICE_ERR_PARAM;
ice_acquire_lock(&hw->fl_profs_locks[blk]);
status = ice_add_vsi_flow(hw, blk, ice_get_hw_vsi_num(hw, vsi_handle),
vsig);
ice_release_lock(&hw->fl_profs_locks[blk]);
return status;
}
/**
* ice_flow_assoc_prof - associate a VSI with a flow profile
* @hw: pointer to the hardware structure
* @blk: classification stage
* @prof: pointer to flow profile
* @vsi_handle: software VSI handle
*
* Assumption: the caller has acquired the lock to the profile list
* and the software VSI handle has been validated
*/
static enum ice_status
ice_flow_assoc_prof(struct ice_hw *hw, enum ice_block blk,
struct ice_flow_prof *prof, u16 vsi_handle)
{
enum ice_status status = ICE_SUCCESS;
if (!ice_is_bit_set(prof->vsis, vsi_handle)) {
status = ice_add_prof_id_flow(hw, blk,
ice_get_hw_vsi_num(hw,
vsi_handle),
prof->id);
if (!status)
ice_set_bit(vsi_handle, prof->vsis);
else
ice_debug(hw, ICE_DBG_FLOW, "HW profile add failed, %d\n",
status);
}
return status;
}
/**
* ice_flow_disassoc_prof - disassociate a VSI from a flow profile
* @hw: pointer to the hardware structure
* @blk: classification stage
* @prof: pointer to flow profile
* @vsi_handle: software VSI handle
*
* Assumption: the caller has acquired the lock to the profile list
* and the software VSI handle has been validated
*/
static enum ice_status
ice_flow_disassoc_prof(struct ice_hw *hw, enum ice_block blk,
struct ice_flow_prof *prof, u16 vsi_handle)
{
enum ice_status status = ICE_SUCCESS;
if (ice_is_bit_set(prof->vsis, vsi_handle)) {
status = ice_rem_prof_id_flow(hw, blk,
ice_get_hw_vsi_num(hw,
vsi_handle),
prof->id);
if (!status)
ice_clear_bit(vsi_handle, prof->vsis);
else
ice_debug(hw, ICE_DBG_FLOW, "HW profile remove failed, %d\n",
status);
}
return status;
}
/**
* ice_flow_add_prof - Add a flow profile for packet segments and matched fields
* @hw: pointer to the HW struct
* @blk: classification stage
* @dir: flow direction
* @prof_id: unique ID to identify this flow profile
* @segs: array of one or more packet segments that describe the flow
* @segs_cnt: number of packet segments provided
* @acts: array of default actions
* @acts_cnt: number of default actions
* @prof: stores the returned flow profile added
*/
static enum ice_status
ice_flow_add_prof(struct ice_hw *hw, enum ice_block blk, enum ice_flow_dir dir,
u64 prof_id, struct ice_flow_seg_info *segs, u8 segs_cnt,
struct ice_flow_action *acts, u8 acts_cnt,
struct ice_flow_prof **prof)
{
enum ice_status status;
if (segs_cnt > ICE_FLOW_SEG_MAX)
return ICE_ERR_MAX_LIMIT;
if (!segs_cnt)
return ICE_ERR_PARAM;
if (!segs)
return ICE_ERR_BAD_PTR;
status = ice_flow_val_hdrs(segs, segs_cnt);
if (status)
return status;
ice_acquire_lock(&hw->fl_profs_locks[blk]);
status = ice_flow_add_prof_sync(hw, blk, dir, prof_id, segs, segs_cnt,
acts, acts_cnt, prof);
if (!status)
LIST_ADD(&(*prof)->l_entry, &hw->fl_profs[blk]);
ice_release_lock(&hw->fl_profs_locks[blk]);
return status;
}
/**
* ice_flow_rem_prof - Remove a flow profile and all entries associated with it
* @hw: pointer to the HW struct
* @blk: the block for which the flow profile is to be removed
* @prof_id: unique ID of the flow profile to be removed
*/
static enum ice_status
ice_flow_rem_prof(struct ice_hw *hw, enum ice_block blk, u64 prof_id)
{
struct ice_flow_prof *prof;
enum ice_status status;
ice_acquire_lock(&hw->fl_profs_locks[blk]);
prof = ice_flow_find_prof_id(hw, blk, prof_id);
if (!prof) {
status = ICE_ERR_DOES_NOT_EXIST;
goto out;
}
/* prof becomes invalid after the call */
status = ice_flow_rem_prof_sync(hw, blk, prof);
out:
ice_release_lock(&hw->fl_profs_locks[blk]);
return status;
}
/**
* ice_flow_set_fld_ext - specifies locations of field from entry's input buffer
* @seg: packet segment the field being set belongs to
* @fld: field to be set
* @field_type: type of the field
* @val_loc: if not ICE_FLOW_FLD_OFF_INVAL, location of the value to match from
* entry's input buffer
* @mask_loc: if not ICE_FLOW_FLD_OFF_INVAL, location of mask value from entry's
* input buffer
* @last_loc: if not ICE_FLOW_FLD_OFF_INVAL, location of last/upper value from
* entry's input buffer
*
* This helper function stores information of a field being matched, including
* the type of the field and the locations of the value to match, the mask, and
* the upper-bound value in the start of the input buffer for a flow entry.
* This function should only be used for fixed-size data structures.
*
* This function also opportunistically determines the protocol headers to be
* present based on the fields being set. Some fields cannot be used alone to
* determine the protocol headers present. Sometimes, fields for particular
* protocol headers are not matched. In those cases, the protocol headers
* must be explicitly set.
*/
static void
ice_flow_set_fld_ext(struct ice_flow_seg_info *seg, enum ice_flow_field fld,
enum ice_flow_fld_match_type field_type, u16 val_loc,
u16 mask_loc, u16 last_loc)
{
u64 bit = BIT_ULL(fld);
seg->match |= bit;
if (field_type == ICE_FLOW_FLD_TYPE_RANGE)
seg->range |= bit;
seg->fields[fld].type = field_type;
seg->fields[fld].src.val = val_loc;
seg->fields[fld].src.mask = mask_loc;
seg->fields[fld].src.last = last_loc;
ICE_FLOW_SET_HDRS(seg, ice_flds_info[fld].hdr);
}
/**
* ice_flow_set_fld - specifies locations of field from entry's input buffer
* @seg: packet segment the field being set belongs to
* @fld: field to be set
* @val_loc: if not ICE_FLOW_FLD_OFF_INVAL, location of the value to match from
* entry's input buffer
* @mask_loc: if not ICE_FLOW_FLD_OFF_INVAL, location of mask value from entry's
* input buffer
* @last_loc: if not ICE_FLOW_FLD_OFF_INVAL, location of last/upper value from
* entry's input buffer
* @range: indicate if field being matched is to be in a range
*
* This function specifies the locations, in the form of byte offsets from the
* start of the input buffer for a flow entry, from where the value to match,
* the mask value, and upper value can be extracted. These locations are then
* stored in the flow profile. When adding a flow entry associated with the
* flow profile, these locations will be used to quickly extract the values and
* create the content of a match entry. This function should only be used for
* fixed-size data structures.
*/
static void
ice_flow_set_fld(struct ice_flow_seg_info *seg, enum ice_flow_field fld,
u16 val_loc, u16 mask_loc, u16 last_loc, bool range)
{
enum ice_flow_fld_match_type t = range ?
ICE_FLOW_FLD_TYPE_RANGE : ICE_FLOW_FLD_TYPE_REG;
ice_flow_set_fld_ext(seg, fld, t, val_loc, mask_loc, last_loc);
}
/**
* ice_flow_set_fld_prefix - sets locations of prefix field from entry's buf
* @seg: packet segment the field being set belongs to
* @fld: field to be set
* @val_loc: if not ICE_FLOW_FLD_OFF_INVAL, location of the value to match from
* entry's input buffer
* @pref_loc: location of prefix value from entry's input buffer
* @pref_sz: size of the location holding the prefix value
*
* This function specifies the locations, in the form of byte offsets from the
* start of the input buffer for a flow entry, from where the value to match
* and the IPv4 prefix value can be extracted. These locations are then stored
* in the flow profile. When adding flow entries to the associated flow profile,
* these locations can be used to quickly extract the values to create the
* content of a match entry. This function should only be used for fixed-size
* data structures.
*/
void
ice_flow_set_fld_prefix(struct ice_flow_seg_info *seg, enum ice_flow_field fld,
u16 val_loc, u16 pref_loc, u8 pref_sz)
{
/* For this type of field, the "mask" location is for the prefix value's
* location and the "last" location is for the size of the location of
* the prefix value.
*/
ice_flow_set_fld_ext(seg, fld, ICE_FLOW_FLD_TYPE_PREFIX, val_loc,
pref_loc, (u16)pref_sz);
}
#define ICE_FLOW_RSS_SEG_HDR_L3_MASKS \
(ICE_FLOW_SEG_HDR_IPV4 | ICE_FLOW_SEG_HDR_IPV6)
#define ICE_FLOW_RSS_SEG_HDR_L4_MASKS \
(ICE_FLOW_SEG_HDR_TCP | ICE_FLOW_SEG_HDR_UDP | ICE_FLOW_SEG_HDR_SCTP)
#define ICE_FLOW_RSS_SEG_HDR_VAL_MASKS \
(ICE_FLOW_RSS_SEG_HDR_L3_MASKS | \
ICE_FLOW_RSS_SEG_HDR_L4_MASKS)
/**
* ice_flow_set_rss_seg_info - setup packet segments for RSS
* @segs: pointer to the flow field segment(s)
* @seg_cnt: segment count
* @cfg: configure parameters
*
* Helper function to extract fields from hash bitmap and use flow
* header value to set flow field segment for further use in flow
* profile entry or removal.
*/
static enum ice_status
ice_flow_set_rss_seg_info(struct ice_flow_seg_info *segs, u8 seg_cnt,
const struct ice_rss_hash_cfg *cfg)
{
struct ice_flow_seg_info *seg;
u64 val;
u8 i;
/* set inner most segment */
seg = &segs[seg_cnt - 1];
ice_for_each_set_bit(i, (const ice_bitmap_t *)&cfg->hash_flds,
ICE_FLOW_FIELD_IDX_MAX)
ice_flow_set_fld(seg, (enum ice_flow_field)i,
ICE_FLOW_FLD_OFF_INVAL, ICE_FLOW_FLD_OFF_INVAL,
ICE_FLOW_FLD_OFF_INVAL, false);
ICE_FLOW_SET_HDRS(seg, cfg->addl_hdrs);
/* set outer most header */
if (cfg->hdr_type == ICE_RSS_INNER_HEADERS_W_OUTER_IPV4)
segs[ICE_RSS_OUTER_HEADERS].hdrs |= ICE_FLOW_SEG_HDR_IPV4 |
ICE_FLOW_SEG_HDR_IPV_FRAG |
ICE_FLOW_SEG_HDR_IPV_OTHER;
else if (cfg->hdr_type == ICE_RSS_INNER_HEADERS_W_OUTER_IPV6)
segs[ICE_RSS_OUTER_HEADERS].hdrs |= ICE_FLOW_SEG_HDR_IPV6 |
ICE_FLOW_SEG_HDR_IPV_FRAG |
ICE_FLOW_SEG_HDR_IPV_OTHER;
else if (cfg->hdr_type == ICE_RSS_INNER_HEADERS_W_OUTER_IPV4_GRE)
segs[ICE_RSS_OUTER_HEADERS].hdrs |= ICE_FLOW_SEG_HDR_IPV4 |
ICE_FLOW_SEG_HDR_GRE |
ICE_FLOW_SEG_HDR_IPV_OTHER;
else if (cfg->hdr_type == ICE_RSS_INNER_HEADERS_W_OUTER_IPV6_GRE)
segs[ICE_RSS_OUTER_HEADERS].hdrs |= ICE_FLOW_SEG_HDR_IPV6 |
ICE_FLOW_SEG_HDR_GRE |
ICE_FLOW_SEG_HDR_IPV_OTHER;
if (seg->hdrs & ~ICE_FLOW_RSS_SEG_HDR_VAL_MASKS)
return ICE_ERR_PARAM;
val = (u64)(seg->hdrs & ICE_FLOW_RSS_SEG_HDR_L3_MASKS);
if (val && !ice_is_pow2(val))
return ICE_ERR_CFG;
val = (u64)(seg->hdrs & ICE_FLOW_RSS_SEG_HDR_L4_MASKS);
if (val && !ice_is_pow2(val))
return ICE_ERR_CFG;
return ICE_SUCCESS;
}
/**
* ice_rem_vsi_rss_list - remove VSI from RSS list
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
*
* Remove the VSI from all RSS configurations in the list.
*/
void ice_rem_vsi_rss_list(struct ice_hw *hw, u16 vsi_handle)
{
struct ice_rss_cfg *r, *tmp;
if (LIST_EMPTY(&hw->rss_list_head))
return;
ice_acquire_lock(&hw->rss_locks);
LIST_FOR_EACH_ENTRY_SAFE(r, tmp, &hw->rss_list_head,
ice_rss_cfg, l_entry)
if (ice_test_and_clear_bit(vsi_handle, r->vsis))
if (!ice_is_any_bit_set(r->vsis, ICE_MAX_VSI)) {
LIST_DEL(&r->l_entry);
ice_free(hw, r);
}
ice_release_lock(&hw->rss_locks);
}
/**
* ice_rem_vsi_rss_cfg - remove RSS configurations associated with VSI
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
*
* This function will iterate through all flow profiles and disassociate
* the VSI from that profile. If the flow profile has no VSIs it will
* be removed.
*/
enum ice_status ice_rem_vsi_rss_cfg(struct ice_hw *hw, u16 vsi_handle)
{
const enum ice_block blk = ICE_BLK_RSS;
struct ice_flow_prof *p, *t;
enum ice_status status = ICE_SUCCESS;
if (!ice_is_vsi_valid(hw, vsi_handle))
return ICE_ERR_PARAM;
if (LIST_EMPTY(&hw->fl_profs[blk]))
return ICE_SUCCESS;
ice_acquire_lock(&hw->rss_locks);
LIST_FOR_EACH_ENTRY_SAFE(p, t, &hw->fl_profs[blk], ice_flow_prof,
l_entry)
if (ice_is_bit_set(p->vsis, vsi_handle)) {
status = ice_flow_disassoc_prof(hw, blk, p, vsi_handle);
if (status)
break;
if (!ice_is_any_bit_set(p->vsis, ICE_MAX_VSI)) {
status = ice_flow_rem_prof(hw, blk, p->id);
if (status)
break;
}
}
ice_release_lock(&hw->rss_locks);
return status;
}
/**
* ice_get_rss_hdr_type - get a RSS profile's header type
* @prof: RSS flow profile
*/
static enum ice_rss_cfg_hdr_type
ice_get_rss_hdr_type(struct ice_flow_prof *prof)
{
enum ice_rss_cfg_hdr_type hdr_type = ICE_RSS_ANY_HEADERS;
if (prof->segs_cnt == ICE_FLOW_SEG_SINGLE) {
hdr_type = ICE_RSS_OUTER_HEADERS;
} else if (prof->segs_cnt == ICE_FLOW_SEG_MAX) {
if (prof->segs[ICE_RSS_OUTER_HEADERS].hdrs == ICE_FLOW_SEG_HDR_NONE)
hdr_type = ICE_RSS_INNER_HEADERS;
if (prof->segs[ICE_RSS_OUTER_HEADERS].hdrs & ICE_FLOW_SEG_HDR_IPV4)
hdr_type = ICE_RSS_INNER_HEADERS_W_OUTER_IPV4;
if (prof->segs[ICE_RSS_OUTER_HEADERS].hdrs & ICE_FLOW_SEG_HDR_IPV6)
hdr_type = ICE_RSS_INNER_HEADERS_W_OUTER_IPV6;
}
return hdr_type;
}
/**
* ice_rem_rss_list - remove RSS configuration from list
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
* @prof: pointer to flow profile
*
* Assumption: lock has already been acquired for RSS list
*/
static void
ice_rem_rss_list(struct ice_hw *hw, u16 vsi_handle, struct ice_flow_prof *prof)
{
enum ice_rss_cfg_hdr_type hdr_type;
struct ice_rss_cfg *r, *tmp;
/* Search for RSS hash fields associated to the VSI that match the
* hash configurations associated to the flow profile. If found
* remove from the RSS entry list of the VSI context and delete entry.
*/
hdr_type = ice_get_rss_hdr_type(prof);
LIST_FOR_EACH_ENTRY_SAFE(r, tmp, &hw->rss_list_head,
ice_rss_cfg, l_entry)
if (r->hash.hash_flds == prof->segs[prof->segs_cnt - 1].match &&
r->hash.addl_hdrs == prof->segs[prof->segs_cnt - 1].hdrs &&
r->hash.hdr_type == hdr_type) {
ice_clear_bit(vsi_handle, r->vsis);
if (!ice_is_any_bit_set(r->vsis, ICE_MAX_VSI)) {
LIST_DEL(&r->l_entry);
ice_free(hw, r);
}
return;
}
}
/**
* ice_add_rss_list - add RSS configuration to list
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
* @prof: pointer to flow profile
*
* Assumption: lock has already been acquired for RSS list
*/
static enum ice_status
ice_add_rss_list(struct ice_hw *hw, u16 vsi_handle, struct ice_flow_prof *prof)
{
enum ice_rss_cfg_hdr_type hdr_type;
struct ice_rss_cfg *r, *rss_cfg;
hdr_type = ice_get_rss_hdr_type(prof);
LIST_FOR_EACH_ENTRY(r, &hw->rss_list_head,
ice_rss_cfg, l_entry)
if (r->hash.hash_flds == prof->segs[prof->segs_cnt - 1].match &&
r->hash.addl_hdrs == prof->segs[prof->segs_cnt - 1].hdrs &&
r->hash.hdr_type == hdr_type) {
ice_set_bit(vsi_handle, r->vsis);
return ICE_SUCCESS;
}
rss_cfg = (struct ice_rss_cfg *)ice_malloc(hw, sizeof(*rss_cfg));
if (!rss_cfg)
return ICE_ERR_NO_MEMORY;
rss_cfg->hash.hash_flds = prof->segs[prof->segs_cnt - 1].match;
rss_cfg->hash.addl_hdrs = prof->segs[prof->segs_cnt - 1].hdrs;
rss_cfg->hash.hdr_type = hdr_type;
rss_cfg->hash.symm = prof->cfg.symm;
ice_set_bit(vsi_handle, rss_cfg->vsis);
LIST_ADD_TAIL(&rss_cfg->l_entry, &hw->rss_list_head);
return ICE_SUCCESS;
}
#define ICE_FLOW_PROF_HASH_S 0
#define ICE_FLOW_PROF_HASH_M (0xFFFFFFFFULL << ICE_FLOW_PROF_HASH_S)
#define ICE_FLOW_PROF_HDR_S 32
#define ICE_FLOW_PROF_HDR_M (0x3FFFFFFFULL << ICE_FLOW_PROF_HDR_S)
#define ICE_FLOW_PROF_ENCAP_S 62
#define ICE_FLOW_PROF_ENCAP_M (0x3ULL << ICE_FLOW_PROF_ENCAP_S)
/* Flow profile ID format:
* [0:31] - Packet match fields
* [32:61] - Protocol header
* [62:63] - Encapsulation flag:
* 0 if non-tunneled
* 1 if tunneled
* 2 for tunneled with outer ipv4
* 3 for tunneled with outer ipv6
*/
#define ICE_FLOW_GEN_PROFID(hash, hdr, encap) \
((u64)(((u64)(hash) & ICE_FLOW_PROF_HASH_M) | \
(((u64)(hdr) << ICE_FLOW_PROF_HDR_S) & ICE_FLOW_PROF_HDR_M) | \
(((u64)(encap) << ICE_FLOW_PROF_ENCAP_S) & ICE_FLOW_PROF_ENCAP_M)))
/**
* ice_add_rss_cfg_sync - add an RSS configuration
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
* @cfg: configure parameters
*
* Assumption: lock has already been acquired for RSS list
*/
static enum ice_status
ice_add_rss_cfg_sync(struct ice_hw *hw, u16 vsi_handle,
const struct ice_rss_hash_cfg *cfg)
{
const enum ice_block blk = ICE_BLK_RSS;
struct ice_flow_prof *prof = NULL;
struct ice_flow_seg_info *segs;
enum ice_status status;
u8 segs_cnt;
if (cfg->symm)
return ICE_ERR_PARAM;
segs_cnt = (cfg->hdr_type == ICE_RSS_OUTER_HEADERS) ?
ICE_FLOW_SEG_SINGLE : ICE_FLOW_SEG_MAX;
segs = (struct ice_flow_seg_info *)ice_calloc(hw, segs_cnt,
sizeof(*segs));
if (!segs)
return ICE_ERR_NO_MEMORY;
/* Construct the packet segment info from the hashed fields */
status = ice_flow_set_rss_seg_info(segs, segs_cnt, cfg);
if (status)
goto exit;
/* Search for a flow profile that has matching headers, hash fields
* and has the input VSI associated to it. If found, no further
* operations required and exit.
*/
prof = ice_flow_find_prof_conds(hw, blk, ICE_FLOW_RX, segs, segs_cnt,
vsi_handle,
ICE_FLOW_FIND_PROF_CHK_FLDS |
ICE_FLOW_FIND_PROF_CHK_VSI);
if (prof)
goto exit;
/* Check if a flow profile exists with the same protocol headers and
* associated with the input VSI. If so disassociate the VSI from
* this profile. The VSI will be added to a new profile created with
* the protocol header and new hash field configuration.
*/
prof = ice_flow_find_prof_conds(hw, blk, ICE_FLOW_RX, segs, segs_cnt,
vsi_handle, ICE_FLOW_FIND_PROF_CHK_VSI);
if (prof) {
status = ice_flow_disassoc_prof(hw, blk, prof, vsi_handle);
if (!status)
ice_rem_rss_list(hw, vsi_handle, prof);
else
goto exit;
/* Remove profile if it has no VSIs associated */
if (!ice_is_any_bit_set(prof->vsis, ICE_MAX_VSI)) {
status = ice_flow_rem_prof(hw, blk, prof->id);
if (status)
goto exit;
}
}
/* Search for a profile that has same match fields only. If this
* exists then associate the VSI to this profile.
*/
prof = ice_flow_find_prof_conds(hw, blk, ICE_FLOW_RX, segs, segs_cnt,
vsi_handle,
ICE_FLOW_FIND_PROF_CHK_FLDS);
if (prof) {
status = ice_flow_assoc_prof(hw, blk, prof, vsi_handle);
if (!status)
status = ice_add_rss_list(hw, vsi_handle, prof);
goto exit;
}
/* Create a new flow profile with generated profile and packet
* segment information.
*/
status = ice_flow_add_prof(hw, blk, ICE_FLOW_RX,
ICE_FLOW_GEN_PROFID(cfg->hash_flds,
segs[segs_cnt - 1].hdrs,
cfg->hdr_type),
segs, segs_cnt, NULL, 0, &prof);
if (status)
goto exit;
status = ice_flow_assoc_prof(hw, blk, prof, vsi_handle);
/* If association to a new flow profile failed then this profile can
* be removed.
*/
if (status) {
ice_flow_rem_prof(hw, blk, prof->id);
goto exit;
}
status = ice_add_rss_list(hw, vsi_handle, prof);
prof->cfg.symm = cfg->symm;
exit:
ice_free(hw, segs);
return status;
}
/**
* ice_add_rss_cfg - add an RSS configuration with specified hashed fields
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
* @cfg: configure parameters
*
* This function will generate a flow profile based on fields associated with
* the input fields to hash on, the flow type and use the VSI number to add
* a flow entry to the profile.
*/
enum ice_status
ice_add_rss_cfg(struct ice_hw *hw, u16 vsi_handle,
const struct ice_rss_hash_cfg *cfg)
{
struct ice_rss_hash_cfg local_cfg;
enum ice_status status;
if (!ice_is_vsi_valid(hw, vsi_handle) ||
!cfg || cfg->hdr_type > ICE_RSS_ANY_HEADERS ||
cfg->hash_flds == ICE_HASH_INVALID)
return ICE_ERR_PARAM;
local_cfg = *cfg;
if (cfg->hdr_type < ICE_RSS_ANY_HEADERS) {
ice_acquire_lock(&hw->rss_locks);
status = ice_add_rss_cfg_sync(hw, vsi_handle, &local_cfg);
ice_release_lock(&hw->rss_locks);
} else {
ice_acquire_lock(&hw->rss_locks);
local_cfg.hdr_type = ICE_RSS_OUTER_HEADERS;
status = ice_add_rss_cfg_sync(hw, vsi_handle, &local_cfg);
if (!status) {
local_cfg.hdr_type = ICE_RSS_INNER_HEADERS;
status = ice_add_rss_cfg_sync(hw, vsi_handle,
&local_cfg);
}
ice_release_lock(&hw->rss_locks);
}
return status;
}
/**
* ice_rem_rss_cfg_sync - remove an existing RSS configuration
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
* @cfg: configure parameters
*
* Assumption: lock has already been acquired for RSS list
*/
static enum ice_status
ice_rem_rss_cfg_sync(struct ice_hw *hw, u16 vsi_handle,
const struct ice_rss_hash_cfg *cfg)
{
const enum ice_block blk = ICE_BLK_RSS;
struct ice_flow_seg_info *segs;
struct ice_flow_prof *prof;
enum ice_status status;
u8 segs_cnt;
segs_cnt = (cfg->hdr_type == ICE_RSS_OUTER_HEADERS) ?
ICE_FLOW_SEG_SINGLE : ICE_FLOW_SEG_MAX;
segs = (struct ice_flow_seg_info *)ice_calloc(hw, segs_cnt,
sizeof(*segs));
if (!segs)
return ICE_ERR_NO_MEMORY;
/* Construct the packet segment info from the hashed fields */
status = ice_flow_set_rss_seg_info(segs, segs_cnt, cfg);
if (status)
goto out;
prof = ice_flow_find_prof_conds(hw, blk, ICE_FLOW_RX, segs, segs_cnt,
vsi_handle,
ICE_FLOW_FIND_PROF_CHK_FLDS);
if (!prof) {
status = ICE_ERR_DOES_NOT_EXIST;
goto out;
}
status = ice_flow_disassoc_prof(hw, blk, prof, vsi_handle);
if (status)
goto out;
/* Remove RSS configuration from VSI context before deleting
* the flow profile.
*/
ice_rem_rss_list(hw, vsi_handle, prof);
if (!ice_is_any_bit_set(prof->vsis, ICE_MAX_VSI))
status = ice_flow_rem_prof(hw, blk, prof->id);
out:
ice_free(hw, segs);
return status;
}
/**
* ice_rem_rss_cfg - remove an existing RSS config with matching hashed fields
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
* @cfg: configure parameters
*
* This function will lookup the flow profile based on the input
* hash field bitmap, iterate through the profile entry list of
* that profile and find entry associated with input VSI to be
* removed. Calls are made to underlying flow apis which will in
* turn build or update buffers for RSS XLT1 section.
*/
enum ice_status
ice_rem_rss_cfg(struct ice_hw *hw, u16 vsi_handle,
const struct ice_rss_hash_cfg *cfg)
{
struct ice_rss_hash_cfg local_cfg;
enum ice_status status;
if (!ice_is_vsi_valid(hw, vsi_handle) ||
!cfg || cfg->hdr_type > ICE_RSS_ANY_HEADERS ||
cfg->hash_flds == ICE_HASH_INVALID)
return ICE_ERR_PARAM;
ice_acquire_lock(&hw->rss_locks);
local_cfg = *cfg;
if (cfg->hdr_type < ICE_RSS_ANY_HEADERS) {
status = ice_rem_rss_cfg_sync(hw, vsi_handle, &local_cfg);
} else {
local_cfg.hdr_type = ICE_RSS_OUTER_HEADERS;
status = ice_rem_rss_cfg_sync(hw, vsi_handle, &local_cfg);
if (!status) {
local_cfg.hdr_type = ICE_RSS_INNER_HEADERS;
status = ice_rem_rss_cfg_sync(hw, vsi_handle,
&local_cfg);
}
}
ice_release_lock(&hw->rss_locks);
return status;
}
/* Mapping of AVF hash bit fields to an L3-L4 hash combination.
* As the ice_flow_avf_hdr_field represent individual bit shifts in a hash,
* convert its values to their appropriate flow L3, L4 values.
*/
#define ICE_FLOW_AVF_RSS_IPV4_MASKS \
(BIT_ULL(ICE_AVF_FLOW_FIELD_IPV4_OTHER) | \
BIT_ULL(ICE_AVF_FLOW_FIELD_FRAG_IPV4))
#define ICE_FLOW_AVF_RSS_TCP_IPV4_MASKS \
(BIT_ULL(ICE_AVF_FLOW_FIELD_IPV4_TCP_SYN_NO_ACK) | \
BIT_ULL(ICE_AVF_FLOW_FIELD_IPV4_TCP))
#define ICE_FLOW_AVF_RSS_UDP_IPV4_MASKS \
(BIT_ULL(ICE_AVF_FLOW_FIELD_UNICAST_IPV4_UDP) | \
BIT_ULL(ICE_AVF_FLOW_FIELD_MULTICAST_IPV4_UDP) | \
BIT_ULL(ICE_AVF_FLOW_FIELD_IPV4_UDP))
#define ICE_FLOW_AVF_RSS_ALL_IPV4_MASKS \
(ICE_FLOW_AVF_RSS_TCP_IPV4_MASKS | ICE_FLOW_AVF_RSS_UDP_IPV4_MASKS | \
ICE_FLOW_AVF_RSS_IPV4_MASKS | BIT_ULL(ICE_AVF_FLOW_FIELD_IPV4_SCTP))
#define ICE_FLOW_AVF_RSS_IPV6_MASKS \
(BIT_ULL(ICE_AVF_FLOW_FIELD_IPV6_OTHER) | \
BIT_ULL(ICE_AVF_FLOW_FIELD_FRAG_IPV6))
#define ICE_FLOW_AVF_RSS_UDP_IPV6_MASKS \
(BIT_ULL(ICE_AVF_FLOW_FIELD_UNICAST_IPV6_UDP) | \
BIT_ULL(ICE_AVF_FLOW_FIELD_MULTICAST_IPV6_UDP) | \
BIT_ULL(ICE_AVF_FLOW_FIELD_IPV6_UDP))
#define ICE_FLOW_AVF_RSS_TCP_IPV6_MASKS \
(BIT_ULL(ICE_AVF_FLOW_FIELD_IPV6_TCP_SYN_NO_ACK) | \
BIT_ULL(ICE_AVF_FLOW_FIELD_IPV6_TCP))
#define ICE_FLOW_AVF_RSS_ALL_IPV6_MASKS \
(ICE_FLOW_AVF_RSS_TCP_IPV6_MASKS | ICE_FLOW_AVF_RSS_UDP_IPV6_MASKS | \
ICE_FLOW_AVF_RSS_IPV6_MASKS | BIT_ULL(ICE_AVF_FLOW_FIELD_IPV6_SCTP))
/**
* ice_add_avf_rss_cfg - add an RSS configuration for AVF driver
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
* @avf_hash: hash bit fields (ICE_AVF_FLOW_FIELD_*) to configure
*
* This function will take the hash bitmap provided by the AVF driver via a
* message, convert it to ICE-compatible values, and configure RSS flow
* profiles.
*/
enum ice_status
ice_add_avf_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 avf_hash)
{
enum ice_status status = ICE_SUCCESS;
struct ice_rss_hash_cfg hcfg;
u64 hash_flds;
if (avf_hash == ICE_AVF_FLOW_FIELD_INVALID ||
!ice_is_vsi_valid(hw, vsi_handle))
return ICE_ERR_PARAM;
/* Make sure no unsupported bits are specified */
if (avf_hash & ~(ICE_FLOW_AVF_RSS_ALL_IPV4_MASKS |
ICE_FLOW_AVF_RSS_ALL_IPV6_MASKS))
return ICE_ERR_CFG;
hash_flds = avf_hash;
/* Always create an L3 RSS configuration for any L4 RSS configuration */
if (hash_flds & ICE_FLOW_AVF_RSS_ALL_IPV4_MASKS)
hash_flds |= ICE_FLOW_AVF_RSS_IPV4_MASKS;
if (hash_flds & ICE_FLOW_AVF_RSS_ALL_IPV6_MASKS)
hash_flds |= ICE_FLOW_AVF_RSS_IPV6_MASKS;
/* Create the corresponding RSS configuration for each valid hash bit */
while (hash_flds) {
u64 rss_hash = ICE_HASH_INVALID;
if (hash_flds & ICE_FLOW_AVF_RSS_ALL_IPV4_MASKS) {
if (hash_flds & ICE_FLOW_AVF_RSS_IPV4_MASKS) {
rss_hash = ICE_FLOW_HASH_IPV4;
hash_flds &= ~ICE_FLOW_AVF_RSS_IPV4_MASKS;
} else if (hash_flds &
ICE_FLOW_AVF_RSS_TCP_IPV4_MASKS) {
rss_hash = ICE_FLOW_HASH_IPV4 |
ICE_FLOW_HASH_TCP_PORT;
hash_flds &= ~ICE_FLOW_AVF_RSS_TCP_IPV4_MASKS;
} else if (hash_flds &
ICE_FLOW_AVF_RSS_UDP_IPV4_MASKS) {
rss_hash = ICE_FLOW_HASH_IPV4 |
ICE_FLOW_HASH_UDP_PORT;
hash_flds &= ~ICE_FLOW_AVF_RSS_UDP_IPV4_MASKS;
} else if (hash_flds &
BIT_ULL(ICE_AVF_FLOW_FIELD_IPV4_SCTP)) {
rss_hash = ICE_FLOW_HASH_IPV4 |
ICE_FLOW_HASH_SCTP_PORT;
hash_flds &=
~BIT_ULL(ICE_AVF_FLOW_FIELD_IPV4_SCTP);
}
} else if (hash_flds & ICE_FLOW_AVF_RSS_ALL_IPV6_MASKS) {
if (hash_flds & ICE_FLOW_AVF_RSS_IPV6_MASKS) {
rss_hash = ICE_FLOW_HASH_IPV6;
hash_flds &= ~ICE_FLOW_AVF_RSS_IPV6_MASKS;
} else if (hash_flds &
ICE_FLOW_AVF_RSS_TCP_IPV6_MASKS) {
rss_hash = ICE_FLOW_HASH_IPV6 |
ICE_FLOW_HASH_TCP_PORT;
hash_flds &= ~ICE_FLOW_AVF_RSS_TCP_IPV6_MASKS;
} else if (hash_flds &
ICE_FLOW_AVF_RSS_UDP_IPV6_MASKS) {
rss_hash = ICE_FLOW_HASH_IPV6 |
ICE_FLOW_HASH_UDP_PORT;
hash_flds &= ~ICE_FLOW_AVF_RSS_UDP_IPV6_MASKS;
} else if (hash_flds &
BIT_ULL(ICE_AVF_FLOW_FIELD_IPV6_SCTP)) {
rss_hash = ICE_FLOW_HASH_IPV6 |
ICE_FLOW_HASH_SCTP_PORT;
hash_flds &=
~BIT_ULL(ICE_AVF_FLOW_FIELD_IPV6_SCTP);
}
}
if (rss_hash == ICE_HASH_INVALID)
return ICE_ERR_OUT_OF_RANGE;
hcfg.addl_hdrs = ICE_FLOW_SEG_HDR_NONE;
hcfg.hash_flds = rss_hash;
hcfg.symm = false;
hcfg.hdr_type = ICE_RSS_ANY_HEADERS;
status = ice_add_rss_cfg(hw, vsi_handle, &hcfg);
if (status)
break;
}
return status;
}
/**
* ice_replay_rss_cfg - replay RSS configurations associated with VSI
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
*/
enum ice_status ice_replay_rss_cfg(struct ice_hw *hw, u16 vsi_handle)
{
enum ice_status status = ICE_SUCCESS;
struct ice_rss_cfg *r;
if (!ice_is_vsi_valid(hw, vsi_handle))
return ICE_ERR_PARAM;
ice_acquire_lock(&hw->rss_locks);
LIST_FOR_EACH_ENTRY(r, &hw->rss_list_head,
ice_rss_cfg, l_entry) {
if (ice_is_bit_set(r->vsis, vsi_handle)) {
status = ice_add_rss_cfg_sync(hw, vsi_handle, &r->hash);
if (status)
break;
}
}
ice_release_lock(&hw->rss_locks);
return status;
}
/**
* ice_get_rss_cfg - returns hashed fields for the given header types
* @hw: pointer to the hardware structure
* @vsi_handle: software VSI handle
* @hdrs: protocol header type
*
* This function will return the match fields of the first instance of flow
* profile having the given header types and containing input VSI
*/
u64 ice_get_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u32 hdrs)
{
u64 rss_hash = ICE_HASH_INVALID;
struct ice_rss_cfg *r;
/* verify if the protocol header is non zero and VSI is valid */
if (hdrs == ICE_FLOW_SEG_HDR_NONE || !ice_is_vsi_valid(hw, vsi_handle))
return ICE_HASH_INVALID;
ice_acquire_lock(&hw->rss_locks);
LIST_FOR_EACH_ENTRY(r, &hw->rss_list_head,
ice_rss_cfg, l_entry)
if (ice_is_bit_set(r->vsis, vsi_handle) &&
r->hash.addl_hdrs == hdrs) {
rss_hash = r->hash.hash_flds;
break;
}
ice_release_lock(&hw->rss_locks);
return rss_hash;
}