/* SPDX-License-Identifier: (GPL-2.0+ OR BSD-3-Clause) */
/* Copyright 2014-2016 Freescale Semiconductor Inc.
* Copyright 2016 NXP
*/
#ifndef __DPAA2_ETH_H
#define __DPAA2_ETH_H
#include <linux/netdevice.h>
#include <linux/if_vlan.h>
#include <linux/fsl/mc.h>
#include "../../fsl-mc/include/dpaa2-io.h"
#include "../../fsl-mc/include/dpaa2-fd.h"
#include "dpni.h"
#include "dpni-cmd.h"
#include "dpaa2-eth-trace.h"
#define DPAA2_WRIOP_VERSION(x, y, z) ((x) << 10 | (y) << 5 | (z) << 0)
#define DPAA2_ETH_STORE_SIZE 16
/* Maximum number of scatter-gather entries in an ingress frame,
* considering the maximum receive frame size is 64K
*/
#define DPAA2_ETH_MAX_SG_ENTRIES ((64 * 1024) / DPAA2_ETH_RX_BUF_SIZE)
/* Maximum acceptable MTU value. It is in direct relation with the hardware
* enforced Max Frame Length (currently 10k).
*/
#define DPAA2_ETH_MFL (10 * 1024)
#define DPAA2_ETH_MAX_MTU (DPAA2_ETH_MFL - VLAN_ETH_HLEN)
/* Convert L3 MTU to L2 MFL */
#define DPAA2_ETH_L2_MAX_FRM(mtu) ((mtu) + VLAN_ETH_HLEN)
/* Set the taildrop threshold (in bytes) to allow the enqueue of several jumbo
* frames in the Rx queues (length of the current frame is not
* taken into account when making the taildrop decision)
*/
#define DPAA2_ETH_TAILDROP_THRESH (64 * 1024)
/* Buffer quota per queue. Must be large enough such that for minimum sized
* frames taildrop kicks in before the bpool gets depleted, so we compute
* how many 64B frames fit inside the taildrop threshold and add a margin
* to accommodate the buffer refill delay.
*/
#define DPAA2_ETH_MAX_FRAMES_PER_QUEUE (DPAA2_ETH_TAILDROP_THRESH / 64)
#define DPAA2_ETH_NUM_BUFS (DPAA2_ETH_MAX_FRAMES_PER_QUEUE + 256)
#define DPAA2_ETH_REFILL_THRESH DPAA2_ETH_MAX_FRAMES_PER_QUEUE
/* Maximum number of buffers that can be acquired/released through a single
* QBMan command
*/
#define DPAA2_ETH_BUFS_PER_CMD 7
/* Hardware requires alignment for ingress/egress buffer addresses */
#define DPAA2_ETH_TX_BUF_ALIGN 64
#define DPAA2_ETH_RX_BUF_SIZE 2048
#define DPAA2_ETH_SKB_SIZE \
(DPAA2_ETH_RX_BUF_SIZE + SKB_DATA_ALIGN(sizeof(struct skb_shared_info)))
/* Hardware annotation area in RX/TX buffers */
#define DPAA2_ETH_RX_HWA_SIZE 64
#define DPAA2_ETH_TX_HWA_SIZE 128
/* PTP nominal frequency 1GHz */
#define DPAA2_PTP_CLK_PERIOD_NS 1
/* Due to a limitation in WRIOP 1.0.0, the RX buffer data must be aligned
* to 256B. For newer revisions, the requirement is only for 64B alignment
*/
#define DPAA2_ETH_RX_BUF_ALIGN_REV1 256
#define DPAA2_ETH_RX_BUF_ALIGN 64
/* We are accommodating a skb backpointer and some S/G info
* in the frame's software annotation. The hardware
* options are either 0 or 64, so we choose the latter.
*/
#define DPAA2_ETH_SWA_SIZE 64
/* Must keep this struct smaller than DPAA2_ETH_SWA_SIZE */
struct dpaa2_eth_swa {
struct sk_buff *skb;
struct scatterlist *scl;
int num_sg;
int sgt_size;
};
/* Annotation valid bits in FD FRC */
#define DPAA2_FD_FRC_FASV 0x8000
#define DPAA2_FD_FRC_FAEADV 0x4000
#define DPAA2_FD_FRC_FAPRV 0x2000
#define DPAA2_FD_FRC_FAIADV 0x1000
#define DPAA2_FD_FRC_FASWOV 0x0800
#define DPAA2_FD_FRC_FAICFDV 0x0400
/* Error bits in FD CTRL */
#define DPAA2_FD_CTRL_UFD 0x00000004
#define DPAA2_FD_CTRL_SBE 0x00000008
#define DPAA2_FD_CTRL_FSE 0x00000020
#define DPAA2_FD_CTRL_FAERR 0x00000040
#define DPAA2_FD_RX_ERR_MASK (DPAA2_FD_CTRL_SBE | \
DPAA2_FD_CTRL_FAERR)
#define DPAA2_FD_TX_ERR_MASK (DPAA2_FD_CTRL_UFD | \
DPAA2_FD_CTRL_SBE | \
DPAA2_FD_CTRL_FSE | \
DPAA2_FD_CTRL_FAERR)
/* Annotation bits in FD CTRL */
#define DPAA2_FD_CTRL_PTA 0x00800000
#define DPAA2_FD_CTRL_PTV1 0x00400000
#define DPAA2_FD_CTRL_ASAL 0x00020000 /* ASAL = 128B */
/* Frame annotation status */
struct dpaa2_fas {
u8 reserved;
u8 ppid;
__le16 ifpid;
__le32 status;
};
/* Frame annotation status word is located in the first 8 bytes
* of the buffer's hardware annoatation area
*/
#define DPAA2_FAS_OFFSET 0
#define DPAA2_FAS_SIZE (sizeof(struct dpaa2_fas))
/* Timestamp is located in the next 8 bytes of the buffer's
* hardware annotation area
*/
#define DPAA2_TS_OFFSET 0x8
/* Frame annotation egress action descriptor */
#define DPAA2_FAEAD_OFFSET 0x58
struct dpaa2_faead {
__le32 conf_fqid;
__le32 ctrl;
};
#define DPAA2_FAEAD_A2V 0x20000000
#define DPAA2_FAEAD_UPDV 0x00001000
#define DPAA2_FAEAD_UPD 0x00000010
/* Accessors for the hardware annotation fields that we use */
static inline void *dpaa2_get_hwa(void *buf_addr, bool swa)
{
return buf_addr + (swa ? DPAA2_ETH_SWA_SIZE : 0);
}
static inline struct dpaa2_fas *dpaa2_get_fas(void *buf_addr, bool swa)
{
return dpaa2_get_hwa(buf_addr, swa) + DPAA2_FAS_OFFSET;
}
static inline __le64 *dpaa2_get_ts(void *buf_addr, bool swa)
{
return dpaa2_get_hwa(buf_addr, swa) + DPAA2_TS_OFFSET;
}
static inline struct dpaa2_faead *dpaa2_get_faead(void *buf_addr, bool swa)
{
return dpaa2_get_hwa(buf_addr, swa) + DPAA2_FAEAD_OFFSET;
}
/* Error and status bits in the frame annotation status word */
/* Debug frame, otherwise supposed to be discarded */
#define DPAA2_FAS_DISC 0x80000000
/* MACSEC frame */
#define DPAA2_FAS_MS 0x40000000
#define DPAA2_FAS_PTP 0x08000000
/* Ethernet multicast frame */
#define DPAA2_FAS_MC 0x04000000
/* Ethernet broadcast frame */
#define DPAA2_FAS_BC 0x02000000
#define DPAA2_FAS_KSE 0x00040000
#define DPAA2_FAS_EOFHE 0x00020000
#define DPAA2_FAS_MNLE 0x00010000
#define DPAA2_FAS_TIDE 0x00008000
#define DPAA2_FAS_PIEE 0x00004000
/* Frame length error */
#define DPAA2_FAS_FLE 0x00002000
/* Frame physical error */
#define DPAA2_FAS_FPE 0x00001000
#define DPAA2_FAS_PTE 0x00000080
#define DPAA2_FAS_ISP 0x00000040
#define DPAA2_FAS_PHE 0x00000020
#define DPAA2_FAS_BLE 0x00000010
/* L3 csum validation performed */
#define DPAA2_FAS_L3CV 0x00000008
/* L3 csum error */
#define DPAA2_FAS_L3CE 0x00000004
/* L4 csum validation performed */
#define DPAA2_FAS_L4CV 0x00000002
/* L4 csum error */
#define DPAA2_FAS_L4CE 0x00000001
/* Possible errors on the ingress path */
#define DPAA2_FAS_RX_ERR_MASK (DPAA2_FAS_KSE | \
DPAA2_FAS_EOFHE | \
DPAA2_FAS_MNLE | \
DPAA2_FAS_TIDE | \
DPAA2_FAS_PIEE | \
DPAA2_FAS_FLE | \
DPAA2_FAS_FPE | \
DPAA2_FAS_PTE | \
DPAA2_FAS_ISP | \
DPAA2_FAS_PHE | \
DPAA2_FAS_BLE | \
DPAA2_FAS_L3CE | \
DPAA2_FAS_L4CE)
/* Time in milliseconds between link state updates */
#define DPAA2_ETH_LINK_STATE_REFRESH 1000
/* Number of times to retry a frame enqueue before giving up.
* Value determined empirically, in order to minimize the number
* of frames dropped on Tx
*/
#define DPAA2_ETH_ENQUEUE_RETRIES 10
/* Driver statistics, other than those in struct rtnl_link_stats64.
* These are usually collected per-CPU and aggregated by ethtool.
*/
struct dpaa2_eth_drv_stats {
__u64 tx_conf_frames;
__u64 tx_conf_bytes;
__u64 tx_sg_frames;
__u64 tx_sg_bytes;
__u64 tx_reallocs;
__u64 rx_sg_frames;
__u64 rx_sg_bytes;
/* Enqueues retried due to portal busy */
__u64 tx_portal_busy;
};
/* Per-FQ statistics */
struct dpaa2_eth_fq_stats {
/* Number of frames received on this queue */
__u64 frames;
};
/* Per-channel statistics */
struct dpaa2_eth_ch_stats {
/* Volatile dequeues retried due to portal busy */
__u64 dequeue_portal_busy;
/* Number of CDANs; useful to estimate avg NAPI len */
__u64 cdan;
/* Number of frames received on queues from this channel */
__u64 frames;
/* Pull errors */
__u64 pull_err;
};
/* Maximum number of queues associated with a DPNI */
#define DPAA2_ETH_MAX_RX_QUEUES 16
#define DPAA2_ETH_MAX_TX_QUEUES 16
#define DPAA2_ETH_MAX_QUEUES (DPAA2_ETH_MAX_RX_QUEUES + \
DPAA2_ETH_MAX_TX_QUEUES)
#define DPAA2_ETH_MAX_DPCONS 16
enum dpaa2_eth_fq_type {
DPAA2_RX_FQ = 0,
DPAA2_TX_CONF_FQ,
};
struct dpaa2_eth_priv;
struct dpaa2_eth_fq {
u32 fqid;
u32 tx_qdbin;
u16 flowid;
int target_cpu;
struct dpaa2_eth_channel *channel;
enum dpaa2_eth_fq_type type;
void (*consume)(struct dpaa2_eth_priv *priv,
struct dpaa2_eth_channel *ch,
const struct dpaa2_fd *fd,
struct napi_struct *napi,
u16 queue_id);
struct dpaa2_eth_fq_stats stats;
};
struct dpaa2_eth_channel {
struct dpaa2_io_notification_ctx nctx;
struct fsl_mc_device *dpcon;
int dpcon_id;
int ch_id;
struct napi_struct napi;
struct dpaa2_io *dpio;
struct dpaa2_io_store *store;
struct dpaa2_eth_priv *priv;
int buf_count;
struct dpaa2_eth_ch_stats stats;
};
struct dpaa2_eth_hash_fields {
u64 rxnfc_field;
enum net_prot cls_prot;
int cls_field;
int size;
};
/* Driver private data */
struct dpaa2_eth_priv {
struct net_device *net_dev;
u8 num_fqs;
struct dpaa2_eth_fq fq[DPAA2_ETH_MAX_QUEUES];
u8 num_channels;
struct dpaa2_eth_channel *channel[DPAA2_ETH_MAX_DPCONS];
struct dpni_attr dpni_attrs;
u16 dpni_ver_major;
u16 dpni_ver_minor;
u16 tx_data_offset;
struct fsl_mc_device *dpbp_dev;
u16 bpid;
struct iommu_domain *iommu_domain;
bool tx_tstamp; /* Tx timestamping enabled */
bool rx_tstamp; /* Rx timestamping enabled */
u16 tx_qdid;
u16 rx_buf_align;
struct fsl_mc_io *mc_io;
/* Cores which have an affine DPIO/DPCON.
* This is the cpu set on which Rx and Tx conf frames are processed
*/
struct cpumask dpio_cpumask;
/* Standard statistics */
struct rtnl_link_stats64 __percpu *percpu_stats;
/* Extra stats, in addition to the ones known by the kernel */
struct dpaa2_eth_drv_stats __percpu *percpu_extras;
u16 mc_token;
struct dpni_link_state link_state;
bool do_link_poll;
struct task_struct *poll_thread;
/* enabled ethtool hashing bits */
u64 rx_hash_fields;
};
#define DPAA2_RXH_SUPPORTED (RXH_L2DA | RXH_VLAN | RXH_L3_PROTO \
| RXH_IP_SRC | RXH_IP_DST | RXH_L4_B_0_1 \
| RXH_L4_B_2_3)
/* default Rx hash options, set during probing */
#define DPAA2_RXH_DEFAULT (RXH_L3_PROTO | RXH_IP_SRC | RXH_IP_DST | \
RXH_L4_B_0_1 | RXH_L4_B_2_3)
#define dpaa2_eth_hash_enabled(priv) \
((priv)->dpni_attrs.num_queues > 1)
/* Required by struct dpni_rx_tc_dist_cfg::key_cfg_iova */
#define DPAA2_CLASSIFIER_DMA_SIZE 256
extern const struct ethtool_ops dpaa2_ethtool_ops;
extern int dpaa2_phc_index;
static inline int dpaa2_eth_cmp_dpni_ver(struct dpaa2_eth_priv *priv,
u16 ver_major, u16 ver_minor)
{
if (priv->dpni_ver_major == ver_major)
return priv->dpni_ver_minor - ver_minor;
return priv->dpni_ver_major - ver_major;
}
/* Hardware only sees DPAA2_ETH_RX_BUF_SIZE, but the skb built around
* the buffer also needs space for its shared info struct, and we need
* to allocate enough to accommodate hardware alignment restrictions
*/
static inline unsigned int dpaa2_eth_buf_raw_size(struct dpaa2_eth_priv *priv)
{
return DPAA2_ETH_SKB_SIZE + priv->rx_buf_align;
}
static inline
unsigned int dpaa2_eth_needed_headroom(struct dpaa2_eth_priv *priv,
struct sk_buff *skb)
{
unsigned int headroom = DPAA2_ETH_SWA_SIZE;
/* For non-linear skbs we have no headroom requirement, as we build a
* SG frame with a newly allocated SGT buffer
*/
if (skb_is_nonlinear(skb))
return 0;
/* If we have Tx timestamping, need 128B hardware annotation */
if (priv->tx_tstamp && skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP)
headroom += DPAA2_ETH_TX_HWA_SIZE;
return headroom;
}
/* Extra headroom space requested to hardware, in order to make sure there's
* no realloc'ing in forwarding scenarios
*/
static inline unsigned int dpaa2_eth_rx_head_room(struct dpaa2_eth_priv *priv)
{
return priv->tx_data_offset + DPAA2_ETH_TX_BUF_ALIGN -
DPAA2_ETH_RX_HWA_SIZE;
}
static int dpaa2_eth_queue_count(struct dpaa2_eth_priv *priv)
{
return priv->dpni_attrs.num_queues;
}
#endif /* __DPAA2_H */
