WSL2-Linux-Kernel/drivers/net/ethernet/freescale/gianfar.h

1379 строки
43 KiB
C
Исходник Обычный вид История

/*
* drivers/net/ethernet/freescale/gianfar.h
*
* Gianfar Ethernet Driver
* Driver for FEC on MPC8540 and TSEC on MPC8540/MPC8560
* Based on 8260_io/fcc_enet.c
*
* Author: Andy Fleming
* Maintainer: Kumar Gala
* Modifier: Sandeep Gopalpet <sandeep.kumar@freescale.com>
*
* Copyright 2002-2009, 2011-2013 Freescale Semiconductor, Inc.
*
* This program is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation; either version 2 of the License, or (at your
* option) any later version.
*
* Still left to do:
* -Add support for module parameters
* -Add patch for ethtool phys id
*/
#ifndef __GIANFAR_H
#define __GIANFAR_H
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/slab.h>
#include <linux/interrupt.h>
#include <linux/delay.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/skbuff.h>
#include <linux/spinlock.h>
#include <linux/mm.h>
#include <linux/mii.h>
#include <linux/phy.h>
#include <asm/io.h>
#include <asm/irq.h>
#include <linux/uaccess.h>
#include <linux/module.h>
#include <linux/crc32.h>
#include <linux/workqueue.h>
#include <linux/ethtool.h>
struct ethtool_flow_spec_container {
struct ethtool_rx_flow_spec fs;
struct list_head list;
};
struct ethtool_rx_list {
struct list_head list;
unsigned int count;
};
/* The maximum number of packets to be handled in one call of gfar_poll */
#define GFAR_DEV_WEIGHT 64
/* Length for FCB */
#define GMAC_FCB_LEN 8
/* Length for TxPAL */
#define GMAC_TXPAL_LEN 16
/* Default padding amount */
#define DEFAULT_PADDING 2
/* Number of bytes to align the rx bufs to */
#define RXBUF_ALIGNMENT 64
#define PHY_INIT_TIMEOUT 100000
#define DRV_NAME "gfar-enet"
extern const char gfar_driver_version[];
/* MAXIMUM NUMBER OF QUEUES SUPPORTED */
#define MAX_TX_QS 0x8
#define MAX_RX_QS 0x8
/* MAXIMUM NUMBER OF GROUPS SUPPORTED */
#define MAXGROUPS 0x2
/* These need to be powers of 2 for this driver */
#define DEFAULT_TX_RING_SIZE 256
#define DEFAULT_RX_RING_SIZE 256
#define GFAR_RX_BUFF_ALLOC 16
#define GFAR_RX_MAX_RING_SIZE 256
#define GFAR_TX_MAX_RING_SIZE 256
#define GFAR_MAX_FIFO_THRESHOLD 511
#define GFAR_MAX_FIFO_STARVE 511
#define GFAR_MAX_FIFO_STARVE_OFF 511
#define FBTHR_SHIFT 24
#define DEFAULT_RX_LFC_THR 16
#define DEFAULT_LFC_PTVVAL 4
/* prevent fragmenation by HW in DSA environments */
#define GFAR_RXB_SIZE roundup(1536 + 8, 64)
#define GFAR_SKBFRAG_SIZE (RXBUF_ALIGNMENT + GFAR_RXB_SIZE \
+ SKB_DATA_ALIGN(sizeof(struct skb_shared_info)))
#define GFAR_RXB_TRUESIZE 2048
#define TX_RING_MOD_MASK(size) (size-1)
#define RX_RING_MOD_MASK(size) (size-1)
#define GFAR_JUMBO_FRAME_SIZE 9600
#define DEFAULT_FIFO_TX_THR 0x100
#define DEFAULT_FIFO_TX_STARVE 0x40
#define DEFAULT_FIFO_TX_STARVE_OFF 0x80
#define DEFAULT_BD_STASH 1
#define DEFAULT_STASH_LENGTH 96
#define DEFAULT_STASH_INDEX 0
/* The number of Exact Match registers */
#define GFAR_EM_NUM 15
/* Latency of interface clock in nanoseconds */
/* Interface clock latency , in this case, means the
* time described by a value of 1 in the interrupt
* coalescing registers' time fields. Since those fields
* refer to the time it takes for 64 clocks to pass, the
* latencies are as such:
* GBIT = 125MHz => 8ns/clock => 8*64 ns / tick
* 100 = 25 MHz => 40ns/clock => 40*64 ns / tick
* 10 = 2.5 MHz => 400ns/clock => 400*64 ns / tick
*/
#define GFAR_GBIT_TIME 512
#define GFAR_100_TIME 2560
#define GFAR_10_TIME 25600
#define DEFAULT_TX_COALESCE 1
#define DEFAULT_TXCOUNT 16
#define DEFAULT_TXTIME 21
#define DEFAULT_RXTIME 21
#define DEFAULT_RX_COALESCE 0
#define DEFAULT_RXCOUNT 0
#define GFAR_SUPPORTED (SUPPORTED_10baseT_Half \
| SUPPORTED_10baseT_Full \
| SUPPORTED_100baseT_Half \
| SUPPORTED_100baseT_Full \
| SUPPORTED_Autoneg \
| SUPPORTED_MII)
#define GFAR_SUPPORTED_GBIT SUPPORTED_1000baseT_Full
/* TBI register addresses */
#define MII_TBICON 0x11
/* TBICON register bit fields */
#define TBICON_CLK_SELECT 0x0020
/* MAC register bits */
#define MACCFG1_SOFT_RESET 0x80000000
#define MACCFG1_RESET_RX_MC 0x00080000
#define MACCFG1_RESET_TX_MC 0x00040000
#define MACCFG1_RESET_RX_FUN 0x00020000
#define MACCFG1_RESET_TX_FUN 0x00010000
#define MACCFG1_LOOPBACK 0x00000100
#define MACCFG1_RX_FLOW 0x00000020
#define MACCFG1_TX_FLOW 0x00000010
#define MACCFG1_SYNCD_RX_EN 0x00000008
#define MACCFG1_RX_EN 0x00000004
#define MACCFG1_SYNCD_TX_EN 0x00000002
#define MACCFG1_TX_EN 0x00000001
#define MACCFG2_INIT_SETTINGS 0x00007205
#define MACCFG2_FULL_DUPLEX 0x00000001
#define MACCFG2_IF 0x00000300
#define MACCFG2_MII 0x00000100
#define MACCFG2_GMII 0x00000200
#define MACCFG2_HUGEFRAME 0x00000020
#define MACCFG2_LENGTHCHECK 0x00000010
#define MACCFG2_MPEN 0x00000008
#define ECNTRL_FIFM 0x00008000
#define ECNTRL_INIT_SETTINGS 0x00001000
#define ECNTRL_TBI_MODE 0x00000020
#define ECNTRL_REDUCED_MODE 0x00000010
#define ECNTRL_R100 0x00000008
#define ECNTRL_REDUCED_MII_MODE 0x00000004
#define ECNTRL_SGMII_MODE 0x00000002
#define MRBLR_INIT_SETTINGS DEFAULT_RX_BUFFER_SIZE
#define MINFLR_INIT_SETTINGS 0x00000040
/* Tqueue control */
#define TQUEUE_EN0 0x00008000
#define TQUEUE_EN1 0x00004000
#define TQUEUE_EN2 0x00002000
#define TQUEUE_EN3 0x00001000
#define TQUEUE_EN4 0x00000800
#define TQUEUE_EN5 0x00000400
#define TQUEUE_EN6 0x00000200
#define TQUEUE_EN7 0x00000100
#define TQUEUE_EN_ALL 0x0000FF00
#define TR03WT_WT0_MASK 0xFF000000
#define TR03WT_WT1_MASK 0x00FF0000
#define TR03WT_WT2_MASK 0x0000FF00
#define TR03WT_WT3_MASK 0x000000FF
#define TR47WT_WT4_MASK 0xFF000000
#define TR47WT_WT5_MASK 0x00FF0000
#define TR47WT_WT6_MASK 0x0000FF00
#define TR47WT_WT7_MASK 0x000000FF
/* Rqueue control */
#define RQUEUE_EX0 0x00800000
#define RQUEUE_EX1 0x00400000
#define RQUEUE_EX2 0x00200000
#define RQUEUE_EX3 0x00100000
#define RQUEUE_EX4 0x00080000
#define RQUEUE_EX5 0x00040000
#define RQUEUE_EX6 0x00020000
#define RQUEUE_EX7 0x00010000
#define RQUEUE_EX_ALL 0x00FF0000
#define RQUEUE_EN0 0x00000080
#define RQUEUE_EN1 0x00000040
#define RQUEUE_EN2 0x00000020
#define RQUEUE_EN3 0x00000010
#define RQUEUE_EN4 0x00000008
#define RQUEUE_EN5 0x00000004
#define RQUEUE_EN6 0x00000002
#define RQUEUE_EN7 0x00000001
#define RQUEUE_EN_ALL 0x000000FF
/* Init to do tx snooping for buffers and descriptors */
#define DMACTRL_INIT_SETTINGS 0x000000c3
#define DMACTRL_GRS 0x00000010
#define DMACTRL_GTS 0x00000008
#define TSTAT_CLEAR_THALT_ALL 0xFF000000
#define TSTAT_CLEAR_THALT 0x80000000
#define TSTAT_CLEAR_THALT0 0x80000000
#define TSTAT_CLEAR_THALT1 0x40000000
#define TSTAT_CLEAR_THALT2 0x20000000
#define TSTAT_CLEAR_THALT3 0x10000000
#define TSTAT_CLEAR_THALT4 0x08000000
#define TSTAT_CLEAR_THALT5 0x04000000
#define TSTAT_CLEAR_THALT6 0x02000000
#define TSTAT_CLEAR_THALT7 0x01000000
/* Interrupt coalescing macros */
#define IC_ICEN 0x80000000
#define IC_ICFT_MASK 0x1fe00000
#define IC_ICFT_SHIFT 21
#define mk_ic_icft(x) \
(((unsigned int)x << IC_ICFT_SHIFT)&IC_ICFT_MASK)
#define IC_ICTT_MASK 0x0000ffff
#define mk_ic_ictt(x) (x&IC_ICTT_MASK)
#define mk_ic_value(count, time) (IC_ICEN | \
mk_ic_icft(count) | \
mk_ic_ictt(time))
#define get_icft_value(ic) (((unsigned long)ic & IC_ICFT_MASK) >> \
IC_ICFT_SHIFT)
#define get_ictt_value(ic) ((unsigned long)ic & IC_ICTT_MASK)
#define DEFAULT_TXIC mk_ic_value(DEFAULT_TXCOUNT, DEFAULT_TXTIME)
#define DEFAULT_RXIC mk_ic_value(DEFAULT_RXCOUNT, DEFAULT_RXTIME)
#define skip_bd(bdp, stride, base, ring_size) ({ \
typeof(bdp) new_bd = (bdp) + (stride); \
(new_bd >= (base) + (ring_size)) ? (new_bd - (ring_size)) : new_bd; })
#define next_bd(bdp, base, ring_size) skip_bd(bdp, 1, base, ring_size)
#define RCTRL_TS_ENABLE 0x01000000
#define RCTRL_PAL_MASK 0x001f0000
#define RCTRL_LFC 0x00004000
#define RCTRL_VLEX 0x00002000
#define RCTRL_FILREN 0x00001000
#define RCTRL_GHTX 0x00000400
#define RCTRL_IPCSEN 0x00000200
#define RCTRL_TUCSEN 0x00000100
#define RCTRL_PRSDEP_MASK 0x000000c0
#define RCTRL_PRSDEP_INIT 0x000000c0
#define RCTRL_PRSFM 0x00000020
#define RCTRL_PROM 0x00000008
#define RCTRL_EMEN 0x00000002
#define RCTRL_REQ_PARSER (RCTRL_VLEX | RCTRL_IPCSEN | \
RCTRL_TUCSEN | RCTRL_FILREN)
#define RCTRL_CHECKSUMMING (RCTRL_IPCSEN | RCTRL_TUCSEN | \
RCTRL_PRSDEP_INIT)
#define RCTRL_EXTHASH (RCTRL_GHTX)
#define RCTRL_VLAN (RCTRL_PRSDEP_INIT)
#define RCTRL_PADDING(x) ((x << 16) & RCTRL_PAL_MASK)
#define RSTAT_CLEAR_RHALT 0x00800000
#define RSTAT_CLEAR_RXF0 0x00000080
#define RSTAT_RXF_MASK 0x000000ff
#define TCTRL_IPCSEN 0x00004000
#define TCTRL_TUCSEN 0x00002000
#define TCTRL_VLINS 0x00001000
#define TCTRL_THDF 0x00000800
#define TCTRL_RFCPAUSE 0x00000010
#define TCTRL_TFCPAUSE 0x00000008
#define TCTRL_TXSCHED_MASK 0x00000006
#define TCTRL_TXSCHED_INIT 0x00000000
/* priority scheduling */
#define TCTRL_TXSCHED_PRIO 0x00000002
/* weighted round-robin scheduling (WRRS) */
#define TCTRL_TXSCHED_WRRS 0x00000004
/* default WRRS weight and policy setting,
* tailored to the tr03wt and tr47wt registers:
* equal weight for all Tx Qs, measured in 64byte units
*/
#define DEFAULT_WRRS_WEIGHT 0x18181818
#define TCTRL_INIT_CSUM (TCTRL_TUCSEN | TCTRL_IPCSEN)
#define IEVENT_INIT_CLEAR 0xffffffff
#define IEVENT_BABR 0x80000000
#define IEVENT_RXC 0x40000000
#define IEVENT_BSY 0x20000000
#define IEVENT_EBERR 0x10000000
#define IEVENT_MSRO 0x04000000
#define IEVENT_GTSC 0x02000000
#define IEVENT_BABT 0x01000000
#define IEVENT_TXC 0x00800000
#define IEVENT_TXE 0x00400000
#define IEVENT_TXB 0x00200000
#define IEVENT_TXF 0x00100000
#define IEVENT_LC 0x00040000
#define IEVENT_CRL 0x00020000
#define IEVENT_XFUN 0x00010000
#define IEVENT_RXB0 0x00008000
#define IEVENT_MAG 0x00000800
#define IEVENT_GRSC 0x00000100
#define IEVENT_RXF0 0x00000080
#define IEVENT_FGPI 0x00000010
#define IEVENT_FIR 0x00000008
#define IEVENT_FIQ 0x00000004
#define IEVENT_DPE 0x00000002
#define IEVENT_PERR 0x00000001
#define IEVENT_RX_MASK (IEVENT_RXB0 | IEVENT_RXF0 | IEVENT_BSY)
#define IEVENT_TX_MASK (IEVENT_TXB | IEVENT_TXF)
#define IEVENT_RTX_MASK (IEVENT_RX_MASK | IEVENT_TX_MASK)
#define IEVENT_ERR_MASK \
(IEVENT_RXC | IEVENT_BSY | IEVENT_EBERR | IEVENT_MSRO | \
IEVENT_BABT | IEVENT_TXC | IEVENT_TXE | IEVENT_LC \
| IEVENT_CRL | IEVENT_XFUN | IEVENT_DPE | IEVENT_PERR \
| IEVENT_MAG | IEVENT_BABR)
#define IMASK_INIT_CLEAR 0x00000000
#define IMASK_BABR 0x80000000
#define IMASK_RXC 0x40000000
#define IMASK_BSY 0x20000000
#define IMASK_EBERR 0x10000000
#define IMASK_MSRO 0x04000000
#define IMASK_GTSC 0x02000000
#define IMASK_BABT 0x01000000
#define IMASK_TXC 0x00800000
#define IMASK_TXEEN 0x00400000
#define IMASK_TXBEN 0x00200000
#define IMASK_TXFEN 0x00100000
#define IMASK_LC 0x00040000
#define IMASK_CRL 0x00020000
#define IMASK_XFUN 0x00010000
#define IMASK_RXB0 0x00008000
#define IMASK_MAG 0x00000800
#define IMASK_GRSC 0x00000100
#define IMASK_RXFEN0 0x00000080
#define IMASK_FGPI 0x00000010
#define IMASK_FIR 0x00000008
#define IMASK_FIQ 0x00000004
#define IMASK_DPE 0x00000002
#define IMASK_PERR 0x00000001
#define IMASK_DEFAULT (IMASK_TXEEN | IMASK_TXFEN | IMASK_TXBEN | \
IMASK_RXFEN0 | IMASK_BSY | IMASK_EBERR | IMASK_BABR | \
IMASK_XFUN | IMASK_RXC | IMASK_BABT | IMASK_DPE \
| IMASK_PERR)
gianfar: Separate out the Tx interrupt handling (Tx NAPI) There are some concurrency issues on devices w/ 2 CPUs related to the handling of Rx and Tx interrupts. eTSEC has separate interrupt lines for Rx and Tx but a single imask register to mask these interrupts and a single NAPI instance to handle both Rx and Tx work. As a result, the Rx and Tx ISRs are identical, both are invoking gfar_schedule_cleanup(), however both handlers can be entered at the same time when the Rx and Tx interrupts are taken by different CPUs. In this case spurrious interrupts (SPU) show up (in /proc/interrupts) indicating a concurrency issue. Also, Tx overruns followed by Tx timeout have been observed under heavy Tx traffic load. To address these issues, the schedule cleanup ISR part has been changed to handle the Rx and Tx interrupts independently. The patch adds a separate NAPI poll routine for Tx cleanup to be triggerred independently by the Tx confirmation interrupts only. Existing poll functions are modified to handle only the Rx path processing. The Tx poll routine does not need a budget, since Tx processing doesn't consume NAPI budget, and hence it is registered with minimum NAPI weight. NAPI scheduling does not require locking since there are different NAPI instances between the Rx and Tx confirmation paths now. So, the patch fixes the occurence of spurrious Rx/Tx interrupts. Tx overruns also occur less frequently now. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:45 +04:00
#define IMASK_RX_DEFAULT (IMASK_RXFEN0 | IMASK_BSY)
#define IMASK_TX_DEFAULT (IMASK_TXFEN | IMASK_TXBEN)
#define IMASK_RX_DISABLED ((~(IMASK_RX_DEFAULT)) & IMASK_DEFAULT)
#define IMASK_TX_DISABLED ((~(IMASK_TX_DEFAULT)) & IMASK_DEFAULT)
/* Fifo management */
#define FIFO_TX_THR_MASK 0x01ff
#define FIFO_TX_STARVE_MASK 0x01ff
#define FIFO_TX_STARVE_OFF_MASK 0x01ff
/* Attribute fields */
/* This enables rx snooping for buffers and descriptors */
#define ATTR_BDSTASH 0x00000800
#define ATTR_BUFSTASH 0x00004000
#define ATTR_SNOOPING 0x000000c0
#define ATTR_INIT_SETTINGS ATTR_SNOOPING
#define ATTRELI_INIT_SETTINGS 0x0
#define ATTRELI_EL_MASK 0x3fff0000
#define ATTRELI_EL(x) (x << 16)
#define ATTRELI_EI_MASK 0x00003fff
#define ATTRELI_EI(x) (x)
#define BD_LFLAG(flags) ((flags) << 16)
#define BD_LENGTH_MASK 0x0000ffff
#define FPR_FILER_MASK 0xFFFFFFFF
#define MAX_FILER_IDX 0xFF
/* This default RIR value directly corresponds
* to the 3-bit hash value generated */
gianfar: Use Single-Queue polling for "fsl,etsec2" For the "fsl,etsec2" compatible models the driver currently supports 8 Tx and Rx DMA rings (aka HW queues). However, there are only 2 pairs of Rx/Tx interrupt lines, as these controllers are integrated in low power SoCs with 2 CPUs at most. As a result, there are at most 2 NAPI instances that have to service multiple Tx and Rx queues for these devices. This complicates the NAPI polling routine having to iterate over the mutiple Rx/Tx queues hooked to the same interrupt lines. And there's also an overhead at HW level, as the controller needs to service all the 8 Tx rings in a round robin manner. The combined overhead shows up for multi parallel Tx flows transmitted by the kernel stack, when the driver usually starts returning NETDEV_TX_BUSY leading to NETDEV WATCHDOG Tx timeout triggering if the Tx path is congested for too long. As an alternative, this patch makes the driver support only one Tx/Rx DMA ring per NAPI instance (per interrupt group or pair of Tx/Rx interrupt lines) by default. The simplified single queue polling routine (gfar_poll_sq) will be the default napi poll routine for the etsec2 devices too. Some adjustments needed to be made to link the Tx/Rx HW queues with each NAPI instance (2 in this case). The gfar_poll_sq() is already successfully used by older SQ_SG_MODE (single interrupt group) controllers. This patch fixes Tx timeout triggering under heavy Tx traffic load (i.e. iperf -c -P 8) for the "fsl,etsec2" (currently the only MQ_MG_MODE devices). There's also a significant memory footprint reduction by supporting 2 Rx/Tx DMA rings (at most), instead of 8, for these devices. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:46 +04:00
#define DEFAULT_8RXQ_RIR0 0x05397700
/* Map even hash values to Q0, and odd ones to Q1 */
#define DEFAULT_2RXQ_RIR0 0x04104100
/* RQFCR register bits */
#define RQFCR_GPI 0x80000000
#define RQFCR_HASHTBL_Q 0x00000000
#define RQFCR_HASHTBL_0 0x00020000
#define RQFCR_HASHTBL_1 0x00040000
#define RQFCR_HASHTBL_2 0x00060000
#define RQFCR_HASHTBL_3 0x00080000
#define RQFCR_HASH 0x00010000
#define RQFCR_QUEUE 0x0000FC00
#define RQFCR_CLE 0x00000200
#define RQFCR_RJE 0x00000100
#define RQFCR_AND 0x00000080
#define RQFCR_CMP_EXACT 0x00000000
#define RQFCR_CMP_MATCH 0x00000020
#define RQFCR_CMP_NOEXACT 0x00000040
#define RQFCR_CMP_NOMATCH 0x00000060
/* RQFCR PID values */
#define RQFCR_PID_MASK 0x00000000
#define RQFCR_PID_PARSE 0x00000001
#define RQFCR_PID_ARB 0x00000002
#define RQFCR_PID_DAH 0x00000003
#define RQFCR_PID_DAL 0x00000004
#define RQFCR_PID_SAH 0x00000005
#define RQFCR_PID_SAL 0x00000006
#define RQFCR_PID_ETY 0x00000007
#define RQFCR_PID_VID 0x00000008
#define RQFCR_PID_PRI 0x00000009
#define RQFCR_PID_TOS 0x0000000A
#define RQFCR_PID_L4P 0x0000000B
#define RQFCR_PID_DIA 0x0000000C
#define RQFCR_PID_SIA 0x0000000D
#define RQFCR_PID_DPT 0x0000000E
#define RQFCR_PID_SPT 0x0000000F
/* RQFPR when PID is 0x0001 */
#define RQFPR_HDR_GE_512 0x00200000
#define RQFPR_LERR 0x00100000
#define RQFPR_RAR 0x00080000
#define RQFPR_RARQ 0x00040000
#define RQFPR_AR 0x00020000
#define RQFPR_ARQ 0x00010000
#define RQFPR_EBC 0x00008000
#define RQFPR_VLN 0x00004000
#define RQFPR_CFI 0x00002000
#define RQFPR_JUM 0x00001000
#define RQFPR_IPF 0x00000800
#define RQFPR_FIF 0x00000400
#define RQFPR_IPV4 0x00000200
#define RQFPR_IPV6 0x00000100
#define RQFPR_ICC 0x00000080
#define RQFPR_ICV 0x00000040
#define RQFPR_TCP 0x00000020
#define RQFPR_UDP 0x00000010
#define RQFPR_TUC 0x00000008
#define RQFPR_TUV 0x00000004
#define RQFPR_PER 0x00000002
#define RQFPR_EER 0x00000001
/* TxBD status field bits */
#define TXBD_READY 0x8000
#define TXBD_PADCRC 0x4000
#define TXBD_WRAP 0x2000
#define TXBD_INTERRUPT 0x1000
#define TXBD_LAST 0x0800
#define TXBD_CRC 0x0400
#define TXBD_DEF 0x0200
#define TXBD_HUGEFRAME 0x0080
#define TXBD_LATECOLLISION 0x0080
#define TXBD_RETRYLIMIT 0x0040
#define TXBD_RETRYCOUNTMASK 0x003c
#define TXBD_UNDERRUN 0x0002
#define TXBD_TOE 0x0002
/* Tx FCB param bits */
#define TXFCB_VLN 0x80
#define TXFCB_IP 0x40
#define TXFCB_IP6 0x20
#define TXFCB_TUP 0x10
#define TXFCB_UDP 0x08
#define TXFCB_CIP 0x04
#define TXFCB_CTU 0x02
#define TXFCB_NPH 0x01
#define TXFCB_DEFAULT (TXFCB_IP|TXFCB_TUP|TXFCB_CTU|TXFCB_NPH)
/* RxBD status field bits */
#define RXBD_EMPTY 0x8000
#define RXBD_RO1 0x4000
#define RXBD_WRAP 0x2000
#define RXBD_INTERRUPT 0x1000
#define RXBD_LAST 0x0800
#define RXBD_FIRST 0x0400
#define RXBD_MISS 0x0100
#define RXBD_BROADCAST 0x0080
#define RXBD_MULTICAST 0x0040
#define RXBD_LARGE 0x0020
#define RXBD_NONOCTET 0x0010
#define RXBD_SHORT 0x0008
#define RXBD_CRCERR 0x0004
#define RXBD_OVERRUN 0x0002
#define RXBD_TRUNCATED 0x0001
#define RXBD_STATS 0x01ff
#define RXBD_ERR (RXBD_LARGE | RXBD_SHORT | RXBD_NONOCTET \
| RXBD_CRCERR | RXBD_OVERRUN \
| RXBD_TRUNCATED)
/* Rx FCB status field bits */
#define RXFCB_VLN 0x8000
#define RXFCB_IP 0x4000
#define RXFCB_IP6 0x2000
#define RXFCB_TUP 0x1000
#define RXFCB_CIP 0x0800
#define RXFCB_CTU 0x0400
#define RXFCB_EIP 0x0200
#define RXFCB_ETU 0x0100
#define RXFCB_CSUM_MASK 0x0f00
#define RXFCB_PERR_MASK 0x000c
#define RXFCB_PERR_BADL3 0x0008
#define GFAR_INT_NAME_MAX (IFNAMSIZ + 6) /* '_g#_xx' */
#define GFAR_WOL_MAGIC 0x00000001
#define GFAR_WOL_FILER_UCAST 0x00000002
struct txbd8
{
union {
struct {
__be16 status; /* Status Fields */
__be16 length; /* Buffer length */
};
__be32 lstatus;
};
__be32 bufPtr; /* Buffer Pointer */
};
struct txfcb {
u8 flags;
u8 ptp; /* Flag to enable tx timestamping */
u8 l4os; /* Level 4 Header Offset */
u8 l3os; /* Level 3 Header Offset */
__be16 phcs; /* Pseudo-header Checksum */
__be16 vlctl; /* VLAN control word */
};
struct rxbd8
{
union {
struct {
__be16 status; /* Status Fields */
__be16 length; /* Buffer Length */
};
__be32 lstatus;
};
__be32 bufPtr; /* Buffer Pointer */
};
struct rxfcb {
__be16 flags;
u8 rq; /* Receive Queue index */
u8 pro; /* Layer 4 Protocol */
u16 reserved;
__be16 vlctl; /* VLAN control word */
};
struct gianfar_skb_cb {
unsigned int bytes_sent; /* bytes-on-wire (i.e. no FCB) */
};
#define GFAR_CB(skb) ((struct gianfar_skb_cb *)((skb)->cb))
struct rmon_mib
{
u32 tr64; /* 0x.680 - Transmit and Receive 64-byte Frame Counter */
u32 tr127; /* 0x.684 - Transmit and Receive 65-127 byte Frame Counter */
u32 tr255; /* 0x.688 - Transmit and Receive 128-255 byte Frame Counter */
u32 tr511; /* 0x.68c - Transmit and Receive 256-511 byte Frame Counter */
u32 tr1k; /* 0x.690 - Transmit and Receive 512-1023 byte Frame Counter */
u32 trmax; /* 0x.694 - Transmit and Receive 1024-1518 byte Frame Counter */
u32 trmgv; /* 0x.698 - Transmit and Receive 1519-1522 byte Good VLAN Frame */
u32 rbyt; /* 0x.69c - Receive Byte Counter */
u32 rpkt; /* 0x.6a0 - Receive Packet Counter */
u32 rfcs; /* 0x.6a4 - Receive FCS Error Counter */
u32 rmca; /* 0x.6a8 - Receive Multicast Packet Counter */
u32 rbca; /* 0x.6ac - Receive Broadcast Packet Counter */
u32 rxcf; /* 0x.6b0 - Receive Control Frame Packet Counter */
u32 rxpf; /* 0x.6b4 - Receive Pause Frame Packet Counter */
u32 rxuo; /* 0x.6b8 - Receive Unknown OP Code Counter */
u32 raln; /* 0x.6bc - Receive Alignment Error Counter */
u32 rflr; /* 0x.6c0 - Receive Frame Length Error Counter */
u32 rcde; /* 0x.6c4 - Receive Code Error Counter */
u32 rcse; /* 0x.6c8 - Receive Carrier Sense Error Counter */
u32 rund; /* 0x.6cc - Receive Undersize Packet Counter */
u32 rovr; /* 0x.6d0 - Receive Oversize Packet Counter */
u32 rfrg; /* 0x.6d4 - Receive Fragments Counter */
u32 rjbr; /* 0x.6d8 - Receive Jabber Counter */
u32 rdrp; /* 0x.6dc - Receive Drop Counter */
u32 tbyt; /* 0x.6e0 - Transmit Byte Counter Counter */
u32 tpkt; /* 0x.6e4 - Transmit Packet Counter */
u32 tmca; /* 0x.6e8 - Transmit Multicast Packet Counter */
u32 tbca; /* 0x.6ec - Transmit Broadcast Packet Counter */
u32 txpf; /* 0x.6f0 - Transmit Pause Control Frame Counter */
u32 tdfr; /* 0x.6f4 - Transmit Deferral Packet Counter */
u32 tedf; /* 0x.6f8 - Transmit Excessive Deferral Packet Counter */
u32 tscl; /* 0x.6fc - Transmit Single Collision Packet Counter */
u32 tmcl; /* 0x.700 - Transmit Multiple Collision Packet Counter */
u32 tlcl; /* 0x.704 - Transmit Late Collision Packet Counter */
u32 txcl; /* 0x.708 - Transmit Excessive Collision Packet Counter */
u32 tncl; /* 0x.70c - Transmit Total Collision Counter */
u8 res1[4];
u32 tdrp; /* 0x.714 - Transmit Drop Frame Counter */
u32 tjbr; /* 0x.718 - Transmit Jabber Frame Counter */
u32 tfcs; /* 0x.71c - Transmit FCS Error Counter */
u32 txcf; /* 0x.720 - Transmit Control Frame Counter */
u32 tovr; /* 0x.724 - Transmit Oversize Frame Counter */
u32 tund; /* 0x.728 - Transmit Undersize Frame Counter */
u32 tfrg; /* 0x.72c - Transmit Fragments Frame Counter */
u32 car1; /* 0x.730 - Carry Register One */
u32 car2; /* 0x.734 - Carry Register Two */
u32 cam1; /* 0x.738 - Carry Mask Register One */
u32 cam2; /* 0x.73c - Carry Mask Register Two */
};
struct gfar_extra_stats {
atomic64_t rx_alloc_err;
atomic64_t rx_large;
atomic64_t rx_short;
atomic64_t rx_nonoctet;
atomic64_t rx_crcerr;
atomic64_t rx_overrun;
atomic64_t rx_bsy;
atomic64_t rx_babr;
atomic64_t rx_trunc;
atomic64_t eberr;
atomic64_t tx_babt;
atomic64_t tx_underrun;
atomic64_t tx_timeout;
};
#define GFAR_RMON_LEN ((sizeof(struct rmon_mib) - 16)/sizeof(u32))
#define GFAR_EXTRA_STATS_LEN \
(sizeof(struct gfar_extra_stats)/sizeof(atomic64_t))
/* Number of stats exported via ethtool */
#define GFAR_STATS_LEN (GFAR_RMON_LEN + GFAR_EXTRA_STATS_LEN)
struct gfar {
u32 tsec_id; /* 0x.000 - Controller ID register */
u32 tsec_id2; /* 0x.004 - Controller ID2 register */
u8 res1[8];
u32 ievent; /* 0x.010 - Interrupt Event Register */
u32 imask; /* 0x.014 - Interrupt Mask Register */
u32 edis; /* 0x.018 - Error Disabled Register */
u32 emapg; /* 0x.01c - Group Error mapping register */
u32 ecntrl; /* 0x.020 - Ethernet Control Register */
u32 minflr; /* 0x.024 - Minimum Frame Length Register */
u32 ptv; /* 0x.028 - Pause Time Value Register */
u32 dmactrl; /* 0x.02c - DMA Control Register */
u32 tbipa; /* 0x.030 - TBI PHY Address Register */
u8 res2[28];
u32 fifo_rx_pause; /* 0x.050 - FIFO receive pause start threshold
register */
u32 fifo_rx_pause_shutoff; /* x.054 - FIFO receive starve shutoff
register */
u32 fifo_rx_alarm; /* 0x.058 - FIFO receive alarm start threshold
register */
u32 fifo_rx_alarm_shutoff; /*0x.05c - FIFO receive alarm starve
shutoff register */
u8 res3[44];
u32 fifo_tx_thr; /* 0x.08c - FIFO transmit threshold register */
u8 res4[8];
u32 fifo_tx_starve; /* 0x.098 - FIFO transmit starve register */
u32 fifo_tx_starve_shutoff; /* 0x.09c - FIFO transmit starve shutoff register */
u8 res5[96];
u32 tctrl; /* 0x.100 - Transmit Control Register */
u32 tstat; /* 0x.104 - Transmit Status Register */
u32 dfvlan; /* 0x.108 - Default VLAN Control word */
u32 tbdlen; /* 0x.10c - Transmit Buffer Descriptor Data Length Register */
u32 txic; /* 0x.110 - Transmit Interrupt Coalescing Configuration Register */
u32 tqueue; /* 0x.114 - Transmit queue control register */
u8 res7[40];
u32 tr03wt; /* 0x.140 - TxBD Rings 0-3 round-robin weightings */
u32 tr47wt; /* 0x.144 - TxBD Rings 4-7 round-robin weightings */
u8 res8[52];
u32 tbdbph; /* 0x.17c - Tx data buffer pointer high */
u8 res9a[4];
u32 tbptr0; /* 0x.184 - TxBD Pointer for ring 0 */
u8 res9b[4];
u32 tbptr1; /* 0x.18c - TxBD Pointer for ring 1 */
u8 res9c[4];
u32 tbptr2; /* 0x.194 - TxBD Pointer for ring 2 */
u8 res9d[4];
u32 tbptr3; /* 0x.19c - TxBD Pointer for ring 3 */
u8 res9e[4];
u32 tbptr4; /* 0x.1a4 - TxBD Pointer for ring 4 */
u8 res9f[4];
u32 tbptr5; /* 0x.1ac - TxBD Pointer for ring 5 */
u8 res9g[4];
u32 tbptr6; /* 0x.1b4 - TxBD Pointer for ring 6 */
u8 res9h[4];
u32 tbptr7; /* 0x.1bc - TxBD Pointer for ring 7 */
u8 res9[64];
u32 tbaseh; /* 0x.200 - TxBD base address high */
u32 tbase0; /* 0x.204 - TxBD Base Address of ring 0 */
u8 res10a[4];
u32 tbase1; /* 0x.20c - TxBD Base Address of ring 1 */
u8 res10b[4];
u32 tbase2; /* 0x.214 - TxBD Base Address of ring 2 */
u8 res10c[4];
u32 tbase3; /* 0x.21c - TxBD Base Address of ring 3 */
u8 res10d[4];
u32 tbase4; /* 0x.224 - TxBD Base Address of ring 4 */
u8 res10e[4];
u32 tbase5; /* 0x.22c - TxBD Base Address of ring 5 */
u8 res10f[4];
u32 tbase6; /* 0x.234 - TxBD Base Address of ring 6 */
u8 res10g[4];
u32 tbase7; /* 0x.23c - TxBD Base Address of ring 7 */
u8 res10[192];
u32 rctrl; /* 0x.300 - Receive Control Register */
u32 rstat; /* 0x.304 - Receive Status Register */
u8 res12[8];
u32 rxic; /* 0x.310 - Receive Interrupt Coalescing Configuration Register */
u32 rqueue; /* 0x.314 - Receive queue control register */
u32 rir0; /* 0x.318 - Ring mapping register 0 */
u32 rir1; /* 0x.31c - Ring mapping register 1 */
u32 rir2; /* 0x.320 - Ring mapping register 2 */
u32 rir3; /* 0x.324 - Ring mapping register 3 */
u8 res13[8];
u32 rbifx; /* 0x.330 - Receive bit field extract control register */
u32 rqfar; /* 0x.334 - Receive queue filing table address register */
u32 rqfcr; /* 0x.338 - Receive queue filing table control register */
u32 rqfpr; /* 0x.33c - Receive queue filing table property register */
u32 mrblr; /* 0x.340 - Maximum Receive Buffer Length Register */
u8 res14[56];
u32 rbdbph; /* 0x.37c - Rx data buffer pointer high */
u8 res15a[4];
u32 rbptr0; /* 0x.384 - RxBD pointer for ring 0 */
u8 res15b[4];
u32 rbptr1; /* 0x.38c - RxBD pointer for ring 1 */
u8 res15c[4];
u32 rbptr2; /* 0x.394 - RxBD pointer for ring 2 */
u8 res15d[4];
u32 rbptr3; /* 0x.39c - RxBD pointer for ring 3 */
u8 res15e[4];
u32 rbptr4; /* 0x.3a4 - RxBD pointer for ring 4 */
u8 res15f[4];
u32 rbptr5; /* 0x.3ac - RxBD pointer for ring 5 */
u8 res15g[4];
u32 rbptr6; /* 0x.3b4 - RxBD pointer for ring 6 */
u8 res15h[4];
u32 rbptr7; /* 0x.3bc - RxBD pointer for ring 7 */
u8 res16[64];
u32 rbaseh; /* 0x.400 - RxBD base address high */
u32 rbase0; /* 0x.404 - RxBD base address of ring 0 */
u8 res17a[4];
u32 rbase1; /* 0x.40c - RxBD base address of ring 1 */
u8 res17b[4];
u32 rbase2; /* 0x.414 - RxBD base address of ring 2 */
u8 res17c[4];
u32 rbase3; /* 0x.41c - RxBD base address of ring 3 */
u8 res17d[4];
u32 rbase4; /* 0x.424 - RxBD base address of ring 4 */
u8 res17e[4];
u32 rbase5; /* 0x.42c - RxBD base address of ring 5 */
u8 res17f[4];
u32 rbase6; /* 0x.434 - RxBD base address of ring 6 */
u8 res17g[4];
u32 rbase7; /* 0x.43c - RxBD base address of ring 7 */
u8 res17[192];
u32 maccfg1; /* 0x.500 - MAC Configuration 1 Register */
u32 maccfg2; /* 0x.504 - MAC Configuration 2 Register */
u32 ipgifg; /* 0x.508 - Inter Packet Gap/Inter Frame Gap Register */
u32 hafdup; /* 0x.50c - Half Duplex Register */
u32 maxfrm; /* 0x.510 - Maximum Frame Length Register */
u8 res18[12];
u8 gfar_mii_regs[24]; /* See gianfar_phy.h */
u32 ifctrl; /* 0x.538 - Interface control register */
u32 ifstat; /* 0x.53c - Interface Status Register */
u32 macstnaddr1; /* 0x.540 - Station Address Part 1 Register */
u32 macstnaddr2; /* 0x.544 - Station Address Part 2 Register */
u32 mac01addr1; /* 0x.548 - MAC exact match address 1, part 1 */
u32 mac01addr2; /* 0x.54c - MAC exact match address 1, part 2 */
u32 mac02addr1; /* 0x.550 - MAC exact match address 2, part 1 */
u32 mac02addr2; /* 0x.554 - MAC exact match address 2, part 2 */
u32 mac03addr1; /* 0x.558 - MAC exact match address 3, part 1 */
u32 mac03addr2; /* 0x.55c - MAC exact match address 3, part 2 */
u32 mac04addr1; /* 0x.560 - MAC exact match address 4, part 1 */
u32 mac04addr2; /* 0x.564 - MAC exact match address 4, part 2 */
u32 mac05addr1; /* 0x.568 - MAC exact match address 5, part 1 */
u32 mac05addr2; /* 0x.56c - MAC exact match address 5, part 2 */
u32 mac06addr1; /* 0x.570 - MAC exact match address 6, part 1 */
u32 mac06addr2; /* 0x.574 - MAC exact match address 6, part 2 */
u32 mac07addr1; /* 0x.578 - MAC exact match address 7, part 1 */
u32 mac07addr2; /* 0x.57c - MAC exact match address 7, part 2 */
u32 mac08addr1; /* 0x.580 - MAC exact match address 8, part 1 */
u32 mac08addr2; /* 0x.584 - MAC exact match address 8, part 2 */
u32 mac09addr1; /* 0x.588 - MAC exact match address 9, part 1 */
u32 mac09addr2; /* 0x.58c - MAC exact match address 9, part 2 */
u32 mac10addr1; /* 0x.590 - MAC exact match address 10, part 1*/
u32 mac10addr2; /* 0x.594 - MAC exact match address 10, part 2*/
u32 mac11addr1; /* 0x.598 - MAC exact match address 11, part 1*/
u32 mac11addr2; /* 0x.59c - MAC exact match address 11, part 2*/
u32 mac12addr1; /* 0x.5a0 - MAC exact match address 12, part 1*/
u32 mac12addr2; /* 0x.5a4 - MAC exact match address 12, part 2*/
u32 mac13addr1; /* 0x.5a8 - MAC exact match address 13, part 1*/
u32 mac13addr2; /* 0x.5ac - MAC exact match address 13, part 2*/
u32 mac14addr1; /* 0x.5b0 - MAC exact match address 14, part 1*/
u32 mac14addr2; /* 0x.5b4 - MAC exact match address 14, part 2*/
u32 mac15addr1; /* 0x.5b8 - MAC exact match address 15, part 1*/
u32 mac15addr2; /* 0x.5bc - MAC exact match address 15, part 2*/
u8 res20[192];
struct rmon_mib rmon; /* 0x.680-0x.73c */
u32 rrej; /* 0x.740 - Receive filer rejected packet counter */
u8 res21[188];
u32 igaddr0; /* 0x.800 - Indivdual/Group address register 0*/
u32 igaddr1; /* 0x.804 - Indivdual/Group address register 1*/
u32 igaddr2; /* 0x.808 - Indivdual/Group address register 2*/
u32 igaddr3; /* 0x.80c - Indivdual/Group address register 3*/
u32 igaddr4; /* 0x.810 - Indivdual/Group address register 4*/
u32 igaddr5; /* 0x.814 - Indivdual/Group address register 5*/
u32 igaddr6; /* 0x.818 - Indivdual/Group address register 6*/
u32 igaddr7; /* 0x.81c - Indivdual/Group address register 7*/
u8 res22[96];
u32 gaddr0; /* 0x.880 - Group address register 0 */
u32 gaddr1; /* 0x.884 - Group address register 1 */
u32 gaddr2; /* 0x.888 - Group address register 2 */
u32 gaddr3; /* 0x.88c - Group address register 3 */
u32 gaddr4; /* 0x.890 - Group address register 4 */
u32 gaddr5; /* 0x.894 - Group address register 5 */
u32 gaddr6; /* 0x.898 - Group address register 6 */
u32 gaddr7; /* 0x.89c - Group address register 7 */
u8 res23a[352];
u32 fifocfg; /* 0x.a00 - FIFO interface config register */
u8 res23b[252];
u8 res23c[248];
u32 attr; /* 0x.bf8 - Attributes Register */
u32 attreli; /* 0x.bfc - Attributes Extract Length and Extract Index Register */
u32 rqprm0; /* 0x.c00 - Receive queue parameters register 0 */
u32 rqprm1; /* 0x.c04 - Receive queue parameters register 1 */
u32 rqprm2; /* 0x.c08 - Receive queue parameters register 2 */
u32 rqprm3; /* 0x.c0c - Receive queue parameters register 3 */
u32 rqprm4; /* 0x.c10 - Receive queue parameters register 4 */
u32 rqprm5; /* 0x.c14 - Receive queue parameters register 5 */
u32 rqprm6; /* 0x.c18 - Receive queue parameters register 6 */
u32 rqprm7; /* 0x.c1c - Receive queue parameters register 7 */
u8 res24[36];
u32 rfbptr0; /* 0x.c44 - Last free RxBD pointer for ring 0 */
u8 res24a[4];
u32 rfbptr1; /* 0x.c4c - Last free RxBD pointer for ring 1 */
u8 res24b[4];
u32 rfbptr2; /* 0x.c54 - Last free RxBD pointer for ring 2 */
u8 res24c[4];
u32 rfbptr3; /* 0x.c5c - Last free RxBD pointer for ring 3 */
u8 res24d[4];
u32 rfbptr4; /* 0x.c64 - Last free RxBD pointer for ring 4 */
u8 res24e[4];
u32 rfbptr5; /* 0x.c6c - Last free RxBD pointer for ring 5 */
u8 res24f[4];
u32 rfbptr6; /* 0x.c74 - Last free RxBD pointer for ring 6 */
u8 res24g[4];
u32 rfbptr7; /* 0x.c7c - Last free RxBD pointer for ring 7 */
u8 res24h[4];
u8 res24x[556];
u32 isrg0; /* 0x.eb0 - Interrupt steering group 0 register */
u32 isrg1; /* 0x.eb4 - Interrupt steering group 1 register */
u32 isrg2; /* 0x.eb8 - Interrupt steering group 2 register */
u32 isrg3; /* 0x.ebc - Interrupt steering group 3 register */
u8 res25[16];
u32 rxic0; /* 0x.ed0 - Ring 0 Rx interrupt coalescing */
u32 rxic1; /* 0x.ed4 - Ring 1 Rx interrupt coalescing */
u32 rxic2; /* 0x.ed8 - Ring 2 Rx interrupt coalescing */
u32 rxic3; /* 0x.edc - Ring 3 Rx interrupt coalescing */
u32 rxic4; /* 0x.ee0 - Ring 4 Rx interrupt coalescing */
u32 rxic5; /* 0x.ee4 - Ring 5 Rx interrupt coalescing */
u32 rxic6; /* 0x.ee8 - Ring 6 Rx interrupt coalescing */
u32 rxic7; /* 0x.eec - Ring 7 Rx interrupt coalescing */
u8 res26[32];
u32 txic0; /* 0x.f10 - Ring 0 Tx interrupt coalescing */
u32 txic1; /* 0x.f14 - Ring 1 Tx interrupt coalescing */
u32 txic2; /* 0x.f18 - Ring 2 Tx interrupt coalescing */
u32 txic3; /* 0x.f1c - Ring 3 Tx interrupt coalescing */
u32 txic4; /* 0x.f20 - Ring 4 Tx interrupt coalescing */
u32 txic5; /* 0x.f24 - Ring 5 Tx interrupt coalescing */
u32 txic6; /* 0x.f28 - Ring 6 Tx interrupt coalescing */
u32 txic7; /* 0x.f2c - Ring 7 Tx interrupt coalescing */
u8 res27[208];
};
/* Flags related to gianfar device features */
#define FSL_GIANFAR_DEV_HAS_GIGABIT 0x00000001
#define FSL_GIANFAR_DEV_HAS_COALESCE 0x00000002
#define FSL_GIANFAR_DEV_HAS_RMON 0x00000004
#define FSL_GIANFAR_DEV_HAS_MULTI_INTR 0x00000008
#define FSL_GIANFAR_DEV_HAS_CSUM 0x00000010
#define FSL_GIANFAR_DEV_HAS_VLAN 0x00000020
#define FSL_GIANFAR_DEV_HAS_EXTENDED_HASH 0x00000040
#define FSL_GIANFAR_DEV_HAS_MAGIC_PACKET 0x00000100
#define FSL_GIANFAR_DEV_HAS_BD_STASHING 0x00000200
#define FSL_GIANFAR_DEV_HAS_BUF_STASHING 0x00000400
#define FSL_GIANFAR_DEV_HAS_TIMER 0x00000800
#define FSL_GIANFAR_DEV_HAS_WAKE_ON_FILER 0x00001000
#define FSL_GIANFAR_DEV_HAS_RX_FILER 0x00002000
#if (MAXGROUPS == 2)
#define DEFAULT_MAPPING 0xAA
#else
#define DEFAULT_MAPPING 0xFF
#endif
#define ISRG_RR0 0x80000000
#define ISRG_TR0 0x00800000
/* The same driver can operate in two modes */
/* SQ_SG_MODE: Single Queue Single Group Mode
* (Backward compatible mode)
* MQ_MG_MODE: Multi Queue Multi Group mode
*/
enum {
SQ_SG_MODE = 0,
MQ_MG_MODE
};
gianfar: Use Single-Queue polling for "fsl,etsec2" For the "fsl,etsec2" compatible models the driver currently supports 8 Tx and Rx DMA rings (aka HW queues). However, there are only 2 pairs of Rx/Tx interrupt lines, as these controllers are integrated in low power SoCs with 2 CPUs at most. As a result, there are at most 2 NAPI instances that have to service multiple Tx and Rx queues for these devices. This complicates the NAPI polling routine having to iterate over the mutiple Rx/Tx queues hooked to the same interrupt lines. And there's also an overhead at HW level, as the controller needs to service all the 8 Tx rings in a round robin manner. The combined overhead shows up for multi parallel Tx flows transmitted by the kernel stack, when the driver usually starts returning NETDEV_TX_BUSY leading to NETDEV WATCHDOG Tx timeout triggering if the Tx path is congested for too long. As an alternative, this patch makes the driver support only one Tx/Rx DMA ring per NAPI instance (per interrupt group or pair of Tx/Rx interrupt lines) by default. The simplified single queue polling routine (gfar_poll_sq) will be the default napi poll routine for the etsec2 devices too. Some adjustments needed to be made to link the Tx/Rx HW queues with each NAPI instance (2 in this case). The gfar_poll_sq() is already successfully used by older SQ_SG_MODE (single interrupt group) controllers. This patch fixes Tx timeout triggering under heavy Tx traffic load (i.e. iperf -c -P 8) for the "fsl,etsec2" (currently the only MQ_MG_MODE devices). There's also a significant memory footprint reduction by supporting 2 Rx/Tx DMA rings (at most), instead of 8, for these devices. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:46 +04:00
/* GFAR_SQ_POLLING: Single Queue NAPI polling mode
* The driver supports a single pair of RX/Tx queues
* per interrupt group (Rx/Tx int line). MQ_MG mode
* devices have 2 interrupt groups, so the device will
* have a total of 2 Tx and 2 Rx queues in this case.
* GFAR_MQ_POLLING: Multi Queue NAPI polling mode
* The driver supports all the 8 Rx and Tx HW queues
* each queue mapped by the Device Tree to one of
* the 2 interrupt groups. This mode implies significant
* processing overhead (CPU and controller level).
*/
enum gfar_poll_mode {
GFAR_SQ_POLLING = 0,
GFAR_MQ_POLLING
};
/*
* Per TX queue stats
*/
struct tx_q_stats {
unsigned long tx_packets;
unsigned long tx_bytes;
};
/**
* struct gfar_priv_tx_q - per tx queue structure
* @txlock: per queue tx spin lock
* @tx_skbuff:skb pointers
* @skb_curtx: to be used skb pointer
* @skb_dirtytx:the last used skb pointer
* @stats: bytes/packets stats
* @qindex: index of this queue
* @dev: back pointer to the dev structure
* @grp: back pointer to the group to which this queue belongs
* @tx_bd_base: First tx buffer descriptor
* @cur_tx: Next free ring entry
* @dirty_tx: First buffer in line to be transmitted
* @tx_ring_size: Tx ring size
* @num_txbdfree: number of free TxBds
* @txcoalescing: enable/disable tx coalescing
* @txic: transmit interrupt coalescing value
* @txcount: coalescing value if based on tx frame count
* @txtime: coalescing value if based on time
*/
struct gfar_priv_tx_q {
/* cacheline 1 */
spinlock_t txlock __attribute__ ((aligned (SMP_CACHE_BYTES)));
struct txbd8 *tx_bd_base;
struct txbd8 *cur_tx;
unsigned int num_txbdfree;
unsigned short skb_curtx;
unsigned short tx_ring_size;
struct tx_q_stats stats;
struct gfar_priv_grp *grp;
/* cacheline 2 */
struct net_device *dev;
struct sk_buff **tx_skbuff;
struct txbd8 *dirty_tx;
unsigned short skb_dirtytx;
unsigned short qindex;
/* Configuration info for the coalescing features */
unsigned int txcoalescing;
unsigned long txic;
dma_addr_t tx_bd_dma_base;
};
/*
* Per RX queue stats
*/
struct rx_q_stats {
unsigned long rx_packets;
unsigned long rx_bytes;
unsigned long rx_dropped;
};
struct gfar_rx_buff {
dma_addr_t dma;
struct page *page;
unsigned int page_offset;
};
/**
* struct gfar_priv_rx_q - per rx queue structure
* @rx_buff: Array of buffer info metadata structs
* @rx_bd_base: First rx buffer descriptor
* @next_to_use: index of the next buffer to be alloc'd
* @next_to_clean: index of the next buffer to be cleaned
* @qindex: index of this queue
* @ndev: back pointer to net_device
* @rx_ring_size: Rx ring size
* @rxcoalescing: enable/disable rx-coalescing
* @rxic: receive interrupt coalescing vlaue
*/
struct gfar_priv_rx_q {
struct gfar_rx_buff *rx_buff __aligned(SMP_CACHE_BYTES);
struct rxbd8 *rx_bd_base;
struct net_device *ndev;
struct device *dev;
u16 rx_ring_size;
u16 qindex;
struct gfar_priv_grp *grp;
u16 next_to_clean;
u16 next_to_use;
u16 next_to_alloc;
struct sk_buff *skb;
struct rx_q_stats stats;
u32 __iomem *rfbptr;
unsigned char rxcoalescing;
unsigned long rxic;
dma_addr_t rx_bd_dma_base;
};
enum gfar_irqinfo_id {
GFAR_TX = 0,
GFAR_RX = 1,
GFAR_ER = 2,
GFAR_NUM_IRQS = 3
};
struct gfar_irqinfo {
unsigned int irq;
char name[GFAR_INT_NAME_MAX];
};
/**
* struct gfar_priv_grp - per group structure
* @napi: the napi poll function
* @priv: back pointer to the priv structure
* @regs: the ioremapped register space for this group
* @irqinfo: TX/RX/ER irq data for this group
*/
struct gfar_priv_grp {
gianfar: Use Single-Queue polling for "fsl,etsec2" For the "fsl,etsec2" compatible models the driver currently supports 8 Tx and Rx DMA rings (aka HW queues). However, there are only 2 pairs of Rx/Tx interrupt lines, as these controllers are integrated in low power SoCs with 2 CPUs at most. As a result, there are at most 2 NAPI instances that have to service multiple Tx and Rx queues for these devices. This complicates the NAPI polling routine having to iterate over the mutiple Rx/Tx queues hooked to the same interrupt lines. And there's also an overhead at HW level, as the controller needs to service all the 8 Tx rings in a round robin manner. The combined overhead shows up for multi parallel Tx flows transmitted by the kernel stack, when the driver usually starts returning NETDEV_TX_BUSY leading to NETDEV WATCHDOG Tx timeout triggering if the Tx path is congested for too long. As an alternative, this patch makes the driver support only one Tx/Rx DMA ring per NAPI instance (per interrupt group or pair of Tx/Rx interrupt lines) by default. The simplified single queue polling routine (gfar_poll_sq) will be the default napi poll routine for the etsec2 devices too. Some adjustments needed to be made to link the Tx/Rx HW queues with each NAPI instance (2 in this case). The gfar_poll_sq() is already successfully used by older SQ_SG_MODE (single interrupt group) controllers. This patch fixes Tx timeout triggering under heavy Tx traffic load (i.e. iperf -c -P 8) for the "fsl,etsec2" (currently the only MQ_MG_MODE devices). There's also a significant memory footprint reduction by supporting 2 Rx/Tx DMA rings (at most), instead of 8, for these devices. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:46 +04:00
spinlock_t grplock __aligned(SMP_CACHE_BYTES);
gianfar: Separate out the Tx interrupt handling (Tx NAPI) There are some concurrency issues on devices w/ 2 CPUs related to the handling of Rx and Tx interrupts. eTSEC has separate interrupt lines for Rx and Tx but a single imask register to mask these interrupts and a single NAPI instance to handle both Rx and Tx work. As a result, the Rx and Tx ISRs are identical, both are invoking gfar_schedule_cleanup(), however both handlers can be entered at the same time when the Rx and Tx interrupts are taken by different CPUs. In this case spurrious interrupts (SPU) show up (in /proc/interrupts) indicating a concurrency issue. Also, Tx overruns followed by Tx timeout have been observed under heavy Tx traffic load. To address these issues, the schedule cleanup ISR part has been changed to handle the Rx and Tx interrupts independently. The patch adds a separate NAPI poll routine for Tx cleanup to be triggerred independently by the Tx confirmation interrupts only. Existing poll functions are modified to handle only the Rx path processing. The Tx poll routine does not need a budget, since Tx processing doesn't consume NAPI budget, and hence it is registered with minimum NAPI weight. NAPI scheduling does not require locking since there are different NAPI instances between the Rx and Tx confirmation paths now. So, the patch fixes the occurence of spurrious Rx/Tx interrupts. Tx overruns also occur less frequently now. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:45 +04:00
struct napi_struct napi_rx;
struct napi_struct napi_tx;
struct gfar __iomem *regs;
gianfar: Use Single-Queue polling for "fsl,etsec2" For the "fsl,etsec2" compatible models the driver currently supports 8 Tx and Rx DMA rings (aka HW queues). However, there are only 2 pairs of Rx/Tx interrupt lines, as these controllers are integrated in low power SoCs with 2 CPUs at most. As a result, there are at most 2 NAPI instances that have to service multiple Tx and Rx queues for these devices. This complicates the NAPI polling routine having to iterate over the mutiple Rx/Tx queues hooked to the same interrupt lines. And there's also an overhead at HW level, as the controller needs to service all the 8 Tx rings in a round robin manner. The combined overhead shows up for multi parallel Tx flows transmitted by the kernel stack, when the driver usually starts returning NETDEV_TX_BUSY leading to NETDEV WATCHDOG Tx timeout triggering if the Tx path is congested for too long. As an alternative, this patch makes the driver support only one Tx/Rx DMA ring per NAPI instance (per interrupt group or pair of Tx/Rx interrupt lines) by default. The simplified single queue polling routine (gfar_poll_sq) will be the default napi poll routine for the etsec2 devices too. Some adjustments needed to be made to link the Tx/Rx HW queues with each NAPI instance (2 in this case). The gfar_poll_sq() is already successfully used by older SQ_SG_MODE (single interrupt group) controllers. This patch fixes Tx timeout triggering under heavy Tx traffic load (i.e. iperf -c -P 8) for the "fsl,etsec2" (currently the only MQ_MG_MODE devices). There's also a significant memory footprint reduction by supporting 2 Rx/Tx DMA rings (at most), instead of 8, for these devices. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:46 +04:00
struct gfar_priv_tx_q *tx_queue;
struct gfar_priv_rx_q *rx_queue;
unsigned int tstat;
gianfar: Use Single-Queue polling for "fsl,etsec2" For the "fsl,etsec2" compatible models the driver currently supports 8 Tx and Rx DMA rings (aka HW queues). However, there are only 2 pairs of Rx/Tx interrupt lines, as these controllers are integrated in low power SoCs with 2 CPUs at most. As a result, there are at most 2 NAPI instances that have to service multiple Tx and Rx queues for these devices. This complicates the NAPI polling routine having to iterate over the mutiple Rx/Tx queues hooked to the same interrupt lines. And there's also an overhead at HW level, as the controller needs to service all the 8 Tx rings in a round robin manner. The combined overhead shows up for multi parallel Tx flows transmitted by the kernel stack, when the driver usually starts returning NETDEV_TX_BUSY leading to NETDEV WATCHDOG Tx timeout triggering if the Tx path is congested for too long. As an alternative, this patch makes the driver support only one Tx/Rx DMA ring per NAPI instance (per interrupt group or pair of Tx/Rx interrupt lines) by default. The simplified single queue polling routine (gfar_poll_sq) will be the default napi poll routine for the etsec2 devices too. Some adjustments needed to be made to link the Tx/Rx HW queues with each NAPI instance (2 in this case). The gfar_poll_sq() is already successfully used by older SQ_SG_MODE (single interrupt group) controllers. This patch fixes Tx timeout triggering under heavy Tx traffic load (i.e. iperf -c -P 8) for the "fsl,etsec2" (currently the only MQ_MG_MODE devices). There's also a significant memory footprint reduction by supporting 2 Rx/Tx DMA rings (at most), instead of 8, for these devices. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:46 +04:00
unsigned int rstat;
struct gfar_private *priv;
unsigned long num_tx_queues;
unsigned long tx_bit_map;
gianfar: Use Single-Queue polling for "fsl,etsec2" For the "fsl,etsec2" compatible models the driver currently supports 8 Tx and Rx DMA rings (aka HW queues). However, there are only 2 pairs of Rx/Tx interrupt lines, as these controllers are integrated in low power SoCs with 2 CPUs at most. As a result, there are at most 2 NAPI instances that have to service multiple Tx and Rx queues for these devices. This complicates the NAPI polling routine having to iterate over the mutiple Rx/Tx queues hooked to the same interrupt lines. And there's also an overhead at HW level, as the controller needs to service all the 8 Tx rings in a round robin manner. The combined overhead shows up for multi parallel Tx flows transmitted by the kernel stack, when the driver usually starts returning NETDEV_TX_BUSY leading to NETDEV WATCHDOG Tx timeout triggering if the Tx path is congested for too long. As an alternative, this patch makes the driver support only one Tx/Rx DMA ring per NAPI instance (per interrupt group or pair of Tx/Rx interrupt lines) by default. The simplified single queue polling routine (gfar_poll_sq) will be the default napi poll routine for the etsec2 devices too. Some adjustments needed to be made to link the Tx/Rx HW queues with each NAPI instance (2 in this case). The gfar_poll_sq() is already successfully used by older SQ_SG_MODE (single interrupt group) controllers. This patch fixes Tx timeout triggering under heavy Tx traffic load (i.e. iperf -c -P 8) for the "fsl,etsec2" (currently the only MQ_MG_MODE devices). There's also a significant memory footprint reduction by supporting 2 Rx/Tx DMA rings (at most), instead of 8, for these devices. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:46 +04:00
unsigned long num_rx_queues;
unsigned long rx_bit_map;
struct gfar_irqinfo *irqinfo[GFAR_NUM_IRQS];
};
#define gfar_irq(grp, ID) \
((grp)->irqinfo[GFAR_##ID])
enum gfar_errata {
GFAR_ERRATA_74 = 0x01,
GFAR_ERRATA_76 = 0x02,
GFAR_ERRATA_A002 = 0x04,
GFAR_ERRATA_12 = 0x08, /* a.k.a errata eTSEC49 */
};
gianfar: Fix device reset races (oops) for Tx The device reset procedure, stop_gfar()/startup_gfar(), has concurrency issues. "Kernel access of bad area" oopses show up during Tx timeout device reset or other reset cases (like changing MTU) that happen while the interface still has traffic. The oopses happen in start_xmit and clean_tx_ring when accessing tx_queue-> tx_skbuff which is NULL. The race comes from de-allocating the tx_skbuff while transmission and napi processing are still active. Though the Tx queues get temoprarily stopped when Tx timeout occurs, they get re-enabled as a result of Tx congestion handling inside the napi context (see clean_tx_ring()). Not disabling the napi during reset is also a bug, because clean_tx_ring() will try to access tx_skbuff while it is being de-alloc'ed and re-alloc'ed. To fix this, stop_gfar() needs to disable napi processing after stopping the Tx queues. However, in order to prevent clean_tx_ring() to re-enable the Tx queue before the napi gets disabled, the device state DOWN has been introduced. It prevents the Tx congestion management from re-enabling the de-congested Tx queue while the device is brought down. An additional locking state, RESETTING, has been introduced to prevent simultaneous resets or to prevent configuring the device while it is resetting. The bogus 'rxlock's (for each Rx queue) have been removed since their purpose is not justified, as they don't prevent nor are suited to prevent device reset/reconfig races (such as this one). Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-02-24 14:13:45 +04:00
enum gfar_dev_state {
GFAR_DOWN = 1,
GFAR_RESETTING
};
/* Struct stolen almost completely (and shamelessly) from the FCC enet source
* (Ok, that's not so true anymore, but there is a family resemblance)
* The GFAR buffer descriptors track the ring buffers. The rx_bd_base
* and tx_bd_base always point to the currently available buffer.
* The dirty_tx tracks the current buffer that is being sent by the
* controller. The cur_tx and dirty_tx are equal under both completely
* empty and completely full conditions. The empty/ready indicator in
* the buffer descriptor determines the actual condition.
*/
struct gfar_private {
struct device *dev;
struct net_device *ndev;
enum gfar_errata errata;
gianfar: Fix and cleanup Rx FCB indication This fixes a less obvious error on one hand, and prevents futher similar errors by disambiguating and optimizing RxFCB indication, on the other hand. The error consists in NETIF_F_HW_VLAN_TX flag being used as an indication of Rx FCB insertion. This happened as soon gfar_uses_fcb(), which despite its name indicates Rx FCB insertion, started incorporating is_vlan_on(). is_vlan_on(), on the other hand, is also a misleading construct because we need to differentiate b/w hw VLAN extraction/VLEX (marked by VLAN_RX flag) and hw VLAN insertion/VLINS (VLAN_TX flag), which are different mechanisms using different types of FCBs. The hw spec for the RxFCB feature is as follows: In the case of RxBD rings, FCBs (Frame Control Block) are inserted by the eTSEC whenever RCTRL[PRSDEP] is set to a non-zero value. Only one FCB is inserted per frame (in the buffer pointed to by the RxBD with bit F set). TOE acceleration for receive is enabled for all rx frames in this case. This patch introduces priv->uses_rxfcb field to quickly signal RxFCB insertion in accordance with the specification above. The dependency on FSL_GIANFAR_DEV_HAS_TIMER was also eliminated as another source of confusion. The actual dependency is to priv->hwts_rx_en. Upon changing priv->hwts_rx_en via IOCTL, the gfar device is being restarted and on init_mac() the priv->hwts_rx_en flag determines RxFCB insertion, and rctrl is programmed accordingly. The patch takes care of this case too. Though maybe not as self documenting as the inlining version uses_fcb(), priv->uses_rxfcb has the main purpose to quickly signal, on the hot path, that the incoming frame has a *Rx* FCB block inserted which needs to be pulled out before passing the skb to the stack. This is a performance critical operation, it needs to happen fast, that's why uses_rxfcb is placed in the first cacheline of gfar_private. This is also why a cached rctrl cannot be used instead: 1) because we don't have 32 bits available in the first cacheline of gfar_priv (but only 16); 2) bit operations are expensive on the hot path. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-02-14 09:00:07 +04:00
u16 uses_rxfcb;
u16 padding;
gianfar: Use Single-Queue polling for "fsl,etsec2" For the "fsl,etsec2" compatible models the driver currently supports 8 Tx and Rx DMA rings (aka HW queues). However, there are only 2 pairs of Rx/Tx interrupt lines, as these controllers are integrated in low power SoCs with 2 CPUs at most. As a result, there are at most 2 NAPI instances that have to service multiple Tx and Rx queues for these devices. This complicates the NAPI polling routine having to iterate over the mutiple Rx/Tx queues hooked to the same interrupt lines. And there's also an overhead at HW level, as the controller needs to service all the 8 Tx rings in a round robin manner. The combined overhead shows up for multi parallel Tx flows transmitted by the kernel stack, when the driver usually starts returning NETDEV_TX_BUSY leading to NETDEV WATCHDOG Tx timeout triggering if the Tx path is congested for too long. As an alternative, this patch makes the driver support only one Tx/Rx DMA ring per NAPI instance (per interrupt group or pair of Tx/Rx interrupt lines) by default. The simplified single queue polling routine (gfar_poll_sq) will be the default napi poll routine for the etsec2 devices too. Some adjustments needed to be made to link the Tx/Rx HW queues with each NAPI instance (2 in this case). The gfar_poll_sq() is already successfully used by older SQ_SG_MODE (single interrupt group) controllers. This patch fixes Tx timeout triggering under heavy Tx traffic load (i.e. iperf -c -P 8) for the "fsl,etsec2" (currently the only MQ_MG_MODE devices). There's also a significant memory footprint reduction by supporting 2 Rx/Tx DMA rings (at most), instead of 8, for these devices. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:46 +04:00
u32 device_flags;
/* HW time stamping enabled flag */
int hwts_rx_en;
int hwts_tx_en;
struct gfar_priv_tx_q *tx_queue[MAX_TX_QS];
struct gfar_priv_rx_q *rx_queue[MAX_RX_QS];
struct gfar_priv_grp gfargrp[MAXGROUPS];
gianfar: Fix device reset races (oops) for Tx The device reset procedure, stop_gfar()/startup_gfar(), has concurrency issues. "Kernel access of bad area" oopses show up during Tx timeout device reset or other reset cases (like changing MTU) that happen while the interface still has traffic. The oopses happen in start_xmit and clean_tx_ring when accessing tx_queue-> tx_skbuff which is NULL. The race comes from de-allocating the tx_skbuff while transmission and napi processing are still active. Though the Tx queues get temoprarily stopped when Tx timeout occurs, they get re-enabled as a result of Tx congestion handling inside the napi context (see clean_tx_ring()). Not disabling the napi during reset is also a bug, because clean_tx_ring() will try to access tx_skbuff while it is being de-alloc'ed and re-alloc'ed. To fix this, stop_gfar() needs to disable napi processing after stopping the Tx queues. However, in order to prevent clean_tx_ring() to re-enable the Tx queue before the napi gets disabled, the device state DOWN has been introduced. It prevents the Tx congestion management from re-enabling the de-congested Tx queue while the device is brought down. An additional locking state, RESETTING, has been introduced to prevent simultaneous resets or to prevent configuring the device while it is resetting. The bogus 'rxlock's (for each Rx queue) have been removed since their purpose is not justified, as they don't prevent nor are suited to prevent device reset/reconfig races (such as this one). Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-02-24 14:13:45 +04:00
unsigned long state;
gianfar: Use Single-Queue polling for "fsl,etsec2" For the "fsl,etsec2" compatible models the driver currently supports 8 Tx and Rx DMA rings (aka HW queues). However, there are only 2 pairs of Rx/Tx interrupt lines, as these controllers are integrated in low power SoCs with 2 CPUs at most. As a result, there are at most 2 NAPI instances that have to service multiple Tx and Rx queues for these devices. This complicates the NAPI polling routine having to iterate over the mutiple Rx/Tx queues hooked to the same interrupt lines. And there's also an overhead at HW level, as the controller needs to service all the 8 Tx rings in a round robin manner. The combined overhead shows up for multi parallel Tx flows transmitted by the kernel stack, when the driver usually starts returning NETDEV_TX_BUSY leading to NETDEV WATCHDOG Tx timeout triggering if the Tx path is congested for too long. As an alternative, this patch makes the driver support only one Tx/Rx DMA ring per NAPI instance (per interrupt group or pair of Tx/Rx interrupt lines) by default. The simplified single queue polling routine (gfar_poll_sq) will be the default napi poll routine for the etsec2 devices too. Some adjustments needed to be made to link the Tx/Rx HW queues with each NAPI instance (2 in this case). The gfar_poll_sq() is already successfully used by older SQ_SG_MODE (single interrupt group) controllers. This patch fixes Tx timeout triggering under heavy Tx traffic load (i.e. iperf -c -P 8) for the "fsl,etsec2" (currently the only MQ_MG_MODE devices). There's also a significant memory footprint reduction by supporting 2 Rx/Tx DMA rings (at most), instead of 8, for these devices. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:46 +04:00
unsigned short mode;
unsigned short poll_mode;
unsigned int num_tx_queues;
gianfar: Use Single-Queue polling for "fsl,etsec2" For the "fsl,etsec2" compatible models the driver currently supports 8 Tx and Rx DMA rings (aka HW queues). However, there are only 2 pairs of Rx/Tx interrupt lines, as these controllers are integrated in low power SoCs with 2 CPUs at most. As a result, there are at most 2 NAPI instances that have to service multiple Tx and Rx queues for these devices. This complicates the NAPI polling routine having to iterate over the mutiple Rx/Tx queues hooked to the same interrupt lines. And there's also an overhead at HW level, as the controller needs to service all the 8 Tx rings in a round robin manner. The combined overhead shows up for multi parallel Tx flows transmitted by the kernel stack, when the driver usually starts returning NETDEV_TX_BUSY leading to NETDEV WATCHDOG Tx timeout triggering if the Tx path is congested for too long. As an alternative, this patch makes the driver support only one Tx/Rx DMA ring per NAPI instance (per interrupt group or pair of Tx/Rx interrupt lines) by default. The simplified single queue polling routine (gfar_poll_sq) will be the default napi poll routine for the etsec2 devices too. Some adjustments needed to be made to link the Tx/Rx HW queues with each NAPI instance (2 in this case). The gfar_poll_sq() is already successfully used by older SQ_SG_MODE (single interrupt group) controllers. This patch fixes Tx timeout triggering under heavy Tx traffic load (i.e. iperf -c -P 8) for the "fsl,etsec2" (currently the only MQ_MG_MODE devices). There's also a significant memory footprint reduction by supporting 2 Rx/Tx DMA rings (at most), instead of 8, for these devices. Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-03-07 16:42:46 +04:00
unsigned int num_rx_queues;
unsigned int num_grps;
int tx_actual_en;
/* Network Statistics */
struct gfar_extra_stats extra_stats;
/* PHY stuff */
phy_interface_t interface;
struct device_node *phy_node;
struct device_node *tbi_node;
struct mii_bus *mii_bus;
int oldspeed;
int oldduplex;
int oldlink;
uint32_t msg_enable;
struct work_struct reset_task;
struct platform_device *ofdev;
unsigned char
extended_hash:1,
bd_stash_en:1,
rx_filer_enable:1,
/* Enable priorty based Tx scheduling in Hw */
prio_sched_en:1,
/* Flow control flags */
pause_aneg_en:1,
tx_pause_en:1,
rx_pause_en:1;
/* The total tx and rx ring size for the enabled queues */
unsigned int total_tx_ring_size;
unsigned int total_rx_ring_size;
u32 rqueue;
u32 tqueue;
/* RX per device parameters */
unsigned int rx_stash_size;
unsigned int rx_stash_index;
u32 cur_filer_idx;
/* RX queue filer rule set*/
struct ethtool_rx_list rx_list;
struct mutex rx_queue_access;
/* Hash registers and their width */
u32 __iomem *hash_regs[16];
int hash_width;
/* wake-on-lan settings */
u16 wol_opts;
u16 wol_supported;
/*Filer table*/
unsigned int ftp_rqfpr[MAX_FILER_IDX + 1];
unsigned int ftp_rqfcr[MAX_FILER_IDX + 1];
};
static inline int gfar_has_errata(struct gfar_private *priv,
enum gfar_errata err)
{
return priv->errata & err;
}
static inline u32 gfar_read(unsigned __iomem *addr)
{
u32 val;
val = ioread32be(addr);
return val;
}
static inline void gfar_write(unsigned __iomem *addr, u32 val)
{
iowrite32be(val, addr);
}
static inline void gfar_write_filer(struct gfar_private *priv,
unsigned int far, unsigned int fcr, unsigned int fpr)
{
struct gfar __iomem *regs = priv->gfargrp[0].regs;
gfar_write(&regs->rqfar, far);
gfar_write(&regs->rqfcr, fcr);
gfar_write(&regs->rqfpr, fpr);
}
static inline void gfar_read_filer(struct gfar_private *priv,
unsigned int far, unsigned int *fcr, unsigned int *fpr)
{
struct gfar __iomem *regs = priv->gfargrp[0].regs;
gfar_write(&regs->rqfar, far);
*fcr = gfar_read(&regs->rqfcr);
*fpr = gfar_read(&regs->rqfpr);
}
static inline void gfar_write_isrg(struct gfar_private *priv)
{
struct gfar __iomem *regs = priv->gfargrp[0].regs;
u32 __iomem *baddr = &regs->isrg0;
u32 isrg = 0;
int grp_idx, i;
for (grp_idx = 0; grp_idx < priv->num_grps; grp_idx++) {
struct gfar_priv_grp *grp = &priv->gfargrp[grp_idx];
for_each_set_bit(i, &grp->rx_bit_map, priv->num_rx_queues) {
isrg |= (ISRG_RR0 >> i);
}
for_each_set_bit(i, &grp->tx_bit_map, priv->num_tx_queues) {
isrg |= (ISRG_TR0 >> i);
}
gfar_write(baddr, isrg);
baddr++;
isrg = 0;
}
}
static inline int gfar_is_dma_stopped(struct gfar_private *priv)
{
struct gfar __iomem *regs = priv->gfargrp[0].regs;
return ((gfar_read(&regs->ievent) & (IEVENT_GRSC | IEVENT_GTSC)) ==
(IEVENT_GRSC | IEVENT_GTSC));
}
static inline int gfar_is_rx_dma_stopped(struct gfar_private *priv)
{
struct gfar __iomem *regs = priv->gfargrp[0].regs;
return gfar_read(&regs->ievent) & IEVENT_GRSC;
}
static inline void gfar_wmb(void)
{
#if defined(CONFIG_PPC)
/* The powerpc-specific eieio() is used, as wmb() has too strong
* semantics (it requires synchronization between cacheable and
* uncacheable mappings, which eieio() doesn't provide and which we
* don't need), thus requiring a more expensive sync instruction. At
* some point, the set of architecture-independent barrier functions
* should be expanded to include weaker barriers.
*/
eieio();
#else
wmb(); /* order write acesses for BD (or FCB) fields */
#endif
}
static inline void gfar_clear_txbd_status(struct txbd8 *bdp)
{
u32 lstatus = be32_to_cpu(bdp->lstatus);
lstatus &= BD_LFLAG(TXBD_WRAP);
bdp->lstatus = cpu_to_be32(lstatus);
}
static inline int gfar_rxbd_unused(struct gfar_priv_rx_q *rxq)
{
if (rxq->next_to_clean > rxq->next_to_use)
return rxq->next_to_clean - rxq->next_to_use - 1;
return rxq->rx_ring_size + rxq->next_to_clean - rxq->next_to_use - 1;
}
static inline u32 gfar_rxbd_dma_lastfree(struct gfar_priv_rx_q *rxq)
{
struct rxbd8 *bdp;
u32 bdp_dma;
int i;
i = rxq->next_to_use ? rxq->next_to_use - 1 : rxq->rx_ring_size - 1;
bdp = &rxq->rx_bd_base[i];
bdp_dma = lower_32_bits(rxq->rx_bd_dma_base);
bdp_dma += (uintptr_t)bdp - (uintptr_t)rxq->rx_bd_base;
return bdp_dma;
}
irqreturn_t gfar_receive(int irq, void *dev_id);
int startup_gfar(struct net_device *dev);
void stop_gfar(struct net_device *dev);
gianfar: Fix device reset races (oops) for Tx The device reset procedure, stop_gfar()/startup_gfar(), has concurrency issues. "Kernel access of bad area" oopses show up during Tx timeout device reset or other reset cases (like changing MTU) that happen while the interface still has traffic. The oopses happen in start_xmit and clean_tx_ring when accessing tx_queue-> tx_skbuff which is NULL. The race comes from de-allocating the tx_skbuff while transmission and napi processing are still active. Though the Tx queues get temoprarily stopped when Tx timeout occurs, they get re-enabled as a result of Tx congestion handling inside the napi context (see clean_tx_ring()). Not disabling the napi during reset is also a bug, because clean_tx_ring() will try to access tx_skbuff while it is being de-alloc'ed and re-alloc'ed. To fix this, stop_gfar() needs to disable napi processing after stopping the Tx queues. However, in order to prevent clean_tx_ring() to re-enable the Tx queue before the napi gets disabled, the device state DOWN has been introduced. It prevents the Tx congestion management from re-enabling the de-congested Tx queue while the device is brought down. An additional locking state, RESETTING, has been introduced to prevent simultaneous resets or to prevent configuring the device while it is resetting. The bogus 'rxlock's (for each Rx queue) have been removed since their purpose is not justified, as they don't prevent nor are suited to prevent device reset/reconfig races (such as this one). Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-02-24 14:13:45 +04:00
void reset_gfar(struct net_device *dev);
void gfar_mac_reset(struct gfar_private *priv);
void gfar_halt(struct gfar_private *priv);
void gfar_start(struct gfar_private *priv);
void gfar_phy_test(struct mii_bus *bus, struct phy_device *phydev, int enable,
u32 regnum, u32 read);
void gfar_configure_coalescing_all(struct gfar_private *priv);
int gfar_set_features(struct net_device *dev, netdev_features_t features);
extern const struct ethtool_ops gfar_ethtool_ops;
#define MAX_FILER_CACHE_IDX (2*(MAX_FILER_IDX))
#define RQFCR_PID_PRI_MASK 0xFFFFFFF8
#define RQFCR_PID_L4P_MASK 0xFFFFFF00
#define RQFCR_PID_VID_MASK 0xFFFFF000
#define RQFCR_PID_PORT_MASK 0xFFFF0000
#define RQFCR_PID_MAC_MASK 0xFF000000
struct gfar_mask_entry {
unsigned int mask; /* The mask value which is valid form start to end */
unsigned int start;
unsigned int end;
unsigned int block; /* Same block values indicate depended entries */
};
/* Represents a receive filer table entry */
struct gfar_filer_entry {
u32 ctrl;
u32 prop;
};
/* The 20 additional entries are a shadow for one extra element */
struct filer_table {
u32 index;
struct gfar_filer_entry fe[MAX_FILER_CACHE_IDX + 20];
};
/* The gianfar_ptp module will set this variable */
extern int gfar_phc_index;
#endif /* __GIANFAR_H */