android_kernel_xiaomi_sm8350/drivers/net/fec_8xx/fec_main.c

1265 lines
29 KiB
C
Raw Normal View History

/*
* Fast Ethernet Controller (FEC) driver for Motorola MPC8xx.
*
* Copyright (c) 2003 Intracom S.A.
* by Pantelis Antoniou <panto@intracom.gr>
*
* Heavily based on original FEC driver by Dan Malek <dan@embeddededge.com>
* and modifications by Joakim Tjernlund <joakim.tjernlund@lumentis.se>
*
* Released under the GPL
*/
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/string.h>
#include <linux/ptrace.h>
#include <linux/errno.h>
#include <linux/ioport.h>
#include <linux/slab.h>
#include <linux/interrupt.h>
#include <linux/init.h>
#include <linux/delay.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/skbuff.h>
#include <linux/spinlock.h>
#include <linux/mii.h>
#include <linux/ethtool.h>
#include <linux/bitops.h>
#include <linux/dma-mapping.h>
#include <asm/8xx_immap.h>
#include <asm/pgtable.h>
#include <asm/mpc8xx.h>
#include <asm/irq.h>
#include <asm/uaccess.h>
#include <asm/commproc.h>
#include "fec_8xx.h"
/*************************************************/
#define FEC_MAX_MULTICAST_ADDRS 64
/*************************************************/
static char version[] __devinitdata =
DRV_MODULE_NAME ".c:v" DRV_MODULE_VERSION " (" DRV_MODULE_RELDATE ")" "\n";
MODULE_AUTHOR("Pantelis Antoniou <panto@intracom.gr>");
MODULE_DESCRIPTION("Motorola 8xx FEC ethernet driver");
MODULE_LICENSE("GPL");
int fec_8xx_debug = -1; /* -1 == use FEC_8XX_DEF_MSG_ENABLE as value */
module_param(fec_8xx_debug, int, 0);
MODULE_PARM_DESC(fec_8xx_debug,
"FEC 8xx bitmapped debugging message enable value");
/*************************************************/
/*
* Delay to wait for FEC reset command to complete (in us)
*/
#define FEC_RESET_DELAY 50
/*****************************************************************************************/
static void fec_whack_reset(fec_t * fecp)
{
int i;
/*
* Whack a reset. We should wait for this.
*/
FW(fecp, ecntrl, FEC_ECNTRL_PINMUX | FEC_ECNTRL_RESET);
for (i = 0;
(FR(fecp, ecntrl) & FEC_ECNTRL_RESET) != 0 && i < FEC_RESET_DELAY;
i++)
udelay(1);
if (i == FEC_RESET_DELAY)
printk(KERN_WARNING "FEC Reset timeout!\n");
}
/****************************************************************************/
/*
* Transmitter timeout.
*/
#define TX_TIMEOUT (2*HZ)
/****************************************************************************/
/*
* Returns the CRC needed when filling in the hash table for
* multicast group filtering
* pAddr must point to a MAC address (6 bytes)
*/
static __u32 fec_mulicast_calc_crc(char *pAddr)
{
u8 byte;
int byte_count;
int bit_count;
__u32 crc = 0xffffffff;
u8 msb;
for (byte_count = 0; byte_count < 6; byte_count++) {
byte = pAddr[byte_count];
for (bit_count = 0; bit_count < 8; bit_count++) {
msb = crc >> 31;
crc <<= 1;
if (msb ^ (byte & 0x1)) {
crc ^= FEC_CRC_POLY;
}
byte >>= 1;
}
}
return (crc);
}
/*
* Set or clear the multicast filter for this adaptor.
* Skeleton taken from sunlance driver.
* The CPM Ethernet implementation allows Multicast as well as individual
* MAC address filtering. Some of the drivers check to make sure it is
* a group multicast address, and discard those that are not. I guess I
* will do the same for now, but just remove the test if you want
* individual filtering as well (do the upper net layers want or support
* this kind of feature?).
*/
static void fec_set_multicast_list(struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
fec_t *fecp = fep->fecp;
struct dev_mc_list *pmc;
__u32 crc;
int temp;
__u32 csrVal;
int hash_index;
__u32 hthi, htlo;
unsigned long flags;
if ((dev->flags & IFF_PROMISC) != 0) {
spin_lock_irqsave(&fep->lock, flags);
FS(fecp, r_cntrl, FEC_RCNTRL_PROM);
spin_unlock_irqrestore(&fep->lock, flags);
/*
* Log any net taps.
*/
printk(KERN_WARNING DRV_MODULE_NAME
": %s: Promiscuous mode enabled.\n", dev->name);
return;
}
if ((dev->flags & IFF_ALLMULTI) != 0 ||
dev->mc_count > FEC_MAX_MULTICAST_ADDRS) {
/*
* Catch all multicast addresses, set the filter to all 1's.
*/
hthi = 0xffffffffU;
htlo = 0xffffffffU;
} else {
hthi = 0;
htlo = 0;
/*
* Now populate the hash table
*/
for (pmc = dev->mc_list; pmc != NULL; pmc = pmc->next) {
crc = fec_mulicast_calc_crc(pmc->dmi_addr);
temp = (crc & 0x3f) >> 1;
hash_index = ((temp & 0x01) << 4) |
((temp & 0x02) << 2) |
((temp & 0x04)) |
((temp & 0x08) >> 2) |
((temp & 0x10) >> 4);
csrVal = (1 << hash_index);
if (crc & 1)
hthi |= csrVal;
else
htlo |= csrVal;
}
}
spin_lock_irqsave(&fep->lock, flags);
FC(fecp, r_cntrl, FEC_RCNTRL_PROM);
FW(fecp, hash_table_high, hthi);
FW(fecp, hash_table_low, htlo);
spin_unlock_irqrestore(&fep->lock, flags);
}
static int fec_set_mac_address(struct net_device *dev, void *addr)
{
struct sockaddr *mac = addr;
struct fec_enet_private *fep = netdev_priv(dev);
struct fec *fecp = fep->fecp;
int i;
__u32 addrhi, addrlo;
unsigned long flags;
/* Get pointer to SCC area in parameter RAM. */
for (i = 0; i < 6; i++)
dev->dev_addr[i] = mac->sa_data[i];
/*
* Set station address.
*/
addrhi = ((__u32) dev->dev_addr[0] << 24) |
((__u32) dev->dev_addr[1] << 16) |
((__u32) dev->dev_addr[2] << 8) |
(__u32) dev->dev_addr[3];
addrlo = ((__u32) dev->dev_addr[4] << 24) |
((__u32) dev->dev_addr[5] << 16);
spin_lock_irqsave(&fep->lock, flags);
FW(fecp, addr_low, addrhi);
FW(fecp, addr_high, addrlo);
spin_unlock_irqrestore(&fep->lock, flags);
return 0;
}
/*
* This function is called to start or restart the FEC during a link
* change. This only happens when switching between half and full
* duplex.
*/
void fec_restart(struct net_device *dev, int duplex, int speed)
{
#ifdef CONFIG_DUET
immap_t *immap = (immap_t *) IMAP_ADDR;
__u32 cptr;
#endif
struct fec_enet_private *fep = netdev_priv(dev);
struct fec *fecp = fep->fecp;
const struct fec_platform_info *fpi = fep->fpi;
cbd_t *bdp;
struct sk_buff *skb;
int i;
__u32 addrhi, addrlo;
fec_whack_reset(fep->fecp);
/*
* Set station address.
*/
addrhi = ((__u32) dev->dev_addr[0] << 24) |
((__u32) dev->dev_addr[1] << 16) |
((__u32) dev->dev_addr[2] << 8) |
(__u32) dev->dev_addr[3];
addrlo = ((__u32) dev->dev_addr[4] << 24) |
((__u32) dev->dev_addr[5] << 16);
FW(fecp, addr_low, addrhi);
FW(fecp, addr_high, addrlo);
/*
* Reset all multicast.
*/
FW(fecp, hash_table_high, 0);
FW(fecp, hash_table_low, 0);
/*
* Set maximum receive buffer size.
*/
FW(fecp, r_buff_size, PKT_MAXBLR_SIZE);
FW(fecp, r_hash, PKT_MAXBUF_SIZE);
/*
* Set receive and transmit descriptor base.
*/
FW(fecp, r_des_start, iopa((__u32) (fep->rx_bd_base)));
FW(fecp, x_des_start, iopa((__u32) (fep->tx_bd_base)));
fep->dirty_tx = fep->cur_tx = fep->tx_bd_base;
fep->tx_free = fep->tx_ring;
fep->cur_rx = fep->rx_bd_base;
/*
* Reset SKB receive buffers
*/
for (i = 0; i < fep->rx_ring; i++) {
if ((skb = fep->rx_skbuff[i]) == NULL)
continue;
fep->rx_skbuff[i] = NULL;
dev_kfree_skb(skb);
}
/*
* Initialize the receive buffer descriptors.
*/
for (i = 0, bdp = fep->rx_bd_base; i < fep->rx_ring; i++, bdp++) {
skb = dev_alloc_skb(ENET_RX_FRSIZE);
if (skb == NULL) {
printk(KERN_WARNING DRV_MODULE_NAME
": %s Memory squeeze, unable to allocate skb\n",
dev->name);
fep->stats.rx_dropped++;
break;
}
fep->rx_skbuff[i] = skb;
skb->dev = dev;
CBDW_BUFADDR(bdp, dma_map_single(NULL, skb->data,
L1_CACHE_ALIGN(PKT_MAXBUF_SIZE),
DMA_FROM_DEVICE));
CBDW_DATLEN(bdp, 0); /* zero */
CBDW_SC(bdp, BD_ENET_RX_EMPTY |
((i < fep->rx_ring - 1) ? 0 : BD_SC_WRAP));
}
/*
* if we failed, fillup remainder
*/
for (; i < fep->rx_ring; i++, bdp++) {
fep->rx_skbuff[i] = NULL;
CBDW_SC(bdp, (i < fep->rx_ring - 1) ? 0 : BD_SC_WRAP);
}
/*
* Reset SKB transmit buffers.
*/
for (i = 0; i < fep->tx_ring; i++) {
if ((skb = fep->tx_skbuff[i]) == NULL)
continue;
fep->tx_skbuff[i] = NULL;
dev_kfree_skb(skb);
}
/*
* ...and the same for transmit.
*/
for (i = 0, bdp = fep->tx_bd_base; i < fep->tx_ring; i++, bdp++) {
fep->tx_skbuff[i] = NULL;
CBDW_BUFADDR(bdp, virt_to_bus(NULL));
CBDW_DATLEN(bdp, 0);
CBDW_SC(bdp, (i < fep->tx_ring - 1) ? 0 : BD_SC_WRAP);
}
/*
* Enable big endian and don't care about SDMA FC.
*/
FW(fecp, fun_code, 0x78000000);
/*
* Set MII speed.
*/
FW(fecp, mii_speed, fep->fec_phy_speed);
/*
* Clear any outstanding interrupt.
*/
FW(fecp, ievent, 0xffc0);
FW(fecp, ivec, (fpi->fec_irq / 2) << 29);
/*
* adjust to speed (only for DUET & RMII)
*/
#ifdef CONFIG_DUET
cptr = in_be32(&immap->im_cpm.cp_cptr);
switch (fpi->fec_no) {
case 0:
/*
* check if in RMII mode
*/
if ((cptr & 0x100) == 0)
break;
if (speed == 10)
cptr |= 0x0000010;
else if (speed == 100)
cptr &= ~0x0000010;
break;
case 1:
/*
* check if in RMII mode
*/
if ((cptr & 0x80) == 0)
break;
if (speed == 10)
cptr |= 0x0000008;
else if (speed == 100)
cptr &= ~0x0000008;
break;
default:
break;
}
out_be32(&immap->im_cpm.cp_cptr, cptr);
#endif
FW(fecp, r_cntrl, FEC_RCNTRL_MII_MODE); /* MII enable */
/*
* adjust to duplex mode
*/
if (duplex) {
FC(fecp, r_cntrl, FEC_RCNTRL_DRT);
FS(fecp, x_cntrl, FEC_TCNTRL_FDEN); /* FD enable */
} else {
FS(fecp, r_cntrl, FEC_RCNTRL_DRT);
FC(fecp, x_cntrl, FEC_TCNTRL_FDEN); /* FD disable */
}
/*
* Enable interrupts we wish to service.
*/
FW(fecp, imask, FEC_ENET_TXF | FEC_ENET_TXB |
FEC_ENET_RXF | FEC_ENET_RXB);
/*
* And last, enable the transmit and receive processing.
*/
FW(fecp, ecntrl, FEC_ECNTRL_PINMUX | FEC_ECNTRL_ETHER_EN);
FW(fecp, r_des_active, 0x01000000);
}
void fec_stop(struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
fec_t *fecp = fep->fecp;
struct sk_buff *skb;
int i;
if ((FR(fecp, ecntrl) & FEC_ECNTRL_ETHER_EN) == 0)
return; /* already down */
FW(fecp, x_cntrl, 0x01); /* Graceful transmit stop */
for (i = 0; ((FR(fecp, ievent) & 0x10000000) == 0) &&
i < FEC_RESET_DELAY; i++)
udelay(1);
if (i == FEC_RESET_DELAY)
printk(KERN_WARNING DRV_MODULE_NAME
": %s FEC timeout on graceful transmit stop\n",
dev->name);
/*
* Disable FEC. Let only MII interrupts.
*/
FW(fecp, imask, 0);
FW(fecp, ecntrl, ~FEC_ECNTRL_ETHER_EN);
/*
* Reset SKB transmit buffers.
*/
for (i = 0; i < fep->tx_ring; i++) {
if ((skb = fep->tx_skbuff[i]) == NULL)
continue;
fep->tx_skbuff[i] = NULL;
dev_kfree_skb(skb);
}
/*
* Reset SKB receive buffers
*/
for (i = 0; i < fep->rx_ring; i++) {
if ((skb = fep->rx_skbuff[i]) == NULL)
continue;
fep->rx_skbuff[i] = NULL;
dev_kfree_skb(skb);
}
}
/* common receive function */
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
static int fec_enet_rx_common(struct fec_enet_private *ep,
struct net_device *dev, int budget)
{
fec_t *fecp = fep->fecp;
const struct fec_platform_info *fpi = fep->fpi;
cbd_t *bdp;
struct sk_buff *skb, *skbn, *skbt;
int received = 0;
__u16 pkt_len, sc;
int curidx;
/*
* First, grab all of the stats for the incoming packet.
* These get messed up if we get called due to a busy condition.
*/
bdp = fep->cur_rx;
/* clear RX status bits for napi*/
if (fpi->use_napi)
FW(fecp, ievent, FEC_ENET_RXF | FEC_ENET_RXB);
while (((sc = CBDR_SC(bdp)) & BD_ENET_RX_EMPTY) == 0) {
curidx = bdp - fep->rx_bd_base;
/*
* Since we have allocated space to hold a complete frame,
* the last indicator should be set.
*/
if ((sc & BD_ENET_RX_LAST) == 0)
printk(KERN_WARNING DRV_MODULE_NAME
": %s rcv is not +last\n",
dev->name);
/*
* Check for errors.
*/
if (sc & (BD_ENET_RX_LG | BD_ENET_RX_SH | BD_ENET_RX_CL |
BD_ENET_RX_NO | BD_ENET_RX_CR | BD_ENET_RX_OV)) {
fep->stats.rx_errors++;
/* Frame too long or too short. */
if (sc & (BD_ENET_RX_LG | BD_ENET_RX_SH))
fep->stats.rx_length_errors++;
/* Frame alignment */
if (sc & (BD_ENET_RX_NO | BD_ENET_RX_CL))
fep->stats.rx_frame_errors++;
/* CRC Error */
if (sc & BD_ENET_RX_CR)
fep->stats.rx_crc_errors++;
/* FIFO overrun */
if (sc & BD_ENET_RX_OV)
fep->stats.rx_crc_errors++;
skbn = fep->rx_skbuff[curidx];
BUG_ON(skbn == NULL);
} else {
skb = fep->rx_skbuff[curidx];
BUG_ON(skb == NULL);
/*
* Process the incoming frame.
*/
fep->stats.rx_packets++;
pkt_len = CBDR_DATLEN(bdp) - 4; /* remove CRC */
fep->stats.rx_bytes += pkt_len + 4;
if (pkt_len <= fpi->rx_copybreak) {
/* +2 to make IP header L1 cache aligned */
skbn = dev_alloc_skb(pkt_len + 2);
if (skbn != NULL) {
skb_reserve(skbn, 2); /* align IP header */
skb_copy_from_linear_data(skb,
skbn->data,
pkt_len);
/* swap */
skbt = skb;
skb = skbn;
skbn = skbt;
}
} else
skbn = dev_alloc_skb(ENET_RX_FRSIZE);
if (skbn != NULL) {
skb_put(skb, pkt_len); /* Make room */
skb->protocol = eth_type_trans(skb, dev);
received++;
if (!fpi->use_napi)
netif_rx(skb);
else
netif_receive_skb(skb);
} else {
printk(KERN_WARNING DRV_MODULE_NAME
": %s Memory squeeze, dropping packet.\n",
dev->name);
fep->stats.rx_dropped++;
skbn = skb;
}
}
fep->rx_skbuff[curidx] = skbn;
CBDW_BUFADDR(bdp, dma_map_single(NULL, skbn->data,
L1_CACHE_ALIGN(PKT_MAXBUF_SIZE),
DMA_FROM_DEVICE));
CBDW_DATLEN(bdp, 0);
CBDW_SC(bdp, (sc & ~BD_ENET_RX_STATS) | BD_ENET_RX_EMPTY);
/*
* Update BD pointer to next entry.
*/
if ((sc & BD_ENET_RX_WRAP) == 0)
bdp++;
else
bdp = fep->rx_bd_base;
/*
* Doing this here will keep the FEC running while we process
* incoming frames. On a heavily loaded network, we should be
* able to keep up at the expense of system resources.
*/
FW(fecp, r_des_active, 0x01000000);
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
if (received >= budget)
break;
}
fep->cur_rx = bdp;
if (fpi->use_napi) {
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
if (received < budget) {
netif_rx_complete(dev, &fep->napi);
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
/* enable RX interrupt bits */
FS(fecp, imask, FEC_ENET_RXF | FEC_ENET_RXB);
}
}
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
return received;
}
static void fec_enet_tx(struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
cbd_t *bdp;
struct sk_buff *skb;
int dirtyidx, do_wake;
__u16 sc;
spin_lock(&fep->lock);
bdp = fep->dirty_tx;
do_wake = 0;
while (((sc = CBDR_SC(bdp)) & BD_ENET_TX_READY) == 0) {
dirtyidx = bdp - fep->tx_bd_base;
if (fep->tx_free == fep->tx_ring)
break;
skb = fep->tx_skbuff[dirtyidx];
/*
* Check for errors.
*/
if (sc & (BD_ENET_TX_HB | BD_ENET_TX_LC |
BD_ENET_TX_RL | BD_ENET_TX_UN | BD_ENET_TX_CSL)) {
fep->stats.tx_errors++;
if (sc & BD_ENET_TX_HB) /* No heartbeat */
fep->stats.tx_heartbeat_errors++;
if (sc & BD_ENET_TX_LC) /* Late collision */
fep->stats.tx_window_errors++;
if (sc & BD_ENET_TX_RL) /* Retrans limit */
fep->stats.tx_aborted_errors++;
if (sc & BD_ENET_TX_UN) /* Underrun */
fep->stats.tx_fifo_errors++;
if (sc & BD_ENET_TX_CSL) /* Carrier lost */
fep->stats.tx_carrier_errors++;
} else
fep->stats.tx_packets++;
if (sc & BD_ENET_TX_READY)
printk(KERN_WARNING DRV_MODULE_NAME
": %s HEY! Enet xmit interrupt and TX_READY.\n",
dev->name);
/*
* Deferred means some collisions occurred during transmit,
* but we eventually sent the packet OK.
*/
if (sc & BD_ENET_TX_DEF)
fep->stats.collisions++;
/*
* Free the sk buffer associated with this last transmit.
*/
dev_kfree_skb_irq(skb);
fep->tx_skbuff[dirtyidx] = NULL;
/*
* Update pointer to next buffer descriptor to be transmitted.
*/
if ((sc & BD_ENET_TX_WRAP) == 0)
bdp++;
else
bdp = fep->tx_bd_base;
/*
* Since we have freed up a buffer, the ring is no longer
* full.
*/
if (!fep->tx_free++)
do_wake = 1;
}
fep->dirty_tx = bdp;
spin_unlock(&fep->lock);
if (do_wake && netif_queue_stopped(dev))
netif_wake_queue(dev);
}
/*
* The interrupt handler.
* This is called from the MPC core interrupt.
*/
static irqreturn_t
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 09:55:46 -04:00
fec_enet_interrupt(int irq, void *dev_id)
{
struct net_device *dev = dev_id;
struct fec_enet_private *fep;
const struct fec_platform_info *fpi;
fec_t *fecp;
__u32 int_events;
__u32 int_events_napi;
if (unlikely(dev == NULL))
return IRQ_NONE;
fep = netdev_priv(dev);
fecp = fep->fecp;
fpi = fep->fpi;
/*
* Get the interrupt events that caused us to be here.
*/
while ((int_events = FR(fecp, ievent) & FR(fecp, imask)) != 0) {
if (!fpi->use_napi)
FW(fecp, ievent, int_events);
else {
int_events_napi = int_events & ~(FEC_ENET_RXF | FEC_ENET_RXB);
FW(fecp, ievent, int_events_napi);
}
if ((int_events & (FEC_ENET_HBERR | FEC_ENET_BABR |
FEC_ENET_BABT | FEC_ENET_EBERR)) != 0)
printk(KERN_WARNING DRV_MODULE_NAME
": %s FEC ERROR(s) 0x%x\n",
dev->name, int_events);
if ((int_events & FEC_ENET_RXF) != 0) {
if (!fpi->use_napi)
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
fec_enet_rx_common(fep, dev, ~0);
else {
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
if (netif_rx_schedule_prep(dev, &fep->napi)) {
/* disable rx interrupts */
FC(fecp, imask, FEC_ENET_RXF | FEC_ENET_RXB);
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
__netif_rx_schedule(dev, &fep->napi);
} else {
printk(KERN_ERR DRV_MODULE_NAME
": %s driver bug! interrupt while in poll!\n",
dev->name);
FC(fecp, imask, FEC_ENET_RXF | FEC_ENET_RXB);
}
}
}
if ((int_events & FEC_ENET_TXF) != 0)
fec_enet_tx(dev);
}
return IRQ_HANDLED;
}
/* This interrupt occurs when the PHY detects a link change. */
static irqreturn_t
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 09:55:46 -04:00
fec_mii_link_interrupt(int irq, void *dev_id)
{
struct net_device *dev = dev_id;
struct fec_enet_private *fep;
const struct fec_platform_info *fpi;
if (unlikely(dev == NULL))
return IRQ_NONE;
fep = netdev_priv(dev);
fpi = fep->fpi;
if (!fpi->use_mdio)
return IRQ_NONE;
/*
* Acknowledge the interrupt if possible. If we have not
* found the PHY yet we can't process or acknowledge the
* interrupt now. Instead we ignore this interrupt for now,
* which we can do since it is edge triggered. It will be
* acknowledged later by fec_enet_open().
*/
if (!fep->phy)
return IRQ_NONE;
fec_mii_ack_int(dev);
fec_mii_link_status_change_check(dev, 0);
return IRQ_HANDLED;
}
/**********************************************************************************/
static int fec_enet_start_xmit(struct sk_buff *skb, struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
fec_t *fecp = fep->fecp;
cbd_t *bdp;
int curidx;
unsigned long flags;
spin_lock_irqsave(&fep->tx_lock, flags);
/*
* Fill in a Tx ring entry
*/
bdp = fep->cur_tx;
if (!fep->tx_free || (CBDR_SC(bdp) & BD_ENET_TX_READY)) {
netif_stop_queue(dev);
spin_unlock_irqrestore(&fep->tx_lock, flags);
/*
* Ooops. All transmit buffers are full. Bail out.
* This should not happen, since the tx queue should be stopped.
*/
printk(KERN_WARNING DRV_MODULE_NAME
": %s tx queue full!.\n", dev->name);
return 1;
}
curidx = bdp - fep->tx_bd_base;
/*
* Clear all of the status flags.
*/
CBDC_SC(bdp, BD_ENET_TX_STATS);
/*
* Save skb pointer.
*/
fep->tx_skbuff[curidx] = skb;
fep->stats.tx_bytes += skb->len;
/*
* Push the data cache so the CPM does not get stale memory data.
*/
CBDW_BUFADDR(bdp, dma_map_single(NULL, skb->data,
skb->len, DMA_TO_DEVICE));
CBDW_DATLEN(bdp, skb->len);
dev->trans_start = jiffies;
/*
* If this was the last BD in the ring, start at the beginning again.
*/
if ((CBDR_SC(bdp) & BD_ENET_TX_WRAP) == 0)
fep->cur_tx++;
else
fep->cur_tx = fep->tx_bd_base;
if (!--fep->tx_free)
netif_stop_queue(dev);
/*
* Trigger transmission start
*/
CBDS_SC(bdp, BD_ENET_TX_READY | BD_ENET_TX_INTR |
BD_ENET_TX_LAST | BD_ENET_TX_TC);
FW(fecp, x_des_active, 0x01000000);
spin_unlock_irqrestore(&fep->tx_lock, flags);
return 0;
}
static void fec_timeout(struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
fep->stats.tx_errors++;
if (fep->tx_free)
netif_wake_queue(dev);
/* check link status again */
fec_mii_link_status_change_check(dev, 0);
}
static int fec_enet_open(struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
const struct fec_platform_info *fpi = fep->fpi;
unsigned long flags;
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
napi_enable(&fep->napi);
/* Install our interrupt handler. */
if (request_irq(fpi->fec_irq, fec_enet_interrupt, 0, "fec", dev) != 0) {
printk(KERN_ERR DRV_MODULE_NAME
": %s Could not allocate FEC IRQ!", dev->name);
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
napi_disable(&fep->napi);
return -EINVAL;
}
/* Install our phy interrupt handler */
if (fpi->phy_irq != -1 &&
request_irq(fpi->phy_irq, fec_mii_link_interrupt, 0, "fec-phy",
dev) != 0) {
printk(KERN_ERR DRV_MODULE_NAME
": %s Could not allocate PHY IRQ!", dev->name);
free_irq(fpi->fec_irq, dev);
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
napi_disable(&fep->napi);
return -EINVAL;
}
if (fpi->use_mdio) {
fec_mii_startup(dev);
netif_carrier_off(dev);
fec_mii_link_status_change_check(dev, 1);
} else {
spin_lock_irqsave(&fep->lock, flags);
fec_restart(dev, 1, 100); /* XXX this sucks */
spin_unlock_irqrestore(&fep->lock, flags);
netif_carrier_on(dev);
netif_start_queue(dev);
}
return 0;
}
static int fec_enet_close(struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
const struct fec_platform_info *fpi = fep->fpi;
unsigned long flags;
netif_stop_queue(dev);
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
napi_disable(&fep->napi);
netif_carrier_off(dev);
if (fpi->use_mdio)
fec_mii_shutdown(dev);
spin_lock_irqsave(&fep->lock, flags);
fec_stop(dev);
spin_unlock_irqrestore(&fep->lock, flags);
/* release any irqs */
if (fpi->phy_irq != -1)
free_irq(fpi->phy_irq, dev);
free_irq(fpi->fec_irq, dev);
return 0;
}
static struct net_device_stats *fec_enet_get_stats(struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
return &fep->stats;
}
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
static int fec_enet_poll(struct napi_struct *napi, int budget)
{
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
struct fec_enet_private *fep = container_of(napi, struct fec_enet_private, napi);
struct net_device *dev = fep->dev;
return fec_enet_rx_common(fep, dev, budget);
}
/*************************************************************************/
static void fec_get_drvinfo(struct net_device *dev,
struct ethtool_drvinfo *info)
{
strcpy(info->driver, DRV_MODULE_NAME);
strcpy(info->version, DRV_MODULE_VERSION);
}
static int fec_get_regs_len(struct net_device *dev)
{
return sizeof(fec_t);
}
static void fec_get_regs(struct net_device *dev, struct ethtool_regs *regs,
void *p)
{
struct fec_enet_private *fep = netdev_priv(dev);
unsigned long flags;
if (regs->len < sizeof(fec_t))
return;
regs->version = 0;
spin_lock_irqsave(&fep->lock, flags);
memcpy_fromio(p, fep->fecp, sizeof(fec_t));
spin_unlock_irqrestore(&fep->lock, flags);
}
static int fec_get_settings(struct net_device *dev, struct ethtool_cmd *cmd)
{
struct fec_enet_private *fep = netdev_priv(dev);
unsigned long flags;
int rc;
spin_lock_irqsave(&fep->lock, flags);
rc = mii_ethtool_gset(&fep->mii_if, cmd);
spin_unlock_irqrestore(&fep->lock, flags);
return rc;
}
static int fec_set_settings(struct net_device *dev, struct ethtool_cmd *cmd)
{
struct fec_enet_private *fep = netdev_priv(dev);
unsigned long flags;
int rc;
spin_lock_irqsave(&fep->lock, flags);
rc = mii_ethtool_sset(&fep->mii_if, cmd);
spin_unlock_irqrestore(&fep->lock, flags);
return rc;
}
static int fec_nway_reset(struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
return mii_nway_restart(&fep->mii_if);
}
static __u32 fec_get_msglevel(struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
return fep->msg_enable;
}
static void fec_set_msglevel(struct net_device *dev, __u32 value)
{
struct fec_enet_private *fep = netdev_priv(dev);
fep->msg_enable = value;
}
static const struct ethtool_ops fec_ethtool_ops = {
.get_drvinfo = fec_get_drvinfo,
.get_regs_len = fec_get_regs_len,
.get_settings = fec_get_settings,
.set_settings = fec_set_settings,
.nway_reset = fec_nway_reset,
.get_link = ethtool_op_get_link,
.get_msglevel = fec_get_msglevel,
.set_msglevel = fec_set_msglevel,
.set_tx_csum = ethtool_op_set_tx_csum, /* local! */
.set_sg = ethtool_op_set_sg,
.get_regs = fec_get_regs,
};
static int fec_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
{
struct fec_enet_private *fep = netdev_priv(dev);
struct mii_ioctl_data *mii = (struct mii_ioctl_data *)&rq->ifr_data;
unsigned long flags;
int rc;
if (!netif_running(dev))
return -EINVAL;
spin_lock_irqsave(&fep->lock, flags);
rc = generic_mii_ioctl(&fep->mii_if, mii, cmd, NULL);
spin_unlock_irqrestore(&fep->lock, flags);
return rc;
}
int fec_8xx_init_one(const struct fec_platform_info *fpi,
struct net_device **devp)
{
immap_t *immap = (immap_t *) IMAP_ADDR;
static int fec_8xx_version_printed = 0;
struct net_device *dev = NULL;
struct fec_enet_private *fep = NULL;
fec_t *fecp = NULL;
int i;
int err = 0;
int registered = 0;
__u32 siel;
*devp = NULL;
switch (fpi->fec_no) {
case 0:
fecp = &((immap_t *) IMAP_ADDR)->im_cpm.cp_fec;
break;
#ifdef CONFIG_DUET
case 1:
fecp = &((immap_t *) IMAP_ADDR)->im_cpm.cp_fec2;
break;
#endif
default:
return -EINVAL;
}
if (fec_8xx_version_printed++ == 0)
printk(KERN_INFO "%s", version);
i = sizeof(*fep) + (sizeof(struct sk_buff **) *
(fpi->rx_ring + fpi->tx_ring));
dev = alloc_etherdev(i);
if (!dev) {
err = -ENOMEM;
goto err;
}
fep = netdev_priv(dev);
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
fep->dev = dev;
/* partial reset of FEC */
fec_whack_reset(fecp);
/* point rx_skbuff, tx_skbuff */
fep->rx_skbuff = (struct sk_buff **)&fep[1];
fep->tx_skbuff = fep->rx_skbuff + fpi->rx_ring;
fep->fecp = fecp;
fep->fpi = fpi;
/* init locks */
spin_lock_init(&fep->lock);
spin_lock_init(&fep->tx_lock);
/*
* Set the Ethernet address.
*/
for (i = 0; i < 6; i++)
dev->dev_addr[i] = fpi->macaddr[i];
fep->ring_base = dma_alloc_coherent(NULL,
(fpi->tx_ring + fpi->rx_ring) *
sizeof(cbd_t), &fep->ring_mem_addr,
GFP_KERNEL);
if (fep->ring_base == NULL) {
printk(KERN_ERR DRV_MODULE_NAME
": %s dma alloc failed.\n", dev->name);
err = -ENOMEM;
goto err;
}
/*
* Set receive and transmit descriptor base.
*/
fep->rx_bd_base = fep->ring_base;
fep->tx_bd_base = fep->rx_bd_base + fpi->rx_ring;
/* initialize ring size variables */
fep->tx_ring = fpi->tx_ring;
fep->rx_ring = fpi->rx_ring;
/* SIU interrupt */
if (fpi->phy_irq != -1 &&
(fpi->phy_irq >= SIU_IRQ0 && fpi->phy_irq < SIU_LEVEL7)) {
siel = in_be32(&immap->im_siu_conf.sc_siel);
if ((fpi->phy_irq & 1) == 0)
siel |= (0x80000000 >> fpi->phy_irq);
else
siel &= ~(0x80000000 >> (fpi->phy_irq & ~1));
out_be32(&immap->im_siu_conf.sc_siel, siel);
}
/*
* The FEC Ethernet specific entries in the device structure.
*/
dev->open = fec_enet_open;
dev->hard_start_xmit = fec_enet_start_xmit;
dev->tx_timeout = fec_timeout;
dev->watchdog_timeo = TX_TIMEOUT;
dev->stop = fec_enet_close;
dev->get_stats = fec_enet_get_stats;
dev->set_multicast_list = fec_set_multicast_list;
dev->set_mac_address = fec_set_mac_address;
[NET]: Make NAPI polling independent of struct net_device objects. Several devices have multiple independant RX queues per net device, and some have a single interrupt doorbell for several queues. In either case, it's easier to support layouts like that if the structure representing the poll is independant from the net device itself. The signature of the ->poll() call back goes from: int foo_poll(struct net_device *dev, int *budget) to int foo_poll(struct napi_struct *napi, int budget) The caller is returned the number of RX packets processed (or the number of "NAPI credits" consumed if you want to get abstract). The callee no longer messes around bumping dev->quota, *budget, etc. because that is all handled in the caller upon return. The napi_struct is to be embedded in the device driver private data structures. Furthermore, it is the driver's responsibility to disable all NAPI instances in it's ->stop() device close handler. Since the napi_struct is privatized into the driver's private data structures, only the driver knows how to get at all of the napi_struct instances it may have per-device. With lots of help and suggestions from Rusty Russell, Roland Dreier, Michael Chan, Jeff Garzik, and Jamal Hadi Salim. Bug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra, Joseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan. [ Ported to current tree and all drivers converted. Integrated Stephen's follow-on kerneldoc additions, and restored poll_list handling to the old style to fix mutual exclusion issues. -DaveM ] Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-03 19:41:36 -04:00
netif_napi_add(dev, &fec->napi,
fec_enet_poll, fpi->napi_weight);
dev->ethtool_ops = &fec_ethtool_ops;
dev->do_ioctl = fec_ioctl;
fep->fec_phy_speed =
((((fpi->sys_clk + 4999999) / 2500000) / 2) & 0x3F) << 1;
init_timer(&fep->phy_timer_list);
/* partial reset of FEC so that only MII works */
FW(fecp, mii_speed, fep->fec_phy_speed);
FW(fecp, ievent, 0xffc0);
FW(fecp, ivec, (fpi->fec_irq / 2) << 29);
FW(fecp, imask, 0);
FW(fecp, r_cntrl, FEC_RCNTRL_MII_MODE); /* MII enable */
FW(fecp, ecntrl, FEC_ECNTRL_PINMUX | FEC_ECNTRL_ETHER_EN);
netif_carrier_off(dev);
err = register_netdev(dev);
if (err != 0)
goto err;
registered = 1;
if (fpi->use_mdio) {
fep->mii_if.dev = dev;
fep->mii_if.mdio_read = fec_mii_read;
fep->mii_if.mdio_write = fec_mii_write;
fep->mii_if.phy_id_mask = 0x1f;
fep->mii_if.reg_num_mask = 0x1f;
fep->mii_if.phy_id = fec_mii_phy_id_detect(dev);
}
*devp = dev;
return 0;
err:
if (dev != NULL) {
if (fecp != NULL)
fec_whack_reset(fecp);
if (registered)
unregister_netdev(dev);
if (fep != NULL) {
if (fep->ring_base)
dma_free_coherent(NULL,
(fpi->tx_ring +
fpi->rx_ring) *
sizeof(cbd_t), fep->ring_base,
fep->ring_mem_addr);
}
free_netdev(dev);
}
return err;
}
int fec_8xx_cleanup_one(struct net_device *dev)
{
struct fec_enet_private *fep = netdev_priv(dev);
fec_t *fecp = fep->fecp;
const struct fec_platform_info *fpi = fep->fpi;
fec_whack_reset(fecp);
unregister_netdev(dev);
dma_free_coherent(NULL, (fpi->tx_ring + fpi->rx_ring) * sizeof(cbd_t),
fep->ring_base, fep->ring_mem_addr);
free_netdev(dev);
return 0;
}
/**************************************************************************************/
/**************************************************************************************/
/**************************************************************************************/
static int __init fec_8xx_init(void)
{
return fec_8xx_platform_init();
}
static void __exit fec_8xx_cleanup(void)
{
fec_8xx_platform_cleanup();
}
/**************************************************************************************/
/**************************************************************************************/
/**************************************************************************************/
module_init(fec_8xx_init);
module_exit(fec_8xx_cleanup);