android_kernel_xiaomi_sm8350/drivers/net/apne.c

644 lines
17 KiB
C
Raw Normal View History

/*
* Amiga Linux/68k 8390 based PCMCIA Ethernet Driver for the Amiga 1200
*
* (C) Copyright 1997 Alain Malek
* (Alain.Malek@cryogen.com)
*
* ----------------------------------------------------------------------------
*
* This program is based on
*
* ne.c: A general non-shared-memory NS8390 ethernet driver for linux
* Written 1992-94 by Donald Becker.
*
* 8390.c: A general NS8390 ethernet driver core for linux.
* Written 1992-94 by Donald Becker.
*
* cnetdevice: A Sana-II ethernet driver for AmigaOS
* Written by Bruce Abbott (bhabbott@inhb.co.nz)
*
* ----------------------------------------------------------------------------
*
* This file is subject to the terms and conditions of the GNU General Public
* License. See the file COPYING in the main directory of the Linux
* distribution for more details.
*
* ----------------------------------------------------------------------------
*
*/
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/pci.h>
#include <linux/init.h>
#include <linux/delay.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/jiffies.h>
#include <asm/system.h>
#include <asm/io.h>
#include <asm/setup.h>
#include <asm/amigaints.h>
#include <asm/amigahw.h>
#include <asm/amigayle.h>
#include <asm/amipcmcia.h>
#include "8390.h"
/* ---- No user-serviceable parts below ---- */
#define DRV_NAME "apne"
#define NE_BASE (dev->base_addr)
#define NE_CMD 0x00
#define NE_DATAPORT 0x10 /* NatSemi-defined port window offset. */
#define NE_RESET 0x1f /* Issue a read to reset, a write to clear. */
#define NE_IO_EXTENT 0x20
#define NE_EN0_ISR 0x07
#define NE_EN0_DCFG 0x0e
#define NE_EN0_RSARLO 0x08
#define NE_EN0_RSARHI 0x09
#define NE_EN0_RCNTLO 0x0a
#define NE_EN0_RXCR 0x0c
#define NE_EN0_TXCR 0x0d
#define NE_EN0_RCNTHI 0x0b
#define NE_EN0_IMR 0x0f
#define NE1SM_START_PG 0x20 /* First page of TX buffer */
#define NE1SM_STOP_PG 0x40 /* Last page +1 of RX ring */
#define NESM_START_PG 0x40 /* First page of TX buffer */
#define NESM_STOP_PG 0x80 /* Last page +1 of RX ring */
struct net_device * __init apne_probe(int unit);
static int apne_probe1(struct net_device *dev, int ioaddr);
static int apne_open(struct net_device *dev);
static int apne_close(struct net_device *dev);
static void apne_reset_8390(struct net_device *dev);
static void apne_get_8390_hdr(struct net_device *dev, struct e8390_pkt_hdr *hdr,
int ring_page);
static void apne_block_input(struct net_device *dev, int count,
struct sk_buff *skb, int ring_offset);
static void apne_block_output(struct net_device *dev, const int count,
const unsigned char *buf, const int start_page);
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
static irqreturn_t apne_interrupt(int irq, void *dev_id);
static int init_pcmcia(void);
/* IO base address used for nic */
#define IOBASE 0x300
/*
use MANUAL_CONFIG and MANUAL_OFFSET for enabling IO by hand
you can find the values to use by looking at the cnet.device
config file example (the default values are for the CNET40BC card)
*/
/*
#define MANUAL_CONFIG 0x20
#define MANUAL_OFFSET 0x3f8
#define MANUAL_HWADDR0 0x00
#define MANUAL_HWADDR1 0x12
#define MANUAL_HWADDR2 0x34
#define MANUAL_HWADDR3 0x56
#define MANUAL_HWADDR4 0x78
#define MANUAL_HWADDR5 0x9a
*/
static const char version[] =
"apne.c:v1.1 7/10/98 Alain Malek (Alain.Malek@cryogen.ch)\n";
static int apne_owned; /* signal if card already owned */
struct net_device * __init apne_probe(int unit)
{
struct net_device *dev;
#ifndef MANUAL_CONFIG
char tuple[8];
#endif
int err;
if (!MACH_IS_AMIGA)
return ERR_PTR(-ENODEV);
if (apne_owned)
return ERR_PTR(-ENODEV);
if ( !(AMIGAHW_PRESENT(PCMCIA)) )
return ERR_PTR(-ENODEV);
printk("Looking for PCMCIA ethernet card : ");
/* check if a card is inserted */
if (!(PCMCIA_INSERTED)) {
printk("NO PCMCIA card inserted\n");
return ERR_PTR(-ENODEV);
}
dev = alloc_ei_netdev();
if (!dev)
return ERR_PTR(-ENOMEM);
if (unit >= 0) {
sprintf(dev->name, "eth%d", unit);
netdev_boot_setup_check(dev);
}
/* disable pcmcia irq for readtuple */
pcmcia_disable_irq();
#ifndef MANUAL_CONFIG
if ((pcmcia_copy_tuple(CISTPL_FUNCID, tuple, 8) < 3) ||
(tuple[2] != CISTPL_FUNCID_NETWORK)) {
printk("not an ethernet card\n");
/* XXX: shouldn't we re-enable irq here? */
free_netdev(dev);
return ERR_PTR(-ENODEV);
}
#endif
printk("ethernet PCMCIA card inserted\n");
if (!init_pcmcia()) {
/* XXX: shouldn't we re-enable irq here? */
free_netdev(dev);
return ERR_PTR(-ENODEV);
}
if (!request_region(IOBASE, 0x20, DRV_NAME)) {
free_netdev(dev);
return ERR_PTR(-EBUSY);
}
err = apne_probe1(dev, IOBASE);
if (err) {
release_region(IOBASE, 0x20);
free_netdev(dev);
return ERR_PTR(err);
}
err = register_netdev(dev);
if (!err)
return dev;
pcmcia_disable_irq();
free_irq(IRQ_AMIGA_PORTS, dev);
pcmcia_reset();
release_region(IOBASE, 0x20);
free_netdev(dev);
return ERR_PTR(err);
}
static int __init apne_probe1(struct net_device *dev, int ioaddr)
{
int i;
unsigned char SA_prom[32];
int wordlength = 2;
const char *name = NULL;
int start_page, stop_page;
#ifndef MANUAL_HWADDR0
int neX000, ctron;
#endif
static unsigned version_printed;
DECLARE_MAC_BUF(mac);
if (ei_debug && version_printed++ == 0)
printk(version);
printk("PCMCIA NE*000 ethercard probe");
/* Reset card. Who knows what dain-bramaged state it was left in. */
{ unsigned long reset_start_time = jiffies;
outb(inb(ioaddr + NE_RESET), ioaddr + NE_RESET);
while ((inb(ioaddr + NE_EN0_ISR) & ENISR_RESET) == 0)
if (time_after(jiffies, reset_start_time + 2*HZ/100)) {
printk(" not found (no reset ack).\n");
return -ENODEV;
}
outb(0xff, ioaddr + NE_EN0_ISR); /* Ack all intr. */
}
#ifndef MANUAL_HWADDR0
/* Read the 16 bytes of station address PROM.
We must first initialize registers, similar to NS8390_init(eifdev, 0).
We can't reliably read the SAPROM address without this.
(I learned the hard way!). */
{
struct {unsigned long value, offset; } program_seq[] = {
{E8390_NODMA+E8390_PAGE0+E8390_STOP, NE_CMD}, /* Select page 0*/
{0x48, NE_EN0_DCFG}, /* Set byte-wide (0x48) access. */
{0x00, NE_EN0_RCNTLO}, /* Clear the count regs. */
{0x00, NE_EN0_RCNTHI},
{0x00, NE_EN0_IMR}, /* Mask completion irq. */
{0xFF, NE_EN0_ISR},
{E8390_RXOFF, NE_EN0_RXCR}, /* 0x20 Set to monitor */
{E8390_TXOFF, NE_EN0_TXCR}, /* 0x02 and loopback mode. */
{32, NE_EN0_RCNTLO},
{0x00, NE_EN0_RCNTHI},
{0x00, NE_EN0_RSARLO}, /* DMA starting at 0x0000. */
{0x00, NE_EN0_RSARHI},
{E8390_RREAD+E8390_START, NE_CMD},
};
for (i = 0; i < ARRAY_SIZE(program_seq); i++) {
outb(program_seq[i].value, ioaddr + program_seq[i].offset);
}
}
for(i = 0; i < 32 /*sizeof(SA_prom)*/; i+=2) {
SA_prom[i] = inb(ioaddr + NE_DATAPORT);
SA_prom[i+1] = inb(ioaddr + NE_DATAPORT);
if (SA_prom[i] != SA_prom[i+1])
wordlength = 1;
}
/* At this point, wordlength *only* tells us if the SA_prom is doubled
up or not because some broken PCI cards don't respect the byte-wide
request in program_seq above, and hence don't have doubled up values.
These broken cards would otherwise be detected as an ne1000. */
if (wordlength == 2)
for (i = 0; i < 16; i++)
SA_prom[i] = SA_prom[i+i];
if (wordlength == 2) {
/* We must set the 8390 for word mode. */
outb(0x49, ioaddr + NE_EN0_DCFG);
start_page = NESM_START_PG;
stop_page = NESM_STOP_PG;
} else {
start_page = NE1SM_START_PG;
stop_page = NE1SM_STOP_PG;
}
neX000 = (SA_prom[14] == 0x57 && SA_prom[15] == 0x57);
ctron = (SA_prom[0] == 0x00 && SA_prom[1] == 0x00 && SA_prom[2] == 0x1d);
/* Set up the rest of the parameters. */
if (neX000) {
name = (wordlength == 2) ? "NE2000" : "NE1000";
} else if (ctron) {
name = (wordlength == 2) ? "Ctron-8" : "Ctron-16";
start_page = 0x01;
stop_page = (wordlength == 2) ? 0x40 : 0x20;
} else {
printk(" not found.\n");
return -ENXIO;
}
#else
wordlength = 2;
/* We must set the 8390 for word mode. */
outb(0x49, ioaddr + NE_EN0_DCFG);
start_page = NESM_START_PG;
stop_page = NESM_STOP_PG;
SA_prom[0] = MANUAL_HWADDR0;
SA_prom[1] = MANUAL_HWADDR1;
SA_prom[2] = MANUAL_HWADDR2;
SA_prom[3] = MANUAL_HWADDR3;
SA_prom[4] = MANUAL_HWADDR4;
SA_prom[5] = MANUAL_HWADDR5;
name = "NE2000";
#endif
dev->base_addr = ioaddr;
dev->irq = IRQ_AMIGA_PORTS;
/* Install the Interrupt handler */
i = request_irq(dev->irq, apne_interrupt, IRQF_SHARED, DRV_NAME, dev);
if (i) return i;
for(i = 0; i < ETHER_ADDR_LEN; i++)
dev->dev_addr[i] = SA_prom[i];
printk(" %s\n", print_mac(mac, dev->dev_addr));
printk("%s: %s found.\n", dev->name, name);
ei_status.name = name;
ei_status.tx_start_page = start_page;
ei_status.stop_page = stop_page;
ei_status.word16 = (wordlength == 2);
ei_status.rx_start_page = start_page + TX_PAGES;
ei_status.reset_8390 = &apne_reset_8390;
ei_status.block_input = &apne_block_input;
ei_status.block_output = &apne_block_output;
ei_status.get_8390_hdr = &apne_get_8390_hdr;
dev->open = &apne_open;
dev->stop = &apne_close;
#ifdef CONFIG_NET_POLL_CONTROLLER
dev->poll_controller = ei_poll;
#endif
NS8390_init(dev, 0);
pcmcia_ack_int(pcmcia_get_intreq()); /* ack PCMCIA int req */
pcmcia_enable_irq();
apne_owned = 1;
return 0;
}
static int
apne_open(struct net_device *dev)
{
ei_open(dev);
return 0;
}
static int
apne_close(struct net_device *dev)
{
if (ei_debug > 1)
printk("%s: Shutting down ethercard.\n", dev->name);
ei_close(dev);
return 0;
}
/* Hard reset the card. This used to pause for the same period that a
8390 reset command required, but that shouldn't be necessary. */
static void
apne_reset_8390(struct net_device *dev)
{
unsigned long reset_start_time = jiffies;
init_pcmcia();
if (ei_debug > 1) printk("resetting the 8390 t=%ld...", jiffies);
outb(inb(NE_BASE + NE_RESET), NE_BASE + NE_RESET);
ei_status.txing = 0;
ei_status.dmaing = 0;
/* This check _should_not_ be necessary, omit eventually. */
while ((inb(NE_BASE+NE_EN0_ISR) & ENISR_RESET) == 0)
if (time_after(jiffies, reset_start_time + 2*HZ/100)) {
printk("%s: ne_reset_8390() did not complete.\n", dev->name);
break;
}
outb(ENISR_RESET, NE_BASE + NE_EN0_ISR); /* Ack intr. */
}
/* Grab the 8390 specific header. Similar to the block_input routine, but
we don't need to be concerned with ring wrap as the header will be at
the start of a page, so we optimize accordingly. */
static void
apne_get_8390_hdr(struct net_device *dev, struct e8390_pkt_hdr *hdr, int ring_page)
{
int nic_base = dev->base_addr;
int cnt;
char *ptrc;
short *ptrs;
/* This *shouldn't* happen. If it does, it's the last thing you'll see */
if (ei_status.dmaing) {
printk("%s: DMAing conflict in ne_get_8390_hdr "
"[DMAstat:%d][irqlock:%d][intr:%d].\n",
dev->name, ei_status.dmaing, ei_status.irqlock, dev->irq);
return;
}
ei_status.dmaing |= 0x01;
outb(E8390_NODMA+E8390_PAGE0+E8390_START, nic_base+ NE_CMD);
outb(ENISR_RDC, nic_base + NE_EN0_ISR);
outb(sizeof(struct e8390_pkt_hdr), nic_base + NE_EN0_RCNTLO);
outb(0, nic_base + NE_EN0_RCNTHI);
outb(0, nic_base + NE_EN0_RSARLO); /* On page boundary */
outb(ring_page, nic_base + NE_EN0_RSARHI);
outb(E8390_RREAD+E8390_START, nic_base + NE_CMD);
if (ei_status.word16) {
ptrs = (short*)hdr;
for(cnt = 0; cnt < (sizeof(struct e8390_pkt_hdr)>>1); cnt++)
*ptrs++ = inw(NE_BASE + NE_DATAPORT);
} else {
ptrc = (char*)hdr;
for(cnt = 0; cnt < sizeof(struct e8390_pkt_hdr); cnt++)
*ptrc++ = inb(NE_BASE + NE_DATAPORT);
}
outb(ENISR_RDC, nic_base + NE_EN0_ISR); /* Ack intr. */
ei_status.dmaing &= ~0x01;
le16_to_cpus(&hdr->count);
}
/* Block input and output, similar to the Crynwr packet driver. If you
are porting to a new ethercard, look at the packet driver source for hints.
The NEx000 doesn't share the on-board packet memory -- you have to put
the packet out through the "remote DMA" dataport using outb. */
static void
apne_block_input(struct net_device *dev, int count, struct sk_buff *skb, int ring_offset)
{
int nic_base = dev->base_addr;
char *buf = skb->data;
char *ptrc;
short *ptrs;
int cnt;
/* This *shouldn't* happen. If it does, it's the last thing you'll see */
if (ei_status.dmaing) {
printk("%s: DMAing conflict in ne_block_input "
"[DMAstat:%d][irqlock:%d][intr:%d].\n",
dev->name, ei_status.dmaing, ei_status.irqlock, dev->irq);
return;
}
ei_status.dmaing |= 0x01;
outb(E8390_NODMA+E8390_PAGE0+E8390_START, nic_base+ NE_CMD);
outb(ENISR_RDC, nic_base + NE_EN0_ISR);
outb(count & 0xff, nic_base + NE_EN0_RCNTLO);
outb(count >> 8, nic_base + NE_EN0_RCNTHI);
outb(ring_offset & 0xff, nic_base + NE_EN0_RSARLO);
outb(ring_offset >> 8, nic_base + NE_EN0_RSARHI);
outb(E8390_RREAD+E8390_START, nic_base + NE_CMD);
if (ei_status.word16) {
ptrs = (short*)buf;
for (cnt = 0; cnt < (count>>1); cnt++)
*ptrs++ = inw(NE_BASE + NE_DATAPORT);
if (count & 0x01) {
buf[count-1] = inb(NE_BASE + NE_DATAPORT);
}
} else {
ptrc = (char*)buf;
for (cnt = 0; cnt < count; cnt++)
*ptrc++ = inb(NE_BASE + NE_DATAPORT);
}
outb(ENISR_RDC, nic_base + NE_EN0_ISR); /* Ack intr. */
ei_status.dmaing &= ~0x01;
}
static void
apne_block_output(struct net_device *dev, int count,
const unsigned char *buf, const int start_page)
{
int nic_base = NE_BASE;
unsigned long dma_start;
char *ptrc;
short *ptrs;
int cnt;
/* Round the count up for word writes. Do we need to do this?
What effect will an odd byte count have on the 8390?
I should check someday. */
if (ei_status.word16 && (count & 0x01))
count++;
/* This *shouldn't* happen. If it does, it's the last thing you'll see */
if (ei_status.dmaing) {
printk("%s: DMAing conflict in ne_block_output."
"[DMAstat:%d][irqlock:%d][intr:%d]\n",
dev->name, ei_status.dmaing, ei_status.irqlock, dev->irq);
return;
}
ei_status.dmaing |= 0x01;
/* We should already be in page 0, but to be safe... */
outb(E8390_PAGE0+E8390_START+E8390_NODMA, nic_base + NE_CMD);
outb(ENISR_RDC, nic_base + NE_EN0_ISR);
/* Now the normal output. */
outb(count & 0xff, nic_base + NE_EN0_RCNTLO);
outb(count >> 8, nic_base + NE_EN0_RCNTHI);
outb(0x00, nic_base + NE_EN0_RSARLO);
outb(start_page, nic_base + NE_EN0_RSARHI);
outb(E8390_RWRITE+E8390_START, nic_base + NE_CMD);
if (ei_status.word16) {
ptrs = (short*)buf;
for (cnt = 0; cnt < count>>1; cnt++)
outw(*ptrs++, NE_BASE+NE_DATAPORT);
} else {
ptrc = (char*)buf;
for (cnt = 0; cnt < count; cnt++)
outb(*ptrc++, NE_BASE + NE_DATAPORT);
}
dma_start = jiffies;
while ((inb(NE_BASE + NE_EN0_ISR) & ENISR_RDC) == 0)
if (time_after(jiffies, dma_start + 2*HZ/100)) { /* 20ms */
printk("%s: timeout waiting for Tx RDC.\n", dev->name);
apne_reset_8390(dev);
NS8390_init(dev,1);
break;
}
outb(ENISR_RDC, nic_base + NE_EN0_ISR); /* Ack intr. */
ei_status.dmaing &= ~0x01;
return;
}
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
static irqreturn_t apne_interrupt(int irq, void *dev_id)
{
unsigned char pcmcia_intreq;
if (!(gayle.inten & GAYLE_IRQ_IRQ))
return IRQ_NONE;
pcmcia_intreq = pcmcia_get_intreq();
if (!(pcmcia_intreq & GAYLE_IRQ_IRQ)) {
pcmcia_ack_int(pcmcia_intreq);
return IRQ_NONE;
}
if (ei_debug > 3)
printk("pcmcia intreq = %x\n", pcmcia_intreq);
pcmcia_disable_irq(); /* to get rid of the sti() within ei_interrupt */
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
ei_interrupt(irq, dev_id);
pcmcia_ack_int(pcmcia_get_intreq());
pcmcia_enable_irq();
return IRQ_HANDLED;
}
#ifdef MODULE
static struct net_device *apne_dev;
static int __init apne_module_init(void)
{
apne_dev = apne_probe(-1);
if (IS_ERR(apne_dev))
return PTR_ERR(apne_dev);
return 0;
}
static void __exit apne_module_exit(void)
{
unregister_netdev(apne_dev);
pcmcia_disable_irq();
free_irq(IRQ_AMIGA_PORTS, apne_dev);
pcmcia_reset();
release_region(IOBASE, 0x20);
free_netdev(apne_dev);
}
module_init(apne_module_init);
module_exit(apne_module_exit);
#endif
static int init_pcmcia(void)
{
u_char config;
#ifndef MANUAL_CONFIG
u_char tuple[32];
int offset_len;
#endif
u_long offset;
pcmcia_reset();
pcmcia_program_voltage(PCMCIA_0V);
pcmcia_access_speed(PCMCIA_SPEED_250NS);
pcmcia_write_enable();
#ifdef MANUAL_CONFIG
config = MANUAL_CONFIG;
#else
/* get and write config byte to enable IO port */
if (pcmcia_copy_tuple(CISTPL_CFTABLE_ENTRY, tuple, 32) < 3)
return 0;
config = tuple[2] & 0x3f;
#endif
#ifdef MANUAL_OFFSET
offset = MANUAL_OFFSET;
#else
if (pcmcia_copy_tuple(CISTPL_CONFIG, tuple, 32) < 6)
return 0;
offset_len = (tuple[2] & 0x3) + 1;
offset = 0;
while(offset_len--) {
offset = (offset << 8) | tuple[4+offset_len];
}
#endif
out_8(GAYLE_ATTRIBUTE+offset, config);
return 1;
}
MODULE_LICENSE("GPL");