mirror of
https://github.com/openwrt/openwrt.git
synced 2024-12-25 08:21:14 +00:00
ec9ffe892f
Signed-off-by: Daniel Golle <daniel@makrotopia.org> SVN-Revision: 47214
316 lines
6.7 KiB
C
316 lines
6.7 KiB
C
/*
|
|
* arch/arm/mach-ox820/platsmp.c
|
|
*
|
|
* Copyright (C) 2002 ARM Ltd.
|
|
* All Rights Reserved
|
|
*
|
|
* This program is free software; you can redistribute it and/or modify
|
|
* it under the terms of the GNU General Public License version 2 as
|
|
* published by the Free Software Foundation.
|
|
*/
|
|
#include <linux/init.h>
|
|
#include <linux/device.h>
|
|
#include <linux/jiffies.h>
|
|
#include <linux/smp.h>
|
|
#include <linux/io.h>
|
|
#include <linux/dma-mapping.h>
|
|
#include <linux/cache.h>
|
|
#include <asm/cacheflush.h>
|
|
#include <asm/smp_scu.h>
|
|
#include <asm/tlbflush.h>
|
|
#include <asm/cputype.h>
|
|
#include <linux/delay.h>
|
|
#include <asm/fiq.h>
|
|
|
|
#include <linux/irqchip/arm-gic.h>
|
|
#include <mach/iomap.h>
|
|
#include <mach/smp.h>
|
|
#include <mach/hardware.h>
|
|
#include <mach/irqs.h>
|
|
|
|
#ifdef CONFIG_DMA_CACHE_FIQ_BROADCAST
|
|
|
|
#define FIQ_GENERATE 0x00000002
|
|
#define OXNAS_MAP_AREA 0x01000000
|
|
#define OXNAS_UNMAP_AREA 0x02000000
|
|
#define OXNAS_FLUSH_RANGE 0x03000000
|
|
|
|
struct fiq_req {
|
|
union {
|
|
struct {
|
|
const void *addr;
|
|
size_t size;
|
|
} map;
|
|
struct {
|
|
const void *addr;
|
|
size_t size;
|
|
} unmap;
|
|
struct {
|
|
const void *start;
|
|
const void *end;
|
|
} flush;
|
|
};
|
|
volatile uint flags;
|
|
void __iomem *reg;
|
|
} ____cacheline_aligned;
|
|
|
|
static struct fiq_handler fh = {
|
|
.name = "oxnas-fiq"
|
|
};
|
|
|
|
DEFINE_PER_CPU(struct fiq_req, fiq_data);
|
|
|
|
static inline void ox820_set_fiq_regs(unsigned int cpu)
|
|
{
|
|
struct pt_regs FIQ_regs;
|
|
struct fiq_req *fiq_req = &per_cpu(fiq_data, !cpu);
|
|
|
|
FIQ_regs.ARM_r8 = 0;
|
|
FIQ_regs.ARM_ip = (unsigned int)fiq_req;
|
|
FIQ_regs.ARM_sp = (int)(cpu ? RPSC_IRQ_SOFT : RPSA_IRQ_SOFT);
|
|
fiq_req->reg = cpu ? RPSC_IRQ_SOFT : RPSA_IRQ_SOFT;
|
|
|
|
set_fiq_regs(&FIQ_regs);
|
|
}
|
|
|
|
static void __init ox820_init_fiq(void)
|
|
{
|
|
void *fiqhandler_start;
|
|
unsigned int fiqhandler_length;
|
|
int ret;
|
|
|
|
fiqhandler_start = &ox820_fiq_start;
|
|
fiqhandler_length = &ox820_fiq_end - &ox820_fiq_start;
|
|
|
|
ret = claim_fiq(&fh);
|
|
|
|
if (ret)
|
|
return;
|
|
|
|
set_fiq_handler(fiqhandler_start, fiqhandler_length);
|
|
|
|
writel(IRQ_SOFT, RPSA_FIQ_IRQ_TO_FIQ);
|
|
writel(1, RPSA_FIQ_ENABLE);
|
|
writel(IRQ_SOFT, RPSC_FIQ_IRQ_TO_FIQ);
|
|
writel(1, RPSC_FIQ_ENABLE);
|
|
}
|
|
|
|
void fiq_dma_map_area(const void *addr, size_t size, int dir)
|
|
{
|
|
unsigned long flags;
|
|
struct fiq_req *req;
|
|
|
|
raw_local_irq_save(flags);
|
|
/* currently, not possible to take cpu0 down, so only check cpu1 */
|
|
if (!cpu_online(1)) {
|
|
raw_local_irq_restore(flags);
|
|
v6_dma_map_area(addr, size, dir);
|
|
return;
|
|
}
|
|
|
|
req = this_cpu_ptr(&fiq_data);
|
|
req->map.addr = addr;
|
|
req->map.size = size;
|
|
req->flags = dir | OXNAS_MAP_AREA;
|
|
smp_mb();
|
|
|
|
writel_relaxed(FIQ_GENERATE, req->reg);
|
|
|
|
v6_dma_map_area(addr, size, dir);
|
|
while (req->flags)
|
|
barrier();
|
|
|
|
raw_local_irq_restore(flags);
|
|
}
|
|
|
|
void fiq_dma_unmap_area(const void *addr, size_t size, int dir)
|
|
{
|
|
unsigned long flags;
|
|
struct fiq_req *req;
|
|
|
|
raw_local_irq_save(flags);
|
|
/* currently, not possible to take cpu0 down, so only check cpu1 */
|
|
if (!cpu_online(1)) {
|
|
raw_local_irq_restore(flags);
|
|
v6_dma_unmap_area(addr, size, dir);
|
|
return;
|
|
}
|
|
|
|
req = this_cpu_ptr(&fiq_data);
|
|
req->unmap.addr = addr;
|
|
req->unmap.size = size;
|
|
req->flags = dir | OXNAS_UNMAP_AREA;
|
|
smp_mb();
|
|
|
|
writel_relaxed(FIQ_GENERATE, req->reg);
|
|
|
|
v6_dma_unmap_area(addr, size, dir);
|
|
while (req->flags)
|
|
barrier();
|
|
|
|
raw_local_irq_restore(flags);
|
|
}
|
|
|
|
void fiq_dma_flush_range(const void *start, const void *end)
|
|
{
|
|
unsigned long flags;
|
|
struct fiq_req *req;
|
|
|
|
raw_local_irq_save(flags);
|
|
/* currently, not possible to take cpu0 down, so only check cpu1 */
|
|
if (!cpu_online(1)) {
|
|
raw_local_irq_restore(flags);
|
|
v6_dma_flush_range(start, end);
|
|
return;
|
|
}
|
|
|
|
req = this_cpu_ptr(&fiq_data);
|
|
|
|
req->flush.start = start;
|
|
req->flush.end = end;
|
|
req->flags = OXNAS_FLUSH_RANGE;
|
|
smp_mb();
|
|
|
|
writel_relaxed(FIQ_GENERATE, req->reg);
|
|
|
|
v6_dma_flush_range(start, end);
|
|
|
|
while (req->flags)
|
|
barrier();
|
|
|
|
raw_local_irq_restore(flags);
|
|
}
|
|
|
|
void fiq_flush_kern_dcache_area(void *addr, size_t size)
|
|
{
|
|
fiq_dma_flush_range(addr, addr + size);
|
|
}
|
|
#else
|
|
|
|
#define ox820_set_fiq_regs(cpu) do {} while (0) /* nothing */
|
|
#define ox820_init_fiq() do {} while (0) /* nothing */
|
|
|
|
#endif /* DMA_CACHE_FIQ_BROADCAST */
|
|
|
|
static DEFINE_SPINLOCK(boot_lock);
|
|
|
|
void ox820_secondary_init(unsigned int cpu)
|
|
{
|
|
/*
|
|
* Setup Secondary Core FIQ regs
|
|
*/
|
|
ox820_set_fiq_regs(1);
|
|
|
|
/*
|
|
* let the primary processor know we're out of the
|
|
* pen, then head off into the C entry point
|
|
*/
|
|
write_pen_release(-1);
|
|
|
|
/*
|
|
* Synchronise with the boot thread.
|
|
*/
|
|
spin_lock(&boot_lock);
|
|
spin_unlock(&boot_lock);
|
|
}
|
|
|
|
int ox820_boot_secondary(unsigned int cpu, struct task_struct *idle)
|
|
{
|
|
unsigned long timeout;
|
|
|
|
/*
|
|
* Set synchronisation state between this boot processor
|
|
* and the secondary one
|
|
*/
|
|
spin_lock(&boot_lock);
|
|
|
|
/*
|
|
* This is really belt and braces; we hold unintended secondary
|
|
* CPUs in the holding pen until we're ready for them. However,
|
|
* since we haven't sent them a soft interrupt, they shouldn't
|
|
* be there.
|
|
*/
|
|
write_pen_release(cpu);
|
|
|
|
writel(1, IOMEM(OXNAS_GICN_BASE_VA(cpu) + GIC_CPU_CTRL));
|
|
|
|
/*
|
|
* Send the secondary CPU a soft interrupt, thereby causing
|
|
* the boot monitor to read the system wide flags register,
|
|
* and branch to the address found there.
|
|
*/
|
|
|
|
arch_send_wakeup_ipi_mask(cpumask_of(cpu));
|
|
timeout = jiffies + (1 * HZ);
|
|
while (time_before(jiffies, timeout)) {
|
|
smp_rmb();
|
|
if (read_pen_release() == -1)
|
|
break;
|
|
|
|
udelay(10);
|
|
}
|
|
|
|
/*
|
|
* now the secondary core is starting up let it run its
|
|
* calibrations, then wait for it to finish
|
|
*/
|
|
spin_unlock(&boot_lock);
|
|
|
|
return read_pen_release() != -1 ? -ENOSYS : 0;
|
|
}
|
|
|
|
void *scu_base_addr(void)
|
|
{
|
|
return IOMEM(OXNAS_SCU_BASE_VA);
|
|
}
|
|
|
|
/*
|
|
* Initialise the CPU possible map early - this describes the CPUs
|
|
* which may be present or become present in the system.
|
|
*/
|
|
static void __init ox820_smp_init_cpus(void)
|
|
{
|
|
void __iomem *scu_base = scu_base_addr();
|
|
unsigned int i, ncores;
|
|
|
|
ncores = scu_base ? scu_get_core_count(scu_base) : 1;
|
|
|
|
/* sanity check */
|
|
if (ncores > nr_cpu_ids) {
|
|
pr_warn("SMP: %u cores greater than maximum (%u), clipping\n",
|
|
ncores, nr_cpu_ids);
|
|
ncores = nr_cpu_ids;
|
|
}
|
|
|
|
for (i = 0; i < ncores; i++)
|
|
set_cpu_possible(i, true);
|
|
}
|
|
|
|
static void __init ox820_smp_prepare_cpus(unsigned int max_cpus)
|
|
{
|
|
|
|
scu_enable(scu_base_addr());
|
|
|
|
/*
|
|
* Write the address of secondary startup into the
|
|
* system-wide flags register. The BootMonitor waits
|
|
* until it receives a soft interrupt, and then the
|
|
* secondary CPU branches to this address.
|
|
*/
|
|
writel(virt_to_phys(ox820_secondary_startup),
|
|
HOLDINGPEN_LOCATION);
|
|
ox820_init_fiq();
|
|
|
|
ox820_set_fiq_regs(0);
|
|
}
|
|
|
|
struct smp_operations ox820_smp_ops __initdata = {
|
|
.smp_init_cpus = ox820_smp_init_cpus,
|
|
.smp_prepare_cpus = ox820_smp_prepare_cpus,
|
|
.smp_secondary_init = ox820_secondary_init,
|
|
.smp_boot_secondary = ox820_boot_secondary,
|
|
#ifdef CONFIG_HOTPLUG_CPU
|
|
.cpu_die = ox820_cpu_die,
|
|
#endif
|
|
};
|