* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*/
-#include <linux/config.h>
+
#include <linux/types.h>
#include <linux/threads.h>
#include <linux/kernel.h>
#include <linux/irq.h>
#include <linux/smp.h>
#include <linux/interrupt.h>
-#include <linux/signal.h>
#include <linux/init.h>
-#include <linux/gfp.h>
#include <linux/radix-tree.h>
#include <linux/cpu.h>
+#include <linux/of.h>
+
#include <asm/firmware.h>
-#include <asm/prom.h>
#include <asm/io.h>
#include <asm/pgtable.h>
#include <asm/smp.h>
#include <asm/rtas.h>
#include <asm/hvcall.h>
#include <asm/machdep.h>
-#include <asm/i8259.h>
#include "xics.h"
+#include "plpar_wrappers.h"
-static unsigned int xics_startup(unsigned int irq);
-static void xics_enable_irq(unsigned int irq);
-static void xics_disable_irq(unsigned int irq);
-static void xics_mask_and_ack_irq(unsigned int irq);
-static void xics_end_irq(unsigned int irq);
-static void xics_set_affinity(unsigned int irq_nr, cpumask_t cpumask);
-
-static struct hw_interrupt_type xics_pic = {
- .typename = " XICS ",
- .startup = xics_startup,
- .enable = xics_enable_irq,
- .disable = xics_disable_irq,
- .ack = xics_mask_and_ack_irq,
- .end = xics_end_irq,
- .set_affinity = xics_set_affinity
-};
-
-/* This is used to map real irq numbers to virtual */
-static struct radix_tree_root irq_map = RADIX_TREE_INIT(GFP_ATOMIC);
+static struct irq_host *xics_host;
#define XICS_IPI 2
#define XICS_IRQ_SPURIOUS 0
/*
* Mark IPIs as higher priority so we can take them inside interrupts that
- * arent marked SA_INTERRUPT
+ * arent marked IRQF_DISABLED
*/
#define IPI_PRIORITY 4
+static unsigned int default_server = 0xFF;
+static unsigned int default_distrib_server = 0;
+static unsigned int interrupt_server_size = 8;
+
+/* RTAS service tokens */
+static int ibm_get_xive;
+static int ibm_set_xive;
+static int ibm_int_on;
+static int ibm_int_off;
+
+
+/* Direct hardware low level accessors */
+
+/* The part of the interrupt presentation layer that we care about */
struct xics_ipl {
union {
u32 word;
static struct xics_ipl __iomem *xics_per_cpu[NR_CPUS];
-static int xics_irq_8259_cascade = 0;
-static int xics_irq_8259_cascade_real = 0;
-static unsigned int default_server = 0xFF;
-static unsigned int default_distrib_server = 0;
-static unsigned int interrupt_server_size = 8;
-
-/*
- * XICS only has a single IPI, so encode the messages per CPU
- */
-struct xics_ipi_struct xics_ipi_message[NR_CPUS] __cacheline_aligned;
-
-/* RTAS service tokens */
-static int ibm_get_xive;
-static int ibm_set_xive;
-static int ibm_int_on;
-static int ibm_int_off;
-
-typedef struct {
- int (*xirr_info_get)(int cpu);
- void (*xirr_info_set)(int cpu, int val);
- void (*cppr_info)(int cpu, u8 val);
- void (*qirr_info)(int cpu, u8 val);
-} xics_ops;
-
-
-/* SMP */
-
-static int pSeries_xirr_info_get(int n_cpu)
+static inline unsigned int direct_xirr_info_get(void)
{
- return in_be32(&xics_per_cpu[n_cpu]->xirr.word);
-}
+ int cpu = smp_processor_id();
-static void pSeries_xirr_info_set(int n_cpu, int value)
-{
- out_be32(&xics_per_cpu[n_cpu]->xirr.word, value);
+ return in_be32(&xics_per_cpu[cpu]->xirr.word);
}
-static void pSeries_cppr_info(int n_cpu, u8 value)
+static inline void direct_xirr_info_set(unsigned int value)
{
- out_8(&xics_per_cpu[n_cpu]->xirr.bytes[0], value);
-}
+ int cpu = smp_processor_id();
-static void pSeries_qirr_info(int n_cpu, u8 value)
-{
- out_8(&xics_per_cpu[n_cpu]->qirr.bytes[0], value);
+ out_be32(&xics_per_cpu[cpu]->xirr.word, value);
}
-static xics_ops pSeries_ops = {
- pSeries_xirr_info_get,
- pSeries_xirr_info_set,
- pSeries_cppr_info,
- pSeries_qirr_info
-};
-
-static xics_ops *ops = &pSeries_ops;
-
-
-/* LPAR */
-
-static inline long plpar_eoi(unsigned long xirr)
+static inline void direct_cppr_info(u8 value)
{
- return plpar_hcall_norets(H_EOI, xirr);
-}
+ int cpu = smp_processor_id();
-static inline long plpar_cppr(unsigned long cppr)
-{
- return plpar_hcall_norets(H_CPPR, cppr);
+ out_8(&xics_per_cpu[cpu]->xirr.bytes[0], value);
}
-static inline long plpar_ipi(unsigned long servernum, unsigned long mfrr)
+static inline void direct_qirr_info(int n_cpu, u8 value)
{
- return plpar_hcall_norets(H_IPI, servernum, mfrr);
+ out_8(&xics_per_cpu[n_cpu]->qirr.bytes[0], value);
}
-static inline long plpar_xirr(unsigned long *xirr_ret)
-{
- unsigned long dummy;
- return plpar_hcall(H_XIRR, 0, 0, 0, 0, xirr_ret, &dummy, &dummy);
-}
-static int pSeriesLP_xirr_info_get(int n_cpu)
+/* LPAR low level accessors */
+
+static inline unsigned int lpar_xirr_info_get(void)
{
unsigned long lpar_rc;
unsigned long return_value;
lpar_rc = plpar_xirr(&return_value);
if (lpar_rc != H_SUCCESS)
panic(" bad return code xirr - rc = %lx \n", lpar_rc);
- return (int)return_value;
+ return (unsigned int)return_value;
}
-static void pSeriesLP_xirr_info_set(int n_cpu, int value)
+static inline void lpar_xirr_info_set(unsigned int value)
{
unsigned long lpar_rc;
- unsigned long val64 = value & 0xffffffff;
- lpar_rc = plpar_eoi(val64);
+ lpar_rc = plpar_eoi(value);
if (lpar_rc != H_SUCCESS)
- panic("bad return code EOI - rc = %ld, value=%lx\n", lpar_rc,
- val64);
+ panic("bad return code EOI - rc = %ld, value=%x\n", lpar_rc,
+ value);
}
-void pSeriesLP_cppr_info(int n_cpu, u8 value)
+static inline void lpar_cppr_info(u8 value)
{
unsigned long lpar_rc;
panic("bad return code cppr - rc = %lx\n", lpar_rc);
}
-static void pSeriesLP_qirr_info(int n_cpu , u8 value)
+static inline void lpar_qirr_info(int n_cpu , u8 value)
{
unsigned long lpar_rc;
panic("bad return code qirr - rc = %lx\n", lpar_rc);
}
-xics_ops pSeriesLP_ops = {
- pSeriesLP_xirr_info_get,
- pSeriesLP_xirr_info_set,
- pSeriesLP_cppr_info,
- pSeriesLP_qirr_info
-};
-static unsigned int xics_startup(unsigned int virq)
-{
- unsigned int irq;
-
- irq = irq_offset_down(virq);
- if (radix_tree_insert(&irq_map, virt_irq_to_real(irq),
- &virt_irq_to_real_map[irq]) == -ENOMEM)
- printk(KERN_CRIT "Out of memory creating real -> virtual"
- " IRQ mapping for irq %u (real 0x%x)\n",
- virq, virt_irq_to_real(irq));
- xics_enable_irq(virq);
- return 0; /* return value is ignored */
-}
-
-static unsigned int real_irq_to_virt(unsigned int real_irq)
-{
- unsigned int *ptr;
-
- ptr = radix_tree_lookup(&irq_map, real_irq);
- if (ptr == NULL)
- return NO_IRQ;
- return ptr - virt_irq_to_real_map;
-}
+/* Interface to generic irq subsystem */
#ifdef CONFIG_SMP
-static int get_irq_server(unsigned int irq)
+static int get_irq_server(unsigned int virq, unsigned int strict_check)
{
- unsigned int server;
+ int server;
/* For the moment only implement delivery to all cpus or one cpu */
- cpumask_t cpumask = irq_affinity[irq];
+ cpumask_t cpumask;
cpumask_t tmp = CPU_MASK_NONE;
+ cpumask_copy(&cpumask, irq_desc[virq].affinity);
if (!distribute_irqs)
return default_server;
- if (cpus_equal(cpumask, CPU_MASK_ALL)) {
- server = default_distrib_server;
- } else {
+ if (!cpus_equal(cpumask, CPU_MASK_ALL)) {
cpus_and(tmp, cpu_online_map, cpumask);
- if (cpus_empty(tmp))
- server = default_distrib_server;
- else
- server = get_hard_smp_processor_id(first_cpu(tmp));
+ server = first_cpu(tmp);
+
+ if (server < NR_CPUS)
+ return get_hard_smp_processor_id(server);
+
+ if (strict_check)
+ return -1;
}
- return server;
+ if (cpus_equal(cpu_online_map, cpu_present_map))
+ return default_distrib_server;
+ return default_server;
}
#else
-static int get_irq_server(unsigned int irq)
+static int get_irq_server(unsigned int virq, unsigned int strict_check)
{
return default_server;
}
#endif
-static void xics_enable_irq(unsigned int virq)
+static void xics_unmask_irq(unsigned int virq)
{
unsigned int irq;
int call_status;
- unsigned int server;
+ int server;
- irq = virt_irq_to_real(irq_offset_down(virq));
- if (irq == XICS_IPI)
+ pr_debug("xics: unmask virq %d\n", virq);
+
+ irq = (unsigned int)irq_map[virq].hwirq;
+ pr_debug(" -> map to hwirq 0x%x\n", irq);
+ if (irq == XICS_IPI || irq == XICS_IRQ_SPURIOUS)
return;
- server = get_irq_server(virq);
+ server = get_irq_server(virq, 0);
+
call_status = rtas_call(ibm_set_xive, 3, 1, NULL, irq, server,
DEFAULT_PRIORITY);
if (call_status != 0) {
- printk(KERN_ERR "xics_enable_irq: irq=%u: ibm_set_xive "
- "returned %d\n", irq, call_status);
- printk("set_xive %x, server %x\n", ibm_set_xive, server);
+ printk(KERN_ERR
+ "%s: ibm_set_xive irq %u server %x returned %d\n",
+ __func__, irq, server, call_status);
return;
}
/* Now unmask the interrupt (often a no-op) */
call_status = rtas_call(ibm_int_on, 1, 1, NULL, irq);
if (call_status != 0) {
- printk(KERN_ERR "xics_enable_irq: irq=%u: ibm_int_on "
- "returned %d\n", irq, call_status);
+ printk(KERN_ERR "%s: ibm_int_on irq=%u returned %d\n",
+ __func__, irq, call_status);
return;
}
}
-static void xics_disable_real_irq(unsigned int irq)
+static unsigned int xics_startup(unsigned int virq)
+{
+ /* unmask it */
+ xics_unmask_irq(virq);
+ return 0;
+}
+
+static void xics_mask_real_irq(unsigned int irq)
{
int call_status;
- unsigned int server;
if (irq == XICS_IPI)
return;
call_status = rtas_call(ibm_int_off, 1, 1, NULL, irq);
if (call_status != 0) {
- printk(KERN_ERR "xics_disable_real_irq: irq=%u: "
- "ibm_int_off returned %d\n", irq, call_status);
+ printk(KERN_ERR "%s: ibm_int_off irq=%u returned %d\n",
+ __func__, irq, call_status);
return;
}
- server = get_irq_server(irq);
/* Have to set XIVE to 0xff to be able to remove a slot */
- call_status = rtas_call(ibm_set_xive, 3, 1, NULL, irq, server, 0xff);
+ call_status = rtas_call(ibm_set_xive, 3, 1, NULL, irq,
+ default_server, 0xff);
if (call_status != 0) {
- printk(KERN_ERR "xics_disable_irq: irq=%u: ibm_set_xive(0xff)"
- " returned %d\n", irq, call_status);
+ printk(KERN_ERR "%s: ibm_set_xive(0xff) irq=%u returned %d\n",
+ __func__, irq, call_status);
return;
}
}
-static void xics_disable_irq(unsigned int virq)
+static void xics_mask_irq(unsigned int virq)
{
unsigned int irq;
- irq = virt_irq_to_real(irq_offset_down(virq));
- xics_disable_real_irq(irq);
+ pr_debug("xics: mask virq %d\n", virq);
+
+ irq = (unsigned int)irq_map[virq].hwirq;
+ if (irq == XICS_IPI || irq == XICS_IRQ_SPURIOUS)
+ return;
+ xics_mask_real_irq(irq);
}
-static void xics_end_irq(unsigned int irq)
+static void xics_mask_unknown_vec(unsigned int vec)
{
- int cpu = smp_processor_id();
+ printk(KERN_ERR "Interrupt %u (real) is invalid, disabling it.\n", vec);
+ xics_mask_real_irq(vec);
+}
+
+static inline unsigned int xics_xirr_vector(unsigned int xirr)
+{
+ /*
+ * The top byte is the old cppr, to be restored on EOI.
+ * The remaining 24 bits are the vector.
+ */
+ return xirr & 0x00ffffff;
+}
+
+static unsigned int xics_get_irq_direct(void)
+{
+ unsigned int xirr = direct_xirr_info_get();
+ unsigned int vec = xics_xirr_vector(xirr);
+ unsigned int irq;
+
+ if (vec == XICS_IRQ_SPURIOUS)
+ return NO_IRQ;
+
+ irq = irq_radix_revmap_lookup(xics_host, vec);
+ if (likely(irq != NO_IRQ))
+ return irq;
+
+ /* We don't have a linux mapping, so have rtas mask it. */
+ xics_mask_unknown_vec(vec);
+
+ /* We might learn about it later, so EOI it */
+ direct_xirr_info_set(xirr);
+ return NO_IRQ;
+}
+
+static unsigned int xics_get_irq_lpar(void)
+{
+ unsigned int xirr = lpar_xirr_info_get();
+ unsigned int vec = xics_xirr_vector(xirr);
+ unsigned int irq;
+
+ if (vec == XICS_IRQ_SPURIOUS)
+ return NO_IRQ;
+
+ irq = irq_radix_revmap_lookup(xics_host, vec);
+ if (likely(irq != NO_IRQ))
+ return irq;
+
+ /* We don't have a linux mapping, so have RTAS mask it. */
+ xics_mask_unknown_vec(vec);
+
+ /* We might learn about it later, so EOI it */
+ lpar_xirr_info_set(xirr);
+ return NO_IRQ;
+}
+
+static void xics_eoi_direct(unsigned int virq)
+{
+ unsigned int irq = (unsigned int)irq_map[virq].hwirq;
iosync();
- ops->xirr_info_set(cpu, ((0xff << 24) |
- (virt_irq_to_real(irq_offset_down(irq)))));
+ direct_xirr_info_set((0xff << 24) | irq);
+}
+
+static void xics_eoi_lpar(unsigned int virq)
+{
+ unsigned int irq = (unsigned int)irq_map[virq].hwirq;
+ iosync();
+ lpar_xirr_info_set((0xff << 24) | irq);
}
-static void xics_mask_and_ack_irq(unsigned int irq)
+static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask)
{
- int cpu = smp_processor_id();
+ unsigned int irq;
+ int status;
+ int xics_status[2];
+ int irq_server;
+
+ irq = (unsigned int)irq_map[virq].hwirq;
+ if (irq == XICS_IPI || irq == XICS_IRQ_SPURIOUS)
+ return;
+
+ status = rtas_call(ibm_get_xive, 1, 3, xics_status, irq);
+
+ if (status) {
+ printk(KERN_ERR "%s: ibm,get-xive irq=%u returns %d\n",
+ __func__, irq, status);
+ return;
+ }
+
+ /*
+ * For the moment only implement delivery to all cpus or one cpu.
+ * Get current irq_server for the given irq
+ */
+ irq_server = get_irq_server(virq, 1);
+ if (irq_server == -1) {
+ char cpulist[128];
+ cpumask_scnprintf(cpulist, sizeof(cpulist), cpumask);
+ printk(KERN_WARNING
+ "%s: No online cpus in the mask %s for irq %d\n",
+ __func__, cpulist, virq);
+ return;
+ }
- if (irq < irq_offset_value()) {
- i8259_pic.ack(irq);
- iosync();
- ops->xirr_info_set(cpu, ((0xff<<24) |
- xics_irq_8259_cascade_real));
- iosync();
+ status = rtas_call(ibm_set_xive, 3, 1, NULL,
+ irq, irq_server, xics_status[1]);
+
+ if (status) {
+ printk(KERN_ERR "%s: ibm,set-xive irq=%u returns %d\n",
+ __func__, irq, status);
+ return;
}
}
-int xics_get_irq(struct pt_regs *regs)
+static struct irq_chip xics_pic_direct = {
+ .typename = " XICS ",
+ .startup = xics_startup,
+ .mask = xics_mask_irq,
+ .unmask = xics_unmask_irq,
+ .eoi = xics_eoi_direct,
+ .set_affinity = xics_set_affinity
+};
+
+static struct irq_chip xics_pic_lpar = {
+ .typename = " XICS ",
+ .startup = xics_startup,
+ .mask = xics_mask_irq,
+ .unmask = xics_unmask_irq,
+ .eoi = xics_eoi_lpar,
+ .set_affinity = xics_set_affinity
+};
+
+
+/* Interface to arch irq controller subsystem layer */
+
+/* Points to the irq_chip we're actually using */
+static struct irq_chip *xics_irq_chip;
+
+static int xics_host_match(struct irq_host *h, struct device_node *node)
{
- unsigned int cpu = smp_processor_id();
- unsigned int vec;
- int irq;
+ /* IBM machines have interrupt parents of various funky types for things
+ * like vdevices, events, etc... The trick we use here is to match
+ * everything here except the legacy 8259 which is compatible "chrp,iic"
+ */
+ return !of_device_is_compatible(node, "chrp,iic");
+}
- vec = ops->xirr_info_get(cpu);
- /* (vec >> 24) == old priority */
- vec &= 0x00ffffff;
+static int xics_host_map(struct irq_host *h, unsigned int virq,
+ irq_hw_number_t hw)
+{
+ pr_debug("xics: map virq %d, hwirq 0x%lx\n", virq, hw);
- /* for sanity, this had better be < NR_IRQS - 16 */
- if (vec == xics_irq_8259_cascade_real) {
- irq = i8259_irq(regs);
- xics_end_irq(irq_offset_up(xics_irq_8259_cascade));
- } else if (vec == XICS_IRQ_SPURIOUS) {
- irq = -1;
- } else {
- irq = real_irq_to_virt(vec);
- if (irq == NO_IRQ)
- irq = real_irq_to_virt_slowpath(vec);
- if (irq == NO_IRQ) {
- printk(KERN_ERR "Interrupt %u (real) is invalid,"
- " disabling it.\n", vec);
- xics_disable_real_irq(vec);
- } else
- irq = irq_offset_up(irq);
- }
- return irq;
+ /* Insert the interrupt mapping into the radix tree for fast lookup */
+ irq_radix_revmap_insert(xics_host, virq, hw);
+
+ get_irq_desc(virq)->status |= IRQ_LEVEL;
+ set_irq_chip_and_handler(virq, xics_irq_chip, handle_fasteoi_irq);
+ return 0;
}
+static int xics_host_xlate(struct irq_host *h, struct device_node *ct,
+ u32 *intspec, unsigned int intsize,
+ irq_hw_number_t *out_hwirq, unsigned int *out_flags)
+
+{
+ /* Current xics implementation translates everything
+ * to level. It is not technically right for MSIs but this
+ * is irrelevant at this point. We might get smarter in the future
+ */
+ *out_hwirq = intspec[0];
+ *out_flags = IRQ_TYPE_LEVEL_LOW;
+
+ return 0;
+}
+
+static struct irq_host_ops xics_host_ops = {
+ .match = xics_host_match,
+ .map = xics_host_map,
+ .xlate = xics_host_xlate,
+};
+
+static void __init xics_init_host(void)
+{
+ if (firmware_has_feature(FW_FEATURE_LPAR))
+ xics_irq_chip = &xics_pic_lpar;
+ else
+ xics_irq_chip = &xics_pic_direct;
+
+ xics_host = irq_alloc_host(NULL, IRQ_HOST_MAP_TREE, 0, &xics_host_ops,
+ XICS_IRQ_SPURIOUS);
+ BUG_ON(xics_host == NULL);
+ irq_set_default_host(xics_host);
+}
+
+
+/* Inter-processor interrupt support */
+
#ifdef CONFIG_SMP
+/*
+ * XICS only has a single IPI, so encode the messages per CPU
+ */
+struct xics_ipi_struct {
+ unsigned long value;
+ } ____cacheline_aligned;
-static irqreturn_t xics_ipi_action(int irq, void *dev_id, struct pt_regs *regs)
+static struct xics_ipi_struct xics_ipi_message[NR_CPUS] __cacheline_aligned;
+
+static inline void smp_xics_do_message(int cpu, int msg)
{
- int cpu = smp_processor_id();
+ set_bit(msg, &xics_ipi_message[cpu].value);
+ mb();
+ if (firmware_has_feature(FW_FEATURE_LPAR))
+ lpar_qirr_info(cpu, IPI_PRIORITY);
+ else
+ direct_qirr_info(cpu, IPI_PRIORITY);
+}
- ops->qirr_info(cpu, 0xff);
+void smp_xics_message_pass(int target, int msg)
+{
+ unsigned int i;
+ if (target < NR_CPUS) {
+ smp_xics_do_message(target, msg);
+ } else {
+ for_each_online_cpu(i) {
+ if (target == MSG_ALL_BUT_SELF
+ && i == smp_processor_id())
+ continue;
+ smp_xics_do_message(i, msg);
+ }
+ }
+}
+
+static irqreturn_t xics_ipi_dispatch(int cpu)
+{
WARN_ON(cpu_is_offline(cpu));
+ mb(); /* order mmio clearing qirr */
while (xics_ipi_message[cpu].value) {
if (test_and_clear_bit(PPC_MSG_CALL_FUNCTION,
&xics_ipi_message[cpu].value)) {
- mb();
- smp_message_recv(PPC_MSG_CALL_FUNCTION, regs);
+ smp_message_recv(PPC_MSG_CALL_FUNCTION);
}
if (test_and_clear_bit(PPC_MSG_RESCHEDULE,
&xics_ipi_message[cpu].value)) {
- mb();
- smp_message_recv(PPC_MSG_RESCHEDULE, regs);
+ smp_message_recv(PPC_MSG_RESCHEDULE);
}
-#if 0
- if (test_and_clear_bit(PPC_MSG_MIGRATE_TASK,
+ if (test_and_clear_bit(PPC_MSG_CALL_FUNC_SINGLE,
&xics_ipi_message[cpu].value)) {
- mb();
- smp_message_recv(PPC_MSG_MIGRATE_TASK, regs);
+ smp_message_recv(PPC_MSG_CALL_FUNC_SINGLE);
}
-#endif
#if defined(CONFIG_DEBUGGER) || defined(CONFIG_KEXEC)
if (test_and_clear_bit(PPC_MSG_DEBUGGER_BREAK,
&xics_ipi_message[cpu].value)) {
- mb();
- smp_message_recv(PPC_MSG_DEBUGGER_BREAK, regs);
+ smp_message_recv(PPC_MSG_DEBUGGER_BREAK);
}
#endif
}
return IRQ_HANDLED;
}
-void xics_cause_IPI(int cpu)
+static irqreturn_t xics_ipi_action_direct(int irq, void *dev_id)
{
- ops->qirr_info(cpu, IPI_PRIORITY);
+ int cpu = smp_processor_id();
+
+ direct_qirr_info(cpu, 0xff);
+
+ return xics_ipi_dispatch(cpu);
}
-#endif /* CONFIG_SMP */
-void xics_setup_cpu(void)
+static irqreturn_t xics_ipi_action_lpar(int irq, void *dev_id)
{
int cpu = smp_processor_id();
- ops->cppr_info(cpu, 0xff);
- iosync();
+ lpar_qirr_info(cpu, 0xff);
+
+ return xics_ipi_dispatch(cpu);
+}
+
+static void xics_request_ipi(void)
+{
+ unsigned int ipi;
+ int rc;
+
+ ipi = irq_create_mapping(xics_host, XICS_IPI);
+ BUG_ON(ipi == NO_IRQ);
/*
- * Put the calling processor into the GIQ. This is really only
- * necessary from a secondary thread as the OF start-cpu interface
- * performs this function for us on primary threads.
- *
- * XXX: undo of teardown on kexec needs this too, as may hotplug
+ * IPIs are marked IRQF_DISABLED as they must run with irqs
+ * disabled
*/
- rtas_set_indicator(GLOBAL_INTERRUPT_QUEUE,
- (1UL << interrupt_server_size) - 1 - default_distrib_server, 1);
+ set_irq_handler(ipi, handle_percpu_irq);
+ if (firmware_has_feature(FW_FEATURE_LPAR))
+ rc = request_irq(ipi, xics_ipi_action_lpar,
+ IRQF_DISABLED|IRQF_PERCPU, "IPI", NULL);
+ else
+ rc = request_irq(ipi, xics_ipi_action_direct,
+ IRQF_DISABLED|IRQF_PERCPU, "IPI", NULL);
+ BUG_ON(rc);
}
-void xics_init_IRQ(void)
+int __init smp_xics_probe(void)
{
- int i;
- unsigned long intr_size = 0;
+ xics_request_ipi();
+
+ return cpus_weight(cpu_possible_map);
+}
+
+#endif /* CONFIG_SMP */
+
+
+/* Initialization */
+
+static void xics_update_irq_servers(void)
+{
+ int i, j;
struct device_node *np;
- uint *ireg, ilen, indx = 0;
- unsigned long intr_base = 0;
- struct xics_interrupt_node {
- unsigned long addr;
- unsigned long size;
- } intnodes[NR_CPUS];
+ u32 ilen;
+ const u32 *ireg;
+ u32 hcpuid;
- ppc64_boot_msg(0x20, "XICS Init");
+ /* Find the server numbers for the boot cpu. */
+ np = of_get_cpu_node(boot_cpuid, NULL);
+ BUG_ON(!np);
- ibm_get_xive = rtas_token("ibm,get-xive");
- ibm_set_xive = rtas_token("ibm,set-xive");
- ibm_int_on = rtas_token("ibm,int-on");
- ibm_int_off = rtas_token("ibm,int-off");
+ ireg = of_get_property(np, "ibm,ppc-interrupt-gserver#s", &ilen);
+ if (!ireg) {
+ of_node_put(np);
+ return;
+ }
- np = of_find_node_by_type(NULL, "PowerPC-External-Interrupt-Presentation");
- if (!np)
- panic("xics_init_IRQ: can't find interrupt presentation");
+ i = ilen / sizeof(int);
+ hcpuid = get_hard_smp_processor_id(boot_cpuid);
+
+ /* Global interrupt distribution server is specified in the last
+ * entry of "ibm,ppc-interrupt-gserver#s" property. Get the last
+ * entry fom this property for current boot cpu id and use it as
+ * default distribution server
+ */
+ for (j = 0; j < i; j += 2) {
+ if (ireg[j] == hcpuid) {
+ default_server = hcpuid;
+ default_distrib_server = ireg[j+1];
+ }
+ }
+
+ of_node_put(np);
+}
+
+static void __init xics_map_one_cpu(int hw_id, unsigned long addr,
+ unsigned long size)
+{
+ int i;
+
+ /* This may look gross but it's good enough for now, we don't quite
+ * have a hard -> linux processor id matching.
+ */
+ for_each_possible_cpu(i) {
+ if (!cpu_present(i))
+ continue;
+ if (hw_id == get_hard_smp_processor_id(i)) {
+ xics_per_cpu[i] = ioremap(addr, size);
+ return;
+ }
+ }
+}
+
+static void __init xics_init_one_node(struct device_node *np,
+ unsigned int *indx)
+{
+ unsigned int ilen;
+ const u32 *ireg;
-nextnode:
- ireg = (uint *)get_property(np, "ibm,interrupt-server-ranges", NULL);
+ /* This code does the theorically broken assumption that the interrupt
+ * server numbers are the same as the hard CPU numbers.
+ * This happens to be the case so far but we are playing with fire...
+ * should be fixed one of these days. -BenH.
+ */
+ ireg = of_get_property(np, "ibm,interrupt-server-ranges", NULL);
+
+ /* Do that ever happen ? we'll know soon enough... but even good'old
+ * f80 does have that property ..
+ */
+ WARN_ON(ireg == NULL);
if (ireg) {
/*
* set node starting index for this node
*/
- indx = *ireg;
+ *indx = *ireg;
}
-
- ireg = (uint *)get_property(np, "reg", &ilen);
+ ireg = of_get_property(np, "reg", &ilen);
if (!ireg)
panic("xics_init_IRQ: can't find interrupt reg property");
- while (ilen) {
- intnodes[indx].addr = (unsigned long)*ireg++ << 32;
- ilen -= sizeof(uint);
- intnodes[indx].addr |= *ireg++;
- ilen -= sizeof(uint);
- intnodes[indx].size = (unsigned long)*ireg++ << 32;
- ilen -= sizeof(uint);
- intnodes[indx].size |= *ireg++;
- ilen -= sizeof(uint);
- indx++;
- if (indx >= NR_CPUS) break;
+ while (ilen >= (4 * sizeof(u32))) {
+ unsigned long addr, size;
+
+ /* XXX Use proper OF parsing code here !!! */
+ addr = (unsigned long)*ireg++ << 32;
+ ilen -= sizeof(u32);
+ addr |= *ireg++;
+ ilen -= sizeof(u32);
+ size = (unsigned long)*ireg++ << 32;
+ ilen -= sizeof(u32);
+ size |= *ireg++;
+ ilen -= sizeof(u32);
+ xics_map_one_cpu(*indx, addr, size);
+ (*indx)++;
}
+}
- np = of_find_node_by_type(np, "PowerPC-External-Interrupt-Presentation");
- if ((indx < NR_CPUS) && np) goto nextnode;
+void __init xics_init_IRQ(void)
+{
+ struct device_node *np;
+ u32 indx = 0;
+ int found = 0;
+ const u32 *isize;
- /* Find the server numbers for the boot cpu. */
- for (np = of_find_node_by_type(NULL, "cpu");
- np;
- np = of_find_node_by_type(np, "cpu")) {
- ireg = (uint *)get_property(np, "reg", &ilen);
- if (ireg && ireg[0] == get_hard_smp_processor_id(boot_cpuid)) {
- ireg = (uint *)get_property(np, "ibm,ppc-interrupt-gserver#s",
- &ilen);
- i = ilen / sizeof(int);
- if (ireg && i > 0) {
- default_server = ireg[0];
- default_distrib_server = ireg[i-1]; /* take last element */
- }
- ireg = (uint *)get_property(np,
- "ibm,interrupt-server#-size", NULL);
- if (ireg)
- interrupt_server_size = *ireg;
- break;
- }
- }
- of_node_put(np);
+ ppc64_boot_msg(0x20, "XICS Init");
- intr_base = intnodes[0].addr;
- intr_size = intnodes[0].size;
+ ibm_get_xive = rtas_token("ibm,get-xive");
+ ibm_set_xive = rtas_token("ibm,set-xive");
+ ibm_int_on = rtas_token("ibm,int-on");
+ ibm_int_off = rtas_token("ibm,int-off");
- np = of_find_node_by_type(NULL, "interrupt-controller");
- if (!np) {
- printk(KERN_DEBUG "xics: no ISA interrupt controller\n");
- xics_irq_8259_cascade_real = -1;
- xics_irq_8259_cascade = -1;
- } else {
- ireg = (uint *) get_property(np, "interrupts", NULL);
- if (!ireg)
- panic("xics_init_IRQ: can't find ISA interrupts property");
-
- xics_irq_8259_cascade_real = *ireg;
- xics_irq_8259_cascade
- = virt_irq_create_mapping(xics_irq_8259_cascade_real);
- i8259_init(0, 0);
- of_node_put(np);
+ for_each_node_by_type(np, "PowerPC-External-Interrupt-Presentation") {
+ found = 1;
+ if (firmware_has_feature(FW_FEATURE_LPAR)) {
+ of_node_put(np);
+ break;
+ }
+ xics_init_one_node(np, &indx);
}
+ if (found == 0)
+ return;
- if (firmware_has_feature(FW_FEATURE_LPAR))
- ops = &pSeriesLP_ops;
- else {
-#ifdef CONFIG_SMP
- for_each_possible_cpu(i) {
- int hard_id;
+ /* get the bit size of server numbers */
+ found = 0;
- /* FIXME: Do this dynamically! --RR */
- if (!cpu_present(i))
- continue;
+ for_each_compatible_node(np, NULL, "ibm,ppc-xics") {
+ isize = of_get_property(np, "ibm,interrupt-server#-size", NULL);
+
+ if (!isize)
+ continue;
- hard_id = get_hard_smp_processor_id(i);
- xics_per_cpu[i] = ioremap(intnodes[hard_id].addr,
- intnodes[hard_id].size);
+ if (!found) {
+ interrupt_server_size = *isize;
+ found = 1;
+ } else if (*isize != interrupt_server_size) {
+ printk(KERN_WARNING "XICS: "
+ "mismatched ibm,interrupt-server#-size\n");
+ interrupt_server_size = max(*isize,
+ interrupt_server_size);
}
-#else
- xics_per_cpu[0] = ioremap(intr_base, intr_size);
-#endif /* CONFIG_SMP */
}
- for (i = irq_offset_value(); i < NR_IRQS; ++i)
- get_irq_desc(i)->handler = &xics_pic;
+ xics_update_irq_servers();
+ xics_init_host();
+
+ if (firmware_has_feature(FW_FEATURE_LPAR))
+ ppc_md.get_irq = xics_get_irq_lpar;
+ else
+ ppc_md.get_irq = xics_get_irq_direct;
xics_setup_cpu();
ppc64_boot_msg(0x21, "XICS Done");
}
-/*
- * We cant do this in init_IRQ because we need the memory subsystem up for
- * request_irq()
- */
-static int __init xics_setup_i8259(void)
-{
- if (ppc64_interrupt_controller == IC_PPC_XIC &&
- xics_irq_8259_cascade != -1) {
- if (request_irq(irq_offset_up(xics_irq_8259_cascade),
- no_action, 0, "8259 cascade", NULL))
- printk(KERN_ERR "xics_setup_i8259: couldn't get 8259 "
- "cascade\n");
- }
- return 0;
-}
-arch_initcall(xics_setup_i8259);
+/* Cpu startup, shutdown, and hotplug */
-#ifdef CONFIG_SMP
-void xics_request_IPIs(void)
+static void xics_set_cpu_priority(unsigned char cppr)
{
- virt_irq_to_real_map[XICS_IPI] = XICS_IPI;
-
- /* IPIs are marked SA_INTERRUPT as they must run with irqs disabled */
- request_irq(irq_offset_up(XICS_IPI), xics_ipi_action, SA_INTERRUPT,
- "IPI", NULL);
- get_irq_desc(irq_offset_up(XICS_IPI))->status |= IRQ_PER_CPU;
+ if (firmware_has_feature(FW_FEATURE_LPAR))
+ lpar_cppr_info(cppr);
+ else
+ direct_cppr_info(cppr);
+ iosync();
}
-#endif
-static void xics_set_affinity(unsigned int virq, cpumask_t cpumask)
+/* Have the calling processor join or leave the specified global queue */
+static void xics_set_cpu_giq(unsigned int gserver, unsigned int join)
{
- unsigned int irq;
+ int index;
int status;
- int xics_status[2];
- unsigned long newmask;
- cpumask_t tmp = CPU_MASK_NONE;
- irq = virt_irq_to_real(irq_offset_down(virq));
- if (irq == XICS_IPI || irq == NO_IRQ)
+ if (!rtas_indicator_present(GLOBAL_INTERRUPT_QUEUE, NULL))
return;
- status = rtas_call(ibm_get_xive, 1, 3, xics_status, irq);
+ index = (1UL << interrupt_server_size) - 1 - gserver;
- if (status) {
- printk(KERN_ERR "xics_set_affinity: irq=%u ibm,get-xive "
- "returns %d\n", irq, status);
- return;
- }
+ status = rtas_set_indicator_fast(GLOBAL_INTERRUPT_QUEUE, index, join);
- /* For the moment only implement delivery to all cpus or one cpu */
- if (cpus_equal(cpumask, CPU_MASK_ALL)) {
- newmask = default_distrib_server;
- } else {
- cpus_and(tmp, cpu_online_map, cpumask);
- if (cpus_empty(tmp))
- return;
- newmask = get_hard_smp_processor_id(first_cpu(tmp));
- }
+ WARN(status < 0, "set-indicator(%d, %d, %u) returned %d\n",
+ GLOBAL_INTERRUPT_QUEUE, index, join, status);
+}
- status = rtas_call(ibm_set_xive, 3, 1, NULL,
- irq, newmask, xics_status[1]);
+void xics_setup_cpu(void)
+{
+ xics_set_cpu_priority(0xff);
- if (status) {
- printk(KERN_ERR "xics_set_affinity: irq=%u ibm,set-xive "
- "returns %d\n", irq, status);
- return;
- }
+ xics_set_cpu_giq(default_distrib_server, 1);
}
-void xics_teardown_cpu(int secondary)
+void xics_teardown_cpu(void)
{
int cpu = smp_processor_id();
- ops->cppr_info(cpu, 0x00);
- iosync();
+ xics_set_cpu_priority(0);
+
+ /* Clear any pending IPI request */
+ if (firmware_has_feature(FW_FEATURE_LPAR))
+ lpar_qirr_info(cpu, 0xff);
+ else
+ direct_qirr_info(cpu, 0xff);
+}
- /* Clear IPI */
- ops->qirr_info(cpu, 0xff);
+void xics_kexec_teardown_cpu(int secondary)
+{
+ xics_teardown_cpu();
/*
- * we need to EOI the IPI if we got here from kexec down IPI
+ * we take the ipi irq but and never return so we
+ * need to EOI the IPI, but want to leave our priority 0
*
- * probably need to check all the other interrupts too
+ * should we check all the other interrupts too?
* should we be flagging idle loop instead?
* or creating some task to be scheduled?
*/
- ops->xirr_info_set(cpu, XICS_IPI);
+
+ if (firmware_has_feature(FW_FEATURE_LPAR))
+ lpar_xirr_info_set((0x00 << 24) | XICS_IPI);
+ else
+ direct_xirr_info_set((0x00 << 24) | XICS_IPI);
/*
* Some machines need to have at least one cpu in the GIQ,
* so leave the master cpu in the group.
*/
if (secondary)
- rtas_set_indicator(GLOBAL_INTERRUPT_QUEUE,
- (1UL << interrupt_server_size) - 1 -
- default_distrib_server, 0);
+ xics_set_cpu_giq(default_distrib_server, 0);
}
#ifdef CONFIG_HOTPLUG_CPU
/* Interrupts are disabled. */
void xics_migrate_irqs_away(void)
{
- int status;
- unsigned int irq, virq, cpu = smp_processor_id();
+ int cpu = smp_processor_id(), hw_cpu = hard_smp_processor_id();
+ unsigned int irq, virq;
+
+ /* If we used to be the default server, move to the new "boot_cpuid" */
+ if (hw_cpu == default_server)
+ xics_update_irq_servers();
/* Reject any interrupt that was queued to us... */
- ops->cppr_info(cpu, 0);
- iosync();
+ xics_set_cpu_priority(0);
- /* remove ourselves from the global interrupt queue */
- status = rtas_set_indicator(GLOBAL_INTERRUPT_QUEUE,
- (1UL << interrupt_server_size) - 1 - default_distrib_server, 0);
- WARN_ON(status < 0);
+ /* Remove ourselves from the global interrupt queue */
+ xics_set_cpu_giq(default_distrib_server, 0);
/* Allow IPIs again... */
- ops->cppr_info(cpu, DEFAULT_PRIORITY);
- iosync();
+ xics_set_cpu_priority(DEFAULT_PRIORITY);
for_each_irq(virq) {
- irq_desc_t *desc;
+ struct irq_desc *desc;
int xics_status[2];
+ int status;
unsigned long flags;
/* We cant set affinity on ISA interrupts */
- if (virq < irq_offset_value())
+ if (virq < NUM_ISA_INTERRUPTS)
continue;
-
- desc = get_irq_desc(virq);
- irq = virt_irq_to_real(irq_offset_down(virq));
-
+ if (irq_map[virq].host != xics_host)
+ continue;
+ irq = (unsigned int)irq_map[virq].hwirq;
/* We need to get IPIs still. */
- if (irq == XICS_IPI || irq == NO_IRQ)
+ if (irq == XICS_IPI || irq == XICS_IRQ_SPURIOUS)
continue;
+ desc = get_irq_desc(virq);
/* We only need to migrate enabled IRQS */
- if (desc == NULL || desc->handler == NULL
+ if (desc == NULL || desc->chip == NULL
|| desc->action == NULL
- || desc->handler->set_affinity == NULL)
+ || desc->chip->set_affinity == NULL)
continue;
spin_lock_irqsave(&desc->lock, flags);
status = rtas_call(ibm_get_xive, 1, 3, xics_status, irq);
if (status) {
- printk(KERN_ERR "migrate_irqs_away: irq=%u "
- "ibm,get-xive returns %d\n",
- virq, status);
+ printk(KERN_ERR "%s: ibm,get-xive irq=%u returns %d\n",
+ __func__, irq, status);
goto unlock;
}
* The irq has to be migrated only in the single cpu
* case.
*/
- if (xics_status[0] != get_hard_smp_processor_id(cpu))
+ if (xics_status[0] != hw_cpu)
goto unlock;
printk(KERN_WARNING "IRQ %u affinity broken off cpu %u\n",
virq, cpu);
/* Reset affinity to all cpus */
- desc->handler->set_affinity(virq, CPU_MASK_ALL);
- irq_affinity[virq] = CPU_MASK_ALL;
+ cpumask_setall(irq_desc[virq].affinity);
+ desc->chip->set_affinity(virq, cpu_all_mask);
unlock:
spin_unlock_irqrestore(&desc->lock, flags);
}