Patches contributed by Eötvös Lorand University
commit 88271e9e43e65537c6390af3718a60c45c787e8f
Author: Ingo Molnar <mingo@elte.hu>
Date: Thu Oct 5 18:47:22 2006 +0200
[PATCH] i386: fix rwsem build bug on CONFIG_M386=y
CONFIG_M386 turns on spinlock-based generic rwsems - which surprises the
semaphore.S rwsem stubs. Tested both with and without CONFIG_M386.
Reported-by: Klaus Knopper <knopper@knopper.net>
Triaged-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andi Kleen <ak@suse.de>
diff --git a/arch/i386/lib/semaphore.S b/arch/i386/lib/semaphore.S
index ef6ad9e1a609..c01eb39c0b43 100644
--- a/arch/i386/lib/semaphore.S
+++ b/arch/i386/lib/semaphore.S
@@ -152,6 +152,8 @@ ENTRY(__read_lock_failed)
#endif
+#ifdef CONFIG_RWSEM_XCHGADD_ALGORITHM
+
/* Fix up special calling conventions */
ENTRY(call_rwsem_down_read_failed)
CFI_STARTPROC
@@ -214,3 +216,4 @@ ENTRY(call_rwsem_downgrade_wake)
CFI_ENDPROC
END(call_rwsem_downgrade_wake)
+#endif
commit f5b9ed7acdcfea4bf73a70dececa7483787503ed
Author: Ingo Molnar <mingo@elte.hu>
Date: Wed Oct 4 02:16:26 2006 -0700
[PATCH] genirq: convert the i386 architecture to irq-chips
This patch converts all the i386 PIC controllers (except VisWS and Voyager,
which I could not test - but which should still work as old-style IRQ layers)
to the new and simpler irq-chip interrupt handling layer.
[akpm@osdl.org: build fix]
[mingo@elte.hu: enable fasteoi handler for i386 level-triggered IO-APIC irqs]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/arch/i386/kernel/i8259.c b/arch/i386/kernel/i8259.c
index ea5f4e7958d8..d07ed31f11e3 100644
--- a/arch/i386/kernel/i8259.c
+++ b/arch/i386/kernel/i8259.c
@@ -34,35 +34,15 @@
* moves to arch independent land
*/
-DEFINE_SPINLOCK(i8259A_lock);
-
-static void end_8259A_irq (unsigned int irq)
-{
- if (!(irq_desc[irq].status & (IRQ_DISABLED|IRQ_INPROGRESS)) &&
- irq_desc[irq].action)
- enable_8259A_irq(irq);
-}
-
-#define shutdown_8259A_irq disable_8259A_irq
-
static int i8259A_auto_eoi;
-
+DEFINE_SPINLOCK(i8259A_lock);
static void mask_and_ack_8259A(unsigned int);
-unsigned int startup_8259A_irq(unsigned int irq)
-{
- enable_8259A_irq(irq);
- return 0; /* never anything pending */
-}
-
-static struct hw_interrupt_type i8259A_irq_type = {
- .typename = "XT-PIC",
- .startup = startup_8259A_irq,
- .shutdown = shutdown_8259A_irq,
- .enable = enable_8259A_irq,
- .disable = disable_8259A_irq,
- .ack = mask_and_ack_8259A,
- .end = end_8259A_irq,
+static struct irq_chip i8259A_chip = {
+ .name = "XT-PIC",
+ .mask = disable_8259A_irq,
+ .unmask = enable_8259A_irq,
+ .mask_ack = mask_and_ack_8259A,
};
/*
@@ -133,7 +113,7 @@ void make_8259A_irq(unsigned int irq)
{
disable_irq_nosync(irq);
io_apic_irqs &= ~(1<<irq);
- irq_desc[irq].chip = &i8259A_irq_type;
+ set_irq_chip_and_handler(irq, &i8259A_chip, handle_level_irq);
enable_irq(irq);
}
@@ -327,12 +307,12 @@ void init_8259A(int auto_eoi)
outb_p(SLAVE_ICW4_DEFAULT, PIC_SLAVE_IMR); /* (slave's support for AEOI in flat mode is to be investigated) */
if (auto_eoi)
/*
- * in AEOI mode we just have to mask the interrupt
+ * In AEOI mode we just have to mask the interrupt
* when acking.
*/
- i8259A_irq_type.ack = disable_8259A_irq;
+ i8259A_chip.mask_ack = disable_8259A_irq;
else
- i8259A_irq_type.ack = mask_and_ack_8259A;
+ i8259A_chip.mask_ack = mask_and_ack_8259A;
udelay(100); /* wait for 8259A to initialize */
@@ -389,12 +369,13 @@ void __init init_ISA_irqs (void)
/*
* 16 old-style INTA-cycle interrupts:
*/
- irq_desc[i].chip = &i8259A_irq_type;
+ set_irq_chip_and_handler(i, &i8259A_chip,
+ handle_level_irq);
} else {
/*
* 'high' PCI IRQs filled in on demand
*/
- irq_desc[i].chip = &no_irq_type;
+ irq_desc[i].chip = &no_irq_chip;
}
}
}
diff --git a/arch/i386/kernel/io_apic.c b/arch/i386/kernel/io_apic.c
index fd0df75cfbda..b3b01894ca69 100644
--- a/arch/i386/kernel/io_apic.c
+++ b/arch/i386/kernel/io_apic.c
@@ -1219,8 +1219,7 @@ int assign_irq_vector(int irq)
return vector;
}
-static struct hw_interrupt_type ioapic_level_type;
-static struct hw_interrupt_type ioapic_edge_type;
+static struct irq_chip ioapic_chip;
#define IOAPIC_AUTO -1
#define IOAPIC_EDGE 0
@@ -1234,9 +1233,11 @@ static void ioapic_register_intr(int irq, int vector, unsigned long trigger)
if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) ||
trigger == IOAPIC_LEVEL)
- irq_desc[idx].chip = &ioapic_level_type;
+ set_irq_chip_and_handler(idx, &ioapic_chip,
+ handle_fasteoi_irq);
else
- irq_desc[idx].chip = &ioapic_edge_type;
+ set_irq_chip_and_handler(idx, &ioapic_chip,
+ handle_edge_irq);
set_intr_gate(vector, interrupt[idx]);
}
@@ -1346,7 +1347,8 @@ static void __init setup_ExtINT_IRQ0_pin(unsigned int apic, unsigned int pin, in
* The timer IRQ doesn't have to know that behind the
* scene we have a 8259A-master in AEOI mode ...
*/
- irq_desc[0].chip = &ioapic_edge_type;
+ irq_desc[0].chip = &ioapic_chip;
+ set_irq_handler(0, handle_edge_irq);
/*
* Add it to the IO-APIC irq-routing table:
@@ -1918,6 +1920,8 @@ static int __init timer_irq_works(void)
*/
/*
+ * Startup quirk:
+ *
* Starting up a edge-triggered IO-APIC interrupt is
* nasty - we need to make sure that we get the edge.
* If it is already asserted for some reason, we need
@@ -1925,8 +1929,10 @@ static int __init timer_irq_works(void)
*
* This is not complete - we should be able to fake
* an edge even if it isn't on the 8259A...
+ *
+ * (We do this for level-triggered IRQs too - it cannot hurt.)
*/
-static unsigned int startup_edge_ioapic_irq(unsigned int irq)
+static unsigned int startup_ioapic_irq(unsigned int irq)
{
int was_pending = 0;
unsigned long flags;
@@ -1943,42 +1949,13 @@ static unsigned int startup_edge_ioapic_irq(unsigned int irq)
return was_pending;
}
-/*
- * Once we have recorded IRQ_PENDING already, we can mask the
- * interrupt for real. This prevents IRQ storms from unhandled
- * devices.
- */
-static void ack_edge_ioapic_irq(unsigned int irq)
+static void ack_ioapic_irq(unsigned int irq)
{
move_irq(irq);
- if ((irq_desc[irq].status & (IRQ_PENDING | IRQ_DISABLED))
- == (IRQ_PENDING | IRQ_DISABLED))
- mask_IO_APIC_irq(irq);
ack_APIC_irq();
}
-/*
- * Level triggered interrupts can just be masked,
- * and shutting down and starting up the interrupt
- * is the same as enabling and disabling them -- except
- * with a startup need to return a "was pending" value.
- *
- * Level triggered interrupts are special because we
- * do not touch any IO-APIC register while handling
- * them. We ack the APIC in the end-IRQ handler, not
- * in the start-IRQ-handler. Protection against reentrance
- * from the same interrupt is still provided, both by the
- * generic IRQ layer and by the fact that an unacked local
- * APIC does not accept IRQs.
- */
-static unsigned int startup_level_ioapic_irq (unsigned int irq)
-{
- unmask_IO_APIC_irq(irq);
-
- return 0; /* don't check for pending */
-}
-
-static void end_level_ioapic_irq (unsigned int irq)
+static void ack_ioapic_quirk_irq(unsigned int irq)
{
unsigned long v;
int i;
@@ -2018,35 +1995,27 @@ static void end_level_ioapic_irq (unsigned int irq)
}
}
-#ifdef CONFIG_PCI_MSI
-static unsigned int startup_edge_ioapic_vector(unsigned int vector)
+static unsigned int startup_ioapic_vector(unsigned int vector)
{
int irq = vector_to_irq(vector);
- return startup_edge_ioapic_irq(irq);
+ return startup_ioapic_irq(irq);
}
-static void ack_edge_ioapic_vector(unsigned int vector)
+static void ack_ioapic_vector(unsigned int vector)
{
int irq = vector_to_irq(vector);
move_native_irq(vector);
- ack_edge_ioapic_irq(irq);
+ ack_ioapic_irq(irq);
}
-static unsigned int startup_level_ioapic_vector (unsigned int vector)
-{
- int irq = vector_to_irq(vector);
-
- return startup_level_ioapic_irq (irq);
-}
-
-static void end_level_ioapic_vector (unsigned int vector)
+static void ack_ioapic_quirk_vector(unsigned int vector)
{
int irq = vector_to_irq(vector);
move_native_irq(vector);
- end_level_ioapic_irq(irq);
+ ack_ioapic_quirk_irq(irq);
}
static void mask_IO_APIC_vector (unsigned int vector)
@@ -2063,7 +2032,12 @@ static void unmask_IO_APIC_vector (unsigned int vector)
unmask_IO_APIC_irq(irq);
}
-#ifdef CONFIG_SMP
+/*
+ * Oh just glorious. If CONFIG_PCI_MSI we've done
+ * #define set_ioapic_affinity set_ioapic_affinity_vector
+ */
+#if defined (CONFIG_SMP) && defined(CONFIG_X86_IO_APIC) && \
+ defined(CONFIG_PCI_MSI)
static void set_ioapic_affinity_vector (unsigned int vector,
cpumask_t cpu_mask)
{
@@ -2073,50 +2047,29 @@ static void set_ioapic_affinity_vector (unsigned int vector,
set_ioapic_affinity_irq(irq, cpu_mask);
}
#endif
-#endif
-static int ioapic_retrigger(unsigned int irq)
+static int ioapic_retrigger_vector(unsigned int vector)
{
+ int irq = vector_to_irq(vector);
+
send_IPI_self(IO_APIC_VECTOR(irq));
return 1;
}
-/*
- * Level and edge triggered IO-APIC interrupts need different handling,
- * so we use two separate IRQ descriptors. Edge triggered IRQs can be
- * handled with the level-triggered descriptor, but that one has slightly
- * more overhead. Level-triggered interrupts cannot be handled with the
- * edge-triggered handler, without risking IRQ storms and other ugly
- * races.
- */
-static struct hw_interrupt_type ioapic_edge_type __read_mostly = {
- .typename = "IO-APIC-edge",
- .startup = startup_edge_ioapic,
- .shutdown = shutdown_edge_ioapic,
- .enable = enable_edge_ioapic,
- .disable = disable_edge_ioapic,
- .ack = ack_edge_ioapic,
- .end = end_edge_ioapic,
+static struct irq_chip ioapic_chip __read_mostly = {
+ .name = "IO-APIC",
+ .startup = startup_ioapic_vector,
+ .mask = mask_IO_APIC_vector,
+ .unmask = unmask_IO_APIC_vector,
+ .ack = ack_ioapic_vector,
+ .eoi = ack_ioapic_quirk_vector,
#ifdef CONFIG_SMP
.set_affinity = set_ioapic_affinity,
#endif
- .retrigger = ioapic_retrigger,
+ .retrigger = ioapic_retrigger_vector,
};
-static struct hw_interrupt_type ioapic_level_type __read_mostly = {
- .typename = "IO-APIC-level",
- .startup = startup_level_ioapic,
- .shutdown = shutdown_level_ioapic,
- .enable = enable_level_ioapic,
- .disable = disable_level_ioapic,
- .ack = mask_and_ack_level_ioapic,
- .end = end_level_ioapic,
-#ifdef CONFIG_SMP
- .set_affinity = set_ioapic_affinity,
-#endif
- .retrigger = ioapic_retrigger,
-};
static inline void init_IO_APIC_traps(void)
{
@@ -2150,20 +2103,21 @@ static inline void init_IO_APIC_traps(void)
make_8259A_irq(irq);
else
/* Strange. Oh, well.. */
- irq_desc[irq].chip = &no_irq_type;
+ irq_desc[irq].chip = &no_irq_chip;
}
}
}
-static void enable_lapic_irq (unsigned int irq)
-{
- unsigned long v;
+/*
+ * The local APIC irq-chip implementation:
+ */
- v = apic_read(APIC_LVT0);
- apic_write_around(APIC_LVT0, v & ~APIC_LVT_MASKED);
+static void ack_apic(unsigned int irq)
+{
+ ack_APIC_irq();
}
-static void disable_lapic_irq (unsigned int irq)
+static void mask_lapic_irq (unsigned int irq)
{
unsigned long v;
@@ -2171,21 +2125,19 @@ static void disable_lapic_irq (unsigned int irq)
apic_write_around(APIC_LVT0, v | APIC_LVT_MASKED);
}
-static void ack_lapic_irq (unsigned int irq)
+static void unmask_lapic_irq (unsigned int irq)
{
- ack_APIC_irq();
-}
+ unsigned long v;
-static void end_lapic_irq (unsigned int i) { /* nothing */ }
+ v = apic_read(APIC_LVT0);
+ apic_write_around(APIC_LVT0, v & ~APIC_LVT_MASKED);
+}
-static struct hw_interrupt_type lapic_irq_type __read_mostly = {
- .typename = "local-APIC-edge",
- .startup = NULL, /* startup_irq() not used for IRQ0 */
- .shutdown = NULL, /* shutdown_irq() not used for IRQ0 */
- .enable = enable_lapic_irq,
- .disable = disable_lapic_irq,
- .ack = ack_lapic_irq,
- .end = end_lapic_irq
+static struct irq_chip lapic_chip __read_mostly = {
+ .name = "local-APIC-edge",
+ .mask = mask_lapic_irq,
+ .unmask = unmask_lapic_irq,
+ .eoi = ack_apic,
};
static void setup_nmi (void)
@@ -2356,7 +2308,7 @@ static inline void check_timer(void)
printk(KERN_INFO "...trying to set up timer as Virtual Wire IRQ...");
disable_8259A_irq(0);
- irq_desc[0].chip = &lapic_irq_type;
+ set_irq_chip_and_handler(0, &lapic_chip, handle_fasteoi_irq);
apic_write_around(APIC_LVT0, APIC_DM_FIXED | vector); /* Fixed mode */
enable_8259A_irq(0);
diff --git a/arch/i386/kernel/irq.c b/arch/i386/kernel/irq.c
index 5fe547cd8f9f..3dd2e180151b 100644
--- a/arch/i386/kernel/irq.c
+++ b/arch/i386/kernel/irq.c
@@ -55,6 +55,7 @@ fastcall unsigned int do_IRQ(struct pt_regs *regs)
{
/* high bit used in ret_from_ code */
int irq = ~regs->orig_eax;
+ struct irq_desc *desc = irq_desc + irq;
#ifdef CONFIG_4KSTACKS
union irq_ctx *curctx, *irqctx;
u32 *isp;
@@ -94,7 +95,7 @@ fastcall unsigned int do_IRQ(struct pt_regs *regs)
* current stack (which is the irq stack already after all)
*/
if (curctx != irqctx) {
- int arg1, arg2, ebx;
+ int arg1, arg2, arg3, ebx;
/* build the stack frame on the IRQ stack */
isp = (u32*) ((char*)irqctx + sizeof(*irqctx));
@@ -110,16 +111,17 @@ fastcall unsigned int do_IRQ(struct pt_regs *regs)
(curctx->tinfo.preempt_count & SOFTIRQ_MASK);
asm volatile(
- " xchgl %%ebx,%%esp \n"
- " call __do_IRQ \n"
+ " xchgl %%ebx,%%esp \n"
+ " call *%%edi \n"
" movl %%ebx,%%esp \n"
- : "=a" (arg1), "=d" (arg2), "=b" (ebx)
- : "0" (irq), "1" (regs), "2" (isp)
- : "memory", "cc", "ecx"
+ : "=a" (arg1), "=d" (arg2), "=c" (arg3), "=b" (ebx)
+ : "0" (irq), "1" (desc), "2" (regs), "3" (isp),
+ "D" (desc->handle_irq)
+ : "memory", "cc"
);
} else
#endif
- __do_IRQ(irq, regs);
+ desc->handle_irq(irq, desc, regs);
irq_exit();
@@ -253,7 +255,8 @@ int show_interrupts(struct seq_file *p, void *v)
for_each_online_cpu(j)
seq_printf(p, "%10u ", kstat_cpu(j).irqs[i]);
#endif
- seq_printf(p, " %14s", irq_desc[i].chip->typename);
+ seq_printf(p, " %8s", irq_desc[i].chip->name);
+ seq_printf(p, "-%s", handle_irq_name(irq_desc[i].handle_irq));
seq_printf(p, " %s", action->name);
for (action=action->next; action; action = action->next)
diff --git a/include/asm-i386/hw_irq.h b/include/asm-i386/hw_irq.h
index 87e5a351d881..8e8fbfaa252a 100644
--- a/include/asm-i386/hw_irq.h
+++ b/include/asm-i386/hw_irq.h
@@ -17,8 +17,6 @@
#include <asm/irq.h>
#include <asm/sections.h>
-struct hw_interrupt_type;
-
#define NMI_VECTOR 0x02
/*
commit f29bd1ba68c8c6a0f50bd678bbd5a26674018f7c
Author: Ingo Molnar <mingo@elte.hu>
Date: Wed Oct 4 02:16:25 2006 -0700
[PATCH] genirq: convert the x86_64 architecture to irq-chips
This patch converts all the x86_64 PIC controllers layers to the new and
simpler irq-chip interrupt handling layer.
[mingo@elte.hu: The patch also enables the fasteoi handler for x86_64]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Roland Dreier <rolandd@cisco.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/arch/x86_64/kernel/i8259.c b/arch/x86_64/kernel/i8259.c
index 2dd51f364ea2..ae1101cd5252 100644
--- a/arch/x86_64/kernel/i8259.c
+++ b/arch/x86_64/kernel/i8259.c
@@ -121,42 +121,15 @@ void (*interrupt[NR_IRQS])(void) = {
* moves to arch independent land
*/
-DEFINE_SPINLOCK(i8259A_lock);
-
static int i8259A_auto_eoi;
-
-static void end_8259A_irq (unsigned int irq)
-{
- if (irq > 256) {
- char var;
- printk("return %p stack %p ti %p\n", __builtin_return_address(0), &var, task_thread_info(current));
-
- BUG();
- }
-
- if (!(irq_desc[irq].status & (IRQ_DISABLED|IRQ_INPROGRESS)) &&
- irq_desc[irq].action)
- enable_8259A_irq(irq);
-}
-
-#define shutdown_8259A_irq disable_8259A_irq
-
+DEFINE_SPINLOCK(i8259A_lock);
static void mask_and_ack_8259A(unsigned int);
-static unsigned int startup_8259A_irq(unsigned int irq)
-{
- enable_8259A_irq(irq);
- return 0; /* never anything pending */
-}
-
-static struct hw_interrupt_type i8259A_irq_type = {
- .typename = "XT-PIC",
- .startup = startup_8259A_irq,
- .shutdown = shutdown_8259A_irq,
- .enable = enable_8259A_irq,
- .disable = disable_8259A_irq,
- .ack = mask_and_ack_8259A,
- .end = end_8259A_irq,
+static struct irq_chip i8259A_chip = {
+ .name = "XT-PIC",
+ .mask = disable_8259A_irq,
+ .unmask = enable_8259A_irq,
+ .mask_ack = mask_and_ack_8259A,
};
/*
@@ -231,7 +204,7 @@ void make_8259A_irq(unsigned int irq)
{
disable_irq_nosync(irq);
io_apic_irqs &= ~(1<<irq);
- irq_desc[irq].chip = &i8259A_irq_type;
+ set_irq_chip_and_handler(irq, &i8259A_chip, handle_level_irq);
enable_irq(irq);
}
@@ -367,9 +340,9 @@ void init_8259A(int auto_eoi)
* in AEOI mode we just have to mask the interrupt
* when acking.
*/
- i8259A_irq_type.ack = disable_8259A_irq;
+ i8259A_chip.mask_ack = disable_8259A_irq;
else
- i8259A_irq_type.ack = mask_and_ack_8259A;
+ i8259A_chip.mask_ack = mask_and_ack_8259A;
udelay(100); /* wait for 8259A to initialize */
@@ -464,12 +437,13 @@ void __init init_ISA_irqs (void)
/*
* 16 old-style INTA-cycle interrupts:
*/
- irq_desc[i].chip = &i8259A_irq_type;
+ set_irq_chip_and_handler(i, &i8259A_chip,
+ handle_level_irq);
} else {
/*
* 'high' PCI IRQs filled in on demand
*/
- irq_desc[i].chip = &no_irq_type;
+ irq_desc[i].chip = &no_irq_chip;
}
}
}
diff --git a/arch/x86_64/kernel/io_apic.c b/arch/x86_64/kernel/io_apic.c
index 0491019d4c8d..e1d1b4217cb2 100644
--- a/arch/x86_64/kernel/io_apic.c
+++ b/arch/x86_64/kernel/io_apic.c
@@ -633,8 +633,8 @@ int assign_irq_vector(int irq)
}
extern void (*interrupt[NR_IRQS])(void);
-static struct hw_interrupt_type ioapic_level_type;
-static struct hw_interrupt_type ioapic_edge_type;
+
+static struct irq_chip ioapic_chip;
#define IOAPIC_AUTO -1
#define IOAPIC_EDGE 0
@@ -648,9 +648,11 @@ static void ioapic_register_intr(int irq, int vector, unsigned long trigger)
if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) ||
trigger == IOAPIC_LEVEL)
- irq_desc[idx].chip = &ioapic_level_type;
+ set_irq_chip_and_handler(idx, &ioapic_chip,
+ handle_fasteoi_irq);
else
- irq_desc[idx].chip = &ioapic_edge_type;
+ set_irq_chip_and_handler(idx, &ioapic_chip,
+ handle_edge_irq);
set_intr_gate(vector, interrupt[idx]);
}
@@ -752,7 +754,7 @@ static void __init setup_ExtINT_IRQ0_pin(unsigned int apic, unsigned int pin, in
* The timer IRQ doesn't have to know that behind the
* scene we have a 8259A-master in AEOI mode ...
*/
- irq_desc[0].chip = &ioapic_edge_type;
+ set_irq_chip_and_handler(0, &ioapic_chip, handle_edge_irq);
/*
* Add it to the IO-APIC irq-routing table:
@@ -1185,7 +1187,7 @@ static int __init timer_irq_works(void)
* an edge even if it isn't on the 8259A...
*/
-static unsigned int startup_edge_ioapic_irq(unsigned int irq)
+static unsigned int startup_ioapic_irq(unsigned int irq)
{
int was_pending = 0;
unsigned long flags;
@@ -1202,86 +1204,21 @@ static unsigned int startup_edge_ioapic_irq(unsigned int irq)
return was_pending;
}
-/*
- * Once we have recorded IRQ_PENDING already, we can mask the
- * interrupt for real. This prevents IRQ storms from unhandled
- * devices.
- */
-static void ack_edge_ioapic_irq(unsigned int irq)
-{
- move_irq(irq);
- if ((irq_desc[irq].status & (IRQ_PENDING | IRQ_DISABLED))
- == (IRQ_PENDING | IRQ_DISABLED))
- mask_IO_APIC_irq(irq);
- ack_APIC_irq();
-}
-
-/*
- * Level triggered interrupts can just be masked,
- * and shutting down and starting up the interrupt
- * is the same as enabling and disabling them -- except
- * with a startup need to return a "was pending" value.
- *
- * Level triggered interrupts are special because we
- * do not touch any IO-APIC register while handling
- * them. We ack the APIC in the end-IRQ handler, not
- * in the start-IRQ-handler. Protection against reentrance
- * from the same interrupt is still provided, both by the
- * generic IRQ layer and by the fact that an unacked local
- * APIC does not accept IRQs.
- */
-static unsigned int startup_level_ioapic_irq (unsigned int irq)
-{
- unmask_IO_APIC_irq(irq);
-
- return 0; /* don't check for pending */
-}
-
-static void end_level_ioapic_irq (unsigned int irq)
-{
- move_irq(irq);
- ack_APIC_irq();
-}
-
-#ifdef CONFIG_PCI_MSI
-static unsigned int startup_edge_ioapic_vector(unsigned int vector)
+static unsigned int startup_ioapic_vector(unsigned int vector)
{
int irq = vector_to_irq(vector);
- return startup_edge_ioapic_irq(irq);
+ return startup_ioapic_irq(irq);
}
-static void ack_edge_ioapic_vector(unsigned int vector)
-{
- int irq = vector_to_irq(vector);
-
- move_native_irq(vector);
- ack_edge_ioapic_irq(irq);
-}
-
-static unsigned int startup_level_ioapic_vector (unsigned int vector)
-{
- int irq = vector_to_irq(vector);
-
- return startup_level_ioapic_irq (irq);
-}
-
-static void end_level_ioapic_vector (unsigned int vector)
-{
- int irq = vector_to_irq(vector);
-
- move_native_irq(vector);
- end_level_ioapic_irq(irq);
-}
-
-static void mask_IO_APIC_vector (unsigned int vector)
+static void mask_ioapic_vector (unsigned int vector)
{
int irq = vector_to_irq(vector);
mask_IO_APIC_irq(irq);
}
-static void unmask_IO_APIC_vector (unsigned int vector)
+static void unmask_ioapic_vector (unsigned int vector)
{
int irq = vector_to_irq(vector);
@@ -1298,10 +1235,11 @@ static void set_ioapic_affinity_vector (unsigned int vector,
set_ioapic_affinity_irq(irq, cpu_mask);
}
#endif // CONFIG_SMP
-#endif // CONFIG_PCI_MSI
-static int ioapic_retrigger(unsigned int irq)
+static int ioapic_retrigger_vector(unsigned int vector)
{
+ int irq = vector_to_irq(vector);
+
send_IPI_self(IO_APIC_VECTOR(irq));
return 1;
@@ -1316,32 +1254,22 @@ static int ioapic_retrigger(unsigned int irq)
* races.
*/
-static struct hw_interrupt_type ioapic_edge_type __read_mostly = {
- .typename = "IO-APIC-edge",
- .startup = startup_edge_ioapic,
- .shutdown = shutdown_edge_ioapic,
- .enable = enable_edge_ioapic,
- .disable = disable_edge_ioapic,
- .ack = ack_edge_ioapic,
- .end = end_edge_ioapic,
-#ifdef CONFIG_SMP
- .set_affinity = set_ioapic_affinity,
-#endif
- .retrigger = ioapic_retrigger,
-};
+static void ack_apic(unsigned int vector)
+{
+ ack_APIC_irq();
+}
-static struct hw_interrupt_type ioapic_level_type __read_mostly = {
- .typename = "IO-APIC-level",
- .startup = startup_level_ioapic,
- .shutdown = shutdown_level_ioapic,
- .enable = enable_level_ioapic,
- .disable = disable_level_ioapic,
- .ack = mask_and_ack_level_ioapic,
- .end = end_level_ioapic,
+static struct irq_chip ioapic_chip __read_mostly = {
+ .name = "IO-APIC",
+ .startup = startup_ioapic_vector,
+ .mask = mask_ioapic_vector,
+ .unmask = unmask_ioapic_vector,
+ .ack = ack_apic,
+ .eoi = ack_apic,
#ifdef CONFIG_SMP
- .set_affinity = set_ioapic_affinity,
+ .set_affinity = set_ioapic_affinity_vector,
#endif
- .retrigger = ioapic_retrigger,
+ .retrigger = ioapic_retrigger_vector,
};
static inline void init_IO_APIC_traps(void)
@@ -1376,7 +1304,7 @@ static inline void init_IO_APIC_traps(void)
make_8259A_irq(irq);
else
/* Strange. Oh, well.. */
- irq_desc[irq].chip = &no_irq_type;
+ irq_desc[irq].chip = &no_irq_chip;
}
}
}
diff --git a/arch/x86_64/kernel/irq.c b/arch/x86_64/kernel/irq.c
index b3677e6ccc6e..609f97153b88 100644
--- a/arch/x86_64/kernel/irq.c
+++ b/arch/x86_64/kernel/irq.c
@@ -74,7 +74,8 @@ int show_interrupts(struct seq_file *p, void *v)
for_each_online_cpu(j)
seq_printf(p, "%10u ", kstat_cpu(j).irqs[i]);
#endif
- seq_printf(p, " %14s", irq_desc[i].chip->typename);
+ seq_printf(p, " %8s", irq_desc[i].chip->name);
+ seq_printf(p, "-%s", handle_irq_name(irq_desc[i].handle_irq));
seq_printf(p, " %s", action->name);
for (action=action->next; action; action = action->next)
@@ -117,7 +118,7 @@ asmlinkage unsigned int do_IRQ(struct pt_regs *regs)
#ifdef CONFIG_DEBUG_STACKOVERFLOW
stack_overflow_check(regs);
#endif
- __do_IRQ(irq, regs);
+ generic_handle_irq(irq, regs);
irq_exit();
return 1;
diff --git a/include/asm-x86_64/hw_irq.h b/include/asm-x86_64/hw_irq.h
index 48a4a5364e85..f5da94af9c41 100644
--- a/include/asm-x86_64/hw_irq.h
+++ b/include/asm-x86_64/hw_irq.h
@@ -19,8 +19,6 @@
#include <asm/irq.h>
#include <linux/profile.h>
#include <linux/smp.h>
-
-struct hw_interrupt_type;
#endif
#define NMI_VECTOR 0x02
commit 03cbc358aab3faf34bfeaa02206fa660127e9b3f
Author: Ingo Molnar <mingo@elte.hu>
Date: Fri Sep 29 02:01:46 2006 -0700
[PATCH] lockdep core: improve the lock-chain-hash
With CONFIG_DEBUG_LOCK_ALLOC turned off i was getting sporadic failures in
the locking self-test:
------------>
| Locking API testsuite:
----------------------------------------------------------------------------
| spin |wlock |rlock |mutex | wsem | rsem |
--------------------------------------------------------------------------
A-A deadlock: ok | ok | ok | ok | ok | ok |
A-B-B-A deadlock: ok | ok | ok | ok | ok | ok |
A-B-B-C-C-A deadlock: ok | ok | ok | ok | ok | ok |
A-B-C-A-B-C deadlock: ok | ok | ok | ok | ok | ok |
A-B-B-C-C-D-D-A deadlock: ok |FAILED| ok | ok | ok | ok |
A-B-C-D-B-D-D-A deadlock: ok | ok | ok | ok | ok | ok |
A-B-C-D-B-C-D-A deadlock: ok | ok | ok | ok | ok |FAILED|
after much debugging it turned out to be caused by accidental chain-hash
key collisions. The current hash is:
#define iterate_chain_key(key1, key2) \
(((key1) << MAX_LOCKDEP_KEYS_BITS/2) ^ \
((key1) >> (64-MAX_LOCKDEP_KEYS_BITS/2)) ^ \
(key2))
where MAX_LOCKDEP_KEYS_BITS is 11. This hash is pretty good as it will
shift by 5 bits in every iteration, where every new ID 'mixed' into the
hash would have up to 11 bits. But because there was a 6 bits overlap
between subsequent IDs and their high bits tended to be similar, there was
a chance for accidental chain-hash collision for a low number of locks
held.
the solution is to shift by 11 bits:
#define iterate_chain_key(key1, key2) \
(((key1) << MAX_LOCKDEP_KEYS_BITS) ^ \
((key1) >> (64-MAX_LOCKDEP_KEYS_BITS)) ^ \
(key2))
This keeps the hash perfect up to 5 locks held, but even above that the
hash is still good because 11 bits is a relative prime to the total 64
bits, so a complete match will only occur after 64 held locks (which doesnt
happen in Linux). Even after 5 locks held, entropy of the 5 IDs mixed into
the hash is already good enough so that overlap doesnt generate a colliding
hash ID.
with this change the false positives went away.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/kernel/lockdep.c b/kernel/lockdep.c
index df1c3594de31..e596525669ed 100644
--- a/kernel/lockdep.c
+++ b/kernel/lockdep.c
@@ -122,8 +122,8 @@ static struct list_head chainhash_table[CHAINHASH_SIZE];
* unique.
*/
#define iterate_chain_key(key1, key2) \
- (((key1) << MAX_LOCKDEP_KEYS_BITS/2) ^ \
- ((key1) >> (64-MAX_LOCKDEP_KEYS_BITS/2)) ^ \
+ (((key1) << MAX_LOCKDEP_KEYS_BITS) ^ \
+ ((key1) >> (64-MAX_LOCKDEP_KEYS_BITS)) ^ \
(key2))
void lockdep_off(void)
commit 86998aa6534e839ec003ed2ef7067d6fe8696ccc
Author: Ingo Molnar <mingo@elte.hu>
Date: Tue Sep 19 11:14:34 2006 +0200
[PATCH] genirq core: fix handle_level_irq()
while porting the -rt tree to 2.6.18-rc7 i noticed the following
screaming-IRQ scenario on an SMP system:
2274 0Dn.:1 0.001ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
2274 0Dn.:1 0.010ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
2274 0Dn.:1 0.020ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
2274 0Dn.:1 0.029ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
2274 0Dn.:1 0.039ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
2274 0Dn.:1 0.048ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
2274 0Dn.:1 0.058ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
2274 0Dn.:1 0.068ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
2274 0Dn.:1 0.077ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
2274 0Dn.:1 0.087ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
2274 0Dn.:1 0.097ms: do_IRQ+0xc/0x103 <= (ret_from_intr+0x0/0xf)
as it turns out, the bug is caused by handle_level_irq(), which if it
races with another CPU already handling this IRQ, it _unmasks_ the IRQ
line on the way out. This is not how 2.6.17 works, and we introduced
this bug in one of the early genirq cleanups right before it went into
-mm. (the bug was not in the genirq patchset for a long time, and we
didnt notice the bug due to the lack of -rt rebase to the new genirq
code. -rt, and hardirq-preemption in particular opens up such races much
wider than anything else.)
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c
index 9336f2e89e40..ac1f850d4937 100644
--- a/kernel/irq/chip.c
+++ b/kernel/irq/chip.c
@@ -252,7 +252,7 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc, struct pt_regs *regs)
mask_ack_irq(desc, irq);
if (unlikely(desc->status & IRQ_INPROGRESS))
- goto out;
+ goto out_unlock;
desc->status &= ~(IRQ_REPLAY | IRQ_WAITING);
kstat_cpu(cpu).irqs[irq]++;
@@ -263,7 +263,7 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc, struct pt_regs *regs)
action = desc->action;
if (unlikely(!action || (desc->status & IRQ_DISABLED))) {
desc->status |= IRQ_PENDING;
- goto out;
+ goto out_unlock;
}
desc->status |= IRQ_INPROGRESS;
@@ -276,9 +276,9 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc, struct pt_regs *regs)
spin_lock(&desc->lock);
desc->status &= ~IRQ_INPROGRESS;
-out:
if (!(desc->status & IRQ_DISABLED) && desc->chip->unmask)
desc->chip->unmask(irq);
+out_unlock:
spin_unlock(&desc->lock);
}
commit 9bb25bf36f0d7b06368432e2324dbbc2e98b5e60
Author: Ingo Molnar <mingo@elte.hu>
Date: Tue Sep 12 20:35:50 2006 -0700
[PATCH] lockdep: double the number of stack-trace entries
Miles Lane reported the "BUG: MAX_STACK_TRACE_ENTRIES too low!" message,
which means that during normal use his system produced enough lockdep
events so that the 128-thousand entries stack-trace array got exhausted.
Double the size of the array.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Miles Lane <miles.lane@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/kernel/lockdep_internals.h b/kernel/lockdep_internals.h
index 0d355f24fe04..eab043c83bb2 100644
--- a/kernel/lockdep_internals.h
+++ b/kernel/lockdep_internals.h
@@ -27,7 +27,7 @@
* Stack-trace: tightly packed array of stack backtrace
* addresses. Protected by the hash_lock.
*/
-#define MAX_STACK_TRACE_ENTRIES 131072UL
+#define MAX_STACK_TRACE_ENTRIES 262144UL
extern struct list_head all_lock_classes;
commit 068c4579fe5c21e84c7cb2ba89db80899e25104e
Author: Ingo Molnar <mingo@elte.hu>
Date: Wed Sep 6 00:03:44 2006 -0700
[PATCH] lockdep: do not touch console state when tainting the kernel
Remove an unintended console_verbose() side-effect from add_taint().
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/kernel/panic.c b/kernel/panic.c
index 9b8dcfd1ca93..8010b9b17aca 100644
--- a/kernel/panic.c
+++ b/kernel/panic.c
@@ -173,7 +173,7 @@ const char *print_tainted(void)
void add_taint(unsigned flag)
{
- debug_locks_off(); /* can't trust the integrity of the kernel anymore */
+ debug_locks = 0; /* can't trust the integrity of the kernel anymore */
tainted |= flag;
}
EXPORT_SYMBOL(add_taint);
commit 4df46240a1312161e3c794f6ace50ef7eb5ff3d7
Author: Ingo Molnar <mingo@elte.hu>
Date: Sun Aug 27 01:23:56 2006 -0700
[PATCH] lockdep: annotate reiserfs
reiserfs seems to have another locking level layer for the i_mutex due to the
xattrs-are-a-directory thing.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c
index 39fedaa88a0c..d935fb9394e3 100644
--- a/fs/reiserfs/xattr.c
+++ b/fs/reiserfs/xattr.c
@@ -424,7 +424,7 @@ int xattr_readdir(struct file *file, filldir_t filler, void *buf)
int res = -ENOTDIR;
if (!file->f_op || !file->f_op->readdir)
goto out;
- mutex_lock(&inode->i_mutex);
+ mutex_lock_nested(&inode->i_mutex, I_MUTEX_XATTR);
// down(&inode->i_zombie);
res = -ENOENT;
if (!IS_DEADDIR(inode)) {
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 25610205c90d..555bc195c420 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -570,13 +570,14 @@ struct inode {
* 3: quota file
*
* The locking order between these classes is
- * parent -> child -> normal -> quota
+ * parent -> child -> normal -> xattr -> quota
*/
enum inode_i_mutex_lock_class
{
I_MUTEX_NORMAL,
I_MUTEX_PARENT,
I_MUTEX_CHILD,
+ I_MUTEX_XATTR,
I_MUTEX_QUOTA
};
commit 36e8e5783297fbb83bdebe7e245ef659958f23cb
Author: Ingo Molnar <mingo@elte.hu>
Date: Sun Aug 27 01:23:56 2006 -0700
[PATCH] lockdep: annotate idescsi_pc_intr()
idescsi_pc_intr() uses local_irq_enable() in IRQ context: annotate it.
(this has no effect on kernels with lockdep disabled. On kernels with lockdep
enabled this means that we wont actually disable interrupts, and the warning
message will go away as well.)
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/drivers/scsi/ide-scsi.c b/drivers/scsi/ide-scsi.c
index f7b5d7372d26..94d1de55607f 100644
--- a/drivers/scsi/ide-scsi.c
+++ b/drivers/scsi/ide-scsi.c
@@ -517,7 +517,7 @@ static ide_startstop_t idescsi_pc_intr (ide_drive_t *drive)
/* No more interrupts */
if (test_bit(IDESCSI_LOG_CMD, &scsi->log))
printk (KERN_INFO "Packet command completed, %d bytes transferred\n", pc->actually_transferred);
- local_irq_enable();
+ local_irq_enable_in_hardirq();
if (status.b.check)
rq->errors++;
idescsi_end_request (drive, 1, 0);
commit 640c41c77a96dbbfb74d40ae86ab75b759afb911
Author: Ingo Molnar <mingo@elte.hu>
Date: Tue Aug 15 00:06:56 2006 -0700
[IPV6] lockdep: annotate __icmpv6_socket
Split off __icmpv6_socket's sk->sk_dst_lock class, because it gets
used from softirqs, which is safe for __icmpv6_sockets (because they
never get directly used via userspace syscalls), but unsafe for normal
sockets.
Has no effect on non-lockdep kernels.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c
index 1044b6fce0d5..3d6e9a351150 100644
--- a/net/ipv6/icmp.c
+++ b/net/ipv6/icmp.c
@@ -712,6 +712,11 @@ static int icmpv6_rcv(struct sk_buff **pskb)
return 0;
}
+/*
+ * Special lock-class for __icmpv6_socket:
+ */
+static struct lock_class_key icmpv6_socket_sk_dst_lock_key;
+
int __init icmpv6_init(struct net_proto_family *ops)
{
struct sock *sk;
@@ -730,6 +735,14 @@ int __init icmpv6_init(struct net_proto_family *ops)
sk = per_cpu(__icmpv6_socket, i)->sk;
sk->sk_allocation = GFP_ATOMIC;
+ /*
+ * Split off their lock-class, because sk->sk_dst_lock
+ * gets used from softirqs, which is safe for
+ * __icmpv6_socket (because those never get directly used
+ * via userspace syscalls), but unsafe for normal sockets.
+ */
+ lockdep_set_class(&sk->sk_dst_lock,
+ &icmpv6_socket_sk_dst_lock_key);
/* Enough space for 2 64K ICMP packets, including
* sk_buff struct overhead.