From d3b060231b2e1eb7e7e9680ff93326a4ae576720 Mon Sep 17 00:00:00 2001
From: Huang Weiyi <weiyi.huang@gmail.com>
Date: Thu, 24 Jul 2008 00:44:51 +1000
Subject: powerpc: Removed duplicated include in stacktrace.c

Removed duplicated include file <linux/module.h> in
arch/powerpc/kernel/stacktrace.c.

Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/stacktrace.c | 1 -
 1 file changed, 1 deletion(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/stacktrace.c b/arch/powerpc/kernel/stacktrace.c
index f2589645870..b0dbb1daa4d 100644
--- a/arch/powerpc/kernel/stacktrace.c
+++ b/arch/powerpc/kernel/stacktrace.c
@@ -13,7 +13,6 @@
 #include <linux/module.h>
 #include <linux/sched.h>
 #include <linux/stacktrace.h>
-#include <linux/module.h>
 #include <asm/ptrace.h>
 #include <asm/processor.h>
 
-- 
cgit v1.2.3


From 2325f0a0c3d76bb515f3312ab2b16afdbffcc594 Mon Sep 17 00:00:00 2001
From: Kumar Gala <galak@kernel.crashing.org>
Date: Sat, 26 Jul 2008 05:27:33 +1000
Subject: powerpc/booke: Clean up the hardware watchpoint support

* CONFIG_BOOKE is selected by CONFIG_44x so we dont need both
* Fixed a few comments
* Go back to only using DBCR0_IDM to determine if we are using
  debug resources.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/entry_32.S | 6 +++---
 arch/powerpc/kernel/process.c  | 8 ++++----
 arch/powerpc/kernel/ptrace.c   | 7 ++++---
 arch/powerpc/kernel/signal.c   | 2 +-
 4 files changed, 12 insertions(+), 11 deletions(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index 81c8324a4a3..da52269aec1 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -148,7 +148,7 @@ transfer_to_handler:
 	/* Check to see if the dbcr0 register is set up to debug.  Use the
 	   internal debug mode bit to do this. */
 	lwz	r12,THREAD_DBCR0(r12)
-	andis.	r12,r12,(DBCR0_IDM  | DBSR_DAC1R | DBSR_DAC1W)@h
+	andis.	r12,r12,DBCR0_IDM@h
 	beq+	3f
 	/* From user and task is ptraced - load up global dbcr0 */
 	li	r12,-1			/* clear all pending debug events */
@@ -292,7 +292,7 @@ syscall_exit_cont:
 	/* If the process has its own DBCR0 value, load it up.  The internal
 	   debug mode bit tells us that dbcr0 should be loaded. */
 	lwz	r0,THREAD+THREAD_DBCR0(r2)
-	andis.	r10,r0,(DBCR0_IDM  | DBSR_DAC1R | DBSR_DAC1W)@h
+	andis.	r10,r0,DBCR0_IDM@h
 	bnel-	load_dbcr0
 #endif
 #ifdef CONFIG_44x
@@ -720,7 +720,7 @@ restore_user:
 	/* Check whether this process has its own DBCR0 value.  The internal
 	   debug mode bit tells us that dbcr0 should be loaded. */
 	lwz	r0,THREAD+THREAD_DBCR0(r2)
-	andis.	r10,r0,(DBCR0_IDM  | DBSR_DAC1R | DBSR_DAC1W)@h
+	andis.	r10,r0,DBCR0_IDM@h
 	bnel-	load_dbcr0
 #endif
 
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index db2497ccc11..e030f3bd502 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -254,7 +254,7 @@ void do_dabr(struct pt_regs *regs, unsigned long address,
 		return;
 
 	/* Clear the DAC and struct entries.  One shot trigger */
-#if (defined(CONFIG_44x) || defined(CONFIG_BOOKE))
+#if defined(CONFIG_BOOKE)
 	mtspr(SPRN_DBCR0, mfspr(SPRN_DBCR0) & ~(DBSR_DAC1R | DBSR_DAC1W
 							| DBCR0_IDM));
 #endif
@@ -286,7 +286,7 @@ int set_dabr(unsigned long dabr)
 	mtspr(SPRN_DABR, dabr);
 #endif
 
-#if defined(CONFIG_44x) || defined(CONFIG_BOOKE)
+#if defined(CONFIG_BOOKE)
 	mtspr(SPRN_DAC1, dabr);
 #endif
 
@@ -373,7 +373,7 @@ struct task_struct *__switch_to(struct task_struct *prev,
 	if (unlikely(__get_cpu_var(current_dabr) != new->thread.dabr))
 		set_dabr(new->thread.dabr);
 
-#if defined(CONFIG_44x) || defined(CONFIG_BOOKE)
+#if defined(CONFIG_BOOKE)
 	/* If new thread DAC (HW breakpoint) is the same then leave it */
 	if (new->thread.dabr)
 		set_dabr(new->thread.dabr);
@@ -568,7 +568,7 @@ void flush_thread(void)
 		current->thread.dabr = 0;
 		set_dabr(0);
 
-#if defined(CONFIG_44x) || defined(CONFIG_BOOKE)
+#if defined(CONFIG_BOOKE)
 		current->thread.dbcr0 &= ~(DBSR_DAC1R | DBSR_DAC1W);
 #endif
 	}
diff --git a/arch/powerpc/kernel/ptrace.c b/arch/powerpc/kernel/ptrace.c
index a5d0e78779c..66204cb51a1 100644
--- a/arch/powerpc/kernel/ptrace.c
+++ b/arch/powerpc/kernel/ptrace.c
@@ -717,7 +717,7 @@ void user_disable_single_step(struct task_struct *task)
 	struct pt_regs *regs = task->thread.regs;
 
 
-#if defined(CONFIG_44x) || defined(CONFIG_BOOKE)
+#if defined(CONFIG_BOOKE)
 	/* If DAC then do not single step, skip */
 	if (task->thread.dabr)
 		return;
@@ -744,10 +744,11 @@ int ptrace_set_debugreg(struct task_struct *task, unsigned long addr,
 	if (addr > 0)
 		return -EINVAL;
 
+	/* The bottom 3 bits in dabr are flags */
 	if ((data & ~0x7UL) >= TASK_SIZE)
 		return -EIO;
 
-#ifdef CONFIG_PPC64
+#ifndef CONFIG_BOOKE
 
 	/* For processors using DABR (i.e. 970), the bottom 3 bits are flags.
 	 *  It was assumed, on previous implementations, that 3 bits were
@@ -769,7 +770,7 @@ int ptrace_set_debugreg(struct task_struct *task, unsigned long addr,
 	task->thread.dabr = data;
 
 #endif
-#if defined(CONFIG_44x) || defined(CONFIG_BOOKE)
+#if defined(CONFIG_BOOKE)
 
 	/* As described above, it was assumed 3 bits were passed with the data
 	 *  address, but we will assume only the mode bits will be passed
diff --git a/arch/powerpc/kernel/signal.c b/arch/powerpc/kernel/signal.c
index 7aada783ec6..2b5eaa6c8f3 100644
--- a/arch/powerpc/kernel/signal.c
+++ b/arch/powerpc/kernel/signal.c
@@ -147,7 +147,7 @@ int do_signal(sigset_t *oldset, struct pt_regs *regs)
 	 */
 	if (current->thread.dabr) {
 		set_dabr(current->thread.dabr);
-#if defined(CONFIG_44x) || defined(CONFIG_BOOKE)
+#if defined(CONFIG_BOOKE)
 		mtspr(SPRN_DBCR0, current->thread.dbcr0);
 #endif
 	}
-- 
cgit v1.2.3


From b9fa49a9a908407d9366b0e1e7222aee81a2df5b Mon Sep 17 00:00:00 2001
From: Nathan Lynch <ntl@pobox.com>
Date: Sat, 26 Jul 2008 09:06:17 +1000
Subject: powerpc: Fix vio build warnings
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

arch/powerpc/kernel/vio.c:1034: warning: function declaration isn’t a prototype
arch/powerpc/kernel/vio.c:1035: warning: function declaration isn’t a prototype

Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/vio.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/vio.c b/arch/powerpc/kernel/vio.c
index ade8aeaa2e7..6952384cf89 100644
--- a/arch/powerpc/kernel/vio.c
+++ b/arch/powerpc/kernel/vio.c
@@ -1031,8 +1031,8 @@ void vio_cmo_set_dev_desired(struct vio_dev *viodev, size_t desired) {}
 static int vio_cmo_bus_probe(struct vio_dev *viodev) { return 0; }
 static void vio_cmo_bus_remove(struct vio_dev *viodev) {}
 static void vio_cmo_set_dma_ops(struct vio_dev *viodev) {}
-static void vio_cmo_bus_init() {}
-static void vio_cmo_sysfs_init() { }
+static void vio_cmo_bus_init(void) {}
+static void vio_cmo_sysfs_init(void) { }
 #endif /* CONFIG_PPC_SMLPAR */
 EXPORT_SYMBOL(vio_cmo_entitlement_update);
 EXPORT_SYMBOL(vio_cmo_set_dev_desired);
-- 
cgit v1.2.3


From 7d2f6075f992d33c7be829c3638b8cb72b782b19 Mon Sep 17 00:00:00 2001
From: Nathan Lynch <ntl@pobox.com>
Date: Sun, 27 Jul 2008 15:24:50 +1000
Subject: powerpc: kill useless SMT code in prom_hold_cpus

This piece of code is broken for >2 threads, and possibly in some
other subtle ways (such as comparing a value obtained from an
"ibm,ppc-interrupt-server#s" property to a value obtained from a
"reg" property) and doesn't seem to have any useful purpose in the
first place other than a dubious warning in case NR_CPUS is too
small, which probably isn't the right place to do so.

Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/prom_init.c | 39 +++------------------------------------
 1 file changed, 3 insertions(+), 36 deletions(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c
index c4ab2195b9c..b72849ac7db 100644
--- a/arch/powerpc/kernel/prom_init.c
+++ b/arch/powerpc/kernel/prom_init.c
@@ -205,8 +205,6 @@ static int __initdata mem_reserve_cnt;
 static cell_t __initdata regbuf[1024];
 
 
-#define MAX_CPU_THREADS 2
-
 /*
  * Error results ... some OF calls will return "-1" on error, some
  * will return 0, some will return either. To simplify, here are
@@ -1339,10 +1337,6 @@ static void __init prom_hold_cpus(void)
 	unsigned int reg;
 	phandle node;
 	char type[64];
-	int cpuid = 0;
-	unsigned int interrupt_server[MAX_CPU_THREADS];
-	unsigned int cpu_threads, hw_cpu_num;
-	int propsize;
 	struct prom_t *_prom = &RELOC(prom);
 	unsigned long *spinloop
 		= (void *) LOW_ADDR(__secondary_hold_spinloop);
@@ -1386,7 +1380,6 @@ static void __init prom_hold_cpus(void)
 		reg = -1;
 		prom_getprop(node, "reg", &reg, sizeof(reg));
 
-		prom_debug("\ncpuid        = 0x%x\n", cpuid);
 		prom_debug("cpu hw idx   = 0x%x\n", reg);
 
 		/* Init the acknowledge var which will be reset by
@@ -1395,28 +1388,9 @@ static void __init prom_hold_cpus(void)
 		 */
 		*acknowledge = (unsigned long)-1;
 
-		propsize = prom_getprop(node, "ibm,ppc-interrupt-server#s",
-					&interrupt_server,
-					sizeof(interrupt_server));
-		if (propsize < 0) {
-			/* no property.  old hardware has no SMT */
-			cpu_threads = 1;
-			interrupt_server[0] = reg; /* fake it with phys id */
-		} else {
-			/* We have a threaded processor */
-			cpu_threads = propsize / sizeof(u32);
-			if (cpu_threads > MAX_CPU_THREADS) {
-				prom_printf("SMT: too many threads!\n"
-					    "SMT: found %x, max is %x\n",
-					    cpu_threads, MAX_CPU_THREADS);
-				cpu_threads = 1; /* ToDo: panic? */
-			}
-		}
-
-		hw_cpu_num = interrupt_server[0];
-		if (hw_cpu_num != _prom->cpu) {
+		if (reg != _prom->cpu) {
 			/* Primary Thread of non-boot cpu */
-			prom_printf("%x : starting cpu hw idx %x... ", cpuid, reg);
+			prom_printf("starting cpu hw idx %x... ", reg);
 			call_prom("start-cpu", 3, 0, node,
 				  secondary_hold, reg);
 
@@ -1431,17 +1405,10 @@ static void __init prom_hold_cpus(void)
 		}
 #ifdef CONFIG_SMP
 		else
-			prom_printf("%x : boot cpu     %x\n", cpuid, reg);
+			prom_printf("boot cpu hw idx %x\n", reg);
 #endif /* CONFIG_SMP */
-
-		/* Reserve cpu #s for secondary threads.   They start later. */
-		cpuid += cpu_threads;
 	}
 
-	if (cpuid > NR_CPUS)
-		prom_printf("WARNING: maximum CPUs (" __stringify(NR_CPUS)
-			    ") exceeded: ignoring extras\n");
-
 	prom_debug("prom_hold_cpus: end...\n");
 }
 
-- 
cgit v1.2.3


From 9ba1984ead5d25c93d241e0ee43f8f6a252f60d9 Mon Sep 17 00:00:00 2001
From: Nathan Lynch <ntl@pobox.com>
Date: Sun, 27 Jul 2008 15:24:51 +1000
Subject: powerpc: register_cpu_online should be __cpuinit

It is called only in cpu online paths.

(caught by CONFIG_DEBUG_SECTION_MISMATCH=y)

Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/sysfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c
index 800e5e9a087..156808086ca 100644
--- a/arch/powerpc/kernel/sysfs.c
+++ b/arch/powerpc/kernel/sysfs.c
@@ -298,7 +298,7 @@ static struct sysdev_attribute pa6t_attrs[] = {
 };
 
 
-static void register_cpu_online(unsigned int cpu)
+static void __cpuinit register_cpu_online(unsigned int cpu)
 {
 	struct cpu *c = &per_cpu(cpu_devices, cpu);
 	struct sys_device *s = &c->sysdev;
-- 
cgit v1.2.3


From e2075f79a99b45a6cc10de021c93f07212098a84 Mon Sep 17 00:00:00 2001
From: Nathan Lynch <ntl@pobox.com>
Date: Sun, 27 Jul 2008 15:24:52 +1000
Subject: powerpc: Update cpu_sibling_maps dynamically

Rather doing one initialization pass over all the per-cpu
cpu_sibling_maps at boot, update the maps at cpu online/offline time.

This is a behavior change -- the thread_siblings attribute now
reflects only online siblings, whereas it would display offline
siblings before.  The new behavior matches that of x86, and is
arguably more useful.

Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/setup-common.c | 24 ------------------------
 arch/powerpc/kernel/setup_64.c     |  3 ---
 arch/powerpc/kernel/smp.c          | 32 +++++++++++++++++++++++++++++---
 3 files changed, 29 insertions(+), 30 deletions(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/setup-common.c b/arch/powerpc/kernel/setup-common.c
index 61a3f413208..9cc5a52711e 100644
--- a/arch/powerpc/kernel/setup-common.c
+++ b/arch/powerpc/kernel/setup-common.c
@@ -367,7 +367,6 @@ static void __init cpu_init_thread_core_maps(int tpc)
  * setup_cpu_maps - initialize the following cpu maps:
  *                  cpu_possible_map
  *                  cpu_present_map
- *                  cpu_sibling_map
  *
  * Having the possible map set up early allows us to restrict allocations
  * of things like irqstacks to num_possible_cpus() rather than NR_CPUS.
@@ -475,29 +474,6 @@ void __init smp_setup_cpu_maps(void)
 	 */
 	cpu_init_thread_core_maps(nthreads);
 }
-
-/*
- * Being that cpu_sibling_map is now a per_cpu array, then it cannot
- * be initialized until the per_cpu areas have been created.  This
- * function is now called from setup_per_cpu_areas().
- */
-void __init smp_setup_cpu_sibling_map(void)
-{
-#ifdef CONFIG_PPC64
-	int i, cpu, base;
-
-	for_each_possible_cpu(cpu) {
-		DBG("Sibling map for CPU %d:", cpu);
-		base = cpu_first_thread_in_core(cpu);
-		for (i = 0; i < threads_per_core; i++) {
-			cpu_set(base + i, per_cpu(cpu_sibling_map, cpu));
-			DBG(" %d", base + i);
-		}
-		DBG("\n");
-	}
-
-#endif /* CONFIG_PPC64 */
-}
 #endif /* CONFIG_SMP */
 
 #ifdef CONFIG_PCSPKR_PLATFORM
diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index 04d8de9f0fc..8b25f51f03b 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -611,9 +611,6 @@ void __init setup_per_cpu_areas(void)
 		paca[i].data_offset = ptr - __per_cpu_start;
 		memcpy(ptr, __per_cpu_start, __per_cpu_end - __per_cpu_start);
 	}
-
-	/* Now that per_cpu is setup, initialize cpu_sibling_map */
-	smp_setup_cpu_sibling_map();
 }
 #endif
 
diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
index f5ae9fa222e..3c4d07e5e06 100644
--- a/arch/powerpc/kernel/smp.c
+++ b/arch/powerpc/kernel/smp.c
@@ -41,6 +41,7 @@
 #include <asm/smp.h>
 #include <asm/time.h>
 #include <asm/machdep.h>
+#include <asm/cputhreads.h>
 #include <asm/cputable.h>
 #include <asm/system.h>
 #include <asm/mpic.h>
@@ -228,6 +229,7 @@ void __devinit smp_prepare_boot_cpu(void)
 	BUG_ON(smp_processor_id() != boot_cpuid);
 
 	cpu_set(boot_cpuid, cpu_online_map);
+	cpu_set(boot_cpuid, per_cpu(cpu_sibling_map, boot_cpuid));
 #ifdef CONFIG_PPC64
 	paca[boot_cpuid].__current = current;
 #endif
@@ -380,6 +382,7 @@ int __cpuinit __cpu_up(unsigned int cpu)
 int __devinit start_secondary(void *unused)
 {
 	unsigned int cpu = smp_processor_id();
+	int i, base;
 
 	atomic_inc(&init_mm.mm_count);
 	current->active_mm = &init_mm;
@@ -400,6 +403,14 @@ int __devinit start_secondary(void *unused)
 
 	ipi_call_lock();
 	cpu_set(cpu, cpu_online_map);
+	/* Update sibling maps */
+	base = cpu_first_thread_in_core(cpu);
+	for (i = 0; i < threads_per_core; i++) {
+		if (cpu_is_offline(base + i))
+			continue;
+		cpu_set(cpu, per_cpu(cpu_sibling_map, base + i));
+		cpu_set(base + i, per_cpu(cpu_sibling_map, cpu));
+	}
 	ipi_call_unlock();
 
 	local_irq_enable();
@@ -437,10 +448,25 @@ void __init smp_cpus_done(unsigned int max_cpus)
 #ifdef CONFIG_HOTPLUG_CPU
 int __cpu_disable(void)
 {
-	if (smp_ops->cpu_disable)
-		return smp_ops->cpu_disable();
+	int cpu = smp_processor_id();
+	int base, i;
+	int err;
 
-	return -ENOSYS;
+	if (!smp_ops->cpu_disable)
+		return -ENOSYS;
+
+	err = smp_ops->cpu_disable();
+	if (err)
+		return err;
+
+	/* Update sibling maps */
+	base = cpu_first_thread_in_core(cpu);
+	for (i = 0; i < threads_per_core; i++) {
+		cpu_clear(cpu, per_cpu(cpu_sibling_map, base + i));
+		cpu_clear(base + i, per_cpu(cpu_sibling_map, cpu));
+	}
+
+	return 0;
 }
 
 void __cpu_die(unsigned int cpu)
-- 
cgit v1.2.3


From 6558ba2b5cc3a2f22039db30616fcd07c1b28ac8 Mon Sep 17 00:00:00 2001
From: Roland McGrath <roland@redhat.com>
Date: Sun, 27 Jul 2008 16:49:50 +1000
Subject: powerpc: Call tracehook_signal_handler() when setting up signal
 frames

This makes the powerpc signal handling code call tracehook_signal_handler()
after a handler is set up.  This means that using PTRACE_SINGLESTEP to
enter a signal handler will report to ptrace on the first instruction of
the handler, instead of the second.  This is consistent with what x86 and
other machines do, and what users and debuggers want.

BenH: Fixed up the test for the trap value.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/signal.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/signal.c b/arch/powerpc/kernel/signal.c
index 2b5eaa6c8f3..e74aa0ed4e9 100644
--- a/arch/powerpc/kernel/signal.c
+++ b/arch/powerpc/kernel/signal.c
@@ -9,7 +9,7 @@
  * this archive for more details.
  */
 
-#include <linux/ptrace.h>
+#include <linux/tracehook.h>
 #include <linux/signal.h>
 #include <asm/uaccess.h>
 #include <asm/unistd.h>
@@ -177,6 +177,12 @@ int do_signal(sigset_t *oldset, struct pt_regs *regs)
 		 * its frame, and we can clear the TLF_RESTORE_SIGMASK flag.
 		 */
 		current_thread_info()->local_flags &= ~_TLF_RESTORE_SIGMASK;
+
+		/*
+		 * Let tracing know that we've done the handler setup.
+		 */
+		tracehook_signal_handler(signr, &info, &ka, regs,
+					 test_thread_flag(TIF_SINGLESTEP));
 	}
 
 	return ret;
-- 
cgit v1.2.3


From 4f72c4279eab1e5f3ed1ac4e55d4527617582392 Mon Sep 17 00:00:00 2001
From: Roland McGrath <roland@redhat.com>
Date: Sun, 27 Jul 2008 16:51:03 +1000
Subject: powerpc: Make syscall tracing use tracehook.h helpers

This changes powerpc syscall tracing to use the new tracehook.h entry
points.  There is no change, only cleanup.

In addition, the assembly changes allow do_syscall_trace_enter() to
abort the syscall without losing the information about the original
r0 value.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/entry_32.S |  7 ++++++-
 arch/powerpc/kernel/entry_64.S |  7 ++++++-
 arch/powerpc/kernel/ptrace.c   | 47 ++++++++++++++++++++----------------------
 3 files changed, 34 insertions(+), 27 deletions(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index da52269aec1..e6fca6a9014 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -343,7 +343,12 @@ syscall_dotrace:
 	stw	r0,_TRAP(r1)
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	do_syscall_trace_enter
-	lwz	r0,GPR0(r1)	/* Restore original registers */
+	/*
+	 * Restore argument registers possibly just changed.
+	 * We use the return value of do_syscall_trace_enter
+	 * for call number to look up in the table (r0).
+	 */
+	mr	r0,r3
 	lwz	r3,GPR3(r1)
 	lwz	r4,GPR4(r1)
 	lwz	r5,GPR5(r1)
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index d7369243ae4..79c089e97ce 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -214,7 +214,12 @@ syscall_dotrace:
 	bl	.save_nvgprs
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	.do_syscall_trace_enter
-	ld	r0,GPR0(r1)	/* Restore original registers */
+	/*
+	 * Restore argument registers possibly just changed.
+	 * We use the return value of do_syscall_trace_enter
+	 * for the call number to look up in the table (r0).
+	 */
+	mr	r0,r3
 	ld	r3,GPR3(r1)
 	ld	r4,GPR4(r1)
 	ld	r5,GPR5(r1)
diff --git a/arch/powerpc/kernel/ptrace.c b/arch/powerpc/kernel/ptrace.c
index 66204cb51a1..6b66cd85b43 100644
--- a/arch/powerpc/kernel/ptrace.c
+++ b/arch/powerpc/kernel/ptrace.c
@@ -22,6 +22,7 @@
 #include <linux/errno.h>
 #include <linux/ptrace.h>
 #include <linux/regset.h>
+#include <linux/tracehook.h>
 #include <linux/elf.h>
 #include <linux/user.h>
 #include <linux/security.h>
@@ -1014,31 +1015,24 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data)
 	return ret;
 }
 
-static void do_syscall_trace(void)
+/*
+ * We must return the syscall number to actually look up in the table.
+ * This can be -1L to skip running any syscall at all.
+ */
+long do_syscall_trace_enter(struct pt_regs *regs)
 {
-	/* the 0x80 provides a way for the tracing parent to distinguish
-	   between a syscall stop and SIGTRAP delivery */
-	ptrace_notify(SIGTRAP | ((current->ptrace & PT_TRACESYSGOOD)
-				 ? 0x80 : 0));
-
-	/*
-	 * this isn't the same as continuing with a signal, but it will do
-	 * for normal use.  strace only continues with a signal if the
-	 * stopping signal is not SIGTRAP.  -brl
-	 */
-	if (current->exit_code) {
-		send_sig(current->exit_code, current, 1);
-		current->exit_code = 0;
-	}
-}
+	long ret = 0;
 
-void do_syscall_trace_enter(struct pt_regs *regs)
-{
 	secure_computing(regs->gpr[0]);
 
-	if (test_thread_flag(TIF_SYSCALL_TRACE)
-	    && (current->ptrace & PT_PTRACED))
-		do_syscall_trace();
+	if (test_thread_flag(TIF_SYSCALL_TRACE) &&
+	    tracehook_report_syscall_entry(regs))
+		/*
+		 * Tracing decided this syscall should not happen.
+		 * We'll return a bogus call number to get an ENOSYS
+		 * error, but leave the original number in regs->gpr[0].
+		 */
+		ret = -1L;
 
 	if (unlikely(current->audit_context)) {
 #ifdef CONFIG_PPC64
@@ -1056,16 +1050,19 @@ void do_syscall_trace_enter(struct pt_regs *regs)
 					    regs->gpr[5] & 0xffffffff,
 					    regs->gpr[6] & 0xffffffff);
 	}
+
+	return ret ?: regs->gpr[0];
 }
 
 void do_syscall_trace_leave(struct pt_regs *regs)
 {
+	int step;
+
 	if (unlikely(current->audit_context))
 		audit_syscall_exit((regs->ccr&0x10000000)?AUDITSC_FAILURE:AUDITSC_SUCCESS,
 				   regs->result);
 
-	if ((test_thread_flag(TIF_SYSCALL_TRACE)
-	     || test_thread_flag(TIF_SINGLESTEP))
-	    && (current->ptrace & PT_PTRACED))
-		do_syscall_trace();
+	step = test_thread_flag(TIF_SINGLESTEP);
+	if (step || test_thread_flag(TIF_SYSCALL_TRACE))
+		tracehook_report_syscall_exit(regs, step);
 }
-- 
cgit v1.2.3


From 7d6d637dac2050f30a1b57b0a3dc5de4a10616ba Mon Sep 17 00:00:00 2001
From: Roland McGrath <roland@redhat.com>
Date: Sun, 27 Jul 2008 16:52:52 +1000
Subject: powerpc: Add TIF_NOTIFY_RESUME support for tracehook

This adds TIF_NOTIFY_RESUME support for powerpc.  When set,
we call tracehook_notify_resume() on the way to user mode.
This overloads do_signal() to do the work, but changes its
arguments to it has the TIF_* bits handy in a register and
drops the useless first argument that was always zero.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/entry_32.S |  4 ++--
 arch/powerpc/kernel/entry_64.S |  3 +--
 arch/powerpc/kernel/signal.c   | 13 ++++++++++++-
 3 files changed, 15 insertions(+), 5 deletions(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index e6fca6a9014..1cbbf703364 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -1060,8 +1060,8 @@ do_user_signal:			/* r10 contains MSR_KERNEL here */
 	SAVE_NVGPRS(r1)
 	rlwinm	r3,r3,0,0,30
 	stw	r3,_TRAP(r1)
-2:	li	r3,0
-	addi	r4,r1,STACK_FRAME_OVERHEAD
+2:	addi	r3,r1,STACK_FRAME_OVERHEAD
+	mr	r4,r9
 	bl	do_signal
 	REST_NVGPRS(r1)
 	b	recheck
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 79c089e97ce..2d802e97097 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -643,8 +643,7 @@ user_work:
 	b	.ret_from_except_lite
 
 1:	bl	.save_nvgprs
-	li	r3,0
-	addi	r4,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	.do_signal
 	b	.ret_from_except
 
diff --git a/arch/powerpc/kernel/signal.c b/arch/powerpc/kernel/signal.c
index e74aa0ed4e9..a54405ebd7b 100644
--- a/arch/powerpc/kernel/signal.c
+++ b/arch/powerpc/kernel/signal.c
@@ -112,7 +112,7 @@ static void check_syscall_restart(struct pt_regs *regs, struct k_sigaction *ka,
 	}
 }
 
-int do_signal(sigset_t *oldset, struct pt_regs *regs)
+static int do_signal_pending(sigset_t *oldset, struct pt_regs *regs)
 {
 	siginfo_t info;
 	int signr;
@@ -188,6 +188,17 @@ int do_signal(sigset_t *oldset, struct pt_regs *regs)
 	return ret;
 }
 
+void do_signal(struct pt_regs *regs, unsigned long thread_info_flags)
+{
+	if (thread_info_flags & _TIF_SIGPENDING)
+		do_signal_pending(NULL, regs);
+
+	if (thread_info_flags & _TIF_NOTIFY_RESUME) {
+		clear_thread_flag(TIF_NOTIFY_RESUME);
+		tracehook_notify_resume(regs);
+	}
+}
+
 long sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss,
 		unsigned long r5, unsigned long r6, unsigned long r7,
 		unsigned long r8, struct pt_regs *regs)
-- 
cgit v1.2.3


From dec2b0d0cc64e495a3c9c5700fa7a3b397b5e855 Mon Sep 17 00:00:00 2001
From: Roland McGrath <roland@redhat.com>
Date: Sun, 27 Jul 2008 16:53:20 +1000
Subject: powerpc: Enable tracehook for the architecture

The powerpc arch code has all the prerequisites, so set HAVE_ARCH_TRACEHOOK.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/Kconfig | 1 +
 1 file changed, 1 insertion(+)

(limited to 'arch')

diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index fe88418167c..587da5e0990 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -117,6 +117,7 @@ config PPC
 	select HAVE_KPROBES
 	select HAVE_ARCH_KGDB
 	select HAVE_KRETPROBES
+	select HAVE_ARCH_TRACEHOOK
 	select HAVE_LMB
 	select HAVE_DMA_ATTRS if PPC64
 	select USE_GENERIC_SMP_HELPERS if SMP
-- 
cgit v1.2.3


From 3cee67f77922721e90c1573d84c07e18c5508713 Mon Sep 17 00:00:00 2001
From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Mon, 28 Jul 2008 00:51:02 +1000
Subject: powerpc/pseries: Fix CMO sysdev attribute API change fallout

Noticed due to these wanings:

arch/powerpc/platforms/pseries/cmm.c:298: warning: initialization from incompatible pointer type
arch/powerpc/platforms/pseries/cmm.c:299: warning: initialization from incompatible pointer type
arch/powerpc/platforms/pseries/cmm.c:320: warning: initialization from incompatible pointer type
arch/powerpc/platforms/pseries/cmm.c:320: warning: initialization from incompatible pointer type

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/platforms/pseries/cmm.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

(limited to 'arch')

diff --git a/arch/powerpc/platforms/pseries/cmm.c b/arch/powerpc/platforms/pseries/cmm.c
index c6b3be03168..38fe32a7cc7 100644
--- a/arch/powerpc/platforms/pseries/cmm.c
+++ b/arch/powerpc/platforms/pseries/cmm.c
@@ -289,7 +289,9 @@ static int cmm_thread(void *dummy)
 }
 
 #define CMM_SHOW(name, format, args...)			\
-	static ssize_t show_##name(struct sys_device *dev, char *buf)	\
+	static ssize_t show_##name(struct sys_device *dev,	\
+				   struct sysdev_attribute *attr,	\
+				   char *buf)			\
 	{							\
 		return sprintf(buf, format, ##args);		\
 	}							\
@@ -298,12 +300,14 @@ static int cmm_thread(void *dummy)
 CMM_SHOW(loaned_kb, "%lu\n", PAGES2KB(loaned_pages));
 CMM_SHOW(loaned_target_kb, "%lu\n", PAGES2KB(loaned_pages_target));
 
-static ssize_t show_oom_pages(struct sys_device *dev, char *buf)
+static ssize_t show_oom_pages(struct sys_device *dev,
+			      struct sysdev_attribute *attr, char *buf)
 {
 	return sprintf(buf, "%lu\n", PAGES2KB(oom_freed_pages));
 }
 
 static ssize_t store_oom_pages(struct sys_device *dev,
+			       struct sysdev_attribute *attr,
 			       const char *buf, size_t count)
 {
 	unsigned long val = simple_strtoul (buf, NULL, 10);
-- 
cgit v1.2.3


From 0764bf63da5466474eebf7d21994cf6b106265a3 Mon Sep 17 00:00:00 2001
From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Mon, 28 Jul 2008 02:22:14 +1000
Subject: powerpc/vio: More fallout from dma_mapping_error API change

arch/powerpc/kernel/vio.c:533: error: too few arguments to function 'dma_mapping_error'

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/vio.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/vio.c b/arch/powerpc/kernel/vio.c
index 6952384cf89..22a3c33fd75 100644
--- a/arch/powerpc/kernel/vio.c
+++ b/arch/powerpc/kernel/vio.c
@@ -530,7 +530,7 @@ static dma_addr_t vio_dma_iommu_map_single(struct device *dev, void *vaddr,
 	}
 
 	ret = dma_iommu_ops.map_single(dev, vaddr, size, direction, attrs);
-	if (unlikely(dma_mapping_error(ret))) {
+	if (unlikely(dma_mapping_error(dev, ret))) {
 		vio_cmo_dealloc(viodev, roundup(size, IOMMU_PAGE_SIZE));
 		atomic_inc(&viodev->cmo.allocs_failed);
 	}
-- 
cgit v1.2.3


From 440a0857e32a05979fb01fc59ea454a723e80e4b Mon Sep 17 00:00:00 2001
From: Nathan Lynch <ntl@pobox.com>
Date: Sun, 27 Jul 2008 15:24:53 +1000
Subject: powerpc: Make core sibling information available to userspace

Implement the notion of "core siblings" for powerpc.  This makes
/sys/devices/system/cpu/cpu*/topology/core_siblings present sensible
values, indicating online CPUs which share an L2 cache.

BenH: Made cpu_to_l2cache() use of_find_node_by_phandle() instead
of IBM-specific open coded search

Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/smp.c | 64 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 64 insertions(+)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
index 3c4d07e5e06..f7a2f81b5b7 100644
--- a/arch/powerpc/kernel/smp.c
+++ b/arch/powerpc/kernel/smp.c
@@ -63,10 +63,12 @@ struct thread_info *secondary_ti;
 cpumask_t cpu_possible_map = CPU_MASK_NONE;
 cpumask_t cpu_online_map = CPU_MASK_NONE;
 DEFINE_PER_CPU(cpumask_t, cpu_sibling_map) = CPU_MASK_NONE;
+DEFINE_PER_CPU(cpumask_t, cpu_core_map) = CPU_MASK_NONE;
 
 EXPORT_SYMBOL(cpu_online_map);
 EXPORT_SYMBOL(cpu_possible_map);
 EXPORT_PER_CPU_SYMBOL(cpu_sibling_map);
+EXPORT_PER_CPU_SYMBOL(cpu_core_map);
 
 /* SMP operations for this machine */
 struct smp_ops_t *smp_ops;
@@ -230,6 +232,7 @@ void __devinit smp_prepare_boot_cpu(void)
 
 	cpu_set(boot_cpuid, cpu_online_map);
 	cpu_set(boot_cpuid, per_cpu(cpu_sibling_map, boot_cpuid));
+	cpu_set(boot_cpuid, per_cpu(cpu_core_map, boot_cpuid));
 #ifdef CONFIG_PPC64
 	paca[boot_cpuid].__current = current;
 #endif
@@ -377,11 +380,36 @@ int __cpuinit __cpu_up(unsigned int cpu)
 	return 0;
 }
 
+/* Must be called when no change can occur to cpu_present_map,
+ * i.e. during cpu online or offline.
+ */
+static struct device_node *cpu_to_l2cache(int cpu)
+{
+	struct device_node *np;
+	const phandle *php;
+	phandle ph;
+
+	if (!cpu_present(cpu))
+		return NULL;
+
+	np = of_get_cpu_node(cpu, NULL);
+	if (np == NULL)
+		return NULL;
+
+	php = of_get_property(np, "l2-cache", NULL);
+	if (php == NULL)
+		return NULL;
+	ph = *php;
+	of_node_put(np);
+
+	return of_find_node_by_phandle(ph);
+}
 
 /* Activate a secondary processor. */
 int __devinit start_secondary(void *unused)
 {
 	unsigned int cpu = smp_processor_id();
+	struct device_node *l2_cache;
 	int i, base;
 
 	atomic_inc(&init_mm.mm_count);
@@ -410,7 +438,26 @@ int __devinit start_secondary(void *unused)
 			continue;
 		cpu_set(cpu, per_cpu(cpu_sibling_map, base + i));
 		cpu_set(base + i, per_cpu(cpu_sibling_map, cpu));
+
+		/* cpu_core_map should be a superset of
+		 * cpu_sibling_map even if we don't have cache
+		 * information, so update the former here, too.
+		 */
+		cpu_set(cpu, per_cpu(cpu_core_map, base +i));
+		cpu_set(base + i, per_cpu(cpu_core_map, cpu));
 	}
+	l2_cache = cpu_to_l2cache(cpu);
+	for_each_online_cpu(i) {
+		struct device_node *np = cpu_to_l2cache(i);
+		if (!np)
+			continue;
+		if (np == l2_cache) {
+			cpu_set(cpu, per_cpu(cpu_core_map, i));
+			cpu_set(i, per_cpu(cpu_core_map, cpu));
+		}
+		of_node_put(np);
+	}
+	of_node_put(l2_cache);
 	ipi_call_unlock();
 
 	local_irq_enable();
@@ -448,6 +495,7 @@ void __init smp_cpus_done(unsigned int max_cpus)
 #ifdef CONFIG_HOTPLUG_CPU
 int __cpu_disable(void)
 {
+	struct device_node *l2_cache;
 	int cpu = smp_processor_id();
 	int base, i;
 	int err;
@@ -464,7 +512,23 @@ int __cpu_disable(void)
 	for (i = 0; i < threads_per_core; i++) {
 		cpu_clear(cpu, per_cpu(cpu_sibling_map, base + i));
 		cpu_clear(base + i, per_cpu(cpu_sibling_map, cpu));
+		cpu_clear(cpu, per_cpu(cpu_core_map, base +i));
+		cpu_clear(base + i, per_cpu(cpu_core_map, cpu));
+	}
+
+	l2_cache = cpu_to_l2cache(cpu);
+	for_each_present_cpu(i) {
+		struct device_node *np = cpu_to_l2cache(i);
+		if (!np)
+			continue;
+		if (np == l2_cache) {
+			cpu_clear(cpu, per_cpu(cpu_core_map, i));
+			cpu_clear(i, per_cpu(cpu_core_map, cpu));
+		}
+		of_node_put(np);
 	}
+	of_node_put(l2_cache);
+
 
 	return 0;
 }
-- 
cgit v1.2.3


From e9efed3b80a83e44b98fc626f3268ae072550b84 Mon Sep 17 00:00:00 2001
From: Nathan Lynch <ntl@pobox.com>
Date: Sun, 27 Jul 2008 15:24:54 +1000
Subject: powerpc: Make core id information available to userspace

Existing Open Firmware practice is to report each processor core as a
separate node in the device tree.  Report the value of the "reg" OF
property corresponding to a logical CPU's device node as the core_id
attribute in /sys/devices/system/cpu/cpu*/topology/core_id.

Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/smp.c | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
index f7a2f81b5b7..5337ca7bb64 100644
--- a/arch/powerpc/kernel/smp.c
+++ b/arch/powerpc/kernel/smp.c
@@ -380,6 +380,29 @@ int __cpuinit __cpu_up(unsigned int cpu)
 	return 0;
 }
 
+/* Return the value of the reg property corresponding to the given
+ * logical cpu.
+ */
+int cpu_to_core_id(int cpu)
+{
+	struct device_node *np;
+	const int *reg;
+	int id = -1;
+
+	np = of_get_cpu_node(cpu, NULL);
+	if (!np)
+		goto out;
+
+	reg = of_get_property(np, "reg", NULL);
+	if (!reg)
+		goto out;
+
+	id = *reg;
+out:
+	of_node_put(np);
+	return id;
+}
+
 /* Must be called when no change can occur to cpu_present_map,
  * i.e. during cpu online or offline.
  */
-- 
cgit v1.2.3


From 124c27d375f72dd857eac27f2932f9f01df76bf4 Mon Sep 17 00:00:00 2001
From: Nathan Lynch <ntl@pobox.com>
Date: Sun, 27 Jul 2008 15:24:55 +1000
Subject: powerpc: Show processor cache information in sysfs

Collect cache information from the OF device tree and display it in
the cpu hierarchy in sysfs.  This is intended to be compatible at the
userspace level with x86's implementation[1], hence some of the funny
attribute names.  The arrangement of cache info is not immediately
intuitive, but (again) it's for compatibility's sake.

The cache attributes exposed are:

type (Data, Instruction, or Unified)
level (1, 2, 3...)
size
coherency_line_size
number_of_sets
ways_of_associativity

All of these can be derived on platforms that follow the OF PowerPC
Processor binding.  The code "publishes" only those attributes for
which it is able to determine values; attributes for values which
cannot be determined are not created at all.

[1] arch/x86/kernel/cpu/intel_cacheinfo.c

BenH: Turned some printk's into pr_debug, added better NULL checking
in a couple of places.

Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/sysfs.c | 309 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 309 insertions(+)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c
index 156808086ca..56d172d16e5 100644
--- a/arch/powerpc/kernel/sysfs.c
+++ b/arch/powerpc/kernel/sysfs.c
@@ -22,6 +22,8 @@
 
 static DEFINE_PER_CPU(struct cpu, cpu_devices);
 
+static DEFINE_PER_CPU(struct kobject *, cache_toplevel);
+
 /* SMT stuff */
 
 #ifdef CONFIG_PPC_MULTIPLATFORM
@@ -297,6 +299,287 @@ static struct sysdev_attribute pa6t_attrs[] = {
 #endif /* CONFIG_DEBUG_KERNEL */
 };
 
+struct cache_desc {
+	struct kobject kobj;
+	struct cache_desc *next;
+	const char *type;	/* Instruction, Data, or Unified */
+	u32 size;		/* total cache size in KB */
+	u32 line_size;		/* in bytes */
+	u32 nr_sets;		/* number of sets */
+	u32 level;		/* e.g. 1, 2, 3... */
+	u32 associativity;	/* e.g. 8-way... 0 is fully associative */
+};
+
+DEFINE_PER_CPU(struct cache_desc *, cache_desc);
+
+static struct cache_desc *kobj_to_cache_desc(struct kobject *k)
+{
+	return container_of(k, struct cache_desc, kobj);
+}
+
+static void cache_desc_release(struct kobject *k)
+{
+	struct cache_desc *desc = kobj_to_cache_desc(k);
+
+	pr_debug("%s: releasing %s\n", __func__, kobject_name(k));
+
+	if (desc->next)
+		kobject_put(&desc->next->kobj);
+
+	kfree(kobj_to_cache_desc(k));
+}
+
+static ssize_t cache_desc_show(struct kobject *k, struct attribute *attr, char *buf)
+{
+	struct kobj_attribute *kobj_attr;
+
+	kobj_attr = container_of(attr, struct kobj_attribute, attr);
+
+	return kobj_attr->show(k, kobj_attr, buf);
+}
+
+static struct sysfs_ops cache_desc_sysfs_ops = {
+	.show = cache_desc_show,
+};
+
+static struct kobj_type cache_desc_type = {
+	.release = cache_desc_release,
+	.sysfs_ops = &cache_desc_sysfs_ops,
+};
+
+static ssize_t cache_size_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
+{
+	struct cache_desc *cache = kobj_to_cache_desc(k);
+
+	return sprintf(buf, "%uK\n", cache->size);
+}
+
+static struct kobj_attribute cache_size_attr =
+	__ATTR(size, 0444, cache_size_show, NULL);
+
+static ssize_t cache_line_size_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
+{
+	struct cache_desc *cache = kobj_to_cache_desc(k);
+
+	return sprintf(buf, "%u\n", cache->line_size);
+}
+
+static struct kobj_attribute cache_line_size_attr =
+	__ATTR(coherency_line_size, 0444, cache_line_size_show, NULL);
+
+static ssize_t cache_nr_sets_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
+{
+	struct cache_desc *cache = kobj_to_cache_desc(k);
+
+	return sprintf(buf, "%u\n", cache->nr_sets);
+}
+
+static struct kobj_attribute cache_nr_sets_attr =
+	__ATTR(number_of_sets, 0444, cache_nr_sets_show, NULL);
+
+static ssize_t cache_type_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
+{
+	struct cache_desc *cache = kobj_to_cache_desc(k);
+
+	return sprintf(buf, "%s\n", cache->type);
+}
+
+static struct kobj_attribute cache_type_attr =
+	__ATTR(type, 0444, cache_type_show, NULL);
+
+static ssize_t cache_level_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
+{
+	struct cache_desc *cache = kobj_to_cache_desc(k);
+
+	return sprintf(buf, "%u\n", cache->level);
+}
+
+static struct kobj_attribute cache_level_attr =
+	__ATTR(level, 0444, cache_level_show, NULL);
+
+static ssize_t cache_assoc_show(struct kobject *k, struct kobj_attribute *attr, char *buf)
+{
+	struct cache_desc *cache = kobj_to_cache_desc(k);
+
+	return sprintf(buf, "%u\n", cache->associativity);
+}
+
+static struct kobj_attribute cache_assoc_attr =
+	__ATTR(ways_of_associativity, 0444, cache_assoc_show, NULL);
+
+struct cache_desc_info {
+	const char *type;
+	const char *size_prop;
+	const char *line_size_prop;
+	const char *nr_sets_prop;
+};
+
+/* PowerPC Processor binding says the [di]-cache-* must be equal on
+ * unified caches, so just use d-cache properties. */
+static struct cache_desc_info ucache_info = {
+	.type = "Unified",
+	.size_prop = "d-cache-size",
+	.line_size_prop = "d-cache-line-size",
+	.nr_sets_prop = "d-cache-sets",
+};
+
+static struct cache_desc_info dcache_info = {
+	.type = "Data",
+	.size_prop = "d-cache-size",
+	.line_size_prop = "d-cache-line-size",
+	.nr_sets_prop = "d-cache-sets",
+};
+
+static struct cache_desc_info icache_info = {
+	.type = "Instruction",
+	.size_prop = "i-cache-size",
+	.line_size_prop = "i-cache-line-size",
+	.nr_sets_prop = "i-cache-sets",
+};
+
+static struct cache_desc * __cpuinit create_cache_desc(struct device_node *np, struct kobject *parent, int index, int level, struct cache_desc_info *info)
+{
+	const u32 *cache_line_size;
+	struct cache_desc *new;
+	const u32 *cache_size;
+	const u32 *nr_sets;
+	int rc;
+
+	new = kzalloc(sizeof(*new), GFP_KERNEL);
+	if (!new)
+		return NULL;
+
+	rc = kobject_init_and_add(&new->kobj, &cache_desc_type, parent,
+				  "index%d", index);
+	if (rc)
+		goto err;
+
+	/* type */
+	new->type = info->type;
+	rc = sysfs_create_file(&new->kobj, &cache_type_attr.attr);
+	WARN_ON(rc);
+
+	/* level */
+	new->level = level;
+	rc = sysfs_create_file(&new->kobj, &cache_level_attr.attr);
+	WARN_ON(rc);
+
+	/* size */
+	cache_size = of_get_property(np, info->size_prop, NULL);
+	if (cache_size) {
+		new->size = *cache_size / 1024;
+		rc = sysfs_create_file(&new->kobj,
+				       &cache_size_attr.attr);
+		WARN_ON(rc);
+	}
+
+	/* coherency_line_size */
+	cache_line_size = of_get_property(np, info->line_size_prop, NULL);
+	if (cache_line_size) {
+		new->line_size = *cache_line_size;
+		rc = sysfs_create_file(&new->kobj,
+				       &cache_line_size_attr.attr);
+		WARN_ON(rc);
+	}
+
+	/* number_of_sets */
+	nr_sets = of_get_property(np, info->nr_sets_prop, NULL);
+	if (nr_sets) {
+		new->nr_sets = *nr_sets;
+		rc = sysfs_create_file(&new->kobj,
+				       &cache_nr_sets_attr.attr);
+		WARN_ON(rc);
+	}
+
+	/* ways_of_associativity */
+	if (new->nr_sets == 1) {
+		/* fully associative */
+		new->associativity = 0;
+		goto create_assoc;
+	}
+
+	if (new->nr_sets && new->size && new->line_size) {
+		/* If we have values for all of these we can derive
+		 * the associativity. */
+		new->associativity =
+			((new->size * 1024) / new->nr_sets) / new->line_size;
+create_assoc:
+		rc = sysfs_create_file(&new->kobj,
+				       &cache_assoc_attr.attr);
+		WARN_ON(rc);
+	}
+
+	return new;
+err:
+	kfree(new);
+	return NULL;
+}
+
+static bool cache_is_unified(struct device_node *np)
+{
+	return of_get_property(np, "cache-unified", NULL);
+}
+
+static struct cache_desc * __cpuinit create_cache_index_info(struct device_node *np, struct kobject *parent, int index, int level)
+{
+	const phandle *next_cache_phandle;
+	struct device_node *next_cache;
+	struct cache_desc *new, **end;
+
+	pr_debug("%s(node = %s, index = %d)\n", __func__, np->full_name, index);
+
+	if (cache_is_unified(np)) {
+		new = create_cache_desc(np, parent, index, level,
+					&ucache_info);
+	} else {
+		new = create_cache_desc(np, parent, index, level,
+					&dcache_info);
+		if (new) {
+			index++;
+			new->next = create_cache_desc(np, parent, index, level,
+						      &icache_info);
+		}
+	}
+	if (!new)
+		return NULL;
+
+	end = &new->next;
+	while (*end)
+		end = &(*end)->next;
+
+	next_cache_phandle = of_get_property(np, "l2-cache", NULL);
+	if (!next_cache_phandle)
+		goto out;
+
+	next_cache = of_find_node_by_phandle(*next_cache_phandle);
+	if (!next_cache)
+		goto out;
+
+	*end = create_cache_index_info(next_cache, parent, ++index, ++level);
+
+	of_node_put(next_cache);
+out:
+	return new;
+}
+
+static void __cpuinit create_cache_info(struct sys_device *sysdev)
+{
+	struct kobject *cache_toplevel;
+	struct device_node *np = NULL;
+	int cpu = sysdev->id;
+
+	cache_toplevel = kobject_create_and_add("cache", &sysdev->kobj);
+	if (!cache_toplevel)
+		return;
+	per_cpu(cache_toplevel, cpu) = cache_toplevel;
+	np = of_get_cpu_node(cpu, NULL);
+	if (np != NULL) {
+		per_cpu(cache_desc, cpu) =
+			create_cache_index_info(np, cache_toplevel, 0, 1);
+		of_node_put(np);
+	}
+	return;
+}
 
 static void __cpuinit register_cpu_online(unsigned int cpu)
 {
@@ -346,9 +629,33 @@ static void __cpuinit register_cpu_online(unsigned int cpu)
 
 	if (cpu_has_feature(CPU_FTR_DSCR))
 		sysdev_create_file(s, &attr_dscr);
+
+	create_cache_info(s);
 }
 
 #ifdef CONFIG_HOTPLUG_CPU
+static void remove_cache_info(struct sys_device *sysdev)
+{
+	struct kobject *cache_toplevel;
+	struct cache_desc *cache_desc;
+	int cpu = sysdev->id;
+
+	cache_desc = per_cpu(cache_desc, cpu);
+	if (cache_desc != NULL) {
+		sysfs_remove_file(&cache_desc->kobj, &cache_size_attr.attr);
+		sysfs_remove_file(&cache_desc->kobj, &cache_line_size_attr.attr);
+		sysfs_remove_file(&cache_desc->kobj, &cache_type_attr.attr);
+		sysfs_remove_file(&cache_desc->kobj, &cache_level_attr.attr);
+		sysfs_remove_file(&cache_desc->kobj, &cache_nr_sets_attr.attr);
+		sysfs_remove_file(&cache_desc->kobj, &cache_assoc_attr.attr);
+
+		kobject_put(&cache_desc->kobj);
+	}
+	cache_toplevel = per_cpu(cache_toplevel, cpu);
+	if (cache_toplevel != NULL)
+		kobject_put(cache_toplevel);
+}
+
 static void unregister_cpu_online(unsigned int cpu)
 {
 	struct cpu *c = &per_cpu(cpu_devices, cpu);
@@ -399,6 +706,8 @@ static void unregister_cpu_online(unsigned int cpu)
 
 	if (cpu_has_feature(CPU_FTR_DSCR))
 		sysdev_remove_file(s, &attr_dscr);
+
+	remove_cache_info(s);
 }
 #endif /* CONFIG_HOTPLUG_CPU */
 
-- 
cgit v1.2.3


From f023bf0f91f1f1b926ec8f5cf0ee24be134bf024 Mon Sep 17 00:00:00 2001
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Date: Mon, 28 Jul 2008 12:06:19 +1000
Subject: powerpc/powermac: Use sane default baudrate for SCC debugging

When using the "sccdbg" option to route early kernel messages and
xmon to the SCC serial port on PowerMacs, when this wasn't the
configured output port of Open Firmware, we initialize the baudrate
to 57600bps. This isn't a very good default on some powermacs where
both the FW and pmac_zilog will default to 38400. This fixes it to
use the same logic as pmac_zilog to pick a default speed.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/platforms/powermac/udbg_scc.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

(limited to 'arch')

diff --git a/arch/powerpc/platforms/powermac/udbg_scc.c b/arch/powerpc/platforms/powermac/udbg_scc.c
index 47de4d3fc16..572771fd846 100644
--- a/arch/powerpc/platforms/powermac/udbg_scc.c
+++ b/arch/powerpc/platforms/powermac/udbg_scc.c
@@ -125,13 +125,23 @@ void udbg_scc_init(int force_scc)
 	out_8(sccc, 0xc0);
 
 	/* If SCC was the OF output port, read the BRG value, else
-	 * Setup for 57600 8N1
+	 * Setup for 38400 or 57600 8N1 depending on the machine
 	 */
 	if (ch_def != NULL) {
 		out_8(sccc, 13);
 		scc_inittab[1] = in_8(sccc);
 		out_8(sccc, 12);
 		scc_inittab[3] = in_8(sccc);
+	} else if (machine_is_compatible("RackMac1,1")
+		   || machine_is_compatible("RackMac1,2")
+		   || machine_is_compatible("MacRISC4")) {
+		/* Xserves and G5s default to 57600 */
+		scc_inittab[1] = 0;
+		scc_inittab[3] = 0;
+	} else {
+		/* Others default to 38400 */
+		scc_inittab[1] = 0;
+		scc_inittab[3] = 1;
 	}
 
 	for (i = 0; i < sizeof(scc_inittab); ++i)
-- 
cgit v1.2.3


From 025d7917a5ede982a5669c6735ef73a227b9827e Mon Sep 17 00:00:00 2001
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Date: Mon, 28 Jul 2008 13:49:15 +1000
Subject: powerpc/powermac: Fixup default serial port device for pmac_zilog

This removes the non-working code in legacy_serial that tried to handle
the powermac SCC ports, and instead add a (now working) function to the
powermac platform code to find the default serial console if any.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/legacy_serial.c     | 44 ++++++++------------
 arch/powerpc/platforms/powermac/setup.c | 72 +++++++++++++++++++++++++++++++++
 2 files changed, 89 insertions(+), 27 deletions(-)

(limited to 'arch')

diff --git a/arch/powerpc/kernel/legacy_serial.c b/arch/powerpc/kernel/legacy_serial.c
index 4d96e1db55e..9ddfaef1a18 100644
--- a/arch/powerpc/kernel/legacy_serial.c
+++ b/arch/powerpc/kernel/legacy_serial.c
@@ -493,18 +493,18 @@ static int __init serial_dev_init(void)
 device_initcall(serial_dev_init);
 
 
+#ifdef CONFIG_SERIAL_8250_CONSOLE
 /*
  * This is called very early, as part of console_init() (typically just after
  * time_init()). This function is respondible for trying to find a good
  * default console on serial ports. It tries to match the open firmware
- * default output with one of the available serial console drivers, either
- * one of the platform serial ports that have been probed earlier by
- * find_legacy_serial_ports() or some more platform specific ones.
+ * default output with one of the available serial console drivers that have
+ * been probed earlier by find_legacy_serial_ports()
  */
 static int __init check_legacy_serial_console(void)
 {
 	struct device_node *prom_stdout = NULL;
-	int speed = 0, offset = 0;
+	int i, speed = 0, offset = 0;
 	const char *name;
 	const u32 *spd;
 
@@ -548,31 +548,20 @@ static int __init check_legacy_serial_console(void)
 	if (spd)
 		speed = *spd;
 
-	if (0)
-		;
-#ifdef CONFIG_SERIAL_8250_CONSOLE
-	else if (strcmp(name, "serial") == 0) {
-		int i;
-		/* Look for it in probed array */
-		for (i = 0; i < legacy_serial_count; i++) {
-			if (prom_stdout != legacy_serial_infos[i].np)
-				continue;
-			offset = i;
-			speed = legacy_serial_infos[i].speed;
-			break;
-		}
-		if (i >= legacy_serial_count)
-			goto not_found;
+	if (strcmp(name, "serial") != 0)
+		goto not_found;
+
+	/* Look for it in probed array */
+	for (i = 0; i < legacy_serial_count; i++) {
+		if (prom_stdout != legacy_serial_infos[i].np)
+			continue;
+		offset = i;
+		speed = legacy_serial_infos[i].speed;
+		break;
 	}
-#endif /* CONFIG_SERIAL_8250_CONSOLE */
-#ifdef CONFIG_SERIAL_PMACZILOG_CONSOLE
-	else if (strcmp(name, "ch-a") == 0)
-		offset = 0;
-	else if (strcmp(name, "ch-b") == 0)
-		offset = 1;
-#endif /* CONFIG_SERIAL_PMACZILOG_CONSOLE */
-	else
+	if (i >= legacy_serial_count)
 		goto not_found;
+
 	of_node_put(prom_stdout);
 
 	DBG("Found serial console at ttyS%d\n", offset);
@@ -591,3 +580,4 @@ static int __init check_legacy_serial_console(void)
 }
 console_initcall(check_legacy_serial_console);
 
+#endif /* CONFIG_SERIAL_8250_CONSOLE */
diff --git a/arch/powerpc/platforms/powermac/setup.c b/arch/powerpc/platforms/powermac/setup.c
index 31635446901..88ccf3a08a9 100644
--- a/arch/powerpc/platforms/powermac/setup.c
+++ b/arch/powerpc/platforms/powermac/setup.c
@@ -541,6 +541,78 @@ static int __init pmac_declare_of_platform_devices(void)
 }
 machine_device_initcall(powermac, pmac_declare_of_platform_devices);
 
+#ifdef CONFIG_SERIAL_PMACZILOG_CONSOLE
+/*
+ * This is called very early, as part of console_init() (typically just after
+ * time_init()). This function is respondible for trying to find a good
+ * default console on serial ports. It tries to match the open firmware
+ * default output with one of the available serial console drivers.
+ */
+static int __init check_pmac_serial_console(void)
+{
+	struct device_node *prom_stdout = NULL;
+	int offset = 0;
+	const char *name;
+#ifdef CONFIG_SERIAL_PMACZILOG_TTYS
+	char *devname = "ttyS";
+#else
+	char *devname = "ttyPZ";
+#endif
+
+	pr_debug(" -> check_pmac_serial_console()\n");
+
+	/* The user has requested a console so this is already set up. */
+	if (strstr(boot_command_line, "console=")) {
+		pr_debug(" console was specified !\n");
+		return -EBUSY;
+	}
+
+	if (!of_chosen) {
+		pr_debug(" of_chosen is NULL !\n");
+		return -ENODEV;
+	}
+
+	/* We are getting a weird phandle from OF ... */
+	/* ... So use the full path instead */
+	name = of_get_property(of_chosen, "linux,stdout-path", NULL);
+	if (name == NULL) {
+		pr_debug(" no linux,stdout-path !\n");
+		return -ENODEV;
+	}
+	prom_stdout = of_find_node_by_path(name);
+	if (!prom_stdout) {
+		pr_debug(" can't find stdout package %s !\n", name);
+		return -ENODEV;
+	}
+	pr_debug("stdout is %s\n", prom_stdout->full_name);
+
+	name = of_get_property(prom_stdout, "name", NULL);
+	if (!name) {
+		pr_debug(" stdout package has no name !\n");
+		goto not_found;
+	}
+
+	if (strcmp(name, "ch-a") == 0)
+		offset = 0;
+	else if (strcmp(name, "ch-b") == 0)
+		offset = 1;
+	else
+		goto not_found;
+	of_node_put(prom_stdout);
+
+	pr_debug("Found serial console at %s%d\n", devname, offset);
+
+	return add_preferred_console(devname, offset, NULL);
+
+ not_found:
+	pr_debug("No preferred console found !\n");
+	of_node_put(prom_stdout);
+	return -ENODEV;
+}
+console_initcall(check_pmac_serial_console);
+
+#endif /* CONFIG_SERIAL_PMACZILOG_CONSOLE */
+
 /*
  * Called very early, MMU is off, device-tree isn't unflattened
  */
-- 
cgit v1.2.3


From 00df438e89a9003895948170e1abf64dd4665872 Mon Sep 17 00:00:00 2001
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Date: Mon, 28 Jul 2008 16:13:18 +1000
Subject: powerpc: Disable 64K hugetlb support when doing 64K SPU mappings

The 64K SPU local store mapping feature is incompatible with the
64K huge pages support due to the inability of some parts of
the memory management to differenciate between them while they
use a different page table format.

For now, disable 64K huge pages when CONFIG_SPU_FS_64K_LS,
in the long run, this can be fixed by making this feature use
the hugetlb page table format.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/mm/hugetlbpage.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

(limited to 'arch')

diff --git a/arch/powerpc/mm/hugetlbpage.c b/arch/powerpc/mm/hugetlbpage.c
index ed0aab0208a..f1c2d55b437 100644
--- a/arch/powerpc/mm/hugetlbpage.c
+++ b/arch/powerpc/mm/hugetlbpage.c
@@ -736,14 +736,21 @@ static int __init hugetlbpage_init(void)
 
 	if (!cpu_has_feature(CPU_FTR_16M_PAGE))
 		return -ENODEV;
+
 	/* Add supported huge page sizes.  Need to change HUGE_MAX_HSTATE
 	 * and adjust PTE_NONCACHE_NUM if the number of supported huge page
 	 * sizes changes.
 	 */
 	set_huge_psize(MMU_PAGE_16M);
-	set_huge_psize(MMU_PAGE_64K);
 	set_huge_psize(MMU_PAGE_16G);
 
+	/* Temporarily disable support for 64K huge pages when 64K SPU local
+	 * store support is enabled as the current implementation conflicts.
+	 */
+#ifndef CONFIG_SPU_FS_64K_LS
+	set_huge_psize(MMU_PAGE_64K);
+#endif
+
 	for (psize = 0; psize < MMU_PAGE_COUNT; ++psize) {
 		if (mmu_huge_psizes[psize]) {
 			huge_pgtable_cache(psize) = kmem_cache_create(
-- 
cgit v1.2.3