From 0366a1c70b89efed4f9d590216bb004a16effbed Mon Sep 17 00:00:00 2001
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Date: Mon, 23 Sep 2013 14:29:11 +1000
Subject: [PATCH 1/4] powerpc/irq: Run softirqs off the top of the irq stack

Nowadays, irq_exit() calls __do_softirq() pretty much directly
instead of calling do_softirq() which switches to the decicated
softirq stack.

This has lead to observed stack overflows on powerpc since we call
irq_enter() and irq_exit() outside of the scope that switches to
the irq stack.

This fixes it by moving the stack switching up a level, making
irq_enter() and irq_exit() run off the irq stack.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/include/asm/irq.h |   4 +-
 arch/powerpc/kernel/irq.c      | 104 ++++++++++++++++-----------------
 arch/powerpc/kernel/misc_32.S  |   9 ++-
 arch/powerpc/kernel/misc_64.S  |  10 ++--
 4 files changed, 62 insertions(+), 65 deletions(-)

diff --git a/arch/powerpc/include/asm/irq.h b/arch/powerpc/include/asm/irq.h
index 0e40843a1c6e..41f13cec8a8f 100644
--- a/arch/powerpc/include/asm/irq.h
+++ b/arch/powerpc/include/asm/irq.h
@@ -69,9 +69,9 @@ extern struct thread_info *softirq_ctx[NR_CPUS];
 
 extern void irq_ctx_init(void);
 extern void call_do_softirq(struct thread_info *tp);
-extern int call_handle_irq(int irq, void *p1,
-			   struct thread_info *tp, void *func);
+extern void call_do_irq(struct pt_regs *regs, struct thread_info *tp);
 extern void do_IRQ(struct pt_regs *regs);
+extern void __do_irq(struct pt_regs *regs);
 
 int irq_choose_cpu(const struct cpumask *mask);
 
diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index c69440cef7af..2234a1276a77 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c
@@ -441,50 +441,6 @@ void migrate_irqs(void)
 }
 #endif
 
-static inline void handle_one_irq(unsigned int irq)
-{
-	struct thread_info *curtp, *irqtp;
-	unsigned long saved_sp_limit;
-	struct irq_desc *desc;
-
-	desc = irq_to_desc(irq);
-	if (!desc)
-		return;
-
-	/* Switch to the irq stack to handle this */
-	curtp = current_thread_info();
-	irqtp = hardirq_ctx[smp_processor_id()];
-
-	if (curtp == irqtp) {
-		/* We're already on the irq stack, just handle it */
-		desc->handle_irq(irq, desc);
-		return;
-	}
-
-	saved_sp_limit = current->thread.ksp_limit;
-
-	irqtp->task = curtp->task;
-	irqtp->flags = 0;
-
-	/* Copy the softirq bits in preempt_count so that the
-	 * softirq checks work in the hardirq context. */
-	irqtp->preempt_count = (irqtp->preempt_count & ~SOFTIRQ_MASK) |
-			       (curtp->preempt_count & SOFTIRQ_MASK);
-
-	current->thread.ksp_limit = (unsigned long)irqtp +
-		_ALIGN_UP(sizeof(struct thread_info), 16);
-
-	call_handle_irq(irq, desc, irqtp, desc->handle_irq);
-	current->thread.ksp_limit = saved_sp_limit;
-	irqtp->task = NULL;
-
-	/* Set any flag that may have been set on the
-	 * alternate stack
-	 */
-	if (irqtp->flags)
-		set_bits(irqtp->flags, &curtp->flags);
-}
-
 static inline void check_stack_overflow(void)
 {
 #ifdef CONFIG_DEBUG_STACKOVERFLOW
@@ -501,9 +457,9 @@ static inline void check_stack_overflow(void)
 #endif
 }
 
-void do_IRQ(struct pt_regs *regs)
+void __do_irq(struct pt_regs *regs)
 {
-	struct pt_regs *old_regs = set_irq_regs(regs);
+	struct irq_desc *desc;
 	unsigned int irq;
 
 	irq_enter();
@@ -519,18 +475,64 @@ void do_IRQ(struct pt_regs *regs)
 	 */
 	irq = ppc_md.get_irq();
 
-	/* We can hard enable interrupts now */
+	/* We can hard enable interrupts now to allow perf interrupts */
 	may_hard_irq_enable();
 
 	/* And finally process it */
-	if (irq != NO_IRQ)
-		handle_one_irq(irq);
-	else
+	if (unlikely(irq == NO_IRQ))
 		__get_cpu_var(irq_stat).spurious_irqs++;
+	else {
+		desc = irq_to_desc(irq);
+		if (likely(desc))
+			desc->handle_irq(irq, desc);
+	}
 
 	trace_irq_exit(regs);
 
 	irq_exit();
+}
+
+void do_IRQ(struct pt_regs *regs)
+{
+	struct pt_regs *old_regs = set_irq_regs(regs);
+	struct thread_info *curtp, *irqtp;
+	unsigned long saved_sp_limit;
+
+	/* Switch to the irq stack to handle this */
+	curtp = current_thread_info();
+	irqtp = hardirq_ctx[raw_smp_processor_id()];
+
+	/* Already there ? */
+	if (unlikely(curtp == irqtp)) {
+		__do_irq(regs);
+		set_irq_regs(old_regs);
+		return;
+	}
+
+	/* Adjust the stack limit */
+	saved_sp_limit = current->thread.ksp_limit;
+	current->thread.ksp_limit = (unsigned long)irqtp +
+		_ALIGN_UP(sizeof(struct thread_info), 16);
+
+
+	/* Prepare the thread_info in the irq stack */
+	irqtp->task = curtp->task;
+	irqtp->flags = 0;
+
+	/* Copy the preempt_count so that the [soft]irq checks work. */
+	irqtp->preempt_count = curtp->preempt_count;
+
+	/* Switch stack and call */
+	call_do_irq(regs, irqtp);
+
+	/* Restore stack limit */
+	current->thread.ksp_limit = saved_sp_limit;
+	irqtp->task = NULL;
+
+	/* Copy back updates to the thread_info */
+	if (irqtp->flags)
+		set_bits(irqtp->flags, &curtp->flags);
+
 	set_irq_regs(old_regs);
 }
 
@@ -592,12 +594,10 @@ void irq_ctx_init(void)
 		memset((void *)softirq_ctx[i], 0, THREAD_SIZE);
 		tp = softirq_ctx[i];
 		tp->cpu = i;
-		tp->preempt_count = 0;
 
 		memset((void *)hardirq_ctx[i], 0, THREAD_SIZE);
 		tp = hardirq_ctx[i];
 		tp->cpu = i;
-		tp->preempt_count = HARDIRQ_OFFSET;
 	}
 }
 
diff --git a/arch/powerpc/kernel/misc_32.S b/arch/powerpc/kernel/misc_32.S
index 777d999f563b..7da3882a3622 100644
--- a/arch/powerpc/kernel/misc_32.S
+++ b/arch/powerpc/kernel/misc_32.S
@@ -47,13 +47,12 @@ _GLOBAL(call_do_softirq)
 	mtlr	r0
 	blr
 
-_GLOBAL(call_handle_irq)
+_GLOBAL(call_do_irq)
 	mflr	r0
 	stw	r0,4(r1)
-	mtctr	r6
-	stwu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r5)
-	mr	r1,r5
-	bctrl
+	stwu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r4)
+	mr	r1,r4
+	bl	__do_irq
 	lwz	r1,0(r1)
 	lwz	r0,4(r1)
 	mtlr	r0
diff --git a/arch/powerpc/kernel/misc_64.S b/arch/powerpc/kernel/misc_64.S
index 971d7e78aff2..e59caf874d05 100644
--- a/arch/powerpc/kernel/misc_64.S
+++ b/arch/powerpc/kernel/misc_64.S
@@ -40,14 +40,12 @@ _GLOBAL(call_do_softirq)
 	mtlr	r0
 	blr
 
-_GLOBAL(call_handle_irq)
-	ld	r8,0(r6)
+_GLOBAL(call_do_irq)
 	mflr	r0
 	std	r0,16(r1)
-	mtctr	r8
-	stdu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r5)
-	mr	r1,r5
-	bctrl
+	stdu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r4)
+	mr	r1,r4
+	bl	.__do_irq
 	ld	r1,0(r1)
 	ld	r0,16(r1)
 	mtlr	r0

From cbc9565ee82694dec31d8137dec975b83175183b Mon Sep 17 00:00:00 2001
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Date: Tue, 24 Sep 2013 15:17:21 +1000
Subject: [PATCH 2/4] powerpc: Remove ksp_limit on ppc64

We've been keeping that field in thread_struct for a while, it contains
the "limit" of the current stack pointer and is meant to be used for
detecting stack overflows.

It has a few problems however:

 - First, it was never actually *used* on 64-bit. Set and updated but
not actually exploited

 - When switching stack to/from irq and softirq stacks, it's update
is racy unless we hard disable interrupts, which is costly. This
is fine on 32-bit as we don't soft-disable there but not on 64-bit.

Thus rather than fixing 2 in order to implement 1 in some hypothetical
future, let's remove the code completely from 64-bit. In order to avoid
a clutter of ifdef's, we remove the updates from C code completely
during interrupt stack switching, and instead maintain it from the
asm helper that is used to do the stack switching in the first place.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/include/asm/processor.h |  4 +---
 arch/powerpc/kernel/asm-offsets.c    |  3 ++-
 arch/powerpc/kernel/irq.c            | 12 ------------
 arch/powerpc/kernel/misc_32.S        | 16 ++++++++++++++++
 arch/powerpc/kernel/process.c        |  3 ++-
 arch/powerpc/lib/sstep.c             |  3 ++-
 6 files changed, 23 insertions(+), 18 deletions(-)

diff --git a/arch/powerpc/include/asm/processor.h b/arch/powerpc/include/asm/processor.h
index e378cccfca55..ce4de5aed7b5 100644
--- a/arch/powerpc/include/asm/processor.h
+++ b/arch/powerpc/include/asm/processor.h
@@ -149,8 +149,6 @@ typedef struct {
 
 struct thread_struct {
 	unsigned long	ksp;		/* Kernel stack pointer */
-	unsigned long	ksp_limit;	/* if ksp <= ksp_limit stack overflow */
-
 #ifdef CONFIG_PPC64
 	unsigned long	ksp_vsid;
 #endif
@@ -162,6 +160,7 @@ struct thread_struct {
 #endif
 #ifdef CONFIG_PPC32
 	void		*pgdir;		/* root of page-table tree */
+	unsigned long	ksp_limit;	/* if ksp <= ksp_limit stack overflow */
 #endif
 #ifdef CONFIG_PPC_ADV_DEBUG_REGS
 	/*
@@ -321,7 +320,6 @@ struct thread_struct {
 #else
 #define INIT_THREAD  { \
 	.ksp = INIT_SP, \
-	.ksp_limit = INIT_SP_LIMIT, \
 	.regs = (struct pt_regs *)INIT_SP - 1, /* XXX bogus, I think */ \
 	.fs = KERNEL_DS, \
 	.fpr = {{0}}, \
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index d8958be5f31a..502c7a4e73f7 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -80,10 +80,11 @@ int main(void)
 	DEFINE(TASKTHREADPPR, offsetof(struct task_struct, thread.ppr));
 #else
 	DEFINE(THREAD_INFO, offsetof(struct task_struct, stack));
+	DEFINE(THREAD_INFO_GAP, _ALIGN_UP(sizeof(struct thread_info), 16));
+	DEFINE(KSP_LIMIT, offsetof(struct thread_struct, ksp_limit));
 #endif /* CONFIG_PPC64 */
 
 	DEFINE(KSP, offsetof(struct thread_struct, ksp));
-	DEFINE(KSP_LIMIT, offsetof(struct thread_struct, ksp_limit));
 	DEFINE(PT_REGS, offsetof(struct thread_struct, regs));
 #ifdef CONFIG_BOOKE
 	DEFINE(THREAD_NORMSAVES, offsetof(struct thread_struct, normsave[0]));
diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index 2234a1276a77..57d286a78f86 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c
@@ -496,7 +496,6 @@ void do_IRQ(struct pt_regs *regs)
 {
 	struct pt_regs *old_regs = set_irq_regs(regs);
 	struct thread_info *curtp, *irqtp;
-	unsigned long saved_sp_limit;
 
 	/* Switch to the irq stack to handle this */
 	curtp = current_thread_info();
@@ -509,12 +508,6 @@ void do_IRQ(struct pt_regs *regs)
 		return;
 	}
 
-	/* Adjust the stack limit */
-	saved_sp_limit = current->thread.ksp_limit;
-	current->thread.ksp_limit = (unsigned long)irqtp +
-		_ALIGN_UP(sizeof(struct thread_info), 16);
-
-
 	/* Prepare the thread_info in the irq stack */
 	irqtp->task = curtp->task;
 	irqtp->flags = 0;
@@ -526,7 +519,6 @@ void do_IRQ(struct pt_regs *regs)
 	call_do_irq(regs, irqtp);
 
 	/* Restore stack limit */
-	current->thread.ksp_limit = saved_sp_limit;
 	irqtp->task = NULL;
 
 	/* Copy back updates to the thread_info */
@@ -604,16 +596,12 @@ void irq_ctx_init(void)
 static inline void do_softirq_onstack(void)
 {
 	struct thread_info *curtp, *irqtp;
-	unsigned long saved_sp_limit = current->thread.ksp_limit;
 
 	curtp = current_thread_info();
 	irqtp = softirq_ctx[smp_processor_id()];
 	irqtp->task = curtp->task;
 	irqtp->flags = 0;
-	current->thread.ksp_limit = (unsigned long)irqtp +
-				    _ALIGN_UP(sizeof(struct thread_info), 16);
 	call_do_softirq(irqtp);
-	current->thread.ksp_limit = saved_sp_limit;
 	irqtp->task = NULL;
 
 	/* Set any flag that may have been set on the
diff --git a/arch/powerpc/kernel/misc_32.S b/arch/powerpc/kernel/misc_32.S
index 7da3882a3622..2b0ad9845363 100644
--- a/arch/powerpc/kernel/misc_32.S
+++ b/arch/powerpc/kernel/misc_32.S
@@ -36,25 +36,41 @@
 
 	.text
 
+/*
+ * We store the saved ksp_limit in the unused part
+ * of the STACK_FRAME_OVERHEAD
+ */
 _GLOBAL(call_do_softirq)
 	mflr	r0
 	stw	r0,4(r1)
+	lwz	r10,THREAD+KSP_LIMIT(r2)
+	addi	r11,r3,THREAD_INFO_GAP
 	stwu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r3)
 	mr	r1,r3
+	stw	r10,8(r1)
+	stw	r11,THREAD+KSP_LIMIT(r2)
 	bl	__do_softirq
+	lwz	r10,8(r1)
 	lwz	r1,0(r1)
 	lwz	r0,4(r1)
+	stw	r10,THREAD+KSP_LIMIT(r2)
 	mtlr	r0
 	blr
 
 _GLOBAL(call_do_irq)
 	mflr	r0
 	stw	r0,4(r1)
+	lwz	r10,THREAD+KSP_LIMIT(r2)
+	addi	r11,r3,THREAD_INFO_GAP
 	stwu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r4)
 	mr	r1,r4
+	stw	r10,8(r1)
+	stw	r11,THREAD+KSP_LIMIT(r2)
 	bl	__do_irq
+	lwz	r10,8(r1)
 	lwz	r1,0(r1)
 	lwz	r0,4(r1)
+	stw	r10,THREAD+KSP_LIMIT(r2)
 	mtlr	r0
 	blr
 
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index 6f428da53e20..96d2fdf3aa9e 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -1000,9 +1000,10 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
 	kregs = (struct pt_regs *) sp;
 	sp -= STACK_FRAME_OVERHEAD;
 	p->thread.ksp = sp;
+#ifdef CONFIG_PPC32
 	p->thread.ksp_limit = (unsigned long)task_stack_page(p) +
 				_ALIGN_UP(sizeof(struct thread_info), 16);
-
+#endif
 #ifdef CONFIG_HAVE_HW_BREAKPOINT
 	p->thread.ptrace_bps[0] = NULL;
 #endif
diff --git a/arch/powerpc/lib/sstep.c b/arch/powerpc/lib/sstep.c
index a7ee978fb860..b1faa1593c90 100644
--- a/arch/powerpc/lib/sstep.c
+++ b/arch/powerpc/lib/sstep.c
@@ -1505,6 +1505,7 @@ int __kprobes emulate_step(struct pt_regs *regs, unsigned int instr)
 		 */
 		if ((ra == 1) && !(regs->msr & MSR_PR) \
 			&& (val3 >= (regs->gpr[1] - STACK_INT_FRAME_SIZE))) {
+#ifdef CONFIG_PPC32
 			/*
 			 * Check if we will touch kernel sack overflow
 			 */
@@ -1513,7 +1514,7 @@ int __kprobes emulate_step(struct pt_regs *regs, unsigned int instr)
 				err = -EINVAL;
 				break;
 			}
-
+#endif /* CONFIG_PPC32 */
 			/*
 			 * Check if we already set since that means we'll
 			 * lose the previous value.

From 0c9fa29149d3726e14262aeb0c8461a948cc9d56 Mon Sep 17 00:00:00 2001
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Date: Tue, 24 Sep 2013 16:10:38 +1000
Subject: [PATCH 3/4] powerpc/zImage: make the "OF" wrapper support ePAPR boot

This makes the "OF" zImage wrapper (zImage.pseries, zImage.pmac,
zImage.maple) work if booted via a flat device-tree (ePAPR boot
mode), and thus potentially usable with kexec.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/boot/Makefile        |  4 ++--
 arch/powerpc/boot/epapr-wrapper.c |  9 +++++++++
 arch/powerpc/boot/epapr.c         |  4 ++--
 arch/powerpc/boot/of.c            | 16 +++++++++++++++-
 arch/powerpc/boot/wrapper         |  9 +++++----
 5 files changed, 33 insertions(+), 9 deletions(-)
 create mode 100644 arch/powerpc/boot/epapr-wrapper.c

diff --git a/arch/powerpc/boot/Makefile b/arch/powerpc/boot/Makefile
index 6a15c968d214..15ca2255f438 100644
--- a/arch/powerpc/boot/Makefile
+++ b/arch/powerpc/boot/Makefile
@@ -74,7 +74,7 @@ src-wlib-$(CONFIG_8xx) += mpc8xx.c planetcore.c
 src-wlib-$(CONFIG_PPC_82xx) += pq2.c fsl-soc.c planetcore.c
 src-wlib-$(CONFIG_EMBEDDED6xx) += mv64x60.c mv64x60_i2c.c ugecon.c
 
-src-plat-y := of.c
+src-plat-y := of.c epapr.c
 src-plat-$(CONFIG_40x) += fixed-head.S ep405.c cuboot-hotfoot.c \
 				treeboot-walnut.c cuboot-acadia.c \
 				cuboot-kilauea.c simpleboot.c \
@@ -97,7 +97,7 @@ src-plat-$(CONFIG_EMBEDDED6xx) += cuboot-pq2.c cuboot-mpc7448hpc2.c \
 					prpmc2800.c
 src-plat-$(CONFIG_AMIGAONE) += cuboot-amigaone.c
 src-plat-$(CONFIG_PPC_PS3) += ps3-head.S ps3-hvcall.S ps3.c
-src-plat-$(CONFIG_EPAPR_BOOT) += epapr.c
+src-plat-$(CONFIG_EPAPR_BOOT) += epapr.c epapr-wrapper.c
 
 src-wlib := $(sort $(src-wlib-y))
 src-plat := $(sort $(src-plat-y))
diff --git a/arch/powerpc/boot/epapr-wrapper.c b/arch/powerpc/boot/epapr-wrapper.c
new file mode 100644
index 000000000000..c10191006673
--- /dev/null
+++ b/arch/powerpc/boot/epapr-wrapper.c
@@ -0,0 +1,9 @@
+extern void epapr_platform_init(unsigned long r3, unsigned long r4,
+				unsigned long r5, unsigned long r6,
+				unsigned long r7);
+
+void platform_init(unsigned long r3, unsigned long r4, unsigned long r5,
+		   unsigned long r6, unsigned long r7)
+{
+	epapr_platform_init(r3, r4, r5, r6, r7);
+}
diff --git a/arch/powerpc/boot/epapr.c b/arch/powerpc/boot/epapr.c
index 06c1961bd124..02e91aa2194a 100644
--- a/arch/powerpc/boot/epapr.c
+++ b/arch/powerpc/boot/epapr.c
@@ -48,8 +48,8 @@ static void platform_fixups(void)
 		       fdt_addr, fdt_totalsize((void *)fdt_addr), ima_size);
 }
 
-void platform_init(unsigned long r3, unsigned long r4, unsigned long r5,
-		   unsigned long r6, unsigned long r7)
+void epapr_platform_init(unsigned long r3, unsigned long r4, unsigned long r5,
+			 unsigned long r6, unsigned long r7)
 {
 	epapr_magic = r6;
 	ima_size = r7;
diff --git a/arch/powerpc/boot/of.c b/arch/powerpc/boot/of.c
index 61d9899aa0d0..62e2f43ec1df 100644
--- a/arch/powerpc/boot/of.c
+++ b/arch/powerpc/boot/of.c
@@ -26,6 +26,9 @@
 
 static unsigned long claim_base;
 
+void epapr_platform_init(unsigned long r3, unsigned long r4, unsigned long r5,
+			 unsigned long r6, unsigned long r7);
+
 static void *of_try_claim(unsigned long size)
 {
 	unsigned long addr = 0;
@@ -61,7 +64,7 @@ static void of_image_hdr(const void *hdr)
 	}
 }
 
-void platform_init(unsigned long a1, unsigned long a2, void *promptr)
+static void of_platform_init(unsigned long a1, unsigned long a2, void *promptr)
 {
 	platform_ops.image_hdr = of_image_hdr;
 	platform_ops.malloc = of_try_claim;
@@ -81,3 +84,14 @@ void platform_init(unsigned long a1, unsigned long a2, void *promptr)
 		loader_info.initrd_size = a2;
 	}
 }
+
+void platform_init(unsigned long r3, unsigned long r4, unsigned long r5,
+		   unsigned long r6, unsigned long r7)
+{
+	/* Detect OF vs. ePAPR boot */
+	if (r5)
+		of_platform_init(r3, r4, (void *)r5);
+	else
+		epapr_platform_init(r3, r4, r5, r6, r7);
+}
+
diff --git a/arch/powerpc/boot/wrapper b/arch/powerpc/boot/wrapper
index 6761c746048d..cd7af841ba05 100755
--- a/arch/powerpc/boot/wrapper
+++ b/arch/powerpc/boot/wrapper
@@ -148,18 +148,18 @@ make_space=y
 
 case "$platform" in
 pseries)
-    platformo=$object/of.o
+    platformo="$object/of.o $object/epapr.o"
     link_address='0x4000000'
     ;;
 maple)
-    platformo=$object/of.o
+    platformo="$object/of.o $object/epapr.o"
     link_address='0x400000'
     ;;
 pmac|chrp)
-    platformo=$object/of.o
+    platformo="$object/of.o $object/epapr.o"
     ;;
 coff)
-    platformo="$object/crt0.o $object/of.o"
+    platformo="$object/crt0.o $object/of.o $object/epapr.o"
     lds=$object/zImage.coff.lds
     link_address='0x500000'
     pie=
@@ -253,6 +253,7 @@ treeboot-iss4xx-mpic)
     platformo="$object/treeboot-iss4xx.o"
     ;;
 epapr)
+    platformo="$object/epapr.o $object/epapr-wrapper.o"
     link_address='0x20000000'
     pie=-pie
     ;;

From dbe78b40118636f2d5d276144239dd4bfd5f04f9 Mon Sep 17 00:00:00 2001
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Date: Wed, 25 Sep 2013 14:02:50 +1000
Subject: [PATCH 4/4] powerpc/pseries: Do not start secondaries in Open
 Firmware

Starting secondary CPUs early on from Open Firmware and placing them
in a holding spin loop slows down the boot process significantly under
some hypervisors such as KVM.

This is also unnecessary when RTAS supports querying the CPU state

So let's not do it.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/prom_init.c      | 21 +++++++++++++++++++++
 arch/powerpc/platforms/pseries/smp.c | 26 ++++++++++++++++----------
 2 files changed, 37 insertions(+), 10 deletions(-)

diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c
index 12e656ffe60e..5fe2842e8bab 100644
--- a/arch/powerpc/kernel/prom_init.c
+++ b/arch/powerpc/kernel/prom_init.c
@@ -196,6 +196,8 @@ static int __initdata mem_reserve_cnt;
 
 static cell_t __initdata regbuf[1024];
 
+static bool rtas_has_query_cpu_stopped;
+
 
 /*
  * Error results ... some OF calls will return "-1" on error, some
@@ -1574,6 +1576,11 @@ static void __init prom_instantiate_rtas(void)
 	prom_setprop(rtas_node, "/rtas", "linux,rtas-entry",
 		     &val, sizeof(val));
 
+	/* Check if it supports "query-cpu-stopped-state" */
+	if (prom_getprop(rtas_node, "query-cpu-stopped-state",
+			 &val, sizeof(val)) != PROM_ERROR)
+		rtas_has_query_cpu_stopped = true;
+
 #if defined(CONFIG_PPC_POWERNV) && defined(__BIG_ENDIAN__)
 	/* PowerVN takeover hack */
 	prom_rtas_data = base;
@@ -1815,6 +1822,18 @@ static void __init prom_hold_cpus(void)
 		= (void *) LOW_ADDR(__secondary_hold_acknowledge);
 	unsigned long secondary_hold = LOW_ADDR(__secondary_hold);
 
+	/*
+	 * On pseries, if RTAS supports "query-cpu-stopped-state",
+	 * we skip this stage, the CPUs will be started by the
+	 * kernel using RTAS.
+	 */
+	if ((of_platform == PLATFORM_PSERIES ||
+	     of_platform == PLATFORM_PSERIES_LPAR) &&
+	    rtas_has_query_cpu_stopped) {
+		prom_printf("prom_hold_cpus: skipped\n");
+		return;
+	}
+
 	prom_debug("prom_hold_cpus: start...\n");
 	prom_debug("    1) spinloop       = 0x%x\n", (unsigned long)spinloop);
 	prom_debug("    1) *spinloop      = 0x%x\n", *spinloop);
@@ -3011,6 +3030,8 @@ unsigned long __init prom_init(unsigned long r3, unsigned long r4,
 	 * On non-powermacs, put all CPUs in spin-loops.
 	 *
 	 * PowerMacs use a different mechanism to spin CPUs
+	 *
+	 * (This must be done after instanciating RTAS)
 	 */
 	if (of_platform != PLATFORM_POWERMAC &&
 	    of_platform != PLATFORM_OPAL)
diff --git a/arch/powerpc/platforms/pseries/smp.c b/arch/powerpc/platforms/pseries/smp.c
index 1c1771a40250..24f58cb0a543 100644
--- a/arch/powerpc/platforms/pseries/smp.c
+++ b/arch/powerpc/platforms/pseries/smp.c
@@ -233,17 +233,23 @@ static void __init smp_init_pseries(void)
 
 	alloc_bootmem_cpumask_var(&of_spin_mask);
 
-	/* Mark threads which are still spinning in hold loops. */
-	if (cpu_has_feature(CPU_FTR_SMT)) {
-		for_each_present_cpu(i) { 
-			if (cpu_thread_in_core(i) == 0)
-				cpumask_set_cpu(i, of_spin_mask);
-		}
-	} else {
-		cpumask_copy(of_spin_mask, cpu_present_mask);
-	}
+	/*
+	 * Mark threads which are still spinning in hold loops
+	 *
+	 * We know prom_init will not have started them if RTAS supports
+	 * query-cpu-stopped-state.
+	 */
+	if (rtas_token("query-cpu-stopped-state") == RTAS_UNKNOWN_SERVICE) {
+		if (cpu_has_feature(CPU_FTR_SMT)) {
+			for_each_present_cpu(i) {
+				if (cpu_thread_in_core(i) == 0)
+					cpumask_set_cpu(i, of_spin_mask);
+			}
+		} else
+			cpumask_copy(of_spin_mask, cpu_present_mask);
 
-	cpumask_clear_cpu(boot_cpuid, of_spin_mask);
+		cpumask_clear_cpu(boot_cpuid, of_spin_mask);
+	}
 
 	/* Non-lpar has additional take/give timebase */
 	if (rtas_token("freeze-time-base") != RTAS_UNKNOWN_SERVICE) {