From 1028ccf560b97adbf272381a61a67e17d44d1054 Mon Sep 17 00:00:00 2001
From: Romeo Cane <romeo.cane.ext@coriant.com>
Date: Thu, 2 Oct 2014 15:41:39 +0100
Subject: [PATCH 01/22] powerpc: Fix sys_call_table declaration to enable
 syscall tracing

Declaring sys_call_table as a pointer causes the compiler to generate
the wrong lookup code in arch_syscall_addr().

     <arch_syscall_addr>:
        lis     r9,-16384
        rlwinm  r3,r3,2,0,29
  -     lwz     r11,30640(r9)
  -     lwzx    r3,r11,r3
  +     addi    r9,r9,30640
  +     lwzx    r3,r9,r3
        blr

The actual sys_call_table symbol, declared in assembler, is an
array. If we lie about that to the compiler we get the wrong code
generated, as above.

This definition seems only to be used by the syscall tracing code in
kernel/trace/trace_syscalls.c. With this patch I can successfully use
the syscall tracepoints:

  bash-3815  [002] ....   333.239082: sys_write -> 0x2
  bash-3815  [002] ....   333.239087: sys_dup2(oldfd: a, newfd: 1)
  bash-3815  [002] ....   333.239088: sys_dup2 -> 0x1
  bash-3815  [002] ....   333.239092: sys_fcntl(fd: a, cmd: 1, arg: 0)
  bash-3815  [002] ....   333.239093: sys_fcntl -> 0x1
  bash-3815  [002] ....   333.239094: sys_close(fd: a)
  bash-3815  [002] ....   333.239094: sys_close -> 0x0

Signed-off-by: Romeo Cane <romeo.cane.ext@coriant.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/include/asm/syscall.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/powerpc/include/asm/syscall.h b/arch/powerpc/include/asm/syscall.h
index b54b2add07be..528ba9d8eed5 100644
--- a/arch/powerpc/include/asm/syscall.h
+++ b/arch/powerpc/include/asm/syscall.h
@@ -17,7 +17,7 @@
 
 /* ftrace syscalls requires exporting the sys_call_table */
 #ifdef CONFIG_FTRACE_SYSCALLS
-extern const unsigned long *sys_call_table;
+extern const unsigned long sys_call_table[];
 #endif /* CONFIG_FTRACE_SYSCALLS */
 
 static inline long syscall_get_nr(struct task_struct *task,

From c675c7db629bd1abf4a9a36c0200686716eeea05 Mon Sep 17 00:00:00 2001
From: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Date: Tue, 7 Oct 2014 15:55:33 +0530
Subject: [PATCH 02/22] powerpc/book3s: Don't clear MSR_RI in hmi handler.

In HMI interrupt handler we don't touch SRR0/SRR1, instead we touch
HSRR0/HSRR1. Hence we don't need to clear MSR_RI bit.

Signed-off-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/kernel/exceptions-64s.S | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index 050f79a4a168..72e783ea0681 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -1270,11 +1270,6 @@ hmi_exception_early:
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	hmi_exception_realmode
 	/* Windup the stack. */
-	/* Clear MSR_RI before setting SRR0 and SRR1. */
-	li	r0,MSR_RI
-	mfmsr	r9			/* get MSR value */
-	andc	r9,r9,r0
-	mtmsrd	r9,1			/* Clear MSR_RI */
 	/* Move original HSRR0 and HSRR1 into the respective regs */
 	ld	r9,_MSR(r1)
 	mtspr	SPRN_HSRR1,r9

From 6507955c9781a75f1b085f0cf0a77b9df06f0197 Mon Sep 17 00:00:00 2001
From: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Date: Fri, 10 Oct 2014 21:28:26 +0530
Subject: [PATCH 03/22] powerpc/powernv: Fallback to old HMI handling behavior
 for old firmware

Recently we moved HMI handling into Linux kernel instead of taking
HMI directly in OPAL. This new change is dependent on new OPAL call
for HMI recovery which was introduced in newer firmware. While this new
change works fine with latest OPAL firmware, we broke the HMI handling
if we run newer kernel on old OPAL firmware that results in system hang.

This patch fixes this issue by falling back to old HMI behavior on older
OPAL firmware.

This patch introduces a check for opal token OPAL_HANDLE_HMI to see
if we are running on newer firmware or old firmware. On newer firmware
this check would return OPAL_TOKEN_PRESENT, otherwise we are running on
old firmware and fallback to old HMI behavior.

Old firmware: POWER8 System Firmware Release as of today <= SV810_087
Action: Let OPAL handle HMIs

Newer firmware: in development/yet to be released.
Action: Let Linux host handle HMIs.

This patch depends on opal check token patch posted at ppc-devel
https://lists.ozlabs.org/pipermail/linuxppc-dev/2014-August/120224.html

Signed-off-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
[mpe: Minor comment and printk rewording]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/platforms/powernv/opal.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/arch/powerpc/platforms/powernv/opal.c b/arch/powerpc/platforms/powernv/opal.c
index cf85ba86bf5e..34d96615db6c 100644
--- a/arch/powerpc/platforms/powernv/opal.c
+++ b/arch/powerpc/platforms/powernv/opal.c
@@ -194,6 +194,27 @@ static int __init opal_register_exception_handlers(void)
 	 * fwnmi area at 0x7000 to provide the glue space to OPAL
 	 */
 	glue = 0x7000;
+
+	/*
+	 * Check if we are running on newer firmware that exports
+	 * OPAL_HANDLE_HMI token. If yes, then don't ask OPAL to patch
+	 * the HMI interrupt and we catch it directly in Linux.
+	 *
+	 * For older firmware (i.e currently released POWER8 System Firmware
+	 * as of today <= SV810_087), we fallback to old behavior and let OPAL
+	 * patch the HMI vector and handle it inside OPAL firmware.
+	 *
+	 * For newer firmware (in development/yet to be released) we will
+	 * start catching/handling HMI directly in Linux.
+	 */
+	if (!opal_check_token(OPAL_HANDLE_HMI)) {
+		pr_info("opal: Old firmware detected, OPAL handles HMIs.\n");
+		opal_register_exception_handler(
+				OPAL_HYPERVISOR_MAINTENANCE_HANDLER,
+				0, glue);
+		glue += 128;
+	}
+
 	opal_register_exception_handler(OPAL_SOFTPATCH_HANDLER, 0, glue);
 #endif
 

From 2d15b9b479512f05680541acffd9acbbc831a47c Mon Sep 17 00:00:00 2001
From: Nishanth Aravamudan <nacc@linux.vnet.ibm.com>
Date: Thu, 9 Oct 2014 16:41:28 -0700
Subject: [PATCH 04/22] powerpc/numa: check error return from proc_create

proc_create can fail, we should check the return value and pass up the
failure.

Suggested-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/mm/numa.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
index 649666d5d1c2..51d707d85b2d 100644
--- a/arch/powerpc/mm/numa.c
+++ b/arch/powerpc/mm/numa.c
@@ -1801,7 +1801,8 @@ static const struct file_operations topology_ops = {
 static int topology_update_init(void)
 {
 	start_topology_update();
-	proc_create("powerpc/topology_updates", 0644, NULL, &topology_ops);
+	if (!proc_create("powerpc/topology_updates", 0644, NULL, &topology_ops))
+		return -ENOMEM;
 
 	return 0;
 }

From 2d73bae12b26db6eba074b70406c707961b6cda9 Mon Sep 17 00:00:00 2001
From: Nishanth Aravamudan <nacc@linux.vnet.ibm.com>
Date: Fri, 10 Oct 2014 09:04:49 -0700
Subject: [PATCH 05/22] powerpc/numa: Add ability to disable and debug topology
 updates

We have hit a few customer issues with the topology update code (VPHN
and PRRN). It would be nice to be able to debug the notifications coming
from the hypervisor in both cases to the LPAR, as well as to disable
responding to the notifications at boot-time, to narrow down the source
of the problems. Add a basic level of such functionality, similar to the
numa= command-line parameter. We already have a toggle in
/proc/powerpc/topology_updates that allows run-time enabling/disabling,
so the updates can be started at run-time if desired. But the bugs we've
run into have occured during boot or very shortly after coming to login,
and have resulted in a broken NUMA topology.

Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 Documentation/kernel-parameters.txt |  6 +++++
 arch/powerpc/mm/numa.c              | 35 ++++++++++++++++++++++++++++-
 2 files changed, 40 insertions(+), 1 deletion(-)

diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt
index 10d51c2f10d7..c536e1c74320 100644
--- a/Documentation/kernel-parameters.txt
+++ b/Documentation/kernel-parameters.txt
@@ -3370,6 +3370,12 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 			e.g. base its process migration decisions on it.
 			Default is on.
 
+	topology_updates= [KNL, PPC, NUMA]
+			Format: {off}
+			Specify if the kernel should ignore (off)
+			topology updates sent by the hypervisor to this
+			LPAR.
+
 	tp720=		[HW,PS2]
 
 	tpm_suspend_pcr=[HW,TPM]
diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
index 51d707d85b2d..177659050fa0 100644
--- a/arch/powerpc/mm/numa.c
+++ b/arch/powerpc/mm/numa.c
@@ -8,6 +8,8 @@
  * as published by the Free Software Foundation; either version
  * 2 of the License, or (at your option) any later version.
  */
+#define pr_fmt(fmt) "numa: " fmt
+
 #include <linux/threads.h>
 #include <linux/bootmem.h>
 #include <linux/init.h>
@@ -1153,6 +1155,22 @@ static int __init early_numa(char *p)
 }
 early_param("numa", early_numa);
 
+static bool topology_updates_enabled = true;
+
+static int __init early_topology_updates(char *p)
+{
+	if (!p)
+		return 0;
+
+	if (!strcmp(p, "off")) {
+		pr_info("Disabling topology updates\n");
+		topology_updates_enabled = false;
+	}
+
+	return 0;
+}
+early_param("topology_updates", early_topology_updates);
+
 #ifdef CONFIG_MEMORY_HOTPLUG
 /*
  * Find the node associated with a hot added memory section for
@@ -1539,6 +1557,9 @@ int arch_update_cpu_topology(void)
 	struct device *dev;
 	int weight, new_nid, i = 0;
 
+	if (!prrn_enabled && !vphn_enabled)
+		return 0;
+
 	weight = cpumask_weight(&cpu_associativity_changes_mask);
 	if (!weight)
 		return 0;
@@ -1592,6 +1613,15 @@ int arch_update_cpu_topology(void)
 		cpu = cpu_last_thread_sibling(cpu);
 	}
 
+	pr_debug("Topology update for the following CPUs:\n");
+	if (cpumask_weight(&updated_cpus)) {
+		for (ud = &updates[0]; ud; ud = ud->next) {
+			pr_debug("cpu %d moving from node %d "
+					  "to %d\n", ud->cpu,
+					  ud->old_nid, ud->new_nid);
+		}
+	}
+
 	/*
 	 * In cases where we have nothing to update (because the updates list
 	 * is too short or because the new topology is same as the old one),
@@ -1800,7 +1830,10 @@ static const struct file_operations topology_ops = {
 
 static int topology_update_init(void)
 {
-	start_topology_update();
+	/* Do not poll for changes if disabled at boot */
+	if (topology_updates_enabled)
+		start_topology_update();
+
 	if (!proc_create("powerpc/topology_updates", 0644, NULL, &topology_ops))
 		return -ENOMEM;
 

From bfe9a2cfe91a1c920f152ce5fd0a9ad74b3daf12 Mon Sep 17 00:00:00 2001
From: Anton Blanchard <anton@samba.org>
Date: Mon, 13 Oct 2014 19:41:38 +1100
Subject: [PATCH 06/22] powerpc: Reimplement __get_SP() as a function not a
 define

Li Zhong points out an issue with our current __get_SP()
implementation. If ftrace function tracing is enabled (ie -pg
profiling using _mcount) we spill a stack frame on 64bit all the
time.

If a function calls __get_SP() and later calls a function that is
tail call optimised, we will pop the stack frame and the value
returned by __get_SP() is no longer valid. An example from Li can
be found in save_stack_trace -> save_context_stack:

c0000000000432c0 <.save_stack_trace>:
c0000000000432c0:       mflr    r0
c0000000000432c4:       std     r0,16(r1)
c0000000000432c8:       stdu    r1,-128(r1) <-- stack frame for _mcount
c0000000000432cc:       std     r3,112(r1)
c0000000000432d0:       bl      <._mcount>
c0000000000432d4:       nop

c0000000000432d8:       mr      r4,r1 <-- __get_SP()

c0000000000432dc:       ld      r5,632(r13)
c0000000000432e0:       ld      r3,112(r1)
c0000000000432e4:       li      r6,1

c0000000000432e8:       addi    r1,r1,128 <-- pop stack frame

c0000000000432ec:       ld      r0,16(r1)
c0000000000432f0:       mtlr    r0
c0000000000432f4:       b       <.save_context_stack> <-- tail call optimized

save_context_stack ends up with a stack pointer below the current
one, and it is likely to be scribbled over.

Fix this by making __get_SP() a function which returns the
callers stack frame. Also replace inline assembly which grabs
the stack pointer in save_stack_trace and show_stack with
__get_SP().

This also fixes an issue with perf_arch_fetch_caller_regs().
It currently unwinds the stack once, which will skip a
valid stack frame on a leaf function. With the __get_SP() fixes
in this patch, we never need to unwind the stack frame to get
to the first interesting frame.

We have to export __get_SP() because perf_arch_fetch_caller_regs()
(which is used in modules) calls it from a header file.

Reported-by: Li Zhong <zhong@linux.vnet.ibm.com>
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/include/asm/perf_event.h | 2 +-
 arch/powerpc/include/asm/reg.h        | 3 +--
 arch/powerpc/kernel/misc.S            | 4 ++++
 arch/powerpc/kernel/ppc_ksyms.c       | 2 ++
 arch/powerpc/kernel/process.c         | 2 +-
 arch/powerpc/kernel/stacktrace.c      | 2 +-
 6 files changed, 10 insertions(+), 5 deletions(-)

diff --git a/arch/powerpc/include/asm/perf_event.h b/arch/powerpc/include/asm/perf_event.h
index 0bb23725b1e7..b058568a850d 100644
--- a/arch/powerpc/include/asm/perf_event.h
+++ b/arch/powerpc/include/asm/perf_event.h
@@ -34,7 +34,7 @@
 	do {							\
 		(regs)->result = 0;				\
 		(regs)->nip = __ip;				\
-		(regs)->gpr[1] = *(unsigned long *)__get_SP();	\
+		(regs)->gpr[1] = __get_SP();			\
 		asm volatile("mfmsr %0" : "=r" ((regs)->msr));	\
 	} while (0)
 #endif
diff --git a/arch/powerpc/include/asm/reg.h b/arch/powerpc/include/asm/reg.h
index fe3f9488f321..e539d7e71451 100644
--- a/arch/powerpc/include/asm/reg.h
+++ b/arch/powerpc/include/asm/reg.h
@@ -1265,8 +1265,7 @@ static inline unsigned long mfvtb (void)
 
 #define proc_trap()	asm volatile("trap")
 
-#define __get_SP()	({unsigned long sp; \
-			asm volatile("mr %0,1": "=r" (sp)); sp;})
+extern unsigned long __get_SP(void);
 
 extern unsigned long scom970_read(unsigned int address);
 extern void scom970_write(unsigned int address, unsigned long value);
diff --git a/arch/powerpc/kernel/misc.S b/arch/powerpc/kernel/misc.S
index 7ce26d45777e..120deb713bc8 100644
--- a/arch/powerpc/kernel/misc.S
+++ b/arch/powerpc/kernel/misc.S
@@ -114,3 +114,7 @@ _GLOBAL(longjmp)
 	mtlr	r0
 	mr	r3,r4
 	blr
+
+_GLOBAL(__get_SP)
+	PPC_LL	r3,0(r1)
+	blr
diff --git a/arch/powerpc/kernel/ppc_ksyms.c b/arch/powerpc/kernel/ppc_ksyms.c
index c4dfff6c2719..9d84efbd7b7a 100644
--- a/arch/powerpc/kernel/ppc_ksyms.c
+++ b/arch/powerpc/kernel/ppc_ksyms.c
@@ -41,3 +41,5 @@ EXPORT_SYMBOL(giveup_spe);
 #ifdef CONFIG_EPAPR_PARAVIRT
 EXPORT_SYMBOL(epapr_hypercall_start);
 #endif
+
+EXPORT_SYMBOL(__get_SP);
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index aa1df89c8b2a..3cc643988101 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -1545,7 +1545,7 @@ void show_stack(struct task_struct *tsk, unsigned long *stack)
 		tsk = current;
 	if (sp == 0) {
 		if (tsk == current)
-			asm("mr %0,1" : "=r" (sp));
+			sp = __get_SP();
 		else
 			sp = tsk->thread.ksp;
 	}
diff --git a/arch/powerpc/kernel/stacktrace.c b/arch/powerpc/kernel/stacktrace.c
index 3d30ef1038e5..7f65baec29f6 100644
--- a/arch/powerpc/kernel/stacktrace.c
+++ b/arch/powerpc/kernel/stacktrace.c
@@ -50,7 +50,7 @@ void save_stack_trace(struct stack_trace *trace)
 {
 	unsigned long sp;
 
-	asm("mr %0,1" : "=r" (sp));
+	sp = __get_SP();
 
 	save_context_stack(trace, sp, current, 1);
 }

From acf620ecf56cfc4edaffaf158250e128539cdd26 Mon Sep 17 00:00:00 2001
From: Anton Blanchard <anton@samba.org>
Date: Mon, 13 Oct 2014 19:41:39 +1100
Subject: [PATCH 07/22] powerpc: Rename __get_SP() to current_stack_pointer()

Michael points out that __get_SP() is a pretty horrible
function name. Let's give it a better name.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/include/asm/perf_event.h | 2 +-
 arch/powerpc/include/asm/reg.h        | 2 +-
 arch/powerpc/kernel/irq.c             | 2 +-
 arch/powerpc/kernel/misc.S            | 2 +-
 arch/powerpc/kernel/ppc_ksyms.c       | 2 +-
 arch/powerpc/kernel/process.c         | 2 +-
 arch/powerpc/kernel/stacktrace.c      | 2 +-
 7 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/arch/powerpc/include/asm/perf_event.h b/arch/powerpc/include/asm/perf_event.h
index b058568a850d..8bf1b6351716 100644
--- a/arch/powerpc/include/asm/perf_event.h
+++ b/arch/powerpc/include/asm/perf_event.h
@@ -34,7 +34,7 @@
 	do {							\
 		(regs)->result = 0;				\
 		(regs)->nip = __ip;				\
-		(regs)->gpr[1] = __get_SP();			\
+		(regs)->gpr[1] = current_stack_pointer();	\
 		asm volatile("mfmsr %0" : "=r" ((regs)->msr));	\
 	} while (0)
 #endif
diff --git a/arch/powerpc/include/asm/reg.h b/arch/powerpc/include/asm/reg.h
index e539d7e71451..c998279bd85b 100644
--- a/arch/powerpc/include/asm/reg.h
+++ b/arch/powerpc/include/asm/reg.h
@@ -1265,7 +1265,7 @@ static inline unsigned long mfvtb (void)
 
 #define proc_trap()	asm volatile("trap")
 
-extern unsigned long __get_SP(void);
+extern unsigned long current_stack_pointer(void);
 
 extern unsigned long scom970_read(unsigned int address);
 extern void scom970_write(unsigned int address, unsigned long value);
diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index 8eb857f216c1..c14383575fe8 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c
@@ -466,7 +466,7 @@ static inline void check_stack_overflow(void)
 #ifdef CONFIG_DEBUG_STACKOVERFLOW
 	long sp;
 
-	sp = __get_SP() & (THREAD_SIZE-1);
+	sp = current_stack_pointer() & (THREAD_SIZE-1);
 
 	/* check for stack overflow: is there less than 2KB free? */
 	if (unlikely(sp < (sizeof(struct thread_info) + 2048))) {
diff --git a/arch/powerpc/kernel/misc.S b/arch/powerpc/kernel/misc.S
index 120deb713bc8..0d432194c018 100644
--- a/arch/powerpc/kernel/misc.S
+++ b/arch/powerpc/kernel/misc.S
@@ -115,6 +115,6 @@ _GLOBAL(longjmp)
 	mr	r3,r4
 	blr
 
-_GLOBAL(__get_SP)
+_GLOBAL(current_stack_pointer)
 	PPC_LL	r3,0(r1)
 	blr
diff --git a/arch/powerpc/kernel/ppc_ksyms.c b/arch/powerpc/kernel/ppc_ksyms.c
index 9d84efbd7b7a..202963ee013a 100644
--- a/arch/powerpc/kernel/ppc_ksyms.c
+++ b/arch/powerpc/kernel/ppc_ksyms.c
@@ -42,4 +42,4 @@ EXPORT_SYMBOL(giveup_spe);
 EXPORT_SYMBOL(epapr_hypercall_start);
 #endif
 
-EXPORT_SYMBOL(__get_SP);
+EXPORT_SYMBOL(current_stack_pointer);
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index 3cc643988101..923cd2daba89 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -1545,7 +1545,7 @@ void show_stack(struct task_struct *tsk, unsigned long *stack)
 		tsk = current;
 	if (sp == 0) {
 		if (tsk == current)
-			sp = __get_SP();
+			sp = current_stack_pointer();
 		else
 			sp = tsk->thread.ksp;
 	}
diff --git a/arch/powerpc/kernel/stacktrace.c b/arch/powerpc/kernel/stacktrace.c
index 7f65baec29f6..ea43a347a104 100644
--- a/arch/powerpc/kernel/stacktrace.c
+++ b/arch/powerpc/kernel/stacktrace.c
@@ -50,7 +50,7 @@ void save_stack_trace(struct stack_trace *trace)
 {
 	unsigned long sp;
 
-	sp = __get_SP();
+	sp = current_stack_pointer();
 
 	save_context_stack(trace, sp, current, 1);
 }

From 4ff52b4dedcd4226cf1f2817eb14e8d733790eae Mon Sep 17 00:00:00 2001
From: Anton Blanchard <anton@samba.org>
Date: Mon, 13 Oct 2014 19:41:40 +1100
Subject: [PATCH 08/22] powerpc/pseries: Use dump_stack instead of show_stack

We can use the simpler dump_stack() instead of
show_stack(current, __get_SP())

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/platforms/pseries/iommu.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/arch/powerpc/platforms/pseries/iommu.c b/arch/powerpc/platforms/pseries/iommu.c
index de1ec54a2a57..e32e00976a94 100644
--- a/arch/powerpc/platforms/pseries/iommu.c
+++ b/arch/powerpc/platforms/pseries/iommu.c
@@ -30,7 +30,6 @@
 #include <linux/mm.h>
 #include <linux/memblock.h>
 #include <linux/spinlock.h>
-#include <linux/sched.h>	/* for show_stack */
 #include <linux/string.h>
 #include <linux/pci.h>
 #include <linux/dma-mapping.h>
@@ -168,7 +167,7 @@ static int tce_build_pSeriesLP(struct iommu_table *tbl, long tcenum,
 			printk("\tindex   = 0x%llx\n", (u64)tbl->it_index);
 			printk("\ttcenum  = 0x%llx\n", (u64)tcenum);
 			printk("\ttce val = 0x%llx\n", tce );
-			show_stack(current, (unsigned long *)__get_SP());
+			dump_stack();
 		}
 
 		tcenum++;
@@ -257,7 +256,7 @@ static int tce_buildmulti_pSeriesLP(struct iommu_table *tbl, long tcenum,
 		printk("\tindex   = 0x%llx\n", (u64)tbl->it_index);
 		printk("\tnpages  = 0x%llx\n", (u64)npages);
 		printk("\ttce[0] val = 0x%llx\n", tcep[0]);
-		show_stack(current, (unsigned long *)__get_SP());
+		dump_stack();
 	}
 	return ret;
 }
@@ -273,7 +272,7 @@ static void tce_free_pSeriesLP(struct iommu_table *tbl, long tcenum, long npages
 			printk("tce_free_pSeriesLP: plpar_tce_put failed. rc=%lld\n", rc);
 			printk("\tindex   = 0x%llx\n", (u64)tbl->it_index);
 			printk("\ttcenum  = 0x%llx\n", (u64)tcenum);
-			show_stack(current, (unsigned long *)__get_SP());
+			dump_stack();
 		}
 
 		tcenum++;
@@ -292,7 +291,7 @@ static void tce_freemulti_pSeriesLP(struct iommu_table *tbl, long tcenum, long n
 		printk("\trc      = %lld\n", rc);
 		printk("\tindex   = 0x%llx\n", (u64)tbl->it_index);
 		printk("\tnpages  = 0x%llx\n", (u64)npages);
-		show_stack(current, (unsigned long *)__get_SP());
+		dump_stack();
 	}
 }
 
@@ -307,7 +306,7 @@ static unsigned long tce_get_pSeriesLP(struct iommu_table *tbl, long tcenum)
 		printk("tce_get_pSeriesLP: plpar_tce_get failed. rc=%lld\n", rc);
 		printk("\tindex   = 0x%llx\n", (u64)tbl->it_index);
 		printk("\ttcenum  = 0x%llx\n", (u64)tcenum);
-		show_stack(current, (unsigned long *)__get_SP());
+		dump_stack();
 	}
 
 	return tce_ret;

From d6f1e7abdb95a7ea031e7604829e4b5514d7e2c1 Mon Sep 17 00:00:00 2001
From: Bharata B Rao <bharata@linux.vnet.ibm.com>
Date: Tue, 16 Sep 2014 15:15:45 -0500
Subject: [PATCH 09/22] powerpc/pseries: Make CPU hotplug path endian safe

- ibm,rtas-configure-connector should treat the RTAS data as big endian.
- Treat ibm,ppc-interrupt-server#s as big-endian when setting
  smp_processor_id during hotplug.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Signed-off-by: Thomas Falcon <tlfalcon@linux.vnet.ibm.com>
Acked-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/platforms/pseries/dlpar.c       | 22 ++++++++++----------
 arch/powerpc/platforms/pseries/hotplug-cpu.c |  4 ++--
 arch/powerpc/platforms/pseries/pseries.h     |  3 ++-
 3 files changed, 15 insertions(+), 14 deletions(-)

diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c
index fdf01b660d59..6ad83bd11fe2 100644
--- a/arch/powerpc/platforms/pseries/dlpar.c
+++ b/arch/powerpc/platforms/pseries/dlpar.c
@@ -25,11 +25,11 @@
 #include <asm/rtas.h>
 
 struct cc_workarea {
-	u32	drc_index;
-	u32	zero;
-	u32	name_offset;
-	u32	prop_length;
-	u32	prop_offset;
+	__be32	drc_index;
+	__be32	zero;
+	__be32	name_offset;
+	__be32	prop_length;
+	__be32	prop_offset;
 };
 
 void dlpar_free_cc_property(struct property *prop)
@@ -49,11 +49,11 @@ static struct property *dlpar_parse_cc_property(struct cc_workarea *ccwa)
 	if (!prop)
 		return NULL;
 
-	name = (char *)ccwa + ccwa->name_offset;
+	name = (char *)ccwa + be32_to_cpu(ccwa->name_offset);
 	prop->name = kstrdup(name, GFP_KERNEL);
 
-	prop->length = ccwa->prop_length;
-	value = (char *)ccwa + ccwa->prop_offset;
+	prop->length = be32_to_cpu(ccwa->prop_length);
+	value = (char *)ccwa + be32_to_cpu(ccwa->prop_offset);
 	prop->value = kmemdup(value, prop->length, GFP_KERNEL);
 	if (!prop->value) {
 		dlpar_free_cc_property(prop);
@@ -79,7 +79,7 @@ static struct device_node *dlpar_parse_cc_node(struct cc_workarea *ccwa,
 	if (!dn)
 		return NULL;
 
-	name = (char *)ccwa + ccwa->name_offset;
+	name = (char *)ccwa + be32_to_cpu(ccwa->name_offset);
 	dn->full_name = kasprintf(GFP_KERNEL, "%s/%s", path, name);
 	if (!dn->full_name) {
 		kfree(dn);
@@ -126,7 +126,7 @@ void dlpar_free_cc_nodes(struct device_node *dn)
 #define CALL_AGAIN	-2
 #define ERR_CFG_USE     -9003
 
-struct device_node *dlpar_configure_connector(u32 drc_index,
+struct device_node *dlpar_configure_connector(__be32 drc_index,
 					      struct device_node *parent)
 {
 	struct device_node *dn;
@@ -414,7 +414,7 @@ static ssize_t dlpar_cpu_probe(const char *buf, size_t count)
 	if (!parent)
 		return -ENODEV;
 
-	dn = dlpar_configure_connector(drc_index, parent);
+	dn = dlpar_configure_connector(cpu_to_be32(drc_index), parent);
 	if (!dn)
 		return -EINVAL;
 
diff --git a/arch/powerpc/platforms/pseries/hotplug-cpu.c b/arch/powerpc/platforms/pseries/hotplug-cpu.c
index b174fa751d26..5c375f93c669 100644
--- a/arch/powerpc/platforms/pseries/hotplug-cpu.c
+++ b/arch/powerpc/platforms/pseries/hotplug-cpu.c
@@ -247,7 +247,7 @@ static int pseries_add_processor(struct device_node *np)
 	unsigned int cpu;
 	cpumask_var_t candidate_mask, tmp;
 	int err = -ENOSPC, len, nthreads, i;
-	const u32 *intserv;
+	const __be32 *intserv;
 
 	intserv = of_get_property(np, "ibm,ppc-interrupt-server#s", &len);
 	if (!intserv)
@@ -293,7 +293,7 @@ static int pseries_add_processor(struct device_node *np)
 	for_each_cpu(cpu, tmp) {
 		BUG_ON(cpu_present(cpu));
 		set_cpu_present(cpu, true);
-		set_hard_smp_processor_id(cpu, *intserv++);
+		set_hard_smp_processor_id(cpu, be32_to_cpu(*intserv++));
 	}
 	err = 0;
 out_unlock:
diff --git a/arch/powerpc/platforms/pseries/pseries.h b/arch/powerpc/platforms/pseries/pseries.h
index 361add62abf1..1796c5438cc6 100644
--- a/arch/powerpc/platforms/pseries/pseries.h
+++ b/arch/powerpc/platforms/pseries/pseries.h
@@ -56,7 +56,8 @@ extern void hvc_vio_init_early(void);
 /* Dynamic logical Partitioning/Mobility */
 extern void dlpar_free_cc_nodes(struct device_node *);
 extern void dlpar_free_cc_property(struct property *);
-extern struct device_node *dlpar_configure_connector(u32, struct device_node *);
+extern struct device_node *dlpar_configure_connector(__be32,
+						struct device_node *);
 extern int dlpar_attach_node(struct device_node *);
 extern int dlpar_detach_node(struct device_node *);
 

From 8315070c07e7ef5f58ce9e317dc91fd727ecd419 Mon Sep 17 00:00:00 2001
From: Gavin Shan <gwshan@linux.vnet.ibm.com>
Date: Wed, 1 Oct 2014 17:07:49 +1000
Subject: [PATCH 10/22] powerpc/eeh: Fix condition for isolated state

Function eeh_pe_state_mark() could possibly have combination of
multiple EEH PE state as its argument. The patch fixes the condition
used to check if EEH_PE_ISOLATED is included.

Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/kernel/eeh_pe.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/powerpc/kernel/eeh_pe.c b/arch/powerpc/kernel/eeh_pe.c
index 53dd0915e690..37f21284809c 100644
--- a/arch/powerpc/kernel/eeh_pe.c
+++ b/arch/powerpc/kernel/eeh_pe.c
@@ -525,7 +525,7 @@ static void *__eeh_pe_state_mark(void *data, void *flag)
 	pe->state |= state;
 
 	/* Offline PCI devices if applicable */
-	if (state != EEH_PE_ISOLATED)
+	if (!(state & EEH_PE_ISOLATED))
 		return NULL;
 
 	eeh_pe_for_each_dev(pe, edev, tmp) {

From 8a6b3710ccc33da1fd5c85144ad3db01c4457552 Mon Sep 17 00:00:00 2001
From: Gavin Shan <gwshan@linux.vnet.ibm.com>
Date: Wed, 1 Oct 2014 17:07:50 +1000
Subject: [PATCH 11/22] powerpc/eeh: Rename flag EEH_PE_RESET to
 EEH_PE_CFG_BLOCKED

The flag EEH_PE_RESET indicates blocking config space of the PE
during reset time. We potentially need block PE's config space
other than reset time. So it's reasonable to replace it with
EEH_PE_CFG_BLOCKED to indicate its usage.

There are no substantial code or logic changes in this patch.

Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/include/asm/eeh.h            |  2 +-
 arch/powerpc/kernel/eeh.c                 | 12 ++++++------
 arch/powerpc/kernel/eeh_driver.c          | 12 ++++++------
 arch/powerpc/kernel/rtas_pci.c            |  4 ++--
 arch/powerpc/platforms/powernv/eeh-ioda.c |  2 +-
 arch/powerpc/platforms/powernv/pci.c      |  2 +-
 6 files changed, 17 insertions(+), 17 deletions(-)

diff --git a/arch/powerpc/include/asm/eeh.h b/arch/powerpc/include/asm/eeh.h
index 3b260efbfbf9..6a2ad90e6d8c 100644
--- a/arch/powerpc/include/asm/eeh.h
+++ b/arch/powerpc/include/asm/eeh.h
@@ -71,7 +71,7 @@ struct device_node;
 
 #define EEH_PE_ISOLATED		(1 << 0)	/* Isolated PE		*/
 #define EEH_PE_RECOVERING	(1 << 1)	/* Recovering PE	*/
-#define EEH_PE_RESET		(1 << 2)	/* PE reset in progress	*/
+#define EEH_PE_CFG_BLOCKED	(1 << 2)	/* Block config access	*/
 
 #define EEH_PE_KEEP		(1 << 8)	/* Keep PE on hotplug	*/
 
diff --git a/arch/powerpc/kernel/eeh.c b/arch/powerpc/kernel/eeh.c
index d543e4179c18..4d83f50cc614 100644
--- a/arch/powerpc/kernel/eeh.c
+++ b/arch/powerpc/kernel/eeh.c
@@ -673,18 +673,18 @@ int pcibios_set_pcie_reset_state(struct pci_dev *dev, enum pcie_reset_state stat
 	switch (state) {
 	case pcie_deassert_reset:
 		eeh_ops->reset(pe, EEH_RESET_DEACTIVATE);
-		eeh_pe_state_clear(pe, EEH_PE_RESET);
+		eeh_pe_state_clear(pe, EEH_PE_CFG_BLOCKED);
 		break;
 	case pcie_hot_reset:
-		eeh_pe_state_mark(pe, EEH_PE_RESET);
+		eeh_pe_state_mark(pe, EEH_PE_CFG_BLOCKED);
 		eeh_ops->reset(pe, EEH_RESET_HOT);
 		break;
 	case pcie_warm_reset:
-		eeh_pe_state_mark(pe, EEH_PE_RESET);
+		eeh_pe_state_mark(pe, EEH_PE_CFG_BLOCKED);
 		eeh_ops->reset(pe, EEH_RESET_FUNDAMENTAL);
 		break;
 	default:
-		eeh_pe_state_clear(pe, EEH_PE_RESET);
+		eeh_pe_state_clear(pe, EEH_PE_CFG_BLOCKED);
 		return -EINVAL;
 	};
 
@@ -1523,7 +1523,7 @@ int eeh_pe_reset(struct eeh_pe *pe, int option)
 	switch (option) {
 	case EEH_RESET_DEACTIVATE:
 		ret = eeh_ops->reset(pe, option);
-		eeh_pe_state_clear(pe, EEH_PE_RESET);
+		eeh_pe_state_clear(pe, EEH_PE_CFG_BLOCKED);
 		if (ret)
 			break;
 
@@ -1538,7 +1538,7 @@ int eeh_pe_reset(struct eeh_pe *pe, int option)
 		 */
 		eeh_ops->set_option(pe, EEH_OPT_FREEZE_PE);
 
-		eeh_pe_state_mark(pe, EEH_PE_RESET);
+		eeh_pe_state_mark(pe, EEH_PE_CFG_BLOCKED);
 		ret = eeh_ops->reset(pe, option);
 		break;
 	default:
diff --git a/arch/powerpc/kernel/eeh_driver.c b/arch/powerpc/kernel/eeh_driver.c
index 3fd514f8e4b2..6535936bdf27 100644
--- a/arch/powerpc/kernel/eeh_driver.c
+++ b/arch/powerpc/kernel/eeh_driver.c
@@ -528,13 +528,13 @@ int eeh_pe_reset_and_recover(struct eeh_pe *pe)
 	eeh_pe_dev_traverse(pe, eeh_report_error, &result);
 
 	/* Issue reset */
-	eeh_pe_state_mark(pe, EEH_PE_RESET);
+	eeh_pe_state_mark(pe, EEH_PE_CFG_BLOCKED);
 	ret = eeh_reset_pe(pe);
 	if (ret) {
-		eeh_pe_state_clear(pe, EEH_PE_RECOVERING | EEH_PE_RESET);
+		eeh_pe_state_clear(pe, EEH_PE_RECOVERING | EEH_PE_CFG_BLOCKED);
 		return ret;
 	}
-	eeh_pe_state_clear(pe, EEH_PE_RESET);
+	eeh_pe_state_clear(pe, EEH_PE_CFG_BLOCKED);
 
 	/* Unfreeze the PE */
 	ret = eeh_clear_pe_frozen_state(pe, true);
@@ -601,10 +601,10 @@ static int eeh_reset_device(struct eeh_pe *pe, struct pci_bus *bus)
 	 * config accesses. So we prefer to block them. However, controlled
 	 * PCI config accesses initiated from EEH itself are allowed.
 	 */
-	eeh_pe_state_mark(pe, EEH_PE_RESET);
+	eeh_pe_state_mark(pe, EEH_PE_CFG_BLOCKED);
 	rc = eeh_reset_pe(pe);
 	if (rc) {
-		eeh_pe_state_clear(pe, EEH_PE_RESET);
+		eeh_pe_state_clear(pe, EEH_PE_CFG_BLOCKED);
 		return rc;
 	}
 
@@ -613,7 +613,7 @@ static int eeh_reset_device(struct eeh_pe *pe, struct pci_bus *bus)
 	/* Restore PE */
 	eeh_ops->configure_bridge(pe);
 	eeh_pe_restore_bars(pe);
-	eeh_pe_state_clear(pe, EEH_PE_RESET);
+	eeh_pe_state_clear(pe, EEH_PE_CFG_BLOCKED);
 
 	/* Clear frozen state */
 	rc = eeh_clear_pe_frozen_state(pe, false);
diff --git a/arch/powerpc/kernel/rtas_pci.c b/arch/powerpc/kernel/rtas_pci.c
index c168337aef9d..ce7c8b6a8602 100644
--- a/arch/powerpc/kernel/rtas_pci.c
+++ b/arch/powerpc/kernel/rtas_pci.c
@@ -111,7 +111,7 @@ static int rtas_pci_read_config(struct pci_bus *bus,
 		return PCIBIOS_DEVICE_NOT_FOUND;
 #ifdef CONFIG_EEH
 	edev = of_node_to_eeh_dev(dn);
-	if (edev && edev->pe && edev->pe->state & EEH_PE_RESET)
+	if (edev && edev->pe && edev->pe->state & EEH_PE_CFG_BLOCKED)
 		return PCIBIOS_DEVICE_NOT_FOUND;
 #endif
 
@@ -175,7 +175,7 @@ static int rtas_pci_write_config(struct pci_bus *bus,
 		return PCIBIOS_DEVICE_NOT_FOUND;
 #ifdef CONFIG_EEH
 	edev = of_node_to_eeh_dev(dn);
-	if (edev && edev->pe && (edev->pe->state & EEH_PE_RESET))
+	if (edev && edev->pe && (edev->pe->state & EEH_PE_CFG_BLOCKED))
 		return PCIBIOS_DEVICE_NOT_FOUND;
 #endif
 	ret = rtas_write_config(pdn, where, size, val);
diff --git a/arch/powerpc/platforms/powernv/eeh-ioda.c b/arch/powerpc/platforms/powernv/eeh-ioda.c
index 426814a2ede3..eba9cb10619c 100644
--- a/arch/powerpc/platforms/powernv/eeh-ioda.c
+++ b/arch/powerpc/platforms/powernv/eeh-ioda.c
@@ -373,7 +373,7 @@ static int ioda_eeh_get_pe_state(struct eeh_pe *pe)
 	 * moving forward, we have to return operational
 	 * state during PE reset.
 	 */
-	if (pe->state & EEH_PE_RESET) {
+	if (pe->state & EEH_PE_CFG_BLOCKED) {
 		result = (EEH_STATE_MMIO_ACTIVE  |
 			  EEH_STATE_DMA_ACTIVE   |
 			  EEH_STATE_MMIO_ENABLED |
diff --git a/arch/powerpc/platforms/powernv/pci.c b/arch/powerpc/platforms/powernv/pci.c
index e9f509bbc078..b1b7ac2d494f 100644
--- a/arch/powerpc/platforms/powernv/pci.c
+++ b/arch/powerpc/platforms/powernv/pci.c
@@ -513,7 +513,7 @@ static bool pnv_pci_cfg_check(struct pci_controller *hose,
 	edev = of_node_to_eeh_dev(dn);
 	if (edev) {
 		if (edev->pe &&
-		    (edev->pe->state & EEH_PE_RESET))
+		    (edev->pe->state & EEH_PE_CFG_BLOCKED))
 			return false;
 
 		if (edev->mode & EEH_DEV_REMOVED)

From d2cfbcd7c8136bc4d944eb64f7f0479dd2b129b8 Mon Sep 17 00:00:00 2001
From: Gavin Shan <gwshan@linux.vnet.ibm.com>
Date: Wed, 1 Oct 2014 17:07:51 +1000
Subject: [PATCH 12/22] powerpc/powernv: Drop config requests in EEH accessors

It's bad idea to access the PCI config registers of the adapters,
which is experiencing reset. It leads to recursive EEH error without
exception. The patch drops PCI config requests in EEH accessors if
the PE has been marked to accept PCI config requests, for example
during PE reseet time.

Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/platforms/powernv/eeh-powernv.c | 37 ++++++++++++++++++--
 1 file changed, 35 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/platforms/powernv/eeh-powernv.c b/arch/powerpc/platforms/powernv/eeh-powernv.c
index 3e89cbf55885..04e42f78a37a 100644
--- a/arch/powerpc/platforms/powernv/eeh-powernv.c
+++ b/arch/powerpc/platforms/powernv/eeh-powernv.c
@@ -383,6 +383,39 @@ static int powernv_eeh_err_inject(struct eeh_pe *pe, int type, int func,
 	return ret;
 }
 
+static inline bool powernv_eeh_cfg_blocked(struct device_node *dn)
+{
+	struct eeh_dev *edev = of_node_to_eeh_dev(dn);
+
+	if (!edev || !edev->pe)
+		return false;
+
+	if (edev->pe->state & EEH_PE_CFG_BLOCKED)
+		return true;
+
+	return false;
+}
+
+static int powernv_eeh_read_config(struct device_node *dn,
+				   int where, int size, u32 *val)
+{
+	if (powernv_eeh_cfg_blocked(dn)) {
+		*val = 0xFFFFFFFF;
+		return PCIBIOS_SET_FAILED;
+	}
+
+	return pnv_pci_cfg_read(dn, where, size, val);
+}
+
+static int powernv_eeh_write_config(struct device_node *dn,
+				    int where, int size, u32 val)
+{
+	if (powernv_eeh_cfg_blocked(dn))
+		return PCIBIOS_SET_FAILED;
+
+	return pnv_pci_cfg_write(dn, where, size, val);
+}
+
 /**
  * powernv_eeh_next_error - Retrieve next EEH error to handle
  * @pe: Affected PE
@@ -440,8 +473,8 @@ static struct eeh_ops powernv_eeh_ops = {
 	.get_log                = powernv_eeh_get_log,
 	.configure_bridge       = powernv_eeh_configure_bridge,
 	.err_inject		= powernv_eeh_err_inject,
-	.read_config            = pnv_pci_cfg_read,
-	.write_config           = pnv_pci_cfg_write,
+	.read_config            = powernv_eeh_read_config,
+	.write_config           = powernv_eeh_write_config,
 	.next_error		= powernv_eeh_next_error,
 	.restore_config		= powernv_eeh_restore_config
 };

From 3409eb4e69e1150202bc4ec61801115da32aa380 Mon Sep 17 00:00:00 2001
From: Gavin Shan <gwshan@linux.vnet.ibm.com>
Date: Wed, 1 Oct 2014 17:07:52 +1000
Subject: [PATCH 13/22] powerpc/pseries: Drop config requests in EEH accessors

The pSeires EEH config accessors rely on rtas_{read, write}_config()
and the condition to check if the PE's config space is blocked
should be moved to those 2 functions so that config requests from
kernel, userland, EEH core can be dropped to avoid recursive EEH error
if necessary.

Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/kernel/rtas_pci.c | 30 +++++++++++-------------------
 1 file changed, 11 insertions(+), 19 deletions(-)

diff --git a/arch/powerpc/kernel/rtas_pci.c b/arch/powerpc/kernel/rtas_pci.c
index ce7c8b6a8602..7c55b86206b3 100644
--- a/arch/powerpc/kernel/rtas_pci.c
+++ b/arch/powerpc/kernel/rtas_pci.c
@@ -66,6 +66,11 @@ int rtas_read_config(struct pci_dn *pdn, int where, int size, u32 *val)
 		return PCIBIOS_DEVICE_NOT_FOUND;
 	if (!config_access_valid(pdn, where))
 		return PCIBIOS_BAD_REGISTER_NUMBER;
+#ifdef CONFIG_EEH
+	if (pdn->edev && pdn->edev->pe &&
+	    (pdn->edev->pe->state & EEH_PE_CFG_BLOCKED))
+		return PCIBIOS_SET_FAILED;
+#endif
 
 	addr = rtas_config_addr(pdn->busno, pdn->devfn, where);
 	buid = pdn->phb->buid;
@@ -90,9 +95,6 @@ static int rtas_pci_read_config(struct pci_bus *bus,
 	struct device_node *busdn, *dn;
 	struct pci_dn *pdn;
 	bool found = false;
-#ifdef CONFIG_EEH
-	struct eeh_dev *edev;
-#endif
 	int ret;
 
 	/* Search only direct children of the bus */
@@ -109,11 +111,6 @@ static int rtas_pci_read_config(struct pci_bus *bus,
 
 	if (!found)
 		return PCIBIOS_DEVICE_NOT_FOUND;
-#ifdef CONFIG_EEH
-	edev = of_node_to_eeh_dev(dn);
-	if (edev && edev->pe && edev->pe->state & EEH_PE_CFG_BLOCKED)
-		return PCIBIOS_DEVICE_NOT_FOUND;
-#endif
 
 	ret = rtas_read_config(pdn, where, size, val);
 	if (*val == EEH_IO_ERROR_VALUE(size) &&
@@ -132,6 +129,11 @@ int rtas_write_config(struct pci_dn *pdn, int where, int size, u32 val)
 		return PCIBIOS_DEVICE_NOT_FOUND;
 	if (!config_access_valid(pdn, where))
 		return PCIBIOS_BAD_REGISTER_NUMBER;
+#ifdef CONFIG_EEH
+	if (pdn->edev && pdn->edev->pe &&
+	    (pdn->edev->pe->state & EEH_PE_CFG_BLOCKED))
+		return PCIBIOS_SET_FAILED;
+#endif
 
 	addr = rtas_config_addr(pdn->busno, pdn->devfn, where);
 	buid = pdn->phb->buid;
@@ -155,10 +157,6 @@ static int rtas_pci_write_config(struct pci_bus *bus,
 	struct device_node *busdn, *dn;
 	struct pci_dn *pdn;
 	bool found = false;
-#ifdef CONFIG_EEH
-	struct eeh_dev *edev;
-#endif
-	int ret;
 
 	/* Search only direct children of the bus */
 	busdn = pci_bus_to_OF_node(bus);
@@ -173,14 +171,8 @@ static int rtas_pci_write_config(struct pci_bus *bus,
 
 	if (!found)
 		return PCIBIOS_DEVICE_NOT_FOUND;
-#ifdef CONFIG_EEH
-	edev = of_node_to_eeh_dev(dn);
-	if (edev && edev->pe && (edev->pe->state & EEH_PE_CFG_BLOCKED))
-		return PCIBIOS_DEVICE_NOT_FOUND;
-#endif
-	ret = rtas_write_config(pdn, where, size, val);
 
-	return ret;
+	return rtas_write_config(pdn, where, size, val);
 }
 
 static struct pci_ops rtas_pci_ops = {

From b6541db1395298b326ae1bf59fae6fbb2c6e8f77 Mon Sep 17 00:00:00 2001
From: Gavin Shan <gwshan@linux.vnet.ibm.com>
Date: Wed, 1 Oct 2014 17:07:53 +1000
Subject: [PATCH 14/22] powerpc/eeh: Block PCI config access upon frozen PE

The problem was found when I tried to inject PCI config error by
PHB3 PAPR error injection registers into Broadcom Austin 4-ports
NIC adapter. The frozen PE was reported successfully and EEH core
started to recover it. However, I run into fenced PHB when dumping
PCI config space as EEH logs. I was told that PCI config requests
should not be progagated to the adapter until PE reset is done
successfully. Otherise, we would run out of PHB internal credits
and trigger PCT (PCIE Completion Timeout), which leads to the
fenced PHB.

The patch introduces another PE flag EEH_PE_CFG_RESTRICTED, which
is set during PE initialization time if the PE includes the specific
PCI devices that need block PCI config access until PE reset is done.
When the PE becomes frozen for the first time, EEH_PE_CFG_BLOCKED is
set if the PE has flag EEH_PE_CFG_RESTRICTED. Then the PCI config
access to the PE will be dropped by platform PCI accessors until
PE reset is done successfully. The mechanism is shared by PowerNV
platform owned PE or userland owned ones. It's not used on pSeries
platform yet.

Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/include/asm/eeh.h               |  1 +
 arch/powerpc/kernel/eeh_pe.c                 |  8 ++++++++
 arch/powerpc/platforms/powernv/eeh-powernv.c | 19 +++++++++++++++++++
 3 files changed, 28 insertions(+)

diff --git a/arch/powerpc/include/asm/eeh.h b/arch/powerpc/include/asm/eeh.h
index 6a2ad90e6d8c..ca07f9c27335 100644
--- a/arch/powerpc/include/asm/eeh.h
+++ b/arch/powerpc/include/asm/eeh.h
@@ -74,6 +74,7 @@ struct device_node;
 #define EEH_PE_CFG_BLOCKED	(1 << 2)	/* Block config access	*/
 
 #define EEH_PE_KEEP		(1 << 8)	/* Keep PE on hotplug	*/
+#define EEH_PE_CFG_RESTRICTED	(1 << 9)	/* Block config on error */
 
 struct eeh_pe {
 	int type;			/* PE type: PHB/Bus/Device	*/
diff --git a/arch/powerpc/kernel/eeh_pe.c b/arch/powerpc/kernel/eeh_pe.c
index 37f21284809c..5a63e2b0f65b 100644
--- a/arch/powerpc/kernel/eeh_pe.c
+++ b/arch/powerpc/kernel/eeh_pe.c
@@ -534,6 +534,10 @@ static void *__eeh_pe_state_mark(void *data, void *flag)
 			pdev->error_state = pci_channel_io_frozen;
 	}
 
+	/* Block PCI config access if required */
+	if (pe->state & EEH_PE_CFG_RESTRICTED)
+		pe->state |= EEH_PE_CFG_BLOCKED;
+
 	return NULL;
 }
 
@@ -611,6 +615,10 @@ static void *__eeh_pe_state_clear(void *data, void *flag)
 		pdev->error_state = pci_channel_io_normal;
 	}
 
+	/* Unblock PCI config access if required */
+	if (pe->state & EEH_PE_CFG_RESTRICTED)
+		pe->state &= ~EEH_PE_CFG_BLOCKED;
+
 	return NULL;
 }
 
diff --git a/arch/powerpc/platforms/powernv/eeh-powernv.c b/arch/powerpc/platforms/powernv/eeh-powernv.c
index 04e42f78a37a..443ce965a5b0 100644
--- a/arch/powerpc/platforms/powernv/eeh-powernv.c
+++ b/arch/powerpc/platforms/powernv/eeh-powernv.c
@@ -168,6 +168,25 @@ static int powernv_eeh_dev_probe(struct pci_dev *dev, void *flag)
 		return ret;
 	}
 
+	/*
+	 * If the PE contains any one of following adapters, the
+	 * PCI config space can't be accessed when dumping EEH log.
+	 * Otherwise, we will run into fenced PHB caused by shortage
+	 * of outbound credits in the adapter. The PCI config access
+	 * should be blocked until PE reset. MMIO access is dropped
+	 * by hardware certainly. In order to drop PCI config requests,
+	 * one more flag (EEH_PE_CFG_RESTRICTED) is introduced, which
+	 * will be checked in the backend for PE state retrival. If
+	 * the PE becomes frozen for the first time and the flag has
+	 * been set for the PE, we will set EEH_PE_CFG_BLOCKED for
+	 * that PE to block its config space.
+	 *
+	 * Broadcom Austin 4-ports NICs (14e4:1657)
+	 */
+	if (dev->vendor == PCI_VENDOR_ID_BROADCOM &&
+	    dev->device == 0x1657)
+		edev->pe->state |= EEH_PE_CFG_RESTRICTED;
+
 	/*
 	 * Cache the PE primary bus, which can't be fetched when
 	 * full hotplug is in progress. In that case, all child

From c59004cc83c3f8b182c32ca9d366d222a59ab63f Mon Sep 17 00:00:00 2001
From: Gavin Shan <gwshan@linux.vnet.ibm.com>
Date: Wed, 1 Oct 2014 17:07:54 +1000
Subject: [PATCH 15/22] powerpc/eeh: Don't collect logs on PE with blocked
 config space

When the PE's config space is marked as blocked, PCI config read
requests always return 0xFF's. It's pointless to collect logs in
this case.

Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/kernel/eeh.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/arch/powerpc/kernel/eeh.c b/arch/powerpc/kernel/eeh.c
index 4d83f50cc614..2248a1999c64 100644
--- a/arch/powerpc/kernel/eeh.c
+++ b/arch/powerpc/kernel/eeh.c
@@ -257,6 +257,13 @@ static void *eeh_dump_pe_log(void *data, void *flag)
 	struct eeh_dev *edev, *tmp;
 	size_t *plen = flag;
 
+	/* If the PE's config space is blocked, 0xFF's will be
+	 * returned. It's pointless to collect the log in this
+	 * case.
+	 */
+	if (pe->state & EEH_PE_CFG_BLOCKED)
+		return NULL;
+
 	eeh_pe_for_each_dev(pe, edev, tmp)
 		*plen += eeh_dump_dev_log(edev, pci_regs_buf + *plen,
 					  EEH_PCI_REGS_LOG_LEN - *plen);

From 179ea48bc7c04dba3526d66d9f358c2f4f3b3776 Mon Sep 17 00:00:00 2001
From: Gavin Shan <gwshan@linux.vnet.ibm.com>
Date: Fri, 3 Oct 2014 14:58:32 +1000
Subject: [PATCH 16/22] powerpc/eeh: Block CFG upon frozen Shiner adapter

The Broadcom Shiner 2-ports 10G ethernet adapter has same problem
commit 6f20bda0 ("powerpc/eeh: Block PCI config access upon frozen
PE") fixes. Put it to the black list as well.

   # lspci -s 0004:01:00.0
   0004:01:00.0 Ethernet controller: Broadcom Corporation \
                NetXtreme II BCM57810 10 Gigabit Ethernet (rev 10)
   # lspci -n -s 0004:01:00.0
   0004:01:00.0 0200: 14e4:168e (rev 10)

Reported-by: John Walthour <jwalthour@us.ibm.com>
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/platforms/powernv/eeh-powernv.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/platforms/powernv/eeh-powernv.c b/arch/powerpc/platforms/powernv/eeh-powernv.c
index 443ce965a5b0..1d19e7917d7f 100644
--- a/arch/powerpc/platforms/powernv/eeh-powernv.c
+++ b/arch/powerpc/platforms/powernv/eeh-powernv.c
@@ -182,9 +182,10 @@ static int powernv_eeh_dev_probe(struct pci_dev *dev, void *flag)
 	 * that PE to block its config space.
 	 *
 	 * Broadcom Austin 4-ports NICs (14e4:1657)
+	 * Broadcom Shiner 2-ports 10G NICs (14e4:168e)
 	 */
-	if (dev->vendor == PCI_VENDOR_ID_BROADCOM &&
-	    dev->device == 0x1657)
+	if ((dev->vendor == PCI_VENDOR_ID_BROADCOM && dev->device == 0x1657) ||
+	    (dev->vendor == PCI_VENDOR_ID_BROADCOM && dev->device == 0x168e))
 		edev->pe->state |= EEH_PE_CFG_RESTRICTED;
 
 	/*

From 695911fb1f0e00aebe6c5636b9c08bf0fd51a2fd Mon Sep 17 00:00:00 2001
From: Michael Ellerman <mpe@ellerman.id.au>
Date: Fri, 10 Oct 2014 19:04:24 +1100
Subject: [PATCH 17/22] powerpc/msi: Fix the msi bitmap alignment tests

When we added the alignment tests recently we failed to check they were
actually passing - oops.

They weren't passing, because the bitmap was full. We should also be a
bit more careful when checking the return code, a negative error return
could by divisible by our alignment value.

Fixes: b0345bbc6d09 ("powerpc/msi: Improve IRQ bitmap allocator")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/sysdev/msi_bitmap.c | 26 ++++++++++++++++++--------
 1 file changed, 18 insertions(+), 8 deletions(-)

diff --git a/arch/powerpc/sysdev/msi_bitmap.c b/arch/powerpc/sysdev/msi_bitmap.c
index 0c75214b6f92..8155d93dee1d 100644
--- a/arch/powerpc/sysdev/msi_bitmap.c
+++ b/arch/powerpc/sysdev/msi_bitmap.c
@@ -151,7 +151,7 @@ void msi_bitmap_free(struct msi_bitmap *bmp)
 static void __init test_basics(void)
 {
 	struct msi_bitmap bmp;
-	int i, size = 512;
+	int rc, i, size = 512;
 
 	/* Can't allocate a bitmap of 0 irqs */
 	check(msi_bitmap_alloc(&bmp, 0, NULL) != 0);
@@ -185,14 +185,24 @@ static void __init test_basics(void)
 	msi_bitmap_free_hwirqs(&bmp, size / 2, 1);
 	check(msi_bitmap_alloc_hwirqs(&bmp, 1) == size / 2);
 
+	/* Free most of them for the alignment tests */
+	msi_bitmap_free_hwirqs(&bmp, 3, size - 3);
+
 	/* Check we get a naturally aligned offset */
-	check(msi_bitmap_alloc_hwirqs(&bmp, 2) % 2 == 0);
-	check(msi_bitmap_alloc_hwirqs(&bmp, 4) % 4 == 0);
-	check(msi_bitmap_alloc_hwirqs(&bmp, 8) % 8 == 0);
-	check(msi_bitmap_alloc_hwirqs(&bmp, 9) % 16 == 0);
-	check(msi_bitmap_alloc_hwirqs(&bmp, 3) % 4 == 0);
-	check(msi_bitmap_alloc_hwirqs(&bmp, 7) % 8 == 0);
-	check(msi_bitmap_alloc_hwirqs(&bmp, 121) % 128 == 0);
+	rc = msi_bitmap_alloc_hwirqs(&bmp, 2);
+	check(rc >= 0 && rc % 2 == 0);
+	rc = msi_bitmap_alloc_hwirqs(&bmp, 4);
+	check(rc >= 0 && rc % 4 == 0);
+	rc = msi_bitmap_alloc_hwirqs(&bmp, 8);
+	check(rc >= 0 && rc % 8 == 0);
+	rc = msi_bitmap_alloc_hwirqs(&bmp, 9);
+	check(rc >= 0 && rc % 16 == 0);
+	rc = msi_bitmap_alloc_hwirqs(&bmp, 3);
+	check(rc >= 0 && rc % 4 == 0);
+	rc = msi_bitmap_alloc_hwirqs(&bmp, 7);
+	check(rc >= 0 && rc % 8 == 0);
+	rc = msi_bitmap_alloc_hwirqs(&bmp, 121);
+	check(rc >= 0 && rc % 128 == 0);
 
 	msi_bitmap_free(&bmp);
 

From 4a77f2bdbdef289a02bd02fac483a9350e039705 Mon Sep 17 00:00:00 2001
From: Michael Ellerman <mpe@ellerman.id.au>
Date: Fri, 10 Oct 2014 19:04:25 +1100
Subject: [PATCH 18/22] powerpc/msi: Use WARN_ON() in msi bitmap selftests

As demonstrated in the previous commit, the failure message from the msi
bitmap selftests is a bit subtle, it's easy to miss a failure in a busy
boot log.

So drop our check() macro and use WARN_ON() instead. This necessitates
inverting all the conditions as well.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/sysdev/msi_bitmap.c | 54 ++++++++++++++------------------
 1 file changed, 24 insertions(+), 30 deletions(-)

diff --git a/arch/powerpc/sysdev/msi_bitmap.c b/arch/powerpc/sysdev/msi_bitmap.c
index 8155d93dee1d..73b64c73505b 100644
--- a/arch/powerpc/sysdev/msi_bitmap.c
+++ b/arch/powerpc/sysdev/msi_bitmap.c
@@ -145,69 +145,64 @@ void msi_bitmap_free(struct msi_bitmap *bmp)
 
 #ifdef CONFIG_MSI_BITMAP_SELFTEST
 
-#define check(x)	\
-	if (!(x)) printk("msi_bitmap: test failed at line %d\n", __LINE__);
-
 static void __init test_basics(void)
 {
 	struct msi_bitmap bmp;
 	int rc, i, size = 512;
 
 	/* Can't allocate a bitmap of 0 irqs */
-	check(msi_bitmap_alloc(&bmp, 0, NULL) != 0);
+	WARN_ON(msi_bitmap_alloc(&bmp, 0, NULL) == 0);
 
 	/* of_node may be NULL */
-	check(0 == msi_bitmap_alloc(&bmp, size, NULL));
+	WARN_ON(msi_bitmap_alloc(&bmp, size, NULL));
 
 	/* Should all be free by default */
-	check(0 == bitmap_find_free_region(bmp.bitmap, size,
-					   get_count_order(size)));
+	WARN_ON(bitmap_find_free_region(bmp.bitmap, size, get_count_order(size)));
 	bitmap_release_region(bmp.bitmap, 0, get_count_order(size));
 
 	/* With no node, there's no msi-available-ranges, so expect > 0 */
-	check(msi_bitmap_reserve_dt_hwirqs(&bmp) > 0);
+	WARN_ON(msi_bitmap_reserve_dt_hwirqs(&bmp) <= 0);
 
 	/* Should all still be free */
-	check(0 == bitmap_find_free_region(bmp.bitmap, size,
-					   get_count_order(size)));
+	WARN_ON(bitmap_find_free_region(bmp.bitmap, size, get_count_order(size)));
 	bitmap_release_region(bmp.bitmap, 0, get_count_order(size));
 
 	/* Check we can fill it up and then no more */
 	for (i = 0; i < size; i++)
-		check(msi_bitmap_alloc_hwirqs(&bmp, 1) >= 0);
+		WARN_ON(msi_bitmap_alloc_hwirqs(&bmp, 1) < 0);
 
-	check(msi_bitmap_alloc_hwirqs(&bmp, 1) < 0);
+	WARN_ON(msi_bitmap_alloc_hwirqs(&bmp, 1) >= 0);
 
 	/* Should all be allocated */
-	check(bitmap_find_free_region(bmp.bitmap, size, 0) < 0);
+	WARN_ON(bitmap_find_free_region(bmp.bitmap, size, 0) >= 0);
 
 	/* And if we free one we can then allocate another */
 	msi_bitmap_free_hwirqs(&bmp, size / 2, 1);
-	check(msi_bitmap_alloc_hwirqs(&bmp, 1) == size / 2);
+	WARN_ON(msi_bitmap_alloc_hwirqs(&bmp, 1) != size / 2);
 
 	/* Free most of them for the alignment tests */
 	msi_bitmap_free_hwirqs(&bmp, 3, size - 3);
 
 	/* Check we get a naturally aligned offset */
 	rc = msi_bitmap_alloc_hwirqs(&bmp, 2);
-	check(rc >= 0 && rc % 2 == 0);
+	WARN_ON(rc < 0 && rc % 2 != 0);
 	rc = msi_bitmap_alloc_hwirqs(&bmp, 4);
-	check(rc >= 0 && rc % 4 == 0);
+	WARN_ON(rc < 0 && rc % 4 != 0);
 	rc = msi_bitmap_alloc_hwirqs(&bmp, 8);
-	check(rc >= 0 && rc % 8 == 0);
+	WARN_ON(rc < 0 && rc % 8 != 0);
 	rc = msi_bitmap_alloc_hwirqs(&bmp, 9);
-	check(rc >= 0 && rc % 16 == 0);
+	WARN_ON(rc < 0 && rc % 16 != 0);
 	rc = msi_bitmap_alloc_hwirqs(&bmp, 3);
-	check(rc >= 0 && rc % 4 == 0);
+	WARN_ON(rc < 0 && rc % 4 != 0);
 	rc = msi_bitmap_alloc_hwirqs(&bmp, 7);
-	check(rc >= 0 && rc % 8 == 0);
+	WARN_ON(rc < 0 && rc % 8 != 0);
 	rc = msi_bitmap_alloc_hwirqs(&bmp, 121);
-	check(rc >= 0 && rc % 128 == 0);
+	WARN_ON(rc < 0 && rc % 128 != 0);
 
 	msi_bitmap_free(&bmp);
 
-	/* Clients may check bitmap == NULL for "not-allocated" */
-	check(bmp.bitmap == NULL);
+	/* Clients may WARN_ON bitmap == NULL for "not-allocated" */
+	WARN_ON(bmp.bitmap != NULL);
 
 	kfree(bmp.bitmap);
 }
@@ -229,14 +224,13 @@ static void __init test_of_node(void)
 	of_node_init(&of_node);
 	of_node.full_name = node_name;
 
-	check(0 == msi_bitmap_alloc(&bmp, size, &of_node));
+	WARN_ON(msi_bitmap_alloc(&bmp, size, &of_node));
 
 	/* No msi-available-ranges, so expect > 0 */
-	check(msi_bitmap_reserve_dt_hwirqs(&bmp) > 0);
+	WARN_ON(msi_bitmap_reserve_dt_hwirqs(&bmp) <= 0);
 
 	/* Should all still be free */
-	check(0 == bitmap_find_free_region(bmp.bitmap, size,
-					   get_count_order(size)));
+	WARN_ON(bitmap_find_free_region(bmp.bitmap, size, get_count_order(size)));
 	bitmap_release_region(bmp.bitmap, 0, get_count_order(size));
 
 	/* Now create a fake msi-available-ranges property */
@@ -250,11 +244,11 @@ static void __init test_of_node(void)
 	of_node.properties = &prop;
 
 	/* msi-available-ranges, so expect == 0 */
-	check(msi_bitmap_reserve_dt_hwirqs(&bmp) == 0);
+	WARN_ON(msi_bitmap_reserve_dt_hwirqs(&bmp));
 
 	/* Check we got the expected result */
-	check(0 == bitmap_parselist(expected_str, expected, size));
-	check(bitmap_equal(expected, bmp.bitmap, size));
+	WARN_ON(bitmap_parselist(expected_str, expected, size));
+	WARN_ON(!bitmap_equal(expected, bmp.bitmap, size));
 
 	msi_bitmap_free(&bmp);
 	kfree(bmp.bitmap);

From 5c9fb1899400096c6818181c525897a31d57e488 Mon Sep 17 00:00:00 2001
From: Greg Kurz <gkurz@linux.vnet.ibm.com>
Date: Wed, 15 Oct 2014 12:42:58 +0200
Subject: [PATCH 19/22] powerpc/vphn: NUMA node code expects big-endian

The associativity domain numbers are obtained from the hypervisor through
registers and written into memory by the guest: the packed array passed to
vphn_unpack_associativity() is then native-endian, unlike what was assumed
in the following commit:

commit b08a2a12e44eaec5024b2b969f4fcb98169d1ca3
Author: Alistair Popple <alistair@popple.id.au>
Date:   Wed Aug 7 02:01:44 2013 +1000

    powerpc: Make NUMA device node code endian safe

This issue fills the topology with bogus data and makes it unusable. It may
lead to severe performance breakdowns.

We should ideally patch the vphn_unpack_associativity() function to do the
64-bit loads, but this requires some more brain storming.

In the meantime, let's go for a suboptimal and temporary bug fix: this patch
converts each 64-bit value of the packed array to big endian, as expected by
the current parsing code in vphn_unpack_associativity().

Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/mm/numa.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
index 177659050fa0..e5236c24dc07 100644
--- a/arch/powerpc/mm/numa.c
+++ b/arch/powerpc/mm/numa.c
@@ -1460,8 +1460,11 @@ static long hcall_vphn(unsigned long cpu, __be32 *associativity)
 	long retbuf[PLPAR_HCALL9_BUFSIZE] = {0};
 	u64 flags = 1;
 	int hwcpu = get_hard_smp_processor_id(cpu);
+	int i;
 
 	rc = plpar_hcall9(H_HOME_NODE_ASSOCIATIVITY, retbuf, flags, hwcpu);
+	for (i = 0; i < 6; i++)
+		retbuf[i] = cpu_to_be64(retbuf[i]);
 	vphn_unpack_associativity(retbuf, associativity);
 
 	return rc;

From 2c186e05a5c6dc8fcfb1e8bf6901ad1598c40db6 Mon Sep 17 00:00:00 2001
From: Anton Blanchard <anton@samba.org>
Date: Mon, 13 Oct 2014 20:21:22 +1100
Subject: [PATCH 20/22] powerpc: Add printk levels to setup_system output

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/kernel/setup_64.c | 32 ++++++++++++++++----------------
 1 file changed, 16 insertions(+), 16 deletions(-)

diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index cd07d79ad21c..4f3cfe1b6a33 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -522,36 +522,36 @@ void __init setup_system(void)
 	smp_release_cpus();
 #endif
 
-	printk("Starting Linux PPC64 %s\n", init_utsname()->version);
+	pr_info("Starting Linux PPC64 %s\n", init_utsname()->version);
 
-	printk("-----------------------------------------------------\n");
-	printk("ppc64_pft_size    = 0x%llx\n", ppc64_pft_size);
-	printk("phys_mem_size     = 0x%llx\n", memblock_phys_mem_size());
+	pr_info("-----------------------------------------------------\n");
+	pr_info("ppc64_pft_size    = 0x%llx\n", ppc64_pft_size);
+	pr_info("phys_mem_size     = 0x%llx\n", memblock_phys_mem_size());
 
 	if (ppc64_caches.dline_size != 0x80)
-		printk("dcache_line_size  = 0x%x\n", ppc64_caches.dline_size);
+		pr_info("dcache_line_size  = 0x%x\n", ppc64_caches.dline_size);
 	if (ppc64_caches.iline_size != 0x80)
-		printk("icache_line_size  = 0x%x\n", ppc64_caches.iline_size);
+		pr_info("icache_line_size  = 0x%x\n", ppc64_caches.iline_size);
 
-	printk("cpu_features      = 0x%016lx\n", cur_cpu_spec->cpu_features);
-	printk("  possible        = 0x%016lx\n", CPU_FTRS_POSSIBLE);
-	printk("  always          = 0x%016lx\n", CPU_FTRS_ALWAYS);
-	printk("cpu_user_features = 0x%08x 0x%08x\n", cur_cpu_spec->cpu_user_features,
+	pr_info("cpu_features      = 0x%016lx\n", cur_cpu_spec->cpu_features);
+	pr_info("  possible        = 0x%016lx\n", CPU_FTRS_POSSIBLE);
+	pr_info("  always          = 0x%016lx\n", CPU_FTRS_ALWAYS);
+	pr_info("cpu_user_features = 0x%08x 0x%08x\n", cur_cpu_spec->cpu_user_features,
 		cur_cpu_spec->cpu_user_features2);
-	printk("mmu_features      = 0x%08x\n", cur_cpu_spec->mmu_features);
-	printk("firmware_features = 0x%016lx\n", powerpc_firmware_features);
+	pr_info("mmu_features      = 0x%08x\n", cur_cpu_spec->mmu_features);
+	pr_info("firmware_features = 0x%016lx\n", powerpc_firmware_features);
 
 #ifdef CONFIG_PPC_STD_MMU_64
 	if (htab_address)
-		printk("htab_address      = 0x%p\n", htab_address);
+		pr_info("htab_address      = 0x%p\n", htab_address);
 
-	printk("htab_hash_mask    = 0x%lx\n", htab_hash_mask);
+	pr_info("htab_hash_mask    = 0x%lx\n", htab_hash_mask);
 #endif
 
 	if (PHYSICAL_START > 0)
-		printk("physical_start    = 0x%llx\n",
+		pr_info("physical_start    = 0x%llx\n",
 		       (unsigned long long)PHYSICAL_START);
-	printk("-----------------------------------------------------\n");
+	pr_info("-----------------------------------------------------\n");
 
 	DBG(" <- setup_system()\n");
 }

From 86be175a730bd98de2b75522eae08160ec2dec91 Mon Sep 17 00:00:00 2001
From: Anton Blanchard <anton@samba.org>
Date: Mon, 13 Oct 2014 20:17:33 +1100
Subject: [PATCH 21/22] powerpc: sync pseries_le_defconfig with
 pseries_defconfig

Now KVM is working on LE, enable it. Also enable transarent
hugepage which has already been enabled on BE.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/configs/pseries_le_defconfig | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/arch/powerpc/configs/pseries_le_defconfig b/arch/powerpc/configs/pseries_le_defconfig
index 4428ee428f4e..96a230cd5ca7 100644
--- a/arch/powerpc/configs/pseries_le_defconfig
+++ b/arch/powerpc/configs/pseries_le_defconfig
@@ -48,7 +48,6 @@ CONFIG_KEXEC=y
 CONFIG_IRQ_ALL_CPUS=y
 CONFIG_MEMORY_HOTPLUG=y
 CONFIG_MEMORY_HOTREMOVE=y
-CONFIG_CMA=y
 CONFIG_PPC_64K_PAGES=y
 CONFIG_PPC_SUBPAGE_PROT=y
 CONFIG_SCHED_SMT=y
@@ -137,6 +136,7 @@ CONFIG_NETCONSOLE=y
 CONFIG_NETPOLL_TRAP=y
 CONFIG_TUN=m
 CONFIG_VIRTIO_NET=m
+CONFIG_VHOST_NET=m
 CONFIG_VORTEX=y
 CONFIG_ACENIC=m
 CONFIG_ACENIC_OMIT_TIGON_I=y
@@ -302,4 +302,9 @@ CONFIG_CRYPTO_LZO=m
 # CONFIG_CRYPTO_ANSI_CPRNG is not set
 CONFIG_CRYPTO_DEV_NX=y
 CONFIG_CRYPTO_DEV_NX_ENCRYPT=m
+CONFIG_VIRTUALIZATION=y
+CONFIG_KVM_BOOK3S_64=m
+CONFIG_KVM_BOOK3S_64_HV=y
+CONFIG_TRANSPARENT_HUGEPAGE=y
+CONFIG_TRANSPARENT_HUGEPAGE_ALWAYS=y
 CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND=y

From e89dafb5ca5022d3bc63602018adfc766c73bc2b Mon Sep 17 00:00:00 2001
From: Michael Ellerman <mpe@ellerman.id.au>
Date: Thu, 16 Oct 2014 17:43:02 +1100
Subject: [PATCH 22/22] powerpc: Only do dynamic DMA zone limits on platforms
 that need it

Scott's patch 1c98025c6c95 "Dynamic DMA zone limits" changed
dma_direct_alloc_coherent() to start using dev->coherent_dma_mask.

That seems fair enough, but it exposes the fact that some of the drivers
we care about on IBM platforms aren't setting the coherent mask.

The proper fix is to have drivers set the coherent mask and also have
the platform code honor it.

For now, just restrict the dynamic DMA zone limits to the platforms that
need it.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Scott Wood <scottwood@freescale.com>
---
 arch/powerpc/kernel/dma.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/arch/powerpc/kernel/dma.c b/arch/powerpc/kernel/dma.c
index adac9dc54aee..484b2d4462c1 100644
--- a/arch/powerpc/kernel/dma.c
+++ b/arch/powerpc/kernel/dma.c
@@ -53,9 +53,16 @@ void *dma_direct_alloc_coherent(struct device *dev, size_t size,
 #else
 	struct page *page;
 	int node = dev_to_node(dev);
+#ifdef CONFIG_FSL_SOC
 	u64 pfn = get_pfn_limit(dev);
 	int zone;
 
+	/*
+	 * This code should be OK on other platforms, but we have drivers that
+	 * don't set coherent_dma_mask. As a workaround we just ifdef it. This
+	 * whole routine needs some serious cleanup.
+	 */
+
 	zone = dma_pfn_limit_to_zone(pfn);
 	if (zone < 0) {
 		dev_err(dev, "%s: No suitable zone for pfn %#llx\n",
@@ -73,6 +80,7 @@ void *dma_direct_alloc_coherent(struct device *dev, size_t size,
 		break;
 #endif
 	};
+#endif /* CONFIG_FSL_SOC */
 
 	/* ignore region specifiers */
 	flag  &= ~(__GFP_HIGHMEM);