Skip to contentSkip to conversation

s390 bux fixes for 2.6.25-rc2.

PATCH In review Martin Schwidefsky 4 participants 21 messages Feb 19, 2008 – Feb 20, 2008

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover s390 bux fixes for 2.6.25-rc2.

Greetings,
this and that for s390. The shortlog:

Cornelia Huck (2):
[S390] cio: Remember to initialize recovery_lock. [S390] cio: Do timed recovery on workqueue.

Heiko Carstens (6):
[S390] Let NR_CPUS default to 32/64 on s390/s390x. [S390] Make sure enabled wait psw is loaded in default_idle. [S390] Initialize per cpu lowcores on cpu hotplug. [S390] qdio: fix qdio_activate timeout handling. [S390] etr: fix compile error on !SMP
[S390] Fix futex_atomic_cmpxchg_std inline assembly.

Martin Schwidefsky (1):
[S390] find bit corner case.

Peter Oberparleiter (1):
[S390] sclp: clean up send/receive naming scheme

Roel Kluin (1):

      [S390] dcss: Fix Unlikely(x) != y

Stefan Weinhuber (1):
[S390] dasd: fix locking in __dasd_device_process_final_queue

Ursula Braun (1):
[S390] qdio: FCP/SCSI write I/O stagnates on LPAR

Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#1 lore raw

PATCH 1/13 cio: Remember to initialize recovery_lock.

From: Cornelia Huck <[email protected]>

---

 drivers/s390/cio/device.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/s390/cio/device.c +1 -1
Index: quilt-2.6/drivers/s390/cio/device.c
===================================================================
@@ -32,7 +32,7 @@
 #include "io_sch.h"
 
 static struct timer_list recovery_timer;
-static spinlock_t recovery_lock;
+static DEFINE_SPINLOCK(recovery_lock);
 static int recovery_phase;
 static const unsigned long recovery_delay[] = { 3, 30, 300 };
 
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#2 lore raw

PATCH 2/13 cio: Do timed recovery on workqueue.

In reply to Martin Schwidefsky

From: Cornelia Huck <[email protected]>

We can't do our recovery in softirq context, so we schedule it from our timer function.

---

 drivers/s390/cio/device.c |   13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)
drivers/s390/cio/device.c +12 -1
Index: quilt-2.6/drivers/s390/cio/device.c
===================================================================
@@ -1535,7 +1535,7 @@ static int recovery_check(struct device 
 	return 0;
 }
 
-static void recovery_func(unsigned long data)
+static void recovery_work_func(struct work_struct *unused)
 {
 	int redo = 0;
 
@@ -1553,6 +1553,17 @@ static void recovery_func(unsigned long 
 		CIO_MSG_EVENT(2, "recovery: end\n");
 }
 
+static DECLARE_WORK(recovery_work, recovery_work_func);
+
+static void recovery_func(unsigned long data)
+{
+	/*
+	 * We can't do our recovery in softirq context and it's not
+	 * performance critical, so we schedule it.
+	 */
+	schedule_work(&recovery_work);
+}
+
 void ccw_device_schedule_recovery(void)
 {
 	unsigned long flags;
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#3 lore raw

PATCH 3/13 Let NR_CPUS default to 32/64 on s390/s390x.

In reply to Martin Schwidefsky

From: Heiko Carstens <[email protected]>

---

 arch/s390/Kconfig |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/s390/Kconfig +2 -1
Index: quilt-2.6/arch/s390/Kconfig
===================================================================
@@ -100,7 +100,8 @@ config NR_CPUS
 	int "Maximum number of CPUs (2-64)"
 	range 2 64
 	depends on SMP
-	default "32"
+	default "32" if !64BIT
+	default "64" if 64BIT
 	help
 	  This allows you to specify the maximum number of CPUs which this
 	  kernel will support.  The maximum supported value is 64 and the
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#4 lore raw

PATCH 4/13 Make sure enabled wait psw is loaded in default_idle.

In reply to Martin Schwidefsky

From: Heiko Carstens <[email protected]>

If both NO_IDLE_HZ and VIRT_TIMER are disabled default_idle won't load an enabled wait psw and busy loop instead. This is because the idle_chain is empty and the return value of atomic_notifier_call_chain will be NOTIFY_DONE, which causes default_idle to return instead of loading an enabled wait psw.
Fix this by calling __atomic_notifier_call_chain instead and add proper return value handling.

---

 arch/s390/kernel/process.c |   15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)
arch/s390/kernel/process.c +9 -6
Index: quilt-2.6/arch/s390/kernel/process.c
===================================================================
@@ -114,24 +114,27 @@ extern void s390_handle_mcck(void);
 static void default_idle(void)
 {
 	int cpu, rc;
+	int nr_calls = 0;
+	void *hcpu;
 #ifdef CONFIG_SMP
 	struct s390_idle_data *idle;
 #endif
 
 	/* CPU is going idle. */
 	cpu = smp_processor_id();
-
+	hcpu = (void *)(long)cpu;
 	local_irq_disable();
 	if (need_resched()) {
 		local_irq_enable();
 		return;
 	}
 
-	rc = atomic_notifier_call_chain(&idle_chain,
-					S390_CPU_IDLE, (void *)(long) cpu);
-	if (rc != NOTIFY_OK && rc != NOTIFY_DONE)
-		BUG();
-	if (rc != NOTIFY_OK) {
+	rc = __atomic_notifier_call_chain(&idle_chain, S390_CPU_IDLE, hcpu, -1,
+					  &nr_calls);
+	if (rc == NOTIFY_BAD) {
+		nr_calls--;
+		__atomic_notifier_call_chain(&idle_chain, S390_CPU_NOT_IDLE,
+					     hcpu, nr_calls, NULL);
 		local_irq_enable();
 		return;
 	}
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#5 lore raw

PATCH 5/13 dasd: fix locking in __dasd_device_process_final_queue

In reply to Martin Schwidefsky

From: Stefan Weinhuber <[email protected]>

After setting the status of the cqr and releasing the lock for the block cqr queue, we call the cqr callback function, which will usually just trigger the dasd_block_tasklet. But when the tasklet is already running the cqr might be processed before we invoke the callback function. In rare cases the callback pointer may already be invalid by the time we want to call it, which will result in a panic.
Solution: Call the callback function first and then release the lock.

---

 drivers/s390/block/dasd.c |   12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/s390/block/dasd.c +6 -6
Index: quilt-2.6/drivers/s390/block/dasd.c
===================================================================
@@ -1149,12 +1149,14 @@ static void __dasd_device_process_final_
 {
 	struct list_head *l, *n;
 	struct dasd_ccw_req *cqr;
+	struct dasd_block *block;
 
 	list_for_each_safe(l, n, final_queue) {
 		cqr = list_entry(l, struct dasd_ccw_req, devlist);
 		list_del_init(&cqr->devlist);
-		if (cqr->block)
-			spin_lock_bh(&cqr->block->queue_lock);
+		block = cqr->block;
+		if (block)
+			spin_lock_bh(&block->queue_lock);
 		switch (cqr->status) {
 		case DASD_CQR_SUCCESS:
 			cqr->status = DASD_CQR_DONE;
@@ -1172,15 +1174,13 @@ static void __dasd_device_process_final_
 				    cqr, cqr->status);
 			BUG();
 		}
-		if (cqr->block)
-			spin_unlock_bh(&cqr->block->queue_lock);
 		if (cqr->callback != NULL)
 			(cqr->callback)(cqr, cqr->callback_data);
+		if (block)
+			spin_unlock_bh(&block->queue_lock);
 	}
 }
 
-
-
 /*
  * Take a look at the first request on the ccw queue and check
  * if it reached its expire time. If so, terminate the IO.
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#6 lore raw

PATCH 6/13 find bit corner case.

In reply to Martin Schwidefsky

From: Martin Schwidefsky <[email protected]>

Fix [ext2_]find_first_[zero_]bit for the corner case of an all clear or all set bit field by always handling that last word of the bit field with __ffz_word/__ffs_word.

---

 include/asm-s390/bitops.h |   20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)
include/asm-s390/bitops.h +12 -8
Index: quilt-2.6/include/asm-s390/bitops.h
===================================================================
@@ -456,16 +456,18 @@ static inline unsigned long __ffz_word_l
 
 	asm volatile(
 #ifndef __s390x__
-		"	ahi	%1,31\n"
-		"	srl	%1,5\n"
+		"	ahi	%1,-1\n"
+		"	sra	%1,5\n"
+		"	jz	1f\n"
 		"0:	c	%2,0(%0,%3)\n"
 		"	jne	1f\n"
 		"	la	%0,4(%0)\n"
 		"	brct	%1,0b\n"
 		"1:\n"
 #else
-		"	aghi	%1,63\n"
-		"	srlg	%1,%1,6\n"
+		"	aghi	%1,-1\n"
+		"	srag	%1,%1,6\n"
+		"	jz	1f\n"
 		"0:	cg	%2,0(%0,%3)\n"
 		"	jne	1f\n"
 		"	la	%0,8(%0)\n"
@@ -491,16 +493,18 @@ static inline unsigned long __ffs_word_l
 
 	asm volatile(
 #ifndef __s390x__
-		"	ahi	%1,31\n"
-		"	srl	%1,5\n"
+		"	ahi	%1,-1\n"
+		"	sra	%1,5\n"
+		"	jz	1f\n"
 		"0:	c	%2,0(%0,%3)\n"
 		"	jne	1f\n"
 		"	la	%0,4(%0)\n"
 		"	brct	%1,0b\n"
 		"1:\n"
 #else
-		"	aghi	%1,63\n"
-		"	srlg	%1,%1,6\n"
+		"	aghi	%1,-1\n"
+		"	srag	%1,%1,6\n"
+		"	jz	1f\n"
 		"0:	cg	%2,0(%0,%3)\n"
 		"	jne	1f\n"
 		"	la	%0,8(%0)\n"
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#7 lore raw

PATCH 7/13 Initialize per cpu lowcores on cpu hotplug.

In reply to Martin Schwidefsky

From: Heiko Carstens <[email protected]>

Just copy the first 512 read-only bytes of the current cpu lowcore if a new cpu gets onlined. The rest is zeroed out and must be explicitly initialized. Current code just copies the entire lowcore and initializes the needed fields.
This should reveal bugs in future enhancements quite early. Also when the lowcore of the first cpu is replaced this is now done atomically (no interrupts, no machine checks).

---

 arch/s390/kernel/smp.c |   53 +++++++++++++++++++++++++++++++++++--------------
 1 file changed, 38 insertions(+), 15 deletions(-)
arch/s390/kernel/smp.c +38 -15
Index: quilt-2.6/arch/s390/kernel/smp.c
===================================================================
@@ -626,13 +626,17 @@ static int __cpuinit smp_alloc_lowcore(i
 	if (!lowcore)
 		return -ENOMEM;
 	async_stack = __get_free_pages(GFP_KERNEL, ASYNC_ORDER);
-	if (!async_stack)
-		goto out_async_stack;
 	panic_stack = __get_free_page(GFP_KERNEL);
-	if (!panic_stack)
-		goto out_panic_stack;
-
-	*lowcore = S390_lowcore;
+	if (!panic_stack || !async_stack)
+		goto out;
+	/*
+	 * Only need to copy the first 512 bytes from address 0. But since
+	 * the compiler emits a warning if src == NULL for memcpy use copy_page
+	 * instead. Copies more than needed but this code is not performance
+	 * critical.
+	 */
+	copy_page(lowcore, &S390_lowcore);
+	memset((void *)lowcore + 512, 0, sizeof(*lowcore) - 512);
 	lowcore->async_stack = async_stack + ASYNC_SIZE;
 	lowcore->panic_stack = panic_stack + PAGE_SIZE;
 
@@ -653,9 +657,8 @@ static int __cpuinit smp_alloc_lowcore(i
 out_save_area:
 	free_page(panic_stack);
 #endif
-out_panic_stack:
+out:
 	free_pages(async_stack, ASYNC_ORDER);
-out_async_stack:
 	free_pages((unsigned long) lowcore, lc_order);
 	return -ENOMEM;
 }
@@ -719,8 +722,8 @@ int __cpuinit __cpu_up(unsigned int cpu)
 	cpu_lowcore->percpu_offset = __per_cpu_offset[cpu];
 	cpu_lowcore->current_task = (unsigned long) idle;
 	cpu_lowcore->cpu_data.cpu_nr = cpu;
-	cpu_lowcore->softirq_pending = 0;
-	cpu_lowcore->ext_call_fast = 0;
+	cpu_lowcore->kernel_asce = S390_lowcore.kernel_asce;
+	cpu_lowcore->ipl_device = S390_lowcore.ipl_device;
 	eieio();
 
 	while (signal_processor(cpu, sigp_restart) == sigp_busy)
@@ -797,23 +800,43 @@ void cpu_die(void)
 
 void __init smp_prepare_cpus(unsigned int max_cpus)
 {
+#ifndef CONFIG_64BIT
+	unsigned long save_area = 0;
+#endif
+	unsigned long async_stack, panic_stack;
+	struct _lowcore *lowcore;
 	unsigned int cpu;
+	int lc_order;
 
 	smp_detect_cpus();
 
 	/* request the 0x1201 emergency signal external interrupt */
 	if (register_external_interrupt(0x1201, do_ext_call_interrupt) != 0)
 		panic("Couldn't request external interrupt 0x1201");
-	memset(lowcore_ptr, 0, sizeof(lowcore_ptr));
 	print_cpu_info(&S390_lowcore.cpu_data);
-	smp_alloc_lowcore(smp_processor_id());
 
+	/* Reallocate current lowcore, but keep its contents. */
+	lc_order = sizeof(long) == 8 ? 1 : 0;
+	lowcore = (void *) __get_free_pages(GFP_KERNEL | GFP_DMA, lc_order);
+	panic_stack = __get_free_page(GFP_KERNEL);
+	async_stack = __get_free_pages(GFP_KERNEL, ASYNC_ORDER);
 #ifndef CONFIG_64BIT
 	if (MACHINE_HAS_IEEE)
-		ctl_set_bit(14, 29); /* enable extended save area */
+		save_area = get_zeroed_page(GFP_KERNEL);
 #endif
-	set_prefix((u32)(unsigned long) lowcore_ptr[smp_processor_id()]);
-
+	local_irq_disable();
+	local_mcck_disable();
+	lowcore_ptr[smp_processor_id()] = lowcore;
+	*lowcore = S390_lowcore;
+	lowcore->panic_stack = panic_stack + PAGE_SIZE;
+	lowcore->async_stack = async_stack + ASYNC_SIZE;
+#ifndef CONFIG_64BIT
+	if (MACHINE_HAS_IEEE)
+		lowcore->extended_save_area_addr = (u32) save_area;
+#endif
+	set_prefix((u32)(unsigned long) lowcore);
+	local_mcck_enable();
+	local_irq_enable();
 	for_each_possible_cpu(cpu)
 		if (cpu != smp_processor_id())
 			smp_create_idle(cpu);
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#8 lore raw

Re: [patch 07/13] Initialize per cpu lowcores on cpu hotplug.

arch/s390/kernel/smp.c static int __cpuinit smp_alloc_lowcore(i · line 638

Quoted 7 lines · Martin Schwidefsky
arch/s390/kernel/smp.c +7 -0
+	/*
+	 * Only need to copy the first 512 bytes from address 0. But since
+	 * the compiler emits a warning if src == NULL for memcpy use copy_page
+	 * instead. Copies more than needed but this code is not performance
+	 * critical.
+	 */
+	copy_page(lowcore, &S390_lowcore);

Boah, workaround alert. Why do you not fix the compiler?

Bastian

Signature · 3 lines
We fight only when there is no other choice.  We prefer the ways of
peaceful contact.
		-- Kirk, "Spectre of the Gun", stardate 4385.3
#9 lore raw

Re: [patch 07/13] Initialize per cpu lowcores on cpu hotplug.

Quoted 10 lines · Bastian Blank

arch/s390/kernel/smp.c static int __cpuinit smp_alloc_lowcore(i · line 638

Quoted 7 lines · Martin Schwidefsky
arch/s390/kernel/smp.c +7 -0
+	/*
+	 * Only need to copy the first 512 bytes from address 0. But since
+	 * the compiler emits a warning if src == NULL for memcpy use copy_page
+	 * instead. Copies more than needed but this code is not performance
+	 * critical.
+	 */
+	copy_page(lowcore, &S390_lowcore);

Boah, workaround alert. Why do you not fix the compiler?

We need to copy from address 0 (that's where the lowcore resides). But gcc insists to complain if memcpy is used with src == NULL.. Now what?

#10 lore raw

Re: [patch 07/13] Initialize per cpu lowcores on cpu hotplug.

Quoted 12 lines · Heiko Carstens
Quoted 10 lines · Bastian Blank

arch/s390/kernel/smp.c static int __cpuinit smp_alloc_lowcore(i · line 638

Quoted 7 lines · Martin Schwidefsky
arch/s390/kernel/smp.c +7 -0
+	/*
+	 * Only need to copy the first 512 bytes from address 0. But since
+	 * the compiler emits a warning if src == NULL for memcpy use copy_page
+	 * instead. Copies more than needed but this code is not performance
+	 * critical.
+	 */
+	copy_page(lowcore, &S390_lowcore);

Boah, workaround alert. Why do you not fix the compiler?

Quoted 2 lines

We need to copy from address 0 (that's where the lowcore resides). But gcc insists to complain if memcpy is used with src == NULL.. Now what?

Erm sorry, misread your question. Usually it's a bug to use memcpy with src == NULL. In this case it's ok. So I think it's perfectly ok if gcc emits a warning.
If you know of a better way to get around this please let me know.

#11 lore raw

Re: [patch 07/13] Initialize per cpu lowcores on cpu hotplug.

Quoted 12 lines
Quoted 12 lines
Quoted 11 lines

arch/s390/kernel/smp.c static int __cpuinit smp_alloc_lowcore(i · line 638

Quoted 9 lines
arch/s390/kernel/smp.c +3 -0
+	/*
+	 * Only need to copy the first 512 bytes from address 0. But since
+	 * the compiler emits a warning if src == NULL for memcpy use 

copy_page

(quoted diff) +1 -0
+	 * instead. Copies more than needed but this code is not 

performance

arch/s390/kernel/smp.c +3 -0
+	 * critical.
+	 */
+	copy_page(lowcore, &S390_lowcore);

Boah, workaround alert. Why do you not fix the compiler?

Quoted 7 lines
Quoted 2 lines

We need to copy from address 0 (that's where the lowcore resides). But gcc insists to complain if memcpy is used with src == NULL.. Now what?

Erm sorry, misread your question. Usually it's a bug to use memcpy with src == NULL. In this case it's ok. So I think it's perfectly ok if gcc emits a warning.
If you know of a better way to get around this please let me know.

-ffreestanding or -Wno-nonnull?

Segher

#12 lore raw

Re: [patch 07/13] Initialize per cpu lowcores on cpu hotplug.

Quoted 6 lines
Quoted 6 lines
Quoted 5 lines
Quoted 1 line

Boah, workaround alert. Why do you not fix the compiler?

We need to copy from address 0 (that's where the lowcore resides). But gcc
insists to complain if memcpy is used with src == NULL.. Now what?

Quoted 6 lines
Quoted 4 lines

Erm sorry, misread your question. Usually it's a bug to use memcpy with src == NULL. In this case it's ok. So I think it's perfectly ok if gcc emits a warning.
If you know of a better way to get around this please let me know.

-ffreestanding or -Wno-nonnull?

Ok, how about the patch below? Everybody happy with it?

---
 arch/s390/kernel/Makefile |    5 +++++
 arch/s390/kernel/smp.c    |    8 +-------
 2 files changed, 6 insertions(+), 7 deletions(-)
arch/s390/kernel/Makefile +5 -0
Index: linux-2.5/arch/s390/kernel/Makefile
===================================================================
@@ -4,6 +4,11 @@
 
 EXTRA_AFLAGS	:= -traditional
 
+#
+# Passing null pointers is ok for smp code, since we access the lowcore here.
+#
+CFLAGS_smp.o	:= -Wno-nonnull
+
 obj-y	:=  bitmap.o traps.o time.o process.o base.o early.o \
             setup.o sys_s390.o ptrace.o signal.o cpcmd.o ebcdic.o \
 	    semaphore.o s390_ext.o debug.o irq.o ipl.o dis.o diag.o
arch/s390/kernel/smp.c +1 -7
Index: linux-2.5/arch/s390/kernel/smp.c
===================================================================
@@ -631,13 +631,7 @@ static int __cpuinit smp_alloc_lowcore(i
 	panic_stack = __get_free_page(GFP_KERNEL);
 	if (!panic_stack || !async_stack)
 		goto out;
-	/*
-	 * Only need to copy the first 512 bytes from address 0. But since
-	 * the compiler emits a warning if src == NULL for memcpy use copy_page
-	 * instead. Copies more than needed but this code is not performance
-	 * critical.
-	 */
-	copy_page(lowcore, &S390_lowcore);
+	memcpy(lowcore, &S390_lowcore, 512);
 	memset((void *)lowcore + 512, 0, sizeof(*lowcore) - 512);
 	lowcore->async_stack = async_stack + ASYNC_SIZE;
 	lowcore->panic_stack = panic_stack + PAGE_SIZE;
#13 lore raw

Re: [patch 07/13] Initialize per cpu lowcores on cpu hotplug.

arch/s390/kernel/smp.c static int __cpuinit smp_alloc_lowcore(i · line 634

Quoted 2 lines · Heiko Carstens
arch/s390/kernel/smp.c +1 -1
-	copy_page(lowcore, &S390_lowcore);
+	memcpy(lowcore, &S390_lowcore, 512);

Okay

Quoted 1 line

memset((void *)lowcore + 512, 0, sizeof(*lowcore) - 512);

Not completely okay. void pointer are not allowed in arithmetic. gcc handles void * as char * in this case, but I think it should usualy be avoided.

Bastian

Signature · 2 lines
Peace was the way.
		-- Kirk, "The City on the Edge of Forever", stardate unknown
#14 lore raw

Re: [patch 07/13] Initialize per cpu lowcores on cpu hotplug.

Quoted 11 lines · Bastian Blank

arch/s390/kernel/smp.c static int __cpuinit smp_alloc_lowcore(i · line 634

Quoted 2 lines · Heiko Carstens
arch/s390/kernel/smp.c +1 -1
-	copy_page(lowcore, &S390_lowcore);
+	memcpy(lowcore, &S390_lowcore, 512);

Okay

Quoted 1 line

memset((void *)lowcore + 512, 0, sizeof(*lowcore) - 512);

Not completely okay. void pointer are not allowed in arithmetic. gcc handles void * as char * in this case, but I think it should usualy be avoided.

There are many places all over the kernel that assume sizeof(void) == 1. That's yet another gcc extension we use... but I'm going to change that to a char * cast anyway.

Thanks for commenting!

#15 lore raw

PATCH 8/13 qdio: fix qdio_activate timeout handling.

In reply to Martin Schwidefsky

From: Heiko Carstens <[email protected]>

Current code in qdio_activate waits for at least 5 seconds until it returns. It may return earlier if an error occurs, but not if everything is ok. This large timeout value became visible with commit dfa77f611ff295598e218aa0eb6efa73a5cf26d0 "qdio: set QDIO_ACTIVATE_TIMEOUT to 5s", which intended to fix the timeout value which was zero. In turn setting an FCP adapter online took 5 seconds.

In practice waiting for 5ms before continuing is sufficient as pointed out by Utz Bacher and Cornelia Huck.

---

 drivers/s390/cio/qdio.c |   10 ++--------
 drivers/s390/cio/qdio.h |    2 +-
 2 files changed, 3 insertions(+), 9 deletions(-)
drivers/s390/cio/qdio.c +2 -8
Index: quilt-2.6/drivers/s390/cio/qdio.c
===================================================================
@@ -32,7 +32,7 @@
 
 #include <linux/module.h>
 #include <linux/init.h>
-
+#include <linux/delay.h>
 #include <linux/slab.h>
 #include <linux/kernel.h>
 #include <linux/proc_fs.h>
@@ -3332,13 +3332,7 @@ qdio_activate(struct ccw_device *cdev, i
 		}
 	}
 
-	wait_event_interruptible_timeout(cdev->private->wait_q,
-					 ((irq_ptr->state ==
-					  QDIO_IRQ_STATE_STOPPED) ||
-					  (irq_ptr->state ==
-					   QDIO_IRQ_STATE_ERR)),
-					 QDIO_ACTIVATE_TIMEOUT);
-
+	msleep(QDIO_ACTIVATE_TIMEOUT);
 	switch (irq_ptr->state) {
 	case QDIO_IRQ_STATE_STOPPED:
 	case QDIO_IRQ_STATE_ERR:
drivers/s390/cio/qdio.h +1 -1
Index: quilt-2.6/drivers/s390/cio/qdio.h
===================================================================
@@ -57,10 +57,10 @@
 					    of the queue to 0 */
 
 #define QDIO_ESTABLISH_TIMEOUT (1*HZ)
-#define QDIO_ACTIVATE_TIMEOUT (5*HZ)
 #define QDIO_CLEANUP_CLEAR_TIMEOUT (20*HZ)
 #define QDIO_CLEANUP_HALT_TIMEOUT (10*HZ)
 #define QDIO_FORCE_CHECK_TIMEOUT (10*HZ)
+#define QDIO_ACTIVATE_TIMEOUT (5) /* 5 ms */
 
 enum qdio_irq_states {
 	QDIO_IRQ_STATE_INACTIVE,
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#16 lore raw

PATCH 9/13 etr: fix compile error on !SMP

In reply to Martin Schwidefsky

From: Heiko Carstens <[email protected]>

Since a5fbb6d1064be885d2a6b82f625186753cf74848
"KVM: fix !SMP build error" smp_call_function isn't a define anymore that folds into nothing but a define that calls up_smp_call_function with all parameters. Hence we cannot #ifdef out the unused code anymore...
This seems to be the preferred method, so do this for s390 as well.

arch/s390/kernel/time.c: In function 'etr_sync_clock':
arch/s390/kernel/time.c:825: error: 'clock_sync_cpu_start' undeclared arch/s390/kernel/time.c:862: error: 'clock_sync_cpu_end' undeclared

---

 arch/s390/kernel/time.c |    2 --
 1 file changed, 2 deletions(-)
arch/s390/kernel/time.c +0 -2
Index: quilt-2.6/arch/s390/kernel/time.c
===================================================================
@@ -744,7 +744,6 @@ static void etr_adjust_time(unsigned lon
 	}
 }
 
-#ifdef CONFIG_SMP
 static void etr_sync_cpu_start(void *dummy)
 {
 	int *in_sync = dummy;
@@ -777,7 +776,6 @@ static void etr_sync_cpu_start(void *dum
 static void etr_sync_cpu_end(void *dummy)
 {
 }
-#endif /* CONFIG_SMP */
 
 /*
  * Sync the TOD clock using the port refered to by aibp. This port
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#17 lore raw

PATCH 10/13 sclp: clean up send/receive naming scheme

In reply to Martin Schwidefsky

From: Peter Oberparleiter <[email protected]>

Make state change events adjust the correct mask by cleaning up naming inconsistencies. Also remove chance for lockup by removing unnecessary mask related check before reading events.

---

 drivers/s390/char/sclp.c         |   12 ++++++------
 drivers/s390/char/sclp.h         |    6 ++++--
 drivers/s390/char/sclp_config.c  |    2 +-
 drivers/s390/char/sclp_cpi_sys.c |    2 +-
 drivers/s390/char/sclp_rw.c      |    4 ++--
 drivers/s390/char/sclp_vt220.c   |    2 +-
 6 files changed, 15 insertions(+), 13 deletions(-)
drivers/s390/char/sclp.c +6 -6
Index: quilt-2.6/drivers/s390/char/sclp.c
===================================================================
@@ -29,10 +29,10 @@ static ext_int_info_t ext_int_info_hwc;
 /* Lock to protect internal data consistency. */
 static DEFINE_SPINLOCK(sclp_lock);
 
-/* Mask of events that we can receive from the sclp interface. */
+/* Mask of events that we can send to the sclp interface. */
 static sccb_mask_t sclp_receive_mask;
 
-/* Mask of events that we can send to the sclp interface. */
+/* Mask of events that we can receive from the sclp interface. */
 static sccb_mask_t sclp_send_mask;
 
 /* List of registered event listeners and senders. */
@@ -380,7 +380,7 @@ sclp_interrupt_handler(__u16 code)
 		}
 		sclp_running_state = sclp_running_state_idle;
 	}
-	if (evbuf_pending && sclp_receive_mask != 0 &&
+	if (evbuf_pending &&
 	    sclp_activation_state == sclp_activation_state_active)
 		__sclp_queue_read_req();
 	spin_unlock(&sclp_lock);
@@ -459,8 +459,8 @@ sclp_dispatch_state_change(void)
 		reg = NULL;
 		list_for_each(l, &sclp_reg_list) {
 			reg = list_entry(l, struct sclp_register, list);
-			receive_mask = reg->receive_mask & sclp_receive_mask;
-			send_mask = reg->send_mask & sclp_send_mask;
+			receive_mask = reg->send_mask & sclp_receive_mask;
+			send_mask = reg->receive_mask & sclp_send_mask;
 			if (reg->sclp_receive_mask != receive_mask ||
 			    reg->sclp_send_mask != send_mask) {
 				reg->sclp_receive_mask = receive_mask;
@@ -615,8 +615,8 @@ struct init_sccb {
 	u16 mask_length;
 	sccb_mask_t receive_mask;
 	sccb_mask_t send_mask;
-	sccb_mask_t sclp_send_mask;
 	sccb_mask_t sclp_receive_mask;
+	sccb_mask_t sclp_send_mask;
 } __attribute__((packed));
 
 /* Prepare init mask request. Called while sclp_lock is locked. */
drivers/s390/char/sclp_config.c +1 -1
Index: quilt-2.6/drivers/s390/char/sclp_config.c
===================================================================
@@ -64,7 +64,7 @@ static int __init sclp_conf_init(void)
 		return rc;
 	}
 
-	if (!(sclp_conf_register.sclp_receive_mask & EVTYP_CONFMGMDATA_MASK)) {
+	if (!(sclp_conf_register.sclp_send_mask & EVTYP_CONFMGMDATA_MASK)) {
 		printk(KERN_WARNING TAG "no configuration management.\n");
 		sclp_unregister(&sclp_conf_register);
 		rc = -ENOSYS;
drivers/s390/char/sclp_cpi_sys.c +1 -1
Index: quilt-2.6/drivers/s390/char/sclp_cpi_sys.c
===================================================================
@@ -129,7 +129,7 @@ static int cpi_req(void)
 			"to hardware console.\n");
 		goto out;
 	}
-	if (!(sclp_cpi_event.sclp_send_mask & EVTYP_CTLPROGIDENT_MASK)) {
+	if (!(sclp_cpi_event.sclp_receive_mask & EVTYP_CTLPROGIDENT_MASK)) {
 		printk(KERN_WARNING "cpi: no control program "
 			"identification support\n");
 		rc = -EOPNOTSUPP;
drivers/s390/char/sclp.h +4 -2
Index: quilt-2.6/drivers/s390/char/sclp.h
===================================================================
@@ -122,11 +122,13 @@ struct sclp_req {
 /* of some routines it wants to be called from the low level driver */
 struct sclp_register {
 	struct list_head list;
-	/* event masks this user is registered for */
+	/* User wants to receive: */
 	sccb_mask_t receive_mask;
+	/* User wants to send: */
 	sccb_mask_t send_mask;
-	/* actually present events */
+	/* H/W can receive: */
 	sccb_mask_t sclp_receive_mask;
+	/* H/W can send: */
 	sccb_mask_t sclp_send_mask;
 	/* called if event type availability changes */
 	void (*state_change_fn)(struct sclp_register *);
drivers/s390/char/sclp_rw.c +2 -2
Index: quilt-2.6/drivers/s390/char/sclp_rw.c
===================================================================
@@ -452,10 +452,10 @@ sclp_emit_buffer(struct sclp_buffer *buf
 		return -EIO;
 
 	sccb = buffer->sccb;
-	if (sclp_rw_event.sclp_send_mask & EVTYP_MSG_MASK)
+	if (sclp_rw_event.sclp_receive_mask & EVTYP_MSG_MASK)
 		/* Use normal write message */
 		sccb->msg_buf.header.type = EVTYP_MSG;
-	else if (sclp_rw_event.sclp_send_mask & EVTYP_PMSGCMD_MASK)
+	else if (sclp_rw_event.sclp_receive_mask & EVTYP_PMSGCMD_MASK)
 		/* Use write priority message */
 		sccb->msg_buf.header.type = EVTYP_PMSGCMD;
 	else
drivers/s390/char/sclp_vt220.c +1 -1
Index: quilt-2.6/drivers/s390/char/sclp_vt220.c
===================================================================
@@ -202,7 +202,7 @@ sclp_vt220_callback(struct sclp_req *req
 static int
 __sclp_vt220_emit(struct sclp_vt220_request *request)
 {
-	if (!(sclp_vt220_register.sclp_send_mask & EVTYP_VT220MSG_MASK)) {
+	if (!(sclp_vt220_register.sclp_receive_mask & EVTYP_VT220MSG_MASK)) {
 		request->sclp_req.status = SCLP_REQ_FAILED;
 		return -EIO;
 	}
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#18 lore raw

PATCH 11/13 dcss: Fix Unlikely(x) != y

In reply to Martin Schwidefsky

From: Roel Kluin <[email protected]>

Fix Unlikely(x) != y

---

 drivers/s390/block/dcssblk.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/s390/block/dcssblk.c +1 -1
Index: quilt-2.6/drivers/s390/block/dcssblk.c
===================================================================
@@ -666,7 +666,7 @@ dcssblk_make_request(struct request_queu
 		page_addr = (unsigned long)
 			page_address(bvec->bv_page) + bvec->bv_offset;
 		source_addr = dev_info->start + (index<<12) + bytes_done;
-		if (unlikely(page_addr & 4095) != 0 || (bvec->bv_len & 4095) != 0)
+		if (unlikely((page_addr & 4095) != 0) || (bvec->bv_len & 4095) != 0)
 			// More paranoia.
 			goto fail;
 		if (bio_data_dir(bio) == READ) {
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#19 lore raw

PATCH 12/13 Fix futex_atomic_cmpxchg_std inline assembly.

In reply to Martin Schwidefsky

From: Heiko Carstens <[email protected]>

Add missing exception table entry so that the kernel can handle proctection exceptions as well on the cs instruction. Currently only specification exceptions are handled correctly.
The missing entry allows user space to crash the kernel.

---

 arch/s390/lib/uaccess_std.c |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
arch/s390/lib/uaccess_std.c +4 -4
Index: quilt-2.6/arch/s390/lib/uaccess_std.c
===================================================================
@@ -293,10 +293,10 @@ int futex_atomic_cmpxchg_std(int __user 
 
 	asm volatile(
 		"   sacf 256\n"
-		"   cs   %1,%4,0(%5)\n"
-		"0: lr   %0,%1\n"
-		"1: sacf 0\n"
-		EX_TABLE(0b,1b)
+		"0: cs   %1,%4,0(%5)\n"
+		"1: lr   %0,%1\n"
+		"2: sacf 0\n"
+		EX_TABLE(0b,2b) EX_TABLE(1b,2b)
 		: "=d" (ret), "+d" (oldval), "=m" (*uaddr)
 		: "0" (-EFAULT), "d" (newval), "a" (uaddr), "m" (*uaddr)
 		: "cc", "memory" );
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.
#20 lore raw

PATCH 13/13 qdio: FCP/SCSI write I/O stagnates on LPAR

In reply to Martin Schwidefsky

From: Ursula Braun <[email protected]>

If running on LPAR, qdio might overlook an incoming buffer in certain scenarios. The patch makes sure that incoming buffers are detected immediately in all situations.

---

 drivers/s390/cio/qdio.c |    3 ---
 1 file changed, 3 deletions(-)
drivers/s390/cio/qdio.c +0 -3
Index: quilt-2.6/drivers/s390/cio/qdio.c
===================================================================
@@ -1215,9 +1215,6 @@ tiqdio_is_inbound_q_done(struct qdio_q *
 
 	if (!no_used)
 		return 1;
-	if (!q->siga_sync && !irq->is_qebsm)
-		/* we'll check for more primed buffers in qeth_stop_polling */
-		return 0;
 	if (irq->is_qebsm) {
 		count = 1;
 		start_buf = q->first_to_check;
Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.