* [PATCH 0/5] cpumask: more cpumask updates
@ 2009-01-07 19:58 Mike Travis
2009-01-07 19:58 ` [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t Mike Travis
` (5 more replies)
0 siblings, 6 replies; 31+ messages in thread
From: Mike Travis @ 2009-01-07 19:58 UTC (permalink / raw)
To: Ingo Molnar, Rusty Russell
Cc: Thomas Gleixner, H. Peter Anvin, Jack Steiner, linux-kernel
One big change to irq_desc and a few smaller maintenance type updates.
cpumask: update irq_desc to use cpumask_var_t
cpumask: Use topology_core_cpumask()/topology_thread_cpumask()
cpumask: convert misc driver functions
cpumask: convert drivers/net/sfc
cpumask: convert other misc kernel functions
Based on tip/cpus4096: v2.6.28-5703-g1d1a70e
(Ingo - I added a blank line before the From: part and that seems to
separate it from the Subject. Hopefully your git-am will still
recognizes it - I did try a git-am here and it seemed to work ok.)
Signed-off-by: Mike Travis <travis@sgi.com>
--
^ permalink raw reply [flat|nested] 31+ messages in thread
* [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t
2009-01-07 19:58 [PATCH 0/5] cpumask: more cpumask updates Mike Travis
@ 2009-01-07 19:58 ` Mike Travis
2009-01-07 20:27 ` Yinghai Lu
2009-01-07 19:58 ` [PATCH 2/5] cpumask: Use topology_core_cpumask()/topology_thread_cpumask() Mike Travis
` (4 subsequent siblings)
5 siblings, 1 reply; 31+ messages in thread
From: Mike Travis @ 2009-01-07 19:58 UTC (permalink / raw)
To: Ingo Molnar, Rusty Russell
Cc: Thomas Gleixner, H. Peter Anvin, Jack Steiner, linux-kernel,
Chris Wright, Jeremy Fitzhardinge, KOSAKI Motohiro,
Venkatesh Pallipadi, virtualization, xen-devel, Yinghai Lu
[-- Attachment #1: cpumask:update-irq_desc-to-use-cpumask_var_t.patch --]
[-- Type: text/plain, Size: 14894 bytes --]
Impact: reduce memory usage, use new cpumask API.
Replace the affinity and pending_masks with cpumask_var_t's. This adds
to the significant size reduction done with the SPARSE_IRQS changes.
The added functions (init_alloc_desc_masks & init_copy_desc_masks) are
in the include file so they can be inlined (and optimized out for the
!CONFIG_CPUMASKS_OFFSTACK case.) [Naming chosen to be consistent with
the other init*irq functions, as well as the backwards arg declaration
of "from, to" instead of the more common "to, from" standard.]
Includes a slight change to the declaration of struct irq_desc to embed
the pending_mask within ifdef(CONFIG_SMP) to be consistent with other
references, and some small changes to Xen.
Tested: sparse/non-sparse/cpumask_offstack/non-cpumask_offstack/nonuma/nosmp on x86_64
Signed-off-by: Mike Travis <travis@sgi.com>
Cc: Chris Wright <chrisw@sous-sol.org>
Cc: Jeremy Fitzhardinge <jeremy@xensource.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Cc: virtualization@lists.osdl.org
Cc: xen-devel@lists.xensource.com
Cc: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/io_apic.c | 20 +++++------
arch/x86/kernel/irq_32.c | 2 -
arch/x86/kernel/irq_64.c | 2 -
drivers/xen/events.c | 4 +-
include/linux/irq.h | 81 ++++++++++++++++++++++++++++++++++++++++++++--
kernel/irq/chip.c | 5 ++
kernel/irq/handle.c | 26 +++++++-------
kernel/irq/manage.c | 12 +++---
kernel/irq/migration.c | 12 +++---
kernel/irq/numa_migrate.c | 12 ++++++
kernel/irq/proc.c | 4 +-
11 files changed, 135 insertions(+), 45 deletions(-)
--- linux-2.6-for-ingo.orig/arch/x86/kernel/io_apic.c
+++ linux-2.6-for-ingo/arch/x86/kernel/io_apic.c
@@ -358,7 +358,7 @@ set_extra_move_desc(struct irq_desc *des
if (!cfg->move_in_progress) {
/* it means that domain is not changed */
- if (!cpumask_intersects(&desc->affinity, mask))
+ if (!cpumask_intersects(desc->affinity, mask))
cfg->move_desc_pending = 1;
}
}
@@ -581,9 +581,9 @@ set_desc_affinity(struct irq_desc *desc,
if (assign_irq_vector(irq, cfg, mask))
return BAD_APICID;
- cpumask_and(&desc->affinity, cfg->domain, mask);
+ cpumask_and(desc->affinity, cfg->domain, mask);
set_extra_move_desc(desc, mask);
- return cpu_mask_to_apicid_and(&desc->affinity, cpu_online_mask);
+ return cpu_mask_to_apicid_and(desc->affinity, cpu_online_mask);
}
static void
@@ -2385,7 +2385,7 @@ migrate_ioapic_irq_desc(struct irq_desc
if (cfg->move_in_progress)
send_cleanup_vector(cfg);
- cpumask_copy(&desc->affinity, mask);
+ cpumask_copy(desc->affinity, mask);
}
static int migrate_irq_remapped_level_desc(struct irq_desc *desc)
@@ -2407,11 +2407,11 @@ static int migrate_irq_remapped_level_de
}
/* everthing is clear. we have right of way */
- migrate_ioapic_irq_desc(desc, &desc->pending_mask);
+ migrate_ioapic_irq_desc(desc, desc->pending_mask);
ret = 0;
desc->status &= ~IRQ_MOVE_PENDING;
- cpumask_clear(&desc->pending_mask);
+ cpumask_clear(desc->pending_mask);
unmask:
unmask_IO_APIC_irq_desc(desc);
@@ -2436,7 +2436,7 @@ static void ir_irq_migration(struct work
continue;
}
- desc->chip->set_affinity(irq, &desc->pending_mask);
+ desc->chip->set_affinity(irq, desc->pending_mask);
spin_unlock_irqrestore(&desc->lock, flags);
}
}
@@ -2450,7 +2450,7 @@ static void set_ir_ioapic_affinity_irq_d
{
if (desc->status & IRQ_LEVEL) {
desc->status |= IRQ_MOVE_PENDING;
- cpumask_copy(&desc->pending_mask, mask);
+ cpumask_copy(desc->pending_mask, mask);
migrate_irq_remapped_level_desc(desc);
return;
}
@@ -2518,7 +2518,7 @@ static void irq_complete_move(struct irq
/* domain has not changed, but affinity did */
me = smp_processor_id();
- if (cpu_isset(me, desc->affinity)) {
+ if (cpumask_test_cpu(me, desc->affinity)) {
*descp = desc = move_irq_desc(desc, me);
/* get the new one */
cfg = desc->chip_data;
@@ -4041,7 +4041,7 @@ void __init setup_ioapic_dest(void)
*/
if (desc->status &
(IRQ_NO_BALANCING | IRQ_AFFINITY_SET))
- mask = &desc->affinity;
+ mask = desc->affinity;
else
mask = TARGET_CPUS;
--- linux-2.6-for-ingo.orig/arch/x86/kernel/irq_32.c
+++ linux-2.6-for-ingo/arch/x86/kernel/irq_32.c
@@ -248,7 +248,7 @@ void fixup_irqs(void)
if (irq == 2)
continue;
- affinity = &desc->affinity;
+ affinity = desc->affinity;
if (cpumask_any_and(affinity, cpu_online_mask) >= nr_cpu_ids) {
printk("Breaking affinity for irq %i\n", irq);
affinity = cpu_all_mask;
--- linux-2.6-for-ingo.orig/arch/x86/kernel/irq_64.c
+++ linux-2.6-for-ingo/arch/x86/kernel/irq_64.c
@@ -100,7 +100,7 @@ void fixup_irqs(void)
/* interrupt's are disabled at this point */
spin_lock(&desc->lock);
- affinity = &desc->affinity;
+ affinity = desc->affinity;
if (!irq_has_action(irq) ||
cpumask_equal(affinity, cpu_online_mask)) {
spin_unlock(&desc->lock);
--- linux-2.6-for-ingo.orig/drivers/xen/events.c
+++ linux-2.6-for-ingo/drivers/xen/events.c
@@ -125,7 +125,7 @@ static void bind_evtchn_to_cpu(unsigned
BUG_ON(irq == -1);
#ifdef CONFIG_SMP
- irq_to_desc(irq)->affinity = cpumask_of_cpu(cpu);
+ cpumask_copy(irq_to_desc(irq)->affinity, cpumask_of(cpu));
#endif
__clear_bit(chn, cpu_evtchn_mask[cpu_evtchn[chn]]);
@@ -142,7 +142,7 @@ static void init_evtchn_cpu_bindings(voi
/* By default all event channels notify CPU#0. */
for_each_irq_desc(i, desc) {
- desc->affinity = cpumask_of_cpu(0);
+ cpumask_copy(desc->affinity, cpumask_of(0));
}
#endif
--- linux-2.6-for-ingo.orig/include/linux/irq.h
+++ linux-2.6-for-ingo/include/linux/irq.h
@@ -182,11 +182,11 @@ struct irq_desc {
unsigned int irqs_unhandled;
spinlock_t lock;
#ifdef CONFIG_SMP
- cpumask_t affinity;
+ cpumask_var_t affinity;
unsigned int cpu;
-#endif
#ifdef CONFIG_GENERIC_PENDING_IRQ
- cpumask_t pending_mask;
+ cpumask_var_t pending_mask;
+#endif
#endif
#ifdef CONFIG_PROC_FS
struct proc_dir_entry *dir;
@@ -422,4 +422,79 @@ extern int set_irq_msi(unsigned int irq,
#endif /* !CONFIG_S390 */
+#ifdef CONFIG_SMP
+/**
+ * init_alloc_desc_masks - allocate cpumasks for irq_desc
+ * @desc: pointer to irq_desc struct
+ * @boot: true if need bootmem
+ *
+ * Allocates affinity and pending_mask cpumask if required.
+ * Returns true if successful (or not required).
+ * Side effect: affinity has all bits set, pending_mask has all bits clear.
+ */
+static inline bool init_alloc_desc_masks(struct irq_desc *desc, int node,
+ bool boot)
+{
+ if (boot) {
+ alloc_bootmem_cpumask_var(&desc->affinity);
+ cpumask_setall(desc->affinity);
+
+#ifdef CONFIG_GENERIC_PENDING_IRQ
+ alloc_bootmem_cpumask_var(&desc->pending_mask);
+ cpumask_clear(desc->pending_mask);
+#endif
+ return true;
+ }
+
+ if (!alloc_cpumask_var_node(&desc->affinity, GFP_ATOMIC, node))
+ return false;
+ cpumask_setall(desc->affinity);
+
+#ifdef CONFIG_GENERIC_PENDING_IRQ
+ if (!alloc_cpumask_var_node(&desc->pending_mask, GFP_ATOMIC, node)) {
+ free_cpumask_var(desc->affinity);
+ return false;
+ }
+ cpumask_clear(desc->pending_mask);
+#endif
+ return true;
+}
+
+/**
+ * init_copy_desc_masks - copy cpumasks for irq_desc
+ * @old_desc: pointer to old irq_desc struct
+ * @new_desc: pointer to new irq_desc struct
+ *
+ * Insures affinity and pending_masks are copied to new irq_desc.
+ * If !CONFIG_CPUMASKS_OFFSTACK the cpumasks are embedded in the
+ * irq_desc struct so the copy is redundant.
+ */
+
+static inline void init_copy_desc_masks(struct irq_desc *old_desc,
+ struct irq_desc *new_desc)
+{
+#ifdef CONFIG_CPUMASKS_OFFSTACK
+ cpumask_copy(new_desc->affinity, old_desc->affinity);
+
+#ifdef CONFIG_GENERIC_PENDING_IRQ
+ cpumask_copy(new_desc->pending_mask, old_desc->pending_mask);
+#endif
+#endif
+}
+
+#else /* !CONFIG_SMP */
+
+static inline bool init_alloc_desc_masks(struct irq_desc *desc, int node,
+ bool boot)
+{
+ return true;
+}
+
+static inline void init_copy_desc_masks(struct irq_desc *old_desc,
+ struct irq_desc *new_desc)
+{
+}
+
+#endif /* CONFIG_SMP */
+
#endif /* _LINUX_IRQ_H */
--- linux-2.6-for-ingo.orig/kernel/irq/chip.c
+++ linux-2.6-for-ingo/kernel/irq/chip.c
@@ -46,7 +46,10 @@ void dynamic_irq_init(unsigned int irq)
desc->irq_count = 0;
desc->irqs_unhandled = 0;
#ifdef CONFIG_SMP
- cpumask_setall(&desc->affinity);
+ cpumask_setall(desc->affinity);
+#ifdef CONFIG_GENERIC_PENDING_IRQ
+ cpumask_clear(desc->pending_mask);
+#endif
#endif
spin_unlock_irqrestore(&desc->lock, flags);
}
--- linux-2.6-for-ingo.orig/kernel/irq/handle.c
+++ linux-2.6-for-ingo/kernel/irq/handle.c
@@ -64,9 +64,6 @@ static struct irq_desc irq_desc_init = {
.handle_irq = handle_bad_irq,
.depth = 1,
.lock = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock),
-#ifdef CONFIG_SMP
- .affinity = CPU_MASK_ALL
-#endif
};
void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr)
@@ -88,6 +85,8 @@ void init_kstat_irqs(struct irq_desc *de
static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu)
{
+ int node = cpu_to_node(cpu);
+
memcpy(desc, &irq_desc_init, sizeof(struct irq_desc));
spin_lock_init(&desc->lock);
@@ -101,6 +100,10 @@ static void init_one_irq_desc(int irq, s
printk(KERN_ERR "can not alloc kstat_irqs\n");
BUG_ON(1);
}
+ if (!init_alloc_desc_masks(desc, node, false)) {
+ printk(KERN_ERR "can not alloc irq_desc cpumasks\n");
+ BUG_ON(1);
+ }
arch_init_chip_data(desc, cpu);
}
@@ -119,9 +122,6 @@ static struct irq_desc irq_desc_legacy[N
.handle_irq = handle_bad_irq,
.depth = 1,
.lock = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock),
-#ifdef CONFIG_SMP
- .affinity = CPU_MASK_ALL
-#endif
}
};
@@ -141,7 +141,7 @@ int __init early_irq_init(void)
desc[i].irq = i;
desc[i].kstat_irqs = kstat_irqs_legacy[i];
lockdep_set_class(&desc[i].lock, &irq_desc_lock_class);
-
+ init_alloc_desc_masks(&desc[i], 0, true);
irq_desc_ptrs[i] = desc + i;
}
@@ -188,6 +188,10 @@ struct irq_desc *irq_to_desc_alloc_cpu(u
printk(KERN_ERR "can not alloc irq_desc\n");
BUG_ON(1);
}
+ if (!init_alloc_desc_masks(desc, node, false)) {
+ printk(KERN_ERR "can not alloc irq_desc cpumasks\n");
+ BUG_ON(1);
+ }
init_one_irq_desc(irq, desc, cpu);
irq_desc_ptrs[irq] = desc;
@@ -207,9 +211,6 @@ struct irq_desc irq_desc[NR_IRQS] __cach
.handle_irq = handle_bad_irq,
.depth = 1,
.lock = __SPIN_LOCK_UNLOCKED(irq_desc->lock),
-#ifdef CONFIG_SMP
- .affinity = CPU_MASK_ALL
-#endif
}
};
@@ -222,9 +223,10 @@ int __init early_irq_init(void)
desc = irq_desc;
count = ARRAY_SIZE(irq_desc);
- for (i = 0; i < count; i++)
+ for (i = 0; i < count; i++) {
desc[i].irq = i;
-
+ init_alloc_desc_masks(&desc[i], 0, true);
+ }
return arch_early_irq_init();
}
--- linux-2.6-for-ingo.orig/kernel/irq/manage.c
+++ linux-2.6-for-ingo/kernel/irq/manage.c
@@ -98,14 +98,14 @@ int irq_set_affinity(unsigned int irq, c
#ifdef CONFIG_GENERIC_PENDING_IRQ
if (desc->status & IRQ_MOVE_PCNTXT || desc->status & IRQ_DISABLED) {
- cpumask_copy(&desc->affinity, cpumask);
+ cpumask_copy(desc->affinity, cpumask);
desc->chip->set_affinity(irq, cpumask);
} else {
desc->status |= IRQ_MOVE_PENDING;
- cpumask_copy(&desc->pending_mask, cpumask);
+ cpumask_copy(desc->pending_mask, cpumask);
}
#else
- cpumask_copy(&desc->affinity, cpumask);
+ cpumask_copy(desc->affinity, cpumask);
desc->chip->set_affinity(irq, cpumask);
#endif
desc->status |= IRQ_AFFINITY_SET;
@@ -127,16 +127,16 @@ int do_irq_select_affinity(unsigned int
* one of the targets is online.
*/
if (desc->status & (IRQ_AFFINITY_SET | IRQ_NO_BALANCING)) {
- if (cpumask_any_and(&desc->affinity, cpu_online_mask)
+ if (cpumask_any_and(desc->affinity, cpu_online_mask)
< nr_cpu_ids)
goto set_affinity;
else
desc->status &= ~IRQ_AFFINITY_SET;
}
- cpumask_and(&desc->affinity, cpu_online_mask, irq_default_affinity);
+ cpumask_and(desc->affinity, cpu_online_mask, irq_default_affinity);
set_affinity:
- desc->chip->set_affinity(irq, &desc->affinity);
+ desc->chip->set_affinity(irq, desc->affinity);
return 0;
}
--- linux-2.6-for-ingo.orig/kernel/irq/migration.c
+++ linux-2.6-for-ingo/kernel/irq/migration.c
@@ -18,7 +18,7 @@ void move_masked_irq(int irq)
desc->status &= ~IRQ_MOVE_PENDING;
- if (unlikely(cpumask_empty(&desc->pending_mask)))
+ if (unlikely(cpumask_empty(desc->pending_mask)))
return;
if (!desc->chip->set_affinity)
@@ -38,13 +38,13 @@ void move_masked_irq(int irq)
* For correct operation this depends on the caller
* masking the irqs.
*/
- if (likely(cpumask_any_and(&desc->pending_mask, cpu_online_mask)
+ if (likely(cpumask_any_and(desc->pending_mask, cpu_online_mask)
< nr_cpu_ids)) {
- cpumask_and(&desc->affinity,
- &desc->pending_mask, cpu_online_mask);
- desc->chip->set_affinity(irq, &desc->affinity);
+ cpumask_and(desc->affinity,
+ desc->pending_mask, cpu_online_mask);
+ desc->chip->set_affinity(irq, desc->affinity);
}
- cpumask_clear(&desc->pending_mask);
+ cpumask_clear(desc->pending_mask);
}
void move_native_irq(int irq)
--- linux-2.6-for-ingo.orig/kernel/irq/numa_migrate.c
+++ linux-2.6-for-ingo/kernel/irq/numa_migrate.c
@@ -46,6 +46,7 @@ static void init_copy_one_irq_desc(int i
desc->cpu = cpu;
lockdep_set_class(&desc->lock, &irq_desc_lock_class);
init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids);
+ init_copy_desc_masks(old_desc, desc);
arch_init_copy_chip_data(old_desc, desc, cpu);
}
@@ -76,11 +77,20 @@ static struct irq_desc *__real_move_irq_
node = cpu_to_node(cpu);
desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node);
if (!desc) {
- printk(KERN_ERR "irq %d: can not get new irq_desc for migration.\n", irq);
+ printk(KERN_ERR "irq %d: can not get new irq_desc "
+ "for migration.\n", irq);
/* still use old one */
desc = old_desc;
goto out_unlock;
}
+ if (!init_alloc_desc_masks(desc, node, false)) {
+ printk(KERN_ERR "irq %d: can not get new irq_desc cpumask "
+ "for migration.\n", irq);
+ /* still use old one */
+ kfree(desc);
+ desc = old_desc;
+ goto out_unlock;
+ }
init_copy_one_irq_desc(irq, old_desc, desc, cpu);
irq_desc_ptrs[irq] = desc;
--- linux-2.6-for-ingo.orig/kernel/irq/proc.c
+++ linux-2.6-for-ingo/kernel/irq/proc.c
@@ -20,11 +20,11 @@ static struct proc_dir_entry *root_irq_d
static int irq_affinity_proc_show(struct seq_file *m, void *v)
{
struct irq_desc *desc = irq_to_desc((long)m->private);
- const struct cpumask *mask = &desc->affinity;
+ const struct cpumask *mask = desc->affinity;
#ifdef CONFIG_GENERIC_PENDING_IRQ
if (desc->status & IRQ_MOVE_PENDING)
- mask = &desc->pending_mask;
+ mask = desc->pending_mask;
#endif
seq_cpumask(m, mask);
seq_putc(m, '\n');
--
^ permalink raw reply [flat|nested] 31+ messages in thread
* [PATCH 2/5] cpumask: Use topology_core_cpumask()/topology_thread_cpumask()
2009-01-07 19:58 [PATCH 0/5] cpumask: more cpumask updates Mike Travis
2009-01-07 19:58 ` [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t Mike Travis
@ 2009-01-07 19:58 ` Mike Travis
2009-01-19 17:11 ` Ben Hutchings
2009-01-07 19:58 ` [PATCH 3/5] cpumask: convert misc driver functions Mike Travis
` (3 subsequent siblings)
5 siblings, 1 reply; 31+ messages in thread
From: Mike Travis @ 2009-01-07 19:58 UTC (permalink / raw)
To: Ingo Molnar, Rusty Russell
Cc: Thomas Gleixner, H. Peter Anvin, Jack Steiner, linux-kernel,
linux-net-drivers
[-- Attachment #1: cpumask:use-topology_core_cpumask-and-topology_thread_cpumask.patch --]
[-- Type: text/plain, Size: 5231 bytes --]
From: Rusty Russell <rusty@rustcorp.com.au>
Impact: reduce stack usage, use new cpumask API.
This actually uses topology_core_cpumask() and
topology_thread_cpumask(), removing the only users of
topology_core_siblings() and topology_thread_siblings()
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Cc: linux-net-drivers@solarflare.com
---
Documentation/cputopology.txt | 6 +++---
drivers/base/topology.c | 35 +++++++++++++++++------------------
drivers/net/sfc/efx.c | 4 ++--
include/linux/topology.h | 6 ++++++
4 files changed, 28 insertions(+), 23 deletions(-)
--- linux-2.6-for-ingo.orig/Documentation/cputopology.txt
+++ linux-2.6-for-ingo/Documentation/cputopology.txt
@@ -18,11 +18,11 @@ For an architecture to support this feat
these macros in include/asm-XXX/topology.h:
#define topology_physical_package_id(cpu)
#define topology_core_id(cpu)
-#define topology_thread_siblings(cpu)
-#define topology_core_siblings(cpu)
+#define topology_thread_cpumask(cpu)
+#define topology_core_cpumask(cpu)
The type of **_id is int.
-The type of siblings is cpumask_t.
+The type of siblings is (const) struct cpumask *.
To be consistent on all architectures, include/linux/topology.h
provides default definitions for any of the above macros that are
--- linux-2.6-for-ingo.orig/drivers/base/topology.c
+++ linux-2.6-for-ingo/drivers/base/topology.c
@@ -30,7 +30,10 @@
#include <linux/module.h>
#include <linux/topology.h>
-#define define_one_ro(_name) \
+#define define_one_ro_named(_name, _func) \
+static SYSDEV_ATTR(_name, 0444, _func, NULL)
+
+#define define_one_ro(_name) \
static SYSDEV_ATTR(_name, 0444, show_##_name, NULL)
#define define_id_show_func(name) \
@@ -41,8 +44,8 @@ static ssize_t show_##name(struct sys_de
return sprintf(buf, "%d\n", topology_##name(cpu)); \
}
-#if defined(topology_thread_siblings) || defined(topology_core_siblings)
-static ssize_t show_cpumap(int type, cpumask_t *mask, char *buf)
+#if defined(topology_thread_cpumask) || defined(topology_core_cpumask)
+static ssize_t show_cpumap(int type, const struct cpumask *mask, char *buf)
{
ptrdiff_t len = PTR_ALIGN(buf + PAGE_SIZE - 1, PAGE_SIZE) - buf;
int n = 0;
@@ -64,7 +67,7 @@ static ssize_t show_##name(struct sys_de
struct sysdev_attribute *attr, char *buf) \
{ \
unsigned int cpu = dev->id; \
- return show_cpumap(0, &(topology_##name(cpu)), buf); \
+ return show_cpumap(0, topology_##name(cpu), buf); \
}
#define define_siblings_show_list(name) \
@@ -73,7 +76,7 @@ static ssize_t show_##name##_list(struct
char *buf) \
{ \
unsigned int cpu = dev->id; \
- return show_cpumap(1, &(topology_##name(cpu)), buf); \
+ return show_cpumap(1, topology_##name(cpu), buf); \
}
#else
@@ -81,9 +84,7 @@ static ssize_t show_##name##_list(struct
static ssize_t show_##name(struct sys_device *dev, \
struct sysdev_attribute *attr, char *buf) \
{ \
- unsigned int cpu = dev->id; \
- cpumask_t mask = topology_##name(cpu); \
- return show_cpumap(0, &mask, buf); \
+ return show_cpumap(0, topology_##name(dev->id), buf); \
}
#define define_siblings_show_list(name) \
@@ -91,9 +92,7 @@ static ssize_t show_##name##_list(struct
struct sysdev_attribute *attr, \
char *buf) \
{ \
- unsigned int cpu = dev->id; \
- cpumask_t mask = topology_##name(cpu); \
- return show_cpumap(1, &mask, buf); \
+ return show_cpumap(1, topology_##name(dev->id), buf); \
}
#endif
@@ -106,13 +105,13 @@ define_one_ro(physical_package_id);
define_id_show_func(core_id);
define_one_ro(core_id);
-define_siblings_show_func(thread_siblings);
-define_one_ro(thread_siblings);
-define_one_ro(thread_siblings_list);
-
-define_siblings_show_func(core_siblings);
-define_one_ro(core_siblings);
-define_one_ro(core_siblings_list);
+define_siblings_show_func(thread_cpumask);
+define_one_ro_named(thread_siblings, show_thread_cpumask);
+define_one_ro_named(thread_siblings_list, show_thread_cpumask_list);
+
+define_siblings_show_func(core_cpumask);
+define_one_ro_named(core_siblings, show_core_cpumask);
+define_one_ro_named(core_siblings_list, show_core_cpumask_list);
static struct attribute *default_attrs[] = {
&attr_physical_package_id.attr,
--- linux-2.6-for-ingo.orig/drivers/net/sfc/efx.c
+++ linux-2.6-for-ingo/drivers/net/sfc/efx.c
@@ -863,8 +863,8 @@ static int efx_wanted_rx_queues(void)
for_each_online_cpu(cpu) {
if (!cpu_isset(cpu, core_mask)) {
++count;
- cpus_or(core_mask, core_mask,
- topology_core_siblings(cpu));
+ cpumask_or(&core_mask, &core_mask,
+ topology_core_cpumask(cpu));
}
}
--- linux-2.6-for-ingo.orig/include/linux/topology.h
+++ linux-2.6-for-ingo/include/linux/topology.h
@@ -193,5 +193,11 @@ int arch_update_cpu_topology(void);
#ifndef topology_core_siblings
#define topology_core_siblings(cpu) cpumask_of_cpu(cpu)
#endif
+#ifndef topology_thread_cpumask
+#define topology_thread_cpumask(cpu) cpumask_of(cpu)
+#endif
+#ifndef topology_core_cpumask
+#define topology_core_cpumask(cpu) cpumask_of(cpu)
+#endif
#endif /* _LINUX_TOPOLOGY_H */
--
^ permalink raw reply [flat|nested] 31+ messages in thread
* [PATCH 3/5] cpumask: convert misc driver functions
2009-01-07 19:58 [PATCH 0/5] cpumask: more cpumask updates Mike Travis
2009-01-07 19:58 ` [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t Mike Travis
2009-01-07 19:58 ` [PATCH 2/5] cpumask: Use topology_core_cpumask()/topology_thread_cpumask() Mike Travis
@ 2009-01-07 19:58 ` Mike Travis
2009-01-10 10:57 ` Rusty Russell
` (2 more replies)
2009-01-07 19:58 ` [PATCH 4/5] cpumask: convert drivers/net/sfc Mike Travis
` (2 subsequent siblings)
5 siblings, 3 replies; 31+ messages in thread
From: Mike Travis @ 2009-01-07 19:58 UTC (permalink / raw)
To: Ingo Molnar, Rusty Russell
Cc: Thomas Gleixner, H. Peter Anvin, Jack Steiner, linux-kernel,
Dean Nelson, Robert Richter, oprofile-list, Jeremy Fitzhardinge,
Chris Wright, virtualization, xen-devel, Ursula Braun, linux390,
linux-s390
[-- Attachment #1: cpumask:convert-drivers.patch --]
[-- Type: text/plain, Size: 7183 bytes --]
From: Rusty Russell <rusty@rustcorp.com.au>
Impact: use new cpumask API.
Convert misc driver functions to use struct cpumask.
To Do:
- Convert iucv_buffer_cpumask to cpumask_var_t.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Acked-by: Dean Nelson <dcn@sgi.com>
Cc: Robert Richter <robert.richter@amd.com>
Cc: oprofile-list@lists.sf.net
Cc: Jeremy Fitzhardinge <jeremy@xensource.com>
Cc: Chris Wright <chrisw@sous-sol.org>
Cc: virtualization@lists.osdl.org
Cc: xen-devel@lists.xensource.com
Cc: Ursula Braun <ursula.braun@de.ibm.com>
Cc: linux390@de.ibm.com
Cc: linux-s390@vger.kernel.org
---
drivers/base/cpu.c | 2 +-
drivers/misc/sgi-xp/xpc_main.c | 2 +-
drivers/oprofile/buffer_sync.c | 22 ++++++++++++++++++----
drivers/oprofile/buffer_sync.h | 4 ++++
drivers/oprofile/oprof.c | 9 ++++++++-
drivers/xen/manage.c | 2 +-
net/iucv/iucv.c | 28 ++++++++++++++++++----------
7 files changed, 51 insertions(+), 18 deletions(-)
--- linux-2.6-for-ingo.orig/drivers/base/cpu.c
+++ linux-2.6-for-ingo/drivers/base/cpu.c
@@ -107,7 +107,7 @@ static SYSDEV_ATTR(crash_notes, 0400, sh
/*
* Print cpu online, possible, present, and system maps
*/
-static ssize_t print_cpus_map(char *buf, cpumask_t *map)
+static ssize_t print_cpus_map(char *buf, const struct cpumask *map)
{
int n = cpulist_scnprintf(buf, PAGE_SIZE-2, map);
--- linux-2.6-for-ingo.orig/drivers/misc/sgi-xp/xpc_main.c
+++ linux-2.6-for-ingo/drivers/misc/sgi-xp/xpc_main.c
@@ -318,7 +318,7 @@ xpc_hb_checker(void *ignore)
/* this thread was marked active by xpc_hb_init() */
- set_cpus_allowed_ptr(current, &cpumask_of_cpu(XPC_HB_CHECK_CPU));
+ set_cpus_allowed_ptr(current, cpumask_of(XPC_HB_CHECK_CPU));
/* set our heartbeating to other partitions into motion */
xpc_hb_check_timeout = jiffies + (xpc_hb_check_interval * HZ);
--- linux-2.6-for-ingo.orig/drivers/oprofile/buffer_sync.c
+++ linux-2.6-for-ingo/drivers/oprofile/buffer_sync.c
@@ -37,7 +37,7 @@
static LIST_HEAD(dying_tasks);
static LIST_HEAD(dead_tasks);
-static cpumask_t marked_cpus = CPU_MASK_NONE;
+static cpumask_var_t marked_cpus;
static DEFINE_SPINLOCK(task_mortuary);
static void process_task_mortuary(void);
@@ -489,10 +489,10 @@ static void mark_done(int cpu)
{
int i;
- cpu_set(cpu, marked_cpus);
+ cpumask_set_cpu(cpu, marked_cpus);
for_each_online_cpu(i) {
- if (!cpu_isset(i, marked_cpus))
+ if (!cpumask_test_cpu(i, marked_cpus))
return;
}
@@ -501,7 +501,7 @@ static void mark_done(int cpu)
*/
process_task_mortuary();
- cpus_clear(marked_cpus);
+ cpumask_clear(marked_cpus);
}
@@ -595,6 +595,20 @@ void sync_buffer(int cpu)
mutex_unlock(&buffer_mutex);
}
+int __init buffer_sync_init(void)
+{
+ if (!alloc_cpumask_var(&marked_cpus, GFP_KERNEL))
+ return -ENOMEM;
+
+ cpumask_clear(marked_cpus);
+ return 0;
+}
+
+void __exit buffer_sync_cleanup(void)
+{
+ free_cpumask_var(marked_cpus);
+}
+
/* The function can be used to add a buffer worth of data directly to
* the kernel buffer. The buffer is assumed to be a circular buffer.
* Take the entries from index start and end at index end, wrapping
--- linux-2.6-for-ingo.orig/drivers/oprofile/buffer_sync.h
+++ linux-2.6-for-ingo/drivers/oprofile/buffer_sync.h
@@ -19,4 +19,8 @@ void sync_stop(void);
/* sync the given CPU's buffer */
void sync_buffer(int cpu);
+/* initialize/destroy the buffer system. */
+int buffer_sync_init(void);
+void buffer_sync_cleanup(void);
+
#endif /* OPROFILE_BUFFER_SYNC_H */
--- linux-2.6-for-ingo.orig/drivers/oprofile/oprof.c
+++ linux-2.6-for-ingo/drivers/oprofile/oprof.c
@@ -183,6 +183,10 @@ static int __init oprofile_init(void)
{
int err;
+ err = buffer_sync_init();
+ if (err)
+ return err;
+
err = oprofile_arch_init(&oprofile_ops);
if (err < 0 || timer) {
@@ -191,8 +195,10 @@ static int __init oprofile_init(void)
}
err = oprofilefs_register();
- if (err)
+ if (err) {
oprofile_arch_exit();
+ buffer_sync_cleanup();
+ }
return err;
}
@@ -202,6 +208,7 @@ static void __exit oprofile_exit(void)
{
oprofilefs_unregister();
oprofile_arch_exit();
+ buffer_sync_cleanup();
}
--- linux-2.6-for-ingo.orig/drivers/xen/manage.c
+++ linux-2.6-for-ingo/drivers/xen/manage.c
@@ -100,7 +100,7 @@ static void do_suspend(void)
/* XXX use normal device tree? */
xenbus_suspend();
- err = stop_machine(xen_suspend, &cancelled, &cpumask_of_cpu(0));
+ err = stop_machine(xen_suspend, &cancelled, cpumask_of(0));
if (err) {
printk(KERN_ERR "failed to start xen_suspend: %d\n", err);
goto out;
--- linux-2.6-for-ingo.orig/net/iucv/iucv.c
+++ linux-2.6-for-ingo/net/iucv/iucv.c
@@ -492,15 +492,14 @@ static void iucv_setmask_mp(void)
*
* Allow iucv interrupts on a single cpu.
*/
-static void iucv_setmask_up(void)
+static void iucv_setmask_up(struct cpumask *cpumask)
{
- cpumask_t cpumask;
int cpu;
/* Disable all cpu but the first in cpu_irq_cpumask. */
- cpumask = iucv_irq_cpumask;
- cpu_clear(first_cpu(iucv_irq_cpumask), cpumask);
- for_each_cpu_mask_nr(cpu, cpumask)
+ cpumask_copy(cpumask, iucv_irq_cpumask);
+ cpumask_clear_cpu(cpumask_first(iucv_irq_cpumask), cpumask);
+ for_each_cpu(cpu, cpumask)
smp_call_function_single(cpu, iucv_block_cpu, NULL, 1);
}
@@ -558,7 +557,7 @@ static void iucv_disable(void)
static int __cpuinit iucv_cpu_notify(struct notifier_block *self,
unsigned long action, void *hcpu)
{
- cpumask_t cpumask;
+ cpumask_var_t cpumask;
long cpu = (long) hcpu;
switch (action) {
@@ -593,15 +592,20 @@ static int __cpuinit iucv_cpu_notify(str
break;
case CPU_DOWN_PREPARE:
case CPU_DOWN_PREPARE_FROZEN:
- cpumask = iucv_buffer_cpumask;
- cpu_clear(cpu, cpumask);
- if (cpus_empty(cpumask))
+ if (!alloc_cpumask_var(&cpumask, GFP_KERNEL))
+ return NOTIFY_BAD;
+ cpumask_copy(cpumask, &iucv_buffer_cpumask);
+ cpumask_clear_cpu(cpu, cpumask);
+ if (cpumask_empty(cpumask)) {
/* Can't offline last IUCV enabled cpu. */
+ free_cpumask_var(cpumask);
return NOTIFY_BAD;
+ }
smp_call_function_single(cpu, iucv_retrieve_cpu, NULL, 1);
if (cpus_empty(iucv_irq_cpumask))
smp_call_function_single(first_cpu(iucv_buffer_cpumask),
iucv_allow_cpu, NULL, 1);
+ free_cpumask_var(cpumask);
break;
}
return NOTIFY_OK;
@@ -686,9 +690,12 @@ static void iucv_cleanup_queue(void)
int iucv_register(struct iucv_handler *handler, int smp)
{
int rc;
+ cpumask_var_t scratch;
if (!iucv_available)
return -ENOSYS;
+ if (!alloc_cpumask_var(&scratch, GFP_KERNEL))
+ return -ENOMEM;
mutex_lock(&iucv_register_mutex);
if (!smp)
iucv_nonsmp_handler++;
@@ -697,7 +704,7 @@ int iucv_register(struct iucv_handler *h
if (rc)
goto out_mutex;
} else if (!smp && iucv_nonsmp_handler == 1)
- iucv_setmask_up();
+ iucv_setmask_up(scratch);
INIT_LIST_HEAD(&handler->paths);
spin_lock_bh(&iucv_table_lock);
@@ -706,6 +713,7 @@ int iucv_register(struct iucv_handler *h
rc = 0;
out_mutex:
mutex_unlock(&iucv_register_mutex);
+ free_cpumask_var(scratch);
return rc;
}
EXPORT_SYMBOL(iucv_register);
--
^ permalink raw reply [flat|nested] 31+ messages in thread
* [PATCH 4/5] cpumask: convert drivers/net/sfc
2009-01-07 19:58 [PATCH 0/5] cpumask: more cpumask updates Mike Travis
` (2 preceding siblings ...)
2009-01-07 19:58 ` [PATCH 3/5] cpumask: convert misc driver functions Mike Travis
@ 2009-01-07 19:58 ` Mike Travis
2009-01-19 17:08 ` Ben Hutchings
2009-01-07 19:58 ` [PATCH 5/5] cpumask: convert other misc kernel functions Mike Travis
2009-01-07 20:52 ` [PATCH 0/5] cpumask: more cpumask updates Ingo Molnar
5 siblings, 1 reply; 31+ messages in thread
From: Mike Travis @ 2009-01-07 19:58 UTC (permalink / raw)
To: Ingo Molnar, Rusty Russell
Cc: Thomas Gleixner, H. Peter Anvin, Jack Steiner, linux-kernel,
Ben Hutchings, linux-net-drivers
[-- Attachment #1: cpumask:convert-drivers_net_sfc_efx.c.patch --]
[-- Type: text/plain, Size: 1390 bytes --]
From: Rusty Russell <rusty@rustcorp.com.au>
Impact: reduce stack usage, use new cpumask API.
Remove a cpumask from the stack. Ben Hutchings indicated that printing
a warning and returning 1 was acceptable for the corner case where allocation
fails.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: linux-net-drivers@solarflare.com
---
drivers/net/sfc/efx.c | 15 +++++++++++----
1 file changed, 11 insertions(+), 4 deletions(-)
diff --git a/drivers/net/sfc/efx.c b/drivers/net/sfc/efx.c
--- a/drivers/net/sfc/efx.c
+++ b/drivers/net/sfc/efx.c
@@ -854,20 +854,27 @@ static void efx_fini_io(struct efx_nic *
* interrupts across them. */
static int efx_wanted_rx_queues(void)
{
- cpumask_t core_mask;
+ cpumask_var_t core_mask;
int count;
int cpu;
- cpus_clear(core_mask);
+ if (!alloc_cpumask_var(&core_mask, GFP_KERNEL)) {
+ printk(KERN_WARNING
+ "efx.c: allocation failure, irq balancing hobbled\n");
+ return 1;
+ }
+
+ cpumask_clear(core_mask);
count = 0;
for_each_online_cpu(cpu) {
- if (!cpu_isset(cpu, core_mask)) {
+ if (!cpumask_test_cpu(cpu, core_mask)) {
++count;
- cpumask_or(&core_mask, &core_mask,
+ cpumask_or(core_mask, core_mask,
topology_core_cpumask(cpu));
}
}
+ free_cpumask_var(core_mask);
return count;
}
--
^ permalink raw reply [flat|nested] 31+ messages in thread
* [PATCH 5/5] cpumask: convert other misc kernel functions
2009-01-07 19:58 [PATCH 0/5] cpumask: more cpumask updates Mike Travis
` (3 preceding siblings ...)
2009-01-07 19:58 ` [PATCH 4/5] cpumask: convert drivers/net/sfc Mike Travis
@ 2009-01-07 19:58 ` Mike Travis
2009-01-07 20:52 ` [PATCH 0/5] cpumask: more cpumask updates Ingo Molnar
5 siblings, 0 replies; 31+ messages in thread
From: Mike Travis @ 2009-01-07 19:58 UTC (permalink / raw)
To: Ingo Molnar, Rusty Russell
Cc: Thomas Gleixner, H. Peter Anvin, Jack Steiner, linux-kernel
[-- Attachment #1: cpumask:convert-misc.patch --]
[-- Type: text/plain, Size: 709 bytes --]
From: Rusty Russell <rusty@rustcorp.com.au>
Impact: use new cpumask API.
Convert other misc kernel functions to use struct cpumask.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
---
lib/smp_processor_id.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- linux-2.6.28.orig/lib/smp_processor_id.c
+++ linux-2.6.28/lib/smp_processor_id.c
@@ -22,7 +22,7 @@ notrace unsigned int debug_smp_processor
* Kernel threads bound to a single CPU can safely use
* smp_processor_id():
*/
- if (cpus_equal(current->cpus_allowed, cpumask_of_cpu(this_cpu)))
+ if (cpumask_equal(¤t->cpus_allowed, cpumask_of(this_cpu)))
goto out;
/*
--
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t
2009-01-07 19:58 ` [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t Mike Travis
@ 2009-01-07 20:27 ` Yinghai Lu
2009-01-07 22:32 ` Mike Travis
0 siblings, 1 reply; 31+ messages in thread
From: Yinghai Lu @ 2009-01-07 20:27 UTC (permalink / raw)
To: Mike Travis
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, Chris Wright, Jeremy Fitzhardinge,
KOSAKI Motohiro, Venkatesh Pallipadi, virtualization, xen-devel
On Wed, Jan 7, 2009 at 11:58 AM, Mike Travis <travis@sgi.com> wrote:
| --- linux-2.6-for-ingo.orig/kernel/irq/numa_migrate.c
| +++ linux-2.6-for-ingo/kernel/irq/numa_migrate.c
| @@ -46,6 +46,7 @@ static void init_copy_one_irq_desc(int i
| desc->cpu = cpu;
| lockdep_set_class(&desc->lock, &irq_desc_lock_class);
| init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids);
|+ init_copy_desc_masks(old_desc, desc);
| arch_init_copy_chip_data(old_desc, desc, cpu);
| }
|
|@@ -76,11 +77,20 @@ static struct irq_desc *__real_move_irq_
| node = cpu_to_node(cpu);
| desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node);
| if (!desc) {
|- printk(KERN_ERR "irq %d: can not get new irq_desc for migration.\n", irq);
|+ printk(KERN_ERR "irq %d: can not get new irq_desc "
|+ "for migration.\n", irq);
| /* still use old one */
| desc = old_desc;
| goto out_unlock;
| }
|+ if (!init_alloc_desc_masks(desc, node, false)) {
|+ printk(KERN_ERR "irq %d: can not get new irq_desc cpumask "
|+ "for migration.\n", irq);
|+ /* still use old one */
|+ kfree(desc);
|+ desc = old_desc;
|+ goto out_unlock;
|+ }
| init_copy_one_irq_desc(irq, old_desc, desc, cpu);
desc new mask_var (allocated) aka the pointer is overwritten here...
you may need to calling move init_alloc_desc_masks() into
init_copy_one_irq_desc()
| irq_desc_ptrs[irq] = desc;
YH
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 0/5] cpumask: more cpumask updates
2009-01-07 19:58 [PATCH 0/5] cpumask: more cpumask updates Mike Travis
` (4 preceding siblings ...)
2009-01-07 19:58 ` [PATCH 5/5] cpumask: convert other misc kernel functions Mike Travis
@ 2009-01-07 20:52 ` Ingo Molnar
2009-01-07 21:20 ` Mike Travis
5 siblings, 1 reply; 31+ messages in thread
From: Ingo Molnar @ 2009-01-07 20:52 UTC (permalink / raw)
To: Mike Travis
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel
* Mike Travis <travis@sgi.com> wrote:
>
> One big change to irq_desc and a few smaller maintenance type updates.
>
> cpumask: update irq_desc to use cpumask_var_t
> cpumask: Use topology_core_cpumask()/topology_thread_cpumask()
> cpumask: convert misc driver functions
> cpumask: convert drivers/net/sfc
> cpumask: convert other misc kernel functions
>
> Based on tip/cpus4096: v2.6.28-5703-g1d1a70e
>
> (Ingo - I added a blank line before the From: part and that seems to
> separate it from the Subject. Hopefully your git-am will still
> recognizes it - I did try a git-am here and it seemed to work ok.)
>
> Signed-off-by: Mike Travis <travis@sgi.com>
Since you have it in a Git tree, why dont you send me a pull request?
I might be more picky about pull requests in general (because it's harder
to fix up a pulled tree than to fix up commits that come from emails - and
it's also more work for you if i refuse to pull than if i have some
comments over email-patches), but it's a much better (more scalable)
workflow in the long run than the email path.
Ingo
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 0/5] cpumask: more cpumask updates
2009-01-07 20:52 ` [PATCH 0/5] cpumask: more cpumask updates Ingo Molnar
@ 2009-01-07 21:20 ` Mike Travis
0 siblings, 0 replies; 31+ messages in thread
From: Mike Travis @ 2009-01-07 21:20 UTC (permalink / raw)
To: Ingo Molnar
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel
Ingo Molnar wrote:
> * Mike Travis <travis@sgi.com> wrote:
>
>> One big change to irq_desc and a few smaller maintenance type updates.
>>
>> cpumask: update irq_desc to use cpumask_var_t
>> cpumask: Use topology_core_cpumask()/topology_thread_cpumask()
>> cpumask: convert misc driver functions
>> cpumask: convert drivers/net/sfc
>> cpumask: convert other misc kernel functions
>>
>> Based on tip/cpus4096: v2.6.28-5703-g1d1a70e
>>
>> (Ingo - I added a blank line before the From: part and that seems to
>> separate it from the Subject. Hopefully your git-am will still
>> recognizes it - I did try a git-am here and it seemed to work ok.)
>>
>> Signed-off-by: Mike Travis <travis@sgi.com>
>
> Since you have it in a Git tree, why dont you send me a pull request?
>
> I might be more picky about pull requests in general (because it's harder
> to fix up a pulled tree than to fix up commits that come from emails - and
> it's also more work for you if i refuse to pull than if i have some
> comments over email-patches), but it's a much better (more scalable)
> workflow in the long run than the email path.
>
> Ingo
Ok, will do.
Thanks,
Mike
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t
2009-01-07 20:27 ` Yinghai Lu
@ 2009-01-07 22:32 ` Mike Travis
2009-01-08 3:15 ` Yinghai Lu
0 siblings, 1 reply; 31+ messages in thread
From: Mike Travis @ 2009-01-07 22:32 UTC (permalink / raw)
To: Yinghai Lu
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, Chris Wright, Jeremy Fitzhardinge,
KOSAKI Motohiro, Venkatesh Pallipadi, virtualization, xen-devel
Yinghai Lu wrote:
> On Wed, Jan 7, 2009 at 11:58 AM, Mike Travis <travis@sgi.com> wrote:
> | --- linux-2.6-for-ingo.orig/kernel/irq/numa_migrate.c
> | +++ linux-2.6-for-ingo/kernel/irq/numa_migrate.c
> | @@ -46,6 +46,7 @@ static void init_copy_one_irq_desc(int i
> | desc->cpu = cpu;
> | lockdep_set_class(&desc->lock, &irq_desc_lock_class);
> | init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids);
> |+ init_copy_desc_masks(old_desc, desc);
> | arch_init_copy_chip_data(old_desc, desc, cpu);
> | }
> |
> |@@ -76,11 +77,20 @@ static struct irq_desc *__real_move_irq_
> | node = cpu_to_node(cpu);
> | desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node);
> | if (!desc) {
> |- printk(KERN_ERR "irq %d: can not get new irq_desc for migration.\n", irq);
> |+ printk(KERN_ERR "irq %d: can not get new irq_desc "
> |+ "for migration.\n", irq);
> | /* still use old one */
> | desc = old_desc;
> | goto out_unlock;
> | }
> |+ if (!init_alloc_desc_masks(desc, node, false)) {
> |+ printk(KERN_ERR "irq %d: can not get new irq_desc cpumask "
> |+ "for migration.\n", irq);
> |+ /* still use old one */
> |+ kfree(desc);
> |+ desc = old_desc;
> |+ goto out_unlock;
> |+ }
> | init_copy_one_irq_desc(irq, old_desc, desc, cpu);
>
> desc new mask_var (allocated) aka the pointer is overwritten here...
> you may need to calling move init_alloc_desc_masks() into
> init_copy_one_irq_desc()
Wouldn't this in init_copy_one_irq_desc() take care of that?
@@ -46,6 +46,7 @@ static void init_copy_one_irq_desc(int i
desc->cpu = cpu;
lockdep_set_class(&desc->lock, &irq_desc_lock_class);
init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids);
+ init_copy_desc_masks(old_desc, desc);
arch_init_copy_chip_data(old_desc, desc, cpu);
where:
static inline void init_copy_desc_masks(struct irq_desc *old_desc,
struct irq_desc *new_desc)
{
#ifdef CONFIG_CPUMASKS_OFFSTACK
cpumask_copy(new_desc->affinity, old_desc->affinity);
#ifdef CONFIG_GENERIC_PENDING_IRQ
cpumask_copy(new_desc->pending_mask, old_desc->pending_mask);
#endif
#endif
}
In other words if the masks are not a cpumask[1] but instead a
*cpumask pointer, then the old masks are copied to the new desc/
Or am I missing your point?
Thanks!
Mike
>
> | irq_desc_ptrs[irq] = desc;
>
> YH
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t
2009-01-07 22:32 ` Mike Travis
@ 2009-01-08 3:15 ` Yinghai Lu
2009-01-08 15:45 ` Mike Travis
2009-01-08 18:40 ` [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t: fix Mike Travis
0 siblings, 2 replies; 31+ messages in thread
From: Yinghai Lu @ 2009-01-08 3:15 UTC (permalink / raw)
To: Mike Travis
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, Chris Wright, Jeremy Fitzhardinge,
KOSAKI Motohiro, Venkatesh Pallipadi, virtualization, xen-devel
Mike Travis wrote:
> Yinghai Lu wrote:
>> On Wed, Jan 7, 2009 at 11:58 AM, Mike Travis <travis@sgi.com> wrote:
>> | --- linux-2.6-for-ingo.orig/kernel/irq/numa_migrate.c
>> | +++ linux-2.6-for-ingo/kernel/irq/numa_migrate.c
>> | @@ -46,6 +46,7 @@ static void init_copy_one_irq_desc(int i
>> | desc->cpu = cpu;
>> | lockdep_set_class(&desc->lock, &irq_desc_lock_class);
>> | init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids);
>> |+ init_copy_desc_masks(old_desc, desc);
>> | arch_init_copy_chip_data(old_desc, desc, cpu);
>> | }
>> |
>> |@@ -76,11 +77,20 @@ static struct irq_desc *__real_move_irq_
>> | node = cpu_to_node(cpu);
>> | desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node);
>> | if (!desc) {
>> |- printk(KERN_ERR "irq %d: can not get new irq_desc for migration.\n", irq);
>> |+ printk(KERN_ERR "irq %d: can not get new irq_desc "
>> |+ "for migration.\n", irq);
>> | /* still use old one */
>> | desc = old_desc;
>> | goto out_unlock;
>> | }
>> |+ if (!init_alloc_desc_masks(desc, node, false)) {
>> |+ printk(KERN_ERR "irq %d: can not get new irq_desc cpumask "
>> |+ "for migration.\n", irq);
>> |+ /* still use old one */
>> |+ kfree(desc);
>> |+ desc = old_desc;
>> |+ goto out_unlock;
>> |+ }
>> | init_copy_one_irq_desc(irq, old_desc, desc, cpu);
>>
>> desc new mask_var (allocated) aka the pointer is overwritten here...
>> you may need to calling move init_alloc_desc_masks() into
>> init_copy_one_irq_desc()
>
> Wouldn't this in init_copy_one_irq_desc() take care of that?
>
> @@ -46,6 +46,7 @@ static void init_copy_one_irq_desc(int i
> desc->cpu = cpu;
> lockdep_set_class(&desc->lock, &irq_desc_lock_class);
> init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids);
> + init_copy_desc_masks(old_desc, desc);
> arch_init_copy_chip_data(old_desc, desc, cpu);
>
> where:
>
> static inline void init_copy_desc_masks(struct irq_desc *old_desc,
> struct irq_desc *new_desc)
> {
> #ifdef CONFIG_CPUMASKS_OFFSTACK
> cpumask_copy(new_desc->affinity, old_desc->affinity);
>
> #ifdef CONFIG_GENERIC_PENDING_IRQ
> cpumask_copy(new_desc->pending_mask, old_desc->pending_mask);
> #endif
> #endif
> }
>
> In other words if the masks are not a cpumask[1] but instead a
> *cpumask pointer, then the old masks are copied to the new desc/
>
> Or am I missing your point?
static void init_copy_one_irq_desc(int irq, struct irq_desc *old_desc,
struct irq_desc *desc, int cpu)
{
memcpy(desc, old_desc, sizeof(struct irq_desc));
will overwrite new_desc->affinity and pending_mask
YH
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t
2009-01-08 3:15 ` Yinghai Lu
@ 2009-01-08 15:45 ` Mike Travis
2009-01-08 19:31 ` Yinghai Lu
2009-01-08 18:40 ` [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t: fix Mike Travis
1 sibling, 1 reply; 31+ messages in thread
From: Mike Travis @ 2009-01-08 15:45 UTC (permalink / raw)
To: Yinghai Lu
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, Chris Wright, Jeremy Fitzhardinge,
KOSAKI Motohiro, Venkatesh Pallipadi, virtualization, xen-devel
Yinghai Lu wrote:
...
>> Or am I missing your point?
>
> static void init_copy_one_irq_desc(int irq, struct irq_desc *old_desc,
> struct irq_desc *desc, int cpu)
> {
> memcpy(desc, old_desc, sizeof(struct irq_desc));
>
>
> will overwrite new_desc->affinity and pending_mask
>
> YH
Yup, apparently I was missing your point (banging on head once more).
Thanks for spotting this!
Ingo - what's the state of tip/cpus4096? Shall I push this as an
append patch, or redo the original so bisectability works.
Also, Yinghai - would you know of a straight forward way to test
the irq migration? I have taken cpus offline and back online
but wasn't sure if any irq's were being moved off of cpu 0 which
cannot be taken offline.
Thanks,
Mike
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t: fix
2009-01-08 3:15 ` Yinghai Lu
2009-01-08 15:45 ` Mike Travis
@ 2009-01-08 18:40 ` Mike Travis
1 sibling, 0 replies; 31+ messages in thread
From: Mike Travis @ 2009-01-08 18:40 UTC (permalink / raw)
To: Yinghai Lu; +Cc: Ingo Molnar, Rusty Russell, Jack Steiner, linux-kernel
Yinghai Lu wrote:
...
> static void init_copy_one_irq_desc(int irq, struct irq_desc *old_desc,
> struct irq_desc *desc, int cpu)
> {
> memcpy(desc, old_desc, sizeof(struct irq_desc));
>
>
> will overwrite new_desc->affinity and pending_mask
>
> YH
I've appended this to the patchset:
Subject: cpumask: fix bug in use cpumask_var_t in irq_desc
Impact: fix bug where new irq_desc uses old cpumask pointers which are freed.
As Yinghai pointed out, init_copy_one_irq_desc() copies the old desc to
the new desc overwriting the cpumask pointers. Since the old_desc and
the cpumask pointers are freed, then memory corruption will occur if
these old pointers are used.
Move the allocation of these pointers to after the copy.
Signed-off-by: Mike Travis <travis@sgi.com>
Cc: Yinghai Lu <yinghai@kernel.org>
---
include/linux/irq.h | 9 +++++++--
kernel/irq/handle.c | 8 +-------
kernel/irq/numa_migrate.c | 13 ++++++++-----
3 files changed, 16 insertions(+), 14 deletions(-)
--- linux-2.6-for-ingo.orig/include/linux/irq.h
+++ linux-2.6-for-ingo/include/linux/irq.h
@@ -426,15 +426,18 @@ extern int set_irq_msi(unsigned int irq,
/**
* init_alloc_desc_masks - allocate cpumasks for irq_desc
* @desc: pointer to irq_desc struct
+ * @cpu: cpu which will be handling the cpumasks
* @boot: true if need bootmem
*
* Allocates affinity and pending_mask cpumask if required.
* Returns true if successful (or not required).
* Side effect: affinity has all bits set, pending_mask has all bits clear.
*/
-static inline bool init_alloc_desc_masks(struct irq_desc *desc, int node,
+static inline bool init_alloc_desc_masks(struct irq_desc *desc, int cpu,
bool boot)
{
+ int node;
+
if (boot) {
alloc_bootmem_cpumask_var(&desc->affinity);
cpumask_setall(desc->affinity);
@@ -446,6 +449,8 @@ static inline bool init_alloc_desc_masks
return true;
}
+ node = cpu_to_node(cpu);
+
if (!alloc_cpumask_var_node(&desc->affinity, GFP_ATOMIC, node))
return false;
cpumask_setall(desc->affinity);
@@ -484,7 +489,7 @@ static inline void init_copy_desc_masks(
#else /* !CONFIG_SMP */
-static inline bool init_alloc_desc_masks(struct irq_desc *desc, int node,
+static inline bool init_alloc_desc_masks(struct irq_desc *desc, int cpu,
bool boot)
{
return true;
--- linux-2.6-for-ingo.orig/kernel/irq/handle.c
+++ linux-2.6-for-ingo/kernel/irq/handle.c
@@ -85,8 +85,6 @@ void init_kstat_irqs(struct irq_desc *de
static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu)
{
- int node = cpu_to_node(cpu);
-
memcpy(desc, &irq_desc_init, sizeof(struct irq_desc));
spin_lock_init(&desc->lock);
@@ -100,7 +98,7 @@ static void init_one_irq_desc(int irq, s
printk(KERN_ERR "can not alloc kstat_irqs\n");
BUG_ON(1);
}
- if (!init_alloc_desc_masks(desc, node, false)) {
+ if (!init_alloc_desc_masks(desc, cpu, false)) {
printk(KERN_ERR "can not alloc irq_desc cpumasks\n");
BUG_ON(1);
}
@@ -188,10 +186,6 @@ struct irq_desc *irq_to_desc_alloc_cpu(u
printk(KERN_ERR "can not alloc irq_desc\n");
BUG_ON(1);
}
- if (!init_alloc_desc_masks(desc, node, false)) {
- printk(KERN_ERR "can not alloc irq_desc cpumasks\n");
- BUG_ON(1);
- }
init_one_irq_desc(irq, desc, cpu);
irq_desc_ptrs[irq] = desc;
--- linux-2.6-for-ingo.orig/kernel/irq/numa_migrate.c
+++ linux-2.6-for-ingo/kernel/irq/numa_migrate.c
@@ -38,16 +38,22 @@ static void free_kstat_irqs(struct irq_d
old_desc->kstat_irqs = NULL;
}
-static void init_copy_one_irq_desc(int irq, struct irq_desc *old_desc,
+static bool init_copy_one_irq_desc(int irq, struct irq_desc *old_desc,
struct irq_desc *desc, int cpu)
{
memcpy(desc, old_desc, sizeof(struct irq_desc));
+ if (!init_alloc_desc_masks(desc, cpu, false)) {
+ printk(KERN_ERR "irq %d: can not get new irq_desc cpumask "
+ "for migration.\n", irq);
+ return false;
+ }
spin_lock_init(&desc->lock);
desc->cpu = cpu;
lockdep_set_class(&desc->lock, &irq_desc_lock_class);
init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids);
init_copy_desc_masks(old_desc, desc);
arch_init_copy_chip_data(old_desc, desc, cpu);
+ return true;
}
static void free_one_irq_desc(struct irq_desc *old_desc, struct irq_desc *desc)
@@ -83,15 +89,12 @@ static struct irq_desc *__real_move_irq_
desc = old_desc;
goto out_unlock;
}
- if (!init_alloc_desc_masks(desc, node, false)) {
- printk(KERN_ERR "irq %d: can not get new irq_desc cpumask "
- "for migration.\n", irq);
+ if (!init_copy_one_irq_desc(irq, old_desc, desc, cpu)) {
/* still use old one */
kfree(desc);
desc = old_desc;
goto out_unlock;
}
- init_copy_one_irq_desc(irq, old_desc, desc, cpu);
irq_desc_ptrs[irq] = desc;
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t
2009-01-08 15:45 ` Mike Travis
@ 2009-01-08 19:31 ` Yinghai Lu
0 siblings, 0 replies; 31+ messages in thread
From: Yinghai Lu @ 2009-01-08 19:31 UTC (permalink / raw)
To: Mike Travis
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, Chris Wright, Jeremy Fitzhardinge,
KOSAKI Motohiro, Venkatesh Pallipadi, virtualization, xen-devel
Mike Travis wrote:
> Yinghai Lu wrote:
> ...
>>> Or am I missing your point?
>> static void init_copy_one_irq_desc(int irq, struct irq_desc *old_desc,
>> struct irq_desc *desc, int cpu)
>> {
>> memcpy(desc, old_desc, sizeof(struct irq_desc));
>>
>>
>> will overwrite new_desc->affinity and pending_mask
>>
>> YH
>
> Yup, apparently I was missing your point (banging on head once more).
>
> Thanks for spotting this!
>
> Ingo - what's the state of tip/cpus4096? Shall I push this as an
> append patch, or redo the original so bisectability works.
>
> Also, Yinghai - would you know of a straight forward way to test
> the irq migration? I have taken cpus offline and back online
> but wasn't sure if any irq's were being moved off of cpu 0 which
> cannot be taken offline.
use echo "mask" > /proc/irq/xxx/smp_affinity to change affinity
and check /proc/interrupts and in /sys/log/messages i will said allocate new irq desc on some other node/cpu.
YH
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 3/5] cpumask: convert misc driver functions
2009-01-07 19:58 ` [PATCH 3/5] cpumask: convert misc driver functions Mike Travis
@ 2009-01-10 10:57 ` Rusty Russell
2009-01-16 18:42 ` Tony Luck
2009-01-22 13:11 ` Robert Richter
2 siblings, 0 replies; 31+ messages in thread
From: Rusty Russell @ 2009-01-10 10:57 UTC (permalink / raw)
To: Mike Travis
Cc: Ingo Molnar, Thomas Gleixner, H. Peter Anvin, Jack Steiner,
linux-kernel, Dean Nelson, Robert Richter, oprofile-list,
Jeremy Fitzhardinge, Chris Wright, virtualization, xen-devel,
Ursula Braun, linux390, linux-s390
On Thursday 08 January 2009 06:28:35 Mike Travis wrote:
> From: Rusty Russell <rusty@rustcorp.com.au>
>
> Impact: use new cpumask API.
>
> Convert misc driver functions to use struct cpumask.
...
> @@ -593,15 +592,20 @@ static int __cpuinit iucv_cpu_notify(str
> break;
> case CPU_DOWN_PREPARE:
> case CPU_DOWN_PREPARE_FROZEN:
> - cpumask = iucv_buffer_cpumask;
> - cpu_clear(cpu, cpumask);
> - if (cpus_empty(cpumask))
> + if (!alloc_cpumask_var(&cpumask, GFP_KERNEL))
> + return NOTIFY_BAD;
> + cpumask_copy(cpumask, &iucv_buffer_cpumask);
> + cpumask_clear_cpu(cpu, cpumask);
> + if (cpumask_empty(cpumask)) {
> /* Can't offline last IUCV enabled cpu. */
> + free_cpumask_var(cpumask);
> return NOTIFY_BAD;
> + }
> smp_call_function_single(cpu, iucv_retrieve_cpu, NULL, 1);
> if (cpus_empty(iucv_irq_cpumask))
> smp_call_function_single(first_cpu(iucv_buffer_cpumask),
> iucv_allow_cpu, NULL, 1);
> + free_cpumask_var(cpumask);
> break;
Minor nit: this can be done better, actually:
@@ -596,9 +594,7 @@ static int __cpuinit iucv_cpu_notify(str
case CPU_DOWN_PREPARE_FROZEN:
if (!iucv_path_table)
break;
- cpumask = iucv_buffer_cpumask;
- cpu_clear(cpu, cpumask);
- if (cpus_empty(cpumask))
+ if (cpumask_any_but(&iucv_buffer_cpumask, cpu) >= nr_cpu_ids)
/* Can't offline last IUCV enabled cpu. */
return NOTIFY_BAD;
smp_call_function_single(cpu, iucv_retrieve_cpu, NULL, 1);
Cheers,
Rusty.
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 3/5] cpumask: convert misc driver functions
2009-01-07 19:58 ` [PATCH 3/5] cpumask: convert misc driver functions Mike Travis
2009-01-10 10:57 ` Rusty Russell
@ 2009-01-16 18:42 ` Tony Luck
2009-01-16 18:55 ` Mike Travis
2009-01-16 22:02 ` Mike Travis
2009-01-22 13:11 ` Robert Richter
2 siblings, 2 replies; 31+ messages in thread
From: Tony Luck @ 2009-01-16 18:42 UTC (permalink / raw)
To: Mike Travis
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, Dean Nelson, Robert Richter,
oprofile-list, Jeremy Fitzhardinge, Chris Wright, virtualization,
xen-devel, Ursula Braun, linux390, linux-s390
+void __exit buffer_sync_cleanup(void)
+{
+ free_cpumask_var(marked_cpus);
+}
This breaks ia64 with CONFIG_OPROFILE=y
`buffer_sync_cleanup' referenced in section `.init.text' of
arch/ia64/oprofile/built-in.o: defined in discarded section
`.exit.text' of arch/ia64/oprofile/built-in.o
make: *** [.tmp_vmlinux1] Error 1
-Tony
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 3/5] cpumask: convert misc driver functions
2009-01-16 18:42 ` Tony Luck
@ 2009-01-16 18:55 ` Mike Travis
2009-01-16 22:02 ` Mike Travis
1 sibling, 0 replies; 31+ messages in thread
From: Mike Travis @ 2009-01-16 18:55 UTC (permalink / raw)
To: Tony Luck
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, Dean Nelson, Robert Richter,
oprofile-list, Jeremy Fitzhardinge, Chris Wright, virtualization,
xen-devel, Ursula Braun, linux390, linux-s390
Tony Luck wrote:
> +void __exit buffer_sync_cleanup(void)
> +{
> + free_cpumask_var(marked_cpus);
> +}
>
> This breaks ia64 with CONFIG_OPROFILE=y
>
> `buffer_sync_cleanup' referenced in section `.init.text' of
> arch/ia64/oprofile/built-in.o: defined in discarded section
> `.exit.text' of arch/ia64/oprofile/built-in.o
> make: *** [.tmp_vmlinux1] Error 1
>
> -Tony
Hi Tony,
Hmm, that is a tricky error. I'll see what can be done to
eliminate it.
Thanks!
Mike
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 3/5] cpumask: convert misc driver functions
2009-01-16 18:42 ` Tony Luck
2009-01-16 18:55 ` Mike Travis
@ 2009-01-16 22:02 ` Mike Travis
1 sibling, 0 replies; 31+ messages in thread
From: Mike Travis @ 2009-01-16 22:02 UTC (permalink / raw)
To: Tony Luck
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, Dean Nelson, Robert Richter,
oprofile-list
Tony Luck wrote:
> +void __exit buffer_sync_cleanup(void)
> +{
> + free_cpumask_var(marked_cpus);
> +}
>
> This breaks ia64 with CONFIG_OPROFILE=y
>
> `buffer_sync_cleanup' referenced in section `.init.text' of
> arch/ia64/oprofile/built-in.o: defined in discarded section
> `.exit.text' of arch/ia64/oprofile/built-in.o
> make: *** [.tmp_vmlinux1] Error 1
>
> -Tony
Hi,
It's a bit hackish but I couldn't think of another workaround.
And curiously, it doesn't affect x86?
Also, CPUMASK_OFFSTACK=n for ia64 so this is really a nop.
(I haven't been able to turn it on for ia64.)
Thanks,
Mike
---
Subject: ia64: fix build error when OPROFILE enabled.
Impact: Fix build error.
Add a buffer_sync_cleanup_init() which duplicates the
buffer_sync_cleanup() function but is placed in the
.init.text section. This eliminates this build error:
`buffer_sync_cleanup' referenced in section `.init.text' of
arch/ia64/oprofile/built-in.o: defined in discarded section
`.exit.text' of arch/ia64/oprofile/built-in.o
Signed-off-by: Mike Travis <travis@sgi.com>
Cc: Tony Luck <tony.luck@gmail.com>
Cc: Robert Richter <robert.richter@amd.com>
Cc: <oprofile-list@lists.sf.net>
---
drivers/oprofile/buffer_sync.c | 6 ++++++
drivers/oprofile/buffer_sync.h | 1 +
drivers/oprofile/oprof.c | 2 +-
3 files changed, 8 insertions(+), 1 deletion(-)
--- linux-2.6-for-ingo.orig/drivers/oprofile/buffer_sync.c
+++ linux-2.6-for-ingo/drivers/oprofile/buffer_sync.c
@@ -579,6 +579,12 @@ void __exit buffer_sync_cleanup(void)
free_cpumask_var(marked_cpus);
}
+/* same as above but in the __init section */
+void __init buffer_sync_cleanup_init(void)
+{
+ free_cpumask_var(marked_cpus);
+}
+
/* The function can be used to add a buffer worth of data directly to
* the kernel buffer. The buffer is assumed to be a circular buffer.
* Take the entries from index start and end at index end, wrapping
--- linux-2.6-for-ingo.orig/drivers/oprofile/buffer_sync.h
+++ linux-2.6-for-ingo/drivers/oprofile/buffer_sync.h
@@ -22,5 +22,6 @@ void sync_buffer(int cpu);
/* initialize/destroy the buffer system. */
int buffer_sync_init(void);
void buffer_sync_cleanup(void);
+void buffer_sync_cleanup_init(void);
#endif /* OPROFILE_BUFFER_SYNC_H */
--- linux-2.6-for-ingo.orig/drivers/oprofile/oprof.c
+++ linux-2.6-for-ingo/drivers/oprofile/oprof.c
@@ -197,7 +197,7 @@ static int __init oprofile_init(void)
err = oprofilefs_register();
if (err) {
oprofile_arch_exit();
- buffer_sync_cleanup();
+ buffer_sync_cleanup_init();
}
return err;
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 4/5] cpumask: convert drivers/net/sfc
2009-01-07 19:58 ` [PATCH 4/5] cpumask: convert drivers/net/sfc Mike Travis
@ 2009-01-19 17:08 ` Ben Hutchings
[not found] ` <200901201209.34854.rusty@rustcorp.com.au>
0 siblings, 1 reply; 31+ messages in thread
From: Ben Hutchings @ 2009-01-19 17:08 UTC (permalink / raw)
To: Mike Travis
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, linux-net-drivers
On Wed, 2009-01-07 at 11:58 -0800, Mike Travis wrote:
> From: Rusty Russell <rusty@rustcorp.com.au>
>
> Impact: reduce stack usage, use new cpumask API.
>
> Remove a cpumask from the stack. Ben Hutchings indicated that printing
> a warning and returning 1 was acceptable for the corner case where allocation
> fails.
>
> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
> Signed-off-by: Mike Travis <travis@sgi.com>
> Cc: Ben Hutchings <bhutchings@solarflare.com>
> Cc: linux-net-drivers@solarflare.com
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
This seems to work.
> ---
> drivers/net/sfc/efx.c | 15 +++++++++++----
> 1 file changed, 11 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/net/sfc/efx.c b/drivers/net/sfc/efx.c
> --- a/drivers/net/sfc/efx.c
> +++ b/drivers/net/sfc/efx.c
> @@ -854,20 +854,27 @@ static void efx_fini_io(struct efx_nic *
> * interrupts across them. */
> static int efx_wanted_rx_queues(void)
> {
> - cpumask_t core_mask;
> + cpumask_var_t core_mask;
> int count;
> int cpu;
>
> - cpus_clear(core_mask);
> + if (!alloc_cpumask_var(&core_mask, GFP_KERNEL)) {
> + printk(KERN_WARNING
> + "efx.c: allocation failure, irq balancing hobbled\n");
[...]
This is an exceedingly unlikely error case so the error message is not
that important, but if you don't mind re-spinning this then please make
it "sfc: RSS disabled due to allocation failure\n".
(Device-related log messages should also include the device address, but
I don't want to add a device parameter to the function just for that.
Maybe we should call this function just once at module init time since
the result is not device-specific.)
Ben.
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 2/5] cpumask: Use topology_core_cpumask()/topology_thread_cpumask()
2009-01-07 19:58 ` [PATCH 2/5] cpumask: Use topology_core_cpumask()/topology_thread_cpumask() Mike Travis
@ 2009-01-19 17:11 ` Ben Hutchings
0 siblings, 0 replies; 31+ messages in thread
From: Ben Hutchings @ 2009-01-19 17:11 UTC (permalink / raw)
To: Mike Travis
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, linux-net-drivers
On Wed, 2009-01-07 at 11:58 -0800, Mike Travis wrote:
> From: Rusty Russell <rusty@rustcorp.com.au>
>
> Impact: reduce stack usage, use new cpumask API.
>
> This actually uses topology_core_cpumask() and
> topology_thread_cpumask(), removing the only users of
> topology_core_siblings() and topology_thread_siblings()
>
> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
> Signed-off-by: Mike Travis <travis@sgi.com>
> Cc: linux-net-drivers@solarflare.com
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
Ben.
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 4/5] cpumask: convert drivers/net/sfc [PATCH supplied]
[not found] ` <200901201209.34854.rusty@rustcorp.com.au>
@ 2009-01-20 18:26 ` Mike Travis
2009-01-26 15:04 ` [PATCH] sfc: modify allocation error message Mike Travis
1 sibling, 0 replies; 31+ messages in thread
From: Mike Travis @ 2009-01-20 18:26 UTC (permalink / raw)
To: Rusty Russell
Cc: Ben Hutchings, Ingo Molnar, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, linux-net-drivers
Rusty Russell wrote:
> On Tuesday 20 January 2009 03:38:30 Ben Hutchings wrote:
>> On Wed, 2009-01-07 at 11:58 -0800, Mike Travis wrote:
>>> From: Rusty Russell <rusty@rustcorp.com.au>
>>>
>>> Impact: reduce stack usage, use new cpumask API.
>>>
>>> Remove a cpumask from the stack. Ben Hutchings indicated that printing
>>> a warning and returning 1 was acceptable for the corner case where allocation
>>> fails.
>>>
>>> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
>>> Signed-off-by: Mike Travis <travis@sgi.com>
>>> Cc: Ben Hutchings <bhutchings@solarflare.com>
>>> Cc: linux-net-drivers@solarflare.com
>> Acked-by: Ben Hutchings <bhutchings@solarflare.com>
>>
>> This seems to work.
>>
>>> ---
>>> drivers/net/sfc/efx.c | 15 +++++++++++----
>>> 1 file changed, 11 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/drivers/net/sfc/efx.c b/drivers/net/sfc/efx.c
>>> --- a/drivers/net/sfc/efx.c
>>> +++ b/drivers/net/sfc/efx.c
>>> @@ -854,20 +854,27 @@ static void efx_fini_io(struct efx_nic *
>>> * interrupts across them. */
>>> static int efx_wanted_rx_queues(void)
>>> {
>>> - cpumask_t core_mask;
>>> + cpumask_var_t core_mask;
>>> int count;
>>> int cpu;
>>>
>>> - cpus_clear(core_mask);
>>> + if (!alloc_cpumask_var(&core_mask, GFP_KERNEL)) {
>>> + printk(KERN_WARNING
>>> + "efx.c: allocation failure, irq balancing hobbled\n");
>> [...]
>>
>> This is an exceedingly unlikely error case so the error message is not
>> that important, but if you don't mind re-spinning this then please make
>> it "sfc: RSS disabled due to allocation failure\n".
>
> OK, I'm away from all my machines, at linux.conf.au. Mike, can you change
> this if you haven't already?
>
> Thanks,
> Rusty.
>
Subject: sfc: modify allocation error message
Impact: change error message when alloc_cpumask_var fails.
Signed-off-by: Mike Travis <travis@sgi.com>
---
drivers/net/sfc/efx.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
--- linux-2.6-for-ingo.orig/drivers/net/sfc/efx.c
+++ linux-2.6-for-ingo/drivers/net/sfc/efx.c
@@ -858,9 +858,9 @@ static int efx_wanted_rx_queues(void)
int count;
int cpu;
- if (!alloc_cpumask_var(&core_mask, GFP_KERNEL)) {
+ if (unlikely(!alloc_cpumask_var(&core_mask, GFP_KERNEL))) {
printk(KERN_WARNING
- "efx.c: allocation failure, irq balancing hobbled\n");
+ "sfc: RSS disabled due to allocation failure\n");
return 1;
}
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH 3/5] cpumask: convert misc driver functions
2009-01-07 19:58 ` [PATCH 3/5] cpumask: convert misc driver functions Mike Travis
2009-01-10 10:57 ` Rusty Russell
2009-01-16 18:42 ` Tony Luck
@ 2009-01-22 13:11 ` Robert Richter
2009-01-22 13:14 ` [PATCH] cpumask: modifiy oprofile initialization Robert Richter
2 siblings, 1 reply; 31+ messages in thread
From: Robert Richter @ 2009-01-22 13:11 UTC (permalink / raw)
To: Mike Travis
Cc: Ingo Molnar, Rusty Russell, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, Dean Nelson, oprofile-list,
Jeremy Fitzhardinge, Chris Wright, virtualization, xen-devel,
Ursula Braun, linux390, linux-s390
On 07.01.09 11:58:35, Mike Travis wrote:
> From: Rusty Russell <rusty@rustcorp.com.au>
>
> Impact: use new cpumask API.
>
> Convert misc driver functions to use struct cpumask.
>
> To Do:
> - Convert iucv_buffer_cpumask to cpumask_var_t.
>
> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
> Signed-off-by: Mike Travis <travis@sgi.com>
> Acked-by: Dean Nelson <dcn@sgi.com>
> Cc: Robert Richter <robert.richter@amd.com>
> Cc: oprofile-list@lists.sf.net
> Cc: Jeremy Fitzhardinge <jeremy@xensource.com>
> Cc: Chris Wright <chrisw@sous-sol.org>
> Cc: virtualization@lists.osdl.org
> Cc: xen-devel@lists.xensource.com
> Cc: Ursula Braun <ursula.braun@de.ibm.com>
> Cc: linux390@de.ibm.com
> Cc: linux-s390@vger.kernel.org
> ---
> drivers/base/cpu.c | 2 +-
> drivers/misc/sgi-xp/xpc_main.c | 2 +-
> drivers/oprofile/buffer_sync.c | 22 ++++++++++++++++++----
> drivers/oprofile/buffer_sync.h | 4 ++++
> drivers/oprofile/oprof.c | 9 ++++++++-
Acked-by: Robert Richter <robert.richter@amd.com>
I will send out a delta patch that changes the initialization.
-Robert
> drivers/xen/manage.c | 2 +-
> net/iucv/iucv.c | 28 ++++++++++++++++++----------
> 7 files changed, 51 insertions(+), 18 deletions(-)
--
Advanced Micro Devices, Inc.
Operating System Research Center
email: robert.richter@amd.com
^ permalink raw reply [flat|nested] 31+ messages in thread
* [PATCH] cpumask: modifiy oprofile initialization
2009-01-22 13:11 ` Robert Richter
@ 2009-01-22 13:14 ` Robert Richter
2009-01-22 13:37 ` Ingo Molnar
2009-01-22 16:56 ` Mike Travis
0 siblings, 2 replies; 31+ messages in thread
From: Robert Richter @ 2009-01-22 13:14 UTC (permalink / raw)
To: Ingo Molnar; +Cc: Tony Luck, LKML, Robert Richter, Rusty Russell, Mike Travis
Delta patch to f7df8ed164996cd2c6aca9674388be6ef78d8b37 for
tip/cpus4096.
Moved initialization to sync_start()/sync_stop(). No changes needed in
buffer_sync.h and oprof.c anymore.
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Mike Travis <travis@sgi.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
---
drivers/oprofile/buffer_sync.c | 20 ++++++--------------
drivers/oprofile/buffer_sync.h | 4 ----
drivers/oprofile/oprof.c | 9 +--------
3 files changed, 7 insertions(+), 26 deletions(-)
diff --git a/drivers/oprofile/buffer_sync.c b/drivers/oprofile/buffer_sync.c
index c3ea5fa..8574622 100644
--- a/drivers/oprofile/buffer_sync.c
+++ b/drivers/oprofile/buffer_sync.c
@@ -154,6 +154,10 @@ int sync_start(void)
{
int err;
+ if (!alloc_cpumask_var(&marked_cpus, GFP_KERNEL))
+ return -ENOMEM;
+ cpumask_clear(marked_cpus);
+
start_cpu_work();
err = task_handoff_register(&task_free_nb);
@@ -179,6 +183,7 @@ out2:
task_handoff_unregister(&task_free_nb);
out1:
end_sync();
+ free_cpumask_var(marked_cpus);
goto out;
}
@@ -190,6 +195,7 @@ void sync_stop(void)
profile_event_unregister(PROFILE_TASK_EXIT, &task_exit_nb);
task_handoff_unregister(&task_free_nb);
end_sync();
+ free_cpumask_var(marked_cpus);
}
@@ -565,20 +571,6 @@ void sync_buffer(int cpu)
mutex_unlock(&buffer_mutex);
}
-int __init buffer_sync_init(void)
-{
- if (!alloc_cpumask_var(&marked_cpus, GFP_KERNEL))
- return -ENOMEM;
-
- cpumask_clear(marked_cpus);
- return 0;
-}
-
-void __exit buffer_sync_cleanup(void)
-{
- free_cpumask_var(marked_cpus);
-}
-
/* The function can be used to add a buffer worth of data directly to
* the kernel buffer. The buffer is assumed to be a circular buffer.
* Take the entries from index start and end at index end, wrapping
diff --git a/drivers/oprofile/buffer_sync.h b/drivers/oprofile/buffer_sync.h
index 0ebf5db..3110732 100644
--- a/drivers/oprofile/buffer_sync.h
+++ b/drivers/oprofile/buffer_sync.h
@@ -19,8 +19,4 @@ void sync_stop(void);
/* sync the given CPU's buffer */
void sync_buffer(int cpu);
-/* initialize/destroy the buffer system. */
-int buffer_sync_init(void);
-void buffer_sync_cleanup(void);
-
#endif /* OPROFILE_BUFFER_SYNC_H */
diff --git a/drivers/oprofile/oprof.c b/drivers/oprofile/oprof.c
index ced39f6..3cffce9 100644
--- a/drivers/oprofile/oprof.c
+++ b/drivers/oprofile/oprof.c
@@ -183,10 +183,6 @@ static int __init oprofile_init(void)
{
int err;
- err = buffer_sync_init();
- if (err)
- return err;
-
err = oprofile_arch_init(&oprofile_ops);
if (err < 0 || timer) {
@@ -195,10 +191,8 @@ static int __init oprofile_init(void)
}
err = oprofilefs_register();
- if (err) {
+ if (err)
oprofile_arch_exit();
- buffer_sync_cleanup();
- }
return err;
}
@@ -208,7 +202,6 @@ static void __exit oprofile_exit(void)
{
oprofilefs_unregister();
oprofile_arch_exit();
- buffer_sync_cleanup();
}
--
1.6.1
^ permalink raw reply related [flat|nested] 31+ messages in thread
* Re: [PATCH] cpumask: modifiy oprofile initialization
2009-01-22 13:14 ` [PATCH] cpumask: modifiy oprofile initialization Robert Richter
@ 2009-01-22 13:37 ` Ingo Molnar
2009-01-22 17:20 ` Mike Travis
2009-01-22 19:41 ` Mike Travis
2009-01-22 16:56 ` Mike Travis
1 sibling, 2 replies; 31+ messages in thread
From: Ingo Molnar @ 2009-01-22 13:37 UTC (permalink / raw)
To: Robert Richter; +Cc: Tony Luck, LKML, Rusty Russell, Mike Travis
* Robert Richter <robert.richter@amd.com> wrote:
> Delta patch to f7df8ed164996cd2c6aca9674388be6ef78d8b37 for
> tip/cpus4096.
>
> Moved initialization to sync_start()/sync_stop(). No changes needed in
> buffer_sync.h and oprof.c anymore.
>
> Cc: Rusty Russell <rusty@rustcorp.com.au>
> Cc: Mike Travis <travis@sgi.com>
> Signed-off-by: Robert Richter <robert.richter@amd.com>
> ---
> drivers/oprofile/buffer_sync.c | 20 ++++++--------------
> drivers/oprofile/buffer_sync.h | 4 ----
> drivers/oprofile/oprof.c | 9 +--------
> 3 files changed, 7 insertions(+), 26 deletions(-)
applied to tip/cpus4096, thanks Robert!
Ingo
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH] cpumask: modifiy oprofile initialization
2009-01-22 13:14 ` [PATCH] cpumask: modifiy oprofile initialization Robert Richter
2009-01-22 13:37 ` Ingo Molnar
@ 2009-01-22 16:56 ` Mike Travis
2009-01-23 14:12 ` Robert Richter
1 sibling, 1 reply; 31+ messages in thread
From: Mike Travis @ 2009-01-22 16:56 UTC (permalink / raw)
To: Robert Richter; +Cc: Ingo Molnar, Tony Luck, LKML, Rusty Russell
Robert Richter wrote:
> Delta patch to f7df8ed164996cd2c6aca9674388be6ef78d8b37 for
> tip/cpus4096.
>
> Moved initialization to sync_start()/sync_stop(). No changes needed in
> buffer_sync.h and oprof.c anymore.
Great, thanks Robert!
Are you pushing this via some upstream queue so I can drop it from mine?
Thanks,
Mike
>
> Cc: Rusty Russell <rusty@rustcorp.com.au>
> Cc: Mike Travis <travis@sgi.com>
> Signed-off-by: Robert Richter <robert.richter@amd.com>
> ---
> drivers/oprofile/buffer_sync.c | 20 ++++++--------------
> drivers/oprofile/buffer_sync.h | 4 ----
> drivers/oprofile/oprof.c | 9 +--------
> 3 files changed, 7 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/oprofile/buffer_sync.c b/drivers/oprofile/buffer_sync.c
> index c3ea5fa..8574622 100644
> --- a/drivers/oprofile/buffer_sync.c
> +++ b/drivers/oprofile/buffer_sync.c
> @@ -154,6 +154,10 @@ int sync_start(void)
> {
> int err;
>
> + if (!alloc_cpumask_var(&marked_cpus, GFP_KERNEL))
> + return -ENOMEM;
> + cpumask_clear(marked_cpus);
> +
> start_cpu_work();
>
> err = task_handoff_register(&task_free_nb);
> @@ -179,6 +183,7 @@ out2:
> task_handoff_unregister(&task_free_nb);
> out1:
> end_sync();
> + free_cpumask_var(marked_cpus);
> goto out;
> }
>
> @@ -190,6 +195,7 @@ void sync_stop(void)
> profile_event_unregister(PROFILE_TASK_EXIT, &task_exit_nb);
> task_handoff_unregister(&task_free_nb);
> end_sync();
> + free_cpumask_var(marked_cpus);
> }
>
>
> @@ -565,20 +571,6 @@ void sync_buffer(int cpu)
> mutex_unlock(&buffer_mutex);
> }
>
> -int __init buffer_sync_init(void)
> -{
> - if (!alloc_cpumask_var(&marked_cpus, GFP_KERNEL))
> - return -ENOMEM;
> -
> - cpumask_clear(marked_cpus);
> - return 0;
> -}
> -
> -void __exit buffer_sync_cleanup(void)
> -{
> - free_cpumask_var(marked_cpus);
> -}
> -
> /* The function can be used to add a buffer worth of data directly to
> * the kernel buffer. The buffer is assumed to be a circular buffer.
> * Take the entries from index start and end at index end, wrapping
> diff --git a/drivers/oprofile/buffer_sync.h b/drivers/oprofile/buffer_sync.h
> index 0ebf5db..3110732 100644
> --- a/drivers/oprofile/buffer_sync.h
> +++ b/drivers/oprofile/buffer_sync.h
> @@ -19,8 +19,4 @@ void sync_stop(void);
> /* sync the given CPU's buffer */
> void sync_buffer(int cpu);
>
> -/* initialize/destroy the buffer system. */
> -int buffer_sync_init(void);
> -void buffer_sync_cleanup(void);
> -
> #endif /* OPROFILE_BUFFER_SYNC_H */
> diff --git a/drivers/oprofile/oprof.c b/drivers/oprofile/oprof.c
> index ced39f6..3cffce9 100644
> --- a/drivers/oprofile/oprof.c
> +++ b/drivers/oprofile/oprof.c
> @@ -183,10 +183,6 @@ static int __init oprofile_init(void)
> {
> int err;
>
> - err = buffer_sync_init();
> - if (err)
> - return err;
> -
> err = oprofile_arch_init(&oprofile_ops);
>
> if (err < 0 || timer) {
> @@ -195,10 +191,8 @@ static int __init oprofile_init(void)
> }
>
> err = oprofilefs_register();
> - if (err) {
> + if (err)
> oprofile_arch_exit();
> - buffer_sync_cleanup();
> - }
>
> return err;
> }
> @@ -208,7 +202,6 @@ static void __exit oprofile_exit(void)
> {
> oprofilefs_unregister();
> oprofile_arch_exit();
> - buffer_sync_cleanup();
> }
>
>
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH] cpumask: modifiy oprofile initialization
2009-01-22 13:37 ` Ingo Molnar
@ 2009-01-22 17:20 ` Mike Travis
2009-01-22 17:40 ` Ingo Molnar
2009-01-22 19:41 ` Mike Travis
1 sibling, 1 reply; 31+ messages in thread
From: Mike Travis @ 2009-01-22 17:20 UTC (permalink / raw)
To: Ingo Molnar; +Cc: Robert Richter, Tony Luck, LKML, Rusty Russell
Ingo Molnar wrote:
> * Robert Richter <robert.richter@amd.com> wrote:
>
>> Delta patch to f7df8ed164996cd2c6aca9674388be6ef78d8b37 for
>> tip/cpus4096.
>>
>> Moved initialization to sync_start()/sync_stop(). No changes needed in
>> buffer_sync.h and oprof.c anymore.
>>
>> Cc: Rusty Russell <rusty@rustcorp.com.au>
>> Cc: Mike Travis <travis@sgi.com>
>> Signed-off-by: Robert Richter <robert.richter@amd.com>
>> ---
>> drivers/oprofile/buffer_sync.c | 20 ++++++--------------
>> drivers/oprofile/buffer_sync.h | 4 ----
>> drivers/oprofile/oprof.c | 9 +--------
>> 3 files changed, 7 insertions(+), 26 deletions(-)
>
> applied to tip/cpus4096, thanks Robert!
>
> Ingo
Hi Ingo,
Doesn't this need to be pushed upstream as the error was
discovered in linux-next on ia64?
Thanks,
Mike
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH] cpumask: modifiy oprofile initialization
2009-01-22 17:20 ` Mike Travis
@ 2009-01-22 17:40 ` Ingo Molnar
0 siblings, 0 replies; 31+ messages in thread
From: Ingo Molnar @ 2009-01-22 17:40 UTC (permalink / raw)
To: Mike Travis; +Cc: Robert Richter, Tony Luck, LKML, Rusty Russell
* Mike Travis <travis@sgi.com> wrote:
> Ingo Molnar wrote:
> > * Robert Richter <robert.richter@amd.com> wrote:
> >
> >> Delta patch to f7df8ed164996cd2c6aca9674388be6ef78d8b37 for
> >> tip/cpus4096.
> >>
> >> Moved initialization to sync_start()/sync_stop(). No changes needed in
> >> buffer_sync.h and oprof.c anymore.
> >>
> >> Cc: Rusty Russell <rusty@rustcorp.com.au>
> >> Cc: Mike Travis <travis@sgi.com>
> >> Signed-off-by: Robert Richter <robert.richter@amd.com>
> >> ---
> >> drivers/oprofile/buffer_sync.c | 20 ++++++--------------
> >> drivers/oprofile/buffer_sync.h | 4 ----
> >> drivers/oprofile/oprof.c | 9 +--------
> >> 3 files changed, 7 insertions(+), 26 deletions(-)
> >
> > applied to tip/cpus4096, thanks Robert!
> >
> > Ingo
>
> Hi Ingo,
>
> Doesn't this need to be pushed upstream as the error was discovered in
> linux-next on ia64?
no, upstream -git does not have these changes yet, they came into
linux-next via tip:cpus4096.
Ingo
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH] cpumask: modifiy oprofile initialization
2009-01-22 13:37 ` Ingo Molnar
2009-01-22 17:20 ` Mike Travis
@ 2009-01-22 19:41 ` Mike Travis
1 sibling, 0 replies; 31+ messages in thread
From: Mike Travis @ 2009-01-22 19:41 UTC (permalink / raw)
To: Ingo Molnar; +Cc: Robert Richter, Tony Luck, LKML, Rusty Russell
Ingo Molnar wrote:
> * Robert Richter <robert.richter@amd.com> wrote:
>
>> Delta patch to f7df8ed164996cd2c6aca9674388be6ef78d8b37 for
>> tip/cpus4096.
>>
>> Moved initialization to sync_start()/sync_stop(). No changes needed in
>> buffer_sync.h and oprof.c anymore.
>>
>> Cc: Rusty Russell <rusty@rustcorp.com.au>
>> Cc: Mike Travis <travis@sgi.com>
>> Signed-off-by: Robert Richter <robert.richter@amd.com>
>> ---
>> drivers/oprofile/buffer_sync.c | 20 ++++++--------------
>> drivers/oprofile/buffer_sync.h | 4 ----
>> drivers/oprofile/oprof.c | 9 +--------
>> 3 files changed, 7 insertions(+), 26 deletions(-)
>
> applied to tip/cpus4096, thanks Robert!
>
> Ingo
Built and tested on an Altix 4700.
Cheers,
Mike
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH] cpumask: modifiy oprofile initialization
2009-01-22 16:56 ` Mike Travis
@ 2009-01-23 14:12 ` Robert Richter
0 siblings, 0 replies; 31+ messages in thread
From: Robert Richter @ 2009-01-23 14:12 UTC (permalink / raw)
To: Mike Travis; +Cc: Ingo Molnar, Tony Luck, LKML, Rusty Russell
On 22.01.09 08:56:51, Mike Travis wrote:
> Robert Richter wrote:
> > Delta patch to f7df8ed164996cd2c6aca9674388be6ef78d8b37 for
> > tip/cpus4096.
> >
> > Moved initialization to sync_start()/sync_stop(). No changes needed in
> > buffer_sync.h and oprof.c anymore.
>
> Great, thanks Robert!
>
> Are you pushing this via some upstream queue so I can drop it from mine?
Mike, it is in the tip tree commit id 4c50d9e (tip/cpus4096).
-Robert
--
Advanced Micro Devices, Inc.
Operating System Research Center
email: robert.richter@amd.com
^ permalink raw reply [flat|nested] 31+ messages in thread
* [PATCH] sfc: modify allocation error message
[not found] ` <200901201209.34854.rusty@rustcorp.com.au>
2009-01-20 18:26 ` [PATCH 4/5] cpumask: convert drivers/net/sfc [PATCH supplied] Mike Travis
@ 2009-01-26 15:04 ` Mike Travis
2009-01-27 12:13 ` Ben Hutchings
1 sibling, 1 reply; 31+ messages in thread
From: Mike Travis @ 2009-01-26 15:04 UTC (permalink / raw)
To: Rusty Russell
Cc: Ben Hutchings, Ingo Molnar, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, linux-net-drivers
Rusty Russell wrote:
> On Tuesday 20 January 2009 03:38:30 Ben Hutchings wrote:
>> On Wed, 2009-01-07 at 11:58 -0800, Mike Travis wrote:
>>> From: Rusty Russell <rusty@rustcorp.com.au>
>>>
>>> Impact: reduce stack usage, use new cpumask API.
>>>
>>> Remove a cpumask from the stack. Ben Hutchings indicated that printing
>>> a warning and returning 1 was acceptable for the corner case where allocation
>>> fails.
>>>
>>> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
>>> Signed-off-by: Mike Travis <travis@sgi.com>
>>> Cc: Ben Hutchings <bhutchings@solarflare.com>
>>> Cc: linux-net-drivers@solarflare.com
>> Acked-by: Ben Hutchings <bhutchings@solarflare.com>
>>
>> This seems to work.
>>
>>> ---
>>> drivers/net/sfc/efx.c | 15 +++++++++++----
>>> 1 file changed, 11 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/drivers/net/sfc/efx.c b/drivers/net/sfc/efx.c
>>> --- a/drivers/net/sfc/efx.c
>>> +++ b/drivers/net/sfc/efx.c
>>> @@ -854,20 +854,27 @@ static void efx_fini_io(struct efx_nic *
>>> * interrupts across them. */
>>> static int efx_wanted_rx_queues(void)
>>> {
>>> - cpumask_t core_mask;
>>> + cpumask_var_t core_mask;
>>> int count;
>>> int cpu;
>>>
>>> - cpus_clear(core_mask);
>>> + if (!alloc_cpumask_var(&core_mask, GFP_KERNEL)) {
>>> + printk(KERN_WARNING
>>> + "efx.c: allocation failure, irq balancing hobbled\n");
>> [...]
>>
>> This is an exceedingly unlikely error case so the error message is not
>> that important, but if you don't mind re-spinning this then please make
>> it "sfc: RSS disabled due to allocation failure\n".
>
> OK, I'm away from all my machines, at linux.conf.au. Mike, can you change
> this if you haven't already?
>
> Thanks,
> Rusty.
>
Subject: sfc: modify allocation error message
Impact: change error message when alloc_cpumask_var fails.
Signed-off-by: Mike Travis <travis@sgi.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/sfc/efx.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
--- linux-2.6-for-ingo.orig/drivers/net/sfc/efx.c
+++ linux-2.6-for-ingo/drivers/net/sfc/efx.c
@@ -858,9 +858,9 @@ static int efx_wanted_rx_queues(void)
int count;
int cpu;
- if (!alloc_cpumask_var(&core_mask, GFP_KERNEL)) {
+ if (unlikely(!alloc_cpumask_var(&core_mask, GFP_KERNEL))) {
printk(KERN_WARNING
- "efx.c: allocation failure, irq balancing hobbled\n");
+ "sfc: RSS disabled due to allocation failure\n");
return 1;
}
^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [PATCH] sfc: modify allocation error message
2009-01-26 15:04 ` [PATCH] sfc: modify allocation error message Mike Travis
@ 2009-01-27 12:13 ` Ben Hutchings
0 siblings, 0 replies; 31+ messages in thread
From: Ben Hutchings @ 2009-01-27 12:13 UTC (permalink / raw)
To: Mike Travis
Cc: Rusty Russell, Ingo Molnar, Thomas Gleixner, H. Peter Anvin,
Jack Steiner, linux-kernel, linux-net-drivers
On Mon, 2009-01-26 at 07:04 -0800, Mike Travis wrote:
[...]
> Subject: sfc: modify allocation error message
>
> Impact: change error message when alloc_cpumask_var fails.
>
> Signed-off-by: Mike Travis <travis@sgi.com>
> Cc: Ben Hutchings <bhutchings@solarflare.com>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
> --
> drivers/net/sfc/efx.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> --- linux-2.6-for-ingo.orig/drivers/net/sfc/efx.c
> +++ linux-2.6-for-ingo/drivers/net/sfc/efx.c
> @@ -858,9 +858,9 @@ static int efx_wanted_rx_queues(void)
> int count;
> int cpu;
>
> - if (!alloc_cpumask_var(&core_mask, GFP_KERNEL)) {
> + if (unlikely(!alloc_cpumask_var(&core_mask, GFP_KERNEL))) {
> printk(KERN_WARNING
> - "efx.c: allocation failure, irq balancing hobbled\n");
> + "sfc: RSS disabled due to allocation failure\n");
> return 1;
> }
>
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply [flat|nested] 31+ messages in thread
end of thread, other threads:[~2009-01-27 12:13 UTC | newest]
Thread overview: 31+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-01-07 19:58 [PATCH 0/5] cpumask: more cpumask updates Mike Travis
2009-01-07 19:58 ` [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t Mike Travis
2009-01-07 20:27 ` Yinghai Lu
2009-01-07 22:32 ` Mike Travis
2009-01-08 3:15 ` Yinghai Lu
2009-01-08 15:45 ` Mike Travis
2009-01-08 19:31 ` Yinghai Lu
2009-01-08 18:40 ` [PATCH 1/5] cpumask: update irq_desc to use cpumask_var_t: fix Mike Travis
2009-01-07 19:58 ` [PATCH 2/5] cpumask: Use topology_core_cpumask()/topology_thread_cpumask() Mike Travis
2009-01-19 17:11 ` Ben Hutchings
2009-01-07 19:58 ` [PATCH 3/5] cpumask: convert misc driver functions Mike Travis
2009-01-10 10:57 ` Rusty Russell
2009-01-16 18:42 ` Tony Luck
2009-01-16 18:55 ` Mike Travis
2009-01-16 22:02 ` Mike Travis
2009-01-22 13:11 ` Robert Richter
2009-01-22 13:14 ` [PATCH] cpumask: modifiy oprofile initialization Robert Richter
2009-01-22 13:37 ` Ingo Molnar
2009-01-22 17:20 ` Mike Travis
2009-01-22 17:40 ` Ingo Molnar
2009-01-22 19:41 ` Mike Travis
2009-01-22 16:56 ` Mike Travis
2009-01-23 14:12 ` Robert Richter
2009-01-07 19:58 ` [PATCH 4/5] cpumask: convert drivers/net/sfc Mike Travis
2009-01-19 17:08 ` Ben Hutchings
[not found] ` <200901201209.34854.rusty@rustcorp.com.au>
2009-01-20 18:26 ` [PATCH 4/5] cpumask: convert drivers/net/sfc [PATCH supplied] Mike Travis
2009-01-26 15:04 ` [PATCH] sfc: modify allocation error message Mike Travis
2009-01-27 12:13 ` Ben Hutchings
2009-01-07 19:58 ` [PATCH 5/5] cpumask: convert other misc kernel functions Mike Travis
2009-01-07 20:52 ` [PATCH 0/5] cpumask: more cpumask updates Ingo Molnar
2009-01-07 21:20 ` Mike Travis
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).