[PATCH UPDATED] percpu: use dynamic percpu allocator as the defaultpercpu allocator

From: Tejun Heo
Date: Mon Mar 30 2009 - 06:10:14 EST

Impact: use dynamic allocator for most archs w/o
CONFIG_HAVE_SETUP_PER_CPU_AREA

This patch makes most !CONFIG_HAVE_SETUP_PER_CPU_AREA archs use
dynamic percpu allocator. The first chunk is allocated using
embedding helper and 8k is reserved for modules. This ensures that
the new allocator behaves almost identically to the original allocator
as long as static percpu variables are concerned, so it shouldn't
introduce much breakage.

s390 and alpha use custom SHIFT_PERCPU_PTR() to work around addressing
range limit the addressing model imposes. Unfortunately, this breaks
if the address is specified using a variable, so for now, the two
archs aren't converted.

The following architectures are affected by this change.

* sh
* arm
* cris
* mips
* sparc(32)
* blackfin
* avr32
* parisc
* m32r
* powerpc(32)

As this change makes the dynamic allocator the default one,
CONFIG_HAVE_DYNAMIC_PER_CPU_AREA is replaced with its invert -
CONFIG_HAVE_LEGACY_PER_CPU_AREA, which is added to yet-to-be converted
archs. These archs implement their own setup_per_cpu_areas() and the
conversion is not trivial.

* powerpc(64)
* sparc(64)
* ia64
* alpha
* s390

Boot and batch alloc/free tests on x86_32 with debug code (x86_32
doesn't use default first chunk initialization). Compile tested on
sparc(32), powerpc(32), arm and alpha.

Signed-off-by: Tejun Heo <tj@xxxxxxxxxx>
Cc: Paul Mundt <lethal@xxxxxxxxxxxx>
Cc: Russell King <rmk@xxxxxxxxxxxxxxxx>
Cc: Mikael Starvik <starvik@xxxxxxxx>
Cc: Ralf Baechle <ralf@xxxxxxxxxxxxxx>
Cc: David S. Miller <davem@xxxxxxxxxxxxx>
Cc: Bryan Wu <cooloney@xxxxxxxxxx>
Cc: Kyle McMartin <kyle@xxxxxxxxxxx>
Cc: Matthew Wilcox <matthew@xxxxxx>
Cc: Grant Grundler <grundler@xxxxxxxxxxxxxxxx>
Cc: Hirokazu Takata <takata@xxxxxxxxxxxxxx>
Cc: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx>
Cc: Richard Henderson <rth@xxxxxxxxxxx>
Cc: Ivan Kokshaysky <ink@xxxxxxxxxxxxxxxxxxxx>
Cc: Martin Schwidefsky <schwidefsky@xxxxxxxxxx>
Cc: Heiko Carstens <heiko.carstens@xxxxxxxxxx>
---
Okay, this should keep s390 and alpha working till proper solution is
found. Martin, can you please verify? Ingo, please feel free to push
this upstream (or -next) once Martin acks.

Thanks.

arch/alpha/Kconfig | 3 +++
arch/ia64/Kconfig | 3 +++
arch/powerpc/Kconfig | 3 +++
arch/s390/Kconfig | 3 +++
arch/sparc/Kconfig | 3 +++
arch/x86/Kconfig | 3 ---
include/linux/percpu.h | 12 +++++++++---
init/main.c | 24 ------------------------
kernel/module.c | 6 +++---
mm/Makefile | 2 +-
mm/allocpercpu.c | 28 ++++++++++++++++++++++++++++
mm/percpu.c | 40 +++++++++++++++++++++++++++++++++++++++-
12 files changed, 95 insertions(+), 35 deletions(-)

diff --git a/arch/alpha/Kconfig b/arch/alpha/Kconfig
index 9fb8aae..05d8640 100644
--- a/arch/alpha/Kconfig
+++ b/arch/alpha/Kconfig
@@ -70,6 +70,9 @@ config AUTO_IRQ_AFFINITY
depends on SMP
default y

+config HAVE_LEGACY_PER_CPU_AREA
+ def_bool y
+
source "init/Kconfig"
source "kernel/Kconfig.freezer"

diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig
index 153e727..f333149 100644
--- a/arch/ia64/Kconfig
+++ b/arch/ia64/Kconfig
@@ -85,6 +85,9 @@ config GENERIC_TIME_VSYSCALL
bool
default y

+config HAVE_LEGACY_PER_CPU_AREA
+ def_bool y
+
config HAVE_SETUP_PER_CPU_AREA
def_bool y

diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index 74cc312..892f1ad 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -42,6 +42,9 @@ config GENERIC_HARDIRQS
bool
default y

+config HAVE_LEGACY_PER_CPU_AREA
+ def_bool PPC64
+
config HAVE_SETUP_PER_CPU_AREA
def_bool PPC64

diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index 6b0a353..c69264a 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -72,6 +72,9 @@ config PGSTE
config VIRT_CPU_ACCOUNTING
def_bool y

+config HAVE_LEGACY_PER_CPU_AREA
+ def_bool y
+
mainmenu "Linux Kernel Configuration"

config S390
diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig
index c3ea215..0125b57 100644
--- a/arch/sparc/Kconfig
+++ b/arch/sparc/Kconfig
@@ -90,6 +90,9 @@ config AUDIT_ARCH
bool
default y

+config HAVE_LEGACY_PER_CPU_AREA
+ def_bool y if SPARC64
+
config HAVE_SETUP_PER_CPU_AREA
def_bool y if SPARC64

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index f5cef3f..572329a 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -138,9 +138,6 @@ config ARCH_HAS_CACHE_LINE_SIZE
config HAVE_SETUP_PER_CPU_AREA
def_bool y

-config HAVE_DYNAMIC_PER_CPU_AREA
- def_bool y
-
config HAVE_CPUMASK_OF_CPU_MAP
def_bool X86_64_SMP

diff --git a/include/linux/percpu.h b/include/linux/percpu.h
index ee5615d..817136a 100644
--- a/include/linux/percpu.h
+++ b/include/linux/percpu.h
@@ -78,7 +78,7 @@

#ifdef CONFIG_SMP

-#ifdef CONFIG_HAVE_DYNAMIC_PER_CPU_AREA
+#ifndef CONFIG_HAVE_LEGACY_PER_CPU_AREA

/* minimum unit size, also is the maximum supported allocation size */
#define PCPU_MIN_UNIT_SIZE PFN_ALIGN(64 << 10)
@@ -124,7 +124,7 @@ extern ssize_t __init pcpu_embed_first_chunk(

extern void *__alloc_reserved_percpu(size_t size, size_t align);

-#else /* CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */
+#else /* CONFIG_HAVE_LEGACY_PER_CPU_AREA */

struct percpu_data {
void *ptrs[1];
@@ -138,11 +138,15 @@ struct percpu_data {
(__typeof__(ptr))__p->ptrs[(cpu)]; \
})

-#endif /* CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */
+#endif /* CONFIG_HAVE_LEGACY_PER_CPU_AREA */

extern void *__alloc_percpu(size_t size, size_t align);
extern void free_percpu(void *__pdata);

+#ifndef CONFIG_HAVE_SETUP_PER_CPU_AREA
+extern void __init setup_per_cpu_areas(void);
+#endif
+
#else /* CONFIG_SMP */

#define per_cpu_ptr(ptr, cpu) ({ (void)(cpu); (ptr); })
@@ -163,6 +167,8 @@ static inline void free_percpu(void *p)
kfree(p);
}

+static inline void __init setup_per_cpu_areas(void) { }
+
#endif /* CONFIG_SMP */

#define alloc_percpu(type) (type *)__alloc_percpu(sizeof(type), \
diff --git a/init/main.c b/init/main.c
index 6441083..42ae3c0 100644
--- a/init/main.c
+++ b/init/main.c
@@ -354,7 +354,6 @@ static void __init smp_init(void)
#define smp_init() do { } while (0)
#endif

-static inline void setup_per_cpu_areas(void) { }
static inline void setup_nr_cpu_ids(void) { }
static inline void smp_prepare_cpus(unsigned int maxcpus) { }

@@ -375,29 +374,6 @@ static void __init setup_nr_cpu_ids(void)
nr_cpu_ids = find_last_bit(cpumask_bits(cpu_possible_mask),NR_CPUS) + 1;
}

-#ifndef CONFIG_HAVE_SETUP_PER_CPU_AREA
-unsigned long __per_cpu_offset[NR_CPUS] __read_mostly;
-
-EXPORT_SYMBOL(__per_cpu_offset);
-
-static void __init setup_per_cpu_areas(void)
-{
- unsigned long size, i;
- char *ptr;
- unsigned long nr_possible_cpus = num_possible_cpus();
-
- /* Copy section for each CPU (we discard the original) */
- size = ALIGN(PERCPU_ENOUGH_ROOM, PAGE_SIZE);
- ptr = alloc_bootmem_pages(size * nr_possible_cpus);
-
- for_each_possible_cpu(i) {
- __per_cpu_offset[i] = ptr - __per_cpu_start;
- memcpy(ptr, __per_cpu_start, __per_cpu_end - __per_cpu_start);
- ptr += size;
- }
-}
-#endif /* CONFIG_HAVE_SETUP_PER_CPU_AREA */
-
/* Called by boot processor to activate the rest. */
static void __init smp_init(void)
{
diff --git a/kernel/module.c b/kernel/module.c
index f0e04d6..f9c7b86 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -368,7 +368,7 @@ static struct module *find_module(const char *name)

#ifdef CONFIG_SMP

-#ifdef CONFIG_HAVE_DYNAMIC_PER_CPU_AREA
+#ifndef CONFIG_HAVE_LEGACY_PER_CPU_AREA

static void *percpu_modalloc(unsigned long size, unsigned long align,
const char *name)
@@ -393,7 +393,7 @@ static void percpu_modfree(void *freeme)
free_percpu(freeme);
}

-#else /* ... !CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */
+#else /* ... CONFIG_HAVE_LEGACY_PER_CPU_AREA */

/* Number of blocks used and allocated. */
static unsigned int pcpu_num_used, pcpu_num_allocated;
@@ -528,7 +528,7 @@ static int percpu_modinit(void)
}
__initcall(percpu_modinit);

-#endif /* CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */
+#endif /* CONFIG_HAVE_LEGACY_PER_CPU_AREA */

static unsigned int find_pcpusec(Elf_Ehdr *hdr,
Elf_Shdr *sechdrs,
diff --git a/mm/Makefile b/mm/Makefile
index 818569b..f3a037b 100644
--- a/mm/Makefile
+++ b/mm/Makefile
@@ -30,7 +30,7 @@ obj-$(CONFIG_FAILSLAB) += failslab.o
obj-$(CONFIG_MEMORY_HOTPLUG) += memory_hotplug.o
obj-$(CONFIG_FS_XIP) += filemap_xip.o
obj-$(CONFIG_MIGRATION) += migrate.o
-ifdef CONFIG_HAVE_DYNAMIC_PER_CPU_AREA
+ifndef CONFIG_HAVE_LEGACY_PER_CPU_AREA
obj-$(CONFIG_SMP) += percpu.o
else
obj-$(CONFIG_SMP) += allocpercpu.o
diff --git a/mm/allocpercpu.c b/mm/allocpercpu.c
index 1882923..fd75eff 100644
--- a/mm/allocpercpu.c
+++ b/mm/allocpercpu.c
@@ -5,6 +5,8 @@
*/
#include <linux/mm.h>
#include <linux/module.h>
+#include <linux/bootmem.h>
+#include <asm/sections.h>

#ifndef cache_line_size
#define cache_line_size() L1_CACHE_BYTES
@@ -147,3 +149,29 @@ void free_percpu(void *__pdata)
kfree(__percpu_disguise(__pdata));
}
EXPORT_SYMBOL_GPL(free_percpu);
+
+/*
+ * Generic percpu area setup.
+ */
+#ifndef CONFIG_HAVE_SETUP_PER_CPU_AREA
+unsigned long __per_cpu_offset[NR_CPUS] __read_mostly;
+
+EXPORT_SYMBOL(__per_cpu_offset);
+
+void __init setup_per_cpu_areas(void)
+{
+ unsigned long size, i;
+ char *ptr;
+ unsigned long nr_possible_cpus = num_possible_cpus();
+
+ /* Copy section for each CPU (we discard the original) */
+ size = ALIGN(PERCPU_ENOUGH_ROOM, PAGE_SIZE);
+ ptr = alloc_bootmem_pages(size * nr_possible_cpus);
+
+ for_each_possible_cpu(i) {
+ __per_cpu_offset[i] = ptr - __per_cpu_start;
+ memcpy(ptr, __per_cpu_start, __per_cpu_end - __per_cpu_start);
+ ptr += size;
+ }
+}
+#endif /* CONFIG_HAVE_SETUP_PER_CPU_AREA */
diff --git a/mm/percpu.c b/mm/percpu.c
index 1aa5d8f..059b57a 100644
--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -43,7 +43,7 @@
*
* To use this allocator, arch code should do the followings.
*
- * - define CONFIG_HAVE_DYNAMIC_PER_CPU_AREA
+ * - drop CONFIG_HAVE_LEGACY_PER_CPU_AREA
*
* - define __addr_to_pcpu_ptr() and __pcpu_ptr_to_addr() to translate
* regular address to percpu pointer and back if they need to be
@@ -1324,3 +1324,41 @@ ssize_t __init pcpu_embed_first_chunk(size_t static_size, size_t reserved_size,
reserved_size, dyn_size,
pcpue_unit_size, pcpue_ptr, NULL);
}
+
+/*
+ * Generic percpu area setup.
+ *
+ * The embedding helper is used because its behavior closely resembles
+ * the original non-dynamic generic percpu area setup. This is
+ * important because many archs have addressing restrictions and might
+ * fail if the percpu area is located far away from the previous
+ * location. As an added bonus, in non-NUMA cases, embedding is
+ * generally a good idea TLB-wise because percpu area can piggy back
+ * on the physical linear memory mapping which uses large page
+ * mappings on applicable archs.
+ */
+#ifndef CONFIG_HAVE_SETUP_PER_CPU_AREA
+unsigned long __per_cpu_offset[NR_CPUS] __read_mostly;
+EXPORT_SYMBOL(__per_cpu_offset);
+
+void __init setup_per_cpu_areas(void)
+{
+ size_t static_size = __per_cpu_end - __per_cpu_start;
+ ssize_t unit_size;
+ unsigned long delta;
+ unsigned int cpu;
+
+ /*
+ * Always reserve area for module percpu variables. That's
+ * what the legacy allocator did.
+ */
+ unit_size = pcpu_embed_first_chunk(static_size, PERCPU_MODULE_RESERVE,
+ PERCPU_DYNAMIC_RESERVE, -1);
+ if (unit_size < 0)
+ panic("Failed to initialized percpu areas.");
+
+ delta = (unsigned long)pcpu_base_addr - (unsigned long)__per_cpu_start;
+ for_each_possible_cpu(cpu)
+ __per_cpu_offset[cpu] = delta + cpu * unit_size;
+}
+#endif /* CONFIG_HAVE_SETUP_PER_CPU_AREA */
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/