* [PATCH] base memory: introduce CONFIG_MEMORY_DEVICE
@ 2016-10-10 11:26 Yisheng Xie
2016-10-14 10:53 ` Yisheng Xie
0 siblings, 1 reply; 5+ messages in thread
From: Yisheng Xie @ 2016-10-10 11:26 UTC (permalink / raw)
To: akpm, andrew, daniel.kiper, srinivas.kandagatla, gregkh, vkuznets
Cc: linux-kernel, ak, guohanjun, qiuxishi
MEMORY_FAILURE do not depend on SPARSEMEM_MANUAL,
nor MEMORY_HOTPLUG_SPARSE. However, when I tried to use sysfs:
/sys/devices/system/memory/soft_offline_page
/sys/devices/system/memory/hard_offline_page
to test memory failure function with FLATMEM_MANUAL && MEMORY_FAILURE
enabled on arch like i386, it failed for no such sysfs.
To make sysfs soft_offline_page usable once MEMORY_FAILURE is enabled,
this patch introduces CONFIG_MEMORY_DEVICE, and selects it when
MEMORY_FAILURE or MEMORY_HOTPLUG_SPARSE is enabled.
Signed-off-by: Yisheng Xie <xieyisheng1@huawei.com>
---
drivers/base/Kconfig | 3 +++
drivers/base/Makefile | 2 +-
drivers/base/memory.c | 32 ++++++++++++++++++++++++++++++--
include/linux/memory.h | 4 ++++
4 files changed, 38 insertions(+), 3 deletions(-)
diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig
index fdf44ca..b4eac4e 100644
--- a/drivers/base/Kconfig
+++ b/drivers/base/Kconfig
@@ -271,6 +271,9 @@ config DMA_CMA
For more information see <include/linux/dma-contiguous.h>.
If unsure, say "n".
+config MEMORY_DEVICE
+ def_bool MEMORY_HOTPLUG_SPARSE || MEMORY_FAILURE
+
if DMA_CMA
comment "Default contiguous memory area size:"
diff --git a/drivers/base/Makefile b/drivers/base/Makefile
index 2609ba2..aafe34b 100644
--- a/drivers/base/Makefile
+++ b/drivers/base/Makefile
@@ -13,7 +13,7 @@ obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o
obj-$(CONFIG_ISA_BUS_API) += isa.o
obj-$(CONFIG_FW_LOADER) += firmware_class.o
obj-$(CONFIG_NUMA) += node.o
-obj-$(CONFIG_MEMORY_HOTPLUG_SPARSE) += memory.o
+obj-$(CONFIG_MEMORY_DEVICE) += memory.o
ifeq ($(CONFIG_SYSFS),y)
obj-$(CONFIG_MODULES) += module.o
endif
diff --git a/drivers/base/memory.c b/drivers/base/memory.c
index dc75de9..fb00965 100644
--- a/drivers/base/memory.c
+++ b/drivers/base/memory.c
@@ -25,10 +25,11 @@
#include <linux/atomic.h>
#include <asm/uaccess.h>
-static DEFINE_MUTEX(mem_sysfs_mutex);
-
#define MEMORY_CLASS_NAME "memory"
+#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
+static DEFINE_MUTEX(mem_sysfs_mutex);
+
#define to_memory_block(dev) container_of(dev, struct memory_block, dev)
static int sections_per_block;
@@ -381,6 +382,7 @@ static ssize_t show_phys_device(struct device *dev,
struct memory_block *mem = to_memory_block(dev);
return sprintf(buf, "%d\n", mem->phys_device);
}
+#endif
#ifdef CONFIG_MEMORY_HOTREMOVE
static ssize_t show_valid_zones(struct device *dev,
@@ -427,6 +429,7 @@ static ssize_t show_valid_zones(struct device *dev,
static DEVICE_ATTR(valid_zones, 0444, show_valid_zones, NULL);
#endif
+#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
static DEVICE_ATTR(phys_index, 0444, show_mem_start_phys_index, NULL);
static DEVICE_ATTR(state, 0644, show_mem_state, store_mem_state);
static DEVICE_ATTR(phys_device, 0444, show_phys_device, NULL);
@@ -474,6 +477,7 @@ store_auto_online_blocks(struct device *dev, struct device_attribute *attr,
static DEVICE_ATTR(auto_online_blocks, 0644, show_auto_online_blocks,
store_auto_online_blocks);
+#endif
/*
* Some architectures will have custom drivers to do this, and
@@ -557,6 +561,7 @@ static DEVICE_ATTR(soft_offline_page, S_IWUSR, NULL, store_soft_offline_page);
static DEVICE_ATTR(hard_offline_page, S_IWUSR, NULL, store_hard_offline_page);
#endif
+#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
/*
* Note that phys_device is optional. It is here to allow for
* differentiation between which *physical* devices each
@@ -723,6 +728,7 @@ out:
mutex_unlock(&mem_sysfs_mutex);
return ret;
}
+#endif
#ifdef CONFIG_MEMORY_HOTREMOVE
static void
@@ -766,11 +772,13 @@ int unregister_memory_section(struct mem_section *section)
}
#endif /* CONFIG_MEMORY_HOTREMOVE */
+#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
/* return true if the memory block is offlined, otherwise, return false */
bool is_memblock_offlined(struct memory_block *mem)
{
return mem->state == MEM_OFFLINE;
}
+#endif
static struct attribute *memory_root_attrs[] = {
#ifdef CONFIG_ARCH_MEMORY_PROBE
@@ -782,8 +790,10 @@ static struct attribute *memory_root_attrs[] = {
&dev_attr_hard_offline_page.attr,
#endif
+#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
&dev_attr_block_size_bytes.attr,
&dev_attr_auto_online_blocks.attr,
+#endif
NULL
};
@@ -799,6 +809,7 @@ static const struct attribute_group *memory_root_attr_groups[] = {
/*
* Initialize the sysfs support for memory devices...
*/
+#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
int __init memory_dev_init(void)
{
unsigned int i;
@@ -830,3 +841,20 @@ out:
printk(KERN_ERR "%s() failed: %d\n", __func__, ret);
return ret;
}
+#else
+static struct bus_type memory_subsys = {
+ .name = MEMORY_CLASS_NAME,
+ .dev_name = MEMORY_CLASS_NAME,
+};
+
+int __init memory_dev_init(void)
+{
+ int ret = 0;
+
+ ret = subsys_system_register(&memory_subsys, memory_root_attr_groups);
+
+ if (ret)
+ pr_err("%s() failed: %d\n", __func__, ret);
+ return ret;
+}
+#endif
diff --git a/include/linux/memory.h b/include/linux/memory.h
index 093607f..9fe1089 100644
--- a/include/linux/memory.h
+++ b/include/linux/memory.h
@@ -77,10 +77,14 @@ struct mem_section;
#define IPC_CALLBACK_PRI 10
#ifndef CONFIG_MEMORY_HOTPLUG_SPARSE
+#ifdef CONFIG_MEMORY_DEVICE
+extern int memory_dev_init(void);
+#else
static inline int memory_dev_init(void)
{
return 0;
}
+#endif
static inline int register_memory_notifier(struct notifier_block *nb)
{
return 0;
--
1.7.12.4
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH] base memory: introduce CONFIG_MEMORY_DEVICE
2016-10-10 11:26 [PATCH] base memory: introduce CONFIG_MEMORY_DEVICE Yisheng Xie
@ 2016-10-14 10:53 ` Yisheng Xie
0 siblings, 0 replies; 5+ messages in thread
From: Yisheng Xie @ 2016-10-14 10:53 UTC (permalink / raw)
To: akpm, andrew, daniel.kiper, srinivas.kandagatla, gregkh, vkuznets
Cc: linux-kernel, ak, guohanjun, qiuxishi, n-horiguchi, linux-mm
+ mm mail list
On 2016/10/10 19:26, Yisheng Xie wrote:
> MEMORY_FAILURE do not depend on SPARSEMEM_MANUAL,
> nor MEMORY_HOTPLUG_SPARSE. However, when I tried to use sysfs:
> /sys/devices/system/memory/soft_offline_page
> /sys/devices/system/memory/hard_offline_page
> to test memory failure function with FLATMEM_MANUAL && MEMORY_FAILURE
> enabled on arch like i386, it failed for no such sysfs.
>
> To make sysfs soft_offline_page usable once MEMORY_FAILURE is enabled,
> this patch introduces CONFIG_MEMORY_DEVICE, and selects it when
> MEMORY_FAILURE or MEMORY_HOTPLUG_SPARSE is enabled.
>
> Signed-off-by: Yisheng Xie <xieyisheng1@huawei.com>
> ---
> drivers/base/Kconfig | 3 +++
> drivers/base/Makefile | 2 +-
> drivers/base/memory.c | 32 ++++++++++++++++++++++++++++++--
> include/linux/memory.h | 4 ++++
> 4 files changed, 38 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig
> index fdf44ca..b4eac4e 100644
> --- a/drivers/base/Kconfig
> +++ b/drivers/base/Kconfig
> @@ -271,6 +271,9 @@ config DMA_CMA
> For more information see <include/linux/dma-contiguous.h>.
> If unsure, say "n".
>
> +config MEMORY_DEVICE
> + def_bool MEMORY_HOTPLUG_SPARSE || MEMORY_FAILURE
> +
> if DMA_CMA
> comment "Default contiguous memory area size:"
>
> diff --git a/drivers/base/Makefile b/drivers/base/Makefile
> index 2609ba2..aafe34b 100644
> --- a/drivers/base/Makefile
> +++ b/drivers/base/Makefile
> @@ -13,7 +13,7 @@ obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o
> obj-$(CONFIG_ISA_BUS_API) += isa.o
> obj-$(CONFIG_FW_LOADER) += firmware_class.o
> obj-$(CONFIG_NUMA) += node.o
> -obj-$(CONFIG_MEMORY_HOTPLUG_SPARSE) += memory.o
> +obj-$(CONFIG_MEMORY_DEVICE) += memory.o
> ifeq ($(CONFIG_SYSFS),y)
> obj-$(CONFIG_MODULES) += module.o
> endif
> diff --git a/drivers/base/memory.c b/drivers/base/memory.c
> index dc75de9..fb00965 100644
> --- a/drivers/base/memory.c
> +++ b/drivers/base/memory.c
> @@ -25,10 +25,11 @@
> #include <linux/atomic.h>
> #include <asm/uaccess.h>
>
> -static DEFINE_MUTEX(mem_sysfs_mutex);
> -
> #define MEMORY_CLASS_NAME "memory"
>
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> +static DEFINE_MUTEX(mem_sysfs_mutex);
> +
> #define to_memory_block(dev) container_of(dev, struct memory_block, dev)
>
> static int sections_per_block;
> @@ -381,6 +382,7 @@ static ssize_t show_phys_device(struct device *dev,
> struct memory_block *mem = to_memory_block(dev);
> return sprintf(buf, "%d\n", mem->phys_device);
> }
> +#endif
>
> #ifdef CONFIG_MEMORY_HOTREMOVE
> static ssize_t show_valid_zones(struct device *dev,
> @@ -427,6 +429,7 @@ static ssize_t show_valid_zones(struct device *dev,
> static DEVICE_ATTR(valid_zones, 0444, show_valid_zones, NULL);
> #endif
>
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> static DEVICE_ATTR(phys_index, 0444, show_mem_start_phys_index, NULL);
> static DEVICE_ATTR(state, 0644, show_mem_state, store_mem_state);
> static DEVICE_ATTR(phys_device, 0444, show_phys_device, NULL);
> @@ -474,6 +477,7 @@ store_auto_online_blocks(struct device *dev, struct device_attribute *attr,
>
> static DEVICE_ATTR(auto_online_blocks, 0644, show_auto_online_blocks,
> store_auto_online_blocks);
> +#endif
>
> /*
> * Some architectures will have custom drivers to do this, and
> @@ -557,6 +561,7 @@ static DEVICE_ATTR(soft_offline_page, S_IWUSR, NULL, store_soft_offline_page);
> static DEVICE_ATTR(hard_offline_page, S_IWUSR, NULL, store_hard_offline_page);
> #endif
>
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> /*
> * Note that phys_device is optional. It is here to allow for
> * differentiation between which *physical* devices each
> @@ -723,6 +728,7 @@ out:
> mutex_unlock(&mem_sysfs_mutex);
> return ret;
> }
> +#endif
>
> #ifdef CONFIG_MEMORY_HOTREMOVE
> static void
> @@ -766,11 +772,13 @@ int unregister_memory_section(struct mem_section *section)
> }
> #endif /* CONFIG_MEMORY_HOTREMOVE */
>
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> /* return true if the memory block is offlined, otherwise, return false */
> bool is_memblock_offlined(struct memory_block *mem)
> {
> return mem->state == MEM_OFFLINE;
> }
> +#endif
>
> static struct attribute *memory_root_attrs[] = {
> #ifdef CONFIG_ARCH_MEMORY_PROBE
> @@ -782,8 +790,10 @@ static struct attribute *memory_root_attrs[] = {
> &dev_attr_hard_offline_page.attr,
> #endif
>
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> &dev_attr_block_size_bytes.attr,
> &dev_attr_auto_online_blocks.attr,
> +#endif
> NULL
> };
>
> @@ -799,6 +809,7 @@ static const struct attribute_group *memory_root_attr_groups[] = {
> /*
> * Initialize the sysfs support for memory devices...
> */
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> int __init memory_dev_init(void)
> {
> unsigned int i;
> @@ -830,3 +841,20 @@ out:
> printk(KERN_ERR "%s() failed: %d\n", __func__, ret);
> return ret;
> }
> +#else
> +static struct bus_type memory_subsys = {
> + .name = MEMORY_CLASS_NAME,
> + .dev_name = MEMORY_CLASS_NAME,
> +};
> +
> +int __init memory_dev_init(void)
> +{
> + int ret = 0;
> +
> + ret = subsys_system_register(&memory_subsys, memory_root_attr_groups);
> +
> + if (ret)
> + pr_err("%s() failed: %d\n", __func__, ret);
> + return ret;
> +}
> +#endif
> diff --git a/include/linux/memory.h b/include/linux/memory.h
> index 093607f..9fe1089 100644
> --- a/include/linux/memory.h
> +++ b/include/linux/memory.h
> @@ -77,10 +77,14 @@ struct mem_section;
> #define IPC_CALLBACK_PRI 10
>
> #ifndef CONFIG_MEMORY_HOTPLUG_SPARSE
> +#ifdef CONFIG_MEMORY_DEVICE
> +extern int memory_dev_init(void);
> +#else
> static inline int memory_dev_init(void)
> {
> return 0;
> }
> +#endif
> static inline int register_memory_notifier(struct notifier_block *nb)
> {
> return 0;
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] base memory: introduce CONFIG_MEMORY_DEVICE
@ 2016-10-14 10:53 ` Yisheng Xie
0 siblings, 0 replies; 5+ messages in thread
From: Yisheng Xie @ 2016-10-14 10:53 UTC (permalink / raw)
To: akpm, andrew, daniel.kiper, srinivas.kandagatla, gregkh, vkuznets
Cc: linux-kernel, ak, guohanjun, qiuxishi, n-horiguchi, linux-mm
+ mm mail list
On 2016/10/10 19:26, Yisheng Xie wrote:
> MEMORY_FAILURE do not depend on SPARSEMEM_MANUAL,
> nor MEMORY_HOTPLUG_SPARSE. However, when I tried to use sysfs:
> /sys/devices/system/memory/soft_offline_page
> /sys/devices/system/memory/hard_offline_page
> to test memory failure function with FLATMEM_MANUAL && MEMORY_FAILURE
> enabled on arch like i386, it failed for no such sysfs.
>
> To make sysfs soft_offline_page usable once MEMORY_FAILURE is enabled,
> this patch introduces CONFIG_MEMORY_DEVICE, and selects it when
> MEMORY_FAILURE or MEMORY_HOTPLUG_SPARSE is enabled.
>
> Signed-off-by: Yisheng Xie <xieyisheng1@huawei.com>
> ---
> drivers/base/Kconfig | 3 +++
> drivers/base/Makefile | 2 +-
> drivers/base/memory.c | 32 ++++++++++++++++++++++++++++++--
> include/linux/memory.h | 4 ++++
> 4 files changed, 38 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig
> index fdf44ca..b4eac4e 100644
> --- a/drivers/base/Kconfig
> +++ b/drivers/base/Kconfig
> @@ -271,6 +271,9 @@ config DMA_CMA
> For more information see <include/linux/dma-contiguous.h>.
> If unsure, say "n".
>
> +config MEMORY_DEVICE
> + def_bool MEMORY_HOTPLUG_SPARSE || MEMORY_FAILURE
> +
> if DMA_CMA
> comment "Default contiguous memory area size:"
>
> diff --git a/drivers/base/Makefile b/drivers/base/Makefile
> index 2609ba2..aafe34b 100644
> --- a/drivers/base/Makefile
> +++ b/drivers/base/Makefile
> @@ -13,7 +13,7 @@ obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o
> obj-$(CONFIG_ISA_BUS_API) += isa.o
> obj-$(CONFIG_FW_LOADER) += firmware_class.o
> obj-$(CONFIG_NUMA) += node.o
> -obj-$(CONFIG_MEMORY_HOTPLUG_SPARSE) += memory.o
> +obj-$(CONFIG_MEMORY_DEVICE) += memory.o
> ifeq ($(CONFIG_SYSFS),y)
> obj-$(CONFIG_MODULES) += module.o
> endif
> diff --git a/drivers/base/memory.c b/drivers/base/memory.c
> index dc75de9..fb00965 100644
> --- a/drivers/base/memory.c
> +++ b/drivers/base/memory.c
> @@ -25,10 +25,11 @@
> #include <linux/atomic.h>
> #include <asm/uaccess.h>
>
> -static DEFINE_MUTEX(mem_sysfs_mutex);
> -
> #define MEMORY_CLASS_NAME "memory"
>
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> +static DEFINE_MUTEX(mem_sysfs_mutex);
> +
> #define to_memory_block(dev) container_of(dev, struct memory_block, dev)
>
> static int sections_per_block;
> @@ -381,6 +382,7 @@ static ssize_t show_phys_device(struct device *dev,
> struct memory_block *mem = to_memory_block(dev);
> return sprintf(buf, "%d\n", mem->phys_device);
> }
> +#endif
>
> #ifdef CONFIG_MEMORY_HOTREMOVE
> static ssize_t show_valid_zones(struct device *dev,
> @@ -427,6 +429,7 @@ static ssize_t show_valid_zones(struct device *dev,
> static DEVICE_ATTR(valid_zones, 0444, show_valid_zones, NULL);
> #endif
>
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> static DEVICE_ATTR(phys_index, 0444, show_mem_start_phys_index, NULL);
> static DEVICE_ATTR(state, 0644, show_mem_state, store_mem_state);
> static DEVICE_ATTR(phys_device, 0444, show_phys_device, NULL);
> @@ -474,6 +477,7 @@ store_auto_online_blocks(struct device *dev, struct device_attribute *attr,
>
> static DEVICE_ATTR(auto_online_blocks, 0644, show_auto_online_blocks,
> store_auto_online_blocks);
> +#endif
>
> /*
> * Some architectures will have custom drivers to do this, and
> @@ -557,6 +561,7 @@ static DEVICE_ATTR(soft_offline_page, S_IWUSR, NULL, store_soft_offline_page);
> static DEVICE_ATTR(hard_offline_page, S_IWUSR, NULL, store_hard_offline_page);
> #endif
>
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> /*
> * Note that phys_device is optional. It is here to allow for
> * differentiation between which *physical* devices each
> @@ -723,6 +728,7 @@ out:
> mutex_unlock(&mem_sysfs_mutex);
> return ret;
> }
> +#endif
>
> #ifdef CONFIG_MEMORY_HOTREMOVE
> static void
> @@ -766,11 +772,13 @@ int unregister_memory_section(struct mem_section *section)
> }
> #endif /* CONFIG_MEMORY_HOTREMOVE */
>
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> /* return true if the memory block is offlined, otherwise, return false */
> bool is_memblock_offlined(struct memory_block *mem)
> {
> return mem->state == MEM_OFFLINE;
> }
> +#endif
>
> static struct attribute *memory_root_attrs[] = {
> #ifdef CONFIG_ARCH_MEMORY_PROBE
> @@ -782,8 +790,10 @@ static struct attribute *memory_root_attrs[] = {
> &dev_attr_hard_offline_page.attr,
> #endif
>
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> &dev_attr_block_size_bytes.attr,
> &dev_attr_auto_online_blocks.attr,
> +#endif
> NULL
> };
>
> @@ -799,6 +809,7 @@ static const struct attribute_group *memory_root_attr_groups[] = {
> /*
> * Initialize the sysfs support for memory devices...
> */
> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
> int __init memory_dev_init(void)
> {
> unsigned int i;
> @@ -830,3 +841,20 @@ out:
> printk(KERN_ERR "%s() failed: %d\n", __func__, ret);
> return ret;
> }
> +#else
> +static struct bus_type memory_subsys = {
> + .name = MEMORY_CLASS_NAME,
> + .dev_name = MEMORY_CLASS_NAME,
> +};
> +
> +int __init memory_dev_init(void)
> +{
> + int ret = 0;
> +
> + ret = subsys_system_register(&memory_subsys, memory_root_attr_groups);
> +
> + if (ret)
> + pr_err("%s() failed: %d\n", __func__, ret);
> + return ret;
> +}
> +#endif
> diff --git a/include/linux/memory.h b/include/linux/memory.h
> index 093607f..9fe1089 100644
> --- a/include/linux/memory.h
> +++ b/include/linux/memory.h
> @@ -77,10 +77,14 @@ struct mem_section;
> #define IPC_CALLBACK_PRI 10
>
> #ifndef CONFIG_MEMORY_HOTPLUG_SPARSE
> +#ifdef CONFIG_MEMORY_DEVICE
> +extern int memory_dev_init(void);
> +#else
> static inline int memory_dev_init(void)
> {
> return 0;
> }
> +#endif
> static inline int register_memory_notifier(struct notifier_block *nb)
> {
> return 0;
>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] base memory: introduce CONFIG_MEMORY_DEVICE
2016-10-14 10:53 ` Yisheng Xie
@ 2016-11-02 5:43 ` Yisheng Xie
-1 siblings, 0 replies; 5+ messages in thread
From: Yisheng Xie @ 2016-11-02 5:43 UTC (permalink / raw)
To: akpm, andrew, daniel.kiper, srinivas.kandagatla, gregkh, vkuznets
Cc: linux-kernel, ak, guohanjun, qiuxishi, n-horiguchi, linux-mm
On 2016/10/14 18:53, Yisheng Xie wrote:
>
ping ....
Could someone help to give some comment? Really thanks for that.
Thanks.
Yisheng
>
> On 2016/10/10 19:26, Yisheng Xie wrote:
>> MEMORY_FAILURE do not depend on SPARSEMEM_MANUAL,
>> nor MEMORY_HOTPLUG_SPARSE. However, when I tried to use sysfs:
>> /sys/devices/system/memory/soft_offline_page
>> /sys/devices/system/memory/hard_offline_page
>> to test memory failure function with FLATMEM_MANUAL && MEMORY_FAILURE
>> enabled on arch like i386, it failed for no such sysfs.
>>
>> To make sysfs soft_offline_page usable once MEMORY_FAILURE is enabled,
>> this patch introduces CONFIG_MEMORY_DEVICE, and selects it when
>> MEMORY_FAILURE or MEMORY_HOTPLUG_SPARSE is enabled.
>>
>> Signed-off-by: Yisheng Xie <xieyisheng1@huawei.com>
>> ---
>> drivers/base/Kconfig | 3 +++
>> drivers/base/Makefile | 2 +-
>> drivers/base/memory.c | 32 ++++++++++++++++++++++++++++++--
>> include/linux/memory.h | 4 ++++
>> 4 files changed, 38 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig
>> index fdf44ca..b4eac4e 100644
>> --- a/drivers/base/Kconfig
>> +++ b/drivers/base/Kconfig
>> @@ -271,6 +271,9 @@ config DMA_CMA
>> For more information see <include/linux/dma-contiguous.h>.
>> If unsure, say "n".
>>
>> +config MEMORY_DEVICE
>> + def_bool MEMORY_HOTPLUG_SPARSE || MEMORY_FAILURE
>> +
>> if DMA_CMA
>> comment "Default contiguous memory area size:"
>>
>> diff --git a/drivers/base/Makefile b/drivers/base/Makefile
>> index 2609ba2..aafe34b 100644
>> --- a/drivers/base/Makefile
>> +++ b/drivers/base/Makefile
>> @@ -13,7 +13,7 @@ obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o
>> obj-$(CONFIG_ISA_BUS_API) += isa.o
>> obj-$(CONFIG_FW_LOADER) += firmware_class.o
>> obj-$(CONFIG_NUMA) += node.o
>> -obj-$(CONFIG_MEMORY_HOTPLUG_SPARSE) += memory.o
>> +obj-$(CONFIG_MEMORY_DEVICE) += memory.o
>> ifeq ($(CONFIG_SYSFS),y)
>> obj-$(CONFIG_MODULES) += module.o
>> endif
>> diff --git a/drivers/base/memory.c b/drivers/base/memory.c
>> index dc75de9..fb00965 100644
>> --- a/drivers/base/memory.c
>> +++ b/drivers/base/memory.c
>> @@ -25,10 +25,11 @@
>> #include <linux/atomic.h>
>> #include <asm/uaccess.h>
>>
>> -static DEFINE_MUTEX(mem_sysfs_mutex);
>> -
>> #define MEMORY_CLASS_NAME "memory"
>>
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> +static DEFINE_MUTEX(mem_sysfs_mutex);
>> +
>> #define to_memory_block(dev) container_of(dev, struct memory_block, dev)
>>
>> static int sections_per_block;
>> @@ -381,6 +382,7 @@ static ssize_t show_phys_device(struct device *dev,
>> struct memory_block *mem = to_memory_block(dev);
>> return sprintf(buf, "%d\n", mem->phys_device);
>> }
>> +#endif
>>
>> #ifdef CONFIG_MEMORY_HOTREMOVE
>> static ssize_t show_valid_zones(struct device *dev,
>> @@ -427,6 +429,7 @@ static ssize_t show_valid_zones(struct device *dev,
>> static DEVICE_ATTR(valid_zones, 0444, show_valid_zones, NULL);
>> #endif
>>
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> static DEVICE_ATTR(phys_index, 0444, show_mem_start_phys_index, NULL);
>> static DEVICE_ATTR(state, 0644, show_mem_state, store_mem_state);
>> static DEVICE_ATTR(phys_device, 0444, show_phys_device, NULL);
>> @@ -474,6 +477,7 @@ store_auto_online_blocks(struct device *dev, struct device_attribute *attr,
>>
>> static DEVICE_ATTR(auto_online_blocks, 0644, show_auto_online_blocks,
>> store_auto_online_blocks);
>> +#endif
>>
>> /*
>> * Some architectures will have custom drivers to do this, and
>> @@ -557,6 +561,7 @@ static DEVICE_ATTR(soft_offline_page, S_IWUSR, NULL, store_soft_offline_page);
>> static DEVICE_ATTR(hard_offline_page, S_IWUSR, NULL, store_hard_offline_page);
>> #endif
>>
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> /*
>> * Note that phys_device is optional. It is here to allow for
>> * differentiation between which *physical* devices each
>> @@ -723,6 +728,7 @@ out:
>> mutex_unlock(&mem_sysfs_mutex);
>> return ret;
>> }
>> +#endif
>>
>> #ifdef CONFIG_MEMORY_HOTREMOVE
>> static void
>> @@ -766,11 +772,13 @@ int unregister_memory_section(struct mem_section *section)
>> }
>> #endif /* CONFIG_MEMORY_HOTREMOVE */
>>
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> /* return true if the memory block is offlined, otherwise, return false */
>> bool is_memblock_offlined(struct memory_block *mem)
>> {
>> return mem->state == MEM_OFFLINE;
>> }
>> +#endif
>>
>> static struct attribute *memory_root_attrs[] = {
>> #ifdef CONFIG_ARCH_MEMORY_PROBE
>> @@ -782,8 +790,10 @@ static struct attribute *memory_root_attrs[] = {
>> &dev_attr_hard_offline_page.attr,
>> #endif
>>
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> &dev_attr_block_size_bytes.attr,
>> &dev_attr_auto_online_blocks.attr,
>> +#endif
>> NULL
>> };
>>
>> @@ -799,6 +809,7 @@ static const struct attribute_group *memory_root_attr_groups[] = {
>> /*
>> * Initialize the sysfs support for memory devices...
>> */
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> int __init memory_dev_init(void)
>> {
>> unsigned int i;
>> @@ -830,3 +841,20 @@ out:
>> printk(KERN_ERR "%s() failed: %d\n", __func__, ret);
>> return ret;
>> }
>> +#else
>> +static struct bus_type memory_subsys = {
>> + .name = MEMORY_CLASS_NAME,
>> + .dev_name = MEMORY_CLASS_NAME,
>> +};
>> +
>> +int __init memory_dev_init(void)
>> +{
>> + int ret = 0;
>> +
>> + ret = subsys_system_register(&memory_subsys, memory_root_attr_groups);
>> +
>> + if (ret)
>> + pr_err("%s() failed: %d\n", __func__, ret);
>> + return ret;
>> +}
>> +#endif
>> diff --git a/include/linux/memory.h b/include/linux/memory.h
>> index 093607f..9fe1089 100644
>> --- a/include/linux/memory.h
>> +++ b/include/linux/memory.h
>> @@ -77,10 +77,14 @@ struct mem_section;
>> #define IPC_CALLBACK_PRI 10
>>
>> #ifndef CONFIG_MEMORY_HOTPLUG_SPARSE
>> +#ifdef CONFIG_MEMORY_DEVICE
>> +extern int memory_dev_init(void);
>> +#else
>> static inline int memory_dev_init(void)
>> {
>> return 0;
>> }
>> +#endif
>> static inline int register_memory_notifier(struct notifier_block *nb)
>> {
>> return 0;
>>
>
>
> .
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] base memory: introduce CONFIG_MEMORY_DEVICE
@ 2016-11-02 5:43 ` Yisheng Xie
0 siblings, 0 replies; 5+ messages in thread
From: Yisheng Xie @ 2016-11-02 5:43 UTC (permalink / raw)
To: akpm, andrew, daniel.kiper, srinivas.kandagatla, gregkh, vkuznets
Cc: linux-kernel, ak, guohanjun, qiuxishi, n-horiguchi, linux-mm
On 2016/10/14 18:53, Yisheng Xie wrote:
>
ping ....
Could someone help to give some comment? Really thanks for that.
Thanks.
Yisheng
>
> On 2016/10/10 19:26, Yisheng Xie wrote:
>> MEMORY_FAILURE do not depend on SPARSEMEM_MANUAL,
>> nor MEMORY_HOTPLUG_SPARSE. However, when I tried to use sysfs:
>> /sys/devices/system/memory/soft_offline_page
>> /sys/devices/system/memory/hard_offline_page
>> to test memory failure function with FLATMEM_MANUAL && MEMORY_FAILURE
>> enabled on arch like i386, it failed for no such sysfs.
>>
>> To make sysfs soft_offline_page usable once MEMORY_FAILURE is enabled,
>> this patch introduces CONFIG_MEMORY_DEVICE, and selects it when
>> MEMORY_FAILURE or MEMORY_HOTPLUG_SPARSE is enabled.
>>
>> Signed-off-by: Yisheng Xie <xieyisheng1@huawei.com>
>> ---
>> drivers/base/Kconfig | 3 +++
>> drivers/base/Makefile | 2 +-
>> drivers/base/memory.c | 32 ++++++++++++++++++++++++++++++--
>> include/linux/memory.h | 4 ++++
>> 4 files changed, 38 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig
>> index fdf44ca..b4eac4e 100644
>> --- a/drivers/base/Kconfig
>> +++ b/drivers/base/Kconfig
>> @@ -271,6 +271,9 @@ config DMA_CMA
>> For more information see <include/linux/dma-contiguous.h>.
>> If unsure, say "n".
>>
>> +config MEMORY_DEVICE
>> + def_bool MEMORY_HOTPLUG_SPARSE || MEMORY_FAILURE
>> +
>> if DMA_CMA
>> comment "Default contiguous memory area size:"
>>
>> diff --git a/drivers/base/Makefile b/drivers/base/Makefile
>> index 2609ba2..aafe34b 100644
>> --- a/drivers/base/Makefile
>> +++ b/drivers/base/Makefile
>> @@ -13,7 +13,7 @@ obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o
>> obj-$(CONFIG_ISA_BUS_API) += isa.o
>> obj-$(CONFIG_FW_LOADER) += firmware_class.o
>> obj-$(CONFIG_NUMA) += node.o
>> -obj-$(CONFIG_MEMORY_HOTPLUG_SPARSE) += memory.o
>> +obj-$(CONFIG_MEMORY_DEVICE) += memory.o
>> ifeq ($(CONFIG_SYSFS),y)
>> obj-$(CONFIG_MODULES) += module.o
>> endif
>> diff --git a/drivers/base/memory.c b/drivers/base/memory.c
>> index dc75de9..fb00965 100644
>> --- a/drivers/base/memory.c
>> +++ b/drivers/base/memory.c
>> @@ -25,10 +25,11 @@
>> #include <linux/atomic.h>
>> #include <asm/uaccess.h>
>>
>> -static DEFINE_MUTEX(mem_sysfs_mutex);
>> -
>> #define MEMORY_CLASS_NAME "memory"
>>
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> +static DEFINE_MUTEX(mem_sysfs_mutex);
>> +
>> #define to_memory_block(dev) container_of(dev, struct memory_block, dev)
>>
>> static int sections_per_block;
>> @@ -381,6 +382,7 @@ static ssize_t show_phys_device(struct device *dev,
>> struct memory_block *mem = to_memory_block(dev);
>> return sprintf(buf, "%d\n", mem->phys_device);
>> }
>> +#endif
>>
>> #ifdef CONFIG_MEMORY_HOTREMOVE
>> static ssize_t show_valid_zones(struct device *dev,
>> @@ -427,6 +429,7 @@ static ssize_t show_valid_zones(struct device *dev,
>> static DEVICE_ATTR(valid_zones, 0444, show_valid_zones, NULL);
>> #endif
>>
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> static DEVICE_ATTR(phys_index, 0444, show_mem_start_phys_index, NULL);
>> static DEVICE_ATTR(state, 0644, show_mem_state, store_mem_state);
>> static DEVICE_ATTR(phys_device, 0444, show_phys_device, NULL);
>> @@ -474,6 +477,7 @@ store_auto_online_blocks(struct device *dev, struct device_attribute *attr,
>>
>> static DEVICE_ATTR(auto_online_blocks, 0644, show_auto_online_blocks,
>> store_auto_online_blocks);
>> +#endif
>>
>> /*
>> * Some architectures will have custom drivers to do this, and
>> @@ -557,6 +561,7 @@ static DEVICE_ATTR(soft_offline_page, S_IWUSR, NULL, store_soft_offline_page);
>> static DEVICE_ATTR(hard_offline_page, S_IWUSR, NULL, store_hard_offline_page);
>> #endif
>>
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> /*
>> * Note that phys_device is optional. It is here to allow for
>> * differentiation between which *physical* devices each
>> @@ -723,6 +728,7 @@ out:
>> mutex_unlock(&mem_sysfs_mutex);
>> return ret;
>> }
>> +#endif
>>
>> #ifdef CONFIG_MEMORY_HOTREMOVE
>> static void
>> @@ -766,11 +772,13 @@ int unregister_memory_section(struct mem_section *section)
>> }
>> #endif /* CONFIG_MEMORY_HOTREMOVE */
>>
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> /* return true if the memory block is offlined, otherwise, return false */
>> bool is_memblock_offlined(struct memory_block *mem)
>> {
>> return mem->state == MEM_OFFLINE;
>> }
>> +#endif
>>
>> static struct attribute *memory_root_attrs[] = {
>> #ifdef CONFIG_ARCH_MEMORY_PROBE
>> @@ -782,8 +790,10 @@ static struct attribute *memory_root_attrs[] = {
>> &dev_attr_hard_offline_page.attr,
>> #endif
>>
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> &dev_attr_block_size_bytes.attr,
>> &dev_attr_auto_online_blocks.attr,
>> +#endif
>> NULL
>> };
>>
>> @@ -799,6 +809,7 @@ static const struct attribute_group *memory_root_attr_groups[] = {
>> /*
>> * Initialize the sysfs support for memory devices...
>> */
>> +#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
>> int __init memory_dev_init(void)
>> {
>> unsigned int i;
>> @@ -830,3 +841,20 @@ out:
>> printk(KERN_ERR "%s() failed: %d\n", __func__, ret);
>> return ret;
>> }
>> +#else
>> +static struct bus_type memory_subsys = {
>> + .name = MEMORY_CLASS_NAME,
>> + .dev_name = MEMORY_CLASS_NAME,
>> +};
>> +
>> +int __init memory_dev_init(void)
>> +{
>> + int ret = 0;
>> +
>> + ret = subsys_system_register(&memory_subsys, memory_root_attr_groups);
>> +
>> + if (ret)
>> + pr_err("%s() failed: %d\n", __func__, ret);
>> + return ret;
>> +}
>> +#endif
>> diff --git a/include/linux/memory.h b/include/linux/memory.h
>> index 093607f..9fe1089 100644
>> --- a/include/linux/memory.h
>> +++ b/include/linux/memory.h
>> @@ -77,10 +77,14 @@ struct mem_section;
>> #define IPC_CALLBACK_PRI 10
>>
>> #ifndef CONFIG_MEMORY_HOTPLUG_SPARSE
>> +#ifdef CONFIG_MEMORY_DEVICE
>> +extern int memory_dev_init(void);
>> +#else
>> static inline int memory_dev_init(void)
>> {
>> return 0;
>> }
>> +#endif
>> static inline int register_memory_notifier(struct notifier_block *nb)
>> {
>> return 0;
>>
>
>
> .
>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2016-11-02 5:45 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-10-10 11:26 [PATCH] base memory: introduce CONFIG_MEMORY_DEVICE Yisheng Xie
2016-10-14 10:53 ` Yisheng Xie
2016-10-14 10:53 ` Yisheng Xie
2016-11-02 5:43 ` Yisheng Xie
2016-11-02 5:43 ` Yisheng Xie
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.