Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
On Fri, 2014-08-22 at 15:16 -0700, Andrew Morton wrote: > On Mon, 18 Aug 2014 11:25:36 +0800 Zhang Zhen > wrote: > > > On 2014/8/16 5:37, Toshi Kani wrote: > > > On Wed, 2014-08-13 at 12:10 +0800, Zhang Zhen wrote: > > >> Currently memory-hotplug has two limits: > > >> 1. If the memory block is in ZONE_NORMAL, you can change it to > > >> ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. > > >> 2. If the memory block is in ZONE_MOVABLE, you can change it to > > >> ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. > > >> > > >> With this patch, we can easy to know a memory block can be onlined to > > >> which zone, and don't need to know the above two limits. > > >> > > >> Updated the related Documentation. > > >> > > >> Change v1 -> v2: > > >> - optimize the implementation following Dave Hansen's suggestion > > >> > > >> Signed-off-by: Zhang Zhen > > >> --- > > >> Documentation/ABI/testing/sysfs-devices-memory | 8 > > >> Documentation/memory-hotplug.txt | 4 +- > > >> drivers/base/memory.c | 62 > > >> ++ > > >> include/linux/memory_hotplug.h | 1 + > > >> mm/memory_hotplug.c| 2 +- > > >> 5 files changed, 75 insertions(+), 2 deletions(-) > > >> > > >> diff --git a/Documentation/ABI/testing/sysfs-devices-memory > > >> b/Documentation/ABI/testing/sysfs-devices-memory > > >> index 7405de2..2b2a1d7 100644 > > >> --- a/Documentation/ABI/testing/sysfs-devices-memory > > >> +++ b/Documentation/ABI/testing/sysfs-devices-memory > > >> @@ -61,6 +61,14 @@ Users:hotplug memory remove tools > > >> > > >> http://www.ibm.com/developerworks/wikis/display/LinuxP/powerpc-utils > > >> > > >> > > >> +What: /sys/devices/system/memory/memoryX/zones_online_to > > > > > > I think this name is a bit confusing. How about "valid_online_types"? > > > > > Thanks for your suggestion. > > > > This patch has been added to -mm tree. > > If most people think so, i would like to modify the interface name. > > If not, let's leave it as it is. > > Yes, the name could be better. Do we actually need "online" in there? > How about "valid_zones"? I suggested using "online" because a user specifies a zone type during an online operation as follows. $ echo online_movable > /sys/devices/system/memory/memoryXXX/state I also like "valid_zones" and it well represents what it is (and the name is shorter :-). I am fine with this name as well. Thanks, -Toshi -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
On 2014/8/23 6:16, Andrew Morton wrote: > On Mon, 18 Aug 2014 11:25:36 +0800 Zhang Zhen > wrote: > >> On 2014/8/16 5:37, Toshi Kani wrote: >>> On Wed, 2014-08-13 at 12:10 +0800, Zhang Zhen wrote: Currently memory-hotplug has two limits: 1. If the memory block is in ZONE_NORMAL, you can change it to ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. 2. If the memory block is in ZONE_MOVABLE, you can change it to ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. With this patch, we can easy to know a memory block can be onlined to which zone, and don't need to know the above two limits. Updated the related Documentation. Change v1 -> v2: - optimize the implementation following Dave Hansen's suggestion Signed-off-by: Zhang Zhen --- Documentation/ABI/testing/sysfs-devices-memory | 8 Documentation/memory-hotplug.txt | 4 +- drivers/base/memory.c | 62 ++ include/linux/memory_hotplug.h | 1 + mm/memory_hotplug.c| 2 +- 5 files changed, 75 insertions(+), 2 deletions(-) diff --git a/Documentation/ABI/testing/sysfs-devices-memory b/Documentation/ABI/testing/sysfs-devices-memory index 7405de2..2b2a1d7 100644 --- a/Documentation/ABI/testing/sysfs-devices-memory +++ b/Documentation/ABI/testing/sysfs-devices-memory @@ -61,6 +61,14 @@ Users: hotplug memory remove tools http://www.ibm.com/developerworks/wikis/display/LinuxP/powerpc-utils +What: /sys/devices/system/memory/memoryX/zones_online_to >>> >>> I think this name is a bit confusing. How about "valid_online_types"? >>> >> Thanks for your suggestion. >> >> This patch has been added to -mm tree. >> If most people think so, i would like to modify the interface name. >> If not, let's leave it as it is. > > Yes, the name could be better. Do we actually need "online" in there? > How about "valid_zones"? Ok, i will change it to valid_zones. > > Also, it's not really clear to me why we need this sysfs file at all. > Do people really read sysfs files, make onlining decisions and manually > type in commands? Or is this stuff all automated? If the latter then > the script can take care of all this? For example, attempt to online > the memory into the desired zone and report failure if that didn't > succeed? Just like Dave Hansen says, the scripts should be changed when we add a new zone type. And ZONE_MOVABLE may be missed by the scripts writer. > > IOW, please update the changelog to show > > a) example output from >/sys/devices/system/memory/memoryX/whatever-we-call-it and > > b) example use-cases which help reviewers understand why this >feature will be valuable to users. Sorry, this patch has been added to -next tree. I can't modify the changelog. > > Also, please do address the error which Yasuaki Ishimatsu identified. > Yeah, i have been waiting for http://ozlabs.org/~akpm/mmots/broken-out/memory-hotplug-add-sysfs-zones_online_to-attribute-fix-2.patch added to -mm tree. So i can send a patch based on -mm tree to address the error which Yasuaki Ishimatsu identified. Otherwise, conflicts may occur. Thanks! > > . > -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
On 08/22/2014 03:16 PM, Andrew Morton wrote: > Also, it's not really clear to me why we need this sysfs file at all. > Do people really read sysfs files, make onlining decisions and manually > type in commands? Or is this stuff all automated? If the latter then > the script can take care of all this? For example, attempt to online > the memory into the desired zone and report failure if that didn't > succeed? I guess we can just iterate over all possible zone types from userspace until we find one. Seems a bit hokey, but it would work at least until we add a new zone type and we have to teach the scripts about the new type. But that's a pretty rare event I guess. Let's hope the script writers get this right, and don't make omissions like ZONE_MOVABLE because it's not that common in practice. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
On Mon, 18 Aug 2014 11:25:36 +0800 Zhang Zhen wrote: > On 2014/8/16 5:37, Toshi Kani wrote: > > On Wed, 2014-08-13 at 12:10 +0800, Zhang Zhen wrote: > >> Currently memory-hotplug has two limits: > >> 1. If the memory block is in ZONE_NORMAL, you can change it to > >> ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. > >> 2. If the memory block is in ZONE_MOVABLE, you can change it to > >> ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. > >> > >> With this patch, we can easy to know a memory block can be onlined to > >> which zone, and don't need to know the above two limits. > >> > >> Updated the related Documentation. > >> > >> Change v1 -> v2: > >> - optimize the implementation following Dave Hansen's suggestion > >> > >> Signed-off-by: Zhang Zhen > >> --- > >> Documentation/ABI/testing/sysfs-devices-memory | 8 > >> Documentation/memory-hotplug.txt | 4 +- > >> drivers/base/memory.c | 62 > >> ++ > >> include/linux/memory_hotplug.h | 1 + > >> mm/memory_hotplug.c| 2 +- > >> 5 files changed, 75 insertions(+), 2 deletions(-) > >> > >> diff --git a/Documentation/ABI/testing/sysfs-devices-memory > >> b/Documentation/ABI/testing/sysfs-devices-memory > >> index 7405de2..2b2a1d7 100644 > >> --- a/Documentation/ABI/testing/sysfs-devices-memory > >> +++ b/Documentation/ABI/testing/sysfs-devices-memory > >> @@ -61,6 +61,14 @@ Users: hotplug memory remove tools > >> > >> http://www.ibm.com/developerworks/wikis/display/LinuxP/powerpc-utils > >> > >> > >> +What: /sys/devices/system/memory/memoryX/zones_online_to > > > > I think this name is a bit confusing. How about "valid_online_types"? > > > Thanks for your suggestion. > > This patch has been added to -mm tree. > If most people think so, i would like to modify the interface name. > If not, let's leave it as it is. Yes, the name could be better. Do we actually need "online" in there? How about "valid_zones"? Also, it's not really clear to me why we need this sysfs file at all. Do people really read sysfs files, make onlining decisions and manually type in commands? Or is this stuff all automated? If the latter then the script can take care of all this? For example, attempt to online the memory into the desired zone and report failure if that didn't succeed? IOW, please update the changelog to show a) example output from /sys/devices/system/memory/memoryX/whatever-we-call-it and b) example use-cases which help reviewers understand why this feature will be valuable to users. Also, please do address the error which Yasuaki Ishimatsu identified. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
(2014/08/19 16:35), Zhang Zhen wrote: On 2014/8/18 14:11, Yasuaki Ishimatsu wrote: (2014/08/13 13:10), Zhang Zhen wrote: Currently memory-hotplug has two limits: 1. If the memory block is in ZONE_NORMAL, you can change it to ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. 2. If the memory block is in ZONE_MOVABLE, you can change it to ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. With this patch, we can easy to know a memory block can be onlined to which zone, and don't need to know the above two limits. Updated the related Documentation. Change v1 -> v2: - optimize the implementation following Dave Hansen's suggestion Signed-off-by: Zhang Zhen --- Documentation/ABI/testing/sysfs-devices-memory | 8 Documentation/memory-hotplug.txt | 4 +- drivers/base/memory.c | 62 ++ include/linux/memory_hotplug.h | 1 + mm/memory_hotplug.c| 2 +- 5 files changed, 75 insertions(+), 2 deletions(-) diff --git a/Documentation/ABI/testing/sysfs-devices-memory b/Documentation/ABI/testing/sysfs-devices-memory index 7405de2..2b2a1d7 100644 --- a/Documentation/ABI/testing/sysfs-devices-memory +++ b/Documentation/ABI/testing/sysfs-devices-memory @@ -61,6 +61,14 @@ Users:hotplug memory remove tools http://www.ibm.com/developerworks/wikis/display/LinuxP/powerpc-utils +What: /sys/devices/system/memory/memoryX/zones_online_to +Date: July 2014 +Contact:Zhang Zhen +Description: +The file /sys/devices/system/memory/memoryX/zones_online_to +is read-only and is designed to show which zone this memory block can +be onlined to. + What:/sys/devices/system/memoryX/nodeY Date:October 2009 Contact:Linux Memory Management list diff --git a/Documentation/memory-hotplug.txt b/Documentation/memory-hotplug.txt index 45134dc..5b34e33 100644 --- a/Documentation/memory-hotplug.txt +++ b/Documentation/memory-hotplug.txt @@ -155,6 +155,7 @@ Under each memory block, you can see 4 files: /sys/devices/system/memory/memoryXXX/phys_device /sys/devices/system/memory/memoryXXX/state /sys/devices/system/memory/memoryXXX/removable +/sys/devices/system/memory/memoryXXX/zones_online_to 'phys_index' : read-only and contains memory block id, same as XXX. 'state' : read-write @@ -170,6 +171,8 @@ Under each memory block, you can see 4 files: block is removable and a value of 0 indicates that it is not removable. A memory block is removable only if every section in the block is removable. +'zones_online_to' : read-only: designed to show which zone this memory block +can be onlined to. NOTE: These directories/files appear after physical memory hotplug phase. @@ -408,7 +411,6 @@ node if necessary. - allowing memory hot-add to ZONE_MOVABLE. maybe we need some switch like sysctl or new control file. - showing memory block and physical device relationship. - - showing memory block is under ZONE_MOVABLE or not - test and make it better memory offlining. - support HugeTLB page migration and offlining. - memmap removing at memory offline. diff --git a/drivers/base/memory.c b/drivers/base/memory.c index a2e13e2..b5d693f 100644 --- a/drivers/base/memory.c +++ b/drivers/base/memory.c @@ -373,10 +373,71 @@ static ssize_t show_phys_device(struct device *dev, return sprintf(buf, "%d\n", mem->phys_device); } +static int __zones_online_to(unsigned long end_pfn, +struct page *first_page, unsigned long nr_pages) +{ +struct zone *zone_next; + +/*The mem block is the last block of memory.*/ +if (!pfn_valid(end_pfn + 1)) +return 1; The check is not enough if memory has hole as follows: PFN 0x00 0xd0 0xe0 0xf0 +-+-+-+ zone type | Normal| hole| Normal| +-+-+-+ In this case, 0xd1 is invalid pfn. But __zones_online_to should return 0 since 0xe0-0xf0 is Normal zone. Thanks, Yasuaki Ishimatsu You are right, it is not enough. Here we should make a check as follows. if ((end_pfn + 1) > zone_end_pfn(page_zone(first_page))) return 1; I will send a patch to fix it. It's looks good to me. Thanks, Yasuaki Ishimatsu Thanks ! +zone_next = page_zone(first_page + nr_pages); +if (zone_idx(zone_next) == ZONE_MOVABLE) +return 1; +return 0; +} + +static ssize_t show_zones_online_to(struct device *dev, +struct device_attribute *attr, char *buf) +{ +struct memory_block *mem = to_memory_block(dev); +unsigned long start_pfn, end_pfn; +unsigned long nr_pages = PAGES_PER_SECTION * sections_per_block; +struct
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
On 2014/8/18 14:11, Yasuaki Ishimatsu wrote: > (2014/08/13 13:10), Zhang Zhen wrote: >> Currently memory-hotplug has two limits: >> 1. If the memory block is in ZONE_NORMAL, you can change it to >> ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. >> 2. If the memory block is in ZONE_MOVABLE, you can change it to >> ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. >> >> With this patch, we can easy to know a memory block can be onlined to >> which zone, and don't need to know the above two limits. >> >> Updated the related Documentation. >> >> Change v1 -> v2: >> - optimize the implementation following Dave Hansen's suggestion >> >> Signed-off-by: Zhang Zhen >> --- >> Documentation/ABI/testing/sysfs-devices-memory | 8 >> Documentation/memory-hotplug.txt | 4 +- >> drivers/base/memory.c | 62 >> ++ >> include/linux/memory_hotplug.h | 1 + >> mm/memory_hotplug.c| 2 +- >> 5 files changed, 75 insertions(+), 2 deletions(-) >> >> diff --git a/Documentation/ABI/testing/sysfs-devices-memory >> b/Documentation/ABI/testing/sysfs-devices-memory >> index 7405de2..2b2a1d7 100644 >> --- a/Documentation/ABI/testing/sysfs-devices-memory >> +++ b/Documentation/ABI/testing/sysfs-devices-memory >> @@ -61,6 +61,14 @@ Users:hotplug memory remove tools >> >> http://www.ibm.com/developerworks/wikis/display/LinuxP/powerpc-utils >> >> >> +What: /sys/devices/system/memory/memoryX/zones_online_to >> +Date: July 2014 >> +Contact:Zhang Zhen >> +Description: >> +The file /sys/devices/system/memory/memoryX/zones_online_to >> +is read-only and is designed to show which zone this memory block >> can >> +be onlined to. >> + >> What:/sys/devices/system/memoryX/nodeY >> Date:October 2009 >> Contact:Linux Memory Management list >> diff --git a/Documentation/memory-hotplug.txt >> b/Documentation/memory-hotplug.txt >> index 45134dc..5b34e33 100644 >> --- a/Documentation/memory-hotplug.txt >> +++ b/Documentation/memory-hotplug.txt >> @@ -155,6 +155,7 @@ Under each memory block, you can see 4 files: >> /sys/devices/system/memory/memoryXXX/phys_device >> /sys/devices/system/memory/memoryXXX/state >> /sys/devices/system/memory/memoryXXX/removable >> +/sys/devices/system/memory/memoryXXX/zones_online_to >> >> 'phys_index' : read-only and contains memory block id, same as XXX. >> 'state' : read-write >> @@ -170,6 +171,8 @@ Under each memory block, you can see 4 files: >> block is removable and a value of 0 indicates that >> it is not removable. A memory block is removable only >> if >> every section in the block is removable. >> +'zones_online_to' : read-only: designed to show which zone this memory block >> +can be onlined to. >> >> NOTE: >> These directories/files appear after physical memory hotplug phase. >> @@ -408,7 +411,6 @@ node if necessary. >> - allowing memory hot-add to ZONE_MOVABLE. maybe we need some switch like >> sysctl or new control file. >> - showing memory block and physical device relationship. >> - - showing memory block is under ZONE_MOVABLE or not >> - test and make it better memory offlining. >> - support HugeTLB page migration and offlining. >> - memmap removing at memory offline. >> diff --git a/drivers/base/memory.c b/drivers/base/memory.c >> index a2e13e2..b5d693f 100644 >> --- a/drivers/base/memory.c >> +++ b/drivers/base/memory.c >> @@ -373,10 +373,71 @@ static ssize_t show_phys_device(struct device *dev, >> return sprintf(buf, "%d\n", mem->phys_device); >> } >> >> +static int __zones_online_to(unsigned long end_pfn, >> +struct page *first_page, unsigned long nr_pages) >> +{ >> +struct zone *zone_next; >> + > >> +/*The mem block is the last block of memory.*/ >> +if (!pfn_valid(end_pfn + 1)) >> +return 1; > > The check is not enough if memory has hole as follows: > > PFN 0x00 0xd0 0xe0 0xf0 > +-+-+-+ > zone type | Normal| hole| Normal| > +-+-+-+ > > In this case, 0xd1 is invalid pfn. But __zones_online_to should return 0 > since 0xe0-0xf0 is Normal zone. > > Thanks, > Yasuaki Ishimatsu > You are right, it is not enough. Here we should make a check as follows. if ((end_pfn + 1) > zone_end_pfn(page_zone(first_page))) return 1; I will send a patch to fix it. Thanks ! > >> +zone_next = page_zone(first_page + nr_pages); >> +if (zone_idx(zone_next) == ZONE_MOVABLE) >> +return 1; >> +return 0; >> +} >> + >> +static ssize_t show_zones_online_to(struct device *dev, >> +struct device_attribute
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
On 2014/8/19 5:48, David Rientjes wrote: > On Wed, 13 Aug 2014, Zhang Zhen wrote: > >> Currently memory-hotplug has two limits: >> 1. If the memory block is in ZONE_NORMAL, you can change it to >> ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. >> 2. If the memory block is in ZONE_MOVABLE, you can change it to >> ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. >> >> With this patch, we can easy to know a memory block can be onlined to >> which zone, and don't need to know the above two limits. >> >> Updated the related Documentation. >> >> Change v1 -> v2: >> - optimize the implementation following Dave Hansen's suggestion >> >> Signed-off-by: Zhang Zhen > > linux-next build failure: > > drivers/built-in.o: In function `show_zones_online_to': > memory.c:(.text+0x13ee09): undefined reference to `test_pages_in_a_zone' > The function implementation in mm/memory_hotplug.c is only built if CONFIG_MEMORY_HOTREMOVE is enabled. A fix has been proposed. http://ozlabs.org/~akpm/mmots/broken-out/memory-hotplug-add-sysfs-zones_online_to-attribute-fix-2.patch Thanks! > -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
On Wed, 13 Aug 2014, Zhang Zhen wrote: > Currently memory-hotplug has two limits: > 1. If the memory block is in ZONE_NORMAL, you can change it to > ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. > 2. If the memory block is in ZONE_MOVABLE, you can change it to > ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. > > With this patch, we can easy to know a memory block can be onlined to > which zone, and don't need to know the above two limits. > > Updated the related Documentation. > > Change v1 -> v2: > - optimize the implementation following Dave Hansen's suggestion > > Signed-off-by: Zhang Zhen linux-next build failure: drivers/built-in.o: In function `show_zones_online_to': memory.c:(.text+0x13ee09): undefined reference to `test_pages_in_a_zone' -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
(2014/08/18 12:25), Zhang Zhen wrote: On 2014/8/16 5:37, Toshi Kani wrote: On Wed, 2014-08-13 at 12:10 +0800, Zhang Zhen wrote: Currently memory-hotplug has two limits: 1. If the memory block is in ZONE_NORMAL, you can change it to ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. 2. If the memory block is in ZONE_MOVABLE, you can change it to ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. With this patch, we can easy to know a memory block can be onlined to which zone, and don't need to know the above two limits. Updated the related Documentation. Change v1 -> v2: - optimize the implementation following Dave Hansen's suggestion Signed-off-by: Zhang Zhen --- Documentation/ABI/testing/sysfs-devices-memory | 8 Documentation/memory-hotplug.txt | 4 +- drivers/base/memory.c | 62 ++ include/linux/memory_hotplug.h | 1 + mm/memory_hotplug.c| 2 +- 5 files changed, 75 insertions(+), 2 deletions(-) diff --git a/Documentation/ABI/testing/sysfs-devices-memory b/Documentation/ABI/testing/sysfs-devices-memory index 7405de2..2b2a1d7 100644 --- a/Documentation/ABI/testing/sysfs-devices-memory +++ b/Documentation/ABI/testing/sysfs-devices-memory @@ -61,6 +61,14 @@ Users: hotplug memory remove tools http://www.ibm.com/developerworks/wikis/display/LinuxP/powerpc-utils +What: /sys/devices/system/memory/memoryX/zones_online_to I think this name is a bit confusing. How about "valid_online_types"? Thanks for your suggestion. This patch has been added to -mm tree. If most people think so, i would like to modify the interface name. I like Toshi's idea (valid_online_types). Thanks, Yasuaki Ishimatsu If not, let's leave it as it is. Best regards! Thanks, -Toshi -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majord...@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: mailto:"d...@kvack.org";> em...@kvack.org -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
(2014/08/13 13:10), Zhang Zhen wrote: Currently memory-hotplug has two limits: 1. If the memory block is in ZONE_NORMAL, you can change it to ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. 2. If the memory block is in ZONE_MOVABLE, you can change it to ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. With this patch, we can easy to know a memory block can be onlined to which zone, and don't need to know the above two limits. Updated the related Documentation. Change v1 -> v2: - optimize the implementation following Dave Hansen's suggestion Signed-off-by: Zhang Zhen --- Documentation/ABI/testing/sysfs-devices-memory | 8 Documentation/memory-hotplug.txt | 4 +- drivers/base/memory.c | 62 ++ include/linux/memory_hotplug.h | 1 + mm/memory_hotplug.c| 2 +- 5 files changed, 75 insertions(+), 2 deletions(-) diff --git a/Documentation/ABI/testing/sysfs-devices-memory b/Documentation/ABI/testing/sysfs-devices-memory index 7405de2..2b2a1d7 100644 --- a/Documentation/ABI/testing/sysfs-devices-memory +++ b/Documentation/ABI/testing/sysfs-devices-memory @@ -61,6 +61,14 @@ Users: hotplug memory remove tools http://www.ibm.com/developerworks/wikis/display/LinuxP/powerpc-utils +What: /sys/devices/system/memory/memoryX/zones_online_to +Date: July 2014 +Contact: Zhang Zhen +Description: + The file /sys/devices/system/memory/memoryX/zones_online_to + is read-only and is designed to show which zone this memory block can + be onlined to. + What: /sys/devices/system/memoryX/nodeY Date: October 2009 Contact: Linux Memory Management list diff --git a/Documentation/memory-hotplug.txt b/Documentation/memory-hotplug.txt index 45134dc..5b34e33 100644 --- a/Documentation/memory-hotplug.txt +++ b/Documentation/memory-hotplug.txt @@ -155,6 +155,7 @@ Under each memory block, you can see 4 files: /sys/devices/system/memory/memoryXXX/phys_device /sys/devices/system/memory/memoryXXX/state /sys/devices/system/memory/memoryXXX/removable +/sys/devices/system/memory/memoryXXX/zones_online_to 'phys_index' : read-only and contains memory block id, same as XXX. 'state' : read-write @@ -170,6 +171,8 @@ Under each memory block, you can see 4 files: block is removable and a value of 0 indicates that it is not removable. A memory block is removable only if every section in the block is removable. +'zones_online_to' : read-only: designed to show which zone this memory block + can be onlined to. NOTE: These directories/files appear after physical memory hotplug phase. @@ -408,7 +411,6 @@ node if necessary. - allowing memory hot-add to ZONE_MOVABLE. maybe we need some switch like sysctl or new control file. - showing memory block and physical device relationship. - - showing memory block is under ZONE_MOVABLE or not - test and make it better memory offlining. - support HugeTLB page migration and offlining. - memmap removing at memory offline. diff --git a/drivers/base/memory.c b/drivers/base/memory.c index a2e13e2..b5d693f 100644 --- a/drivers/base/memory.c +++ b/drivers/base/memory.c @@ -373,10 +373,71 @@ static ssize_t show_phys_device(struct device *dev, return sprintf(buf, "%d\n", mem->phys_device); } +static int __zones_online_to(unsigned long end_pfn, + struct page *first_page, unsigned long nr_pages) +{ + struct zone *zone_next; + + /*The mem block is the last block of memory.*/ + if (!pfn_valid(end_pfn + 1)) + return 1; The check is not enough if memory has hole as follows: PFN 0x00 0xd0 0xe0 0xf0 +-+-+-+ zone type | Normal| hole| Normal| +-+-+-+ In this case, 0xd1 is invalid pfn. But __zones_online_to should return 0 since 0xe0-0xf0 is Normal zone. Thanks, Yasuaki Ishimatsu + zone_next = page_zone(first_page + nr_pages); + if (zone_idx(zone_next) == ZONE_MOVABLE) + return 1; + return 0; +} + +static ssize_t show_zones_online_to(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct memory_block *mem = to_memory_block(dev); + unsigned long start_pfn, end_pfn; + unsigned long nr_pages = PAGES_PER_SECTION * sections_per_block; + struct page *first_page; + struct zone *zone, *zone_prev; + + start_pfn = section_nr_to_pfn(mem->start_section_nr); + end_pfn = start_pfn + nr_pages; + first_page = pfn_to_page(start_pfn); + + /*The block contains m
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
On 2014/8/16 5:37, Toshi Kani wrote: > On Wed, 2014-08-13 at 12:10 +0800, Zhang Zhen wrote: >> Currently memory-hotplug has two limits: >> 1. If the memory block is in ZONE_NORMAL, you can change it to >> ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. >> 2. If the memory block is in ZONE_MOVABLE, you can change it to >> ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. >> >> With this patch, we can easy to know a memory block can be onlined to >> which zone, and don't need to know the above two limits. >> >> Updated the related Documentation. >> >> Change v1 -> v2: >> - optimize the implementation following Dave Hansen's suggestion >> >> Signed-off-by: Zhang Zhen >> --- >> Documentation/ABI/testing/sysfs-devices-memory | 8 >> Documentation/memory-hotplug.txt | 4 +- >> drivers/base/memory.c | 62 >> ++ >> include/linux/memory_hotplug.h | 1 + >> mm/memory_hotplug.c| 2 +- >> 5 files changed, 75 insertions(+), 2 deletions(-) >> >> diff --git a/Documentation/ABI/testing/sysfs-devices-memory >> b/Documentation/ABI/testing/sysfs-devices-memory >> index 7405de2..2b2a1d7 100644 >> --- a/Documentation/ABI/testing/sysfs-devices-memory >> +++ b/Documentation/ABI/testing/sysfs-devices-memory >> @@ -61,6 +61,14 @@ Users:hotplug memory remove tools >> >> http://www.ibm.com/developerworks/wikis/display/LinuxP/powerpc-utils >> >> >> +What: /sys/devices/system/memory/memoryX/zones_online_to > > I think this name is a bit confusing. How about "valid_online_types"? > Thanks for your suggestion. This patch has been added to -mm tree. If most people think so, i would like to modify the interface name. If not, let's leave it as it is. Best regards! > Thanks, > -Toshi > > > > -- > To unsubscribe, send a message with 'unsubscribe linux-mm' in > the body to majord...@kvack.org. For more info on Linux MM, > see: http://www.linux-mm.org/ . > Don't email: mailto:"d...@kvack.org";> em...@kvack.org > > -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
On Wed, 2014-08-13 at 12:10 +0800, Zhang Zhen wrote: > Currently memory-hotplug has two limits: > 1. If the memory block is in ZONE_NORMAL, you can change it to > ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. > 2. If the memory block is in ZONE_MOVABLE, you can change it to > ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. > > With this patch, we can easy to know a memory block can be onlined to > which zone, and don't need to know the above two limits. > > Updated the related Documentation. > > Change v1 -> v2: > - optimize the implementation following Dave Hansen's suggestion > > Signed-off-by: Zhang Zhen > --- > Documentation/ABI/testing/sysfs-devices-memory | 8 > Documentation/memory-hotplug.txt | 4 +- > drivers/base/memory.c | 62 > ++ > include/linux/memory_hotplug.h | 1 + > mm/memory_hotplug.c| 2 +- > 5 files changed, 75 insertions(+), 2 deletions(-) > > diff --git a/Documentation/ABI/testing/sysfs-devices-memory > b/Documentation/ABI/testing/sysfs-devices-memory > index 7405de2..2b2a1d7 100644 > --- a/Documentation/ABI/testing/sysfs-devices-memory > +++ b/Documentation/ABI/testing/sysfs-devices-memory > @@ -61,6 +61,14 @@ Users: hotplug memory remove tools > > http://www.ibm.com/developerworks/wikis/display/LinuxP/powerpc-utils > > > +What: /sys/devices/system/memory/memoryX/zones_online_to I think this name is a bit confusing. How about "valid_online_types"? Thanks, -Toshi -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[PATCH v2] memory-hotplug: add sysfs zones_online_to attribute
Currently memory-hotplug has two limits: 1. If the memory block is in ZONE_NORMAL, you can change it to ZONE_MOVABLE, but this memory block must be adjacent to ZONE_MOVABLE. 2. If the memory block is in ZONE_MOVABLE, you can change it to ZONE_NORMAL, but this memory block must be adjacent to ZONE_NORMAL. With this patch, we can easy to know a memory block can be onlined to which zone, and don't need to know the above two limits. Updated the related Documentation. Change v1 -> v2: - optimize the implementation following Dave Hansen's suggestion Signed-off-by: Zhang Zhen --- Documentation/ABI/testing/sysfs-devices-memory | 8 Documentation/memory-hotplug.txt | 4 +- drivers/base/memory.c | 62 ++ include/linux/memory_hotplug.h | 1 + mm/memory_hotplug.c| 2 +- 5 files changed, 75 insertions(+), 2 deletions(-) diff --git a/Documentation/ABI/testing/sysfs-devices-memory b/Documentation/ABI/testing/sysfs-devices-memory index 7405de2..2b2a1d7 100644 --- a/Documentation/ABI/testing/sysfs-devices-memory +++ b/Documentation/ABI/testing/sysfs-devices-memory @@ -61,6 +61,14 @@ Users: hotplug memory remove tools http://www.ibm.com/developerworks/wikis/display/LinuxP/powerpc-utils +What: /sys/devices/system/memory/memoryX/zones_online_to +Date: July 2014 +Contact: Zhang Zhen +Description: + The file /sys/devices/system/memory/memoryX/zones_online_to + is read-only and is designed to show which zone this memory block can + be onlined to. + What: /sys/devices/system/memoryX/nodeY Date: October 2009 Contact: Linux Memory Management list diff --git a/Documentation/memory-hotplug.txt b/Documentation/memory-hotplug.txt index 45134dc..5b34e33 100644 --- a/Documentation/memory-hotplug.txt +++ b/Documentation/memory-hotplug.txt @@ -155,6 +155,7 @@ Under each memory block, you can see 4 files: /sys/devices/system/memory/memoryXXX/phys_device /sys/devices/system/memory/memoryXXX/state /sys/devices/system/memory/memoryXXX/removable +/sys/devices/system/memory/memoryXXX/zones_online_to 'phys_index' : read-only and contains memory block id, same as XXX. 'state' : read-write @@ -170,6 +171,8 @@ Under each memory block, you can see 4 files: block is removable and a value of 0 indicates that it is not removable. A memory block is removable only if every section in the block is removable. +'zones_online_to' : read-only: designed to show which zone this memory block + can be onlined to. NOTE: These directories/files appear after physical memory hotplug phase. @@ -408,7 +411,6 @@ node if necessary. - allowing memory hot-add to ZONE_MOVABLE. maybe we need some switch like sysctl or new control file. - showing memory block and physical device relationship. - - showing memory block is under ZONE_MOVABLE or not - test and make it better memory offlining. - support HugeTLB page migration and offlining. - memmap removing at memory offline. diff --git a/drivers/base/memory.c b/drivers/base/memory.c index a2e13e2..b5d693f 100644 --- a/drivers/base/memory.c +++ b/drivers/base/memory.c @@ -373,10 +373,71 @@ static ssize_t show_phys_device(struct device *dev, return sprintf(buf, "%d\n", mem->phys_device); } +static int __zones_online_to(unsigned long end_pfn, + struct page *first_page, unsigned long nr_pages) +{ + struct zone *zone_next; + + /*The mem block is the last block of memory.*/ + if (!pfn_valid(end_pfn + 1)) + return 1; + zone_next = page_zone(first_page + nr_pages); + if (zone_idx(zone_next) == ZONE_MOVABLE) + return 1; + return 0; +} + +static ssize_t show_zones_online_to(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct memory_block *mem = to_memory_block(dev); + unsigned long start_pfn, end_pfn; + unsigned long nr_pages = PAGES_PER_SECTION * sections_per_block; + struct page *first_page; + struct zone *zone, *zone_prev; + + start_pfn = section_nr_to_pfn(mem->start_section_nr); + end_pfn = start_pfn + nr_pages; + first_page = pfn_to_page(start_pfn); + + /*The block contains more than one zone can not be offlined.*/ + if (!test_pages_in_a_zone(start_pfn, end_pfn)) + return sprintf(buf, "none\n"); + + zone = page_zone(first_page); + +#ifdef CONFIG_HIGHMEM + if (zone_idx(zone) == ZONE_HIGHMEM) { + if (__zones_online_to(end_pfn, first_page, nr_pages)) + return sprintf(buf, "%s %s\n", + zone->name, (zone + 1)->name); + } +#else + if (zone_i