| Message ID | 20190423203843.2898-1-pasha.tatashin@soleen.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | arm64: configurable sparsemem section size | expand |
On 04/24/2019 02:08 AM, Pavel Tatashin wrote: > sparsemem section size determines the maximum size and alignment that > is allowed to offline/online memory block. The bigger the size the less > the clutter in /sys/devices/system/memory/*. On the other hand, however, > there is less flexability in what granules of memory can be added and > removed. Is there any scenario where less than a 1GB needs to be added on arm64 ? > > Recently, it was enabled in Linux to hotadd persistent memory that > can be either real NV device, or reserved from regular System RAM > and has identity of devdax. devdax (even ZONE_DEVICE) support has not been enabled on arm64 yet. > > The problem is that because ARM64's section size is 1G, and devdax must > have 2M label section, the first 1G is always missed when device is > attached, because it is not 1G aligned. devdax has to be 2M aligned ? Does Linux enforce that right now ? > > Allow, better flexibility by making section size configurable. Unless 2M is being enforced from Linux not sure why this is necessary at the moment. > > Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com> > --- > arch/arm64/Kconfig | 10 ++++++++++ > arch/arm64/include/asm/sparsemem.h | 2 +- > 2 files changed, 11 insertions(+), 1 deletion(-) > > diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig > index b5d8cf57e220..a0c5b9d13a7f 100644 > --- a/arch/arm64/Kconfig > +++ b/arch/arm64/Kconfig > @@ -801,6 +801,16 @@ config ARM64_PA_BITS > default 48 if ARM64_PA_BITS_48 > default 52 if ARM64_PA_BITS_52 > > +config ARM64_SECTION_SIZE_BITS > + int "sparsemem section size shift" > + range 27 30 27 and 28 do not even compile for ARM64_64_PAGES because of MAX_ORDER and SECTION_SIZE mismatch.
<resending> from original email On Wed, Apr 24, 2019 at 3:48 PM Pavel Tatashin <patatash@linux.microsoft.com> wrote: > > On Wed, Apr 24, 2019 at 5:07 AM Anshuman Khandual > <anshuman.khandual@arm.com> wrote: > > > > On 04/24/2019 02:08 AM, Pavel Tatashin wrote: > > > sparsemem section size determines the maximum size and alignment that > > > is allowed to offline/online memory block. The bigger the size the less > > > the clutter in /sys/devices/system/memory/*. On the other hand, however, > > > there is less flexability in what granules of memory can be added and > > > removed. > > > > Is there any scenario where less than a 1GB needs to be added on arm64 ? > > Yes, DAX hotplug loses 1G of memory without allowing smaller sections. > Machines on which we are going to be using this functionality have 8G > of System RAM, therefore losing 1G is a big problem. > > For details about using scenario see this cover letter: > https://lore.kernel.org/lkml/20190421014429.31206-1-pasha.tatashin@soleen.com/ > > > > > > > > > Recently, it was enabled in Linux to hotadd persistent memory that > > > can be either real NV device, or reserved from regular System RAM > > > and has identity of devdax. > > > > devdax (even ZONE_DEVICE) support has not been enabled on arm64 yet. > > Correct, I use your patches to enable ZONE_DEVICE, and thus devdax on ARM64: > https://lore.kernel.org/lkml/1554265806-11501-1-git-send-email-anshuman.khandual@arm.com/ > > > > > > > > > The problem is that because ARM64's section size is 1G, and devdax must > > > have 2M label section, the first 1G is always missed when device is > > > attached, because it is not 1G aligned. > > > > devdax has to be 2M aligned ? Does Linux enforce that right now ? > > Unfortunately, there is no way around this. Part of the memory can be > reserved as persistent memory via device tree. > memory@40000000 { > device_type = "memory"; > reg = < 0x00000000 0x40000000 > 0x00000002 0x00000000 >; > }; > > pmem@1c0000000 { > compatible = "pmem-region"; > reg = <0x00000001 0xc0000000 > 0x00000000 0x80000000>; > volatile; > numa-node-id = <0>; > }; > > So, while pmem is section aligned, as it should be, the dax device is > going to be pmem start address + label size, which is 2M. The actual > DAX device starts at: > 0x1c0000000 + 2M. > > Because section size is 1G, the hotplug will able to add only memory > starting from > 0x1c0000000 + 1G > > > 27 and 28 do not even compile for ARM64_64_PAGES because of MAX_ORDER and > > SECTION_SIZE mismatch. > > Can you please elaborate what configs are you using? I have no > problems compiling with 27 and 28 bit. > > Thank you, > Pasha
On Wed, Apr 24, 2019 at 12:54 PM Pavel Tatashin <pasha.tatashin@soleen.com> wrote: > > <resending> from original email > > On Wed, Apr 24, 2019 at 3:48 PM Pavel Tatashin > <patatash@linux.microsoft.com> wrote: > > > > On Wed, Apr 24, 2019 at 5:07 AM Anshuman Khandual > > <anshuman.khandual@arm.com> wrote: > > > > > > On 04/24/2019 02:08 AM, Pavel Tatashin wrote: > > > > sparsemem section size determines the maximum size and alignment that > > > > is allowed to offline/online memory block. The bigger the size the less > > > > the clutter in /sys/devices/system/memory/*. On the other hand, however, > > > > there is less flexability in what granules of memory can be added and > > > > removed. > > > > > > Is there any scenario where less than a 1GB needs to be added on arm64 ? > > > > Yes, DAX hotplug loses 1G of memory without allowing smaller sections. > > Machines on which we are going to be using this functionality have 8G > > of System RAM, therefore losing 1G is a big problem. > > > > For details about using scenario see this cover letter: > > https://lore.kernel.org/lkml/20190421014429.31206-1-pasha.tatashin@soleen.com/ > > > > > > > > > > > > > Recently, it was enabled in Linux to hotadd persistent memory that > > > > can be either real NV device, or reserved from regular System RAM > > > > and has identity of devdax. > > > > > > devdax (even ZONE_DEVICE) support has not been enabled on arm64 yet. > > > > Correct, I use your patches to enable ZONE_DEVICE, and thus devdax on ARM64: > > https://lore.kernel.org/lkml/1554265806-11501-1-git-send-email-anshuman.khandual@arm.com/ > > > > > > > > > > > > > The problem is that because ARM64's section size is 1G, and devdax must > > > > have 2M label section, the first 1G is always missed when device is > > > > attached, because it is not 1G aligned. > > > > > > devdax has to be 2M aligned ? Does Linux enforce that right now ? > > > > Unfortunately, there is no way around this. Part of the memory can be > > reserved as persistent memory via device tree. > > memory@40000000 { > > device_type = "memory"; > > reg = < 0x00000000 0x40000000 > > 0x00000002 0x00000000 >; > > }; > > > > pmem@1c0000000 { > > compatible = "pmem-region"; > > reg = <0x00000001 0xc0000000 > > 0x00000000 0x80000000>; > > volatile; > > numa-node-id = <0>; > > }; > > > > So, while pmem is section aligned, as it should be, the dax device is > > going to be pmem start address + label size, which is 2M. The actual > > DAX device starts at: > > 0x1c0000000 + 2M. > > > > Because section size is 1G, the hotplug will able to add only memory > > starting from > > 0x1c0000000 + 1G This is yet another example of where we need to break down the section alignment requirement for arch_add_memory(). https://lore.kernel.org/lkml/155552633539.2015392.2477781120122237934.stgit@dwillia2-desk3.amr.corp.intel.com/
> This is yet another example of where we need to break down the section > alignment requirement for arch_add_memory(). > > https://lore.kernel.org/lkml/155552633539.2015392.2477781120122237934.stgit@dwillia2-desk3.amr.corp.intel.com/ Hi Dan, Yes, that is exactly what I am trying to solve with this patch. I will test if your series works with ARM64, and if does not I will let you know what is broken. But, I think, this patch is not needed if your patches are accepted into mainline. Thank you, Pasha
On 04/25/2019 01:18 AM, Pavel Tatashin wrote: > On Wed, Apr 24, 2019 at 5:07 AM Anshuman Khandual > <anshuman.khandual@arm.com> wrote: >> >> On 04/24/2019 02:08 AM, Pavel Tatashin wrote: >>> sparsemem section size determines the maximum size and alignment that >>> is allowed to offline/online memory block. The bigger the size the less >>> the clutter in /sys/devices/system/memory/*. On the other hand, however, >>> there is less flexability in what granules of memory can be added and >>> removed. >> >> Is there any scenario where less than a 1GB needs to be added on arm64 ? > > Yes, DAX hotplug loses 1G of memory without allowing smaller sections. > Machines on which we are going to be using this functionality have 8G > of System RAM, therefore losing 1G is a big problem. > > For details about using scenario see this cover letter: > https://lore.kernel.org/lkml/20190421014429.31206-1-pasha.tatashin@soleen.com/ Its loosing 1GB because devdax has 2M alignment ? IIRC from Dan's subsection memory hot add series 2M comes from persistent memory HW controller's limitations. Does that limitation applicable across all platforms including arm64 for all possible persistent memory vendors. I mean is it universal ? IIUC subsection memory hot plug series is still getting reviewed. Hence should not we wait for it to get merged before enabling applicable platforms to accommodate these 2M limitations. > >> >>> >>> Recently, it was enabled in Linux to hotadd persistent memory that >>> can be either real NV device, or reserved from regular System RAM >>> and has identity of devdax. >> >> devdax (even ZONE_DEVICE) support has not been enabled on arm64 yet. > > Correct, I use your patches to enable ZONE_DEVICE, and thus devdax on ARM64: > https://lore.kernel.org/lkml/1554265806-11501-1-git-send-email-anshuman.khandual@arm.com/ > >> >>> >>> The problem is that because ARM64's section size is 1G, and devdax must >>> have 2M label section, the first 1G is always missed when device is >>> attached, because it is not 1G aligned. >> >> devdax has to be 2M aligned ? Does Linux enforce that right now ? > > Unfortunately, there is no way around this. Part of the memory can be > reserved as persistent memory via device tree. > memory@40000000 { > device_type = "memory"; > reg = < 0x00000000 0x40000000 > 0x00000002 0x00000000 >; > }; > > pmem@1c0000000 { > compatible = "pmem-region"; > reg = <0x00000001 0xc0000000 > 0x00000000 0x80000000>; > volatile; > numa-node-id = <0>; > }; > > So, while pmem is section aligned, as it should be, the dax device is > going to be pmem start address + label size, which is 2M. The actual Forgive my ignorance here but why dax device label size is 2M aligned. Again is that because of some persistent memory HW controller limitations ? > DAX device starts at: > 0x1c0000000 + 2M. > > Because section size is 1G, the hotplug will able to add only memory > starting from > 0x1c0000000 + 1G Got it but as mentioned before we will have to make sure that 2M alignment requirement is universal else we will be adjusting this multiple times. > >> 27 and 28 do not even compile for ARM64_64_PAGES because of MAX_ORDER and >> SECTION_SIZE mismatch. Even with 27 bits its 128 MB section size. How does it solve the problem with 2M ? The patch just wanted to reduce the memory wastage ? > > Can you please elaborate what configs are you using? I have no > problems compiling with 27 and 28 bit. After applying your patch [1] on current mainline kernel [2]. $make defconfig CONFIG_ARM64_64K_PAGES=y CONFIG_ARM64_VA_BITS_48=y CONFIG_ARM64_VA_BITS=48 CONFIG_ARM64_PA_BITS_48=y CONFIG_ARM64_PA_BITS=48 CONFIG_ARM64_SECTION_SIZE_BITS=27 [1] https://patchwork.kernel.org/patch/10913737/ [2] git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git It fails with CC arch/arm64/kernel/asm-offsets.s In file included from ./include/linux/gfp.h:6, from ./include/linux/slab.h:15, from ./include/linux/resource_ext.h:19, from ./include/linux/acpi.h:26, from ./include/acpi/apei.h:9, from ./include/acpi/ghes.h:5, from ./include/linux/arm_sdei.h:14, from arch/arm64/kernel/asm-offsets.c:21: ./include/linux/mmzone.h:1095:2: error: #error Allocator MAX_ORDER exceeds SECTION_SIZE #error Allocator MAX_ORDER exceeds SECTION_SIZE
On Tue 23-04-19 16:38:43, Pavel Tatashin wrote: > sparsemem section size determines the maximum size and alignment that > is allowed to offline/online memory block. The bigger the size the less > the clutter in /sys/devices/system/memory/*. On the other hand, however, > there is less flexability in what granules of memory can be added and > removed. > > Recently, it was enabled in Linux to hotadd persistent memory that > can be either real NV device, or reserved from regular System RAM > and has identity of devdax. > > The problem is that because ARM64's section size is 1G, and devdax must > have 2M label section, the first 1G is always missed when device is > attached, because it is not 1G aligned. > > Allow, better flexibility by making section size configurable. Is there any inherent reason (64k page size?) that enforces such a large memsection?
On Thu, Apr 25, 2019 at 05:25:50PM +0200, Michal Hocko wrote: > On Tue 23-04-19 16:38:43, Pavel Tatashin wrote: > > sparsemem section size determines the maximum size and alignment that > > is allowed to offline/online memory block. The bigger the size the less > > the clutter in /sys/devices/system/memory/*. On the other hand, however, > > there is less flexability in what granules of memory can be added and > > removed. > > > > Recently, it was enabled in Linux to hotadd persistent memory that > > can be either real NV device, or reserved from regular System RAM > > and has identity of devdax. > > > > The problem is that because ARM64's section size is 1G, and devdax must > > have 2M label section, the first 1G is always missed when device is > > attached, because it is not 1G aligned. > > > > Allow, better flexibility by making section size configurable. > > Is there any inherent reason (64k page size?) that enforces such a large > memsection? I gave *vague* memories of running out of bits in the page flags if we changed this, but that was a while back. If that's no longer the case, then I'm open to changing the value, but I really don't want to expose it as a Kconfig option as proposed in this patch. People won't have a clue what to set and it doesn't help at all with the single-Image effort. Will
On Thu 25-04-19 16:31:38, Will Deacon wrote: > On Thu, Apr 25, 2019 at 05:25:50PM +0200, Michal Hocko wrote: > > On Tue 23-04-19 16:38:43, Pavel Tatashin wrote: > > > sparsemem section size determines the maximum size and alignment that > > > is allowed to offline/online memory block. The bigger the size the less > > > the clutter in /sys/devices/system/memory/*. On the other hand, however, > > > there is less flexability in what granules of memory can be added and > > > removed. > > > > > > Recently, it was enabled in Linux to hotadd persistent memory that > > > can be either real NV device, or reserved from regular System RAM > > > and has identity of devdax. > > > > > > The problem is that because ARM64's section size is 1G, and devdax must > > > have 2M label section, the first 1G is always missed when device is > > > attached, because it is not 1G aligned. > > > > > > Allow, better flexibility by making section size configurable. > > > > Is there any inherent reason (64k page size?) that enforces such a large > > memsection? > > I gave *vague* memories of running out of bits in the page flags if we > changed this, but that was a while back. If that's no longer the case, > then I'm open to changing the value, but I really don't want to expose > it as a Kconfig option as proposed in this patch. People won't have a > clue what to set and it doesn't help at all with the single-Image effort. Ohh, I absolutely agree about the config option part JFTR. 1GB section loos quite excessive. I am not really sure a standard arm64 memory layout looks though.
> > I gave *vague* memories of running out of bits in the page flags if we > > changed this, but that was a while back. If that's no longer the case, > > then I'm open to changing the value, but I really don't want to expose > > it as a Kconfig option as proposed in this patch. People won't have a > > clue what to set and it doesn't help at all with the single-Image effort. > > Ohh, I absolutely agree about the config option part JFTR. 1GB section > loos quite excessive. I am not really sure a standard arm64 memory > layout looks though. I am now looking to use Dan's patches "mm: Sub-section memory hotplug support" to solve this problem. I think this patch can be ignored. Pasha
On Thu 25-04-19 13:57:25, Pavel Tatashin wrote: > > > I gave *vague* memories of running out of bits in the page flags if we > > > changed this, but that was a while back. If that's no longer the case, > > > then I'm open to changing the value, but I really don't want to expose > > > it as a Kconfig option as proposed in this patch. People won't have a > > > clue what to set and it doesn't help at all with the single-Image effort. > > > > Ohh, I absolutely agree about the config option part JFTR. 1GB section > > loos quite excessive. I am not really sure a standard arm64 memory > > layout looks though. > > I am now looking to use Dan's patches "mm: Sub-section memory hotplug > support" to solve this problem. I think this patch can be ignored. Even if the subsection memory hotplug is going to be used then the underlying question remains. If there is no real reason to use large memsections then it would be better to use smaller ones.
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index b5d8cf57e220..a0c5b9d13a7f 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -801,6 +801,16 @@ config ARM64_PA_BITS default 48 if ARM64_PA_BITS_48 default 52 if ARM64_PA_BITS_52 +config ARM64_SECTION_SIZE_BITS + int "sparsemem section size shift" + range 27 30 + default "30" + depends on SPARSEMEM + help + Specify section size in bits. Section size determines the hotplug + hotremove granularity. The current size can be determined from + /sys/devices/system/memory/block_size_bytes + config CPU_BIG_ENDIAN bool "Build big-endian kernel" help diff --git a/arch/arm64/include/asm/sparsemem.h b/arch/arm64/include/asm/sparsemem.h index b299929fe56c..810db34d7038 100644 --- a/arch/arm64/include/asm/sparsemem.h +++ b/arch/arm64/include/asm/sparsemem.h @@ -18,7 +18,7 @@ #ifdef CONFIG_SPARSEMEM #define MAX_PHYSMEM_BITS CONFIG_ARM64_PA_BITS -#define SECTION_SIZE_BITS 30 +#define SECTION_SIZE_BITS CONFIG_ARM64_SECTION_SIZE_BITS #endif #endif
sparsemem section size determines the maximum size and alignment that is allowed to offline/online memory block. The bigger the size the less the clutter in /sys/devices/system/memory/*. On the other hand, however, there is less flexability in what granules of memory can be added and removed. Recently, it was enabled in Linux to hotadd persistent memory that can be either real NV device, or reserved from regular System RAM and has identity of devdax. The problem is that because ARM64's section size is 1G, and devdax must have 2M label section, the first 1G is always missed when device is attached, because it is not 1G aligned. Allow, better flexibility by making section size configurable. Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com> --- arch/arm64/Kconfig | 10 ++++++++++ arch/arm64/include/asm/sparsemem.h | 2 +- 2 files changed, 11 insertions(+), 1 deletion(-)