NUMA emulation (aka fake NUMA) distributes the available memory to nodes
without using real topology information about the physical memory of the
machine.
Splitting the system memory into nodes replicates the memory management
structures for each node. Particularly each node has its own "mm locks"
and its own "kswapd" task.
For large systems, under certain conditions, this results in improved
system performance and/or latency based on reduced pressure on the mm
locks and the kswapd tasks.
NUMA emulation distributes CPUs to nodes while respecting the original
machine topology information. This is done by trying to avoid to separate
CPUs which reside on the same book or even on the same MC. Because the
current Linux scheduler code requires a stable cpu to node mapping, cores
are pinned to nodes when the first CPU thread is set online.
This patch is based on the initial implementation from Philipp Hachtmann.
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The SCLP interface to query, configure and deconfigure CPUs actually
operates on cores. For a machine without the multi-threading faciltiy
a CPU and a core are equivalent but starting with System z13 a core
can have multiple hardware threads, also referred to as logical CPUs.
To avoid confusion replace the word 'cpu' with 'core' in the SCLP
interface. Also replace MAX_CPU_ADDRESS with SCLP_MAX_CORES.
The core-id is an 8-bit field, the maximum thread id is in the range
0-31. The theoretical limit for the CPU address is therefore 8191.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Use the generic mechanism to declare a bitmap instead of unsigned long.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Let's also move the facilities into the sclp struct, so we can avoid
another separate external variable.
Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Let's unify basic access to sclp fields by storing the data in an external
struct in asm/sclp.h.
The values can now directly be accessed by other components, so there is
no need for most accessor functions and external variables anymore.
The mtid, mtid_max and facility part will be cleaned up separately.
Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Commit 27356f54c8 ("mm/hotplug: verify hotplug memory range")
introduced a check that makes add_memory() only accept section size
aligned memory.
Therefore on z/VM systems, where standby memory is not aligned, no
standby memory is registered at all.
Example:
#cp def store 3504M standby 2336M
00: CP Q V STORE
00: STORAGE = 3504M MAX = 6G INC = 8M STANDBY = 2336M RESERVED = 0
For this setup the following error message is printed:
Section-unaligned hotplug range: start 0xdb000000, size 0x92000000
So fix this and register aligned memory in "sclp_cmd.c". This means
that for the corner cases where the standby memory is not aligned we
loose some memory.
In order to inform the user about the potential loss of standby memory,
we add a new message for each added standby block and print how
much of the standby memory is usable, for example:
sclp_cmd.4336b4: Standby memory at 0x50000000 (256M of 256M usable)
sclp_cmd.4336b4: Standby memory at 0xb0000000 (256M of 256M usable)
sclp_cmd.4336b4: Standby memory at 0xdb000000 (2048M of 2336M usable)
We also ensure that a potential memory block that contains both "assigned"
and "standby" memory cannot be setup offline.
Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
PTR_RET is deprecated. Do not recommend its usage anymore.
Use PTR_ERR_OR_ZERO instead.
Signed-off-by: Duan Jiong <duanj.fnst@cn.fujitsu.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
This patch adds a timeout option for queued requests and introduces
sclp_sync_request_timeout() to use this timer. With this, blocking the
system too long, e.g. during an SE reboot, can be avoided in critical
situations like CPU and memory hotplug.
Since there is no way to cancel a running request, this timeout only
applies to queued requests that have not yet been started.
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
To provide access to the set-partition-resource-parameter interface
to user space add a new attribute to hypfs/debugfs:
* s390_hypsfs/diag_304
The data for the query-partition-resource-parameters command can
be access by a read on the attribute. All other diagnose 0x304
requests need to be submitted via ioctl with CAP_SYS_ADMIN rights.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The early sclp detect functions gather the available SCLP facility
information. The sclp_early_read_info_sccb_valid indicates whether the
early sclp request was valid. However, one external reference to it
checks for particular sclp facility bits and this should be sufficient.
Another occurance is in the sclp_get_ipl_info() function that is called
later. Because all information are available at the early stage, save
the ipl information when detecting the sclp facilities. Hence, no more
checks for sclp_early_read_info_sccb_valid are required.
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Reviewed-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The early SCLP driver code in sclp_cmd.c belongs to sclp_early.c
because it is independent from the 'normal' SCLP driver. So move
it to sclp_early.c
Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Currently we have hardcoded the HSA size to 32 MiB. With this patch the
HSA size is determined dynamically via SCLP in early.c.
Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
To detect a line mode console we need a message event type
for the receive mask and a command event type for the send mask.
Checking for a MSG event in the sclp send mask was wrong and
might result in line mode consoles not being detected.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Add a character misc device "sclp_ctl" that allows to run SCCBs
from user space using the SCLP_CTL_SCCB ioctl.
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Given that the memory block size is always larger or equal to
the memory increment size there is no need for a usecount anymore.
Each memory increment belongs to exactly one memory block.
So get rid of the usecount.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
In case a machine supports memory hotplug all active memory increments
present at IPL time have been initialized with a "usecount" of 1.
This is wrong if the memory increment size is larger than the memory
section size of the memory hotplug code. If that is the case the
usecount must be initialized with the number of memory sections that
fit into one memory increment.
Otherwise it is possible to put a memory increment into standby state
even if there are still active sections.
Afterwards addressing exceptions might happen which cause the kernel
to panic.
However even worse, if a memory increment was put into standby state
and afterwards into active state again, it's contents would have been
zeroed, leading to memory corruption.
This was only an issue for machines that support standby memory and
have at least 256GB memory.
This is broken since commit fdb1bb15 "[S390] sclp/memory hotplug: fix
initial usecount of increments".
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: stable@vger.kernel.org # 2.6.39+
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Standby memory that is located outside [0,OLDMEM_SIZE] is currently used
by the s390 memory detection. This leads to additional memory consumption
due to allocation of page structures. To fix this, we now do not
add standby memory if the kernel is started in kdump mode.
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The s390 architecture is unique in respect to dirty page detection,
it uses the change bit in the per-page storage key to track page
modifications. All other architectures track dirty bits by means
of page table entries. This property of s390 has caused numerous
problems in the past, e.g. see git commit ef5d437f71
"mm: fix XFS oops due to dirty pages without buffers on s390".
To avoid future issues in regard to per-page dirty bits convert
s390 to a fault based software dirty bit detection mechanism. All
user page table entries which are marked as clean will be hardware
read-only, even if the pte is supposed to be writable. A write by
the user process will trigger a protection fault which will cause
the user pte to be marked as dirty and the hardware read-only bit
is removed.
With this change the dirty bit in the storage key is irrelevant
for Linux as a host, but the storage key is still required for
KVM guests. The effect is that page_test_and_clear_dirty and the
related code can be removed. The referenced bit in the storage
key is still used by the page_test_and_clear_young primitive to
provide page age information.
For page cache pages of mappings with mapping_cap_account_dirty
there will not be any change in behavior as the dirty bit tracking
already uses read-only ptes to control the amount of dirty pages.
Only for swap cache pages and pages of mappings without
mapping_cap_account_dirty there can be additional protection faults.
To avoid an excessive number of additional faults the mk_pte
primitive checks for PageDirty if the pgprot value allows for writes
and pre-dirties the pte. That avoids all additional faults for
tmpfs and shmem pages until these pages are added to the swap cache.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Add SCLP PCI configure/deconfigure and implement a PCI hotplug
controller (s390_pci_hpc). The hotplug controller creates a slot
for every PCI function in stand-by or configured state. The PCI
functions are named after the PCI function ID (fid). By writing to
the power attribute in /sys/bus/pci/slots/<fid>/power the PCI function
is moved to stand-by or configured state. If moved to the configured
state the device is automatically scanned by the s390 PCI layer.
Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Move and rename init_storage_keys() to pageattr.c, so it can also be
used from the sclp memory hotplug code in order to initialize
storage keys.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
For processing under KVM it is required to detect
the actual SCLP console type in order to set it as
preferred console.
Signed-off-by: Heinz Graalfs <graalfs@linux.vnet.ibm.com>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Acked-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Pull KVM changes from Avi Kivity:
"Changes include additional instruction emulation, page-crossing MMIO,
faster dirty logging, preventing the watchdog from killing a stopped
guest, module autoload, a new MSI ABI, and some minor optimizations
and fixes. Outside x86 we have a small s390 and a very large ppc
update.
Regarding the new (for kvm) rebaseless workflow, some of the patches
that were merged before we switch trees had to be rebased, while
others are true pulls. In either case the signoffs should be correct
now."
Fix up trivial conflicts in Documentation/feature-removal-schedule.txt
arch/powerpc/kvm/book3s_segment.S and arch/x86/include/asm/kvm_para.h.
I suspect the kvm_para.h resolution ends up doing the "do I have cpuid"
check effectively twice (it was done differently in two different
commits), but better safe than sorry ;)
* 'next' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (125 commits)
KVM: make asm-generic/kvm_para.h have an ifdef __KERNEL__ block
KVM: s390: onereg for timer related registers
KVM: s390: epoch difference and TOD programmable field
KVM: s390: KVM_GET/SET_ONEREG for s390
KVM: s390: add capability indicating COW support
KVM: Fix mmu_reload() clash with nested vmx event injection
KVM: MMU: Don't use RCU for lockless shadow walking
KVM: VMX: Optimize %ds, %es reload
KVM: VMX: Fix %ds/%es clobber
KVM: x86 emulator: convert bsf/bsr instructions to emulate_2op_SrcV_nobyte()
KVM: VMX: unlike vmcs on fail path
KVM: PPC: Emulator: clean up SPR reads and writes
KVM: PPC: Emulator: clean up instruction parsing
kvm/powerpc: Add new ioctl to retreive server MMU infos
kvm/book3s: Make kernel emulated H_PUT_TCE available for "PR" KVM
KVM: PPC: bookehv: Fix r8/r13 storing in level exception handler
KVM: PPC: Book3S: Enable IRQs during exit handling
KVM: PPC: Fix PR KVM on POWER7 bare metal
KVM: PPC: Fix stbux emulation
KVM: PPC: bookehv: Use lwz/stw instead of PPC_LL/PPC_STL for 32-bit fields
...
Currently qemu/kvm on s390 uses a guest mapping that does not
allow the guest backing page table to be write-protected to
support older systems. On those older systems a host write
protection fault will be delivered to the guest.
Newer systems allow to write-protect the guest backing memory
and let the fault be delivered to the host, thus allowing COW.
Use a capability bit to tell qemu if that is possible.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Add missing storage key initialization when memory comes online.
Didn't hurt until now, since the kernel always uses access key 0.
However for debugging purposes we sometimes use a different key
and it might be good if everything is properly initialized...
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)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=G9mT
-----END PGP SIGNATURE-----
Merge tag 'split-asm_system_h-for-linus-20120328' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-asm_system
Pull "Disintegrate and delete asm/system.h" from David Howells:
"Here are a bunch of patches to disintegrate asm/system.h into a set of
separate bits to relieve the problem of circular inclusion
dependencies.
I've built all the working defconfigs from all the arches that I can
and made sure that they don't break.
The reason for these patches is that I recently encountered a circular
dependency problem that came about when I produced some patches to
optimise get_order() by rewriting it to use ilog2().
This uses bitops - and on the SH arch asm/bitops.h drags in
asm-generic/get_order.h by a circuituous route involving asm/system.h.
The main difficulty seems to be asm/system.h. It holds a number of
low level bits with no/few dependencies that are commonly used (eg.
memory barriers) and a number of bits with more dependencies that
aren't used in many places (eg. switch_to()).
These patches break asm/system.h up into the following core pieces:
(1) asm/barrier.h
Move memory barriers here. This already done for MIPS and Alpha.
(2) asm/switch_to.h
Move switch_to() and related stuff here.
(3) asm/exec.h
Move arch_align_stack() here. Other process execution related bits
could perhaps go here from asm/processor.h.
(4) asm/cmpxchg.h
Move xchg() and cmpxchg() here as they're full word atomic ops and
frequently used by atomic_xchg() and atomic_cmpxchg().
(5) asm/bug.h
Move die() and related bits.
(6) asm/auxvec.h
Move AT_VECTOR_SIZE_ARCH here.
Other arch headers are created as needed on a per-arch basis."
Fixed up some conflicts from other header file cleanups and moving code
around that has happened in the meantime, so David's testing is somewhat
weakened by that. We'll find out anything that got broken and fix it..
* tag 'split-asm_system_h-for-linus-20120328' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-asm_system: (38 commits)
Delete all instances of asm/system.h
Remove all #inclusions of asm/system.h
Add #includes needed to permit the removal of asm/system.h
Move all declarations of free_initmem() to linux/mm.h
Disintegrate asm/system.h for OpenRISC
Split arch_align_stack() out from asm-generic/system.h
Split the switch_to() wrapper out of asm-generic/system.h
Move the asm-generic/system.h xchg() implementation to asm-generic/cmpxchg.h
Create asm-generic/barrier.h
Make asm-generic/cmpxchg.h #include asm-generic/cmpxchg-local.h
Disintegrate asm/system.h for Xtensa
Disintegrate asm/system.h for Unicore32 [based on ver #3, changed by gxt]
Disintegrate asm/system.h for Tile
Disintegrate asm/system.h for Sparc
Disintegrate asm/system.h for SH
Disintegrate asm/system.h for Score
Disintegrate asm/system.h for S390
Disintegrate asm/system.h for PowerPC
Disintegrate asm/system.h for PA-RISC
Disintegrate asm/system.h for MN10300
...
Use for_each_clear_bit() to iterate over all the cleared bit in a
memory region.
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Split out addressing mode bits from PSW_BASE_BITS, rename PSW_BASE_BITS
to PSW_MASK_BASE, get rid of psw_user32_bits, remove unused function
enabled_wait(), introduce PSW_MASK_USER, and drop PSW_MASK_MERGE macros.
Change psw_kernel_bits / psw_user_bits to contain only the bits that
are always set in the respective mode.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Make sure that only assigned storage increments are unassigned when
attaching a storage element.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Fix initial usecount of attached and assigned storage increments so
they can be set offline.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Avoid 64 bit division to fix this compile error on 32 bit:
drivers/s390/char/sclp_cmd.c:711: undefined reference to `__udivdi3'
Also move the whole arch_get_memory_phys_device function to the
memory hotplug related functions.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Implement arch specific arch_get_memory_phys_device function and
initialize phys_device for each memory section. That way we finally can
tell which piece of memory belongs to which physical device.
This makes s390's /sys/devices/system/memory/memoryX/phys_device display
the correct thing?
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Standby memory detected with the sclp interface gets always registered
with add_memory calls without considering the limitationt that the
"mem=" kernel paramater implies.
So fix this and only register standby memory that is below the specified
limit.
This fixes zfcpdump since it uses "mem=32M". In case there is appr.
2GB standby memory present all of usable memory would be used for the
struct pages needed for standby memory.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Move memory detection code to own file and also simplify it.
Also add an interface which can be called at any time to get the
current memory layout. This interface is needed by our kernel
internal system dumper.
Cc: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Cc: Michael Holzheu <holzheu@de.ibm.com>
Cc: Frank Munzert <munzert@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
It caused only a lot of confusion. From now on cpu hotplug of up to
NR_CPUS will work by default. If somebody wants to limit that then
the possible_cpus parameter can be used.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Add a new interface so that cpus can be put into standby state and
configured state.
Only offline cpus can be put into standby state or configured state.
For that the new percpu sysfs attribute "configure" must be used.
To put a cpu in standby state a "0" must be written to the attribute.
In order to switch it into configured state a "1" must be written to
the attribute.
Only cpus in configured state can be brought online.
In addition this patch introduces a static mapping of physical to
logical cpus. As a result only the sysfs directories of present cpus
will be created. To scan for new cpus the new sysfs attribute "rescan"
must be used.
Writing to /sys/devices/system/cpu/rescan will trigger a rescan of
cpus and will create directories for new cpus.
On IPL only configured cpus will be used. And on reboot/shutdown all
cpus will remain in their current state (configured/standby).
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
