Common: Optional support for adding a small amount of polling on each HLT
instruction executed in the guest (or equivalent for other architectures).
This can improve latency up to 50% on some scenarios (e.g. O_DSYNC writes
or TCP_RR netperf tests). This also has to be enabled manually for now,
but the plan is to auto-tune this in the future.
ARM/ARM64: the highlights are support for GICv3 emulation and dirty page
tracking
s390: several optimizations and bugfixes. Also a first: a feature
exposed by KVM (UUID and long guest name in /proc/sysinfo) before
it is available in IBM's hypervisor! :)
MIPS: Bugfixes.
x86: Support for PML (page modification logging, a new feature in
Broadwell Xeons that speeds up dirty page tracking), nested virtualization
improvements (nested APICv---a nice optimization), usual round of emulation
fixes. There is also a new option to reduce latency of the TSC deadline
timer in the guest; this needs to be tuned manually.
Some commits are common between this pull and Catalin's; I see you
have already included his tree.
ARM has other conflicts where functions are added in the same place
by 3.19-rc and 3.20 patches. These are not large though, and entirely
within KVM.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
iQEcBAABAgAGBQJU28rkAAoJEL/70l94x66DXqQH/1TDOfJIjW7P2kb0Sw7Fy1wi
cEX1KO/VFxAqc8R0E/0Wb55CXyPjQJM6xBXuFr5cUDaIjQ8ULSktL4pEwXyyv/s5
DBDkN65mriry2w5VuEaRLVcuX9Wy+tqLQXWNkEySfyb4uhZChWWHvKEcgw5SqCyg
NlpeHurYESIoNyov3jWqvBjr4OmaQENyv7t2c6q5ErIgG02V+iCux5QGbphM2IC9
LFtPKxoqhfeB2xFxTOIt8HJiXrZNwflsTejIlCl/NSEiDVLLxxHCxK2tWK/tUXMn
JfLD9ytXBWtNMwInvtFm4fPmDouv2VDyR0xnK2db+/axsJZnbxqjGu1um4Dqbak=
=7gdx
-----END PGP SIGNATURE-----
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull KVM update from Paolo Bonzini:
"Fairly small update, but there are some interesting new features.
Common:
Optional support for adding a small amount of polling on each HLT
instruction executed in the guest (or equivalent for other
architectures). This can improve latency up to 50% on some
scenarios (e.g. O_DSYNC writes or TCP_RR netperf tests). This
also has to be enabled manually for now, but the plan is to
auto-tune this in the future.
ARM/ARM64:
The highlights are support for GICv3 emulation and dirty page
tracking
s390:
Several optimizations and bugfixes. Also a first: a feature
exposed by KVM (UUID and long guest name in /proc/sysinfo) before
it is available in IBM's hypervisor! :)
MIPS:
Bugfixes.
x86:
Support for PML (page modification logging, a new feature in
Broadwell Xeons that speeds up dirty page tracking), nested
virtualization improvements (nested APICv---a nice optimization),
usual round of emulation fixes.
There is also a new option to reduce latency of the TSC deadline
timer in the guest; this needs to be tuned manually.
Some commits are common between this pull and Catalin's; I see you
have already included his tree.
Powerpc:
Nothing yet.
The KVM/PPC changes will come in through the PPC maintainers,
because I haven't received them yet and I might end up being
offline for some part of next week"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (130 commits)
KVM: ia64: drop kvm.h from installed user headers
KVM: x86: fix build with !CONFIG_SMP
KVM: x86: emulate: correct page fault error code for NoWrite instructions
KVM: Disable compat ioctl for s390
KVM: s390: add cpu model support
KVM: s390: use facilities and cpu_id per KVM
KVM: s390/CPACF: Choose crypto control block format
s390/kernel: Update /proc/sysinfo file with Extended Name and UUID
KVM: s390: reenable LPP facility
KVM: s390: floating irqs: fix user triggerable endless loop
kvm: add halt_poll_ns module parameter
kvm: remove KVM_MMIO_SIZE
KVM: MIPS: Don't leak FPU/DSP to guest
KVM: MIPS: Disable HTW while in guest
KVM: nVMX: Enable nested posted interrupt processing
KVM: nVMX: Enable nested virtual interrupt delivery
KVM: nVMX: Enable nested apic register virtualization
KVM: nVMX: Make nested control MSRs per-cpu
KVM: nVMX: Enable nested virtualize x2apic mode
KVM: nVMX: Prepare for using hardware MSR bitmap
...
Now that all in-tree users of strnicmp have been converted to
strncasecmp, the wrapper can be removed.
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc: David Howells <dhowells@redhat.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>
If an attacker can cause a controlled kernel stack overflow, overwriting
the restart block is a very juicy exploit target. This is because the
restart_block is held in the same memory allocation as the kernel stack.
Moving the restart block to struct task_struct prevents this exploit by
making the restart_block harder to locate.
Note that there are other fields in thread_info that are also easy
targets, at least on some architectures.
It's also a decent simplification, since the restart code is more or less
identical on all architectures.
[james.hogan@imgtec.com: metag: align thread_info::supervisor_stack]
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: David Miller <davem@davemloft.net>
Acked-by: Richard Weinberger <richard@nod.at>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Haavard Skinnemoen <hskinnemoen@gmail.com>
Cc: Hans-Christian Egtvedt <egtvedt@samfundet.no>
Cc: Steven Miao <realmz6@gmail.com>
Cc: Mark Salter <msalter@redhat.com>
Cc: Aurelien Jacquiot <a-jacquiot@ti.com>
Cc: Mikael Starvik <starvik@axis.com>
Cc: Jesper Nilsson <jesper.nilsson@axis.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Richard Kuo <rkuo@codeaurora.org>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: "James E.J. Bottomley" <jejb@parisc-linux.org>
Cc: Helge Deller <deller@gmx.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Acked-by: Michael Ellerman <mpe@ellerman.id.au> (powerpc)
Tested-by: Michael Ellerman <mpe@ellerman.id.au> (powerpc)
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Chen Liqin <liqin.linux@gmail.com>
Cc: Lennox Wu <lennox.wu@gmail.com>
Cc: Chris Metcalf <cmetcalf@ezchip.com>
Cc: Guan Xuetao <gxt@mprc.pku.edu.cn>
Cc: Chris Zankel <chris@zankel.net>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge second set of updates from Andrew Morton:
"More of MM"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (83 commits)
mm/nommu.c: fix arithmetic overflow in __vm_enough_memory()
mm/mmap.c: fix arithmetic overflow in __vm_enough_memory()
vmstat: Reduce time interval to stat update on idle cpu
mm/page_owner.c: remove unnecessary stack_trace field
Documentation/filesystems/proc.txt: describe /proc/<pid>/map_files
mm: incorporate read-only pages into transparent huge pages
vmstat: do not use deferrable delayed work for vmstat_update
mm: more aggressive page stealing for UNMOVABLE allocations
mm: always steal split buddies in fallback allocations
mm: when stealing freepages, also take pages created by splitting buddy page
mincore: apply page table walker on do_mincore()
mm: /proc/pid/clear_refs: avoid split_huge_page()
mm: pagewalk: fix misbehavior of walk_page_range for vma(VM_PFNMAP)
mempolicy: apply page table walker on queue_pages_range()
arch/powerpc/mm/subpage-prot.c: use walk->vma and walk_page_vma()
memcg: cleanup preparation for page table walk
numa_maps: remove numa_maps->vma
numa_maps: fix typo in gather_hugetbl_stats
pagemap: use walk->vma instead of calling find_vma()
clear_refs: remove clear_refs_private->vma and introduce clear_refs_test_walk()
...
Pull s390 updates from Martin Schwidefsky:
- The remaining patches for the z13 machine support: kernel build
option for z13, the cache synonym avoidance, SMT support,
compare-and-delay for spinloops and the CES5S crypto adapater.
- The ftrace support for function tracing with the gcc hotpatch option.
This touches common code Makefiles, Steven is ok with the changes.
- The hypfs file system gets an extension to access diagnose 0x0c data
in user space for performance analysis for Linux running under z/VM.
- The iucv hvc console gets wildcard spport for the user id filtering.
- The cacheinfo code is converted to use the generic infrastructure.
- Cleanup and bug fixes.
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (42 commits)
s390/process: free vx save area when releasing tasks
s390/hypfs: Eliminate hypfs interval
s390/hypfs: Add diagnose 0c support
s390/cacheinfo: don't use smp_processor_id() in preemptible context
s390/zcrypt: fixed domain scanning problem (again)
s390/smp: increase maximum value of NR_CPUS to 512
s390/jump label: use different nop instruction
s390/jump label: add sanity checks
s390/mm: correct missing space when reporting user process faults
s390/dasd: cleanup profiling
s390/dasd: add locking for global_profile access
s390/ftrace: hotpatch support for function tracing
ftrace: let notrace function attribute disable hotpatching if necessary
ftrace: allow architectures to specify ftrace compile options
s390: reintroduce diag 44 calls for cpu_relax()
s390/zcrypt: Add support for new crypto express (CEX5S) adapter.
s390/zcrypt: Number of supported ap domains is not retrievable.
s390/spinlock: add compare-and-delay to lock wait loops
s390/tape: remove redundant if statement
s390/hvc_iucv: add simple wildcard matches to the iucv allow filter
...
This allows the get_user_pages_fast slow path to release the mmap_sem
before blocking.
Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Reviewed-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andres Lagar-Cavilla <andreslc@google.com>
Cc: Peter Feiner <pfeiner@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
LKP has triggered a compiler warning after my recent patch "mm: account
pmd page tables to the process":
mm/mmap.c: In function 'exit_mmap':
>> mm/mmap.c:2857:2: warning: right shift count >= width of type [enabled by default]
The code:
> 2857 WARN_ON(mm_nr_pmds(mm) >
2858 round_up(FIRST_USER_ADDRESS, PUD_SIZE) >> PUD_SHIFT);
In this, on tile, we have FIRST_USER_ADDRESS defined as 0. round_up() has
the same type -- int. PUD_SHIFT.
I think the best way to fix it is to define FIRST_USER_ADDRESS as unsigned
long. On every arch for consistency.
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reported-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently we have many duplicates in definitions around
follow_huge_addr(), follow_huge_pmd(), and follow_huge_pud(), so this
patch tries to remove the m. The basic idea is to put the default
implementation for these functions in mm/hugetlb.c as weak symbols
(regardless of CONFIG_ARCH_WANT_GENERAL_HUGETL B), and to implement
arch-specific code only when the arch needs it.
For follow_huge_addr(), only powerpc and ia64 have their own
implementation, and in all other architectures this function just returns
ERR_PTR(-EINVAL). So this patch sets returning ERR_PTR(-EINVAL) as
default.
As for follow_huge_(pmd|pud)(), if (pmd|pud)_huge() is implemented to
always return 0 in your architecture (like in ia64 or sparc,) it's never
called (the callsite is optimized away) no matter how implemented it is.
So in such architectures, we don't need arch-specific implementation.
In some architecture (like mips, s390 and tile,) their current
arch-specific follow_huge_(pmd|pud)() are effectively identical with the
common code, so this patch lets these architecture use the common code.
One exception is metag, where pmd_huge() could return non-zero but it
expects follow_huge_pmd() to always return NULL. This means that we need
arch-specific implementation which returns NULL. This behavior looks
strange to me (because non-zero pmd_huge() implies that the architecture
supports PMD-based hugepage, so follow_huge_pmd() can/should return some
relevant value,) but that's beyond this cleanup patch, so let's keep it.
Justification of non-trivial changes:
- in s390, follow_huge_pmd() checks !MACHINE_HAS_HPAGE at first, and this
patch removes the check. This is OK because we can assume MACHINE_HAS_HPAGE
is true when follow_huge_pmd() can be called (note that pmd_huge() has
the same check and always returns 0 for !MACHINE_HAS_HPAGE.)
- in s390 and mips, we use HPAGE_MASK instead of PMD_MASK as done in common
code. This patch forces these archs use PMD_MASK, but it's OK because
they are identical in both archs.
In s390, both of HPAGE_SHIFT and PMD_SHIFT are 20.
In mips, HPAGE_SHIFT is defined as (PAGE_SHIFT + PAGE_SHIFT - 3) and
PMD_SHIFT is define as (PAGE_SHIFT + PAGE_SHIFT + PTE_ORDER - 3), but
PTE_ORDER is always 0, so these are identical.
Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Acked-by: Hugh Dickins <hughd@google.com>
Cc: James Hogan <james.hogan@imgtec.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Luiz Capitulino <lcapitulino@redhat.com>
Cc: Nishanth Aravamudan <nacc@linux.vnet.ibm.com>
Cc: Lee Schermerhorn <lee.schermerhorn@hp.com>
Cc: Steve Capper <steve.capper@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
If a task uses vector registers, a save area is allocated to save/restore
register states. Free the save area when releasing the task.
Found the Memory leak with kmemleak:
unreferenced object 0x72885e00 (size 512):
comm "vx-test", pid 26123, jiffies 4294945635 (age 256.810s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 01 db 71 06 41 .............q.A
00 00 00 00 00 00 00 00 24 f7 a9 a7 51 94 79 bb ........$...Q.y.
backtrace:
[<00000000002d1c8a>] kmem_cache_alloc_trace+0x272/0x3d0
[<00000000001014ac>] alloc_vector_registers+0x54/0x138
[<00000000001017c8>] data_exception+0x158/0x1b0
[<00000000008b551e>] pgm_check_handler+0x13e/0x180
[<00000000800008b6>] 0x800008b6
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Currently the binary hypfs interfaces provides new data only once within
an interval time of one second. This patch removes this restriction and
now new data is returned immediately on every read on a hypfs binary file.
This is done in order to allow more consistent snapshots for programs
that read multiple hypfs binary files.
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
With this feature, you can read the CPU performance metrics provided by the
z/VM diagnose 0C. This then allows to get the management time for each
online CPU of the guest where the diagnose is executed.
The new debugfs file /sys/kernel/debug/s390_hypfs/diag_0c exports the
diag0C binary data to user space via an open/read/close interface.
The binary data consists out of a header structure followed by an
array that contains the diagnose 0c data for each online CPU.
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
show_cacheinfo() needs to access the cacheinfo structure of any online
cpu. This was done with using smp_processor_id() as in index while in
preemtible context.
This means the cpu could be offline and the data be gone when it would
be accessed.
Better use any online cpu address and protect the data by get_online_cpus()
and put_online_cpus().
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Pull RCU updates from Ingo Molnar:
"The main RCU changes in this cycle are:
- Documentation updates.
- Miscellaneous fixes.
- Preemptible-RCU fixes, including fixing an old bug in the
interaction of RCU priority boosting and CPU hotplug.
- SRCU updates.
- RCU CPU stall-warning updates.
- RCU torture-test updates"
* 'core-rcu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (54 commits)
rcu: Initialize tiny RCU stall-warning timeouts at boot
rcu: Fix RCU CPU stall detection in tiny implementation
rcu: Add GP-kthread-starvation checks to CPU stall warnings
rcu: Make cond_resched_rcu_qs() apply to normal RCU flavors
rcu: Optionally run grace-period kthreads at real-time priority
ksoftirqd: Use new cond_resched_rcu_qs() function
ksoftirqd: Enable IRQs and call cond_resched() before poking RCU
rcutorture: Add more diagnostics in rcu_barrier() test failure case
torture: Flag console.log file to prevent holdovers from earlier runs
torture: Add "-enable-kvm -soundhw pcspk" to qemu command line
rcutorture: Handle different mpstat versions
rcutorture: Check from beginning to end of grace period
rcu: Remove redundant rcu_batches_completed() declaration
rcutorture: Drop rcu_torture_completed() and friends
rcu: Provide rcu_batches_completed_sched() for TINY_RCU
rcutorture: Use unsigned for Reader Batch computations
rcutorture: Make build-output parsing correctly flag RCU's warnings
rcu: Make _batches_completed() functions return unsigned long
rcutorture: Issue warnings on close calls due to Reader Batch blows
documentation: Fix smp typo in memory-barriers.txt
...
This patch enables cpu model support in kvm/s390 via the vm attribute
interface.
During KVM initialization, the host properties cpuid, IBC value and the
facility list are stored in the architecture specific cpu model structure.
During vcpu setup, these properties are taken to initialize the related SIE
state. This mechanism allows to adjust the properties from user space and thus
to implement different selectable cpu models.
This patch uses the IBC functionality to block instructions that have not
been implemented at the requested CPU type and GA level compared to the
full host capability.
Userspace has to initialize the cpu model before vcpu creation. A cpu model
change of running vcpus is not possible.
Signed-off-by: Michael Mueller <mimu@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
The patch introduces facilities and cpu_ids per virtual machine.
Different virtual machines may want to expose different facilities and
cpu ids to the guest, so let's make them per-vm instead of global.
Signed-off-by: Michael Mueller <mimu@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
We need to specify a different format for the crypto control block
depending on whether the APXA facility is installed or not. Let's
test for it by executing the PQAP(QCI) function and use either a
format-1 or a format-2 crypto control block accordingly. This is a
host only change for z13 and does not affect the guest view.
Signed-off-by: Tony Krowiak <akrowiak@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
A new architecture extends STSI 3.2.2 with UUID and long names. KVM
will provide the first implementation. This patch adds the additional
data fields (Extended Name and UUID) from the 4KB block returned by
the STSI 3.2.2 command and reflect this information in the
/proc/sysinfo file accordingly.
Signed-off-by: Ekaterina Tumanova <tumanova@linux.vnet.ibm.com>
Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
commit 7be81a4669 ("KVM: s390/facilities: allow TOD-CLOCK steering
facility bit") accidentially disabled the "load program parameter"
facility bit during rebase for upstream submission (my fault).
Re-add that bit.
As this is only for a performance measurement helper instruction
(used by KVM itself) cc stable is not necessary see
http://www-01.ibm.com/support/docview.wss?uid=isg26fcd1cc32246f4c8852574ce0044734a
(SA23-2260 The Load-Program-Parameter and CPU-Measurement Facilities)
for details about LPP and its usecase.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Fixes: 7be81a4669 ("KVM: s390/facilities: allow TOD-CLOCK steering")
If a vm with no VCPUs is created, the injection of a floating irq
leads to an endless loop in the kernel.
Let's skip the search for a destination VCPU for a floating irq if no
VCPUs were created.
Reviewed-by: Dominik Dingel <dingel@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Cc: stable@vger.kernel.org # v3.15+
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
This patch introduces a new module parameter for the KVM module; when it
is present, KVM attempts a bit of polling on every HLT before scheduling
itself out via kvm_vcpu_block.
This parameter helps a lot for latency-bound workloads---in particular
I tested it with O_DSYNC writes with a battery-backed disk in the host.
In this case, writes are fast (because the data doesn't have to go all
the way to the platters) but they cannot be merged by either the host or
the guest. KVM's performance here is usually around 30% of bare metal,
or 50% if you use cache=directsync or cache=writethrough (these
parameters avoid that the guest sends pointless flush requests, and
at the same time they are not slow because of the battery-backed cache).
The bad performance happens because on every halt the host CPU decides
to halt itself too. When the interrupt comes, the vCPU thread is then
migrated to a new physical CPU, and in general the latency is horrible
because the vCPU thread has to be scheduled back in.
With this patch performance reaches 60-65% of bare metal and, more
important, 99% of what you get if you use idle=poll in the guest. This
means that the tunable gets rid of this particular bottleneck, and more
work can be done to improve performance in the kernel or QEMU.
Of course there is some price to pay; every time an otherwise idle vCPUs
is interrupted by an interrupt, it will poll unnecessarily and thus
impose a little load on the host. The above results were obtained with
a mostly random value of the parameter (500000), and the load was around
1.5-2.5% CPU usage on one of the host's core for each idle guest vCPU.
The patch also adds a new stat, /sys/kernel/debug/kvm/halt_successful_poll,
that can be used to tune the parameter. It counts how many HLT
instructions received an interrupt during the polling period; each
successful poll avoids that Linux schedules the VCPU thread out and back
in, and may also avoid a likely trip to C1 and back for the physical CPU.
While the VM is idle, a Linux 4 VCPU VM halts around 10 times per second.
Of these halts, almost all are failed polls. During the benchmark,
instead, basically all halts end within the polling period, except a more
or less constant stream of 50 per second coming from vCPUs that are not
running the benchmark. The wasted time is thus very low. Things may
be slightly different for Windows VMs, which have a ~10 ms timer tick.
The effect is also visible on Marcelo's recently-introduced latency
test for the TSC deadline timer. Though of course a non-RT kernel has
awful latency bounds, the latency of the timer is around 8000-10000 clock
cycles compared to 20000-120000 without setting halt_poll_ns. For the TSC
deadline timer, thus, the effect is both a smaller average latency and
a smaller variance.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Kim Phillips reported following build failure.
LD init/built-in.o
mm/built-in.o: In function `free_pages_prepare':
mm/page_alloc.c:770: undefined reference to `.kernel_map_pages'
mm/built-in.o: In function `prep_new_page':
mm/page_alloc.c:933: undefined reference to `.kernel_map_pages'
mm/built-in.o: In function `map_pages':
mm/compaction.c:61: undefined reference to `.kernel_map_pages'
make: *** [vmlinux] Error 1
Reason for this problem is that commit 031bc5743f
("mm/debug-pagealloc: make debug-pagealloc boottime configurable")
forgot to remove the old declaration of kernel_map_pages() for some
architectures. This patch removes them to fix build failure.
Reported-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: David Miller <davem@davemloft.net>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
With SMT we can have more than 256 CPUs. Let's make them available.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The core VM already knows about VM_FAULT_SIGBUS, but cannot return a
"you should SIGSEGV" error, because the SIGSEGV case was generally
handled by the caller - usually the architecture fault handler.
That results in lots of duplication - all the architecture fault
handlers end up doing very similar "look up vma, check permissions, do
retries etc" - but it generally works. However, there are cases where
the VM actually wants to SIGSEGV, and applications _expect_ SIGSEGV.
In particular, when accessing the stack guard page, libsigsegv expects a
SIGSEGV. And it usually got one, because the stack growth is handled by
that duplicated architecture fault handler.
However, when the generic VM layer started propagating the error return
from the stack expansion in commit fee7e49d45 ("mm: propagate error
from stack expansion even for guard page"), that now exposed the
existing VM_FAULT_SIGBUS result to user space. And user space really
expected SIGSEGV, not SIGBUS.
To fix that case, we need to add a VM_FAULT_SIGSEGV, and teach all those
duplicate architecture fault handlers about it. They all already have
the code to handle SIGSEGV, so it's about just tying that new return
value to the existing code, but it's all a bit annoying.
This is the mindless minimal patch to do this. A more extensive patch
would be to try to gather up the mostly shared fault handling logic into
one generic helper routine, and long-term we really should do that
cleanup.
Just from this patch, you can generally see that most architectures just
copied (directly or indirectly) the old x86 way of doing things, but in
the meantime that original x86 model has been improved to hold the VM
semaphore for shorter times etc and to handle VM_FAULT_RETRY and other
"newer" things, so it would be a good idea to bring all those
improvements to the generic case and teach other architectures about
them too.
Reported-and-tested-by: Takashi Iwai <tiwai@suse.de>
Tested-by: Jan Engelhardt <jengelh@inai.de>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> # "s390 still compiles and boots"
Cc: linux-arch@vger.kernel.org
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Use a brcl 0,2 instruction for jump label nops during compile time,
so we don't mix up the different nops during mcount/hotpatch call
site detection.
The initial jump label code instruction replacement will exchange
these instructions with either a branch or a brcl 0,0 instruction.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Add sanity checks to verify that only expected code will be replaced.
If the code patterns do not match print the code patterns and panic,
since something went terribly wrong.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Make use of gcc's hotpatch support to generate better code for ftrace
function tracing.
The generated code now contains only a six byte nop in each function
prologue instead of a 24 byte code block which will be runtime patched to
support function tracing.
With the new code generation the runtime overhead for supporting function
tracing is close to zero, while the original code did show a significant
performance impact.
Acked-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Christian Borntraeger reported that the now missing diag 44 calls (voluntary
time slice end) does cause a performance regression for stop_machine() calls
if a machine has more virtual cpus than the host has physical cpus.
This patch mainly reverts 57f2ffe14f ("s390: remove diag 44 calls from
cpu_relax()") with the exception that we still do not issue diag 44 calls if
running with smt enabled. Due to group scheduling algorithms when running in
LPAR this would lead to significant latencies.
However, when running in LPAR we do not have more virtual than physical cpus.
Reported-and-tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Add the compare-and-delay instruction to the spin-lock and rw-lock
retry loops. A CPU executing the compare-and-delay instruction stops
until the lock value has changed. This is done to make the locking
code for contended locks to behave better in regard to the multi-
hreading facility. A thread of a core executing a compare-and-delay
will allow the other threads of a core to get a larger share of the
core resources.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
When we convert interrupt data from struct kvm_s390_interrupt to
struct kvm_s390_irq we need to check the data in the input parameter
not the output parameter.
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
We have to delete the allocated interrupt info if __inject_vm() fails.
Otherwise user space can keep flooding kvm with floating interrupts and
provoke more and more memory leaks.
Reported-by: Dominik Dingel <dingel@linux.vnet.ibm.com>
Reviewed-by: Dominik Dingel <dingel@linux.vnet.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Cc: stable@vger.kernel.org # v3.15+
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Created new KVM device attributes for indicating whether the AES and
DES/TDES protected key functions are available for programs running
on the KVM guest. The attributes are used to set up the controls in
the guest SIE block that specify whether programs running on the
guest will be given access to the protected key functions available
on the s390 hardware.
Signed-off-by: Tony Krowiak <akrowiak@linux.vnet.ibm.com>
Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Michael Mueller <mimu@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
[split MSA4/protected key into two patches]
Provide controls for setting/getting the guest TOD clock based on the VM
attribute interface.
Provide TOD and TOD_HIGH vm attributes on s390 for managing guest Time Of
Day clock value.
TOD_HIGH is presently always set to 0. In the future it will contain a high
order expansion of the tod clock value after it overflows the 64-bits of
the TOD.
Signed-off-by: Jason J. Herne <jjherne@linux.vnet.ibm.com>
Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
When injecting SIGP set prefix or a machine check, we trace
the values in our per-vcpu local_int data structure instead
of the parameters passed to the function.
Fix this by changing the trace statement to use the correct values.
Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Currently we are always setting the wrong bit in the
bitmap for pending emergency signals. Instead of using
emerg.code from the passed in irq parameter, we use the
value in our per-vcpu local_int structure, which is always zero.
That means all emergency signals will have address 0 as parameter.
If two CPUs send a SIGP to the same target, one might be lost.
Let's fix this by using the value from the parameter and
also trace the correct value.
Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
The handler for MVPG partial execution interception does not take
the current CPU addressing mode into account yet, so addresses are
always treated as 64-bit addresses. For correct behaviour, we should
properly handle 24-bit and 31-bit addresses, too.
Since MVPG is defined to work with logical addresses, we can simply
use guest_translate_address() to achieve the required behaviour
(since DAT is disabled here, guest_translate_address() skips the MMU
translation and only translates the address via kvm_s390_logical_to_effective()
and kvm_s390_real_to_abs(), which is exactly what we want here).
Signed-off-by: Thomas Huth <thuth@linux.vnet.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
The kvm mutex was (probably) used to protect against cpu hotplug.
The current code no longer needs to protect against that, as we only
rely on CPU data structures that are guaranteed to be available
if we can access the CPU. (e.g. vcpu_create will put the cpu
in the array AFTER the cpu is ready).
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Jens Freimann <jfrei@linux.vnet.ibm.com>
Most SIGP orders are handled partially in kernel and partially in
user space. In order to:
- Get a correct SIGP SET PREFIX handler that informs user space
- Avoid race conditions between concurrently executed SIGP orders
- Serialize SIGP orders per VCPU
We need to handle all "slow" SIGP orders in user space. The remaining
ones to be handled completely in kernel are:
- SENSE
- SENSE RUNNING
- EXTERNAL CALL
- EMERGENCY SIGNAL
- CONDITIONAL EMERGENCY SIGNAL
According to the PoP, they have to be fast. They can be executed
without conflicting to the actions of other pending/concurrently
executing orders (e.g. STOP vs. START).
This patch introduces a new capability that will - when enabled -
forward all but the mentioned SIGP orders to user space. The
instruction counters in the kernel are still updated.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
We need a way to clear the async pfault queue from user space (e.g.
for resets and SIGP SET ARCHITECTURE).
This patch simply clears the queue as soon as user space sets the
invalid pfault token. The definition of the invalid token is moved
to uapi.
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Only one external call may be pending at a vcpu at a time. For this
reason, we have to detect whether the SIGP externcal call interpretation
facility is available. If so, all external calls have to be injected
using this mechanism.
SIGP EXTERNAL CALL orders have to return whether another external
call is already pending. This check was missing until now.
SIGP SENSE hasn't returned yet in all conditions whether an external
call was pending.
If a SIGP EXTERNAL CALL irq is to be injected and one is already
pending, -EBUSY is returned.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
This patch introduces the infrastructure to check whether the SIGP
Interpretation Facility is installed on all VCPUs in the configuration.
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
This patch cleanes up the the SIGP SET PREFIX code.
A SIGP SET PREFIX irq may only be injected if the target vcpu is
stopped. Let's move the checking code into the injection code and
return -EBUSY if the target vcpu is not stopped.
Reviewed-by: Jens Freimann <jfrei@linux.vnet.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
As a SIGP STOP is an interrupt with the least priority, it may only result
in stop of the vcpu when no other interrupts are left pending.
To detect whether a non-stop irq is pending, we need a way to mask out
stop irqs from the general kvm_cpu_has_interrupt() function. For this
reason, the existing function (with an outdated name) is replaced by
kvm_s390_vcpu_has_irq() which allows to mask out pending stop irqs.
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
This patch removes the famous action_bits and moves the handling of
SIGP STOP AND STORE STATUS directly into the SIGP STOP interrupt.
The new local interrupt infrastructure is used to track pending stop
requests.
STOP irqs are the only irqs that don't get actively delivered. They
remain pending until the stop function is executed (=stop intercept).
If another STOP irq is already pending, -EBUSY will now be returned
(needed for the SIGP handling code).
Migration of pending SIGP STOP (AND STORE STATUS) orders should now
be supported out of the box.
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
In order to get rid of the action_flags and to properly migrate pending SIGP
STOP irqs triggered e.g. by SIGP STOP AND STORE STATUS, we need to remember
whether to store the status when stopping.
For this reason, a new parameter (flags) for the SIGP STOP irq is introduced.
These flags further define details of the requested STOP and can be easily
migrated.
Reviewed-by: Thomas Huth <thuth@linux.vnet.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Patch 0759d0681c ("KVM: s390: cleanup handle_wait by reusing
kvm_vcpu_block") changed the way pending guest clock comparator
interrupts are detected. It was assumed that as soon as the hrtimer
wakes up, the condition for the guest ckc is satisfied.
This is however only true as long as adjclock() doesn't speed
up the monotonic clock. Reason is that the hrtimer is based on
CLOCK_MONOTONIC, the guest clock comparator detection is based
on the raw TOD clock. If CLOCK_MONOTONIC runs faster than the
TOD clock, the hrtimer wakes the target VCPU up too early and
the target VCPU will not detect any pending interrupts, therefore
going back to sleep. It will never be woken up again because the
hrtimer has finished. The VCPU is stuck.
As a quick fix, we have to forward the hrtimer until the guest
clock comparator is really due, to guarantee properly timed wake
ups.
As the hrtimer callback might be triggered on another cpu, we
have to make sure that the timer is really stopped and not currently
executing the callback on another cpu. This can happen if the vcpu
thread is scheduled onto another physical cpu, but the timer base
is not migrated. So lets use hrtimer_cancel instead of try_to_cancel.
A proper fix might be to introduce a RAW based hrtimer.
Reported-by: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: stable@vger.kernel.org
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
The hrtimer that handles the wait with enabled timer interrupts
should not be disturbed by changes of the host time.
This patch changes our hrtimer to be based on a monotonic clock.
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Cc: stable@vger.kernel.org
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
We sometimes get an underflow for the sleep duration, which most
likely won't result in the short sleep time we wanted.
So let's check for sleep duration underflows and directly continue
to run the guest if we get one.
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Linus Torvalds
Rasmus Villemoes
Andy Lutomirski
Andrea Arcangeli
Kirill A. Shutemov
Naoya Horiguchi
Hendrik Brueckner
Michael Holzheu
Heiko Carstens
Michael Mueller
Tony Krowiak
Ekaterina Tumanova
Christian Borntraeger
David Hildenbrand
Paolo Bonzini
Joonsoo Kim
Martin Schwidefsky
Jens Freimann
Jason J. Herne
Thomas Huth