Wanpeng Li [Thu, 3 Sep 2015 14:07:38 +0000 (22:07 +0800)]
KVM: dynamic halt-polling
There is a downside of always-poll since poll is still happened for idle
vCPUs which can waste cpu usage. This patchset add the ability to adjust
halt_poll_ns dynamically, to grow halt_poll_ns when shot halt is detected,
and to shrink halt_poll_ns when long halt is detected.
There are two new kernel parameters for changing the halt_poll_ns:
halt_poll_ns_grow and halt_poll_ns_shrink.
"Idle (X) vCPU %c0" is the percent of time the physical cpu spent in
c0 over 60 seconds (each vCPU is pinned to a pCPU). (nohz) means the
guest was tickless. (250HZ) means the guest was ticking at 250HZ.
The big win is with ticking operating systems. Running the linux guest
with nohz=off (and HZ=250), we save 3.4%~12.8% CPUs/second and get close
to no-polling overhead levels by using the dynamic-poll. The savings
should be even higher for higher frequency ticks.
Suggested-by: David Matlack <dmatlack@google.com> Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
[Simplify the patch. - Paolo] Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Valdis Kletnieks [Sat, 29 Aug 2015 21:49:16 +0000 (17:49 -0400)]
Silence compiler warning in arch/x86/kvm/emulate.c
Compiler warning:
CC [M] arch/x86/kvm/emulate.o
arch/x86/kvm/emulate.c: In function "__do_insn_fetch_bytes":
arch/x86/kvm/emulate.c:814:9: warning: "linear" may be used uninitialized in this function [-Wmaybe-uninitialized]
GCC is smart enough to realize that the inlined __linearize may return before
setting the value of linear, but not smart enough to realize the same
X86EMU_CONTINUE blocks actual use of the value. However, the value of
'linear' can only be set to one value, so hoisting the one line of code
upwards makes GCC happy with the code.
kvm: compile process_smi_save_seg_64() only for x86_64
The process_smi_save_seg_64() function called only in the
process_smi_save_state_64() if the CONFIG_X86_64 is set. This
patch adds #ifdef CONFIG_X86_64 around process_smi_save_seg_64()
to prevent following warning message:
arch/x86/kvm/x86.c:5946:13: warning: ‘process_smi_save_seg_64’ defined but not used [-Wunused-function]
static void process_smi_save_seg_64(struct kvm_vcpu *vcpu, char *buf, int n)
^
Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Sun, 6 Sep 2015 14:24:50 +0000 (16:24 +0200)]
KVM: x86: avoid uninitialized variable warning
This does not show up on all compiler versions, so it sneaked into the
first 4.3 pull request. The fix is to mimic the logic of the "print
sptes" loop in the "fill array" loop. Then leaf and root can be
both initialized unconditionally.
Note that "leaf" now points to the first unused element of the array,
not the last filled element.
Reported-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Thomas Huth [Wed, 2 Sep 2015 09:14:48 +0000 (11:14 +0200)]
KVM: PPC: Book3S: Fix size of the PSPB register
The size of the Problem State Priority Boost Register is only
32 bits, but the kvm_vcpu_arch->pspb variable is declared as
"ulong", ie. 64-bit. However, the assembler code accesses this
variable with 32-bit accesses, and the KVM_REG_PPC_PSPB macro
is defined with SIZE_U32, too, so that the current code is
broken on big endian hosts: kvmppc_get_one_reg_hv() will only
return zero for this register since it is using the wrong half
of the pspb variable. Let's fix this problem by adjusting the
size of the pspb field in the kvm_vcpu_arch structure.
Signed-off-by: Thomas Huth <thuth@redhat.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
KVM: PPC: Book3S HV: Exit on H_DOORBELL if HOST_IPI is set
The code that handles the case when we receive a H_DOORBELL interrupt
has a comment which says "Hypervisor doorbell - exit only if host IPI
flag set". However, the current code does not actually check if the
host IPI flag is set. This is due to a comparison instruction that
got missed.
As a result, the current code performs the exit to host only
if some sibling thread or a sibling sub-core is exiting to the
host. This implies that, an IPI sent to a sibling core in
(subcores-per-core != 1) mode will be missed by the host unless the
sibling core is on the exit path to the host.
This patch adds the missing comparison operation which will ensure
that when HOST_IPI flag is set, we unconditionally exit to the host.
Fixes: 66feed61cdf6 Cc: stable@vger.kernel.org # v4.1+ Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
KVM: PPC: Book3S HV: Fix race in starting secondary threads
The current dynamic micro-threading code has a race due to which a
secondary thread naps when it is supposed to be running a vcpu. As a
side effect of this, on a guest exit, the primary thread in
kvmppc_wait_for_nap() finds that this secondary thread hasn't cleared
its vcore pointer. This results in "CPU X seems to be stuck!"
warnings.
The race is possible since the primary thread on exiting the guests
only waits for all the secondaries to clear its vcore pointer. It
subsequently expects the secondary threads to enter nap while it
unsplits the core. A secondary thread which hasn't yet entered the nap
will loop in kvm_no_guest until its vcore pointer and the do_nap flag
are unset. Once the core has been unsplit, a new vcpu thread can grab
the core and set the do_nap flag *before* setting the vcore pointers
of the secondary. As a result, the secondary thread will now enter nap
via kvm_unsplit_nap instead of running the guest vcpu.
Fix this by setting the do_nap flag after setting the vcore pointer in
the PACA of the secondary in kvmppc_run_core. Also, ensure that a
secondary thread doesn't nap in kvm_unsplit_nap when the vcore pointer
in its PACA struct is set.
Fixes: b4deba5c41e9 Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
Sam bobroff [Tue, 26 May 2015 23:56:57 +0000 (09:56 +1000)]
KVM: PPC: Book3S: correct width in XER handling
In 64 bit kernels, the Fixed Point Exception Register (XER) is a 64
bit field (e.g. in kvm_regs and kvm_vcpu_arch) and in most places it is
accessed as such.
This patch corrects places where it is accessed as a 32 bit field by a
64 bit kernel. In some cases this is via a 32 bit load or store
instruction which, depending on endianness, will cause either the
lower or upper 32 bits to be missed. In another case it is cast as a
u32, causing the upper 32 bits to be cleared.
This patch corrects those places by extending the access methods to
64 bits.
Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com> Reviewed-by: Laurent Vivier <lvivier@redhat.com> Reviewed-by: Thomas Huth <thuth@redhat.com> Tested-by: Thomas Huth <thuth@redhat.com> Signed-off-by: Alexander Graf <agraf@suse.de>
Paul Mackerras [Thu, 16 Jul 2015 07:11:14 +0000 (17:11 +1000)]
KVM: PPC: Book3S HV: Fix preempted vcore stolen time calculation
Whenever a vcore state is VCORE_PREEMPT we need to be counting stolen
time for it. This currently isn't the case when we have a vcore that
no longer has any runnable threads in it but still has a runner task,
so we do an explicit call to kvmppc_core_start_stolen() in that case.
Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
Paul Mackerras [Thu, 16 Jul 2015 07:11:13 +0000 (17:11 +1000)]
KVM: PPC: Book3S HV: Fix preempted vcore list locking
When a vcore gets preempted, we put it on the preempted vcore list for
the current CPU. The runner task then calls schedule() and comes back
some time later and takes itself off the list. We need to be careful
to lock the list that it was put onto, which may not be the list for the
current CPU since the runner task may have moved to another CPU.
Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
Paul Mackerras [Wed, 24 Jun 2015 11:18:07 +0000 (21:18 +1000)]
KVM: PPC: Book3S HV: Implement H_CLEAR_REF and H_CLEAR_MOD
This adds implementations for the H_CLEAR_REF (test and clear reference
bit) and H_CLEAR_MOD (test and clear changed bit) hypercalls.
When clearing the reference or change bit in the guest view of the HPTE,
we also have to clear it in the real HPTE so that we can detect future
references or changes. When we do so, we transfer the R or C bit value
to the rmap entry for the underlying host page so that kvm_age_hva_hv(),
kvm_test_age_hva_hv() and kvmppc_hv_get_dirty_log() know that the page
has been referenced and/or changed.
These hypercalls are not used by Linux guests. These implementations
have been tested using a FreeBSD guest.
Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
Paul Mackerras [Wed, 24 Jun 2015 11:18:06 +0000 (21:18 +1000)]
KVM: PPC: Book3S HV: Fix bug in dirty page tracking
This fixes a bug in the tracking of pages that get modified by the
guest. If the guest creates a large-page HPTE, writes to memory
somewhere within the large page, and then removes the HPTE, we only
record the modified state for the first normal page within the large
page, when in fact the guest might have modified some other normal
page within the large page.
To fix this we use some unused bits in the rmap entry to record the
order (log base 2) of the size of the page that was modified, when
removing an HPTE. Then in kvm_test_clear_dirty_npages() we use that
order to return the correct number of modified pages.
The same thing could in principle happen when removing a HPTE at the
host's request, i.e. when paging out a page, except that we never
page out large pages, and the guest can only create large-page HPTEs
if the guest RAM is backed by large pages. However, we also fix
this case for the sake of future-proofing.
The reference bit is also subject to the same loss of information. We
don't make the same fix here for the reference bit because there isn't
an interface for userspace to find out which pages the guest has
referenced, whereas there is one for userspace to find out which pages
the guest has modified. Because of this loss of information, the
kvm_age_hva_hv() and kvm_test_age_hva_hv() functions might incorrectly
say that a page has not been referenced when it has, but that doesn't
matter greatly because we never page or swap out large pages.
Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
Paul Mackerras [Wed, 24 Jun 2015 11:18:05 +0000 (21:18 +1000)]
KVM: PPC: Book3S HV: Fix race in reading change bit when removing HPTE
The reference (R) and change (C) bits in a HPT entry can be set by
hardware at any time up until the HPTE is invalidated and the TLB
invalidation sequence has completed. This means that when removing
a HPTE, we need to read the HPTE after the invalidation sequence has
completed in order to obtain reliable values of R and C. The code
in kvmppc_do_h_remove() used to do this. However, commit 6f22bd3265fb
("KVM: PPC: Book3S HV: Make HTAB code LE host aware") removed the
read after invalidation as a side effect of other changes. This
restores the read of the HPTE after invalidation.
The user-visible effect of this bug would be that when migrating a
guest, there is a small probability that a page modified by the guest
and then unmapped by the guest might not get re-transmitted and thus
the destination might end up with a stale copy of the page.
Fixes: 6f22bd3265fb Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
Paul Mackerras [Thu, 2 Jul 2015 10:38:16 +0000 (20:38 +1000)]
KVM: PPC: Book3S HV: Implement dynamic micro-threading on POWER8
This builds on the ability to run more than one vcore on a physical
core by using the micro-threading (split-core) modes of the POWER8
chip. Previously, only vcores from the same VM could be run together,
and (on POWER8) only if they had just one thread per core. With the
ability to split the core on guest entry and unsplit it on guest exit,
we can run up to 8 vcpu threads from up to 4 different VMs, and we can
run multiple vcores with 2 or 4 vcpus per vcore.
Dynamic micro-threading is only available if the static configuration
of the cores is whole-core mode (unsplit), and only on POWER8.
To manage this, we introduce a new kvm_split_mode struct which is
shared across all of the subcores in the core, with a pointer in the
paca on each thread. In addition we extend the core_info struct to
have information on each subcore. When deciding whether to add a
vcore to the set already on the core, we now have two possibilities:
(a) piggyback the vcore onto an existing subcore, or (b) start a new
subcore.
Currently, when any vcpu needs to exit the guest and switch to host
virtual mode, we interrupt all the threads in all subcores and switch
the core back to whole-core mode. It may be possible in future to
allow some of the subcores to keep executing in the guest while
subcore 0 switches to the host, but that is not implemented in this
patch.
This adds a module parameter called dynamic_mt_modes which controls
which micro-threading (split-core) modes the code will consider, as a
bitmap. In other words, if it is 0, no micro-threading mode is
considered; if it is 2, only 2-way micro-threading is considered; if
it is 4, only 4-way, and if it is 6, both 2-way and 4-way
micro-threading mode will be considered. The default is 6.
With this, we now have secondary threads which are the primary thread
for their subcore and therefore need to do the MMU switch. These
threads will need to be started even if they have no vcpu to run, so
we use the vcore pointer in the PACA rather than the vcpu pointer to
trigger them.
It is now possible for thread 0 to find that an exit has been
requested before it gets to switch the subcore state to the guest. In
that case we haven't added the guest's timebase offset to the
timebase, so we need to be careful not to subtract the offset in the
guest exit path. In fact we just skip the whole path that switches
back to host context, since we haven't switched to the guest context.
Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
Paul Mackerras [Wed, 24 Jun 2015 11:18:03 +0000 (21:18 +1000)]
KVM: PPC: Book3S HV: Make use of unused threads when running guests
When running a virtual core of a guest that is configured with fewer
threads per core than the physical cores have, the extra physical
threads are currently unused. This makes it possible to use them to
run one or more other virtual cores from the same guest when certain
conditions are met. This applies on POWER7, and on POWER8 to guests
with one thread per virtual core. (It doesn't apply to POWER8 guests
with multiple threads per vcore because they require a 1-1 virtual to
physical thread mapping in order to be able to use msgsndp and the
TIR.)
The idea is that we maintain a list of preempted vcores for each
physical cpu (i.e. each core, since the host runs single-threaded).
Then, when a vcore is about to run, it checks to see if there are
any vcores on the list for its physical cpu that could be
piggybacked onto this vcore's execution. If so, those additional
vcores are put into state VCORE_PIGGYBACK and their runnable VCPU
threads are started as well as the original vcore, which is called
the master vcore.
After the vcores have exited the guest, the extra ones are put back
onto the preempted list if any of their VCPUs are still runnable and
not idle.
This means that vcpu->arch.ptid is no longer necessarily the same as
the physical thread that the vcpu runs on. In order to make it easier
for code that wants to send an IPI to know which CPU to target, we
now store that in a new field in struct vcpu_arch, called thread_cpu.
Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Tested-by: Laurent Vivier <lvivier@redhat.com> Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
Thomas Huth [Fri, 22 May 2015 07:25:02 +0000 (09:25 +0200)]
KVM: PPC: Fix warnings from sparse
When compiling the KVM code for POWER with "make C=1", sparse
complains about functions missing proper prototypes and a 64-bit
constant missing the ULL prefix. Let's fix this by making the
functions static or by including the proper header with the
prototypes, and by appending a ULL prefix to the constant
PPC_MPPE_ADDRESS_MASK.
Signed-off-by: Thomas Huth <thuth@redhat.com> Signed-off-by: Alexander Graf <agraf@suse.de>
Thomas Huth [Fri, 22 May 2015 09:41:01 +0000 (11:41 +0200)]
KVM: PPC: Remove PPC970 from KVM_BOOK3S_64_HV text in Kconfig
Since the PPC970 support has been removed from the kvm-hv kernel
module recently, we should also reflect this change in the help
text of the corresponding Kconfig option.
Signed-off-by: Thomas Huth <thuth@redhat.com> Signed-off-by: Alexander Graf <agraf@suse.de>
Mario Smarduch [Thu, 16 Jul 2015 21:29:38 +0000 (22:29 +0100)]
arm: KVM: keep arm vfp/simd exit handling consistent with arm64
After enhancing arm64 FP/SIMD exit handling, ARMv7 VFP exit branch is moved
to guest trap handling. This allows us to keep exit handling flow between both
architectures consistent.
Signed-off-by: Mario Smarduch <m.smarduch@samsung.com> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Mario Smarduch [Thu, 16 Jul 2015 21:29:37 +0000 (22:29 +0100)]
arm64: KVM: Optimize arm64 skip 30-50% vfp/simd save/restore on exits
This patch only saves and restores FP/SIMD registers on Guest access. To do
this cptr_el2 FP/SIMD trap is set on Guest entry and later checked on exit.
lmbench, hackbench show significant improvements, for 30-50% exits FP/SIMD
context is not saved/restored
[chazy/maz: fixed save/restore logic for 32bit guests]
Signed-off-by: Mario Smarduch <m.smarduch@samsung.com> Signed-off-by: Christoffer Dall <christoffer.dall@linaro.org> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Paolo Bonzini [Thu, 13 Aug 2015 09:51:50 +0000 (11:51 +0200)]
Merge tag 'kvm-s390-next-20150812' of git://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into HEAD
KVM: s390: fix and feature for kvm/next (4.3)
1. error handling for irq routes
2. Gracefully handle STP time changes
s390 supports a protocol for syncing different systems via the stp
protocol that will steer the TOD clocks to keep all participating
clocks below the round trip time between the system. In case of
specific out of sync event Linux can opt-in to accept sync checks.
This will result in non-monotonic jumps of the TOD clock, which
Linux will correct via time offsets to keep the wall clock time
monotonic. Now: KVM guests also base their time on the host TOD,
so we need to fixup the offset for them as well.
Marc Zyngier [Mon, 23 Jun 2014 12:59:13 +0000 (13:59 +0100)]
KVM: arm/arm64: timer: Allow the timer to control the active state
In order to remove the crude hack where we sneak the masked bit
into the timer's control register, make use of the phys_irq_map
API control the active state of the interrupt.
This causes some limited changes to allow for potential error
propagation.
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Marc Zyngier [Fri, 24 Jul 2015 10:30:43 +0000 (11:30 +0100)]
KVM: arm/arm64: vgic: Prevent userspace injection of a mapped interrupt
Virtual interrupts mapped to a HW interrupt should only be triggered
from inside the kernel. Otherwise, you could end up confusing the
kernel (and the GIC's) state machine.
Rearrange the injection path so that kvm_vgic_inject_irq is
used for non-mapped interrupts, and kvm_vgic_inject_mapped_irq is
used for mapped interrupts. The latter should only be called from
inside the kernel (timer, irqfd).
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Marc Zyngier [Mon, 8 Jun 2015 15:06:13 +0000 (16:06 +0100)]
KVM: arm/arm64: vgic: Allow HW interrupts to be queued to a guest
To allow a HW interrupt to be injected into a guest, we lookup the
guest virtual interrupt in the irq_phys_map list, and if we have
a match, encode both interrupts in the LR.
We also mark the interrupt as "active" at the host distributor level.
On guest EOI on the virtual interrupt, the host interrupt will be
deactivated.
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Marc Zyngier [Mon, 8 Jun 2015 14:21:32 +0000 (15:21 +0100)]
KVM: arm/arm64: vgic: Convert struct vgic_lr to use bitfields
As we're about to cram more information in the vgic_lr structure
(HW interrupt number and additional state information), we switch
to a layout similar to the HW's:
- use bitfields to save space (we don't need more than 10 bits
to represent the irq numbers)
- source CPU and HW interrupt can share the same field, as
a SGI doesn't have a physical line.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Marc Zyngier [Mon, 8 Jun 2015 14:00:28 +0000 (15:00 +0100)]
arm/arm64: KVM: Move vgic handling to a non-preemptible section
As we're about to introduce some serious GIC-poking to the vgic code,
it is important to make sure that we're going to poke the part of
the GIC that belongs to the CPU we're about to run on (otherwise,
we'd end up with some unexpected interrupts firing)...
Introducing a non-preemptible section in kvm_arch_vcpu_ioctl_run
prevents the problem from occuring.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Marc Zyngier [Fri, 5 Jun 2015 08:33:28 +0000 (09:33 +0100)]
arm/arm64: KVM: Fix ordering of timer/GIC on guest entry
As we now inject the timer interrupt when we're about to enter
the guest, it makes a lot more sense to make sure this happens
before the vgic code queues the pending interrupts.
Otherwise, we get the interrupt on the following exit, which is
not great for latency (and leads to all kind of bizarre issues
when using with active interrupts at the HW level).
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Vladimir Murzin [Tue, 28 Jul 2015 09:42:23 +0000 (10:42 +0100)]
arm64: KVM: remove remaining reference to vgic_sr_vectors
Since commit 8a14849 (arm64: KVM: Switch vgic save/restore to
alternative_insn) vgic_sr_vectors is not used anymore, so remove
remaining leftovers and kill the structure.
Signed-off-by: Vladimir Murzin <vladimir.murzin@arm.com> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
This patch adds a generic ARM v8 KVM target cpu type for use
by the new CPUs which eventualy ends up using the common sys_reg
table. For backward compatibility the existing targets have been
preserved. Any new target CPU that can be covered by generic v8
sys_reg tables should make use of the new generic target.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com> Acked-by: Marc Zyngier <Marc.Zyngier@arm.com> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Wei Huang [Fri, 7 Aug 2015 19:53:30 +0000 (15:53 -0400)]
KVM: x86/vPMU: Fix unnecessary signed extension for AMD PERFCTRn
According to AMD programmer's manual, AMD PERFCTRn is 64-bit MSR which,
unlike Intel perf counters, doesn't require signed extension. This
patch removes the unnecessary conversion in SVM vPMU code when PERFCTRn
is being updated.
Signed-off-by: Wei Huang <wei@redhat.com> Reviewed-by: Andrew Jones <drjones@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Nicholas Krause [Wed, 5 Aug 2015 14:44:40 +0000 (10:44 -0400)]
kvm: x86: Fix error handling in the function kvm_lapic_sync_from_vapic
This fixes error handling in the function kvm_lapic_sync_from_vapic
by checking if the call to kvm_read_guest_cached has returned a
error code to signal to its caller the call to this function has
failed and due to this we must immediately return to the caller
of kvm_lapic_sync_from_vapic to avoid incorrectly call apic_set_tpc
if a error has occurred here.
Signed-off-by: Nicholas Krause <xerofoify@gmail.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Nicholas Krause [Thu, 6 Aug 2015 17:05:54 +0000 (13:05 -0400)]
KVM: s390: Fix assumption that kvm_set_irq_routing is always run successfully
This fixes the assumption that kvm_set_irq_routing is always run
successfully by instead making it equal to the variable r which
we use for returning in the function kvm_arch_vm_ioctl instead
of making r equal to zero when calling this particular function
and incorrectly making the caller of kvm_arch_vm_ioctl think
the function has run successfully.
Signed-off-by: Nicholas Krause <xerofoify@gmail.com>
Message-Id: <1438880754-27149-1-git-send-email-xerofoify@gmail.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Xiao Guangrong [Wed, 5 Aug 2015 04:04:26 +0000 (12:04 +0800)]
KVM: MMU: fully check zero bits for sptes
The #PF with PFEC.RSV = 1 is designed to speed MMIO emulation, however,
it is possible that the RSV #PF is caused by real BUG by mis-configure
shadow page table entries
This patch enables full check for the zero bits on shadow page table
entries (which includes not only bits reserved by the hardware, but also
bits that will never be set in the SPTE), then dump the shadow page table
hierarchy.
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Xiao Guangrong [Wed, 5 Aug 2015 04:04:25 +0000 (12:04 +0800)]
KVM: MMU: introduce is_shadow_zero_bits_set()
We have the same data struct to check reserved bits on guest page tables
and shadow page tables, split is_rsvd_bits_set() so that the logic can be
shared between these two paths
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Xiao Guangrong [Wed, 5 Aug 2015 04:04:24 +0000 (12:04 +0800)]
KVM: MMU: introduce the framework to check zero bits on sptes
We have abstracted the data struct and functions which are used to check
reserved bit on guest page tables, now we extend the logic to check
zero bits on shadow page tables
The zero bits on sptes include not only reserved bits on hardware but also
the bits that SPTEs willnever use. For example, shadow pages will never
use GB pages unless the guest uses them too.
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Xiao Guangrong [Wed, 5 Aug 2015 04:04:23 +0000 (12:04 +0800)]
KVM: MMU: split reset_rsvds_bits_mask_ept
Since shadow ept page tables and Intel nested guest page tables have the
same format, split reset_rsvds_bits_mask_ept so that the logic can be
reused by later patches which check zero bits on sptes
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Xiao Guangrong [Wed, 5 Aug 2015 04:04:22 +0000 (12:04 +0800)]
KVM: MMU: split reset_rsvds_bits_mask
Since softmmu & AMD nested shadow page tables and guest page tables have
the same format, split reset_rsvds_bits_mask so that the logic can be
reused by later patches which check zero bits on sptes
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Xiao Guangrong [Wed, 5 Aug 2015 04:04:21 +0000 (12:04 +0800)]
KVM: MMU: introduce rsvd_bits_validate
These two fields, rsvd_bits_mask and bad_mt_xwr, in "struct kvm_mmu" are
used to check if reserved bits set on guest ptes, move them to a data
struct so that the approach can be applied to check host shadow page
table entries as well
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Xiao Guangrong [Wed, 5 Aug 2015 04:04:19 +0000 (12:04 +0800)]
KVM: MMU: fix validation of mmio page fault
We got the bug that qemu complained with "KVM: unknown exit, hardware
reason 31" and KVM shown these info:
[84245.284948] EPT: Misconfiguration.
[84245.285056] EPT: GPA: 0xfeda848
[84245.285154] ept_misconfig_inspect_spte: spte 0x5eaef50107 level 4
[84245.285344] ept_misconfig_inspect_spte: spte 0x5f5fadc107 level 3
[84245.285532] ept_misconfig_inspect_spte: spte 0x5141d18107 level 2
[84245.285723] ept_misconfig_inspect_spte: spte 0x52e40dad77 level 1
This is because we got a mmio #PF and the handler see the mmio spte becomes
normal (points to the ram page)
However, this is valid after introducing fast mmio spte invalidation which
increases the generation-number instead of zapping mmio sptes, a example
is as follows:
1. QEMU drops mmio region by adding a new memslot
2. invalidate all mmio sptes
3.
VCPU 0 VCPU 1
access the invalid mmio spte
access the region originally was MMIO before
set the spte to the normal ram map
mmio #PF
check the spte and see it becomes normal ram mapping !!!
This patch fixes the bug just by dropping the check in mmio handler, it's
good for backport. Full check will be introduced in later patches
Reported-by: Pavel Shirshov <ru.pchel@gmail.com> Tested-by: Pavel Shirshov <ru.pchel@gmail.com> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com> Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Alex Williamson [Tue, 4 Aug 2015 16:58:26 +0000 (10:58 -0600)]
KVM: MTRR: Use default type for non-MTRR-covered gfn before WARN_ON
The patch was munged on commit to re-order these tests resulting in
excessive warnings when trying to do device assignment. Return to
original ordering: https://lkml.org/lkml/2015/7/15/769
Fixes: 3e5d2fdceda1 ("KVM: MTRR: simplify kvm_mtrr_get_guest_memory_type") Signed-off-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: Xiao Guangrong <guangrong.xiao@linux.intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Fan Zhang [Wed, 13 May 2015 08:58:41 +0000 (10:58 +0200)]
KVM: s390: host STP toleration for VMs
If the host has STP enabled, the TOD of the host will be changed during
synchronization phases. These are performed during a stop_machine() call.
As the guest TOD is based on the host TOD, we have to make sure that:
- no VCPU is in the SIE (implicitly guaranteed via stop_machine())
- manual guest TOD calculations are not affected
"Epoch" is the guest TOD clock delta to the host TOD clock. We have to
adjust that value during the STP synchronization and make sure that code
that accesses the epoch won't get interrupted in between (via disabling
preemption).
Signed-off-by: Fan Zhang <zhangfan@linux.vnet.ibm.com> Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com> Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com> Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Paolo Bonzini [Wed, 29 Jul 2015 10:31:15 +0000 (12:31 +0200)]
KVM: x86: clean/fix memory barriers in irqchip_in_kernel
The memory barriers are trying to protect against concurrent RCU-based
interrupt injection, but the IRQ routing table is not valid at the time
kvm->arch.vpic is written. Fix this by writing kvm->arch.vpic last.
kvm_destroy_pic then need not set kvm->arch.vpic to NULL; modify it
to take a struct kvm_pic* and reuse it if the IOAPIC creation fails.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Wed, 29 Jul 2015 09:06:34 +0000 (11:06 +0200)]
KVM: x86: remove unnecessary memory barriers for shared MSRs
There is no smp_rmb matching the smp_wmb. shared_msr_update is called from
hardware_enable, which in turn is called via on_each_cpu. on_each_cpu
and must imply a read memory barrier (on x86 the rmb is achieved simply
through asm volatile in native_apic_mem_write).
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Wed, 29 Jul 2015 10:53:58 +0000 (12:53 +0200)]
Merge tag 'kvm-s390-next-20150728' of git://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into kvm-next
KVM: s390: Fixes and features for kvm/next (4.3)
1. Rework logging infrastructure (s390dbf) to integrate feedback learned
when debugging performance and test issues
2. Some cleanups and simplifications for CMMA handling
3. Fix gdb debugging and single stepping on some instructions
4. Error handling for storage key setup
KVM: s390: log capability enablement and vm attribute changes
Depending on user space, some capabilities and vm attributes are
enabled at runtime. Let's log those events and while we're at it,
log querying the vm attributes as well.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
KVM: s390: adapt debug entries for instruction handling
Use the default log level 3 for state changing and/or seldom events,
use 4 for others. Also change some numbers from %x to %d and vice versa
to match documentation. If hex, let's prepend the numbers with 0x.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
We do not use the exception logger, so the 2nd area is unused.
Just have one area that is bigger (32 pages).
At the same time we can limit the debug feature size to 7
longs, as the largest user has 3 parameters + string + boiler
plate (vCPU, PSW mask, PSW addr)
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
This patch adds names for missing irq types to the trace events.
In order to identify adapter irqs, the define is moved from
interrupt.c to the other basic irq defines in uapi/linux/kvm.h.
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>
KVM: s390: Fixup interrupt vcpu event messages and levels
This reworks the debug logging for interrupt related logs.
Several changes:
- unify program int/irq
- improve decoding (e.g. use mcic instead of parm64 for machine
check injection)
- remove useless interrupt type number (the name is enough)
- rename "interrupt:" to "deliver:" as the other side is called "inject"
- use log level 3 for state changing and/or seldom events (like machine
checks, restart..)
- use log level 4 for frequent events
- use 0x prefix for hex numbers
- add pfault done logging
- move some tracing outside spinlock
KVM: s390: add more debug data for the pfault diagnoses
We're not only interested in the address of the control block, but
also in the requested subcommand and for the token subcommand, in the
specified token address and masks.
Suggested-by: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com> Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
KVM: s390: remove "from (user|kernel)" from irq injection messages
The "from user"/"from kernel" part of the log/trace messages is not
always correct anymore and therefore not really helpful.
Let's remove that part from the log + trace messages. For program
interrupts, we can now move the logging/tracing part into the real
injection function, as already done for the other injection functions.
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>
SPX (SET PREFIX) and SIGP (Set prefix) can change the prefix
register of a CPU. As sigp set prefix may be handled in user
space (KVM_CAP_S390_USER_SIGP), we would not log the changes
triggered via SIGP in that case. Let's have just one VCPU_EVENT
at the central location that tracks prefix changes.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com> Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
KVM: s390: Improve vcpu event debugging for diagnoses
Let's add a vcpu event for the page reference handling and change
the default debugging level for the ipl diagnose. Both are not
frequent AND change the global state, so lets log them always.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Dominik Dingel [Thu, 7 May 2015 13:41:57 +0000 (15:41 +0200)]
KVM: s390: clean up cmma_enable check
As we already only enable CMMA when userspace requests it, we can
safely move the additional checks to the request handler and avoid
doing them multiple times. This also tells userspace if CMMA is
available.
Signed-off-by: Dominik Dingel <dingel@linux.vnet.ibm.com> Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
KVM: s390: filter space-switch events when PER is enforced
When guest debugging is active, space-switch events might be enforced
by PER. While the PER events are correctly filtered out,
space-switch-events could be forwarded to the guest, although from a
guest point of view, they should not have been reported.
Therefore we have to filter out space-switch events being concurrently
reported with a PER event, if the PER event got filtered out. To do so,
we theoretically have to know which instruction was responsible for the
event. As the applicable instructions modify the PSW address, the
address space set in the PSW and even the address space in cr1, we
can't figure out the instruction that way.
For this reason, we have to rely on the information about the old and
new address space, in order to guess the responsible instruction type
and do appropriate checks for space-switch events.
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Paolo Bonzini [Fri, 10 Jul 2015 11:32:13 +0000 (13:32 +0200)]
KVM: svm: handle KVM_X86_QUIRK_CD_NW_CLEARED in svm_get_mt_mask
We can disable CD unconditionally when there is no assigned device.
KVM now forces guest PAT to all-writeback in that case, so it makes
sense to also force CR0.CD=0.
When there are assigned devices, emulate cache-disabled operation
through the page tables. This behavior is consistent with VMX
microcode, where CD/NW are not touched by vmentry/vmexit. However,
keep this dependent on the quirk because OVMF enables the caches
too late.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
kvm/x86: add sending hyper-v crash notification to user space
Sending of notification is done by exiting vcpu to user space
if KVM_REQ_HV_CRASH is enabled for vcpu. At exit to user space
the kvm_run structure contains system_event with type
KVM_SYSTEM_EVENT_CRASH to notify about guest crash occurred.
Signed-off-by: Andrey Smetanin <asmetanin@virtuozzo.com> Signed-off-by: Denis V. Lunev <den@openvz.org> Reviewed-by: Peter Hornyack <peterhornyack@google.com> CC: Paolo Bonzini <pbonzini@redhat.com> CC: Gleb Natapov <gleb@kernel.org> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
kvm/x86: move Hyper-V MSR's/hypercall code into hyperv.c file
This patch introduce Hyper-V related source code file - hyperv.c and
per vm and per vcpu hyperv context structures.
All Hyper-V MSR's and hypercall code moved into hyperv.c.
All Hyper-V kvm/vcpu fields moved into appropriate hyperv context
structures. Copyrights and authors information copied from x86.c
to hyperv.c.
Signed-off-by: Andrey Smetanin <asmetanin@virtuozzo.com> Signed-off-by: Denis V. Lunev <den@openvz.org> Reviewed-by: Peter Hornyack <peterhornyack@google.com> CC: Paolo Bonzini <pbonzini@redhat.com> CC: Gleb Natapov <gleb@kernel.org> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
KVM: nVMX: VMX instructions: add checks for #GP/#SS exceptions
According to Intel SDM several checks must be applied for memory operands
of VMX instructions.
Long mode: #GP(0) or #SS(0) depending on the segment must be thrown
if the memory address is in a non-canonical form.
Protected mode, checks in chronological order:
- The segment type must be checked with access type (read or write) taken
into account.
For write access: #GP(0) must be generated if the destination operand
is located in a read-only data segment or any code segment.
For read access: #GP(0) must be generated if if the source operand is
located in an execute-only code segment.
- Usability of the segment must be checked. #GP(0) or #SS(0) depending on the
segment must be thrown if the segment is unusable.
- Limit check. #GP(0) or #SS(0) depending on the segment must be
thrown if the memory operand effective address is outside the segment
limit.
Signed-off-by: Eugene Korenevsky <ekorenevsky@gmail.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
kvm_mtrr_get_guest_memory_type never returns -1 which is implied
in the current code since if @type = -1 (means no MTRR contains the
range), iter.partial_map must be true
Simplify the code to indicate this fact
Signed-off-by: Xiao Guangrong <guangrong.xiao@intel.com> Tested-by: Alex Williamson <alex.williamson@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
1) Don't use shared bluetooth antenna in iwlwifi driver for management
frames, from Emmanuel Grumbach.
2) Fix device ID check in ath9k driver, from Felix Fietkau.
3) Off by one in xen-netback BUG checks, from Dan Carpenter.
4) Fix IFLA_VF_PORT netlink attribute validation, from Daniel Borkmann.
5) Fix races in setting peeked bit flag in SKBs during datagram
receive. If it's shared we have to clone it otherwise the value can
easily be corrupted. Fix from Herbert Xu.
7) Fix use after free in fq_codel and sfq packet schedulers, from WANG
Cong.
8) ipvlan bug fixes (memory leaks, missing rcu_dereference_bh, etc.)
from WANG Cong and Konstantin Khlebnikov.
9) Memory leak in act_bpf packet action, from Alexei Starovoitov.
10) ARM bpf JIT bug fixes from Nicolas Schichan.
11) Fix backwards compat of ANY_LAYOUT in virtio_net driver, from
Michael S Tsirkin.
12) Destruction of bond with different ARP header types not handled
correctly, fix from Nikolay Aleksandrov.
13) Revert GRO receive support in ipv6 SIT tunnel driver, causes
regressions because the GRO packets created cannot be processed
properly on the GSO side if we forward the frame. From Herbert Xu.
14) TCCR update race and other fixes to ravb driver from Sergei
Shtylyov.
15) Fix SKB leaks in caif_queue_rcv_skb(), from Eric Dumazet.
16) Fix panics on packet scheduler filter replace, from Daniel Borkmann.
17) Make sure AF_PACKET sees properly IP headers in defragmented frames
(via PACKET_FANOUT_FLAG_DEFRAG option), from Edward Hyunkoo Jee.
18) AF_NETLINK cannot hold mutex in RCU callback, fix from Florian
Westphal.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (84 commits)
ravb: fix ring memory allocation
net: phy: dp83867: Fix warning check for setting the internal delay
openvswitch: allocate nr_node_ids flow_stats instead of num_possible_nodes
netlink: don't hold mutex in rcu callback when releasing mmapd ring
ARM: net: fix vlan access instructions in ARM JIT.
ARM: net: handle negative offsets in BPF JIT.
ARM: net: fix condition for load_order > 0 when translating load instructions.
tcp: suppress a division by zero warning
drivers: net: cpsw: remove tx event processing in rx napi poll
inet: frags: fix defragmented packet's IP header for af_packet
net: mvneta: fix refilling for Rx DMA buffers
stmmac: fix setting of driver data in stmmac_dvr_probe
sched: cls_flow: fix panic on filter replace
sched: cls_flower: fix panic on filter replace
sched: cls_bpf: fix panic on filter replace
net/mdio: fix mdio_bus_match for c45 PHY
net: ratelimit warnings about dst entry refcount underflow or overflow
caif: fix leaks and race in caif_queue_rcv_skb()
qmi_wwan: add the second QMI/network interface for Sierra Wireless MC7305/MC7355
ravb: fix race updating TCCR
...
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull ARM64 fixes from Catalin Marinas:
- arm64 build fix following the move of the thread_struct to the end of
task_struct and the asm offsets becoming too large for the AArch64
ISA
- preparatory patch for moving irq_data struct members (applied now to
reduce dependency for the next merging window)
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
ARM64/irq: Use access helper irq_data_get_affinity_mask()
arm64: switch_to: calculate cpu context pointer using separate register
Will Deacon [Mon, 20 Jul 2015 14:14:53 +0000 (15:14 +0100)]
arm64: switch_to: calculate cpu context pointer using separate register
Commit 0c8c0f03e3a2 ("x86/fpu, sched: Dynamically allocate 'struct fpu'")
moved the thread_struct to the bottom of task_struct. As a result, the
offset is now too large to be used in an immediate add on arm64 with
some kernel configs:
arch/arm64/kernel/entry.S: Assembler messages:
arch/arm64/kernel/entry.S:588: Error: immediate out of range
arch/arm64/kernel/entry.S:597: Error: immediate out of range
This patch calculates the offset using an additional register instead of
an immediate offset.
Sergei Shtylyov [Tue, 21 Jul 2015 22:31:59 +0000 (01:31 +0300)]
ravb: fix ring memory allocation
The driver is written as if it can adapt to a low memory situation allocating
less RX skbs and TX aligned buffers than the respective RX/TX ring sizes. In
reality though the driver would malfunction in this case. Stop being overly
smart and just fail in such situation -- this is achieved by moving the memory
allocation from ravb_ring_format() to ravb_ring_init().
We leave dma_map_single() calls in place but make their failure non-fatal
by marking the corresponding RX descriptors with zero data size which should
prevent DMA to an invalid addresses.
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Dan Murphy [Tue, 21 Jul 2015 17:06:45 +0000 (12:06 -0500)]
net: phy: dp83867: Fix warning check for setting the internal delay
Fix warning: logical ‘or’ of collectively exhaustive tests is always true
Change the internal delay check from an 'or' condition to an 'and'
condition.
Reported-by: David Binderman <dcb314@hotmail.com> Signed-off-by: Dan Murphy <dmurphy@ti.com> Acked-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
openvswitch: allocate nr_node_ids flow_stats instead of num_possible_nodes
Some architectures like POWER can have a NUMA node_possible_map that
contains sparse entries. This causes memory corruption with openvswitch
since it allocates flow_cache with a multiple of num_possible_nodes() and
assumes the node variable returned by for_each_node will index into
flow->stats[node].
Use nr_node_ids to allocate a maximal sparse array instead of
num_possible_nodes().
The crash was noticed after 3af229f2 was applied as it changed the
node_possible_map to match node_online_map on boot. Fixes: 3af229f2071f5b5cb31664be6109561fbe19c861 Signed-off-by: Chris J Arges <chris.j.arges@canonical.com> Acked-by: Pravin B Shelar <pshelar@nicira.com> Acked-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The socket should be dead at this point. It might be simpler to
add a netlink_release_ring() function which doesn't require
locking at all.
Reported-by: "Kirill A. Shutemov" <kirill@shutemov.name> Diagnosed-by: Cong Wang <cwang@twopensource.com> Suggested-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 22 Jul 2015 05:19:55 +0000 (22:19 -0700)]
Merge branch 'arm-bpf-fixes'
Nicolas Schichan says:
====================
BPF JIT fixes for ARM
These patches are fixing bugs in the ARM JIT and should probably find
their way to a stable kernel. All 60 test_bpf tests in Linux 4.1 release
are now passing OK (was 54 out of 60 before).
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Nicolas Schichan [Tue, 21 Jul 2015 12:14:14 +0000 (14:14 +0200)]
ARM: net: fix vlan access instructions in ARM JIT.
This makes BPF_ANC | SKF_AD_VLAN_TAG and BPF_ANC | SKF_AD_VLAN_TAG_PRESENT
have the same behaviour as the in kernel VM and makes the test_bpf LD_VLAN_TAG
and LD_VLAN_TAG_PRESENT tests pass.
Signed-off-by: Nicolas Schichan <nschichan@freebox.fr> Signed-off-by: David S. Miller <davem@davemloft.net>
Nicolas Schichan [Tue, 21 Jul 2015 12:14:13 +0000 (14:14 +0200)]
ARM: net: handle negative offsets in BPF JIT.
Previously, the JIT would reject negative offsets known during code
generation and mishandle negative offsets provided at runtime.
Fix that by calling bpf_internal_load_pointer_neg_helper()
appropriately in the jit_get_skb_{b,h,w} slow path helpers and by forcing
the execution flow to the slow path helpers when the offset is
negative.
Signed-off-by: Nicolas Schichan <nschichan@freebox.fr> Signed-off-by: David S. Miller <davem@davemloft.net>
Nicolas Schichan [Tue, 21 Jul 2015 12:14:12 +0000 (14:14 +0200)]
ARM: net: fix condition for load_order > 0 when translating load instructions.
To check whether the load should take the fast path or not, the code
would check that (r_skb_hlen - load_order) is greater than the offset
of the access using an "Unsigned higher or same" condition. For
halfword accesses and an skb length of 1 at offset 0, that test is
valid, as we end up comparing 0xffffffff(-1) and 0, so the fast path
is taken and the filter allows the load to wrongly succeed. A similar
issue exists for word loads at offset 0 and an skb length of less than
4.
Fix that by using the condition "Signed greater than or equal"
condition for the fast path code for load orders greater than 0.
Signed-off-by: Nicolas Schichan <nschichan@freebox.fr> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Wed, 22 Jul 2015 05:02:00 +0000 (07:02 +0200)]
tcp: suppress a division by zero warning
Andrew Morton reported following warning on one ARM build
with gcc-4.4 :
net/ipv4/inet_hashtables.c: In function 'inet_ehash_locks_alloc':
net/ipv4/inet_hashtables.c:617: warning: division by zero
Even guarded with a test on sizeof(spinlock_t), compiler does not
like current construct on a !CONFIG_SMP build.
Remove the warning by using a temporary variable.
Fixes: 095dc8e0c368 ("tcp: fix/cleanup inet_ehash_locks_alloc()") Reported-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kinglong Mee reports a memory leak with that patch, and Jan Kara confirms:
"Thanks for report! You are right that my patch introduces a race
between fsnotify kthread and fsnotify_destroy_group() which can result
in leaking inotify event on group destruction.
I haven't yet decided whether the right fix is not to queue events for
dying notification group (as that is pointless anyway) or whether we
should just fix the original problem differently... Whenever I look
at fsnotify code mark handling I get lost in the maze of locks, lists,
and subtle differences between how different notification systems
handle notification marks :( I'll think about it over night"
and after thinking about it, Jan says:
"OK, I have looked into the code some more and I found another
relatively simple way of fixing the original oops. It will be IMHO
better than trying to fixup this issue which has more potential for
breakage. I'll ask Linus to revert the fsnotify fix he already merged
and send a new fix"
Reported-by: Kinglong Mee <kinglongmee@gmail.com> Requested-by: Jan Kara <jack@suse.cz> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>