pandora-kernel.git
12 years agoKVM: s390: Make psw available on all exits, not just a subset
Carsten Otte [Thu, 19 Nov 2009 13:21:16 +0000 (14:21 +0100)]
KVM: s390: Make psw available on all exits, not just a subset

This patch moves s390 processor status word into the base kvm_run
struct and keeps it up-to date on all userspace exits.

The userspace ABI is broken by this, however there are no applications
in the wild using this.  A capability check is provided so users can
verify the updated API exists.

Cc: stable@kernel.org
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86: Add KVM_GET/SET_VCPU_EVENTS
Jan Kiszka [Thu, 12 Nov 2009 00:04:25 +0000 (01:04 +0100)]
KVM: x86: Add KVM_GET/SET_VCPU_EVENTS

This new IOCTL exports all yet user-invisible states related to
exceptions, interrupts, and NMIs. Together with appropriate user space
changes, this fixes sporadic problems of vmsave/restore, live migration
and system reset.

[avi: future-proof abi by adding a flags field]

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: VMX: Report unexpected simultaneous exceptions as internal errors
Avi Kivity [Wed, 4 Nov 2009 09:59:01 +0000 (11:59 +0200)]
KVM: VMX: Report unexpected simultaneous exceptions as internal errors

These happen when we trap an exception when another exception is being
delivered; we only expect these with MCEs and page faults.  If something
unexpected happens, things probably went south and we're better off reporting
an internal error and freezing.

Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Allow internal errors reported to userspace to carry extra data
Avi Kivity [Wed, 4 Nov 2009 09:54:59 +0000 (11:54 +0200)]
KVM: Allow internal errors reported to userspace to carry extra data

Usually userspace will freeze the guest so we can inspect it, but some
internal state is not available.  Add extra data to internal error
reporting so we can expose it to the debugger.  Extra data is specific
to the suberror.

Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Reorder IOCTLs in main kvm.h
Jan Kiszka [Mon, 2 Nov 2009 16:20:28 +0000 (17:20 +0100)]
KVM: Reorder IOCTLs in main kvm.h

Obviously, people tend to extend this header at the bottom - more or
less blindly. Ensure that deprecated stuff gets its own corner again by
moving things to the top. Also add some comments and reindent IOCTLs to
make them more readable and reduce the risk of number collisions.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86: Polish exception injection via KVM_SET_GUEST_DEBUG
Jan Kiszka [Fri, 30 Oct 2009 11:46:59 +0000 (12:46 +0100)]
KVM: x86: Polish exception injection via KVM_SET_GUEST_DEBUG

Decouple KVM_GUESTDBG_INJECT_DB and KVM_GUESTDBG_INJECT_BP from
KVM_GUESTDBG_ENABLE, their are actually orthogonal. At this chance,
avoid triggering the WARN_ON in kvm_queue_exception if there is already
an exception pending and reject such invalid requests.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: only clear irq_source_id if irqchip is present
Marcelo Tosatti [Thu, 29 Oct 2009 15:44:17 +0000 (13:44 -0200)]
KVM: only clear irq_source_id if irqchip is present

Otherwise kvm might attempt to dereference a NULL pointer.

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86: disallow KVM_{SET,GET}_LAPIC without allocated in-kernel lapic
Marcelo Tosatti [Thu, 29 Oct 2009 15:44:16 +0000 (13:44 -0200)]
KVM: x86: disallow KVM_{SET,GET}_LAPIC without allocated in-kernel lapic

Otherwise kvm might attempt to dereference a NULL pointer.

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86: disallow multiple KVM_CREATE_IRQCHIP
Marcelo Tosatti [Thu, 29 Oct 2009 15:44:15 +0000 (13:44 -0200)]
KVM: x86: disallow multiple KVM_CREATE_IRQCHIP

Otherwise kvm will leak memory on multiple KVM_CREATE_IRQCHIP.
Also serialize multiple accesses with kvm->lock.

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: VMX: Remove vmx->msr_offset_efer
Avi Kivity [Thu, 29 Oct 2009 09:00:16 +0000 (11:00 +0200)]
KVM: VMX: Remove vmx->msr_offset_efer

This variable is used to communicate between a caller and a callee; switch
to a function argument instead.

Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: MMU: update invlpg handler comment
Marcelo Tosatti [Mon, 26 Oct 2009 18:50:14 +0000 (16:50 -0200)]
KVM: MMU: update invlpg handler comment

Large page translations are always synchronized (either in level 3
or level 2), so its not necessary to properly deal with them
in the invlpg handler.

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: VMX: move CR3/PDPTR update to vmx_set_cr3
Marcelo Tosatti [Mon, 26 Oct 2009 18:48:33 +0000 (16:48 -0200)]
KVM: VMX: move CR3/PDPTR update to vmx_set_cr3

GUEST_CR3 is updated via kvm_set_cr3 whenever CR3 is modified from
outside guest context. Similarly pdptrs are updated via load_pdptrs.

Let kvm_set_cr3 perform the update, removing it from the vcpu_run
fast path.

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Acked-by: Acked-by: Sheng Yang <sheng@linux.intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: remove duplicated task_switch check
Gleb Natapov [Sun, 25 Oct 2009 15:45:07 +0000 (17:45 +0200)]
KVM: remove duplicated task_switch check

Probably introduced by a bad merge.

Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: powerpc: Fix BUILD_BUG_ON condition
Hollis Blanchard [Fri, 23 Oct 2009 00:35:30 +0000 (00:35 +0000)]
KVM: powerpc: Fix BUILD_BUG_ON condition

The old BUILD_BUG_ON implementation didn't work with __builtin_constant_p().
Fixing that revealed this test had been inverted for a long time without
anybody noticing...

Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: VMX: Use shared msr infrastructure
Avi Kivity [Mon, 7 Sep 2009 08:14:12 +0000 (11:14 +0300)]
KVM: VMX: Use shared msr infrastructure

Instead of reloading syscall MSRs on every preemption, use the new shared
msr infrastructure to reload them at the last possible minute (just before
exit to userspace).

Improves vcpu/idle/vcpu switches by about 2000 cycles (when EFER needs to be
reloaded as well).

[jan: fix slot index missing indirection]

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86 shared msr infrastructure
Avi Kivity [Mon, 7 Sep 2009 08:12:18 +0000 (11:12 +0300)]
KVM: x86 shared msr infrastructure

The various syscall-related MSRs are fairly expensive to switch.  Currently
we switch them on every vcpu preemption, which is far too often:

- if we're switching to a kernel thread (idle task, threaded interrupt,
  kernel-mode virtio server (vhost-net), for example) and back, then
  there's no need to switch those MSRs since kernel threasd won't
  be exiting to userspace.

- if we're switching to another guest running an identical OS, most likely
  those MSRs will have the same value, so there's little point in reloading
  them.

- if we're running the same OS on the guest and host, the MSRs will have
  identical values and reloading is unnecessary.

This patch uses the new user return notifiers to implement last-minute
switching, and checks the msr values to avoid unnecessary reloading.

Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: VMX: Move MSR_KERNEL_GS_BASE out of the vmx autoload msr area
Avi Kivity [Sun, 6 Sep 2009 12:55:37 +0000 (15:55 +0300)]
KVM: VMX: Move MSR_KERNEL_GS_BASE out of the vmx autoload msr area

Currently MSR_KERNEL_GS_BASE is saved and restored as part of the
guest/host msr reloading.  Since we wish to lazy-restore all the other
msrs, save and reload MSR_KERNEL_GS_BASE explicitly instead of using
the common code.

Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: SVM: init_vmcb(): remove redundant save->cr0 initialization
Eduardo Habkost [Sat, 24 Oct 2009 04:50:00 +0000 (02:50 -0200)]
KVM: SVM: init_vmcb(): remove redundant save->cr0 initialization

The svm_set_cr0() call will initialize save->cr0 properly even when npt is
enabled, clearing the NW and CD bits as expected, so we don't need to
initialize it manually for npt_enabled anymore.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: SVM: Reset cr0 properly on vcpu reset
Eduardo Habkost [Sat, 24 Oct 2009 04:49:59 +0000 (02:49 -0200)]
KVM: SVM: Reset cr0 properly on vcpu reset

svm_vcpu_reset() was not properly resetting the contents of the guest-visible
cr0 register, causing the following issue:
https://bugzilla.redhat.com/show_bug.cgi?id=525699

Without resetting cr0 properly, the vcpu was running the SIPI bootstrap routine
with paging enabled, making the vcpu get a pagefault exception while trying to
run it.

Instead of setting vmcb->save.cr0 directly, the new code just resets
kvm->arch.cr0 and calls kvm_set_cr0(). The bits that were set/cleared on
vmcb->save.cr0 (PG, WP, !CD, !NW) will be set properly by svm_set_cr0().

kvm_set_cr0() is used instead of calling svm_set_cr0() directly to make sure
kvm_mmu_reset_context() is called to reset the mmu to nonpaging mode.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: VMX: Use macros instead of hex value on cr0 initialization
Eduardo Habkost [Sat, 24 Oct 2009 04:49:58 +0000 (02:49 -0200)]
KVM: VMX: Use macros instead of hex value on cr0 initialization

This should have no effect, it is just to make the code clearer.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Enable 32bit dirty log pointers on 64bit host
Arnd Bergmann [Thu, 22 Oct 2009 12:19:27 +0000 (14:19 +0200)]
KVM: Enable 32bit dirty log pointers on 64bit host

With big endian userspace, we can't quite figure out if a pointer
is 32 bit (shifted >> 32) or 64 bit when we read a 64 bit pointer.

This is what happens with dirty logging. To get the pointer interpreted
correctly, we thus need Arnd's patch to implement a compat layer for
the ioctl:

A better way to do this is to add a separate compat_ioctl() method that
converts this for you.

Based on initial patch from Arnd Bergmann.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: allow userspace to adjust kvmclock offset
Glauber Costa [Fri, 16 Oct 2009 19:28:36 +0000 (15:28 -0400)]
KVM: allow userspace to adjust kvmclock offset

When we migrate a kvm guest that uses pvclock between two hosts, we may
suffer a large skew. This is because there can be significant differences
between the monotonic clock of the hosts involved. When a new host with
a much larger monotonic time starts running the guest, the view of time
will be significantly impacted.

Situation is much worse when we do the opposite, and migrate to a host with
a smaller monotonic clock.

This proposed ioctl will allow userspace to inform us what is the monotonic
clock value in the source host, so we can keep the time skew short, and
more importantly, never goes backwards. Userspace may also need to trigger
the current data, since from the first migration onwards, it won't be
reflected by a simple call to clock_gettime() anymore.

[marcelo: future-proof abi with a flags field]
[jan: fix KVM_GET_CLOCK by clearing flags field instead of checking it]

Signed-off-by: Glauber Costa <glommer@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: fix irq_source_id size verification
Marcelo Tosatti [Sun, 18 Oct 2009 01:47:23 +0000 (22:47 -0300)]
KVM: fix irq_source_id size verification

find_first_zero_bit works with bit numbers, not bytes.

Fixes

https://sourceforge.net/tracker/?func=detail&aid=2847560&group_id=180599&atid=893831

Reported-by: "Xu, Jiajun" <jiajun.xu@intel.com>
Cc: stable@kernel.org
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Cleanup NMI singlestep
Jan Kiszka [Sun, 18 Oct 2009 11:24:54 +0000 (13:24 +0200)]
KVM: SVM: Cleanup NMI singlestep

Push the NMI-related singlestep variable into vcpu_svm. It's dealing
with an AMD-specific deficit, nothing generic for x86.

Acked-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
 arch/x86/include/asm/kvm_host.h |    1 -
 arch/x86/kvm/svm.c              |   12 +++++++-----
 2 files changed, 7 insertions(+), 6 deletions(-)
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: x86: Fix guest single-stepping while interruptible
Jan Kiszka [Sun, 18 Oct 2009 11:24:44 +0000 (13:24 +0200)]
KVM: x86: Fix guest single-stepping while interruptible

Commit 705c5323 opened the doors of hell by unconditionally injecting
single-step flags as long as guest_debug signaled this. This doesn't
work when the guest branches into some interrupt or exception handler
and triggers a vmexit with flag reloading.

Fix it by saving cs:rip when user space requests single-stepping and
restricting the trace flag injection to this guest code position.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: Xen PV-on-HVM guest support
Ed Swierk [Thu, 15 Oct 2009 22:21:43 +0000 (15:21 -0700)]
KVM: Xen PV-on-HVM guest support

Support for Xen PV-on-HVM guests can be implemented almost entirely in
userspace, except for handling one annoying MSR that maps a Xen
hypercall blob into guest address space.

A generic mechanism to delegate MSR writes to userspace seems overkill
and risks encouraging similar MSR abuse in the future.  Thus this patch
adds special support for the Xen HVM MSR.

I implemented a new ioctl, KVM_XEN_HVM_CONFIG, that lets userspace tell
KVM which MSR the guest will write to, as well as the starting address
and size of the hypercall blobs (one each for 32-bit and 64-bit) that
userspace has loaded from files.  When the guest writes to the MSR, KVM
copies one page of the blob from userspace to the guest.

I've tested this patch with a hacked-up version of Gerd's userspace
code, booting a number of guests (CentOS 5.3 i386 and x86_64, and
FreeBSD 8.0-RC1 amd64) and exercising PV network and block devices.

[jan: fix i386 build warning]
[avi: future proof abi with a flags field]

Signed-off-by: Ed Swierk <eswierk@aristanetworks.com>
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86: Drop unneeded CONFIG_HAS_IOMEM check
Jan Kiszka [Mon, 12 Oct 2009 06:51:40 +0000 (08:51 +0200)]
KVM: x86: Drop unneeded CONFIG_HAS_IOMEM check

This (broken) check dates back to the days when this code was shared
across architectures. x86 has IOMEM, so drop it.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: VMX: fix handle_pause declaration
Marcelo Tosatti [Mon, 12 Oct 2009 22:37:31 +0000 (19:37 -0300)]
KVM: VMX: fix handle_pause declaration

There's no kvm_run argument anymore.

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: x86: Harden against cpufreq
Zachary Amsden [Sat, 10 Oct 2009 02:26:08 +0000 (16:26 -1000)]
KVM: x86: Harden against cpufreq

If cpufreq can't determine the CPU khz, or cpufreq is not compiled in,
we should fallback to the measured TSC khz.

Signed-off-by: Zachary Amsden <zamsden@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Support Pause Filter in AMD processors
Mark Langsdorf [Tue, 6 Oct 2009 19:25:02 +0000 (14:25 -0500)]
KVM: SVM: Support Pause Filter in AMD processors

New AMD processors (Family 0x10 models 8+) support the Pause
Filter Feature.  This feature creates a new field in the VMCB
called Pause Filter Count.  If Pause Filter Count is greater
than 0 and intercepting PAUSEs is enabled, the processor will
increment an internal counter when a PAUSE instruction occurs
instead of intercepting.  When the internal counter reaches the
Pause Filter Count value, a PAUSE intercept will occur.

This feature can be used to detect contended spinlocks,
especially when the lock holding VCPU is not scheduled.
Rescheduling another VCPU prevents the VCPU seeking the
lock from wasting its quantum by spinning idly.

Experimental results show that most spinlocks are held
for less than 1000 PAUSE cycles or more than a few
thousand.  Default the Pause Filter Counter to 3000 to
detect the contended spinlocks.

Processor support for this feature is indicated by a CPUID
bit.

On a 24 core system running 4 guests each with 16 VCPUs,
this patch improved overall performance of each guest's
32 job kernbench by approximately 3-5% when combined
with a scheduler algorithm thati caused the VCPU to
sleep for a brief period. Further performance improvement
may be possible with a more sophisticated yield algorithm.

Signed-off-by: Mark Langsdorf <mark.langsdorf@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: VMX: Add support for Pause-Loop Exiting
Zhai, Edwin [Fri, 9 Oct 2009 10:03:20 +0000 (18:03 +0800)]
KVM: VMX: Add support for Pause-Loop Exiting

New NHM processors will support Pause-Loop Exiting by adding 2 VM-execution
control fields:
PLE_Gap    - upper bound on the amount of time between two successive
             executions of PAUSE in a loop.
PLE_Window - upper bound on the amount of time a guest is allowed to execute in
             a PAUSE loop

If the time, between this execution of PAUSE and previous one, exceeds the
PLE_Gap, processor consider this PAUSE belongs to a new loop.
Otherwise, processor determins the the total execution time of this loop(since
1st PAUSE in this loop), and triggers a VM exit if total time exceeds the
PLE_Window.
* Refer SDM volume 3b section 21.6.13 & 22.1.3.

Pause-Loop Exiting can be used to detect Lock-Holder Preemption, where one VP
is sched-out after hold a spinlock, then other VPs for same lock are sched-in
to waste the CPU time.

Our tests indicate that most spinlocks are held for less than 212 cycles.
Performance tests show that with 2X LP over-commitment we can get +2% perf
improvement for kernel build(Even more perf gain with more LPs).

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: introduce kvm_vcpu_on_spin
Zhai, Edwin [Fri, 9 Oct 2009 10:03:20 +0000 (18:03 +0800)]
KVM: introduce kvm_vcpu_on_spin

Introduce kvm_vcpu_on_spin, to be used by VMX/SVM to yield processing
once the cpu detects pause-based looping.

Signed-off-by: "Zhai, Edwin" <edwin.zhai@intel.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Remove nsvm_printk debugging code
Joerg Roedel [Fri, 9 Oct 2009 14:08:33 +0000 (16:08 +0200)]
KVM: SVM: Remove nsvm_printk debugging code

With all important informations now delivered through
tracepoints we can savely remove the nsvm_printk debugging
code for nested svm.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Add tracepoint for skinit instruction
Joerg Roedel [Fri, 9 Oct 2009 14:08:32 +0000 (16:08 +0200)]
KVM: SVM: Add tracepoint for skinit instruction

This patch adds a tracepoint for the event that the guest
executed the SKINIT instruction. This information is
important because SKINIT is an SVM extenstion not yet
implemented by nested SVM and we may need this information
for debugging hypervisors that do not yet run on nested SVM.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Add tracepoint for invlpga instruction
Joerg Roedel [Fri, 9 Oct 2009 14:08:31 +0000 (16:08 +0200)]
KVM: SVM: Add tracepoint for invlpga instruction

This patch adds a tracepoint for the event that the guest
executed the INVLPGA instruction.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Add tracepoint for #vmexit because intr pending
Joerg Roedel [Fri, 9 Oct 2009 14:08:30 +0000 (16:08 +0200)]
KVM: SVM: Add tracepoint for #vmexit because intr pending

This patch adds a special tracepoint for the event that a
nested #vmexit is injected because kvm wants to inject an
interrupt into the guest.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Add tracepoint for injected #vmexit
Joerg Roedel [Fri, 9 Oct 2009 14:08:29 +0000 (16:08 +0200)]
KVM: SVM: Add tracepoint for injected #vmexit

This patch adds a tracepoint for a nested #vmexit that gets
re-injected to the guest.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Add tracepoint for nested #vmexit
Joerg Roedel [Fri, 9 Oct 2009 14:08:28 +0000 (16:08 +0200)]
KVM: SVM: Add tracepoint for nested #vmexit

This patch adds a tracepoint for every #vmexit we get from a
nested guest.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Add tracepoint for nested vmrun
Joerg Roedel [Fri, 9 Oct 2009 14:08:27 +0000 (16:08 +0200)]
KVM: SVM: Add tracepoint for nested vmrun

This patch adds a dedicated kvm tracepoint for a nested
vmrun.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Move INTR vmexit out of atomic code
Joerg Roedel [Fri, 9 Oct 2009 14:08:26 +0000 (16:08 +0200)]
KVM: SVM: Move INTR vmexit out of atomic code

The nested SVM code emulates a #vmexit caused by a request
to open the irq window right in the request function. This
is a bug because the request function runs with preemption
and interrupts disabled but the #vmexit emulation might
sleep. This can cause a schedule()-while-atomic bug and is
fixed with this patch.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: Notify nested hypervisor of lost event injections
Alexander Graf [Fri, 9 Oct 2009 14:08:25 +0000 (16:08 +0200)]
KVM: SVM: Notify nested hypervisor of lost event injections

If event_inj is valid on a #vmexit the host CPU would write
the contents to exit_int_info, so the hypervisor knows that
the event wasn't injected.

We don't do this in nested SVM by now which is a bug and
fixed by this patch.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: x86: include pvclock MSRs in msrs_to_save
Glauber Costa [Tue, 6 Oct 2009 17:24:50 +0000 (13:24 -0400)]
KVM: x86: include pvclock MSRs in msrs_to_save

For a while now, we are issuing a rdmsr instruction to find out which
msrs in our save list are really supported by the underlying machine.
However, it fails to account for kvm-specific msrs, such as the pvclock
ones.

This patch moves then to the beginning of the list, and skip testing them.

Cc: stable@kernel.org
Signed-off-by: Glauber Costa <glommer@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: x86: Rework guest single-step flag injection and filtering
Jan Kiszka [Mon, 5 Oct 2009 11:07:21 +0000 (13:07 +0200)]
KVM: x86: Rework guest single-step flag injection and filtering

Push TF and RF injection and filtering on guest single-stepping into the
vender get/set_rflags callbacks. This makes the whole mechanism more
robust wrt user space IOCTL order and instruction emulations.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86: disable paravirt mmu reporting
Marcelo Tosatti [Thu, 1 Oct 2009 22:28:39 +0000 (19:28 -0300)]
KVM: x86: disable paravirt mmu reporting

Disable paravirt MMU capability reporting, so that new (or rebooted)
guests switch to native operation.

Paravirt MMU is a burden to maintain and does not bring significant
advantages compared to shadow anymore.

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86: Refactor guest debug IOCTL handling
Jan Kiszka [Fri, 2 Oct 2009 22:31:21 +0000 (00:31 +0200)]
KVM: x86: Refactor guest debug IOCTL handling

Much of so far vendor-specific code for setting up guest debug can
actually be handled by the generic code. This also fixes a minor deficit
in the SVM part /wrt processing KVM_GUESTDBG_ENABLE.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: remove pre_task_link setting in save_state_to_tss16
Juan Quintela [Wed, 30 Sep 2009 15:39:07 +0000 (17:39 +0200)]
KVM: remove pre_task_link setting in save_state_to_tss16

Now, also remove pre_task_link setting in save_state_to_tss16.

  commit b237ac37a149e8b56436fabf093532483bff13b0
  Author: Gleb Natapov <gleb@redhat.com>
  Date:   Mon Mar 30 16:03:24 2009 +0300

    KVM: Fix task switch back link handling.

CC: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: Fix hotplug of CPUs
Zachary Amsden [Tue, 29 Sep 2009 21:38:37 +0000 (11:38 -1000)]
KVM: Fix hotplug of CPUs

Both VMX and SVM require per-cpu memory allocation, which is done at module
init time, for only online cpus.

Backend was not allocating enough structure for all possible CPUs, so
new CPUs coming online could not be hardware enabled.

Signed-off-by: Zachary Amsden <zamsden@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: Fix printk name error in svm.c
Zachary Amsden [Tue, 29 Sep 2009 21:38:36 +0000 (11:38 -1000)]
KVM: Fix printk name error in svm.c

Signed-off-by: Zachary Amsden <zamsden@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: Kill the confusing tsc_ref_khz and ref_freq variables
Zachary Amsden [Tue, 29 Sep 2009 21:38:35 +0000 (11:38 -1000)]
KVM: Kill the confusing tsc_ref_khz and ref_freq variables

They are globals, not clearly protected by any ordering or locking, and
vulnerable to various startup races.

Instead, for variable TSC machines, register the cpufreq notifier and get
the TSC frequency directly from the cpufreq machinery.  Not only is it
always right, it is also perfectly accurate, as no error prone measurement
is required.

On such machines, when a new CPU online is brought online, it isn't clear what
frequency it will start with, and it may not correspond to the reference, thus
in hardware_enable we clear the cpu_tsc_khz variable to zero and make sure
it is set before running on a VCPU.

Signed-off-by: Zachary Amsden <zamsden@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: Separate timer intialization into an indepedent function
Zachary Amsden [Tue, 29 Sep 2009 21:38:34 +0000 (11:38 -1000)]
KVM: Separate timer intialization into an indepedent function

Signed-off-by: Zachary Amsden <zamsden@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: fix lock imbalance in kvm_*_irq_source_id()
Jiri Slaby [Fri, 25 Sep 2009 07:33:38 +0000 (09:33 +0200)]
KVM: fix lock imbalance in kvm_*_irq_source_id()

Stanse found 2 lock imbalances in kvm_request_irq_source_id and
kvm_free_irq_source_id. They omit to unlock kvm->irq_lock on fail paths.

Fix that by adding unlock labels at the end of the functions and jump
there from the fail paths.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: SVM: Remove remaining occurences of rdtscll
Joerg Roedel [Wed, 16 Sep 2009 13:24:19 +0000 (15:24 +0200)]
KVM: SVM: Remove remaining occurences of rdtscll

This patch replaces them with native_read_tsc() which can
also be used in expressions and saves a variable on the
stack in this case.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: don't copy exit_int_info on nested vmrun
Joerg Roedel [Wed, 16 Sep 2009 13:24:16 +0000 (15:24 +0200)]
KVM: SVM: don't copy exit_int_info on nested vmrun

The exit_int_info field is only written by the hardware and
never read. So it does not need to be copied on a vmrun
emulation.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: SVM: reorganize svm_interrupt_allowed
Joerg Roedel [Wed, 16 Sep 2009 13:24:15 +0000 (15:24 +0200)]
KVM: SVM: reorganize svm_interrupt_allowed

This patch reorganizes the logic in svm_interrupt_allowed to
make it better to read. This is important because the logic
is a lot more complicated with Nested SVM.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: remove duplicated #include
Huang Weiyi [Wed, 16 Sep 2009 13:09:39 +0000 (21:09 +0800)]
KVM: remove duplicated #include

Remove duplicated #include('s) in
  arch/x86/kvm/lapic.c

Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: Activate Virtualization On Demand
Alexander Graf [Tue, 15 Sep 2009 09:37:46 +0000 (11:37 +0200)]
KVM: Activate Virtualization On Demand

X86 CPUs need to have some magic happening to enable the virtualization
extensions on them. This magic can result in unpleasant results for
users, like blocking other VMMs from working (vmx) or using invalid TLB
entries (svm).

Currently KVM activates virtualization when the respective kernel module
is loaded. This blocks us from autoloading KVM modules without breaking
other VMMs.

To circumvent this problem at least a bit, this patch introduces on
demand activation of virtualization. This means, that instead
virtualization is enabled on creation of the first virtual machine
and disabled on destruction of the last one.

So using this, KVM can be easily autoloaded, while keeping other
hypervisors usable.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: SVM: remove needless mmap_sem acquision from nested_svm_map
Marcelo Tosatti [Tue, 8 Sep 2009 17:47:38 +0000 (14:47 -0300)]
KVM: SVM: remove needless mmap_sem acquision from nested_svm_map

nested_svm_map unnecessarily takes mmap_sem around gfn_to_page, since
gfn_to_page / get_user_pages are responsible for it.

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Acked-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: VMX: Enhance invalid guest state emulation
Mohammed Gamal [Tue, 1 Sep 2009 10:48:18 +0000 (12:48 +0200)]
KVM: VMX: Enhance invalid guest state emulation

- Change returned handle_invalid_guest_state() to return relevant exit codes
- Move triggering the emulation from vmx_vcpu_run() to vmx_handle_exit()
- Return to userspace instead of repeatedly trying to emulate instructions that have already failed

Signed-off-by: Mohammed Gamal <m.gamal005@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
12 years agoKVM: x86 emulator: Add pusha and popa instructions
Mohammed Gamal [Tue, 1 Sep 2009 13:28:11 +0000 (15:28 +0200)]
KVM: x86 emulator: Add pusha and popa instructions

This adds pusha and popa instructions (opcodes 0x60-0x61), this enables booting
MINIX with invalid guest state emulation on.

[marcelo: remove unused variable]

Signed-off-by: Mohammed Gamal <m.gamal005@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86 emulator: Add missing decoder flags for 'or' instructions
Mohammed Gamal [Fri, 28 Aug 2009 14:41:44 +0000 (16:41 +0200)]
KVM: x86 emulator: Add missing decoder flags for 'or' instructions

Add missing decoder flags for or instructions (0xc-0xd).

Signed-off-by: Mohammed Gamal <m.gamal005@gmail.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Move assigned device code to own file
Avi Kivity [Wed, 26 Aug 2009 11:57:50 +0000 (14:57 +0300)]
KVM: Move assigned device code to own file

Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Return -ENOTTY on unrecognized ioctls
Avi Kivity [Wed, 26 Aug 2009 11:57:07 +0000 (14:57 +0300)]
KVM: Return -ENOTTY on unrecognized ioctls

Not the incorrect -EINVAL.

Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Drop kvm->irq_lock lock from irq injection path
Gleb Natapov [Mon, 24 Aug 2009 08:54:26 +0000 (11:54 +0300)]
KVM: Drop kvm->irq_lock lock from irq injection path

The only thing it protects now is interrupt injection into lapic and
this can work lockless. Even now with kvm->irq_lock in place access
to lapic is not entirely serialized since vcpu access doesn't take
kvm->irq_lock.

Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Move IO APIC to its own lock
Gleb Natapov [Mon, 24 Aug 2009 08:54:25 +0000 (11:54 +0300)]
KVM: Move IO APIC to its own lock

The allows removal of irq_lock from the injection path.

Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Convert irq notifiers lists to RCU locking
Gleb Natapov [Mon, 24 Aug 2009 08:54:24 +0000 (11:54 +0300)]
KVM: Convert irq notifiers lists to RCU locking

Use RCU locking for mask/ack notifiers lists.

Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Move irq ack notifier list to arch independent code
Gleb Natapov [Mon, 24 Aug 2009 08:54:23 +0000 (11:54 +0300)]
KVM: Move irq ack notifier list to arch independent code

Mask irq notifier list is already there.

Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Move irq routing data structure to rcu locking
Gleb Natapov [Mon, 24 Aug 2009 08:54:22 +0000 (11:54 +0300)]
KVM: Move irq routing data structure to rcu locking

Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Maintain back mapping from irqchip/pin to gsi
Gleb Natapov [Mon, 24 Aug 2009 08:54:21 +0000 (11:54 +0300)]
KVM: Maintain back mapping from irqchip/pin to gsi

Maintain back mapping from irqchip/pin to gsi to speedup
interrupt acknowledgment notifications.

[avi: build fix on non-x86/ia64]

Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Change irq routing table to use gsi indexed array
Gleb Natapov [Mon, 24 Aug 2009 08:54:20 +0000 (11:54 +0300)]
KVM: Change irq routing table to use gsi indexed array

Use gsi indexed array instead of scanning all entries on each interrupt
injection.

Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Move irq sharing information to irqchip level
Gleb Natapov [Mon, 24 Aug 2009 08:54:19 +0000 (11:54 +0300)]
KVM: Move irq sharing information to irqchip level

This removes assumptions that max GSIs is smaller than number of pins.
Sharing is tracked on pin level not GSI level.

[avi: no PIC on ia64]

Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Call pic_clear_isr() on pic reset to reuse logic there
Gleb Natapov [Mon, 24 Aug 2009 08:54:18 +0000 (11:54 +0300)]
KVM: Call pic_clear_isr() on pic reset to reuse logic there

Also move call of ack notifiers after pic state change.

Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Don't pass kvm_run arguments
Avi Kivity [Mon, 24 Aug 2009 08:10:17 +0000 (11:10 +0300)]
KVM: Don't pass kvm_run arguments

They're just copies of vcpu->run, which is readily accessible.

Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86 emulator: Introduce No64 decode option
Mohammed Gamal [Sun, 23 Aug 2009 11:24:25 +0000 (14:24 +0300)]
KVM: x86 emulator: Introduce No64 decode option

Introduces a new decode option "No64", which is used for instructions that are
invalid in long mode.

Signed-off-by: Mohammed Gamal <m.gamal005@gmail.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: x86 emulator: Add 'push/pop sreg' instructions
Mohammed Gamal [Sun, 23 Aug 2009 11:24:24 +0000 (14:24 +0300)]
KVM: x86 emulator: Add 'push/pop sreg' instructions

[avi: avoid buffer overflow]

Signed-off-by: Mohammed Gamal <m.gamal005@gmail.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoKVM: Don't wrap schedule() with vcpu_put()/vcpu_load()
Avi Kivity [Sun, 23 Aug 2009 14:46:21 +0000 (17:46 +0300)]
KVM: Don't wrap schedule() with vcpu_put()/vcpu_load()

Preemption notifiers will do that for us automatically.

Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoMerge remote branch 'tip/x86/entry' into kvm-updates/2.6.33
Avi Kivity [Thu, 3 Dec 2009 07:30:06 +0000 (09:30 +0200)]
Merge remote branch 'tip/x86/entry' into kvm-updates/2.6.33

Signed-off-by: Avi Kivity <avi@redhat.com>
12 years agoLinux 2.6.32 v2.6.32
Linus Torvalds [Thu, 3 Dec 2009 03:51:21 +0000 (19:51 -0800)]
Linux 2.6.32

12 years agoVIDEO: Correct use of request_region/request_mem_region
Julia Lawall [Sun, 9 Aug 2009 09:42:32 +0000 (11:42 +0200)]
VIDEO: Correct use of request_region/request_mem_region

request_region should be used with release_region, not request_mem_region.

Geert Uytterhoeven pointed out that in the case of drivers/video/gbefb.c,
the problem is actually the other way around; request_mem_region should be
used instead of request_region.

The semantic patch that finds/fixes this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@r1@
expression start;
@@

request_region(start,...)

@b1@
expression r1.start;
@@

request_mem_region(start,...)

@depends on !b1@
expression r1.start;
expression E;
@@

- release_mem_region
+ release_region
  (start,E)
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoSPI: spi_txx9: Fix bit rate calculation
Atsushi Nemoto [Thu, 3 Sep 2009 13:59:01 +0000 (22:59 +0900)]
SPI: spi_txx9: Fix bit rate calculation

TXx9 SPI bit rate is calculated by:
        fBR = (spi-baseclk) / (n + 1)
Fix calculation of min_speed_hz, max_speed_hz and n.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sameo/mfd-2.6
Linus Torvalds [Wed, 2 Dec 2009 23:41:49 +0000 (15:41 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/sameo/mfd-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sameo/mfd-2.6:
  mfd: Correct WM831X_MAX_ISEL_VALUE

12 years agoInput: i8042 - add Dell Vostro 1320, 1520 and 1720 to the reset list
Anisse Astier [Tue, 1 Dec 2009 09:14:25 +0000 (01:14 -0800)]
Input: i8042 - add Dell Vostro 1320, 1520 and 1720 to the reset list

These laptops often leave i8042 in a wierd state resulting in non-
operational touchpad and keyboard.

Signed-off-by: Anisse Astier <anisse@astier.eu>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoMerge branch 'for-linus' of git://neil.brown.name/md
Linus Torvalds [Wed, 2 Dec 2009 23:40:37 +0000 (15:40 -0800)]
Merge branch 'for-linus' of git://neil.brown.name/md

* 'for-linus' of git://neil.brown.name/md:
  md: revert incorrect fix for read error handling in raid1.

12 years agoparam: don't complain about unused module parameters.
Rusty Russell [Tue, 1 Dec 2009 04:26:44 +0000 (14:56 +1030)]
param: don't complain about unused module parameters.

Jon confirms that recent modprobe will look in /proc/cmdline, so these
cmdline options can still be used.

See http://bugzilla.kernel.org/show_bug.cgi?id=14164

Reported-by: Adam Williamson <awilliam@redhat.com>
Cc: stable@kernel.org
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Wed, 2 Dec 2009 23:39:20 +0000 (15:39 -0800)]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  MIPS: RB532: Fix devices.c compilation.
  MIPS: Fix MIPS I build.

12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog
Linus Torvalds [Wed, 2 Dec 2009 23:38:49 +0000 (15:38 -0800)]
Merge git://git./linux/kernel/git/wim/linux-2.6-watchdog

* git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog:
  [PATCH] rc32434_wdt: fix compilation failure
  [WATCHDOG] rc32434_wdt.c: use resource_size()

12 years agomodules: don't export section names of empty sections via sysfs
Helge Deller [Wed, 2 Dec 2009 23:29:15 +0000 (00:29 +0100)]
modules: don't export section names of empty sections via sysfs

On the parisc architecture we face for each and every loaded kernel module
this kernel "badness warning":
  sysfs: cannot create duplicate filename '/module/ac97_bus/sections/.text'
  Badness at fs/sysfs/dir.c:487

Reason for that is, that on parisc all kernel modules do have multiple
.text sections due to the usage of the -ffunction-sections compiler flag
which is needed to reach all jump targets on this platform.

An objdump on such a kernel module gives:
Sections:
Idx Name          Size      VMA       LMA       File off  Algn
  0 .note.gnu.build-id 00000024  00000000  00000000  00000034  2**2
                  CONTENTS, ALLOC, LOAD, READONLY, DATA
  1 .text         00000000  00000000  00000000  00000058  2**0
                  CONTENTS, ALLOC, LOAD, READONLY, CODE
  2 .text.ac97_bus_match 0000001c  00000000  00000000  00000058  2**2
                  CONTENTS, ALLOC, LOAD, READONLY, CODE
  3 .text         00000000  00000000  00000000  000000d4  2**0
                  CONTENTS, ALLOC, LOAD, READONLY, CODE
...
Since the .text sections are empty (size of 0 bytes) and won't be
loaded by the kernel module loader anyway, I don't see a reason
why such sections need to be listed under
/sys/module/<module_name>/sections/<section_name> either.

The attached patch does solve this issue by not exporting section
names which are empty.

This fixes bugzilla http://bugzilla.kernel.org/show_bug.cgi?id=14703

Signed-off-by: Helge Deller <deller@gmx.de>
CC: rusty@rustcorp.com.au
CC: akpm@linux-foundation.org
CC: James.Bottomley@HansenPartnership.com
CC: roland@redhat.com
CC: dave@hiauly1.hia.nrc.ca
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lrg/voltage-2.6
Linus Torvalds [Wed, 2 Dec 2009 20:44:42 +0000 (12:44 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/lrg/voltage-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lrg/voltage-2.6:
  regulator: Initialise wm831x structure pointor for ISINK driver

12 years agoregulator: Initialise wm831x structure pointor for ISINK driver
Mark Brown [Mon, 30 Nov 2009 14:01:56 +0000 (14:01 +0000)]
regulator: Initialise wm831x structure pointor for ISINK driver

The version that made it into mainline missed the initialisation of the
chip handle.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
12 years agoMIPS: RB532: Fix devices.c compilation.
Florian Fainelli [Wed, 2 Dec 2009 12:07:01 +0000 (13:07 +0100)]
MIPS: RB532: Fix devices.c compilation.

We should now use dev_set_drvdata to set the driver driver_data field.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/747/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMIPS: Fix MIPS I build.
Ralf Baechle [Wed, 2 Dec 2009 11:33:03 +0000 (11:33 +0000)]
MIPS: Fix MIPS I build.

Broken by d63c63e889bbeeaa461a8addf1245f89f3ce4ece (lmo) rsp.
f1e39a4a616cd9981a9decfd5332fd07a01abb8b (kernel.org).

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/746/

12 years agoMerge branch 'fix' of git://git.kernel.org/pub/scm/linux/kernel/git/ycmiao/pxa-linux-2.6
Linus Torvalds [Wed, 2 Dec 2009 16:21:58 +0000 (08:21 -0800)]
Merge branch 'fix' of git://git./linux/kernel/git/ycmiao/pxa-linux-2.6

* 'fix' of git://git.kernel.org/pub/scm/linux/kernel/git/ycmiao/pxa-linux-2.6:
  [ARM] pxamci: call mmc_remove_host() before freeing resources

12 years ago[ARM] pxamci: call mmc_remove_host() before freeing resources
Daniel Mack [Tue, 1 Dec 2009 17:17:18 +0000 (18:17 +0100)]
[ARM] pxamci: call mmc_remove_host() before freeing resources

mmc_remove_host() will cause the mmc core to switch off the bus power by
eventually calling pxamci_set_ios(). This function uses the regulator or
the GPIO which have been freed already.

This causes the following Oops on module unload.

[   49.519649] Unable to handle kernel paging request at virtual address 30303a70
[   49.526878] pgd = c7084000
[   49.529563] [30303a70] *pgd=00000000
[   49.533136] Internal error: Oops: 5 [#1]
[   49.537025] last sysfs file: /sys/devices/platform/pxa27x-ohci/usb1/1-1/1-1:1.0/host0/target0:0:0/0:0:0:0/scsi_level
[   49.547471] Modules linked in: pxamci(-) eeti_ts
[   49.552061] CPU: 0    Not tainted  (2.6.32-rc8 #322)
[   49.557001] PC is at regulator_is_enabled+0x3c/0xbc
[   49.561846] LR is at regulator_is_enabled+0x30/0xbc
[   49.566691] pc : [<c01a2448>]    lr : [<c01a243c>]    psr: 60000013
[   49.566702] sp : c7083e70  ip : 30303a30  fp : 00000000
[   49.578093] r10: c705e200  r9 : c7082000  r8 : c705e2e0
[   49.583280] r7 : c7061340  r6 : c7061340  r5 : c7083e70  r4 : 00000000
[   49.589759] r3 : c04dc434  r2 : c04dc434  r1 : c03eecea  r0 : 00000047
[   49.596241] Flags: nZCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment user
[   49.603329] Control: 0000397f  Table: a7084018  DAC: 00000015
[   49.609031] Process rmmod (pid: 1101, stack limit = 0xc7082278)
[   49.614908] Stack: (0xc7083e70 to 0xc7084000)
[   49.619238] 3e60:                                     c7082000 c703c4f8 c705ea00 c04f4074
[   49.627366] 3e80: 00000000 c705e3a0 ffffffff c0247ddc c70361a0 00000000 c705e3a0 ffffffff
[   49.635499] 3ea0: c705e200 bf006400 c78c4f00 c705e200 c705e3a0 ffffffff c705e200 ffffffff
[   49.643633] 3ec0: c04d8ac8 c02476d0 ffffffff c0247c60 c705e200 c0248678 c705e200 c0249064
[   49.651765] 3ee0: ffffffff bf006204 c04d8ad0 c04d8ad0 c04d8ac8 bf007490 00000880 c00440c4
[   49.659898] 3f00: 0000b748 c01c5708 bf007490 c01c44c8 c04d8ac8 c04d8afc bf007490 c01c4570
[   49.668031] 3f20: bf007490 bf00750c c04f4258 c01c37a4 00000000 bf00750c c7083f44 c007b014
[   49.676162] 3f40: 4000d000 6d617870 08006963 00000001 00000000 c7085000 00000001 00000000
[   49.684287] 3f60: 4000d000 c7083f8c 00000001 bea01a54 00005401 c7ab1400 c00440c4 00082000
[   49.692420] 3f80: bf00750c 00000880 c7083f8c 00000000 4000cfa8 00000000 00000880 bea01cc8
[   49.700552] 3fa0: 00000081 c0043f40 00000000 00000880 bea01cc8 00000880 00000006 00000000
[   49.708677] 3fc0: 00000000 00000880 bea01cc8 00000081 00000097 0000cca4 0000b748 00000000
[   49.716802] 3fe0: 4001a4f0 bea01cc0 00018bf4 4001a4fc 20000010 bea01cc8 a063e021 a063e421
[   49.724958] [<c01a2448>] (regulator_is_enabled+0x3c/0xbc) from [<c0247ddc>] (mmc_regulator_set_ocr+0x14/0xd8)
[   49.734836] [<c0247ddc>] (mmc_regulator_set_ocr+0x14/0xd8) from [<bf006400>] (pxamci_set_ios+0xd8/0x17c [pxamci])
[   49.745044] [<bf006400>] (pxamci_set_ios+0xd8/0x17c [pxamci]) from [<c02476d0>] (mmc_power_off+0x50/0x58)
[   49.754555] [<c02476d0>] (mmc_power_off+0x50/0x58) from [<c0247c60>] (mmc_detach_bus+0x68/0xc4)
[   49.763207] [<c0247c60>] (mmc_detach_bus+0x68/0xc4) from [<c0248678>] (mmc_stop_host+0xd4/0x1bc)
[   49.771944] [<c0248678>] (mmc_stop_host+0xd4/0x1bc) from [<c0249064>] (mmc_remove_host+0xc/0x20)
[   49.780681] [<c0249064>] (mmc_remove_host+0xc/0x20) from [<bf006204>] (pxamci_remove+0xc8/0x174 [pxamci])
[   49.790211] [<bf006204>] (pxamci_remove+0xc8/0x174 [pxamci]) from [<c01c5708>] (platform_drv_remove+0x1c/0x24)
[   49.800164] [<c01c5708>] (platform_drv_remove+0x1c/0x24) from [<c01c44c8>] (__device_release_driver+0x7c/0xc4)
[   49.810110] [<c01c44c8>] (__device_release_driver+0x7c/0xc4) from [<c01c4570>] (driver_detach+0x60/0x8c)
[   49.819535] [<c01c4570>] (driver_detach+0x60/0x8c) from [<c01c37a4>] (bus_remove_driver+0x90/0xcc)
[   49.828452] [<c01c37a4>] (bus_remove_driver+0x90/0xcc) from [<c007b014>] (sys_delete_module+0x1d8/0x254)
[   49.837891] [<c007b014>] (sys_delete_module+0x1d8/0x254) from [<c0043f40>] (ret_fast_syscall+0x0/0x28)
[   49.847145] Code: eb06c53a e596c030 e1a0500d e59f106c (e59c0040)
[   49.853566] ---[ end trace b5fa66a00cea142f ]---

Signed-off-by: Daniel Mack <daniel@caiaq.de>
Reported-by: Sven Neumann <s.neumann@raumfeld.com>
Cc: Pierre Ossman <pierre@ossman.eu>
Cc: linux-mmc@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: stable@kernel.org
Signed-off-by: Eric Miao <eric.y.miao@gmail.com>
12 years ago[PATCH] rc32434_wdt: fix compilation failure
Florian Fainelli [Wed, 2 Dec 2009 12:21:23 +0000 (13:21 +0100)]
[PATCH] rc32434_wdt: fix compilation failure

This patch fixes the compilation failure of
rc32434 due to a bad module parameter description.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
12 years ago[WATCHDOG] rc32434_wdt.c: use resource_size()
H Hartley Sweeten [Wed, 25 Nov 2009 02:06:26 +0000 (21:06 -0500)]
[WATCHDOG] rc32434_wdt.c: use resource_size()

The size value passed to ioremap_nocache() is not correct.
Use resource_size() to get the correct value.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Phil Sutter <n0-1@freewrt.org>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
12 years agosysfs: fix SYSFS_DEPRECATED_V2 prompt
Randy Dunlap [Tue, 1 Dec 2009 21:17:50 +0000 (13:17 -0800)]
sysfs: fix SYSFS_DEPRECATED_V2 prompt

The SYSFS_DEPRECATED_V2 says "remove" older, deprecated features, but it
actually enables them, so correct this confusing, backwards text.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agortc-x1205: reset clock to sane state after power failure
Johannes Weiner [Tue, 1 Dec 2009 21:17:49 +0000 (13:17 -0800)]
rtc-x1205: reset clock to sane state after power failure

When detecting power failure, the probe function would reset the clock
time to defined state.

However, the clock's _date_ might still be bogus and a subsequent probe
fails when sanity-checking these values.

Change the power-failure fixup code to do a full setting of rtc_time,
including a valid date.

Signed-off-by: Johannes Weiner <jw@emlix.com>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Cc: Paul Gortmaker <p_gortmaker@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agortc-x1205: fix rtc_time to y2k register value conversion
Johannes Weiner [Tue, 1 Dec 2009 21:17:48 +0000 (13:17 -0800)]
rtc-x1205: fix rtc_time to y2k register value conversion

The possible CCR_Y2K register values are 19 or 20 and struct rtc_time's
tm_year is in years since 1900.

The function translating rtc_time to register values assumes tm_year to be
years since first christmas, though, and we end up storing 0 or 1 in the
CCR_Y2K register, which the hardware does not refuse to do.

A subsequent probing of the clock fails due to the invalid value range in
the register, though.

[ And if it didn't, reading the clock would yield a bogus year because
  the function translating registers to tm_year is assuming a register
  value of 19 or 20. ]

This fixes the conversion from years since 1900 in tm_year to the
corresponding CCR_Y2K value of 19 or 20.

Signed-off-by: Johannes Weiner <jw@emlix.com>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Cc: Paul Gortmaker <p_gortmaker@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoaoe: prevent cache aliases
Peter Horton [Tue, 1 Dec 2009 21:17:46 +0000 (13:17 -0800)]
aoe: prevent cache aliases

Prevent the AoE block driver from creating cache aliases of page cache
pages on machines with virtually indexed caches.

Building kernels on an AT91SAM9G20 board without this patch fails with
segmentation faults after a couple of passes.

Signed-off-by: Peter Horton <zero@colonel-panic.org>
Cc: "Ed L. Cashin" <ecashin@coraid.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agogpio: Langwell GPIO driver bugfixes
Alek Du [Tue, 1 Dec 2009 21:17:45 +0000 (13:17 -0800)]
gpio: Langwell GPIO driver bugfixes

- Remove wrong and unnecessary unmask operation

- Remove extra GEDR reading

This fixes the loss of interrupts which occurs when two or more pins are
triggered in close succession.

Signed-off-by: Alek Du <alek.du@intel.com>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agokbuild: stepping down as maintainer
Sam Ravnborg [Tue, 1 Dec 2009 21:17:44 +0000 (13:17 -0800)]
kbuild: stepping down as maintainer

It has been fun but the last year or more it has been a duty and a burden.
So I leave it open for others to take over.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Anibal Monsalve Salazar <anibal@debian.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Michal Marek <mmarek@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>