*nopage() in kvm_main.c should only store the type of mmap() fault if
the pointers are not NULL. This patch fixes the problem.
Signed-off-by: Nguyen Anh Quynh <aquynh@gmail.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
What guest drivers?
Cc: Jan Engelhardt <jengelh@computergmbh.de>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
A guest context switch to an uncached cr3 can require allocation of
shadow pages, but we only recycle shadow pages in kvm_mmu_page_fault().
Move shadow page recycling to mmu_topup_memory_caches(), which is called
from both the page fault handler and from guest cr3 reload.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When taking a cpu down, we need to hardware_disable() it.
Unfortunately, the CPU_DYING notifier is called with interrupts
disabled, which means we can't use smp_call_function_single().
Fortunately, the CPU_DYING notifier is always called on the dying cpu,
so we don't need to use the function at all and can simply call
hardware_disable() directly.
Tested-by: Paolo Ornati <ornati@fastwebnet.it>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
More fallout from the writeback fixes: debug register transfer
instructions do their own writeback and thus need to disable the general
writeback mechanism.
This fixes oopses and some guest failures on AMD machines (the Intel
variant decodes the instruction in hardware and thus does not need
emulation).
Cc: Alistair John Strachan <alistair@devzero.co.uk>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
0x0f 0x01 instructions (ie lgdt, lidt, smsw, lmsw and invlpg) does
not use writeback. This patch set no_wb=1 when emulating those
instructions.
This fixes a regression booting the FreeBSD kernel on AMD.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Testing the wrong bit caused kvm not to disable nx on the guest when it is
disabled on the host (an mmu optimization relies on the nx bits being the
same in the guest and host).
This allows Windows to boot when nx is disabled on te host (e.g. when
host pae is disabled).
Signed-off-by: Avi Kivity <avi@qumranet.com>
This reverts commit a3c870bdce. While it
does save useless updates, it (probably) defeats the fork detector, causing
a massive performance loss.
Signed-off-by: Avi Kivity <avi@qumranet.com>
We add the kvm to the vm_list before initializing the vcpu mutexes,
which can be mutex_trylock()'ed by decache_vcpus_on_cpu().
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Writes that are contiguous in virtual memory may not be contiguous in
physical memory; so split writes that straddle a page boundary.
Thanks to Aurelien for reporting the bug, patient testing, and a fix
to this very patch.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Avi Kivity <avi@qumranet.com>
__free_page() wants a struct page, not a virtual address.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The kvm mmu uses page->private on shadow page tables; so does slub, and
an oops result. Fix by allocating regular pages for shadows instead of
using slub.
Tested-by: S.Çağlar Onur <caglar@pardus.org.tr>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Allow real-mode emulation of rdmsr and wrmsr. This allows smp Windows to
boot, presumably for its sipi trampoline.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The memory slot management functions were oriented against vcpu 0, where
they should be kvm-wide. This causes hangs starting X on guest smp.
Fix by making the functions (and resultant tail in the mmu) non-vcpu-specific.
Unfortunately this reduces the efficiency of the mmu object cache a bit. We
may have to revisit this later.
Signed-off-by: Avi Kivity <avi@qumranet.com>
We need to distinguish between large page shadows which have the nx bit set
and those which don't. The problem shows up when booting a newer smp Linux
kernel, where the trampoline page (which is in real mode, which uses the
same shadow pages as large pages) is using the same mapping as a kernel data
page, which is mapped using nx, causing kvm to spin on that page.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Slab destructors were no longer supported after Christoph's
c59def9f22 change. They've been
BUGs for both slab and slub, and slob never supported them
either.
This rips out support for the dtor pointer from kmem_cache_create()
completely and fixes up every single callsite in the kernel (there were
about 224, not including the slab allocator definitions themselves,
or the documentation references).
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Currently, CONFIG_X86_CMPXCHG64 both enables boot-time checking of
the cmpxchg64b feature and enables compilation of the set_64bit() family.
Since the option is dependent on PAE, and since KVM depends on set_64bit(),
this effectively disables KVM on i386 nopae.
Simplify by removing the config option altogether: the boot check is made
dependent on CONFIG_X86_PAE directly, and the set_64bit() family is exposed
without constraints. It is up to users to check for the feature flag (KVM
does not as virtualiation extensions imply its existence).
Signed-off-by: Avi Kivity <avi@qumranet.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Only at the CPU_DYING stage can we be sure that no user process will
be scheduled onto the cpu and oops when trying to use virtualization
extensions.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The hotplug IPIs can be called from the cpu on which we are currently
running on, so use on_cpu(). Similarly, drop on_each_cpu() for the
suspend/resume callbacks, as we're in atomic context here and only one
cpu is up anyway.
Signed-off-by: Avi Kivity <avi@qumranet.com>
By keeping track of which cpus have virtualization enabled, we
prevent double-enable or double-disable during hotplug, which is a
very fatal oops.
Signed-off-by: Avi Kivity <avi@qumranet.com>
kvm uses a pseudo filesystem, kvmfs, to generate inodes, a job that the
new anonymous inodes source does much better.
Cc: Davide Libenzi <davidel@xmailserver.org>
Signed-off-by: Avi Kivity <avi@qumranet.com>
This patch adds an implementation to the svm is_disabled function to
detect reliably if the BIOS disabled the SVM feature in the CPU. This
fixes the issues with kernel panics when loading the kvm-amd module on
machines where SVM is available but disabled.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Protected mode code may have corrupted the real-mode tss, so re-initialize
it when switching to real mode.
Signed-off-by: Avi Kivity <avi@qumranet.com>
When writing to normal memory and the memory area is unchanged the write
can be safely skipped, avoiding the costly kvm_mmu_pte_write.
Signed-Off-By: Luca Tettamanti <kronos.it@gmail.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
When the old value and new one are the same the emulator skips the
write; this is undesirable when the destination is a MMIO area and the
write shall be performed regardless of the previous value. This
optimization breaks e.g. a Linux guest APIC compiled without
X86_GOOD_APIC.
Remove the check and perform the writeback stage in the emulation unless
it's explicitly disabled (currently push and some 2 bytes instructions
may disable the writeback).
Signed-Off-By: Luca Tettamanti <kronos.it@gmail.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
With kernel-injected interrupts, we need to check for interrupts on
lightweight exits too.
Signed-off-by: Gregory Haskins <ghaskins@novell.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
If the time stamp counter goes backwards, a guest delay loop can become
infinite. This can happen if a vcpu is migrated to another cpu, where
the counter has a lower value than the first cpu.
Since we're doing an IPI to the first cpu anyway, we can use that to pick
up the old tsc, and use that to calculate the adjustment we need to make
to the tsc offset.
Signed-off-by: Avi Kivity <avi@qumranet.com>
When a vcpu causes a shadow tlb entry to have reduced permissions, it
must also clear the tlb on remote vcpus. We do that by:
- setting a bit on the vcpu that requests a tlb flush before the next entry
- if the vcpu is currently executing, we send an ipi to make sure it
exits before we continue
Signed-off-by: Avi Kivity <avi@qumranet.com>
A vcpu can pin up to four mmu shadow pages, which means the freeing
loop will never terminate. Fix by first unpinning shadow pages on
all vcpus, then freeing shadow pages.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Switch guest paging context may require us to allocate memory, which
might fail. Instead of wiring up error paths everywhere, make context
switching lazy and actually do the switch before the next guest entry,
where we can return an error if allocation fails.
Signed-off-by: Avi Kivity <avi@qumranet.com>
This was once used to avoid accessing the guest pte when upgrading
the shadow pte from read-only to read-write. But usually we need
to set the guest pte dirty or accessed bits anyway, so this wasn't
really exploited.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Always set the accessed and dirty bit (since having them cleared causes
a read-modify-write cycle), always set the present bit, and copy the
nx bit from the guest.
Signed-off-by: Avi Kivity <avi@qumranet.com>
With guest smp, a second vcpu might see partial updates when the first
vcpu services a page fault. So delay all updates until we have figured
out what the pte should look like.
Note that on i386, this is still not completely atomic as a 64-bit write
will be split into two on a 32-bit machine.
Signed-off-by: Avi Kivity <avi@qumranet.com>
This prevents some work from being performed twice, and, more importantly,
reduces the number of places where we modify shadow ptes.
Signed-off-by: Avi Kivity <avi@qumranet.com>
We will need the accessed bit (in addition to the dirty bit) and
also write access (for setting the dirty bit) in a future patch.
Signed-off-by: Avi Kivity <avi@qumranet.com>
KVM compilation fails for some .configs. This fixes it.
Signed-off-by: Markus Rechberger <markus.rechberger@amd.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Make a "menuconfig" out of the Kconfig objects "menu, ..., endmenu",
so that the user can disable all the options in that menu at once
instead of having to disable each option separately.
Signed-off-by: Jan Engelhardt <jengelh@gmx.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Avi Kivity <avi@qumranet.com>
MSR_EFER.LME/LMA bits are automatically save/restored by VMX
hardware, KVM only needs to save NX/SCE bits at time of heavy
weight VM Exit. But clearing NX bits in host envirnment may
cause system hang if the host page table is using EXB bits,
thus we leave NX bits as it is. If Host NX=1 and guest NX=0, we
can do guest page table EXB bits check before inserting a shadow
pte (though no guest is expecting to see this kind of gp fault).
If host NX=0, we present guest no Execute-Disable feature to guest,
thus no host NX=0, guest NX=1 combination.
This patch reduces raw vmexit time by ~27%.
Me: fix compile warnings on i386.
Signed-off-by: Yaozu (Eddie) Dong <eddie.dong@intel.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
In a lightweight exit (where we exit and reenter the guest without
scheduling or exiting to userspace in between), we don't need various
msrs on the host, and avoiding shuffling them around reduces raw exit
time by 8%.
i386 compile fix by Daniel Hecken <dh@bahntechnik.de>.
Signed-off-by: Yaozu (Eddie) Dong <eddie.dong@intel.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Instructions with address size override prefix opcode 0x67
Cause the #SS fault with 0 error code in VM86 mode. Forward
them to the emulator.
Signed-Off-By: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
kunmap() expects a struct page, not a virtual address. Fixes an oops loading
kvm-intel.ko on i386 with CONFIG_HIGHMEM.
Thanks to Michael Ivanov <deruhu@peterstar.ru> for reporting.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The real mode tr needs to be set to a specific tss so that I/O
instructions can function. Divert the new tr values to the real
mode save area from where they will be restored on transition to
protected mode.
This fixes some crashes on reboot when the bios accesses an I/O
instruction.
Signed-off-by: Avi Kivity <avi@qumranet.com>
If we set an msr via an ioctl() instead of by handling a guest exit, we
have the host state loaded, so reloading the msrs would clobber host
state instead of guest state.
This fixes a host oops (and loss of a cpu) on a guest reboot.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Attempting to boot the default 'bsd' kernel of OpenBSD 4.1 i386 in a guest
fails early in the kernel init inside p3_get_bus_clock while trying to read
the IA32_EBL_CR_POWERON MSR. KVM logs an 'unhandled MSR' message and the
guest kernel faults.
This patch is sufficient to allow OpenBSD to boot, after which it seems to
run fine. I'm not sure if this is the correct solution for dealing with
this particular MSR, but it works for me.
Signed-off-by: Matthew Gregan <kinetik@flim.org>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Everyone owns a piece of the exception bitmap, but they happily write to
the entire thing like there's no tomorrow. Centralize handling in
update_exception_bitmap() and have everyone call that.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The lightweight vmexit path avoids saving and reloading certain host
state. However in certain cases lightweight vmexit handling can schedule()
which requires reloading the host state.
So we store the host state in the vcpu structure, and reloaded it if we
relinquish the vcpu.
Signed-off-by: Avi Kivity <avi@qumranet.com>
A typical demand page/copy on write pattern is:
- page fault on vaddr
- kvm propagates fault to guest
- guest handles fault, updates pte
- kvm traps write, clears shadow pte, resumes guest
- guest returns to userspace, re-faults on same vaddr
- kvm installs shadow pte, resumes guest
- guest continues
So, three vmexits for a single guest page fault. But if instead of clearing
the page table entry, we update to correspond to the value that the guest
has just written, we eliminate the third vmexit.
This patch does exactly that, reducing kbuild time by about 10%.
Signed-off-by: Avi Kivity <avi@qumranet.com>
When a guest writes to a page that has an mmu shadow, we have to clear
the shadow pte corresponding to the memory location touched by the guest.
Now, in nonpae mode, a single guest page may have two or four shadow
pages (because a nonpae page maps 4MB or 4GB, whereas the pae shadow maps
2MB or 1GB), so we when we look up the page we find up to three additional
aliases for the page. Since we _clear_ the shadow pte, it doesn't matter
except for a slight performance penalty, but if we want to _update_ the
shadow pte instead of clearing it, it is vital that we don't modify the
aliases.
Fortunately, exactly which page is needed (the "quadrant") is easily
computed, and is accessible in the shadow page header. All we need is
to ignore shadow pages from the wrong quadrants.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Instead of calling two functions and repeating expensive checks, call one
function and provide it with before/after information.
Signed-off-by: Avi Kivity <avi@qumranet.com>
i386 wants fs for accessing the pda even on a lightweight exit, so ensure
we can always restore it. This fixes a regression on i386 introduced by
the lightweight vmexit patch.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The kvm mmu tries to detects forks by looking for repeated writes to a
page table. If it sees a fork, it unshadows the page table so the page
table copying can proceed at native speed instead of being emulated.
However, the detector also triggered on simple demand paging access patterns:
a linear walk of memory would of course cause repeated writes to the same
pagetable page, causing it to unshadow prematurely.
Fix by resetting the fork detector if we detect a demand fault.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Many msrs and the like will only be used by the host if we schedule() or
return to userspace. Therefore, we avoid saving them if we handle the
exit within the kernel, and if a reschedule is not requested.
Based on a patch from Eddie Dong <eddie.dong@intel.com> with a couple of
fixes by me.
Signed-off-by: Yaozu(Eddie) Dong <eddie.dong@intel.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
This allows us to remove write protection earlier than otherwise. Should
some mad OS choose to use byte writes to update pagetables, it will suffer
a performance hit, but still work correctly.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The PC debug port is used for IO delay and does not require emulation.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
This patch enables IO bitmaps control on vmx and unmask the 0x80 port to
avoid VMEXITs caused by accessing port 0x80. 0x80 is used as delays (see
include/asm/io.h), and handling VMEXITs on its access is unnecessary but
slows things down. This patch improves kernel build test at around
3%~5%.
Because every VM uses the same io bitmap, it is shared between
all VMs rather than a per-VM data structure.
Signed-off-by: Qing He <qing.he@intel.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
The lazy fpu changes did not take into account that some vmexit handlers
can sleep. Move loading the guest state into the inner loop so that it
can be reloaded if necessary, and move loading the host state into
vmx_vcpu_put() so it can be performed whenever we relinquish the vcpu.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Fix following section mismatch warning in kvm-intel.o:
WARNING: o-i386/drivers/kvm/kvm-intel.o(.init.text+0xbd): Section mismatch: reference to .exit.text: (between 'hardware_setup' and 'vmx_disabled_by_bios')
The function free_kvm_area is used in the function alloc_kvm_area which
is marked __init.
The __exit area is discarded by some archs during link-time if a
module is built-in resulting in an oops.
Note: This warning is only seen by my local copy of modpost
but the change will soon hit upstream.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Avi Kivity <avi@qumranet.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
First thing mm.h does is including sched.h solely for can_do_mlock() inline
function which has "current" dereference inside. By dealing with can_do_mlock()
mm.h can be detached from sched.h which is good. See below, why.
This patch
a) removes unconditional inclusion of sched.h from mm.h
b) makes can_do_mlock() normal function in mm/mlock.c
c) exports can_do_mlock() to not break compilation
d) adds sched.h inclusions back to files that were getting it indirectly.
e) adds less bloated headers to some files (asm/signal.h, jiffies.h) that were
getting them indirectly
Net result is:
a) mm.h users would get less code to open, read, preprocess, parse, ... if
they don't need sched.h
b) sched.h stops being dependency for significant number of files:
on x86_64 allmodconfig touching sched.h results in recompile of 4083 files,
after patch it's only 3744 (-8.3%).
Cross-compile tested on
all arm defconfigs, all mips defconfigs, all powerpc defconfigs,
alpha alpha-up
arm
i386 i386-up i386-defconfig i386-allnoconfig
ia64 ia64-up
m68k
mips
parisc parisc-up
powerpc powerpc-up
s390 s390-up
sparc sparc-up
sparc64 sparc64-up
um-x86_64
x86_64 x86_64-up x86_64-defconfig x86_64-allnoconfig
as well as my two usual configs.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Refine some depends statements to limit their visibility to the
environments that are actually supported.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Since nonboot CPUs are now disabled after tasks and devices have been
frozen and the CPU hotplug infrastructure is used for this purpose, we need
special CPU hotplug notifications that will help the CPU-hotplug-aware
subsystems distinguish normal CPU hotplug events from CPU hotplug events
related to a system-wide suspend or resume operation in progress. This
patch introduces such notifications and causes them to be used during
suspend and resume transitions. It also changes all of the
CPU-hotplug-aware subsystems to take these notifications into consideration
(for now they are handled in the same way as the corresponding "normal"
ones).
[oleg@tv-sign.ru: cleanups]
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Gautham R Shenoy <ego@in.ibm.com>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It is illegal not to return from a pio or mmio request without completing
it, as mmio or pio is an atomic operation. Therefore, we can simplify
the userspace interface by avoiding the completion indication.
Signed-off-by: Avi Kivity <avi@qumranet.com>
When emulating an mmio read, we actually emulate twice: once to determine
the physical address of the mmio, and, after we've exited to userspace to
get the mmio value, we emulate again to place the value in the result
register and update any flags.
But we don't really need to enter the guest again for that, only to take
an immediate vmexit. So, if we detect that we're doing an mmio read,
emulate a single instruction before entering the guest again.
Signed-off-by: Avi Kivity <avi@qumranet.com>
We only have to save/restore MSR_GS_BASE on every VMEXIT. The rest can be
saved/restored when we leave the VCPU. Since we don't emulate the DEBUGCTL
MSRs and the guest cannot write to them, we don't have to worry about
saving/restoring them at all.
This shaves a whopping 40% off raw vmexit costs on AMD.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
It might have worked in this case since PT_PRESENT_MASK is 1, but let's
express this correctly.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Only save/restore the FPU host state when the guest is actually using the
FPU.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Set all of the host mask bits for CR0 so that we can maintain a proper
shadow of CR0. This exposes CR0.TS, paving the way for lazy fpu handling.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Avoid saving and restoring the guest fpu state on every exit. This
shaves ~100 cycles off the guest/host switch.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Make the exit statistics per-vcpu instead of global. This gives a 3.5%
boost when running one virtual machine per core on my two socket dual core
(4 cores total) machine.
Signed-off-by: Avi Kivity <avi@qumranet.com>
By checking if a reschedule is needed, we avoid dropping the vcpu.
[With changes by me, based on Anthony Liguori's observations]
Signed-off-by: Avi Kivity <avi@qumranet.com>
Intel hosts only support syscall/sysret in long more (and only if efer.sce
is enabled), so only reload the related MSR_K6_STAR if the guest will
actually be able to use it.
This reduces vmexit cost by about 500 cycles (6400 -> 5870) on my setup.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Some msrs are only used by x86_64 instructions, and are therefore
not needed when the guest is legacy mode. By not bothering to switch
them, we reduce vmexit latency by 2400 cycles (from about 8800) when
running a 32-bt guest on a 64-bit host.
Signed-off-by: Avi Kivity <avi@qumranet.com>
THe automatically switched msrs are never changed on the host (with
the exception of MSR_KERNEL_GS_BASE) and thus there is no need to save
them on every vm entry.
This reduces vmexit latency by ~400 cycles on i386 and by ~900 cycles (10%)
on x86_64.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Usually, guest page faults are detected by the kvm page fault handler,
which detects if they are shadow faults, mmio faults, pagetable faults,
or normal guest page faults.
However, in ceratin circumstances, we can detect a page fault much later.
One of these events is the following combination:
- A two memory operand instruction (e.g. movsb) is executed.
- The first operand is in mmio space (which is the fault reported to kvm)
- The second operand is in an ummaped address (e.g. a guest page fault)
The Windows 2000 installer does such an access, an promptly hangs. Fix
by adding the missing page fault injection on that path.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Some guests (Solaris) do not set up all four pdptrs, but leave some invalid.
kvm incorrectly treated these as valid page directories, pinning the
wrong pages and causing general confusion.
Fix by checking the valid bit of a pae pdpte. This closes sourceforge bug
1698922.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Solaris panics if it sees a cpu with no fpu, and it seems to rely on this
bit. Closes sourceforge bug 1698920.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The expression
sp - 6 < sp
where sp is a u16 is undefined in C since 'sp - 6' is promoted to int,
and signed overflow is undefined in C. gcc 4.2 actually warns about it.
Replace with a simpler test.
Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: Avi Kivity <avi@qumranet.com>
This patch enables the virtualization of the last branch record MSRs on
SVM if this feature is available in hardware. It also introduces a small
and simple check feature for specific SVM extensions.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
With this, we can specify that accesses to one physical memory range will
be remapped to another. This is useful for the vga window at 0xa0000 which
is used as a movable window into the (much larger) framebuffer.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Mapping a guest page to a host page is a common operation. Currently,
one has first to find the memory slot where the page belongs (gfn_to_memslot),
then locate the page itself (gfn_to_page()).
This is clumsy, and also won't work well with memory aliases. So simplify
gfn_to_page() not to require memory slot translation first, and instead do it
internally.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Functions that play around with the physical memory map
need a way to clear mappings to possibly nonexistent or
invalid memory. Both the mmu cache and the processor tlb
are cleared.
Signed-off-by: Dor Laor <dor.laor@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
On x86, bit operations operate on a string of bits that can reside in
multiple words. For example, 'btsl %eax, (blah)' will touch the word
at blah+4 if %eax is between 32 and 63.
The x86 emulator compensates for that by advancing the operand address
by (bit offset / BITS_PER_LONG) and truncating the bit offset to the
range (0..BITS_PER_LONG-1). This has a side effect of forcing the operand
size to 8 bytes on 64-bit hosts.
Now, a 32-bit guest goes and fork()s a process. It write protects a stack
page at 0xbffff000 using the 'btr' instruction, at offset 0xffc in the page
table, with bit offset 1 (for the write permission bit).
The emulator now forces the operand size to 8 bytes as previously described,
and an innocent page table update turns into a cross-page-boundary write,
which is assumed by the mmu code not to be a page table, so it doesn't
actually clear the corresponding shadow page table entry. The guest and
host permissions are out of sync and guest memory is corrupted soon
afterwards, leading to guest failure.
Fix by not using BITS_PER_LONG as the word size; instead use the actual
operand size, so we get a 32-bit write in that case.
Note we still have to teach the mmu to handle cross-page-boundary writes
to guest page table; but for now this allows Damn Small Linux 0.4 (2.4.20)
to boot.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Remove unused function
CC drivers/kvm/svm.o
drivers/kvm/svm.c:207: warning: ‘inject_db’ defined but not used
Signed-off-by: Michal Piotrowski <michal.k.k.piotrowski@gmail.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
When a vcpu is migrated from one cpu to another, its timestamp counter
may lose its monotonic property if the host has unsynced timestamp counters.
This can confuse the guest, sometimes to the point of refusing to boot.
As the rdtsc instruction is rather fast on AMD processors (7-10 cycles),
we can simply record the last host tsc when we drop the cpu, and adjust
the vcpu tsc offset when we detect that we've migrated to a different cpu.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The kvm mmu keeps a shadow page for hugepage pdes; if several such pdes map
the same physical address, they share the same shadow page. This is a fairly
common case (kernel mappings on i386 nonpae Linux, for example).
However, if the two pdes map the same memory but with different permissions, kvm
will happily use the cached shadow page. If the access through the more
permissive pde will occur after the access to the strict pde, an endless pagefault
loop will be generated and the guest will make no progress.
Fix by making the access permissions part of the cache lookup key.
The fix allows Xen pae to boot on kvm and run guest domains.
Thanks to Jeremy Fitzhardinge for reporting the bug and testing the fix.
Signed-off-by: Avi Kivity <avi@qumranet.com>
This patch forbids the guest to execute monitor/mwait instructions on
SVM. This is necessary because the guest can execute these instructions
if they are available even if the kvm cpuid doesn't report its
existence.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Some older (~2.6.7) kernels write MCG_STATUS register during kernel
boot (mce_clear_all() function, called from mce_init()). It's not
currently handled by kvm and will cause it to inject a GPF.
Following patch adds a "nop" handler for this.
Signed-off-by: Sergey Kiselev <sergey.kiselev@intel.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
As usual, we need to mangle segment registers when emulating real mode
as vm86 has specific constraints. We special case the reset segment base,
and set the "access rights" (or descriptor flags) to vm86 comaptible values.
This fixes reboot on vmx.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The SET_SREGS ioctl modifies both cr0.pe (real mode/protected mode) and
guest segment registers. Since segment handling is modified by the mode on
Intel procesors, update the segment registers after the mode switch has taken
place.
Signed-off-by: Avi Kivity <avi@qumranet.com>
set_cr0_no_modeswitch() was a hack to avoid corrupting segment registers.
As we now cache the protected mode values on entry to real mode, this
isn't an issue anymore, and it interferes with reboot (which usually _is_
a modeswitch).
Signed-off-by: Avi Kivity <avi@qumranet.com>
The reset state has cs.selector == 0xf000 and cs.base == 0xffff0000,
which aren't compatible with vm86 mode, which is used for real mode
virtualization.
When we create a vcpu, we set cs.base to 0xf0000, but if we get there by
way of a reset, the values are inconsistent and vmx refuses to enter
guest mode.
Workaround by detecting the state and munging it appropriately.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The initial, noncaching, version of the kvm mmu flushed the all nonglobal
shadow page table translations (much like a native tlb flush). The new
implementation flushes translations only when they change, rendering global
pte tracking superfluous.
This removes the unused tracking mechanism and storage space.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The current string pio interface communicates using guest virtual addresses,
relying on userspace to translate addresses and to check permissions. This
interface cannot fully support guest smp, as the check needs to take into
account two pages at one in case an unaligned string transfer straddles a
page boundary.
Change the interface not to communicate guest addresses at all; instead use
a buffer page (mmaped by userspace) and do transfers there. The kernel
manages the virtual to physical translation and can perform the checks
atomically by taking the appropriate locks.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Some ioctls ignore their arguments. By requiring them to be zero now,
we allow a nonzero value to have some special meaning in the future.
Signed-off-by: Avi Kivity <avi@qumranet.com>
This allows us to store offsets in the kernel/user kvm_run area, and be
sure that userspace has them mapped. As offsets can be outside the
kvm_run struct, userspace has no way of knowing how much to mmap.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Allow a special signal mask to be used while executing in guest mode. This
allows signals to be used to interrupt a vcpu without requiring signal
delivery to a userspace handler, which is quite expensive. Userspace still
receives -EINTR and can get the signal via sigwait().
Signed-off-by: Avi Kivity <avi@qumranet.com>
This is redundant, as we also return -EINTR from the ioctl, but it
allows us to examine the exit_reason field on resume without seeing
old data.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Currently, userspace is told about the nature of the last exit from the
guest using two fields, exit_type and exit_reason, where exit_type has
just two enumerations (and no need for more). So fold exit_type into
exit_reason, reducing the complexity of determining what really happened.
Signed-off-by: Avi Kivity <avi@qumranet.com>
KVM used to handle cpuid by letting userspace decide what values to
return to the guest. We now handle cpuid completely in the kernel. We
still let userspace decide which values the guest will see by having
userspace set up the value table beforehand (this is necessary to allow
management software to set the cpu features to the least common denominator,
so that live migration can work).
The motivation for the change is that kvm kernel code can be impacted by
cpuid features, for example the x86 emulator.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Currently when passing the a PIO emulation request to userspace, we
rely on userspace updating %rax (on 'in' instructions) and %rsi/%rdi/%rcx
(on string instructions). This (a) requires two extra ioctls for getting
and setting the registers and (b) is unfriendly to non-x86 archs, when
they get kvm ports.
So fix by doing the register fixups in the kernel and passing to userspace
only an abstract description of the PIO to be done.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Instead of passing a 'struct kvm_run' back and forth between the kernel and
userspace, allocate a page and allow the user to mmap() it. This reduces
needless copying and makes the interface expandable by providing lots of
free space.
Signed-off-by: Avi Kivity <avi@qumranet.com>
When auditing a 32-bit guest on a 64-bit host, sign extension of the page
table directory pointer table index caused bogus addresses to be shown on
audit errors.
Fix by declaring the index unsigned.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Instead of twiddling the rip registers directly, use the
skip_emulated_instruction() function to do that for us.
Signed-off-by: Dor Laor <dor.laor@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
The hypercall code mixes up the ->cache_regs() and ->decache_regs()
callbacks, resulting in guest register corruption.
Signed-off-by: Dor Laor <dor.laor@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Nonpae guest pdes are shadowed by two pae ptes, so we double the offset
twice: once to account for the pte size difference, and once because we
need to shadow pdes for a single guest pde.
But when writing to the upper guest pde we also need to truncate the
lower bits, otherwise the multiply shifts these bits into the pde index
and causes an access to the wrong shadow pde. If we're at the end of the
page (accessing the very last guest pde) we can even overflow into the
next host page and oops.
Signed-off-by: Avi Kivity <avi@qumranet.com>
failed VM entry on VMX might still change %fs or %gs, thus make sure
that KVM always reloads the segment selectors. This is crutial on both
x86 and x86_64: x86 has __KERNEL_PDA in %fs on which things like
'current' depends and x86_64 has 0 there and needs MSR_GS_BASE to work.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Intel virtualization extensions do not support virtualizing real mode. So
kvm uses virtualized vm86 mode to run real mode code. Unfortunately, this
virtualized vm86 mode does not support the so called "big real" mode, where
the segment selector and base do not agree with each other according to the
real mode rules (base == selector << 4).
To work around this, kvm checks whether a selector/base pair violates the
virtualized vm86 rules, and if so, forces it into conformance. On a
transition back to protected mode, if we see that the guest did not touch
a forced segment, we restore it back to the original protected mode value.
This pile of hacks breaks down if the gdt has changed in real mode, as it
can cause a segment selector to point to a system descriptor instead of a
normal data segment. In fact, this happens with the Windows bootloader
and the qemu acpi bios, where a protected mode memcpy routine issues an
innocent 'pop %es' and traps on an attempt to load a system descriptor.
"Fix" by checking if the to-be-restored selector points at a system segment,
and if so, coercing it into a normal data segment. The long term solution,
of course, is to abandon vm86 mode and use emulation for big real mode.
Signed-off-by: Avi Kivity <avi@qumranet.com>
PAGE_MASK is an unsigned long, so using it to mask physical addresses on
i386 (which are 64-bit wide) leads to truncation. This can result in
page->private of unrelated memory pages being modified, with disasterous
results.
Fix by not using PAGE_MASK for physical addresses; instead calculate
the correct value directly from PAGE_SIZE. Also fix a similar BUG_ON().
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Avi Kivity <avi@qumranet.com>
KVM shadow page tables are always in pae mode, regardless of the guest
setting. This means that a guest pde (mapping 4MB of memory) is mapped
to two shadow pdes (mapping 2MB each).
When the guest writes to a pte or pde, we intercept the write and emulate it.
We also remove any shadowed mappings corresponding to the write. Since the
mmu did not account for the doubling in the number of pdes, it removed the
wrong entry, resulting in a mismatch between shadow page tables and guest
page tables, followed shortly by guest memory corruption.
This patch fixes the problem by detecting the special case of writing to
a non-pae pde and adjusting the address and number of shadow pdes zapped
accordingly.
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Avi Kivity <avi@qumranet.com>
The vmx code currently treats the guest's sysenter support msrs as 32-bit
values, which breaks 32-bit compat mode userspace on 64-bit guests. Fix by
using the native word width of the machine.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Otherwise, the core module thinks the arch module is loaded, and won't
let you reload it after you've fixed the bug.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Use the standard magic.h for kvmfs.
Cc: Avi Kivity <avi@qumranet.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Avi Kivity <avi@qumranet.com>
A bogus 'return r' can cause an otherwise successful module load to fail.
This both denies users the use of kvm, and it also denies them the use of
their machine, as it leaves a filesystem registered with its callbacks
pointing into now-freed module memory.
Fix by returning a zero like a good module.
Thanks to Richard Lucassen <mailinglists@lucassen.org> (?) for reporting
the problem and for providing access to a machine which exhibited it.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Enabling dirty page logging is done using KVM_SET_MEMORY_REGION ioctl.
If the memory region already exists, we need to remove write accesses,
so writes will be caught, and dirty pages will be logged.
Signed-off-by: Uri Lublin <uril@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Since dirty_bitmap is an unsigned long array, the alignment and size need
to take that into account.
Signed-off-by: Uri Lublin <uril@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
A few places where we modify guest memory fail to call mark_page_dirty(),
causing live migration to fail. This adds the missing calls.
Signed-off-by: Uri Lublin <uril@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Allocate a distinct inode for every vcpu in a VM. This has the following
benefits:
- the filp cachelines are no longer bounced when f_count is incremented on
every ioctl()
- the API and internal code are distinctly clearer; for example, on the
KVM_GET_REGS ioctl, there is no need to copy the vcpu number from
userspace and then copy the registers back; the vcpu identity is derived
from the fd used to make the call
Right now the performance benefits are completely theoretical since (a) we
don't support more than one vcpu per VM and (b) virtualization hardware
inefficiencies completely everwhelm any cacheline bouncing effects. But
both of these will change, and we need to prepare the API today.
Signed-off-by: Avi Kivity <avi@qumranet.com>
This reflects the changed scope, from device-wide to single vm (previously
every device open created a virtual machine).
Signed-off-by: Avi Kivity <avi@qumranet.com>
This avoids having filp->f_op and the corresponding inode->i_fop different,
which is a little unorthodox.
The ioctl list is split into two: global kvm ioctls and per-vm ioctls. A new
ioctl, KVM_CREATE_VM, is used to create VMs and return the VM fd.
Signed-off-by: Avi Kivity <avi@qumranet.com>
The kvmfs inodes will represent virtual machines and vcpus, as necessary,
reducing cacheline bouncing due to inodes and filps being shared.
Signed-off-by: Avi Kivity <avi@qumranet.com>
This patch changes the SVM code to intercept SMIs and handle it
outside the guest.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
This adds a special MSR based hypercall API to KVM. This is to be
used by paravirtual kernels and virtual drivers.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Besides using an established api, this allows using kvm in older kernels.
Signed-off-by: Markus Rechberger <markus.rechberger@amd.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
The whole thing is rotten, but this allows vmx to boot with the guest reboot
fix.
Signed-off-by: Markus Rechberger <markus.rechberger@amd.com>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
We fail to mark a page dirty in three cases:
- setting the accessed bit in a pte
- setting the dirty bit in a pte
- emulating a write into a pagetable
This fix adds the missing cases.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Convert the PDA code to use %fs rather than %gs as the segment for
per-processor data. This is because some processors show a small but
measurable performance gain for reloading a NULL segment selector (as %fs
generally is in user-space) versus a non-NULL one (as %gs generally is).
On modern processors the difference is very small, perhaps undetectable.
Some old AMD "K6 3D+" processors are noticably slower when %fs is used
rather than %gs; I have no idea why this might be, but I think they're
sufficiently rare that it doesn't matter much.
This patch also fixes the math emulator, which had not been adjusted to
match the changed struct pt_regs.
[frederik.deweerdt@gmail.com: fixit with gdb]
[mingo@elte.hu: Fix KVM too]
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Ian Campbell <Ian.Campbell@XenSource.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Zachary Amsden <zach@vmware.com>
Cc: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Add the necessary callbacks to suspend and resume a host running kvm. This is
just a repeat of the cpu hotplug/unplug work.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
On hotplug, we execute the hardware extension enable sequence. On unplug, we
decache any vcpus that last ran on the exiting cpu, and execute the hardware
extension disable sequence.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Like the inline code it replaces, this function decaches the vmcs from the cpu
it last executed on. in addition:
- vcpu_clear() works if the last cpu is also the cpu we're running on
- it is faster on larger smps by virtue of using smp_call_function_single()
Includes fix from Ingo Molnar.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This will allow us to iterate over all vcpus and see which cpus they are
running on.
[akpm@osdl.org: use standard (ugly) initialisers]
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
vcpu_load() can return NULL and it sometimes does in failure paths (for
example when the userspace ABI version is too old) - causing a preemption
count underflow in the ->vcpu_free() later on. So check for NULL.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Or 32-bit userspace will get confused.
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We report the value of cr8 to userspace on an exit. Also let userspace change
cr8 when we re-enter the guest. The lets 64-bit guest code maintain the tpr
correctly.
Thanks for Yaniv Kamay for the idea.
Signed-off-by: Dor Laor <dor.laor@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
