What it is: vhost net is a character device that can be used to reduce
the number of system calls involved in virtio networking.
Existing virtio net code is used in the guest without modification.
There's similarity with vringfd, with some differences and reduced scope
- uses eventfd for signalling
- structures can be moved around in memory at any time (good for
migration, bug work-arounds in userspace)
- write logging is supported (good for migration)
- support memory table and not just an offset (needed for kvm)
common virtio related code has been put in a separate file vhost.c and
can be made into a separate module if/when more backends appear. I used
Rusty's lguest.c as the source for developing this part : this supplied
me with witty comments I wouldn't be able to write myself.
What it is not: vhost net is not a bus, and not a generic new system
call. No assumptions are made on how guest performs hypercalls.
Userspace hypervisors are supported as well as kvm.
How it works: Basically, we connect virtio frontend (configured by
userspace) to a backend. The backend could be a network device, or a tap
device. Backend is also configured by userspace, including vlan/mac
etc.
Status: This works for me, and I haven't see any crashes.
Compared to userspace, people reported improved latency (as I save up to
4 system calls per packet), as well as better bandwidth and CPU
utilization.
Features that I plan to look at in the future:
- mergeable buffers
- zero copy
- scalability tuning: figure out the best threading model to use
Note on RCU usage (this is also documented in vhost.h, near
private_pointer which is the value protected by this variant of RCU):
what is happening is that the rcu_dereference() is being used in a
workqueue item. The role of rcu_read_lock() is taken on by the start of
execution of the workqueue item, of rcu_read_unlock() by the end of
execution of the workqueue item, and of synchronize_rcu() by
flush_workqueue()/flush_work(). In the future we might need to apply
some gcc attribute or sparse annotation to the function passed to
INIT_WORK(). Paul's ack below is for this RCU usage.
(Includes fixes by Alan Cox <alan@linux.intel.com>,
David L Stevens <dlstevens@us.ibm.com>,
Chris Wright <chrisw@redhat.com>)
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Len Brown pointed out that allmodconfig is broken for
ia64 because of:
arch/ia64/kvm/vmm.c: In function 'vmm_spin_unlock':
arch/ia64/kvm/vmm.c:70: error: 'spinlock_t' has no member named 'raw_lock'
KVM has it's own spinlock routines. It should not depend on the base kernel
spinlock_t type (which changed when ia64 switched to ticket locks). Define
its own vmm_spinlock_t type.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
kbuild.h forces include of autoconf.h on the
commandline using -include - so we do not need to
include the file explicit.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Michal Marek <mmarek@suse.cz>
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>
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>
Remove kvm_cpu_has_interrupt() and kvm_arch_interrupt_allowed() from
interface between general code and arch code. kvm_arch_vcpu_runnable()
checks for interrupts instead.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
This changes bus accesses to use high-level kvm_io_bus_read/kvm_io_bus_write
functions. in_range now becomes unused so it is removed from device ops in
favor of read/write callbacks performing range checks internally.
This allows aliasing (mostly for in-kernel virtio), as well as better error
handling by making it possible to pass errors up to userspace.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Archs are free to use vcpu_id as they see fit. For x86 it is used as
vcpu's apic id. New ioctl is added to configure boot vcpu id that was
assumed to be 0 till now.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Protect irq injection/acking data structures with a separate irq_lock
mutex. This fixes the following deadlock:
CPU A CPU B
kvm_vm_ioctl_deassign_dev_irq()
mutex_lock(&kvm->lock); worker_thread()
-> kvm_deassign_irq() -> kvm_assigned_dev_interrupt_work_handler()
-> deassign_host_irq() mutex_lock(&kvm->lock);
-> cancel_work_sync() [blocked]
[gleb: fix ia64 path]
Reported-by: Alex Williamson <alex.williamson@hp.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Init the itc_offset for all possible vCPUs. The current code by
mistake ends up only initializing the offset on vCPU 0.
Spotted by Gleb Natapov.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by : Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Fix problem with double export of certain symbols from vsprintf.c
which we do not wish to export from the kvm-intel.ko module.
In addition, we do not have access to kallsyms_lookup() from the
module, so make sure to #undef CONFIG_KALLSYMS
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
KVM uses a function call IPI to cause the exit of a guest running on a
physical cpu. For virtual interrupt notification there is no need to
wait on IPI receival, or to execute any function.
This is exactly what the reschedule IPI does, without the overhead
of function IPI. So use it instead of smp_call_function_single in
kvm_vcpu_kick.
Also change the "guest_mode" variable to a bit in vcpu->requests, and
use that to collapse multiple IPI's that would be issued between the
first one and zeroing of guest mode.
This allows kvm_vcpu_kick to called with interrupts disabled.
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Restore local irq enabled state before calling kvm_arch_vcpu_init(),
which calls kmalloc(GFP_KERNEL).
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Preserve interrupt status around call to kvm_insert_vmm_mappin()
in kvm_vcpu_pre_transition().
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Make sure to restore the psr after calling kvm_insert_vmm_mapping()
which calls ia64_itr_entry() as it disables local interrupts and
kvm_vcpu_init() may sleep.
Avoids a warning from the lock debugging code.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by : Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
vti_vcpu_load() doesn't do anything, so lets get rid of it.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by : Xiantao Zhang<xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Reorder locking to avoid holding the slots_lock when entering
the guest.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by : Xiantao Zhang<xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Since kvm-intel modue can't be built-in, make kvm depend on
CONFIG_MODULES.
Signed-off-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Currently, the interrupt enable bit is cleared when in
the vmm. This patch sets the bit and the external interrupts can
be dealt with when in the vmm. This improves the I/O performance.
Signed-off-by: Yang Zhang <yang.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
kvm_vcpu_block() unhalts vpu on an interrupt/timer without checking
if interrupt window is actually opened.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Currently timer events are processed before entering guest mode. Move it
to main vcpu event loop since timer events should be processed even while
vcpu is halted. Timer may cause interrupt/nmi to be injected and only then
vcpu will be unhalted.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Copy in SN2 RTC based ITC emulation for fast exit. The two versions
have the same size, so a dropin is simpler than patching the branch
instruction to hit the SN2 version.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
On SN2 do not pass down the real ITC frequency, but rather patch the
values to match the SN2 RTC frequency.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Move all reading of special register 'AR_ITC' into two functions, one
in the kernel and one in the VMM module. When running on SN2, base the
result on the RTC rather the system ITC, as the ITC isn't
synchronized.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
On SN2, map in the SN2 RTC registers to the VMM module, needed for ITC
emulation.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
The new way does not require additional loop over vcpus to calculate
the one with lowest priority as one is chosen during delivery bitmap
construction.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Use kvm_apic_match_dest() in kvm_get_intr_delivery_bitmask() instead
of duplicating the same code. Use kvm_get_intr_delivery_bitmask() in
apic_send_ipi() to figure out ipi destination instead of reimplementing
the logic.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Get rid of ioapic_inj_irq() and ioapic_inj_nmi() functions.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Modify the arg of kvm_get_lowest_prio_vcpu().
Make it consistent with its declaration.
Signed-off-by: Yang Zhang <yang.zhang@intel.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Reorder locking as down_read() may return with local interrupts enabled,
which means we could go into vti_vcpu_run() with interrupts enabled.
This caused random crashes on the Altix as the timer interrupt tried
to read a memory mapped clock source, for which the TLB had not yet been
reinstated in the exit, before ipsr was retored.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
This patch removes the following warnings and related ones.
Plus some cosmetics.
arch/ia64/kernel/patch.c:112: warning: passing argument 1 of 'paravirt_fc' makes integer from pointer without a cast
arch/ia64/kernel/patch.c:135: warning: passing argument 1 of 'paravirt_fc' makes integer from pointer without a cast
arch/ia64/kernel/patch.c:166: warning: passing argument 1 of 'paravirt_fc' makes integer from pointer without a cast
arch/ia64/kernel/patch.c:202: warning: passing argument 1 of 'paravirt_fc' makes integer from pointer without a cast
arch/ia64/kernel/patch.c:220: warning: passing argument 1 of 'paravirt_fc' makes integer from pointer without a cast
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
Signed-off-by: Tony Luck <tony.luck@intel.com>
define paravirt_dv_serialize_data() and insert it to suppress
false positive warnings.
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Include the newly introduced msidef.h to solve the build issues.
Signed-off-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
IRQ injection status is either -1 (if there was no CPU found
that should except the interrupt because IRQ was masked or
ioapic was misconfigured or ...) or >= 0 in that case the
number indicates to how many CPUs interrupt was injected.
If the value is 0 it means that the interrupt was coalesced
and probably should be reinjected.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Account for number of online cpus and use that in loops iterating over
the list of vpus instead of scanning the full array unconditionally.
This patch is a building block to facilitate allowing to bump up
the size of MAX_VCPUS significantly.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Acked-by : Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
For windows 2008, it needs more pal calls to implement for booting.
In addition, also changes the name of set_{sal, pal}_call_result to
get_{sal,pal}_call_result for readability.
Signed-off-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Add the support to translate PAL Call's pointer args.
Signed-off-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
