This patch makes icount warp use the new QEMU_CLOCK_VIRTUAL_RT clock.
This way, icount's QEMU_CLOCK_VIRTUAL will never count time during which
the virtual machine is stopped.
Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgaluk@ispras.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This patch introduces new QEMU_CLOCK_VIRTUAL_RT clock, which
should be used for icount warping. In the next patch, it
will be used to avoid a huge icount warp when a virtual
machine is stopped for a long time.
Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgaluk@ispras.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Separate accessing the instruction counter from the compensation for
speed and halting that are introduced by qemu_icount_bias. This
introduces new infrastructure used by the record/replay patches.
Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgaluk@ispras.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The goal is to sleep qemu whenever the guest clock
is in advance compared to the host clock (we use
the monotonic clocks). The amount of time to sleep
is calculated in the execution loop in cpu_exec.
At first, we tried to approximate at each for loop the real time elapsed
while searching for a TB (generating or retrieving from cache) and
executing it. We would then approximate the virtual time corresponding
to the number of virtual instructions executed. The difference between
these 2 values would allow us to know if the guest is in advance or delayed.
However, the function used for measuring the real time
(qemu_clock_get_ns(QEMU_CLOCK_REALTIME)) proved to be very expensive.
We had an added overhead of 13% of the total run time.
Therefore, we modified the algorithm and only take into account the
difference between the 2 clocks at the begining of the cpu_exec function.
During the for loop we try to reduce the advance of the guest only by
computing the virtual time elapsed and sleeping if necessary. The overhead
is thus reduced to 3%. Even though this method still has a noticeable
overhead, it no longer is a bottleneck in trying to achieve a better
guest frequency for which the guest clock is faster than the host one.
As for the the alignement of the 2 clocks, with the first algorithm
the guest clock was oscillating between -1 and 1ms compared to the host clock.
Using the second algorithm we notice that the guest is 5ms behind the host, which
is still acceptable for our use case.
The tests where conducted using fio and stress. The host machine in an i5 CPU at
3.10GHz running Debian Jessie (kernel 3.12). The guest machine is an arm versatile-pb
built with buildroot.
Currently, on our test machine, the lowest icount we can achieve that is suitable for
aligning the 2 clocks is 6. However, we observe that the IO tests (using fio) are
slower than the cpu tests (using stress).
Signed-off-by: Sebastian Tanase <sebastian.tanase@openwide.fr>
Tested-by: Camille Bégué <camille.begue@openwide.fr>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This adds cpu_icount_to_ns function which is needed for reverse execution.
It returns the time for a specific instruction.
Signed-off-by: KONRAD Frederic <fred.konrad@greensocs.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
These let a user anticipate the deadline of a timer, atomically with
other sites that call the function. This helps avoiding complicated
lock hierarchies.
Reviewed-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
After disabling the QemuClock, we should make sure that no QemuTimers
are still in flight. To implement that with light overhead, we resort
to QemuEvent. The caller of disabling will wait on QemuEvent of each
timerlist.
Note, qemu_clock_enable(foo,false) can _not_ be called from timer's cb.
Also, the callers of qemu_clock_enable() should be protected by the BQL.
Signed-off-by: Liu Ping Fan <pingfank@linux.vnet.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
QEMU_CLOCK_VIRTUAL may be read outside BQL. This will make its
foundation, i.e. cpu_clock_offset exposed to race condition.
Using private lock to protect it.
After this patch, reading QEMU_CLOCK_VIRTUAL is thread safe
unless use_icount is true, in which case the existing callers
still rely on the BQL.
Lock rule: private lock innermost, ie BQL->"this lock"
Signed-off-by: Liu Ping Fan <pingfank@linux.vnet.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Introduce QEMUTimerList->active_timers_lock to protect the linked list
of active timers. This allows qemu_timer_mod_ns() to be called from any
thread.
Note that vm_clock is not thread-safe and its use of
qemu_clock_has_timers() works fine today but is also not thread-safe.
The purpose of this patch is to eventually let device models set or
cancel timers from a vcpu thread without holding the global mutex.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Remove the legacy interface from include/qemu/timers.h.
Ensure struct QEMUClock is not exposed at all.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Now we have timerlistgroups implemented and main_loop_tlg, we
no longer need the concept of a default timer list associated
with each clock. Remove it and simplify initialisation of
clocks and timer lists.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Rearrange timer.h so it is in order by function type.
Make legacy functions call non-legacy functions rather than vice-versa.
Convert cpus.c to use new API.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Add utility functions qemu_clock_get_ms and qemu_clock_get_us
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Remove qemu_clock_deadline and qemu_timerlist_deadline now we are using
the ns functions throughout.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Remove alarm timers from qemu-timers.c now we use g_poll / ppoll
instead.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Add documentation for existing qemu timer calls. Add new format
calls of the format timer_XXX rather than qemu_XXX_timer
for consistency.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Notify all timerlists derived from vm_clock in icount warp
calculations.
When calculating timer delay based on vm_clock deadline, use
all timerlists.
For compatibility, maintain an apparent bug where when using
icount, if no vm_clock timer was set, qemu_clock_deadline
would return INT32_MAX and always set an icount clock expiry
about 2 seconds ahead.
NB: thread safety - when different timerlists sit on different
threads, this will need some locking.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Introduce new API for creating timers - timer_new and
_ns, _ms, _us derivatives.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
On qemu_mod_timer_ns, ensure qemu_notify or aio_notify is called to
end the appropriate poll(), irrespective of use_icount value.
On qemu_clock_enable, ensure qemu_notify or aio_notify is called for
all QEMUTimerLists attached to the QEMUClock.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Add a notify pointer to QEMUTimerList so it knows what to notify
on a timer change.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Add QEMUTimerListGroup and helper functions, to represent
a QEMUTimerList associated with each clock. Add a default
QEMUTimerListGroup representing the default timer lists
which are not associated with any other object (e.g.
an AioContext as added by future patches).
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
include/qemu/timer.h has no need to include main-loop.h and
doing so causes an issue for the next patch. Unfortunately
various files assume including timers.h will pull in main-loop.h.
Untangle this mess.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Split QEMUClock into QEMUClock and QEMUTimerList so that we can
have more than one QEMUTimerList associated with the same clock.
Introduce a main_loop_timerlist concept and make existing
qemu_clock_* calls that actually should operate on a QEMUTimerList
call the relevant QEMUTimerList implementations, using the clock's
default timerlist. This vastly reduces the invasiveness of this
change and means the API stays constant for existing users.
Introduce a list of QEMUTimerLists associated with each clock
so that reenabling the clock can cause all the notifiers
to be called. Note the code to do the notifications is added
in a later patch.
Switch QEMUClockType to an enum. Remove global variables vm_clock,
host_clock and rt_clock and add compatibility defines. Do not
fix qemu_next_alarm_deadline as it's going to be deleted.
Add qemu_clock_use_for_deadline to indicate whether a particular
clock should be used for deadline calculations. When use_icount
is true, vm_clock should not be used for deadline calculations
as it does not contain a nanosecond count. Instead, icount
timeouts come from the execution thread doing aio_notify or
qemu_notify as appropriate. This function is used in the next
patch.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Make qemu_run_timers and qemu_run_all_timers return progress
so that aio_poll etc. can determine whether a timer has been
run.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Add qemu_poll_ns which works like g_poll but takes a nanosecond
timeout.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Add utility functions to qemu-timer.c for nanosecond timing.
Add qemu_clock_deadline_ns to calculate deadlines to
nanosecond accuracy.
Add utility function qemu_soonest_timeout to calculate soonest deadline.
Add qemu_timeout_ns_to_ms to convert a timeout in nanoseconds back to
milliseconds for when ppoll is not used.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Rename qemu_new_clock to qemu_clock_new.
Expose clock types.
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Rename four functions in preparation for new API.
Rename qemu_timer_expired to timer_expired
Rename qemu_timer_expire_time_ns to timer_expire_time_ns
Rename qemu_timer_pending to timer_pending
Rename qemu_timer_expired_ns to timer_expired_ns
Signed-off-by: Alex Bligh <alex@alex.org.uk>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
sys/param.h was included to define __FreeBSD_version, but the conditional
using it was removed by commit d05ef16045
(Brad Smith, "Allow clock_gettime() monotonic clock to be utilized on more
OS's"), so the include is no longer needed here.
Signed-off-by: Ed Maste <emaste@freebsd.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1366906631-2680-1-git-send-email-emaste@freebsd.org
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Allow the clock_gettime() code using monotonic clock to be utilized on
more POSIX compliannt OS's. This started as a fix for OpenBSD which was
listed in one function as part of the previous hard coded list of OS's
for the functions to support but not in the other.
Signed-off-by: Brad Smith <brad@comstyle.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20130405003748.GH884@rox.home.comstyle.com
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Paolo Bonzini