genode/repos/base/src/lib/timeout/timer_connection.cc

162 lines
4.1 KiB
C++
Raw Normal View History

os/timer: interpolate time via timestamps Previously, the Genode::Timer::curr_time always used the Timer_session::elapsed_ms RPC as back end. Now, Genode::Timer reads this remote time only in a periodic fashion independently from the calls to Genode::Timer::curr_time. If now one calls Genode::Timer::curr_time, the function takes the last read remote time value and adapts it using the timestamp difference since the remote-time read. The conversion factor from timestamps to time is estimated on every remote-time read using the last read remote-time value and the timestamp difference since the last remote time read. This commit also re-works the timeout test. The test now has two stages. In the first stage, it tests fast polling of the Genode::Timer::curr_time. This stage checks the error between locally interpolated and timer-driver time as well as wether the locally interpolated time is monotone and sufficiently homogeneous. In the second stage several periodic and one-shot timeouts are scheduled at once. This stage checks if the timeouts trigger sufficiently precise. This commit adds the new Kernel::time syscall to base-hw. The syscall is solely used by the Genode::Timer on base-hw as substitute for the timestamp. This is because on ARM, the timestamp function uses the ARM performance counter that stops counting when the WFI (wait for interrupt) instruction is active. This instruction, however is used by the base-hw idle contexts that get active when no user thread needs to be scheduled. Thus, the ARM performance counter is not a good choice for time interpolation and we use the kernel internal time instead. With this commit, the timeout library becomes a basic library. That means that it is linked against the LDSO which then provides it to the program it serves. Furthermore, you can't use the timeout library anymore without the LDSO because through the kernel-dependent LDSO make-files we can achieve a kernel-dependent timeout implementation. This commit introduces a structured Duration type that shall successively replace the use of Microseconds, Milliseconds, and integer types for duration values. Open issues: * The timeout test fails on Raspberry PI because of precision errors in the first stage. However, this does not render the framework unusable in general on the RPI but merely is an issue when speaking of microseconds precision. * If we run on ARM with another Kernel than HW the timestamp speed may continuously vary from almost 0 up to CPU speed. The Timer, however, only uses interpolation if the timestamp speed remained stable (12.5% tolerance) for at least 3 observation periods. Currently, one period is 100ms, so its 300ms. As long as this is not the case, Timer_session::elapsed_ms is called instead. Anyway, it might happen that the CPU load was stable for some time so interpolation becomes active and now the timestamp speed drops. In the worst case, we would now have 100ms of slowed down time. The bad thing about it would be, that this also affects the timeout of the period. Thus, it might "freeze" the local time for more than 100ms. On the other hand, if the timestamp speed suddenly raises after some stable time, interpolated time can get too fast. This would shorten the period but nonetheless may result in drifting away into the far future. Now we would have the problem that we can't deliver the real time anymore until it has caught up because the output of Timer::curr_time shall be monotone. So, effectively local time might "freeze" again for more than 100ms. It would be a solution to not use the Trace::timestamp on ARM w/o HW but a function whose return value causes the Timer to never use interpolation because of its stability policy. Fixes #2400
2017-04-22 00:52:23 +02:00
/*
* \brief Connection to timer service and timeout scheduler
* \author Martin Stein
* \date 2016-11-04
*/
/*
* Copyright (C) 2016-2017 Genode Labs GmbH
*
* This file is part of the Genode OS framework, which is distributed
* under the terms of the GNU Affero General Public License version 3.
*/
/* Genode includes */
#include <timer_session/connection.h>
#include <base/internal/globals.h>
using namespace Genode;
using namespace Genode::Trace;
void Timer::Connection::_update_interpolation_quality(uint64_t min_factor,
uint64_t max_factor)
os/timer: interpolate time via timestamps Previously, the Genode::Timer::curr_time always used the Timer_session::elapsed_ms RPC as back end. Now, Genode::Timer reads this remote time only in a periodic fashion independently from the calls to Genode::Timer::curr_time. If now one calls Genode::Timer::curr_time, the function takes the last read remote time value and adapts it using the timestamp difference since the remote-time read. The conversion factor from timestamps to time is estimated on every remote-time read using the last read remote-time value and the timestamp difference since the last remote time read. This commit also re-works the timeout test. The test now has two stages. In the first stage, it tests fast polling of the Genode::Timer::curr_time. This stage checks the error between locally interpolated and timer-driver time as well as wether the locally interpolated time is monotone and sufficiently homogeneous. In the second stage several periodic and one-shot timeouts are scheduled at once. This stage checks if the timeouts trigger sufficiently precise. This commit adds the new Kernel::time syscall to base-hw. The syscall is solely used by the Genode::Timer on base-hw as substitute for the timestamp. This is because on ARM, the timestamp function uses the ARM performance counter that stops counting when the WFI (wait for interrupt) instruction is active. This instruction, however is used by the base-hw idle contexts that get active when no user thread needs to be scheduled. Thus, the ARM performance counter is not a good choice for time interpolation and we use the kernel internal time instead. With this commit, the timeout library becomes a basic library. That means that it is linked against the LDSO which then provides it to the program it serves. Furthermore, you can't use the timeout library anymore without the LDSO because through the kernel-dependent LDSO make-files we can achieve a kernel-dependent timeout implementation. This commit introduces a structured Duration type that shall successively replace the use of Microseconds, Milliseconds, and integer types for duration values. Open issues: * The timeout test fails on Raspberry PI because of precision errors in the first stage. However, this does not render the framework unusable in general on the RPI but merely is an issue when speaking of microseconds precision. * If we run on ARM with another Kernel than HW the timestamp speed may continuously vary from almost 0 up to CPU speed. The Timer, however, only uses interpolation if the timestamp speed remained stable (12.5% tolerance) for at least 3 observation periods. Currently, one period is 100ms, so its 300ms. As long as this is not the case, Timer_session::elapsed_ms is called instead. Anyway, it might happen that the CPU load was stable for some time so interpolation becomes active and now the timestamp speed drops. In the worst case, we would now have 100ms of slowed down time. The bad thing about it would be, that this also affects the timeout of the period. Thus, it might "freeze" the local time for more than 100ms. On the other hand, if the timestamp speed suddenly raises after some stable time, interpolated time can get too fast. This would shorten the period but nonetheless may result in drifting away into the far future. Now we would have the problem that we can't deliver the real time anymore until it has caught up because the output of Timer::curr_time shall be monotone. So, effectively local time might "freeze" again for more than 100ms. It would be a solution to not use the Trace::timestamp on ARM w/o HW but a function whose return value causes the Timer to never use interpolation because of its stability policy. Fixes #2400
2017-04-22 00:52:23 +02:00
{
/*
* If the factor gets adapted less than 12.5%, we raise the
* interpolation-quality value. Otherwise, we reset it to zero.
*/
if ((max_factor - min_factor) < (max_factor >> 3)) {
if (_interpolation_quality < MAX_INTERPOLATION_QUALITY) {
_interpolation_quality++; }
} else if (_interpolation_quality) {
_interpolation_quality = 0;
}
}
uint64_t Timer::Connection::_ts_to_us_ratio(Timestamp ts,
uint64_t us,
unsigned shift)
os/timer: interpolate time via timestamps Previously, the Genode::Timer::curr_time always used the Timer_session::elapsed_ms RPC as back end. Now, Genode::Timer reads this remote time only in a periodic fashion independently from the calls to Genode::Timer::curr_time. If now one calls Genode::Timer::curr_time, the function takes the last read remote time value and adapts it using the timestamp difference since the remote-time read. The conversion factor from timestamps to time is estimated on every remote-time read using the last read remote-time value and the timestamp difference since the last remote time read. This commit also re-works the timeout test. The test now has two stages. In the first stage, it tests fast polling of the Genode::Timer::curr_time. This stage checks the error between locally interpolated and timer-driver time as well as wether the locally interpolated time is monotone and sufficiently homogeneous. In the second stage several periodic and one-shot timeouts are scheduled at once. This stage checks if the timeouts trigger sufficiently precise. This commit adds the new Kernel::time syscall to base-hw. The syscall is solely used by the Genode::Timer on base-hw as substitute for the timestamp. This is because on ARM, the timestamp function uses the ARM performance counter that stops counting when the WFI (wait for interrupt) instruction is active. This instruction, however is used by the base-hw idle contexts that get active when no user thread needs to be scheduled. Thus, the ARM performance counter is not a good choice for time interpolation and we use the kernel internal time instead. With this commit, the timeout library becomes a basic library. That means that it is linked against the LDSO which then provides it to the program it serves. Furthermore, you can't use the timeout library anymore without the LDSO because through the kernel-dependent LDSO make-files we can achieve a kernel-dependent timeout implementation. This commit introduces a structured Duration type that shall successively replace the use of Microseconds, Milliseconds, and integer types for duration values. Open issues: * The timeout test fails on Raspberry PI because of precision errors in the first stage. However, this does not render the framework unusable in general on the RPI but merely is an issue when speaking of microseconds precision. * If we run on ARM with another Kernel than HW the timestamp speed may continuously vary from almost 0 up to CPU speed. The Timer, however, only uses interpolation if the timestamp speed remained stable (12.5% tolerance) for at least 3 observation periods. Currently, one period is 100ms, so its 300ms. As long as this is not the case, Timer_session::elapsed_ms is called instead. Anyway, it might happen that the CPU load was stable for some time so interpolation becomes active and now the timestamp speed drops. In the worst case, we would now have 100ms of slowed down time. The bad thing about it would be, that this also affects the timeout of the period. Thus, it might "freeze" the local time for more than 100ms. On the other hand, if the timestamp speed suddenly raises after some stable time, interpolated time can get too fast. This would shorten the period but nonetheless may result in drifting away into the far future. Now we would have the problem that we can't deliver the real time anymore until it has caught up because the output of Timer::curr_time shall be monotone. So, effectively local time might "freeze" again for more than 100ms. It would be a solution to not use the Trace::timestamp on ARM w/o HW but a function whose return value causes the Timer to never use interpolation because of its stability policy. Fixes #2400
2017-04-22 00:52:23 +02:00
{
/*
* If the timestamp difference is to big to do the following
* factor calculation without an overflow, scale both timestamp
* and time difference down equally. This should neither happen
* often nor have much effect on the resulting factor.
*/
Timestamp const max_ts = ~(Timestamp)0ULL >> shift;
while (ts > max_ts) {
os/timer: interpolate time via timestamps Previously, the Genode::Timer::curr_time always used the Timer_session::elapsed_ms RPC as back end. Now, Genode::Timer reads this remote time only in a periodic fashion independently from the calls to Genode::Timer::curr_time. If now one calls Genode::Timer::curr_time, the function takes the last read remote time value and adapts it using the timestamp difference since the remote-time read. The conversion factor from timestamps to time is estimated on every remote-time read using the last read remote-time value and the timestamp difference since the last remote time read. This commit also re-works the timeout test. The test now has two stages. In the first stage, it tests fast polling of the Genode::Timer::curr_time. This stage checks the error between locally interpolated and timer-driver time as well as wether the locally interpolated time is monotone and sufficiently homogeneous. In the second stage several periodic and one-shot timeouts are scheduled at once. This stage checks if the timeouts trigger sufficiently precise. This commit adds the new Kernel::time syscall to base-hw. The syscall is solely used by the Genode::Timer on base-hw as substitute for the timestamp. This is because on ARM, the timestamp function uses the ARM performance counter that stops counting when the WFI (wait for interrupt) instruction is active. This instruction, however is used by the base-hw idle contexts that get active when no user thread needs to be scheduled. Thus, the ARM performance counter is not a good choice for time interpolation and we use the kernel internal time instead. With this commit, the timeout library becomes a basic library. That means that it is linked against the LDSO which then provides it to the program it serves. Furthermore, you can't use the timeout library anymore without the LDSO because through the kernel-dependent LDSO make-files we can achieve a kernel-dependent timeout implementation. This commit introduces a structured Duration type that shall successively replace the use of Microseconds, Milliseconds, and integer types for duration values. Open issues: * The timeout test fails on Raspberry PI because of precision errors in the first stage. However, this does not render the framework unusable in general on the RPI but merely is an issue when speaking of microseconds precision. * If we run on ARM with another Kernel than HW the timestamp speed may continuously vary from almost 0 up to CPU speed. The Timer, however, only uses interpolation if the timestamp speed remained stable (12.5% tolerance) for at least 3 observation periods. Currently, one period is 100ms, so its 300ms. As long as this is not the case, Timer_session::elapsed_ms is called instead. Anyway, it might happen that the CPU load was stable for some time so interpolation becomes active and now the timestamp speed drops. In the worst case, we would now have 100ms of slowed down time. The bad thing about it would be, that this also affects the timeout of the period. Thus, it might "freeze" the local time for more than 100ms. On the other hand, if the timestamp speed suddenly raises after some stable time, interpolated time can get too fast. This would shorten the period but nonetheless may result in drifting away into the far future. Now we would have the problem that we can't deliver the real time anymore until it has caught up because the output of Timer::curr_time shall be monotone. So, effectively local time might "freeze" again for more than 100ms. It would be a solution to not use the Trace::timestamp on ARM w/o HW but a function whose return value causes the Timer to never use interpolation because of its stability policy. Fixes #2400
2017-04-22 00:52:23 +02:00
warning("timestamp value too big");
ts >>= 1;
us >>= 1;
}
if (!us) { us = 1; }
if (!ts) { ts = 1; }
/*
* To make the result more precise, we scale up the numerator of
* the calculation. This upscaling must be considered when using
* the result.
*/
Timestamp const result = (ts << shift) / us;
uint64_t const result_ul = (uint64_t)result;
if (result != result_ul) {
warning("Timestamp-to-time ratio too big");
return ~0UL;
}
return result_ul;
os/timer: interpolate time via timestamps Previously, the Genode::Timer::curr_time always used the Timer_session::elapsed_ms RPC as back end. Now, Genode::Timer reads this remote time only in a periodic fashion independently from the calls to Genode::Timer::curr_time. If now one calls Genode::Timer::curr_time, the function takes the last read remote time value and adapts it using the timestamp difference since the remote-time read. The conversion factor from timestamps to time is estimated on every remote-time read using the last read remote-time value and the timestamp difference since the last remote time read. This commit also re-works the timeout test. The test now has two stages. In the first stage, it tests fast polling of the Genode::Timer::curr_time. This stage checks the error between locally interpolated and timer-driver time as well as wether the locally interpolated time is monotone and sufficiently homogeneous. In the second stage several periodic and one-shot timeouts are scheduled at once. This stage checks if the timeouts trigger sufficiently precise. This commit adds the new Kernel::time syscall to base-hw. The syscall is solely used by the Genode::Timer on base-hw as substitute for the timestamp. This is because on ARM, the timestamp function uses the ARM performance counter that stops counting when the WFI (wait for interrupt) instruction is active. This instruction, however is used by the base-hw idle contexts that get active when no user thread needs to be scheduled. Thus, the ARM performance counter is not a good choice for time interpolation and we use the kernel internal time instead. With this commit, the timeout library becomes a basic library. That means that it is linked against the LDSO which then provides it to the program it serves. Furthermore, you can't use the timeout library anymore without the LDSO because through the kernel-dependent LDSO make-files we can achieve a kernel-dependent timeout implementation. This commit introduces a structured Duration type that shall successively replace the use of Microseconds, Milliseconds, and integer types for duration values. Open issues: * The timeout test fails on Raspberry PI because of precision errors in the first stage. However, this does not render the framework unusable in general on the RPI but merely is an issue when speaking of microseconds precision. * If we run on ARM with another Kernel than HW the timestamp speed may continuously vary from almost 0 up to CPU speed. The Timer, however, only uses interpolation if the timestamp speed remained stable (12.5% tolerance) for at least 3 observation periods. Currently, one period is 100ms, so its 300ms. As long as this is not the case, Timer_session::elapsed_ms is called instead. Anyway, it might happen that the CPU load was stable for some time so interpolation becomes active and now the timestamp speed drops. In the worst case, we would now have 100ms of slowed down time. The bad thing about it would be, that this also affects the timeout of the period. Thus, it might "freeze" the local time for more than 100ms. On the other hand, if the timestamp speed suddenly raises after some stable time, interpolated time can get too fast. This would shorten the period but nonetheless may result in drifting away into the far future. Now we would have the problem that we can't deliver the real time anymore until it has caught up because the output of Timer::curr_time shall be monotone. So, effectively local time might "freeze" again for more than 100ms. It would be a solution to not use the Trace::timestamp on ARM w/o HW but a function whose return value causes the Timer to never use interpolation because of its stability policy. Fixes #2400
2017-04-22 00:52:23 +02:00
}
Duration Timer::Connection::_update_interpolated_time(Duration &interpolated_time)
{
/*
* The new interpolated time value may be smaller than a
* previously interpolated time value (based on an older real time
* value and factor). In this case, we don't want the user time to
* jump back but to freeze at the higher value until the new
* interpolation has caught up.
*/
if (_interpolated_time.less_than(interpolated_time)) {
os/timer: interpolate time via timestamps Previously, the Genode::Timer::curr_time always used the Timer_session::elapsed_ms RPC as back end. Now, Genode::Timer reads this remote time only in a periodic fashion independently from the calls to Genode::Timer::curr_time. If now one calls Genode::Timer::curr_time, the function takes the last read remote time value and adapts it using the timestamp difference since the remote-time read. The conversion factor from timestamps to time is estimated on every remote-time read using the last read remote-time value and the timestamp difference since the last remote time read. This commit also re-works the timeout test. The test now has two stages. In the first stage, it tests fast polling of the Genode::Timer::curr_time. This stage checks the error between locally interpolated and timer-driver time as well as wether the locally interpolated time is monotone and sufficiently homogeneous. In the second stage several periodic and one-shot timeouts are scheduled at once. This stage checks if the timeouts trigger sufficiently precise. This commit adds the new Kernel::time syscall to base-hw. The syscall is solely used by the Genode::Timer on base-hw as substitute for the timestamp. This is because on ARM, the timestamp function uses the ARM performance counter that stops counting when the WFI (wait for interrupt) instruction is active. This instruction, however is used by the base-hw idle contexts that get active when no user thread needs to be scheduled. Thus, the ARM performance counter is not a good choice for time interpolation and we use the kernel internal time instead. With this commit, the timeout library becomes a basic library. That means that it is linked against the LDSO which then provides it to the program it serves. Furthermore, you can't use the timeout library anymore without the LDSO because through the kernel-dependent LDSO make-files we can achieve a kernel-dependent timeout implementation. This commit introduces a structured Duration type that shall successively replace the use of Microseconds, Milliseconds, and integer types for duration values. Open issues: * The timeout test fails on Raspberry PI because of precision errors in the first stage. However, this does not render the framework unusable in general on the RPI but merely is an issue when speaking of microseconds precision. * If we run on ARM with another Kernel than HW the timestamp speed may continuously vary from almost 0 up to CPU speed. The Timer, however, only uses interpolation if the timestamp speed remained stable (12.5% tolerance) for at least 3 observation periods. Currently, one period is 100ms, so its 300ms. As long as this is not the case, Timer_session::elapsed_ms is called instead. Anyway, it might happen that the CPU load was stable for some time so interpolation becomes active and now the timestamp speed drops. In the worst case, we would now have 100ms of slowed down time. The bad thing about it would be, that this also affects the timeout of the period. Thus, it might "freeze" the local time for more than 100ms. On the other hand, if the timestamp speed suddenly raises after some stable time, interpolated time can get too fast. This would shorten the period but nonetheless may result in drifting away into the far future. Now we would have the problem that we can't deliver the real time anymore until it has caught up because the output of Timer::curr_time shall be monotone. So, effectively local time might "freeze" again for more than 100ms. It would be a solution to not use the Trace::timestamp on ARM w/o HW but a function whose return value causes the Timer to never use interpolation because of its stability policy. Fixes #2400
2017-04-22 00:52:23 +02:00
_interpolated_time = interpolated_time; }
return _interpolated_time;
}
void Timer::Connection::_handle_timeout()
{
uint64_t const us = elapsed_us();
if (us - _us > REAL_TIME_UPDATE_PERIOD_US) {
os/timer: interpolate time via timestamps Previously, the Genode::Timer::curr_time always used the Timer_session::elapsed_ms RPC as back end. Now, Genode::Timer reads this remote time only in a periodic fashion independently from the calls to Genode::Timer::curr_time. If now one calls Genode::Timer::curr_time, the function takes the last read remote time value and adapts it using the timestamp difference since the remote-time read. The conversion factor from timestamps to time is estimated on every remote-time read using the last read remote-time value and the timestamp difference since the last remote time read. This commit also re-works the timeout test. The test now has two stages. In the first stage, it tests fast polling of the Genode::Timer::curr_time. This stage checks the error between locally interpolated and timer-driver time as well as wether the locally interpolated time is monotone and sufficiently homogeneous. In the second stage several periodic and one-shot timeouts are scheduled at once. This stage checks if the timeouts trigger sufficiently precise. This commit adds the new Kernel::time syscall to base-hw. The syscall is solely used by the Genode::Timer on base-hw as substitute for the timestamp. This is because on ARM, the timestamp function uses the ARM performance counter that stops counting when the WFI (wait for interrupt) instruction is active. This instruction, however is used by the base-hw idle contexts that get active when no user thread needs to be scheduled. Thus, the ARM performance counter is not a good choice for time interpolation and we use the kernel internal time instead. With this commit, the timeout library becomes a basic library. That means that it is linked against the LDSO which then provides it to the program it serves. Furthermore, you can't use the timeout library anymore without the LDSO because through the kernel-dependent LDSO make-files we can achieve a kernel-dependent timeout implementation. This commit introduces a structured Duration type that shall successively replace the use of Microseconds, Milliseconds, and integer types for duration values. Open issues: * The timeout test fails on Raspberry PI because of precision errors in the first stage. However, this does not render the framework unusable in general on the RPI but merely is an issue when speaking of microseconds precision. * If we run on ARM with another Kernel than HW the timestamp speed may continuously vary from almost 0 up to CPU speed. The Timer, however, only uses interpolation if the timestamp speed remained stable (12.5% tolerance) for at least 3 observation periods. Currently, one period is 100ms, so its 300ms. As long as this is not the case, Timer_session::elapsed_ms is called instead. Anyway, it might happen that the CPU load was stable for some time so interpolation becomes active and now the timestamp speed drops. In the worst case, we would now have 100ms of slowed down time. The bad thing about it would be, that this also affects the timeout of the period. Thus, it might "freeze" the local time for more than 100ms. On the other hand, if the timestamp speed suddenly raises after some stable time, interpolated time can get too fast. This would shorten the period but nonetheless may result in drifting away into the far future. Now we would have the problem that we can't deliver the real time anymore until it has caught up because the output of Timer::curr_time shall be monotone. So, effectively local time might "freeze" again for more than 100ms. It would be a solution to not use the Trace::timestamp on ARM w/o HW but a function whose return value causes the Timer to never use interpolation because of its stability policy. Fixes #2400
2017-04-22 00:52:23 +02:00
_update_real_time();
}
if (_handler) {
_handler->handle_timeout(curr_time());
os/timer: interpolate time via timestamps Previously, the Genode::Timer::curr_time always used the Timer_session::elapsed_ms RPC as back end. Now, Genode::Timer reads this remote time only in a periodic fashion independently from the calls to Genode::Timer::curr_time. If now one calls Genode::Timer::curr_time, the function takes the last read remote time value and adapts it using the timestamp difference since the remote-time read. The conversion factor from timestamps to time is estimated on every remote-time read using the last read remote-time value and the timestamp difference since the last remote time read. This commit also re-works the timeout test. The test now has two stages. In the first stage, it tests fast polling of the Genode::Timer::curr_time. This stage checks the error between locally interpolated and timer-driver time as well as wether the locally interpolated time is monotone and sufficiently homogeneous. In the second stage several periodic and one-shot timeouts are scheduled at once. This stage checks if the timeouts trigger sufficiently precise. This commit adds the new Kernel::time syscall to base-hw. The syscall is solely used by the Genode::Timer on base-hw as substitute for the timestamp. This is because on ARM, the timestamp function uses the ARM performance counter that stops counting when the WFI (wait for interrupt) instruction is active. This instruction, however is used by the base-hw idle contexts that get active when no user thread needs to be scheduled. Thus, the ARM performance counter is not a good choice for time interpolation and we use the kernel internal time instead. With this commit, the timeout library becomes a basic library. That means that it is linked against the LDSO which then provides it to the program it serves. Furthermore, you can't use the timeout library anymore without the LDSO because through the kernel-dependent LDSO make-files we can achieve a kernel-dependent timeout implementation. This commit introduces a structured Duration type that shall successively replace the use of Microseconds, Milliseconds, and integer types for duration values. Open issues: * The timeout test fails on Raspberry PI because of precision errors in the first stage. However, this does not render the framework unusable in general on the RPI but merely is an issue when speaking of microseconds precision. * If we run on ARM with another Kernel than HW the timestamp speed may continuously vary from almost 0 up to CPU speed. The Timer, however, only uses interpolation if the timestamp speed remained stable (12.5% tolerance) for at least 3 observation periods. Currently, one period is 100ms, so its 300ms. As long as this is not the case, Timer_session::elapsed_ms is called instead. Anyway, it might happen that the CPU load was stable for some time so interpolation becomes active and now the timestamp speed drops. In the worst case, we would now have 100ms of slowed down time. The bad thing about it would be, that this also affects the timeout of the period. Thus, it might "freeze" the local time for more than 100ms. On the other hand, if the timestamp speed suddenly raises after some stable time, interpolated time can get too fast. This would shorten the period but nonetheless may result in drifting away into the far future. Now we would have the problem that we can't deliver the real time anymore until it has caught up because the output of Timer::curr_time shall be monotone. So, effectively local time might "freeze" again for more than 100ms. It would be a solution to not use the Trace::timestamp on ARM w/o HW but a function whose return value causes the Timer to never use interpolation because of its stability policy. Fixes #2400
2017-04-22 00:52:23 +02:00
}
}
void Timer::Connection::schedule_timeout(Microseconds duration,
Timeout_handler &handler)
{
if (duration.value < MIN_TIMEOUT_US)
duration.value = MIN_TIMEOUT_US;
if (duration.value > REAL_TIME_UPDATE_PERIOD_US)
duration.value = REAL_TIME_UPDATE_PERIOD_US;
_handler = &handler;
trigger_once(duration.value);
}
void Timer::Connection::_enable_modern_mode()
{
if (_mode == MODERN) {
return;
}
_mode = MODERN;
_sigh(_signal_handler);
_scheduler._enable();
/* do initial calibration burst to make interpolation available earlier */
for (unsigned i = 0; i < NR_OF_INITIAL_CALIBRATIONS; i++) {
_update_real_time();
}
os/timer: interpolate time via timestamps Previously, the Genode::Timer::curr_time always used the Timer_session::elapsed_ms RPC as back end. Now, Genode::Timer reads this remote time only in a periodic fashion independently from the calls to Genode::Timer::curr_time. If now one calls Genode::Timer::curr_time, the function takes the last read remote time value and adapts it using the timestamp difference since the remote-time read. The conversion factor from timestamps to time is estimated on every remote-time read using the last read remote-time value and the timestamp difference since the last remote time read. This commit also re-works the timeout test. The test now has two stages. In the first stage, it tests fast polling of the Genode::Timer::curr_time. This stage checks the error between locally interpolated and timer-driver time as well as wether the locally interpolated time is monotone and sufficiently homogeneous. In the second stage several periodic and one-shot timeouts are scheduled at once. This stage checks if the timeouts trigger sufficiently precise. This commit adds the new Kernel::time syscall to base-hw. The syscall is solely used by the Genode::Timer on base-hw as substitute for the timestamp. This is because on ARM, the timestamp function uses the ARM performance counter that stops counting when the WFI (wait for interrupt) instruction is active. This instruction, however is used by the base-hw idle contexts that get active when no user thread needs to be scheduled. Thus, the ARM performance counter is not a good choice for time interpolation and we use the kernel internal time instead. With this commit, the timeout library becomes a basic library. That means that it is linked against the LDSO which then provides it to the program it serves. Furthermore, you can't use the timeout library anymore without the LDSO because through the kernel-dependent LDSO make-files we can achieve a kernel-dependent timeout implementation. This commit introduces a structured Duration type that shall successively replace the use of Microseconds, Milliseconds, and integer types for duration values. Open issues: * The timeout test fails on Raspberry PI because of precision errors in the first stage. However, this does not render the framework unusable in general on the RPI but merely is an issue when speaking of microseconds precision. * If we run on ARM with another Kernel than HW the timestamp speed may continuously vary from almost 0 up to CPU speed. The Timer, however, only uses interpolation if the timestamp speed remained stable (12.5% tolerance) for at least 3 observation periods. Currently, one period is 100ms, so its 300ms. As long as this is not the case, Timer_session::elapsed_ms is called instead. Anyway, it might happen that the CPU load was stable for some time so interpolation becomes active and now the timestamp speed drops. In the worst case, we would now have 100ms of slowed down time. The bad thing about it would be, that this also affects the timeout of the period. Thus, it might "freeze" the local time for more than 100ms. On the other hand, if the timestamp speed suddenly raises after some stable time, interpolated time can get too fast. This would shorten the period but nonetheless may result in drifting away into the far future. Now we would have the problem that we can't deliver the real time anymore until it has caught up because the output of Timer::curr_time shall be monotone. So, effectively local time might "freeze" again for more than 100ms. It would be a solution to not use the Trace::timestamp on ARM w/o HW but a function whose return value causes the Timer to never use interpolation because of its stability policy. Fixes #2400
2017-04-22 00:52:23 +02:00
}
Timer::Connection::Connection(Genode::Env &env, char const *label)
:
Genode::Connection<Session>(env, session(env.parent(),
"ram_quota=10K, cap_quota=%u, label=\"%s\"",
CAP_QUOTA, label)),
Session_client(cap()),
_signal_handler(env.ep(), *this, &Connection::_handle_timeout)
{
/* register default signal handler */
Session_client::sigh(_default_sigh_cap);
}
void Timer::Connection::_schedule_one_shot(Timeout &timeout, Microseconds duration)
{
_enable_modern_mode();
_scheduler._schedule_one_shot(timeout, duration);
};
void Timer::Connection::_schedule_periodic(Timeout &timeout, Microseconds duration)
{
_enable_modern_mode();
_scheduler._schedule_periodic(timeout, duration);
};
void Timer::Connection::_discard(Timeout &timeout)
{
_enable_modern_mode();
_scheduler._discard(timeout);
}