2016-12-09 22:09:17 +00:00
|
|
|
#
|
|
|
|
|
# \brief Genode application binary interface (ABI)
|
|
|
|
|
# \author Norman Feske
|
|
|
|
|
# \date 2016-12-20
|
|
|
|
|
#
|
|
|
|
|
# This file contains the binary application interface (ABI) provided by
|
|
|
|
|
# Genode's dynamic linker. Each line contains the name of a symbol followed
|
|
|
|
|
# by its type (according to the encoding used by binutil's 'nm' tool). Data
|
|
|
|
|
# symbols are furher annotated by the size of their corresponding data object.
|
|
|
|
|
# The latter is only needed on ARM.
|
|
|
|
|
#
|
|
|
|
|
# On the ARM architecture, copy relocations are created for read-only data
|
|
|
|
|
# objects that are present in shared libraries. For each data object, the
|
|
|
|
|
# linker preserves a slot in the program's BSS according to the object size.
|
|
|
|
|
# At runtime, the dynamic linker copies the data from the shared library's
|
|
|
|
|
# read-only segment to these slots. The copy relocations for a given binary
|
|
|
|
|
# can be inspected via 'objdump -R'. The size of data symbols as present in a
|
|
|
|
|
# shared library (like 'ld-hw.lib.a') can be inspected via 'nm --format posix'.
|
|
|
|
|
# The data-object sizes as annotated here must always be at least as big as the
|
|
|
|
|
# corresponding data objects present in the dynamic linker.
|
|
|
|
|
#
|
|
|
|
|
# The original version of this file is based on the output of the
|
|
|
|
|
# 'tool/abi_symbols' tool with 'ld-<platform>.lib.so' used as argument.
|
|
|
|
|
# However, this tool was solely used as a starting point for the - now
|
|
|
|
|
# manually maintained - file.
|
|
|
|
|
#
|
|
|
|
|
# Note that not all symbols present in this list are provided by each variant
|
|
|
|
|
# of the dynamic linker. I.e., there are a few symbols that are specific for a
|
|
|
|
|
# particular kernel or the C++ ABI of a specific architecture.
|
|
|
|
|
#
|
|
|
|
|
# Please keep the file sorted via 'LC_COLLATE=C sort'.
|
|
|
|
|
#
|
|
|
|
|
|
|
|
|
|
#
|
2017-02-20 12:23:52 +00:00
|
|
|
# Copyright (C) 2016-2017 Genode Labs GmbH
|
2016-12-09 22:09:17 +00:00
|
|
|
#
|
|
|
|
|
# This file is part of the Genode OS framework, which is distributed
|
2017-02-20 12:23:52 +00:00
|
|
|
# under the terms of the GNU Affero General Public License version 3.
|
2016-12-09 22:09:17 +00:00
|
|
|
#
|
|
|
|
|
|
|
|
|
|
_Unwind_Complete T
|
|
|
|
|
_Unwind_DeleteException T
|
|
|
|
|
_Unwind_Resume T
|
|
|
|
|
_Z11genode_exiti T
|
|
|
|
|
_Z13genode_atexitPFvvE T
|
|
|
|
|
_Z16main_thread_utcbv T
|
|
|
|
|
_Z21genode___cxa_finalizePv T
|
|
|
|
|
_Z22__ldso_raise_exceptionv T
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZN5Timer10Connection16schedule_timeoutEN6Genode12MicrosecondsERNS1_11Time_source15Timeout_handlerE T
|
|
|
|
|
_ZN5Timer10Connection18_schedule_one_shotERN6Genode7TimeoutENS1_12MicrosecondsE T
|
|
|
|
|
_ZN5Timer10Connection18_schedule_periodicERN6Genode7TimeoutENS1_12MicrosecondsE T
|
|
|
|
|
_ZN5Timer10Connection8_discardERN6Genode7TimeoutE T
|
|
|
|
|
_ZN5Timer10Connection9curr_timeEv T
|
|
|
|
|
_ZN5Timer10ConnectionC1ERN6Genode3EnvEPKc T
|
|
|
|
|
_ZN5Timer10ConnectionC1Ev T
|
|
|
|
|
_ZN5Timer10ConnectionC2ERN6Genode3EnvEPKc T
|
|
|
|
|
_ZN5Timer10ConnectionC2Ev T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode10Entrypoint16_dispatch_signalERNS_6SignalE T
|
|
|
|
|
_ZN6Genode10Entrypoint16schedule_suspendEPFvvES2_ T
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZN6Genode10Entrypoint22Signal_proxy_component6signalEv T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode10Entrypoint25_process_incoming_signalsEv T
|
2017-04-03 08:45:51 +00:00
|
|
|
_ZN6Genode10Entrypoint31wait_and_dispatch_one_io_signalEv T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode10Entrypoint6manageERNS_22Signal_dispatcher_baseE T
|
|
|
|
|
_ZN6Genode10Entrypoint8dissolveERNS_22Signal_dispatcher_baseE T
|
|
|
|
|
_ZN6Genode10EntrypointC1ERNS_3EnvE T
|
|
|
|
|
_ZN6Genode10EntrypointC1ERNS_3EnvEmPKc T
|
|
|
|
|
_ZN6Genode10EntrypointC2ERNS_3EnvE T
|
|
|
|
|
_ZN6Genode10EntrypointC2ERNS_3EnvEmPKc T
|
|
|
|
|
_ZN6Genode10Ipc_serverC1Ev T
|
|
|
|
|
_ZN6Genode10Ipc_serverC2Ev T
|
|
|
|
|
_ZN6Genode10Ipc_serverD1Ev T
|
|
|
|
|
_ZN6Genode10Ipc_serverD2Ev T
|
|
|
|
|
_ZN6Genode11Sliced_heap4freeEPvm T
|
|
|
|
|
_ZN6Genode11Sliced_heap5allocEmPPv T
|
2017-05-07 21:49:43 +00:00
|
|
|
_ZN6Genode11Sliced_heapC1ERNS_13Ram_allocatorERNS_10Region_mapE T
|
|
|
|
|
_ZN6Genode11Sliced_heapC2ERNS_13Ram_allocatorERNS_10Region_mapE T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode11Sliced_heapD0Ev T
|
|
|
|
|
_ZN6Genode11Sliced_heapD1Ev T
|
|
|
|
|
_ZN6Genode11Sliced_heapD2Ev T
|
|
|
|
|
_ZN6Genode12Address_infoC1Em T
|
|
|
|
|
_ZN6Genode12Address_infoC2Em T
|
|
|
|
|
_ZN6Genode13Avl_node_base15_rotate_subtreeEPS0_bRNS0_6PolicyE T
|
|
|
|
|
_ZN6Genode13Avl_node_base18_rebalance_subtreeEPS0_RNS0_6PolicyE T
|
|
|
|
|
_ZN6Genode13Avl_node_base6_adoptEPS0_bRNS0_6PolicyE T
|
|
|
|
|
_ZN6Genode13Avl_node_base6insertEPS0_RNS0_6PolicyE T
|
|
|
|
|
_ZN6Genode13Avl_node_base6removeERNS0_6PolicyE T
|
|
|
|
|
_ZN6Genode13Avl_node_baseC1Ev T
|
|
|
|
|
_ZN6Genode13Avl_node_baseC2Ev T
|
|
|
|
|
_ZN6Genode13Registry_base10_processedENS0_4KeepERNS_4ListINS0_7ElementEEERS3_PS3_ T
|
|
|
|
|
_ZN6Genode13Registry_base7ElementC1ERS0_Pv T
|
|
|
|
|
_ZN6Genode13Registry_base7ElementC2ERS0_Pv T
|
|
|
|
|
_ZN6Genode13Registry_base7ElementD1Ev T
|
|
|
|
|
_ZN6Genode13Registry_base7ElementD2Ev T
|
|
|
|
|
_ZN6Genode13Registry_base7_insertERNS0_7ElementE T
|
|
|
|
|
_ZN6Genode13Registry_base7_removeERNS0_7ElementE T
|
|
|
|
|
_ZN6Genode13Registry_base9_for_eachERNS0_15Untyped_functorE T
|
|
|
|
|
_ZN6Genode13Session_state7destroyEv T
|
2017-02-14 16:38:09 +00:00
|
|
|
_ZN6Genode13Session_stateC1ERNS_7ServiceERNS_8Id_spaceINS_6Parent6ClientEEENS6_2IdERKNS_13Session_labelERKNS_6StringILm256EEERKNS_8AffinityE T
|
|
|
|
|
_ZN6Genode13Session_stateC2ERNS_7ServiceERNS_8Id_spaceINS_6Parent6ClientEEENS6_2IdERKNS_13Session_labelERKNS_6StringILm256EEERKNS_8AffinityE T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode13Shared_objectC1ERNS_3EnvERNS_9AllocatorEPKcNS0_4BindENS0_4KeepE T
|
|
|
|
|
_ZN6Genode13Shared_objectC2ERNS_3EnvERNS_9AllocatorEPKcNS0_4BindENS0_4KeepE T
|
|
|
|
|
_ZN6Genode13Shared_objectD1Ev T
|
|
|
|
|
_ZN6Genode13Shared_objectD2Ev T
|
|
|
|
|
_ZN6Genode13sleep_foreverEv T
|
|
|
|
|
_ZN6Genode14Capability_map6insertEmm T
|
|
|
|
|
_ZN6Genode14Parent_service15_env_deprecatedEv T
|
|
|
|
|
_ZN6Genode14Rpc_entrypoint13_free_rpc_capERNS_10Pd_sessionENS_17Native_capabilityE T
|
|
|
|
|
_ZN6Genode14Rpc_entrypoint14_alloc_rpc_capERNS_10Pd_sessionENS_17Native_capabilityEm T
|
|
|
|
|
_ZN6Genode14Rpc_entrypoint17_activation_entryEv T
|
|
|
|
|
_ZN6Genode14Rpc_entrypoint17reply_signal_infoENS_17Native_capabilityEmm T
|
|
|
|
|
_ZN6Genode14Rpc_entrypoint22_block_until_cap_validEv T
|
|
|
|
|
_ZN6Genode14Rpc_entrypoint5entryEv T
|
|
|
|
|
_ZN6Genode14Rpc_entrypoint7_manageEPNS_15Rpc_object_baseE T
|
|
|
|
|
_ZN6Genode14Rpc_entrypoint8activateEv T
|
|
|
|
|
_ZN6Genode14Rpc_entrypoint9_dissolveEPNS_15Rpc_object_baseE T
|
|
|
|
|
_ZN6Genode14Rpc_entrypointC1EPNS_10Pd_sessionEmPKcbNS_8Affinity8LocationE T
|
|
|
|
|
_ZN6Genode14Rpc_entrypointC2EPNS_10Pd_sessionEmPKcbNS_8Affinity8LocationE T
|
|
|
|
|
_ZN6Genode14Rpc_entrypointD0Ev T
|
|
|
|
|
_ZN6Genode14Rpc_entrypointD1Ev T
|
|
|
|
|
_ZN6Genode14Rpc_entrypointD2Ev T
|
|
|
|
|
_ZN6Genode14Signal_context6submitEj T
|
|
|
|
|
_ZN6Genode14Signal_contextD0Ev T
|
|
|
|
|
_ZN6Genode14Signal_contextD1Ev T
|
|
|
|
|
_ZN6Genode14Signal_contextD2Ev T
|
|
|
|
|
_ZN6Genode14Timeout_thread11alarm_timerEv T
|
|
|
|
|
_ZN6Genode14Timeout_thread5entryEv T
|
|
|
|
|
_ZN6Genode14cache_coherentEmm T
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZN6Genode14env_deprecatedEv T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode14ipc_reply_waitERKNS_17Native_capabilityENS_18Rpc_exception_codeERNS_11Msgbuf_baseES5_ T
|
|
|
|
|
_ZN6Genode15Alarm_scheduler12_setup_alarmERNS_5AlarmEmm T
|
|
|
|
|
_ZN6Genode15Alarm_scheduler13next_deadlineEPm T
|
|
|
|
|
_ZN6Genode15Alarm_scheduler17schedule_absoluteEPNS_5AlarmEm T
|
|
|
|
|
_ZN6Genode15Alarm_scheduler18_get_pending_alarmEv T
|
|
|
|
|
_ZN6Genode15Alarm_scheduler23_unsynchronized_dequeueEPNS_5AlarmE T
|
|
|
|
|
_ZN6Genode15Alarm_scheduler23_unsynchronized_enqueueEPNS_5AlarmE T
|
|
|
|
|
_ZN6Genode15Alarm_scheduler6handleEm T
|
|
|
|
|
_ZN6Genode15Alarm_scheduler7discardEPNS_5AlarmE T
|
|
|
|
|
_ZN6Genode15Alarm_scheduler8scheduleEPNS_5AlarmEm T
|
|
|
|
|
_ZN6Genode15Alarm_schedulerD1Ev T
|
|
|
|
|
_ZN6Genode15Alarm_schedulerD2Ev T
|
|
|
|
|
_ZN6Genode15Cancelable_lock4lockEv T
|
|
|
|
|
_ZN6Genode15Cancelable_lock6unlockEv T
|
|
|
|
|
_ZN6Genode15Cancelable_lock9Applicant7wake_upEv T
|
|
|
|
|
_ZN6Genode15Cancelable_lockC1ENS0_5StateE T
|
|
|
|
|
_ZN6Genode15Cancelable_lockC2ENS0_5StateE T
|
|
|
|
|
_ZN6Genode15Connection_baseC1Ev T
|
|
|
|
|
_ZN6Genode15Connection_baseC2Ev T
|
|
|
|
|
_ZN6Genode15Signal_receiver12local_submitENS_6Signal4DataE T
|
|
|
|
|
_ZN6Genode15Signal_receiver14pending_signalEv T
|
|
|
|
|
_ZN6Genode15Signal_receiver15wait_for_signalEv T
|
|
|
|
|
_ZN6Genode15Signal_receiver16block_for_signalEv T
|
|
|
|
|
_ZN6Genode15Signal_receiver6manageEPNS_14Signal_contextE T
|
|
|
|
|
_ZN6Genode15Signal_receiver7pendingEv T
|
|
|
|
|
_ZN6Genode15Signal_receiver8dissolveEPNS_14Signal_contextE T
|
|
|
|
|
_ZN6Genode15Signal_receiverC1Ev T
|
|
|
|
|
_ZN6Genode15Signal_receiverC2Ev T
|
|
|
|
|
_ZN6Genode15Signal_receiverD1Ev T
|
|
|
|
|
_ZN6Genode15Signal_receiverD2Ev T
|
|
|
|
|
_ZN6Genode16raw_write_stringEPKc T
|
|
|
|
|
_ZN6Genode17Native_capability4_decEv T
|
|
|
|
|
_ZN6Genode17Native_capability4_incEv T
|
|
|
|
|
_ZN6Genode17Native_capabilityC1Ev T
|
|
|
|
|
_ZN6Genode17Native_capabilityC2Ev T
|
|
|
|
|
_ZN6Genode17Region_map_client13fault_handlerENS_10CapabilityINS_14Signal_contextEEE T
|
|
|
|
|
_ZN6Genode17Region_map_client5stateEv T
|
|
|
|
|
_ZN6Genode17Region_map_client6attachENS_10CapabilityINS_9DataspaceEEEmlbNS_10Region_map10Local_addrEb T
|
|
|
|
|
_ZN6Genode17Region_map_client6detachENS_10Region_map10Local_addrE T
|
|
|
|
|
_ZN6Genode17Region_map_client9dataspaceEv T
|
|
|
|
|
_ZN6Genode17Region_map_clientC1ENS_10CapabilityINS_10Region_mapEEE T
|
|
|
|
|
_ZN6Genode17Region_map_clientC2ENS_10CapabilityINS_10Region_mapEEE T
|
|
|
|
|
_ZN6Genode17Rm_session_client6createEm T
|
|
|
|
|
_ZN6Genode17Rm_session_client7destroyENS_10CapabilityINS_10Region_mapEEE T
|
|
|
|
|
_ZN6Genode17Rm_session_clientC1ENS_10CapabilityINS_10Rm_sessionEEE T
|
|
|
|
|
_ZN6Genode17Rm_session_clientC2ENS_10CapabilityINS_10Rm_sessionEEE T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base10_add_blockEPNS0_5BlockEmmb T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base10alloc_addrEmm T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base12remove_rangeEmm T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base13alloc_alignedEmPPvimm T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base14_destroy_blockEPNS0_5BlockE T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base14any_block_addrEPm T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base15_cut_from_blockEPNS0_5BlockEmmS2_S2_ T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base20_find_any_used_blockEPNS0_5BlockE T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base21_alloc_block_metadataEv T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base26_alloc_two_blocks_metadataEPPNS0_5BlockES3_ T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base30_revert_allocations_and_rangesEv T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base4freeEPv T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base5Block13find_best_fitEmjmm T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base5Block15find_by_addressEmmb T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base5Block16avail_in_subtreeEv T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base5Block9recomputeEv T
|
|
|
|
|
_ZN6Genode18Allocator_avl_base9add_rangeEmm T
|
|
|
|
|
_ZN6Genode18Signal_transmitter6submitEj T
|
|
|
|
|
_ZN6Genode18Signal_transmitter7contextENS_10CapabilityINS_14Signal_contextEEE T
|
|
|
|
|
_ZN6Genode18Signal_transmitter7contextEv T
|
|
|
|
|
_ZN6Genode18Signal_transmitterC1ENS_10CapabilityINS_14Signal_contextEEE T
|
|
|
|
|
_ZN6Genode18Signal_transmitterC2ENS_10CapabilityINS_14Signal_contextEEE T
|
|
|
|
|
_ZN6Genode18server_socket_pairEv T
|
|
|
|
|
_ZN6Genode20env_session_id_spaceEv T
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZN6Genode23Alarm_timeout_scheduler14handle_timeoutENS_8DurationE T
|
|
|
|
|
_ZN6Genode23Alarm_timeout_scheduler18_schedule_one_shotERNS_7TimeoutENS_12MicrosecondsE T
|
|
|
|
|
_ZN6Genode23Alarm_timeout_scheduler18_schedule_periodicERNS_7TimeoutENS_12MicrosecondsE T
|
|
|
|
|
_ZN6Genode23Alarm_timeout_scheduler7_enableEv T
|
|
|
|
|
_ZN6Genode23Alarm_timeout_schedulerC1ERNS_11Time_sourceE T
|
|
|
|
|
_ZN6Genode23Alarm_timeout_schedulerC2ERNS_11Time_sourceE T
|
2017-05-11 17:39:46 +00:00
|
|
|
_ZN6Genode25env_stack_area_region_mapE B 8
|
|
|
|
|
_ZN6Genode28env_stack_area_ram_allocatorE B 8
|
Capability quota accounting and trading
This patch mirrors the accounting and trading scheme that Genode employs
for physical memory to the accounting of capability allocations.
Capability quotas must now be explicitly assigned to subsystems by
specifying a 'caps=<amount>' attribute to init's start nodes.
Analogously to RAM quotas, cap quotas can be traded between clients and
servers as part of the session protocol. The capability budget of each
component is maintained by the component's corresponding PD session at
core.
At the current stage, the accounting is applied to RPC capabilities,
signal-context capabilities, and dataspace capabilities. Capabilities
that are dynamically allocated via core's CPU and TRACE service are not
yet covered. Also, the capabilities allocated by resource multiplexers
outside of core (like nitpicker) must be accounted by the respective
servers, which is not covered yet.
If a component runs out of capabilities, core's PD service prints a
warning to the log. To observe the consumption of capabilities per
component in detail, the PD service is equipped with a diagnostic
mode, which can be enabled via the 'diag' attribute in the target
node of init's routing rules. E.g., the following route enables the
diagnostic mode for the PD session of the "timer" component:
<default-route>
<service name="PD" unscoped_label="timer">
<parent diag="yes"/>
</service>
...
</default-route>
For subsystems based on a sub-init instance, init can be configured
to report the capability-quota information of its subsystems by
adding the attribute 'child_caps="yes"' to init's '<report>'
config node. Init's own capability quota can be reported by adding
the attribute 'init_caps="yes"'.
Fixes #2398
2017-05-08 19:35:43 +00:00
|
|
|
_ZN6Genode29upgrade_pd_quota_non_blockingENS_9Ram_quotaENS_9Cap_quotaE T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode3Log3logEv T
|
|
|
|
|
_ZN6Genode3Log8_acquireENS0_4TypeE T
|
|
|
|
|
_ZN6Genode3Log8_releaseEv T
|
|
|
|
|
_ZN6Genode3Raw7_outputEv T
|
|
|
|
|
_ZN6Genode3Raw8_acquireEv T
|
|
|
|
|
_ZN6Genode3Raw8_releaseEv T
|
|
|
|
|
_ZN6Genode4Heap11quota_limitEm T
|
|
|
|
|
_ZN6Genode4Heap4freeEPvm T
|
|
|
|
|
_ZN6Genode4Heap5allocEmPPv T
|
2017-05-07 21:49:43 +00:00
|
|
|
_ZN6Genode4HeapC1EPNS_13Ram_allocatorEPNS_10Region_mapEmPvm T
|
|
|
|
|
_ZN6Genode4HeapC2EPNS_13Ram_allocatorEPNS_10Region_mapEmPvm T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode4HeapD0Ev T
|
|
|
|
|
_ZN6Genode4HeapD1Ev T
|
|
|
|
|
_ZN6Genode4HeapD2Ev T
|
|
|
|
|
_ZN6Genode4Slab13any_used_elemEv T
|
|
|
|
|
_ZN6Genode4Slab5Block11_slab_entryEi T
|
|
|
|
|
_ZN6Genode4Slab5Block14any_used_entryEv T
|
|
|
|
|
_ZN6Genode4Slab5Block5allocEv T
|
|
|
|
|
_ZN6Genode4Slab5Block9inc_availERNS0_5EntryE T
|
|
|
|
|
_ZN6Genode4Slab5_freeEPv T
|
|
|
|
|
_ZN6Genode4Slab5allocEmPPv T
|
|
|
|
|
_ZN6Genode4Slab9insert_sbEPv T
|
|
|
|
|
_ZN6Genode4SlabC1EmmPvPNS_9AllocatorE T
|
|
|
|
|
_ZN6Genode4SlabC2EmmPvPNS_9AllocatorE T
|
|
|
|
|
_ZN6Genode4SlabD0Ev T
|
|
|
|
|
_ZN6Genode4SlabD1Ev T
|
|
|
|
|
_ZN6Genode4SlabD2Ev T
|
|
|
|
|
_ZN6Genode5AlarmD0Ev T
|
|
|
|
|
_ZN6Genode5AlarmD1Ev T
|
|
|
|
|
_ZN6Genode5AlarmD2Ev T
|
|
|
|
|
_ZN6Genode5Child10yield_sighENS_10CapabilityINS_14Signal_contextEEE T
|
|
|
|
|
_ZN6Genode5Child11session_capENS_8Id_spaceINS_6Parent6ClientEE2IdE T
|
|
|
|
|
_ZN6Genode5Child12session_sighENS_10CapabilityINS_14Signal_contextEEE T
|
|
|
|
|
_ZN6Genode5Child13session_readyERNS_13Session_stateE T
|
|
|
|
|
_ZN6Genode5Child13yield_requestEv T
|
|
|
|
|
_ZN6Genode5Child14session_closedERNS_13Session_stateE T
|
|
|
|
|
_ZN6Genode5Child14yield_responseEv T
|
|
|
|
|
_ZN6Genode5Child16resource_requestERKNS_6StringILm160EEE T
|
|
|
|
|
_ZN6Genode5Child16session_responseENS_8Id_spaceINS_6Parent6ServerEE2IdENS2_16Session_responseE T
|
|
|
|
|
_ZN6Genode5Child19deliver_session_capENS_8Id_spaceINS_6Parent6ServerEE2IdENS_10CapabilityINS_7SessionEEE T
|
|
|
|
|
_ZN6Genode5Child19resource_avail_sighENS_10CapabilityINS_14Signal_contextEEE T
|
2017-02-19 20:40:52 +00:00
|
|
|
_ZN6Genode5Child21initiate_env_sessionsEv T
|
|
|
|
|
_ZN6Genode5Child24initiate_env_ram_sessionEv T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode5Child4exitEi T
|
|
|
|
|
_ZN6Genode5Child5closeENS_8Id_spaceINS_6Parent6ClientEE2IdE T
|
|
|
|
|
_ZN6Genode5Child5yieldERKNS_6StringILm160EEE T
|
|
|
|
|
_ZN6Genode5Child7sessionENS_8Id_spaceINS_6Parent6ClientEE2IdERKNS_13Rpc_in_bufferILm64EEERKNS6_ILm160EEERKNS_8AffinityE T
|
|
|
|
|
_ZN6Genode5Child7upgradeENS_8Id_spaceINS_6Parent6ClientEE2IdERKNS_13Rpc_in_bufferILm160EEE T
|
|
|
|
|
_ZN6Genode5Child8announceERKNS_13Rpc_in_bufferILm64EEE T
|
|
|
|
|
_ZN6Genode5ChildC1ERNS_10Region_mapERNS_14Rpc_entrypointERNS_12Child_policyE T
|
|
|
|
|
_ZN6Genode5ChildC2ERNS_10Region_mapERNS_14Rpc_entrypointERNS_12Child_policyE T
|
|
|
|
|
_ZN6Genode5ChildD0Ev T
|
|
|
|
|
_ZN6Genode5ChildD1Ev T
|
|
|
|
|
_ZN6Genode5ChildD2Ev T
|
|
|
|
|
_ZN6Genode5Stack4sizeEm T
|
|
|
|
|
_ZN6Genode5Trace6Logger17_evaluate_controlEv T
|
|
|
|
|
_ZN6Genode5Trace6Logger3logEPKcm T
|
|
|
|
|
_ZN6Genode5Trace6LoggerC1Ev T
|
|
|
|
|
_ZN6Genode5Trace6LoggerC2Ev T
|
|
|
|
|
_ZN6Genode5printERNS_6OutputEPKc T
|
|
|
|
|
_ZN6Genode5printERNS_6OutputEPKv T
|
|
|
|
|
_ZN6Genode5printERNS_6OutputEd T
|
|
|
|
|
_ZN6Genode5printERNS_6OutputEf T
|
|
|
|
|
_ZN6Genode5printERNS_6OutputEl T
|
|
|
|
|
_ZN6Genode5printERNS_6OutputEm T
|
|
|
|
|
_ZN6Genode5printERNS_6OutputEx T
|
|
|
|
|
_ZN6Genode5printERNS_6OutputEy T
|
|
|
|
|
_ZN6Genode6Config4sighENS_10CapabilityINS_14Signal_contextEEE T
|
|
|
|
|
_ZN6Genode6Config6reloadEv T
|
|
|
|
|
_ZN6Genode6Config8xml_nodeEv T
|
|
|
|
|
_ZN6Genode6ConfigC1Ev T
|
|
|
|
|
_ZN6Genode6ConfigC2Ev T
|
|
|
|
|
_ZN6Genode6Output10out_stringEPKcm T
|
|
|
|
|
_ZN6Genode6Parent8announceERKNS_13Rpc_in_bufferILm64EEENS_10CapabilityINS_4RootEEE T
|
|
|
|
|
_ZN6Genode6Signal19_dec_ref_and_unlockEv T
|
|
|
|
|
_ZN6Genode6Signal8_inc_refEv T
|
|
|
|
|
_ZN6Genode6SignalC1ENS0_4DataE T
|
|
|
|
|
_ZN6Genode6SignalC1ERKS0_ T
|
|
|
|
|
_ZN6Genode6SignalC2ENS0_4DataE T
|
|
|
|
|
_ZN6Genode6SignalC2ERKS0_ T
|
|
|
|
|
_ZN6Genode6SignalD1Ev T
|
|
|
|
|
_ZN6Genode6SignalD2Ev T
|
|
|
|
|
_ZN6Genode6SignalaSERKS0_ T
|
|
|
|
|
_ZN6Genode6Thread10stack_sizeEm T
|
|
|
|
|
_ZN6Genode6Thread13native_threadEv T
|
|
|
|
|
_ZN6Genode6Thread15cancel_blockingEv T
|
|
|
|
|
_ZN6Genode6Thread18stack_virtual_sizeEv T
|
|
|
|
|
_ZN6Genode6Thread20free_secondary_stackEPv T
|
|
|
|
|
_ZN6Genode6Thread21alloc_secondary_stackEPKcm T
|
|
|
|
|
_ZN6Genode6Thread23stack_area_virtual_baseEv T
|
|
|
|
|
_ZN6Genode6Thread23stack_area_virtual_sizeEv T
|
|
|
|
|
_ZN6Genode6Thread4joinEv T
|
|
|
|
|
_ZN6Genode6Thread4nameEPcm T
|
|
|
|
|
_ZN6Genode6Thread4utcbEv T
|
|
|
|
|
_ZN6Genode6Thread5startEv T
|
|
|
|
|
_ZN6Genode6Thread6myselfEv T
|
|
|
|
|
_ZN6Genode6Thread7_loggerEv T
|
|
|
|
|
_ZN6Genode6Thread7mystackEv T
|
|
|
|
|
_ZN6Genode6ThreadC1ERNS_3EnvERKNS_6StringILm32EEEm T
|
|
|
|
|
_ZN6Genode6ThreadC1ERNS_3EnvERKNS_6StringILm32EEEmNS_8Affinity8LocationENS_11Cpu_session6WeightERS9_ T
|
|
|
|
|
_ZN6Genode6ThreadC1EmPKcmNS0_4TypeENS_8Affinity8LocationE T
|
|
|
|
|
_ZN6Genode6ThreadC1EmPKcmNS0_4TypeEPNS_11Cpu_sessionENS_8Affinity8LocationE T
|
|
|
|
|
_ZN6Genode6ThreadC2ERNS_3EnvERKNS_6StringILm32EEEm T
|
|
|
|
|
_ZN6Genode6ThreadC2ERNS_3EnvERKNS_6StringILm32EEEmNS_8Affinity8LocationENS_11Cpu_session6WeightERS9_ T
|
|
|
|
|
_ZN6Genode6ThreadC2EmPKcmNS0_4TypeENS_8Affinity8LocationE T
|
|
|
|
|
_ZN6Genode6ThreadC2EmPKcmNS0_4TypeEPNS_11Cpu_sessionENS_8Affinity8LocationE T
|
|
|
|
|
_ZN6Genode6ThreadD0Ev T
|
|
|
|
|
_ZN6Genode6ThreadD1Ev T
|
|
|
|
|
_ZN6Genode6ThreadD2Ev T
|
|
|
|
|
_ZN6Genode6configEv T
|
|
|
|
|
_ZN6Genode6printfEPKcz T
|
|
|
|
|
_ZN6Genode7Console11_out_stringEPKc T
|
|
|
|
|
_ZN6Genode7Console6printfEPKcz T
|
|
|
|
|
_ZN6Genode7Console7vprintfEPKcP13__va_list_tag T
|
|
|
|
|
_ZN6Genode7Console7vprintfEPKcPc T
|
|
|
|
|
_ZN6Genode7Console7vprintfEPKcPv T
|
|
|
|
|
_ZN6Genode7Console7vprintfEPKcSt9__va_list T
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZN6Genode7Timeout17schedule_one_shotENS_12MicrosecondsERNS0_7HandlerE T
|
|
|
|
|
_ZN6Genode7Timeout17schedule_periodicENS_12MicrosecondsERNS0_7HandlerE T
|
|
|
|
|
_ZN6Genode7Timeout5AlarmD0Ev W
|
|
|
|
|
_ZN6Genode7Timeout5AlarmD1Ev W
|
|
|
|
|
_ZN6Genode7Timeout5AlarmD2Ev W
|
|
|
|
|
_ZN6Genode7Timeout7discardEv T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode7cap_mapEv T
|
|
|
|
|
_ZN6Genode7vprintfEPKcP13__va_list_tag T
|
|
|
|
|
_ZN6Genode7vprintfEPKcPc T
|
|
|
|
|
_ZN6Genode7vprintfEPKcSt9__va_list T
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZN6Genode8DurationpLENS_12MicrosecondsE T
|
|
|
|
|
_ZN6Genode8DurationpLENS_12MillisecondsE T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZN6Genode8ipc_callENS_17Native_capabilityERNS_11Msgbuf_baseES2_m T
|
|
|
|
|
_ZN6Genode9ipc_replyENS_17Native_capabilityENS_18Rpc_exception_codeERNS_11Msgbuf_baseE T
|
2017-08-24 14:24:04 +00:00
|
|
|
_ZN9Component10stack_sizeEv T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZNK6Genode11Sliced_heap8overheadEm T
|
|
|
|
|
_ZNK6Genode13Session_state24generate_session_requestERNS_13Xml_generatorE T
|
2017-02-17 14:25:52 +00:00
|
|
|
_ZNK6Genode13Session_state25generate_client_side_infoERNS_13Xml_generatorENS0_6DetailE T
|
|
|
|
|
_ZNK6Genode13Session_state25generate_server_side_infoERNS_13Xml_generatorENS0_6DetailE T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZNK6Genode13Session_state5printERNS_6OutputE T
|
|
|
|
|
_ZNK6Genode13Shared_object7_lookupEPKc T
|
|
|
|
|
_ZNK6Genode13Shared_object8link_mapEv T
|
|
|
|
|
_ZNK6Genode14Rpc_entrypoint9is_myselfEv T
|
|
|
|
|
_ZNK6Genode17Native_capability10local_nameEv T
|
|
|
|
|
_ZNK6Genode17Native_capability3rawEv T
|
|
|
|
|
_ZNK6Genode17Native_capability5printERNS_6OutputE T
|
|
|
|
|
_ZNK6Genode17Native_capability5validEv T
|
|
|
|
|
_ZNK6Genode18Allocator_avl_base10valid_addrEm T
|
|
|
|
|
_ZNK6Genode18Allocator_avl_base5availEv T
|
|
|
|
|
_ZNK6Genode18Allocator_avl_base7size_atEPKv T
|
|
|
|
|
_ZNK6Genode3Hex5printERNS_6OutputE T
|
|
|
|
|
_ZNK6Genode4Slab8consumedEv T
|
|
|
|
|
_ZNK6Genode5Child15main_thread_capEv T
|
|
|
|
|
_ZNK6Genode5Child21notify_resource_availEv T
|
|
|
|
|
_ZNK6Genode6Thread10stack_baseEv T
|
|
|
|
|
_ZNK6Genode6Thread4nameEv T
|
|
|
|
|
_ZNK6Genode6Thread9stack_topEv T
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZNK6Genode8Duration17trunc_to_plain_usEv T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZNKSt13bad_exception4whatEv T
|
|
|
|
|
_ZNKSt9exception4whatEv T
|
|
|
|
|
_ZNSt13bad_exceptionD0Ev T
|
|
|
|
|
_ZNSt13bad_exceptionD1Ev T
|
|
|
|
|
_ZNSt13bad_exceptionD2Ev T
|
|
|
|
|
_ZNSt9exceptionD0Ev T
|
|
|
|
|
_ZNSt9exceptionD1Ev T
|
|
|
|
|
_ZNSt9exceptionD2Ev T
|
|
|
|
|
_ZSt18uncaught_exceptionv T
|
|
|
|
|
_ZSt9terminatev T
|
|
|
|
|
_ZTIDd D 8
|
|
|
|
|
_ZTIDe D 8
|
|
|
|
|
_ZTIDf D 8
|
|
|
|
|
_ZTIDi D 8
|
|
|
|
|
_ZTIDn D 8
|
|
|
|
|
_ZTIDs D 8
|
|
|
|
|
_ZTIN10__cxxabiv115__forced_unwindE D 8
|
|
|
|
|
_ZTIN10__cxxabiv116__enum_type_infoE D 12
|
|
|
|
|
_ZTIN10__cxxabiv117__class_type_infoE D 12
|
|
|
|
|
_ZTIN10__cxxabiv117__pbase_type_infoE D 12
|
|
|
|
|
_ZTIN10__cxxabiv119__foreign_exceptionE D 8
|
|
|
|
|
_ZTIN10__cxxabiv119__pointer_type_infoE D 12
|
|
|
|
|
_ZTIN10__cxxabiv120__function_type_infoE D 12
|
|
|
|
|
_ZTIN10__cxxabiv120__si_class_type_infoE D 12
|
|
|
|
|
_ZTIN10__cxxabiv121__vmi_class_type_infoE D 12
|
|
|
|
|
_ZTIN10__cxxabiv123__fundamental_type_infoE D 12
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZTIN5Timer10ConnectionE D 48
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZTIN6Genode11Sliced_heapE D 12
|
|
|
|
|
_ZTIN6Genode14Rpc_entrypointE D 2
|
|
|
|
|
_ZTIN6Genode14Rpc_entrypointE D 32
|
|
|
|
|
_ZTIN6Genode14Signal_contextE D 8
|
|
|
|
|
_ZTIN6Genode14Timeout_threadE D 2
|
|
|
|
|
_ZTIN6Genode14Timeout_threadE D 32
|
|
|
|
|
_ZTIN6Genode17Region_map_clientE D 12
|
|
|
|
|
_ZTIN6Genode17Rm_session_clientE D 12
|
|
|
|
|
_ZTIN6Genode18Allocator_avl_baseE D 12
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZTIN6Genode23Alarm_timeout_schedulerE D 80
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZTIN6Genode4HeapE D 12
|
|
|
|
|
_ZTIN6Genode4SlabE D 12
|
|
|
|
|
_ZTIN6Genode5AlarmE D 8
|
|
|
|
|
_ZTIN6Genode5Child14Initial_threadE D 12
|
|
|
|
|
_ZTIN6Genode5ChildE D 2
|
|
|
|
|
_ZTIN6Genode5ChildE D 40
|
|
|
|
|
_ZTIN6Genode6OutputE D 8
|
|
|
|
|
_ZTIN6Genode6ThreadE D 8
|
|
|
|
|
_ZTIN6Genode7ConsoleE D 8
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZTIN6Genode7Timeout5AlarmE D 24
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZTIPDd D 16
|
|
|
|
|
_ZTIPDe D 16
|
|
|
|
|
_ZTIPDf D 16
|
|
|
|
|
_ZTIPDi D 16
|
|
|
|
|
_ZTIPDn D 16
|
|
|
|
|
_ZTIPDs D 16
|
|
|
|
|
_ZTIPKDd D 16
|
|
|
|
|
_ZTIPKDe D 16
|
|
|
|
|
_ZTIPKDf D 16
|
|
|
|
|
_ZTIPKDi D 16
|
|
|
|
|
_ZTIPKDn D 16
|
|
|
|
|
_ZTIPKDs D 16
|
|
|
|
|
_ZTIPKa D 16
|
|
|
|
|
_ZTIPKb D 16
|
|
|
|
|
_ZTIPKc D 16
|
|
|
|
|
_ZTIPKd D 16
|
|
|
|
|
_ZTIPKe D 16
|
|
|
|
|
_ZTIPKf D 16
|
|
|
|
|
_ZTIPKh D 16
|
|
|
|
|
_ZTIPKi D 16
|
|
|
|
|
_ZTIPKj D 16
|
|
|
|
|
_ZTIPKl D 16
|
|
|
|
|
_ZTIPKm D 16
|
|
|
|
|
_ZTIPKs D 16
|
|
|
|
|
_ZTIPKt D 16
|
|
|
|
|
_ZTIPKx D 16
|
|
|
|
|
_ZTIPKy D 16
|
|
|
|
|
_ZTIPVKa D 16
|
|
|
|
|
_ZTIPVKb D 16
|
|
|
|
|
_ZTIPVKc D 16
|
|
|
|
|
_ZTIPVKd D 16
|
|
|
|
|
_ZTIPVKe D 16
|
|
|
|
|
_ZTIPVKf D 16
|
|
|
|
|
_ZTIPVKh D 16
|
|
|
|
|
_ZTIPVKi D 16
|
|
|
|
|
_ZTIPVKj D 16
|
|
|
|
|
_ZTIPVKl D 16
|
|
|
|
|
_ZTIPVKm D 16
|
|
|
|
|
_ZTIPVKs D 16
|
|
|
|
|
_ZTIPVKt D 16
|
|
|
|
|
_ZTIPVKx D 16
|
|
|
|
|
_ZTIPVKy D 16
|
|
|
|
|
_ZTIPVa D 16
|
|
|
|
|
_ZTIPVb D 16
|
|
|
|
|
_ZTIPVc D 16
|
|
|
|
|
_ZTIPVd D 16
|
|
|
|
|
_ZTIPVe D 16
|
|
|
|
|
_ZTIPVf D 16
|
|
|
|
|
_ZTIPVh D 16
|
|
|
|
|
_ZTIPVi D 16
|
|
|
|
|
_ZTIPVj D 16
|
|
|
|
|
_ZTIPVl D 16
|
|
|
|
|
_ZTIPVm D 16
|
|
|
|
|
_ZTIPVs D 16
|
|
|
|
|
_ZTIPVt D 16
|
|
|
|
|
_ZTIPVx D 16
|
|
|
|
|
_ZTIPVy D 16
|
|
|
|
|
_ZTIPa D 16
|
|
|
|
|
_ZTIPb D 16
|
|
|
|
|
_ZTIPc D 16
|
|
|
|
|
_ZTIPd D 16
|
|
|
|
|
_ZTIPe D 16
|
|
|
|
|
_ZTIPf D 16
|
|
|
|
|
_ZTIPh D 16
|
|
|
|
|
_ZTIPi D 16
|
|
|
|
|
_ZTIPj D 16
|
|
|
|
|
_ZTIPl D 16
|
|
|
|
|
_ZTIPm D 16
|
|
|
|
|
_ZTIPs D 16
|
|
|
|
|
_ZTIPt D 16
|
|
|
|
|
_ZTIPx D 16
|
|
|
|
|
_ZTIPy D 16
|
|
|
|
|
_ZTISt10bad_typeid D 12
|
|
|
|
|
_ZTISt13bad_exception D 12
|
|
|
|
|
_ZTISt16bad_array_length D 12
|
|
|
|
|
_ZTISt20bad_array_new_length D 12
|
|
|
|
|
_ZTISt8bad_cast D 12
|
|
|
|
|
_ZTISt9bad_alloc D 12
|
|
|
|
|
_ZTISt9exception D 8
|
|
|
|
|
_ZTISt9type_info D 8
|
|
|
|
|
_ZTIa D 8
|
|
|
|
|
_ZTIb D 8
|
|
|
|
|
_ZTIc D 8
|
|
|
|
|
_ZTId D 8
|
|
|
|
|
_ZTIe D 8
|
|
|
|
|
_ZTIf D 8
|
|
|
|
|
_ZTIh D 8
|
|
|
|
|
_ZTIi D 8
|
|
|
|
|
_ZTIj D 8
|
|
|
|
|
_ZTIl D 8
|
|
|
|
|
_ZTIm D 8
|
|
|
|
|
_ZTIs D 8
|
|
|
|
|
_ZTIt D 8
|
|
|
|
|
_ZTIx D 8
|
|
|
|
|
_ZTIy D 8
|
|
|
|
|
_ZTSN10__cxxabiv115__forced_unwindE R
|
|
|
|
|
_ZTSN10__cxxabiv116__enum_type_infoE R
|
|
|
|
|
_ZTSN10__cxxabiv117__class_type_infoE R
|
|
|
|
|
_ZTSN10__cxxabiv117__pbase_type_infoE R
|
|
|
|
|
_ZTSN10__cxxabiv119__foreign_exceptionE R
|
|
|
|
|
_ZTSN10__cxxabiv119__pointer_type_infoE R
|
|
|
|
|
_ZTSN10__cxxabiv120__function_type_infoE R
|
|
|
|
|
_ZTSN10__cxxabiv120__si_class_type_infoE R
|
|
|
|
|
_ZTSN10__cxxabiv121__vmi_class_type_infoE R
|
|
|
|
|
_ZTSN10__cxxabiv123__fundamental_type_infoE R
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZTSN5Timer10ConnectionE R
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZTSN6Genode11Sliced_heapE R
|
|
|
|
|
_ZTSN6Genode14Rpc_entrypointE R
|
|
|
|
|
_ZTSN6Genode14Signal_contextE R
|
|
|
|
|
_ZTSN6Genode14Timeout_threadE R
|
|
|
|
|
_ZTSN6Genode17Region_map_clientE R
|
|
|
|
|
_ZTSN6Genode17Rm_session_clientE R
|
|
|
|
|
_ZTSN6Genode18Allocator_avl_baseE R
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZTSN6Genode23Alarm_timeout_schedulerE R
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZTSN6Genode4HeapE R
|
|
|
|
|
_ZTSN6Genode4SlabE R
|
|
|
|
|
_ZTSN6Genode5AlarmE R
|
|
|
|
|
_ZTSN6Genode5Child14Initial_threadE R
|
|
|
|
|
_ZTSN6Genode5ChildE R
|
|
|
|
|
_ZTSN6Genode6OutputE R
|
|
|
|
|
_ZTSN6Genode6ThreadE R
|
|
|
|
|
_ZTSN6Genode7ConsoleE R
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZTSN6Genode7Timeout5AlarmE R
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZTSSt10bad_typeid R
|
|
|
|
|
_ZTSSt13bad_exception R
|
|
|
|
|
_ZTSSt16bad_array_length R
|
|
|
|
|
_ZTSSt20bad_array_new_length R
|
|
|
|
|
_ZTSSt8bad_cast R
|
|
|
|
|
_ZTSSt9bad_alloc R
|
|
|
|
|
_ZTSSt9exception R
|
|
|
|
|
_ZTSSt9type_info R
|
|
|
|
|
_ZTVN10__cxxabiv115__forced_unwindE D 1
|
|
|
|
|
_ZTVN10__cxxabiv115__forced_unwindE D 20
|
|
|
|
|
_ZTVN10__cxxabiv116__enum_type_infoE D 2
|
|
|
|
|
_ZTVN10__cxxabiv116__enum_type_infoE D 32
|
|
|
|
|
_ZTVN10__cxxabiv117__class_type_infoE D 2
|
|
|
|
|
_ZTVN10__cxxabiv117__class_type_infoE D 44
|
|
|
|
|
_ZTVN10__cxxabiv117__pbase_type_infoE D 2
|
|
|
|
|
_ZTVN10__cxxabiv117__pbase_type_infoE D 36
|
|
|
|
|
_ZTVN10__cxxabiv119__foreign_exceptionE D 1
|
|
|
|
|
_ZTVN10__cxxabiv119__foreign_exceptionE D 20
|
|
|
|
|
_ZTVN10__cxxabiv119__pointer_type_infoE D 2
|
|
|
|
|
_ZTVN10__cxxabiv119__pointer_type_infoE D 36
|
|
|
|
|
_ZTVN10__cxxabiv120__function_type_infoE D 2
|
|
|
|
|
_ZTVN10__cxxabiv120__function_type_infoE D 32
|
|
|
|
|
_ZTVN10__cxxabiv120__si_class_type_infoE D 2
|
|
|
|
|
_ZTVN10__cxxabiv120__si_class_type_infoE D 44
|
|
|
|
|
_ZTVN10__cxxabiv121__vmi_class_type_infoE D 2
|
|
|
|
|
_ZTVN10__cxxabiv121__vmi_class_type_infoE D 44
|
|
|
|
|
_ZTVN10__cxxabiv123__fundamental_type_infoE D 2
|
|
|
|
|
_ZTVN10__cxxabiv123__fundamental_type_infoE D 32
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZTVN5Timer10ConnectionE D 116
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZTVN6Genode11Sliced_heapE D 2
|
|
|
|
|
_ZTVN6Genode11Sliced_heapE D 36
|
|
|
|
|
_ZTVN6Genode14Rpc_entrypointE D 1
|
|
|
|
|
_ZTVN6Genode14Rpc_entrypointE D 24
|
|
|
|
|
_ZTVN6Genode14Signal_contextE D 1
|
|
|
|
|
_ZTVN6Genode14Signal_contextE D 16
|
|
|
|
|
_ZTVN6Genode14Timeout_threadE D 1
|
|
|
|
|
_ZTVN6Genode14Timeout_threadE D 24
|
|
|
|
|
_ZTVN6Genode17Region_map_clientE D 1
|
|
|
|
|
_ZTVN6Genode17Region_map_clientE D 28
|
|
|
|
|
_ZTVN6Genode17Rm_session_clientE D 1
|
|
|
|
|
_ZTVN6Genode17Rm_session_clientE D 24
|
|
|
|
|
_ZTVN6Genode18Allocator_avl_baseE D 4
|
|
|
|
|
_ZTVN6Genode18Allocator_avl_baseE D 64
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZTVN6Genode23Alarm_timeout_schedulerE D 80
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZTVN6Genode4HeapE D 2
|
|
|
|
|
_ZTVN6Genode4HeapE D 36
|
|
|
|
|
_ZTVN6Genode4SlabE D 2
|
|
|
|
|
_ZTVN6Genode4SlabE D 36
|
|
|
|
|
_ZTVN6Genode5AlarmE D 1
|
|
|
|
|
_ZTVN6Genode5AlarmE D 20
|
|
|
|
|
_ZTVN6Genode5Child14Initial_threadE D 1
|
|
|
|
|
_ZTVN6Genode5Child14Initial_threadE D 16
|
|
|
|
|
_ZTVN6Genode5ChildE D 11
|
|
|
|
|
_ZTVN6Genode5ChildE D 188
|
|
|
|
|
_ZTVN6Genode6OutputE D 1
|
|
|
|
|
_ZTVN6Genode6OutputE D 16
|
|
|
|
|
_ZTVN6Genode6ThreadE D 1
|
|
|
|
|
_ZTVN6Genode6ThreadE D 24
|
|
|
|
|
_ZTVN6Genode7ConsoleE D 1
|
|
|
|
|
_ZTVN6Genode7ConsoleE D 24
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZTVN6Genode7Timeout5AlarmE D 40
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZTVSt10bad_typeid D 1
|
|
|
|
|
_ZTVSt10bad_typeid D 20
|
|
|
|
|
_ZTVSt13bad_exception D 1
|
|
|
|
|
_ZTVSt13bad_exception D 20
|
|
|
|
|
_ZTVSt16bad_array_length D 1
|
|
|
|
|
_ZTVSt16bad_array_length D 20
|
|
|
|
|
_ZTVSt20bad_array_new_length D 1
|
|
|
|
|
_ZTVSt20bad_array_new_length D 20
|
|
|
|
|
_ZTVSt8bad_cast D 1
|
|
|
|
|
_ZTVSt8bad_cast D 20
|
|
|
|
|
_ZTVSt9bad_alloc D 1
|
|
|
|
|
_ZTVSt9bad_alloc D 20
|
|
|
|
|
_ZTVSt9exception D 1
|
|
|
|
|
_ZTVSt9exception D 20
|
|
|
|
|
_ZTVSt9type_info D 2
|
|
|
|
|
_ZTVSt9type_info D 32
|
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
2017-04-21 22:52:23 +00:00
|
|
|
_ZThn232_N5Timer10Connection16schedule_timeoutEN6Genode12MicrosecondsERNS1_11Time_source15Timeout_handlerE T
|
|
|
|
|
_ZThn232_N5Timer10Connection9curr_timeEv T
|
|
|
|
|
_ZThn236_N5Timer10Connection18_schedule_one_shotERN6Genode7TimeoutENS1_12MicrosecondsE T
|
|
|
|
|
_ZThn236_N5Timer10Connection18_schedule_periodicERN6Genode7TimeoutENS1_12MicrosecondsE T
|
|
|
|
|
_ZThn236_N5Timer10Connection8_discardERN6Genode7TimeoutE T
|
|
|
|
|
_ZThn236_N5Timer10Connection9curr_timeEv T
|
|
|
|
|
_ZThn280_N5Timer10Connection16schedule_timeoutEN6Genode12MicrosecondsERNS1_11Time_source15Timeout_handlerE T
|
|
|
|
|
_ZThn280_N5Timer10Connection9curr_timeEv T
|
|
|
|
|
_ZThn288_N5Timer10Connection18_schedule_one_shotERN6Genode7TimeoutENS1_12MicrosecondsE T
|
|
|
|
|
_ZThn288_N5Timer10Connection18_schedule_periodicERN6Genode7TimeoutENS1_12MicrosecondsE T
|
|
|
|
|
_ZThn288_N5Timer10Connection8_discardERN6Genode7TimeoutE T
|
|
|
|
|
_ZThn288_N5Timer10Connection9curr_timeEv T
|
|
|
|
|
_ZThn4_N6Genode23Alarm_timeout_scheduler14handle_timeoutENS_8DurationE T
|
|
|
|
|
_ZThn8_N6Genode23Alarm_timeout_scheduler14handle_timeoutENS_8DurationE T
|
2016-12-09 22:09:17 +00:00
|
|
|
_ZZN6Genode18Allocator_avl_base5BlockC4EmmbE10num_blocks V
|
|
|
|
|
_ZdlPv W
|
|
|
|
|
_ZdlPvPN6Genode11DeallocatorE T
|
|
|
|
|
_ZdlPvPN6Genode9AllocatorE W
|
|
|
|
|
_ZdlPvRN6Genode11DeallocatorE T
|
|
|
|
|
_ZdlPvRN6Genode9AllocatorE W
|
|
|
|
|
_ZnajPN6Genode9AllocatorE T
|
|
|
|
|
_ZnajRN6Genode9AllocatorE T
|
|
|
|
|
_ZnamPN6Genode9AllocatorE T
|
|
|
|
|
_ZnamRN6Genode9AllocatorE T
|
|
|
|
|
_ZnwjPN6Genode9AllocatorE T
|
|
|
|
|
_ZnwjRN6Genode9AllocatorE T
|
|
|
|
|
_ZnwmPN6Genode9AllocatorE T
|
|
|
|
|
_ZnwmRN6Genode9AllocatorE T
|
|
|
|
|
__aeabi_atexit T
|
|
|
|
|
__aeabi_unwind_cpp_pr0 T
|
|
|
|
|
__aeabi_unwind_cpp_pr1 T
|
|
|
|
|
__cxa_allocate_dependent_exception T
|
|
|
|
|
__cxa_allocate_exception T
|
|
|
|
|
__cxa_atexit T
|
|
|
|
|
__cxa_bad_cast T
|
|
|
|
|
__cxa_bad_typeid T
|
|
|
|
|
__cxa_begin_catch T
|
|
|
|
|
__cxa_begin_cleanup T
|
|
|
|
|
__cxa_call_terminate T
|
|
|
|
|
__cxa_call_unexpected T
|
|
|
|
|
__cxa_current_exception_type T
|
|
|
|
|
__cxa_demangle T
|
|
|
|
|
__cxa_end_catch T
|
|
|
|
|
__cxa_end_cleanup T
|
|
|
|
|
__cxa_finalize T
|
|
|
|
|
__cxa_free_dependent_exception T
|
|
|
|
|
__cxa_free_exception T
|
|
|
|
|
__cxa_get_exception_ptr T
|
|
|
|
|
__cxa_get_globals T
|
|
|
|
|
__cxa_get_globals_fast T
|
|
|
|
|
__cxa_guard_abort T
|
|
|
|
|
__cxa_guard_acquire T
|
|
|
|
|
__cxa_guard_release T
|
|
|
|
|
__cxa_pure_virtual T
|
|
|
|
|
__cxa_rethrow T
|
|
|
|
|
__cxa_throw T
|
|
|
|
|
__cxa_throw_bad_array_length T
|
|
|
|
|
__cxa_throw_bad_array_new_length T
|
|
|
|
|
__cxa_type_match T
|
|
|
|
|
__dynamic_cast T
|
|
|
|
|
__gxx_personality_v0 T
|
|
|
|
|
dl_iterate_phdr T
|
|
|
|
|
dl_unwind_find_exidx T
|
|
|
|
|
genode_argc D 4
|
|
|
|
|
genode_argv D 8
|
2016-12-29 17:27:45 +00:00
|
|
|
genode_envp B 4
|
|
|
|
|
lx_environ B 4
|
2016-12-09 22:09:17 +00:00
|
|
|
memcpy W
|
|
|
|
|
memmove W
|
|
|
|
|
memset W
|
|
|
|
|
stdout_reconnect T
|
|
|
|
|
stdout_write T
|
|
|
|
|
wait_for_continue T
|
