eive/fsfw
eive/fsfw
15
1
Fork 0
forked from fsfw/fsfw
Code Issues Pull Requests Releases 1 Wiki Activity

implement relative timeshift #173

Merged
muellerr merged 10 commits from time-service-relative-timeshift into develop 2024-04-09 13:31:13 +02:00
Conversation 1 Commits 10 Files Changed 3 +24 -15
2 changed files with 24 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files
Showing only changes of commit 31d4b85523 - Show all commits

33
src/fsfw/pus/Service9TimeManagement.cpp
View File

@ -2,10 +2,9 @@
#include <cmath>
#include "fsfw/events/EventManagerIF.h"
#include "fsfw/pus/servicepackets/Service9Packets.h"
#include "fsfw/returnvalues/returnvalue.h"
#include "fsfw/serialize/SerializeAdapter.h"
#include "fsfw/serviceinterface/ServiceInterface.h"
#include "fsfw/timemanager/CCSDSTime.h"
Service9TimeManagement::Service9TimeManagement(PsbParams params) : PusServiceBase(params) {
@ -19,7 +18,10 @@ ReturnValue_t Service9TimeManagement::performService() { return returnvalue::OK;
ReturnValue_t Service9TimeManagement::handleRequest(uint8_t subservice) {
switch (subservice) {
case Subservice::SET_TIME: {
return setTime();
reportCurrentTime();
ReturnValue_t result = setTime();
reportCurrentTime();
meggert commented 2024-04-08 13:35:00 +02:00
Outdated
Review
Copy Link

if we are doing this twice, why did we not use the CLOCK_SET event then? it already contained old and new time. or are we really interessted in microseconds here?

if we are doing this twice, why did we not use the CLOCK_SET event then? it already contained old and new time. or are we really interessted in microseconds here?
muellerr commented 2024-04-08 13:42:12 +02:00
Outdated
Review
Copy Link

events are cheap, and why lose resolution here? The correct approach probably would be to delete the weird CLOCK_SET event. If we still want to distinguish events which come from time changes, we could introduce, TIME_DUMP_BEFORE_CLOCK_UPDATE and TIME_DUMP_AFTER_CLOCK_UPDATE..

events are cheap, and why lose resolution here? The correct approach probably would be to delete the weird `CLOCK_SET` event. If we still want to distinguish events which come from time changes, we could introduce, `TIME_DUMP_BEFORE_CLOCK_UPDATE` and `TIME_DUMP_AFTER_CLOCK_UPDATE`..
meggert commented 2024-04-08 13:50:42 +02:00
Outdated
Review
Copy Link

I would advise splitting the events in those two types yes. Otherwise, this might be confusing for the operator. Calculating the time delta on ground is going to be more tricky now (but still possible if we store the parameter tm as actual tm). Maybe ask OPS what they prefer here

I would advise splitting the events in those two types yes. Otherwise, this might be confusing for the operator. Calculating the time delta on ground is going to be more tricky now (but still possible if we store the parameter tm as actual tm). Maybe ask OPS what they prefer here
muellerr commented 2024-04-09 11:11:19 +02:00
Outdated
Review
Copy Link

done

done
return result;
}
case Subservice::DUMP_TIME: {
timeval newTime;
@ -36,6 +38,7 @@ ReturnValue_t Service9TimeManagement::handleRequest(uint8_t subservice) {
if (result != returnvalue::OK) {
return result;
}
triggerEvent(CLOCK_DUMP, currentTime.tv_sec, currentTime.tv_usec);
if (currentPacket.getUserDataLen() != 8) {
return AcceptsTelecommandsIF::ILLEGAL_APPLICATION_DATA;
@ -63,8 +66,7 @@ ReturnValue_t Service9TimeManagement::handleRequest(uint8_t subservice) {
}
result = Clock::setClock(&newTime);
muellerr marked this conversation as resolved Outdated
meggert commented 2024-04-08 09:03:19 +02:00
Outdated
Review
Copy Link

see above

see above
if (result == returnvalue::OK) {
// Report new time as event.
triggerEvent(CLOCK_DUMP, newTime.tv_sec, newTime.tv_usec);
reportTime(newTime);
meggert commented 2024-04-08 13:35:55 +02:00
Outdated
Review
Copy Link

the same here

the same here
}
return result;
}
@ -83,17 +85,20 @@ ReturnValue_t Service9TimeManagement::setTime() {
return result;
}
timeval time;
Clock::getClock_timeval(&time);
result = Clock::setClock(&timeToSet);
if (result == returnvalue::OK) {
timeval newTime;
Clock::getClock_timeval(&newTime);
triggerEvent(CLOCK_SET, time.tv_sec, newTime.tv_sec);
return returnvalue::OK;
} else {
if (result != returnvalue::OK) {
triggerEvent(CLOCK_SET_FAILURE, result, 0);
return returnvalue::FAILED;
}
return result;
}
void Service9TimeManagement::reportCurrentTime() {
timeval currentTime{};
Clock::getClock_timeval(&currentTime);
triggerEvent(CLOCK_DUMP, currentTime.tv_sec, currentTime.tv_usec);
}
void Service9TimeManagement::reportTime(timeval time) {
triggerEvent(CLOCK_DUMP, time.tv_sec, time.tv_usec);
meggert commented 2024-04-08 13:36:40 +02:00
Outdated
Review
Copy Link

best function ever

best function ever
🚀 1
}

6
src/fsfw/pus/Service9TimeManagement.h
View File

@ -1,13 +1,14 @@
#ifndef FSFW_PUS_SERVICE9TIMEMANAGEMENT_H_
#define FSFW_PUS_SERVICE9TIMEMANAGEMENT_H_
#include "fsfw/returnvalues/returnvalue.h"
#include "fsfw/tmtcservices/PusServiceBase.h"
class Service9TimeManagement : public PusServiceBase {
public:
static constexpr uint8_t SUBSYSTEM_ID = SUBSYSTEM_ID::PUS_SERVICE_9;
static constexpr uint32_t NANOS_PER_SECOND= 1'000'000'000;
static constexpr uint32_t NANOS_PER_SECOND = 1'000'000'000;
//!< Clock has been set. P1: old timeval seconds. P2: new timeval seconds.
static constexpr Event CLOCK_SET = MAKE_EVENT(0, severity::INFO);
@ -32,6 +33,9 @@ class Service9TimeManagement : public PusServiceBase {
*/
ReturnValue_t handleRequest(uint8_t subservice) override;
void reportCurrentTime();
void reportTime(timeval time);
virtual ReturnValue_t setTime();
private: