matteo/sgpemv2
matteo/sgpemv2
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

fixed a bug involving scheduler keeping a running thread in the ready queue

Browse source 
when that thread had just been given the cpu.
fixed a bug involving text-simulation displaying the wrong timestamp for
each snaphot. Now the first snapshot printed is correctly tagged "-1".

added (for the sake of lazyness) a method to readyqueue letting anyone
delete the first element on the queue. Not having this would imply
rebuilding the whole queue each time a thread was selected to run.

Small commits avoid complex mergings.


git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@814 3ecf2c5c-341e-0410-92b4-d18e462d057c
This commit is contained in:
matrevis 2006-08-02 23:12:32 +00:00
parent 01490e24ac
commit 56a7ce1221
4 changed files with 21 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

6
src/backend/ready_queue.cc
View file

@ -70,3 +70,9 @@ ReadyQueue::append(Thread& thread)
{
_scheds.push_back(&thread);
}
void
ReadyQueue::erase_first()
{
_scheds.erase(_scheds.begin());
}

1
src/backend/ready_queue.hh
View file

@ -42,6 +42,7 @@ namespace sgpem
Thread& get_item_at(position index) throw (std::out_of_range);
const Thread& get_item_at(position index) const throw (std::out_of_range);
void append(Thread& schedulable);
void erase_first();
private:
typedef std::vector<Thread*> Threads;

4
src/backend/scheduler.cc
View file

@ -314,6 +314,10 @@ Scheduler::step_forward(History& history, CPUPolicy& cpu_policy) throw(UserInter
DynamicThread& new_running = (DynamicThread&) _ready_queue->get_item_at(0);
new_running.set_state(Schedulable::state_running);
new_running.set_last_acquisition(current_instant);
// removes running element from the ready queue
// since no method was provided to erase an item in the queue, we should rebuild it.
// this is pointless. I just added the method.
_ready_queue->erase_first();
}