synchronization

In the reader/writer synchronization scheme, when some processes are reading

• a new writer can write

• a new writer cannot write

• a new reader can read

• a new reader cannot read

In the reader/writer synchronization scheme, when a process is writing

• a new writer can write

• a new writer cannot write

• a new reader can read

• a new reader cannot read

With Posix monitors, when you call pthread_cond_signal(&condition);

• the next thread which calls pthread_cond_wait() on the condition will not block

• one thread which is blocked on the condition is resumed

• if no thread is blocked on the condition, nothing happens and the signal is lost

• all the threads which are blocked on the condition are resumed

With semaphores, the P() function

• suspends the calling process if the counter is positive

• suspends the calling process if the counter is negative

• increments the counter

• decrements the counter

In a Posix monitor, what happens if we don't use the mutex lock ?

• the executing thread can block

• the shared data may become unconsistent

• we can have a deadlock

• all the threads can access the shared data at the same time

In the prodcons labwork with the products and freeslots conditions

• possibly no threads are blocked on any of the two conditions

• possibly threads are blocked on both conditions

• if threads are blocked, they are all blocked on the same condition

• a thread can block at the same time on both conditions

In the prodcons labwork, the buffer will be most of the time

• empty if we create more producer threads

• full if we create more producer threads

• empty if we create more consumer threads

• full if we create more consumer threads

In the prodcons labwork with the products and freeslots conditions, if I produce, I must:

• test if the buffer is empty

• test if the buffer is full

• possibly block on the freeslots condition

• possibly block on the products condition