badcnt c An Improperly Synchronized Threaded Program 15 213 The course that gives CMU its Zip shared volatile unsigned int cnt 0 define NITERS 100000000 Synchronization December 4 2007 int main pthread t tid1 tid2 Pthread create tid1 count Pthread create tid2 count Topics NULL NULL NULL NULL thread routine void count void arg int i for i 0 i NITERS i cnt return NULL linux badcnt BOOM cnt 198841183 linux badcnt BOOM cnt 198261801 Pthread join tid1 NULL Pthread join tid2 NULL Synchronizing with semaphores Races and deadlocks Thread safety and reentrancy if cnt unsigned NITERS 2 printf BOOM cnt d n cnt else printf OK cnt d n cnt lecture 26 ppt linux badcnt BOOM cnt 198269672 cnt should be equal to 200 000 000 What went wrong 2 Assembly Code for Counter Loop Concurrent Execution Key idea In general any sequentially consistent interleaving is possible but some are incorrect C code for counter loop for i 0 i NITERS i cnt Corresponding asm code Ii denotes that thread i executes instruction I eaxi is the contents of eax in thread i s context L9 movl 4 ebp eax cmpl 99999999 eax jle L12 jmp L10 Head Hi L12 Load cnt Li Update cnt Ui Store cnt Si movl cnt eax leal 1 eax edx movl edx cnt Load Update Store L11 movl 4 ebp eax leal 1 eax edx movl edx 4 ebp jmp L9 Tail Ti L10 3 15 213 F 07 i thread instri eax1 eax2 cnt 1 1 1 1 2 2 2 2 2 1 H1 L1 U1 S1 H2 L2 U2 S2 T2 T1 0 1 1 1 1 2 2 2 0 0 0 1 1 1 1 2 2 2 4 15 213 F 07 15 213 F 07 Concurrent Execution cont Concurrent Execution cont Incorrect ordering two threads increment the counter but the result is 1 instead of 2 How about this ordering 5 i thread instri eax1 eax2 cnt 1 1 1 2 2 1 1 2 2 2 H1 L1 U1 H2 L2 S1 T1 U2 S2 T2 0 1 1 1 0 1 1 1 0 0 0 0 0 1 1 1 1 1 Oops OK i thread instri 1 1 2 2 2 2 1 1 1 2 H1 L1 H2 L2 U2 S2 U1 S1 T1 T2 eax1 eax2 cnt We can clarify our understanding of concurrent execution with the help of the progress graph 6 15 213 F 07 Page 1 15 213 F 07 Progress Graphs Trajectories in Progress Graphs A progress graph depicts the discrete execution state space of concurrent threads Thread 2 T2 Thread 2 T2 L1 S2 Each axis corresponds to the sequential order of instructions in a thread S2 U2 S2 Example U2 Each point corresponds to a possible execution state Inst1 Inst2 L2 H2 H1 L1 U1 S1 Thread 1 T1 H2 H1 Critical Sections and Unsafe Regions Unsafe region U2 L1 U1 S1 T1 critical section wrt cnt Def A trajectory is safe iff it doesn t touch any Unsafe part of an unsafe region trajectory Unsafe region Claim A trajectory is correct wrt cnt iff it is safe U2 L2 H2 Thread 1 H1 L1 U1 S1 T1 Thread 1 critical section wrt cnt critical section wrt cnt 9 Safe trajectory S2 H2 H1 Thread 1 T1 15 213 F 07 T2 Sets of states where such interleaving occurs form unsafe regions L2 S1 Thread 2 Instructions in critical sections wrt to some shared variable should not be interleaved S2 critical section wrt cnt U1 Safe and Unsafe Trajectories L U and S form a critical section with respect to the shared variable cnt T2 L1 8 15 213 F 07 Thread 2 H1 L1 U1 H2 L2 S1 T1 U2 S2 T2 L2 E g L1 S2 denotes state where thread 1 has completed L1 and thread 2 has completed S2 7 A trajectory is a sequence of legal state transitions that describes one possible concurrent execution of the threads 10 15 213 F 07 15 213 F 07 Semaphores Safe Sharing with Semaphores Question How can we guarantee a safe trajectory Here is how we would use P and V operations to synchronize the threads that update cnt We must synchronize the threads so that they never enter an unsafe state Classic solution Dijkstra s P and V operations on semaphores Semaphore s is initially 1 semaphore non negative integer synchronization variable Thread routine void count void arg int i P s while s 0 wait s Dutch for Proberen test V s s Dutch for Verhogen increment for i 0 i NITERS i P s cnt V s return NULL OS guarantees that operations between brackets are executed indivisibly Only one P or V operation at a time can modify s When while loop in P terminates only that P can decrement s Semaphore invariant s 0 11 12 15 213 F 07 Page 2 15 213 F 07 Safe Sharing With Semaphores Wrappers on POSIX Semaphores Thread 2 1 1 0 0 0 0 1 1 1 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 V s Forbidden region 1 S2 U2 L2 P s 1 1 1 0 0 0 0 1 1 0 0 1 1 1 1 0 0 0 0 1 1 1 0 0 0 0 1 1 1 1 1 1 Unsafe region 1 1 1 1 Initialize semaphore sem to value pshared 0 if thread pshared 1 if process void Sem init sem t sem int pshared unsigned int value if sem init sem pshared value 0 unix error Sem init Provide mutually exclusive access to shared variable by surrounding critical section with P and V operations on semaphore s initially set to 1 T2 P operation on semaphore sem void P sem t sem if sem wait sem unix error P Semaphore invariant creates a forbidden region that encloses unsafe region and is never touched by any trajectory V operation on semaphore sem void V sem t sem if sem post sem unix error V H2 1 H1 P s L1 U1 S1 V s T1 Thread 1 Initially s 1 13 14 15 213 F 07 Sharing With POSIX Semaphores properly sync d counter program include csapp h define NITERS 10000000 A race occurs when the correctness of the program depends on one thread reaching point x before another thread reaches point y for i 0 i NITERS i P sem cnt V sem return NULL int main pthread t tid1 tid2 Sem init sem 0 1 sem 1 Races thread routine void count void arg int i volatile unsigned int cnt sem t sem semaphore a threaded program with a race int main pthread t tid N int i for i 0 i N i Pthread create tid i NULL thread i for i 0 i N i Pthread join tid i NULL exit 0 create 2 threads and wait if cnt unsigned NITERS 2 printf BOOM cnt d n cnt else printf OK cnt d n cnt exit 0 thread …
View Full Document
Unlocking...