Talking to Threads. Administriva Grade summary today.

Talking to Threads

Administriva•Grade summary today

Grade Summ: Histogram

Project trends

Main routine: Dijkstra’sfunction dijkstraSSP(G,w,s):

// G: complete graph// w: weight function// s: source node (start location)initializeSSPDataStructs();Set S=new Set();Queue Q=new PriorityQueue(G.V,d);while (!Q.isEmpty()) {

Vertex u=Q.extractMin();S.add(u);for (Iterator i=u.neighbors();i.hasNext();)

{Vertex v=i.next();

relax(u,v,w);}

}

Group questions:•Give an example of this alg in action on a

(small) map from JCiv•How long does this take to run?•What real data structs should you use

for S and Q?•Can you take any shortcuts?•What’s wrong w/ negative-weight edges?•Is this guaranteed to get the correct

shortest path? Why? (Proof!)•Show that d[v] is the correct shortest

path when v is inserted in Set S and that d[v] never changes thereafter.

•Hint: think proof by contradiction...

Group questions:•Give an example of this alg in action on a

(small) map from JCiv•How long does this take to run?•What real data structs should you use

for S and Q?•Can you take any shortcuts?•What’s wrong w/ negative-weight edges?•Is this guaranteed to get the correct

shortest path? Why? (Proof!)•Show that d[v] is the correct shortest

path when v is inserted in Set S and that d[v] never changes thereafter.

•Hint: think proof by contradiction...

Back to Threads

wait()ing for Godot... sleep(n) is good if you know how long you

want to delay What if a thread wants to wait

indefinately, or for some other thread to do something?

Check out Object API: Object o.wait() -- wait until some other

object tells thread to wake up o.notify() -- wake up one (random)

thread that is waiting on o o.notifyAll() -- wake up every thread

that is waiting on o.

More BtS Java•Every Object in Java has a (single) “wait list”.

•When a thread adds itself to an Object’s wait list, the thread suspends

•Some other thread of execution can then call that Object’s notify() method

•One thread wakes up and can now take action

•Does not generate an exception; does not change the status of a sleep()ing thread

•Must be synchronized on target object to wait() or notify()

•Prevents collisions while messing w/ wait list

Use of wait listsDataBucket b;Thread t=new Thread(new Runnable(b) { public Runnable(DataBucket buck) { _bucket=b; } public void run() { while (!bored) { while (_bucket.isEmpty()) { synchronized(_bucket) { _bucket.wait(); } } _bucket.getData(); } }});

// elsewhere...b.addData(data);synchronized(b) { b.notify(); }

Race Cond. & Security•Atomicity failures can sometimes be exploited

to break security on multiprocessing systems

•One of the top 10 classes of exploits since... mid-1980’s, at least

•100’s (or more) of reported vulnerabilities

•Independent of language: Java will not save you!

•Hostile program grabs a shared resource (e.g., file) before it is secured

•Beware when writing privileged code!

•N.b.: Sometimes your never-intended-to-be- secure code will be run in privileged context!

Basic Race Cond. Exploitpriv proc

Basic Race Cond. Exploitpriv proc

file/tmp/foo

write()

read()

close()

unlink()

open(“/tmp/foo”, O_RDWR | O_CREAT);

Basic Race Cond. Exploitpriv proc

open(“/tmp/foo”, O_RDWR | O_CREAT);

file/tmp/foo

write()

read()

close()

unlink()

hostile proc

open(...)

read()

Basic Race Cond. Exploitpriv proc

open(“/tmp/foo”, O_RDWR | O_CREAT);

file/tmp/foo

write()

read()

close()

unlink()

hostile proc

chmod()

Basic Race Cond. Exploitpriv proc

open(“/tmp/foo”, O_RDWR | O_CREAT);

file/tmp/foo

write()

read()

close()

unlink()

hostile proc

chmod()

open(...)

Basic Race Cond. Exploitpriv proc

open(“/tmp/foo”, O_RDWR | O_CREAT);

file/tmp/foo

write()

read()

close()

unlink()

hostile procumask()

Basic Race Cond. Exploitpriv proc

open(“/tmp/foo”, O_RDWR | O_CREAT);

file/tmp/foo

write()

read()

close()

unlink()

hostile procumask() open(...)

read()

Basic Race Cond. Exploitpriv proc

open(“/tmp/foo”, O_RDWR | O_CREAT);

file/tmp/foo

write()

read()

close()

unlink()

hostile procumask()

symlink(“/tmp/foo”, “/etc/passwd”)

Basic Race Cond. Exploitpriv proc

stat(“/tmp/foo”);if (!exists) {

open(“/tmp/foo”, O_RDWR | O_CREAT);} else { error(); }

file/tmp/foo

write()

read()

close()

unlink()

hostile procumask()

Basic Race Cond. Exploitpriv proc

stat(“/tmp/foo”);if (!exists) {

open(“/tmp/foo”, O_RDWR | O_CREAT);} else { error(); }

file/tmp/foo

write()

read()

close()

unlink()

hostile procumask()

symlink(“/tmp/foo”, “/etc/passwd”)

Preventing FS Race Conds•Could create “foo” in dir owned/writable only

by owner of proc

•Can be hard to ensure this

•Still have to watch out for filename collisions

•Could make file names hard to predict (e.g., picked randomly)

•Exploit still possible; hard to make fnames really random

•Ultimate answer: use OS atomicity facilities

• open(“/tmp/foo”, O_RDWR | O_CREAT | O_EXCL)

•Always be on guard!