Aren/haiku
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Added new kernel thread "undertaker" which gets rid of dead thread

Browse Source 
remains. This replaces the previous mechanism of switching the thread to
a dedicated death stack. We might consider moving more cleanup work to
the undertaker, but that seems a little more involved.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25704 a95241bf-73f2-0310-859d-f6bbb57e9c96
This commit is contained in:
Ingo Weinhold 2008-05-29 14:28:31 +00:00
parent 2620242fd5
commit 9101535322
3 changed files with 72 additions and 199 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
headers/private/kernel/thread_types.h
View File

@ -23,8 +23,6 @@ extern spinlock thread_spinlock;
#define GRAB_THREAD_LOCK() acquire_spinlock(&thread_spinlock)
#define RELEASE_THREAD_LOCK() release_spinlock(&thread_spinlock)
extern struct thread_queue dead_q;
extern spinlock team_spinlock;
// NOTE: TEAM lock can be held over a THREAD lock acquisition,
// but not the other way (to avoid deadlock)

3
src/system/kernel/scheduler.cpp
View File

@ -173,9 +173,6 @@ scheduler_reschedule(void)
TRACE(("reschedule(): suspending thread %ld\n", oldThread->id));
break;
case THREAD_STATE_FREE_ON_RESCHED:
// This will hopefully be eliminated once the slab
// allocator is done
thread_enqueue(oldThread, &dead_q);
break;
default:
TRACE(("not enqueueing thread %ld into run q. next_state = %ld\n", oldThread->id, oldThread->next_state));

266
src/system/kernel/thread.cpp
View File

@ -54,15 +54,6 @@
#define THREAD_MAX_MESSAGE_SIZE 65536
// used to pass messages between thread_exit and thread_exit2
struct thread_exit_args {
struct thread *thread;
area_id old_kernel_stack;
uint32 death_stack;
sem_id death_sem;
team_id original_team_id;
};
struct thread_key {
thread_id id;
@ -81,22 +72,27 @@ static thread_id sNextThreadID = 1;
static int32 sMaxThreads = 4096;
static int32 sUsedThreads = 0;
// death stacks - used temporarily as a thread cleans itself up
struct death_stack {
area_id area;
addr_t address;
bool in_use;
struct UndertakerEntry : DoublyLinkedListLinkImpl<UndertakerEntry> {
struct thread* thread;
team_id teamID;
sem_id deathSem;
UndertakerEntry(struct thread* thread, team_id teamID, sem_id deathSem)
:
thread(thread),
teamID(teamID),
deathSem(deathSem)
{
}
};
static struct death_stack *sDeathStacks;
static unsigned int sNumDeathStacks;
static unsigned int volatile sDeathStackBitmap;
static sem_id sDeathStackSem;
static spinlock sDeathStackLock = 0;
static DoublyLinkedList<UndertakerEntry> sUndertakerEntries;
static ConditionVariable sUndertakerCondition;
// The dead queue is used as a pool from which to retrieve and reuse previously
// allocated thread structs when creating a new thread. It should be gone once
// the slab allocator is in.
struct thread_queue dead_q;
static struct thread_queue dead_q;
static void thread_kthread_entry(void);
static void thread_kthread_exit(void);
@ -539,143 +535,58 @@ create_thread(thread_creation_attributes& attributes, bool kernel)
}
/*!
Finds a free death stack for us and allocates it.
Must be called with interrupts enabled.
*/
static uint32
get_death_stack(void)
static status_t
undertaker(void* /*args*/)
{
cpu_status state;
uint32 bit;
int32 i;
while (true) {
// wait for a thread to bury
ConditionVariableEntry conditionEntry;
acquire_sem(sDeathStackSem);
InterruptsSpinLocker locker(thread_spinlock);
sUndertakerCondition.Add(&conditionEntry);
locker.Unlock();
// grab the death stack and thread locks, find a free spot and release
conditionEntry.Wait();
state = disable_interrupts();
locker.Lock();
UndertakerEntry* _entry = sUndertakerEntries.RemoveHead();
locker.Unlock();
acquire_spinlock(&sDeathStackLock);
GRAB_THREAD_LOCK();
if (_entry == NULL)
continue;
bit = sDeathStackBitmap;
bit = (~bit) & ~((~bit) - 1);
sDeathStackBitmap |= bit;
UndertakerEntry entry = *_entry;
// we need a copy, since the original entry is on the thread's stack
RELEASE_THREAD_LOCK();
release_spinlock(&sDeathStackLock);
// we've got an entry
struct thread* thread = entry.thread;
restore_interrupts(state);
// delete the old kernel stack area
delete_area(thread->kernel_stack_area);
// sanity checks
if (!bit)
panic("get_death_stack: couldn't find free stack!\n");
// remove this thread from all of the global lists
disable_interrupts();
GRAB_TEAM_LOCK();
if (bit & (bit - 1))
panic("get_death_stack: impossible bitmap result!\n");
remove_thread_from_team(team_get_kernel_team(), thread);
// bit to number
for (i = -1; bit; i++) {
bit >>= 1;
RELEASE_TEAM_LOCK();
enable_interrupts();
// needed for the debugger notification below
if (entry.deathSem >= 0)
release_sem_etc(entry.deathSem, 1, B_DO_NOT_RESCHEDULE);
// notify the debugger
if (entry.teamID >= 0
&& entry.teamID != team_get_kernel_team_id()) {
user_debug_thread_deleted(entry.teamID, thread->id);
}
// free the thread structure
thread_enqueue(thread, &dead_q);
// TODO: Use the slab allocator!
}
TRACE(("get_death_stack: returning %#lx\n", sDeathStacks[i].address));
return (uint32)i;
}
/*! Returns the thread's death stack to the pool.
Interrupts must be disabled and the sDeathStackLock be held.
*/
static void
put_death_stack(uint32 index)
{
TRACE(("put_death_stack...: passed %lu\n", index));
if (index >= sNumDeathStacks)
panic("put_death_stack: passed invalid stack index %ld\n", index);
if (!(sDeathStackBitmap & (1 << index)))
panic("put_death_stack: passed invalid stack index %ld\n", index);
GRAB_THREAD_LOCK();
sDeathStackBitmap &= ~(1 << index);
RELEASE_THREAD_LOCK();
release_sem_etc(sDeathStackSem, 1, B_DO_NOT_RESCHEDULE);
// we must not hold the thread lock when releasing a semaphore
}
static void
thread_exit2(void *_args)
{
struct thread_exit_args args;
// copy the arguments over, since the source is probably on the kernel
// stack we're about to delete
memcpy(&args, _args, sizeof(struct thread_exit_args));
// we can't let the interrupts disabled at this point
enable_interrupts();
TRACE(("thread_exit2, running on death stack %#lx\n", args.death_stack));
// delete the old kernel stack area
TRACE(("thread_exit2: deleting old kernel stack id %ld for thread %ld\n",
args.old_kernel_stack, args.thread->id));
delete_area(args.old_kernel_stack);
// remove this thread from all of the global lists
TRACE(("thread_exit2: removing thread %ld from global lists\n",
args.thread->id));
disable_interrupts();
GRAB_TEAM_LOCK();
remove_thread_from_team(team_get_kernel_team(), args.thread);
RELEASE_TEAM_LOCK();
enable_interrupts();
// needed for the debugger notification below
TRACE(("thread_exit2: done removing thread from lists\n"));
if (args.death_sem >= 0)
release_sem_etc(args.death_sem, 1, B_DO_NOT_RESCHEDULE);
// notify the debugger
if (args.original_team_id >= 0
&& args.original_team_id != team_get_kernel_team_id()) {
user_debug_thread_deleted(args.original_team_id, args.thread->id);
}
disable_interrupts();
// Set the next state to be gone: this will cause the thread structure
// to be returned to a ready pool upon reschedule.
// Note, we need to have disabled interrupts at this point, or else
// we could get rescheduled too early.
args.thread->next_state = THREAD_STATE_FREE_ON_RESCHED;
// return the death stack and reschedule one last time
// Note that we need to hold sDeathStackLock until we've got the thread
// lock. Otherwise someone else might grab our stack in the meantime.
acquire_spinlock(&sDeathStackLock);
put_death_stack(args.death_stack);
GRAB_THREAD_LOCK();
release_spinlock(&sDeathStackLock);
scheduler_reschedule();
// requires thread lock to be held
// never get to here
panic("thread_exit2: made it where it shouldn't have!\n");
}
@ -1591,28 +1502,17 @@ thread_exit(void)
delete_sem(cachedExitSem);
}
{
struct thread_exit_args args;
// enqueue in the undertaker list and reschedule for the last time
UndertakerEntry undertakerEntry(thread, teamID, cachedDeathSem);
args.thread = thread;
args.old_kernel_stack = thread->kernel_stack_area;
args.death_stack = get_death_stack();
args.death_sem = cachedDeathSem;
args.original_team_id = teamID;
disable_interrupts();
GRAB_THREAD_LOCK();
sUndertakerEntries.Add(&undertakerEntry);
sUndertakerCondition.NotifyOne(true);
disable_interrupts();
// set the new kernel stack officially to the death stack, it won't be
// switched until the next function is called. This must be done now
// before a context switch, or we'll stay on the old stack
thread->kernel_stack_area = sDeathStacks[args.death_stack].area;
thread->kernel_stack_base = sDeathStacks[args.death_stack].address;
// we will continue in thread_exit2(), on the new stack
arch_thread_switch_kstack_and_call(thread, thread->kernel_stack_base
+ KERNEL_STACK_SIZE, thread_exit2, &args);
}
thread->next_state = THREAD_STATE_FREE_ON_RESCHED;
scheduler_reschedule();
panic("never can get here\n");
}
@ -2108,37 +2008,15 @@ thread_init(kernel_args *args)
}
sUsedThreads = args->num_cpus;
// create a set of death stacks
// start the undertaker thread
new(&sUndertakerEntries) DoublyLinkedList<UndertakerEntry>();
sUndertakerCondition.Init(&sUndertakerEntries, "undertaker entries");
sNumDeathStacks = smp_get_num_cpus();
if (sNumDeathStacks > 8 * sizeof(sDeathStackBitmap)) {
// clamp values for really beefy machines
sNumDeathStacks = 8 * sizeof(sDeathStackBitmap);
}
sDeathStackBitmap = 0;
sDeathStacks = (struct death_stack *)malloc(sNumDeathStacks
* sizeof(struct death_stack));
if (sDeathStacks == NULL) {
panic("error creating death stacks\n");
return B_NO_MEMORY;
}
{
char temp[64];
for (i = 0; i < sNumDeathStacks; i++) {
sprintf(temp, "death stack %lu", i);
sDeathStacks[i].area = create_area(temp,
(void **)&sDeathStacks[i].address, B_ANY_KERNEL_ADDRESS,
KERNEL_STACK_SIZE, B_FULL_LOCK,
B_KERNEL_READ_AREA | B_KERNEL_WRITE_AREA | B_KERNEL_STACK_AREA);
if (sDeathStacks[i].area < 0) {
panic("error creating death stacks\n");
return sDeathStacks[i].area;
}
sDeathStacks[i].in_use = false;
}
}
sDeathStackSem = create_sem(sNumDeathStacks, "death stack availability");
thread_id undertakerThread = spawn_kernel_thread(&undertaker, "undertaker",
B_DISPLAY_PRIORITY, NULL);
if (undertakerThread < 0)
panic("Failed to create undertaker thread!");
resume_thread(undertakerThread);
// set up some debugger commands
add_debugger_command_etc("threads", &dump_thread_list, "List all threads",