| Rev |
Age |
Author |
Path |
Log message |
Diff |
| 2446 |
6375 d 10 h |
jermar |
/trunk/kernel/generic/ |
More efficient and simpler task termination.
Based on the assumption, that after its creation, only the task itself can create more threads for itself,
the last thread with userspace context to execute thread_exit() will perform futex and IPC cleanup. When
the task has no threads, it is destroyed. Both the cleanup and destruction is controlled by reference
counting.
As for userspace threads, even though there could be a global garbage collector for joining threads, it is
much simpler if the uinit thread detaches itself before switching to userspace.
task_kill() is now an idempotent operation. It just instructs the threads within a task to exit.
Change in the name of a thread state: Undead -> JoinMe. |
|
| 2440 |
6376 d 5 h |
jermar |
/trunk/kernel/generic/ |
Fix the problem with sys_thread_create() by splitting the create and attach functionality of
thread_create(). Now it is possible to specify a flag that will cause thread_create() to only allocate and
initialize the thread structure. A call to thread_attach() will make the thread visible to the system.
This arrangement makes it easier to undo creation of a thread in case of a failure in sys_thread_create(). |
|
| 2436 |
6377 d 6 h |
jermar |
/trunk/ |
Fix two memory leaks.
In kernel, kernel_uarg structure needs to be deallocated when a thread
with userspace context is destroyed.
In userspace, the return value of the SYS_THREAD_CREATE must be checked
for error conditions and in case of error, uarg and stack must be freed
up. |
|
| 2268 |
6419 d 7 h |
jermar |
/trunk/kernel/generic/ |
Document and assert proper usage of THREAD_FLAG_WIRED.
Original patch by Jan Hudecek. |
|
| 2229 |
6428 d 12 h |
jermar |
/trunk/ |
Revert some of the changes introduced in revision 2209.
I think it is not correct to remove serialization of pseudo threads in printf_core.c.
With thread-level futex serialization, several pseudo threads running in one thread could easily deadlock
the task.
Add a dedicated futex serialization to thread1.c test only. |
|
| 2216 |
6430 d 6 h |
decky |
/trunk/ |
make thread ID 64 bit (task ID is 64 bit already)
cleanup thread syscalls |
|
| 2187 |
6431 d 13 h |
decky |
/trunk/kernel/generic/ |
add syscall for getting current thread ID |
|
| 2183 |
6432 d 10 h |
jermar |
/trunk/kernel/generic/ |
Continue to de-oversynchronize the kernel.
- replace as->refcount with an atomic counter; accesses to this
reference counter are not to be done when the as->lock mutex is held;
this gets us rid of mutex_lock_active();
Remove the possibility of a deadlock between TLB shootdown and asidlock.
- get rid of mutex_lock_active() on as->lock
- when locking the asidlock spinlock, always do it conditionally and with
preemption disabled; in the unsuccessful case, enable interrupts and try again
- there should be no deadlock between TLB shootdown and the as->lock mutexes
- PLEASE REVIEW !!!
Add DEADLOCK_PROBE's to places where we have spinlock_trylock() loops. |
|
| 2118 |
6491 d 11 h |
decky |
/trunk/kernel/generic/src/ |
explicit typecast, fix signed/unsigned comparison |
|
| 2109 |
6492 d 14 h |
jermar |
/trunk/kernel/generic/ |
Revert thread_interrupt_sleep() to waitq_interrupt_sleep().
I'd prefer that this, IMO, waitq related stuff stays together. |
|
| 2089 |
6499 d 13 h |
decky |
/trunk/ |
huge type system cleanup
remove cyclical type dependencies across multiple header files
many minor coding style fixes |
|
| 2087 |
6500 d 9 h |
jermar |
/trunk/kernel/ |
Indentaion and formatting changes even Martin will like :-) |
|
| 2071 |
6510 d 6 h |
jermar |
/trunk/ |
(c) versus (C) |
|
| 2067 |
6520 d 12 h |
jermar |
/trunk/kernel/ |
More formatting changes. |
|
| 2050 |
6539 d 16 h |
decky |
/trunk/kernel/ |
benchmarking with statistics (initial) |
|
| 2042 |
6544 d 9 h |
decky |
/trunk/kernel/ |
introduce uncounted threads, whose accounting doesn't affect accumulated task accounting
run tests in kconsole thread again |
|
| 2039 |
6544 d 14 h |
decky |
/trunk/kernel/generic/ |
accumulate task accounting, run tests as separate kernel task |
|
| 2032 |
6545 d 12 h |
decky |
/trunk/kernel/generic/src/proc/ |
proper initialization of last_cycle |
|
| 2030 |
6545 d 14 h |
decky |
/trunk/kernel/ |
thread CPU cycles accounting |
|
| 1854 |
6651 d 15 h |
jermar |
/trunk/kernel/ |
Define architecture-specific thread sub-constructors and sub-destructors on all architectures.
Define the THREAD_FLAG_USPACE which means that the thread runs in user space.
The forementioned changes allow for allocating of user window buffer on sparc64
threads that execute in userspace.
A lot of formatting and indentation fixes. |
|