|
|
@d23712e
|
9 months |
zarevucky.jiri |
Use thread state variable instead of a cpu local variable to pass state
|
|
|
@dfa4be62
|
9 months |
zarevucky.jiri |
Thread lock is no longer necessary
|
|
|
@3fa4e22a
|
9 months |
zarevucky.jiri |
Only do fpu_cleanup() once we are sure we are switching
|
|
|
@11909ce3
|
9 months |
zarevucky.jiri |
Make thread cycle statistics atomic
|
|
|
@33e15a0
|
9 months |
zarevucky.jiri |
Make thread→btrace weakly atomic, to avoid need for thread lock
|
|
|
@7364e2d1
|
9 months |
zarevucky.jiri |
Don't try to determine internal thread state when requesting stack …
|
|
|
@41bfc64
|
9 months |
zarevucky.jiri |
Make thread→state weakly atomic so we don't need to hold thread lock
|
|
|
@3d84734
|
9 months |
zarevucky.jiri |
Make thread→priority weakly atomic to avoid need for locking
|
|
|
@efed95a3
|
9 months |
zarevucky.jiri |
Make thread→cpu weakly atomic, to avoid need for thread lock
|
|
|
@286da52
|
9 months |
zarevucky.jiri |
Streamline requeuing threads
Split thread_ready() into different …
|
|
|
@6a0e568
|
9 months |
zarevucky.jiri |
Allow fast-switch to another runnable thread
When a thread is …
|
|
|
@c1eaec4
|
9 months |
zarevucky.jiri |
Shuffle some locks around
|
|
|
@25939997
|
9 months |
zarevucky.jiri |
Make separate-stack-scheduler a loop with persistent context
We can …
|
|
|
@29029ac0
|
9 months |
zarevucky.jiri |
Call after_thread_ran_arch() before context switch
This is only used …
|
|
|
@ed7e057
|
9 months |
zarevucky.jiri |
Add functions context_create(), context_replace() and context_swap()
…
|
|
|
@4ed7870
|
9 months |
zarevucky.jiri |
Revert part of commit 5861b602
|
|
|
@a064d4f
|
9 months |
zarevucky.jiri |
Make thread_join() imply thread_put()
This makes the function more …
|
|
|
@5861b60
|
9 months |
zarevucky.jiri |
Lift actions that do not need separate context out of …
|
|
|
@151c050
|
9 months |
zarevucky.jiri |
Rethink scheduler entry points
Changes the way scheduler is entered, …
|
|
|
@8996582
|
9 months |
zarevucky.jiri |
Move context switch preparation to a new separate function
This puts …
|
|
|
@6e49dab
|
9 months |
zarevucky.jiri |
Extract post-thread cleanup into a separate function
|
|
|
@23f36a3
|
9 months |
zarevucky.jiri |
Wrap fpu handling code in named functions
|
|
|
@5663872
|
9 months |
zarevucky.jiri |
Move stuff around for thread sleep
Only mark the thread as ready for …
|
|
|
@4760793
|
9 months |
zarevucky.jiri |
Add CPU_LOCAL alongside CPU and segregate fields that are only used …
|
|
|
@3b68542
|
9 months |
zarevucky.jiri |
Eliminate an undefined behavior in scheduler debug printout
The …
|
|
|
@ec8ef12
|
18 months |
zarevucky.jiri |
Split find_best_thread() into two functions
try_find_thread() …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@117ad5a2
|
18 months |
zarevucky.jiri |
Get thread priority from find_best_thread(), instead of locking thread …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@f3dbe27
|
18 months |
zarevucky.jiri |
Reduce locking further with lazy FPU
It turns out we only need a lock …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@06f81c4
|
18 months |
zarevucky.jiri |
Check cpu_t::fpu_owner directly instead of …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@fbaf6ac
|
18 months |
zarevucky.jiri |
Streamline thread load-balancing
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@169815e
|
18 months |
zarevucky.jiri |
Split cpu_t::lock into fpu_lock and tlb_lock
For all other purposes, …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@dd218ea
|
18 months |
zarevucky.jiri |
Remove unnecessary thread_t::wired
There's already …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@c680333
|
18 months |
zarevucky.jiri |
Move task switch handling into a separate function
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@3118355
|
19 months |
zarevucky.jiri |
Optimize relink_rq()
Iterate backwards and only lock each runqueue …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@bea6233
|
20 months |
zarevucky.jiri |
Replace cpu_sleep() with cpu_interruptible_sleep()
The new function …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@111b9b9
|
20 months |
zarevucky.jiri |
Reimplement waitq using thread_wait/wakeup
This adds a few functions …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@1871118
|
20 months |
zarevucky.jiri |
Make thread_t reference counted
This simplifies interaction between …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@daadfa6
|
20 months |
zarevucky.jiri |
Remove unnecessary context field from cpu_t structure
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@0366d09d
|
20 months |
zarevucky.jiri |
Avoid separate allocation for FPU context
With _Alignas, we can …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@c030818
|
20 months |
zarevucky.jiri |
Move saved interrupt level from context_t to thread_t
Unlike the rest …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@011c79a
|
20 months |
zarevucky.jiri |
Replace CPU→needs_relink with CPU→relink_deadline
This removes a …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@aae2869
|
20 months |
zarevucky.jiri |
Replace THREAD→ticks with CPU→preempt_deadline
This removes a bit …
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|
|
|
@31e15be
|
3 years |
5887334+jxsvoboda |
kernel: deprecate atomic_t
serial
ticket/834-toolchain-update
topic/msim-upgrade
topic/simplify-dev-export
|