Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

调整pcb的sched_info和rwlock,以避免调度器死锁问题 #341

Conversation

Chiichen
Copy link
Member

No description provided.

@@ -496,7 +520,10 @@ impl<'rwlock, T> Drop for RwLockWriteGuard<'rwlock, T> {
self.inner
.lock
.fetch_and(!(WRITER | UPGRADED), Ordering::Release);

if let Some(_) = self.irq_guard {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

使用option的take方法即可

}

pub fn state(&self) -> ProcessState {
return self.state;
}

fn set_state(&mut self, state: ProcessState) {
pub fn set_state(&mut self, state: ProcessState) {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

这里还没改哈哈


if pcb.sched_info().state() != ProcessState::Runnable {
pcb.sched_info_mut_irqsave()
.set_state(ProcessState::Runnable);
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  1. 这里的代码移动到Processmanager里面的wakeup
  2. 这里有问题,更改state的工作,不应该由wakeup函数负责,而是在sched_enqueue里面改才会更合理

然后的话,获得写锁之后一定要再次确认它的state是否真的不为Runnable。
因为内核是处处存在并发的,很有可能,756、757两行之间,另外一个cpu加上了锁,然后改了state。只有写锁才是排它的。因此需要二次确认

@fslongjin fslongjin merged commit a365281 into DragonOS-Community:patch-refactor-process-management Aug 26, 2023
2 checks passed
@Chiichen Chiichen deleted the refactor-process-deadlock-avoid branch August 27, 2023 02:59
fslongjin added a commit that referenced this pull request Sep 15, 2023
* 添加新版pcb的数据结构 (#273)

* 将pcb中的内容分类,分别加锁 (#305)

* 进程管理重构:完成fork的主体逻辑 (#309)

1.完成fork的主体逻辑
2.将文件系统接到新的pcb上
3.经过思考,暂时弃用signal机制,待进程管理重构完成后,重写signal机制.原因是原本的signal机制太烂了

* chdir getcwd pid pgid ppid (#310)


---------

Co-authored-by: longjin <[email protected]>

* 删除旧的fork以及signal的代码,并调整fork/vfork/execve系统调用 (#325)

1.删除旧的fork
2.删除signal相关代码,等进程管理重构结束之后,再重新写.
3.调整了fork/vfork/execve系统调用

* 实现切换进程的代码 (#331)



* 实现切换进程的代码

* Patch modify preempt (#332)

* 修改设置preempt的代码

* 删除rust的list和refcount

* 为每个核心初始化idle进程 (#333)

* 为每个核心初始化idle进程

* 完成了新的内核线程机制 (#335)

* 调度器的pcb替换为新的Arc<ProcessControlBlock>,把调度器队列锁从 RwSpinLock 替换为了 SpinLock (#336)

* 把调度器的pcb替换为新的Arc<ProcessControlBlock>

* 把调度器队列锁从 RwSpinLock 替换为了 SpinLock ,修改了签名以通过编译

* 修正一些双重加锁、细节问题

---------

Co-authored-by: longjin <[email protected]>

* github workflow自动检查代码是否格式化

* cache toolchain yml

* 调整rust版本的waitqueue中的pcb为新版的pcb (#343)

* 解决设置rust workspace带来的“工具链不一致”的问题 (#344)


* 解决设置rust workspace带来的“工具链不一致”的问题

更改workflow

* 调整pcb的sched_info和rwlock,以避免调度器死锁问题 (#341)

* 调整pcb的sched_info和rwlock,以避免调度器死锁问题

* 修改为在 WriterGuard 中维护 Irq_guard

* 修正了 write_irqsave方法

* 优化了代码

* 把 set state 操作从 wakup 移动到 sched_enqueue 中

* 修正为在 wakeup 中设置 running ,以保留 set_state 的私有性

* 移除了 process_wakeup

* 实现进程退出的逻辑 (#340)

实现进程退出的逻辑

* 标志进程sleep

* 修复wakeup的问题

---------

Co-authored-by: longjin <[email protected]>

* rust 重构 completion (#350)

* 完成了completion的基本结构,待完善上级调用

* 用SpinLock保护结构体并发安全

* 修改原子变量为u32,修复符号错误

* irq guard

* 修改为具有内部可变性的结构体

* temp fix

* 修复了由于进程持有自旋锁导致的不被调度的问题

* 对 complete 系列方法上锁,保护 done 数据并发安全

* 移除了未使用的依赖

* 重写显示刷新驱动 (#363)

* 重构显示刷新驱动

* Patch refactor process management (#366)

* 维护进程树

* 维护进程树

* 更改代码结构

* 新建进程时,设置cwd

* 调整adopt childern函数,降低开销

---------

Co-authored-by: longjin <[email protected]>

* waitqueue兼容C部分 (#351)

* PATH

* safe init

* waitqueue兼容C部分

* waitqueue兼容C部分

* 删除semaphore.c,在ps2_keyboard中使用waitqueue

* 删除semaphore.c,在ps2_keyboard中使用waitqueue

* current_pcb的C兼容

* current_pcb的C兼容

* current_pcb的C兼容

* fmt

* current_pcb的兼容

* 针对修改

* 调整代码

* fmt

* 删除pcb的set flags

* 更改函数名

---------

Co-authored-by: longjin <[email protected]>

* merge master

* Patch debug process management refactor (#372)

* 能够调通,执行完textui_init

* 能跑到initial kernel thread

* fmt

* 能够正常初始化所有服务(尚未能切换到用户程序)

* 删除部分无用的extern

* 存在问题:ap处理器启动后,bsp的smp_init函数return之后就出错了,怀疑是栈损坏

* 解决smp启动由于未换栈导致的内存访问错误

* debug

* 1

* 1

* lock no preempt

* 调通

* 优化代码,删除一些调试日志

* fix

* 使用rust重写wait4 (#377)

* 维护进程树

* 维护进程树

* 更改代码结构

* 新建进程时,设置cwd

* 调整adopt childern函数,降低开销

* wait4

* 删除c_sys_wait4

* 使用userbuffer保护裸指针

---------

Co-authored-by: longjin <[email protected]>

* 消除warning

* 1. 修正未设置cpu executing的问题

* 修正kthread机制可能存在的内存泄露问题

* 删除pcb文档

* 删除C的tss struct

---------

Co-authored-by: Bullet <[email protected]>
Co-authored-by: Chiichen <[email protected]>
Co-authored-by: hanjiezhou <[email protected]>
Co-authored-by: GnoCiYeH <[email protected]>
Co-authored-by: houmkh <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants