Age | Commit message (Collapse) | Author |
|
|
|
It's an ugly fix.
|
|
|
|
|
|
|
|
|
|
This reverts commit e0571ec558b13cbf400f2a0c55787238fcf3d858.
|
|
it's not work in dde yet.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
I don't know if it helps fix any bugs, but it should be a right thing to
do.
|
|
In order to avoid dead lock caused by spin_lock_irq or
spin_lock_irqsave, I remove irq disabling in them. It's really
unnecessary to do spin_lock_irq and spin_lock_irqsave any more because
interrupt isn't handled in a real interrupt context.
|
|
|
|
This is a temporary solution. I should implement a correct mechanism to
simulate cli/sti.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
eliminate a potential race condition in thread_wait and thread_wakeup.
|
|
|
|
|
|
|
|
|