The
mi_switch() function implements the machine-independent prelude to an LWP context switch. It is called from only a few distinguished places in the kernel code as a result of the principle of non-preemptable kernel mode execution. The three major uses of
mi_switch() can be enumerated as follows:
1.
From within
cv_wait(9) and associated methods when the current LWP voluntarily relinquishes the CPU to wait for some resource to become available.
2.
From within
preempt(9) when the current LWP voluntarily relinquishes the CPU or when the kernel prepares a return to user-mode execution.
3.
In the signal handling code if a signal is delivered that causes an LWP to stop (see
issignal(9)).
mi_switch() records the amount of time the current LWP has been running in the LWP structure and checks this value against the CPU time limits allocated to the LWP (see
getrlimit(2)). Exceeding the soft limit results in a
SIGXCPU signal to be posted to the LWP, while exceeding the hard limit will cause a
SIGKILL.
Unless
l->l_switchto is not
NULL,
mi_switch() will call
sched_nextlwp() to select a new LWP from the scheduler's runqueue structures. If no runnable LWP is found, the idle LWP is used. If the new LWP is not equal to the current one,
mi_switch() will hand over control to the machine-dependent function
cpu_switchto(9) to switch to the new LWP.
mi_switch() has to be called with the LWP lock held (through calling
lwp_lock() first) and at the
splsched(9) interrupt protection level. It returns with the LWP lock released.