arm64平台进程全局页表pgd的创建和切换

进程创建pgd页表

分配每个进程的pgd。

mm_alloc：fs/exec.c:    bprm->mm = mm = mm_alloc(); //exec加载新二进制程序
    mm_init ：kernel/fork.c: 
        mm_alloc_pgd
            pgd_alloc

pgd_alloc函数: arch/arm64/mm/pgd.c

pgd_t *pgd_alloc(struct mm_struct *mm)
{
        if (PGD_SIZE == PAGE_SIZE)
                return (pgd_t *)__get_free_page(PGALLOC_GFP);
        else
                return kmem_cache_alloc(pgd_cache, PGALLOC_GFP);
}

#define PGD_SIZE        (PTRS_PER_PGD * sizeof(pgd_t))
上述PGD_SIZE在PA_BITS为48bit的机器上:PTRS_PER_PGD 64 *  sizeof(pgd_t) 8  =  512。因此会用kmem_cache_create接口slub来分配小块内存。

返回pgd表虚拟地址。

pgd_cache 的kmeme_cache对象初始化在arch/arm64/mm/pgd.c#L31

void __init pgd_cache_init(void)
{
	if (PGD_SIZE == PAGE_SIZE)
		return;

#ifdef CONFIG_ARM64_PA_BITS_52
	/*
	 * With 52-bit physical addresses, the architecture requires the
	 * top-level table to be aligned to at least 64 bytes.
	 */
	BUILD_BUG_ON(PGD_SIZE < 64);
#endif

	/*
	 * Naturally aligned pgds required by the architecture.
	 */
	pgd_cache = kmem_cache_create("pgd_cache", PGD_SIZE, PGD_SIZE,
				      SLAB_PANIC, NULL);
}

因此在该arm64平台下分配的pgd表512字节。其中有64个pgd_t 项。

pgd切换

static __always_inline struct rq *
context_switch(struct rq *rq, struct task_struct *prev,
	       struct task_struct *next, struct rq_flags *rf)
{

	if (!next->mm) {                                // to kernel
		enter_lazy_tlb(prev->active_mm, next);

		next->active_mm = prev->active_mm;
		if (prev->mm)                           // from user
			mmgrab(prev->active_mm);
		else
			prev->active_mm = NULL;
	} else {                                        // to user
		membarrier_switch_mm(rq, prev->active_mm, next->mm);
		/*
		 * sys_membarrier() requires an smp_mb() between setting
		 * rq->curr / membarrier_switch_mm() and returning to userspace.
		 *
		 * The below provides this either through switch_mm(), or in
		 * case 'prev->active_mm == next->mm' through
		 * finish_task_switch()'s mmdrop().
		 */
		switch_mm_irqs_off(prev->active_mm, next->mm, next);

		if (!prev->mm) {                        // from kernel
			/* will mmdrop() in finish_task_switch(). */
			rq->prev_mm = prev->active_mm;
			prev->active_mm = NULL;
		}
	}

	/* Here we just switch the register state and the stack. */
	switch_to(prev, next, prev);
	barrier();

	return finish_task_switch(prev);
}

__schedule
    context_switch
        0: to kernel
            ....
        1: to user //切换用户进程需要切换内存
            switch_mm_irqs_off(=switch_mm) arg:next  
                __switch_mm
                    check_and_switch_context
                        cpu_switch_mm
                            static inline void cpu_switch_mm(pgd_t *pgd, struct mm_struct *mm)
                            {
                                BUG_ON(pgd == swapper_pg_dir);
                                cpu_set_reserved_ttbr0();
                                cpu_do_switch_mm(virt_to_phys(pgd),mm);
                            }
    switch_to(prev, next, prev);
    finish_task_switch(prev);

ENTRY(cpu_do_switch_mm)
        mrs     x2, ttbr1_el1      //保存最高位为1(内核地址空间)的ttbr1_el1寄存器
        mmid    x1, x1                          // get mm->context.id
        phys_to_ttbr x3, x0
#ifdef CONFIG_ARM64_SW_TTBR0_PAN
        bfi     x3, x1, #48, #16                // set the ASID field in TTBR0
#endif
        bfi     x2, x1, #48, #16                // set the ASID
        msr     ttbr1_el1, x2                   // in TTBR1 (since TCR.A1 is set)
        isb
        msr     ttbr0_el1, x3                   // now update TTBR0  更新用户空间ttbr0寄存器
        isb
        b       post_ttbr_update_workaround     // Back to C code...
ENDPROC(cpu_do_switch_mm)

内核线程和用户进程不同点在于内核线程的mm