ARM MMU page table framework

Back to the alloc_init_pte function, line 605 sets the page table entry (i.e. pte) located by addr in the L2 page table. Its main task is to fill in the physical address of the corresponding physical page for the mmu hardware to implement address translation.

Line 604~line 607 loops through filling two hwpte page tables and completes the creation of a 2M physical memory mapping table.

Line 608 will eventually call the following function: static inline void __pmd_populate(pmd_t *pmdp, phys_addr_t pte, pmdval_t prot)

Before executing this function, two L2 page tables have been created. The function is to set the corresponding table entries of the L1 page table to point to the two newly created L2 page tables (hwpte0, hwpte1). As mentioned earlier, since the L1 page table entry of Linux is 8 bytes in size, so:

Line133 points the first 4 bytes to the hwpte0 page table.

Line135 points the last 4 bytes to the hwpte1 page table. So far, the association of L1---〉L2 page table has been established.

Line 137 refreshes the TLB buffer so that the system's CPU can see the changes in the mapping

At this point, the mapping relationship between the virtual address and the physical address specified by the struct map_desc *md structure has been established, so that the hardware MMU can automatically translate the virtual address to the physical address.

In the above process, some details are selectively omitted due to space limitations. If you understand the process, you can read the relevant code in detail. For example, the set_pte_ext function above will call the assembly function to implement the setting of the pte table entry.