libs/rt-thread
libs/rt-thread
4
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

feat: arm64: mmu: auto-sensing of best paging stride

Browse Source 
Improves the memory mapping process by dynamically selecting the optimal
paging stride (4K or 2M) based on virtual address alignment and mapping
size. This eliminates the need for upfront stride determination, enhancing
flexibility and maintainability in memory management.

Changes:
- Replaced fixed stride selection logic with a dynamic decision loop.
- Removed `npages` calculation and replaced with `remaining_sz` to track
  unprocessed memory size.
- Added assertions to ensure `size` is properly aligned to the smallest
  page size.
- Adjusted loop to dynamically determine and apply the appropriate stride
  (4K or 2M) for each mapping iteration.
- Updated virtual and physical address increments to use the dynamically
  selected stride.

Signed-off-by: Shell <smokewood@qq.com>
This commit is contained in:
Shell 2024-12-02 18:19:47 +08:00 committed by Meco Man
parent 02a114985a
commit 7ff75e21fa
1 changed files with 18 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
libcpu/aarch64/common/mmu.c
View File

@ -275,27 +275,27 @@ void *rt_hw_mmu_map(rt_aspace_t aspace, void *v_addr, void *p_addr, size_t size,
int ret = -1;
void *unmap_va = v_addr;
size_t npages;
size_t remaining_sz = size;
size_t stride;
int (*mapper)(unsigned long *lv0_tbl, void *vaddr, void *paddr, unsigned long attr);
if (((rt_ubase_t)v_addr & ARCH_SECTION_MASK) || (size & ARCH_SECTION_MASK))
{
/* legacy 4k mapping */
npages = size >> ARCH_PAGE_SHIFT;
stride = ARCH_PAGE_SIZE;
mapper = _kernel_map_4K;
}
else
{
/* 2m huge page */
npages = size >> ARCH_SECTION_SHIFT;
stride = ARCH_SECTION_SIZE;
mapper = _kernel_map_2M;
}
RT_ASSERT(!(size & ARCH_PAGE_MASK));
while (npages--)
while (remaining_sz)
{
if (((rt_ubase_t)v_addr & ARCH_SECTION_MASK) || (remaining_sz < ARCH_SECTION_SIZE))
{
/* legacy 4k mapping */
stride = ARCH_PAGE_SIZE;
mapper = _kernel_map_4K;
}
else
{
/* 2m huge page */
stride = ARCH_SECTION_SIZE;
mapper = _kernel_map_2M;
}
MM_PGTBL_LOCK(aspace);
ret = mapper(aspace->page_table, v_addr, p_addr, attr);
MM_PGTBL_UNLOCK(aspace);
@ -314,6 +314,8 @@ void *rt_hw_mmu_map(rt_aspace_t aspace, void *v_addr, void *p_addr, size_t size,
}
break;
}
remaining_sz -= stride;
v_addr = (char *)v_addr + stride;
p_addr = (char *)p_addr + stride;
}