On 2012/12/7 14:43, Tang Chen wrote:

On 11/27/2012 06:00 PM, Wen Congyang wrote:

quoted

For hot removing memory, we sholud remove page table about the memory.
So the patch searches a page table about the removed memory, and clear
page table.

(snip)

quoted

+void __meminit
+kernel_physical_mapping_remove(unsigned long start, unsigned long end)
+{
+    unsigned long next;
+    bool pgd_changed = false;
+
+    start = (unsigned long)__va(start);
+    end = (unsigned long)__va(end);

Hi Wu,

Here, you expect start and end are physical addresses. But in
phys_xxx_remove() function, I think using virtual addresses is just
fine. Functions like pmd_addr_end() and pud_index() only calculate
an offset.

Hi Tang,

 

Virtual addresses will work fine, I used physical addresses in order to
keep consistent with phys_pud[pmd/pte]_init(), So I think we should keep this.

Thanks,
Jianguo Wu

So, would you please tell me if we have to use physical addresses here ?

Thanks. :)

quoted

+
+    for (; start<  end; start = next) {
+        pgd_t *pgd = pgd_offset_k(start);
+        pud_t *pud;
+
+        next = pgd_addr_end(start, end);
+
+        if (!pgd_present(*pgd))
+            continue;
+
+        pud = map_low_page((pud_t *)pgd_page_vaddr(*pgd));
+        phys_pud_remove(pud, __pa(start), __pa(next));
+        if (free_pud_table(pud, pgd))
+            pgd_changed = true;
+        unmap_low_page(pud);
+    }
+
+    if (pgd_changed)
+        sync_global_pgds(start, end - 1);
+
+    flush_tlb_all();
+}
+
  #ifdef CONFIG_MEMORY_HOTREMOVE
  int __ref arch_remove_memory(u64 start, u64 size)
  {

@@ -692,6 +921,8 @@ int __ref arch_remove_memory(u64 start, u64 size)
      ret = __remove_pages(zone, start_pfn, nr_pages);
      WARN_ON_ONCE(ret);

+    kernel_physical_mapping_remove(start, start + size);
+
      return ret;
  }
  #endif

.