x86/mm: Drop usage of __flush_tlb_all() in kernel_physical_mapping_in…

…it()

commit ba6f508 upstream.

Commit:

  f77084d "x86/mm/pat: Disable preemption around __flush_tlb_all()"

addressed a case where __flush_tlb_all() is called without preemption
being disabled. It also left a warning to catch other cases where
preemption is not disabled.

That warning triggers for the memory hotplug path which is also used for
persistent memory enabling:

 WARNING: CPU: 35 PID: 911 at ./arch/x86/include/asm/tlbflush.h:460
 RIP: 0010:__flush_tlb_all+0x1b/0x3a
 [..]
 Call Trace:
  phys_pud_init+0x29c/0x2bb
  kernel_physical_mapping_init+0xfc/0x219
  init_memory_mapping+0x1a5/0x3b0
  arch_add_memory+0x2c/0x50
  devm_memremap_pages+0x3aa/0x610
  pmem_attach_disk+0x585/0x700 [nd_pmem]

Andy wondered why a path that can sleep was using __flush_tlb_all() [1]
and Dave confirmed the expectation for TLB flush is for modifying /
invalidating existing PTE entries, but not initial population [2]. Drop
the usage of __flush_tlb_all() in phys_{p4d,pud,pmd}_init() on the
expectation that this path is only ever populating empty entries for the
linear map. Note, at linear map teardown time there is a call to the
all-cpu flush_tlb_all() to invalidate the removed mappings.

[1]: https://lkml.kernel.org/r/[email protected]
[2]: https://lkml.kernel.org/r/[email protected]

[ mingo: Minor readability edits. ]

Suggested-by: Dave Hansen <[email protected]>
Reported-by: Andy Lutomirski <[email protected]>
Signed-off-by: Dan Williams <[email protected]>
Acked-by: Peter Zijlstra (Intel) <[email protected]>
Acked-by: Kirill A. Shutemov <[email protected]>
Cc: <[email protected]>
Cc: Borislav Petkov <[email protected]>
Cc: H. Peter Anvin <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Sebastian Andrzej Siewior <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: [email protected]
Fixes: f77084d ("x86/mm/pat: Disable preemption around __flush_tlb_all()")
Link: http://lkml.kernel.org/r/154395944713.32119.15611079023837132638.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Ingo Molnar <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

arch/x86/mm/init_64.c

@@ -574,7 +574,6 @@ phys_pud_init(pud_t *pud_page, unsigned long paddr, unsigned long paddr_end,
 							   paddr_end,
 							   page_size_mask,
 							   prot);
-				__flush_tlb_all();
 				continue;
 			}
 			/*
@@ -617,7 +616,6 @@ phys_pud_init(pud_t *pud_page, unsigned long paddr, unsigned long paddr_end,
 		pud_populate(&init_mm, pud, pmd);
 		spin_unlock(&init_mm.page_table_lock);
 	}
-	__flush_tlb_all();
 
 	update_page_count(PG_LEVEL_1G, pages);
 
@@ -658,7 +656,6 @@ phys_p4d_init(p4d_t *p4d_page, unsigned long paddr, unsigned long paddr_end,
 			paddr_last = phys_pud_init(pud, paddr,
 					paddr_end,
 					page_size_mask);
-			__flush_tlb_all();
 			continue;
 		}
 
@@ -670,7 +667,6 @@ phys_p4d_init(p4d_t *p4d_page, unsigned long paddr, unsigned long paddr_end,
 		p4d_populate(&init_mm, p4d, pud);
 		spin_unlock(&init_mm.page_table_lock);
 	}
-	__flush_tlb_all();
 
 	return paddr_last;
 }
@@ -723,8 +719,6 @@ kernel_physical_mapping_init(unsigned long paddr_start,
 	if (pgd_changed)
 		sync_global_pgds(vaddr_start, vaddr_end - 1);
 
-	__flush_tlb_all();
-
 	return paddr_last;
 }