x86/mm: Split vmalloc_sync_all()

Commit Message

Joerg Roedel Oct. 7, 2019, 3:16 p.m. UTC

From: Joerg Roedel <[e-mail]>

Git commit 3f8fd02b1bf1 ("mm/vmalloc: Sync unmappings in
__purge_vmap_area_lazy()") introduced a call to vmalloc_sync_all() in
the vunmap() code-path.  While this change was necessary to maintain
correctness on x86-32-pae kernels, it also adds additional cycles for
architectures that don't need it.

Specifically on x86-64 with CONFIG_VMAP_STACK=y some people reported
severe performance regressions in micro-benchmarks because it now also
calls the x86-64 implementation of vmalloc_sync_all() on vunmap(). But
the vmalloc_sync_all() implementation on x86-64 is only needed for
newly created mappings.

To avoid the unnecessary work on x86-64 and to gain the performance
back, split up vmalloc_sync_all() into two functions:

	* vmalloc_sync_mappings(), and
	* vmalloc_sync_unmappings()

Most call-sites to vmalloc_sync_all() only care about new mappings
being synchronized. The only exception is the new call-site
added in the above mentioned commit.

Reported-by: kernel test robot <[e-mail]>
Fixes: 3f8fd02b1bf1 ("mm/vmalloc: Sync unmappings in __purge_vmap_area_lazy()")
Signed-off-by: Joerg Roedel <[e-mail]>
---
 arch/x86/mm/fault.c      | 14 ++++++++++++--
 drivers/acpi/apei/ghes.c |  2 +-
 include/linux/vmalloc.h  |  5 +++--
 kernel/notifier.c        |  2 +-
 mm/nommu.c               | 10 +++++++---
 mm/vmalloc.c             | 11 +++++++----
 6 files changed, 31 insertions(+), 13 deletions(-)

Comments

Dave Hansen Oct. 7, 2019, 3:30 p.m. UTC | #1

On 10/7/19 8:16 AM, Joerg Roedel wrote:
> @@ -318,7 +328,7 @@ static void dump_pagetable(unsigned long address)
>  
>  #else /* CONFIG_X86_64: */
>  
> -void vmalloc_sync_all(void)
> +void vmalloc_sync_mappings(void)
>  {
>  	sync_global_pgds(VMALLOC_START & PGDIR_MASK, VMALLOC_END);
>  }

FWIW, I generally detest the use of __weak. :)

In this case, it ends up letting us gloss over the fact that we have a
32/64-bit asymmetry.  It would probably be nice to actually have a
64-bit implementation that comes along with a nice comment.  Maybe this
in vmalloc_sync_mappings():

	/*
	 * 64-bit mappings might allocate new p4d/pud pages
	 * that need to be propagated to all tasks' PGDs.
	 */

which would pair nicely with:

void vmalloc_sync_unmappings(void)
{
	/*
	 * Unmappings never allocate or free p4d/pud pages.
	 * No work is required here.
	 */
}

Joerg Roedel Oct. 7, 2019, 3:48 p.m. UTC | #2

Hi Dave,

thanks for your review!

On Mon, Oct 07, 2019 at 08:30:51AM -0700, Dave Hansen wrote:
> On 10/7/19 8:16 AM, Joerg Roedel wrote:
> > @@ -318,7 +328,7 @@ static void dump_pagetable(unsigned long address)
> >  
> >  #else /* CONFIG_X86_64: */
> >  
> > -void vmalloc_sync_all(void)
> > +void vmalloc_sync_mappings(void)
> >  {
> >  	sync_global_pgds(VMALLOC_START & PGDIR_MASK, VMALLOC_END);
> >  }
> 
> FWIW, I generally detest the use of __weak. :)

Yeah, I don't like it either, but in this case it is probably better
than empty stubs in all architectures besides x86 :)

> In this case, it ends up letting us gloss over the fact that we have a
> 32/64-bit asymmetry.  It would probably be nice to actually have a
> 64-bit implementation that comes along with a nice comment.  Maybe this
> in vmalloc_sync_mappings():
> 
> 	/*
> 	 * 64-bit mappings might allocate new p4d/pud pages
> 	 * that need to be propagated to all tasks' PGDs.
> 	 */
> 
> which would pair nicely with:
> 
> void vmalloc_sync_unmappings(void)
> {
> 	/*
> 	 * Unmappings never allocate or free p4d/pud pages.
> 	 * No work is required here.
> 	 */
> }

Yes, that makes, I will add these comments in V2.

Thanks,

	Joerg

Patch

diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 9ceacd1156db..889c87ecc751 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -189,7 +189,7 @@  static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address)
 	return pmd_k;
 }
 
-void vmalloc_sync_all(void)
+static void vmalloc_sync(void)
 {
 	unsigned long address;
 
@@ -216,6 +216,16 @@  void vmalloc_sync_all(void)
 	}
 }
 
+void vmalloc_sync_mappings(void)
+{
+	vmalloc_sync();
+}
+
+void vmalloc_sync_unmappings(void)
+{
+	vmalloc_sync();
+}
+
 /*
  * 32-bit:
  *
@@ -318,7 +328,7 @@  static void dump_pagetable(unsigned long address)
 
 #else /* CONFIG_X86_64: */
 
-void vmalloc_sync_all(void)
+void vmalloc_sync_mappings(void)
 {
 	sync_global_pgds(VMALLOC_START & PGDIR_MASK, VMALLOC_END);
 }
diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
index 777f6f7122b4..e0d82fab1f44 100644
--- a/drivers/acpi/apei/ghes.c
+++ b/drivers/acpi/apei/ghes.c
@@ -171,7 +171,7 @@  int ghes_estatus_pool_init(int num_ghes)
 	 * New allocation must be visible in all pgd before it can be found by
 	 * an NMI allocating from the pool.
 	 */
-	vmalloc_sync_all();
+	vmalloc_sync_mappings();
 
 	rc = gen_pool_add(ghes_estatus_pool, addr, PAGE_ALIGN(len), -1);
 	if (rc)
diff --git a/include/linux/vmalloc.h b/include/linux/vmalloc.h
index 4e7809408073..decac0790fc1 100644
--- a/include/linux/vmalloc.h
+++ b/include/linux/vmalloc.h
@@ -126,8 +126,9 @@  extern int remap_vmalloc_range_partial(struct vm_area_struct *vma,
 
 extern int remap_vmalloc_range(struct vm_area_struct *vma, void *addr,
 							unsigned long pgoff);
-void vmalloc_sync_all(void);
- 
+void vmalloc_sync_mappings(void);
+void vmalloc_sync_unmappings(void);
+
 /*
  *	Lowlevel-APIs (not for driver use!)
  */
diff --git a/kernel/notifier.c b/kernel/notifier.c
index d9f5081d578d..157d7c29f720 100644
--- a/kernel/notifier.c
+++ b/kernel/notifier.c
@@ -554,7 +554,7 @@  NOKPROBE_SYMBOL(notify_die);
 
 int register_die_notifier(struct notifier_block *nb)
 {
-	vmalloc_sync_all();
+	vmalloc_sync_mappings();
 	return atomic_notifier_chain_register(&die_chain, nb);
 }
 EXPORT_SYMBOL_GPL(register_die_notifier);
diff --git a/mm/nommu.c b/mm/nommu.c
index 99b7ec318824..3b67bd20c2af 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -359,10 +359,14 @@  void vm_unmap_aliases(void)
 EXPORT_SYMBOL_GPL(vm_unmap_aliases);
 
 /*
- * Implement a stub for vmalloc_sync_all() if the architecture chose not to
- * have one.
+ * Implement a stub for vmalloc_sync_[un]mapping() if the architecture
+ * chose not to have one.
  */
-void __weak vmalloc_sync_all(void)
+void __weak vmalloc_sync_mappings(void)
+{
+}
+
+void __weak vmalloc_sync_unmappings(void)
 {
 }
 
diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index a3c70e275f4e..c0be707db434 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -1259,7 +1259,7 @@  static bool __purge_vmap_area_lazy(unsigned long start, unsigned long end)
 	 * First make sure the mappings are removed from all page-tables
 	 * before they are freed.
 	 */
-	vmalloc_sync_all();
+	vmalloc_sync_unmappings();
 
 	/*
 	 * TODO: to calculate a flush range without looping.
@@ -3050,16 +3050,19 @@  int remap_vmalloc_range(struct vm_area_struct *vma, void *addr,
 EXPORT_SYMBOL(remap_vmalloc_range);
 
 /*
- * Implement a stub for vmalloc_sync_all() if the architecture chose not to
- * have one.
+ * Implement stubs for vmalloc_sync_[un]mappings () if the architecture chose
+ * not to have one.
  *
  * The purpose of this function is to make sure the vmalloc area
  * mappings are identical in all page-tables in the system.
  */
-void __weak vmalloc_sync_all(void)
+void __weak vmalloc_sync_mappings(void)
 {
 }
 
+void __weak vmalloc_sync_unmappings(void)
+{
+}
 
 static int f(pte_t *pte, unsigned long addr, void *data)
 {