[PATCH V8] mm/debug: Add tests validating architecture page table helpers

[Anshuman Khandual]

On 11/06/2019 04:00 AM, Vineet Gupta wrote:
> Hi Anshuman,

Hello Vineet,

> 
> On 11/4/19 4:00 PM, Anshuman Khandual wrote:
>> On 10/28/2019 10:59 AM, Anshuman Khandual wrote:
>>> This adds tests which will validate architecture page table helpers and
>>> other accessors in their compliance with expected generic MM semantics.
>>> This will help various architectures in validating changes to existing
>>> page table helpers or addition of new ones.
>>>
>>> This test covers basic page table entry transformations including but not
>>> limited to old, young, dirty, clean, write, write protect etc at various
>>> level along with populating intermediate entries with next page table page
>>> and validating them.
>>>
>>> Test page table pages are allocated from system memory with required size
>>> and alignments. The mapped pfns at page table levels are derived from a
>>> real pfn representing a valid kernel text symbol. This test gets called
>>> right after page_alloc_init_late().
>>>
>>> This gets build and run when CONFIG_DEBUG_VM_PGTABLE is selected along with
>>> CONFIG_VM_DEBUG. Architectures willing to subscribe this test also need to
>>> select CONFIG_ARCH_HAS_DEBUG_VM_PGTABLE which for now is limited to x86 and
>>> arm64. Going forward, other architectures too can enable this after fixing
>>> build or runtime problems (if any) with their page table helpers.
>>>
>>> Folks interested in making sure that a given platform's page table helpers
>>> conform to expected generic MM semantics should enable the above config
>>> which will just trigger this test during boot. Any non conformity here will
>>> be reported as an warning which would need to be fixed. This test will help
>>> catch any changes to the agreed upon semantics expected from generic MM and
>>> enable platforms to accommodate it thereafter.
>>>
> 
> I tried enabling this on ARC and ran into a build issue
> 
> ../mm/debug_vm_pgtable.c: In function ‘pmd_basic_tests’:
> ../mm/debug_vm_pgtable.c:73:14: error: implicit declaration of function ‘pfn_pmd’;
> did you mean ‘pfn_pte’? [-Werror=implicit-function-declaration]
>   pmd_t pmd = pfn_pmd(pfn, prot);
>               ^~~~~~~
> 
> The reason being THP was not enabled (although ARC supports THP) - for the
> combination below
> CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE=y
> # CONFIG_TRANSPARENT_HUGEPAGE is not set
> 
> I think you need to use latter for guarding pmd_basic_tests()

So the build complains that pfn_pmd() is not defined when the following config
combination is in place.

CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE=y
CONFIG_TRANSPARENT_HUGEPAGE=n

But should not pfn_pmd() be encapsulated inside HAVE_ARCH_TRANSPARENT_HUGEPAGE
at the minimum (but I would say it should be available always, nonetheless) when
the platform subscribes to THP irrespective of whether THP is enabled or not.

I could see in the file (arch/arc/include/asm/pgtable.h) that fetching pfn_pmd()
and all other basic PMD definitions is conditional on CONFIG_TRANSPARENT_HUGEPAGE.

#ifdef CONFIG_TRANSPARENT_HUGEPAGE
#include <asm/hugepage.h>
#endif

IIUC, CONFIG_TRANSPARENT_HUGEPAGE should only encapsulate PMD page table helpers
which are expected from generic THP code (pmd_trans_huge, pmdp_set_access_flags
etc) but not the basic PMD helpers like pmd_pfn, pmd_mkyoung, pmd_mkdirty,
pmd_mkclean etc. Hence wondering will it be possible to accommodate following
code change on arc platform (not even compiled) in order to fix the problem ?

diff --git a/arch/arc/include/asm/hugepage.h b/arch/arc/include/asm/hugepage.h
index 9a74ce7..2ae15a8 100644
--- a/arch/arc/include/asm/hugepage.h
+++ b/arch/arc/include/asm/hugepage.h
@@ -36,11 +36,11 @@ static inline pmd_t pte_pmd(pte_t pte)
 #define pmd_dirty(pmd)         pte_dirty(pmd_pte(pmd))
 
 #define mk_pmd(page, prot)     pte_pmd(mk_pte(page, prot))
+#define pfn_pmd(pfn, prot)     (__pmd(((pfn) << PAGE_SHIFT) | pgprot_val(prot)))
 
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
 #define pmd_trans_huge(pmd)    (pmd_val(pmd) & _PAGE_HW_SZ)
 
-#define pfn_pmd(pfn, prot)     (__pmd(((pfn) << PAGE_SHIFT) | pgprot_val(prot)))
-
 static inline pmd_t pmd_modify(pmd_t pmd, pgprot_t newprot)
 {
         /*
@@ -73,5 +73,6 @@ extern void flush_pmd_tlb_range(struct vm_area_struct *vma, unsigned long start,
 
 /* We don't have hardware dirty/accessed bits, generic_pmdp_establish is fine.*/
 #define pmdp_establish generic_pmdp_establish
+#endif
 
 #endif
diff --git a/arch/arc/include/asm/pgtable.h b/arch/arc/include/asm/pgtable.h
index 9019ed9..20395f1 100644
--- a/arch/arc/include/asm/pgtable.h
+++ b/arch/arc/include/asm/pgtable.h
@@ -385,7 +385,7 @@ void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
  * remap a physical page `pfn' of size `size' with page protection `prot'
  * into virtual address `from'
  */
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+#ifdef CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE
 #include <asm/hugepage.h>
 #endif

> 
> Other than that the tests pass for !THP and THP too. So once fixed, you could

Glad that it works on arc platform as well.

> enable that for ARC as well> Thx for doing this.
> 
> -Vineet
>