Skip to contentSkip to conversation

x86-64: adjust pmd_bad()

PATCH In review Jan Beulich 3 participants 5 messages Nov 15, 2006

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86-64: adjust pmd_bad()

Make pmd_bad() symmetrical to pgd_bad() and pud_bad(). At once, simplify them all.

include/asm-x86_64/pgtable.h +10 -12
@@ -221,20 +221,19 @@ static inline pte_t ptep_get_and_clear_f
 #define __S110	PAGE_SHARED_EXEC
 #define __S111	PAGE_SHARED_EXEC
 
-static inline unsigned long pgd_bad(pgd_t pgd) 
-{ 
-       unsigned long val = pgd_val(pgd);
-       val &= ~PTE_MASK; 
-       val &= ~(_PAGE_USER | _PAGE_DIRTY); 
-       return val & ~(_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED);      
-} 
+static inline unsigned long pgd_bad(pgd_t pgd)
+{
+	return pgd_val(pgd) & ~(PTE_MASK | _KERNPG_TABLE | _PAGE_USER);
+}
 
 static inline unsigned long pud_bad(pud_t pud)
 {
-       unsigned long val = pud_val(pud);
-       val &= ~PTE_MASK;
-       val &= ~(_PAGE_USER | _PAGE_DIRTY);
-       return val & ~(_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED);
+	return pud_val(pud) & ~(PTE_MASK | _KERNPG_TABLE | _PAGE_USER);
+}
+
+static inline unsigned long pmd_bad(pmd_t pmd)
+{
+	return pmd_val(pmd) & ~(PTE_MASK | _KERNPG_TABLE | _PAGE_USER);
 }
 
 #define pte_none(x)	(!pte_val(x))
@@ -347,7 +346,6 @@ static inline int pmd_large(pmd_t pte) {
 #define pmd_none(x)	(!pmd_val(x))
 #define pmd_present(x)	(pmd_val(x) & _PAGE_PRESENT)
 #define pmd_clear(xp)	do { set_pmd(xp, __pmd(0)); } while (0)
-#define	pmd_bad(x)	((pmd_val(x) & (~PTE_MASK & ~_PAGE_USER)) != _KERNPG_TABLE )
 #define pfn_pmd(nr,prot) (__pmd(((nr) << PAGE_SHIFT) | pgprot_val(prot)))
 #define pmd_pfn(x)  ((pmd_val(x) & __PHYSICAL_MASK) >> PAGE_SHIFT)
 
#1 lore raw
Quoted 2 lines · Jan Beulich

Make pmd_bad() symmetrical to pgd_bad() and pud_bad(). At once, simplify them all.

Got them all, thanks. Including the two i386 patches on l-k.

I didn't include the ACPI warning patch -- that should go via Len Brown's tree. Perhaps you resend it to him/linux-acpi to make sure he gets it.

-Andi

#2 lore raw
In reply to Jan Beulich
Quoted 2 lines · Jan Beulich

Make pmd_bad() symmetrical to pgd_bad() and pud_bad(). At once, simplify them all.

Symmetrical and simpler, yes, but you're weakening the pmd_bad() test:
no longer requires that all those _KERNPG_TABLE bits be set. Wouldn't it be better to go the other way and strengthen pgd_bad, pud_bad?

Hugh

#3 lore raw
Quoted 9 lines
Quoted 1 line
Quoted 1 line

Hugh Dickins <[email protected]> 15.11.06 18:01 >>>

Quoted 2 lines · Jan Beulich

Make pmd_bad() symmetrical to pgd_bad() and pud_bad(). At once, simplify them all.

Symmetrical and simpler, yes, but you're weakening the pmd_bad() test:
no longer requires that all those _KERNPG_TABLE bits be set. Wouldn't it be better to go the other way and strengthen pgd_bad, pud_bad?

Maybe, but there must have been a reason for not doing so. It could certainly be a follow-up patch - if it works.

Jan

#4 lore raw
In reply to Hugh Dickins
Quoted 8 lines · Hugh Dickins
Quoted 2 lines · Jan Beulich

Make pmd_bad() symmetrical to pgd_bad() and pud_bad(). At once, simplify them all.

Symmetrical and simpler, yes, but you're weakening the pmd_bad() test:
no longer requires that all those _KERNPG_TABLE bits be set. Wouldn't it be better to go the other way and strengthen pgd_bad, pud_bad?

That's a good point. Yes that would be better. If it works :)

They can't be completely the same because we don't set large page bits on PGDs (on PUDs we will eventually with 1GB pages)

-Andi