diff mbox series

[RESEND] mm: Move set_pxd_safe() helpers from generic to platform

Message ID 20241003044842.246016-1-anshuman.khandual@arm.com (mailing list archive)
State New
Headers show
Series [RESEND] mm: Move set_pxd_safe() helpers from generic to platform | expand

Commit Message

Anshuman Khandual Oct. 3, 2024, 4:48 a.m. UTC
set_pxd_safe() helpers that serve a specific purpose for both x86 and riscv
platforms, do not need to be in the common memory code. Otherwise they just
unnecessarily make the common API more complicated. This moves the helpers
from common code to platform instead.

Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: x86@kernel.org
Cc: linux-mm@kvack.org
Cc: linux-riscv@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Suggested-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
---
This patch applies on v6.12-rc1

This patch is just a resend for an earlier patch

https://lore.kernel.org/all/20240920053017.2514920-1-anshuman.khandual@arm.com/

 arch/riscv/include/asm/pgtable.h | 19 ++++++++++++++++
 arch/x86/include/asm/pgtable.h   | 37 +++++++++++++++++++++++++++++++
 include/linux/pgtable.h          | 38 --------------------------------
 3 files changed, 56 insertions(+), 38 deletions(-)

Comments

Dave Hansen Oct. 3, 2024, 2:09 p.m. UTC | #1
On 10/2/24 21:48, Anshuman Khandual wrote:
> set_pxd_safe() helpers that serve a specific purpose for both x86 and riscv
> platforms, do not need to be in the common memory code. Otherwise they just
> unnecessarily make the common API more complicated. This moves the helpers
> from common code to platform instead.

I looked into the x86 side a bit.  I'm pretty sure we can just remove
the _safe variants.  All they do is spew warnings that don't seem to be
doing any good.  They're not really safer in any meaningful way.

I've got a series cooked up to zap them, but I need to get some eyeballs
on it.

But if Andrew wants to pick this up any carry it in the meantime since
it's cross-arch I'm fine with it:

Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
David Hildenbrand Oct. 4, 2024, 7:55 a.m. UTC | #2
On 03.10.24 16:09, Dave Hansen wrote:
> On 10/2/24 21:48, Anshuman Khandual wrote:
>> set_pxd_safe() helpers that serve a specific purpose for both x86 and riscv
>> platforms, do not need to be in the common memory code. Otherwise they just
>> unnecessarily make the common API more complicated. This moves the helpers
>> from common code to platform instead.
> 
> I looked into the x86 side a bit.  I'm pretty sure we can just remove
> the _safe variants.  All they do is spew warnings that don't seem to be
> doing any good.  They're not really safer in any meaningful way.

Yes, one reason I am happy that we are removing them from core code :)

> 
> I've got a series cooked up to zap them, but I need to get some eyeballs
> on it.

Feel free to CC me. If you're busy, let me know and I might have someone 
to tackle this.


Acked-by: David Hildenbrand <david@redhat.com>
diff mbox series

Patch

diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
index e79f15293492..5d7f3e8c2e50 100644
--- a/arch/riscv/include/asm/pgtable.h
+++ b/arch/riscv/include/asm/pgtable.h
@@ -963,6 +963,25 @@  void misc_mem_init(void);
 extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
 #define ZERO_PAGE(vaddr) (virt_to_page(empty_zero_page))
 
+/*
+ * Use set_p*_safe(), and elide TLB flushing, when confident that *no*
+ * TLB flush will be required as a result of the "set". For example, use
+ * in scenarios where it is known ahead of time that the routine is
+ * setting non-present entries, or re-setting an existing entry to the
+ * same value. Otherwise, use the typical "set" helpers and flush the
+ * TLB.
+ */
+#define set_p4d_safe(p4dp, p4d) \
+({ \
+	WARN_ON_ONCE(p4d_present(*p4dp) && !p4d_same(*p4dp, p4d)); \
+	set_p4d(p4dp, p4d); \
+})
+
+#define set_pgd_safe(pgdp, pgd) \
+({ \
+	WARN_ON_ONCE(pgd_present(*pgdp) && !pgd_same(*pgdp, pgd)); \
+	set_pgd(pgdp, pgd); \
+})
 #endif /* !__ASSEMBLY__ */
 
 #endif /* _ASM_RISCV_PGTABLE_H */
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 4c2d080d26b4..593f10aabd45 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -1775,6 +1775,43 @@  bool arch_is_platform_page(u64 paddr);
 #define arch_is_platform_page arch_is_platform_page
 #endif
 
+/*
+ * Use set_p*_safe(), and elide TLB flushing, when confident that *no*
+ * TLB flush will be required as a result of the "set". For example, use
+ * in scenarios where it is known ahead of time that the routine is
+ * setting non-present entries, or re-setting an existing entry to the
+ * same value. Otherwise, use the typical "set" helpers and flush the
+ * TLB.
+ */
+#define set_pte_safe(ptep, pte) \
+({ \
+	WARN_ON_ONCE(pte_present(*ptep) && !pte_same(*ptep, pte)); \
+	set_pte(ptep, pte); \
+})
+
+#define set_pmd_safe(pmdp, pmd) \
+({ \
+	WARN_ON_ONCE(pmd_present(*pmdp) && !pmd_same(*pmdp, pmd)); \
+	set_pmd(pmdp, pmd); \
+})
+
+#define set_pud_safe(pudp, pud) \
+({ \
+	WARN_ON_ONCE(pud_present(*pudp) && !pud_same(*pudp, pud)); \
+	set_pud(pudp, pud); \
+})
+
+#define set_p4d_safe(p4dp, p4d) \
+({ \
+	WARN_ON_ONCE(p4d_present(*p4dp) && !p4d_same(*p4dp, p4d)); \
+	set_p4d(p4dp, p4d); \
+})
+
+#define set_pgd_safe(pgdp, pgd) \
+({ \
+	WARN_ON_ONCE(pgd_present(*pgdp) && !pgd_same(*pgdp, pgd)); \
+	set_pgd(pgdp, pgd); \
+})
 #endif	/* __ASSEMBLY__ */
 
 #endif /* _ASM_X86_PGTABLE_H */
diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
index e8b2ac6bd2ae..23aeffd89a4e 100644
--- a/include/linux/pgtable.h
+++ b/include/linux/pgtable.h
@@ -1056,44 +1056,6 @@  static inline int pgd_same(pgd_t pgd_a, pgd_t pgd_b)
 }
 #endif
 
-/*
- * Use set_p*_safe(), and elide TLB flushing, when confident that *no*
- * TLB flush will be required as a result of the "set". For example, use
- * in scenarios where it is known ahead of time that the routine is
- * setting non-present entries, or re-setting an existing entry to the
- * same value. Otherwise, use the typical "set" helpers and flush the
- * TLB.
- */
-#define set_pte_safe(ptep, pte) \
-({ \
-	WARN_ON_ONCE(pte_present(*ptep) && !pte_same(*ptep, pte)); \
-	set_pte(ptep, pte); \
-})
-
-#define set_pmd_safe(pmdp, pmd) \
-({ \
-	WARN_ON_ONCE(pmd_present(*pmdp) && !pmd_same(*pmdp, pmd)); \
-	set_pmd(pmdp, pmd); \
-})
-
-#define set_pud_safe(pudp, pud) \
-({ \
-	WARN_ON_ONCE(pud_present(*pudp) && !pud_same(*pudp, pud)); \
-	set_pud(pudp, pud); \
-})
-
-#define set_p4d_safe(p4dp, p4d) \
-({ \
-	WARN_ON_ONCE(p4d_present(*p4dp) && !p4d_same(*p4dp, p4d)); \
-	set_p4d(p4dp, p4d); \
-})
-
-#define set_pgd_safe(pgdp, pgd) \
-({ \
-	WARN_ON_ONCE(pgd_present(*pgdp) && !pgd_same(*pgdp, pgd)); \
-	set_pgd(pgdp, pgd); \
-})
-
 #ifndef __HAVE_ARCH_DO_SWAP_PAGE
 static inline void arch_do_swap_page_nr(struct mm_struct *mm,
 				     struct vm_area_struct *vma,