* [PATCH 2/5] powerpc/mm: move FSL_BOOK3 version of update_mmu_cache()
2019-08-16 5:41 [PATCH 1/5] powerpc/mm: define empty update_mmu_cache() as static inline Christophe Leroy
@ 2019-08-16 5:41 ` Christophe Leroy
2019-08-16 5:41 ` [PATCH 3/5] powerpc/mm: move update_mmu_cache() into book3s hash utils Christophe Leroy
` (3 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Christophe Leroy @ 2019-08-16 5:41 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Paul Mackerras, Michael Ellerman
Cc: linuxppc-dev, linux-kernel
Move FSL_BOOK3E version of update_mmu_cache() at the same
place as book3e_hugetlb_preload() as update_mmu_cache() is
the only user of book3e_hugetlb_preload().
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
---
arch/powerpc/include/asm/hugetlb.h | 3 ---
arch/powerpc/mm/mem.c | 8 --------
arch/powerpc/mm/nohash/book3e_hugetlbpage.c | 16 ++++++++++++++--
3 files changed, 14 insertions(+), 13 deletions(-)
diff --git a/arch/powerpc/include/asm/hugetlb.h b/arch/powerpc/include/asm/hugetlb.h
index 20a101046cff..bd6504c28c2f 100644
--- a/arch/powerpc/include/asm/hugetlb.h
+++ b/arch/powerpc/include/asm/hugetlb.h
@@ -31,9 +31,6 @@ static inline int is_hugepage_only_range(struct mm_struct *mm,
return 0;
}
-void book3e_hugetlb_preload(struct vm_area_struct *vma, unsigned long ea,
- pte_t pte);
-
#define __HAVE_ARCH_HUGETLB_FREE_PGD_RANGE
void hugetlb_free_pgd_range(struct mmu_gather *tlb, unsigned long addr,
unsigned long end, unsigned long floor,
diff --git a/arch/powerpc/mm/mem.c b/arch/powerpc/mm/mem.c
index 582ad728ac9d..c45d44538ddb 100644
--- a/arch/powerpc/mm/mem.c
+++ b/arch/powerpc/mm/mem.c
@@ -457,14 +457,6 @@ void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
hash_preload(vma->vm_mm, address, is_exec, trap);
}
#endif /* CONFIG_PPC_BOOK3S */
-#if defined(CONFIG_PPC_FSL_BOOK3E) && defined(CONFIG_HUGETLB_PAGE)
-void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
- pte_t *ptep)
-{
- if (is_vm_hugetlb_page(vma))
- book3e_hugetlb_preload(vma, address, *ptep);
-}
-#endif
/*
* System memory should not be in /proc/iomem but various tools expect it
diff --git a/arch/powerpc/mm/nohash/book3e_hugetlbpage.c b/arch/powerpc/mm/nohash/book3e_hugetlbpage.c
index 61915f4d3c7f..8b88be91b622 100644
--- a/arch/powerpc/mm/nohash/book3e_hugetlbpage.c
+++ b/arch/powerpc/mm/nohash/book3e_hugetlbpage.c
@@ -122,8 +122,8 @@ static inline int book3e_tlb_exists(unsigned long ea, unsigned long pid)
return found;
}
-void book3e_hugetlb_preload(struct vm_area_struct *vma, unsigned long ea,
- pte_t pte)
+static void
+book3e_hugetlb_preload(struct vm_area_struct *vma, unsigned long ea, pte_t pte)
{
unsigned long mas1, mas2;
u64 mas7_3;
@@ -183,6 +183,18 @@ void book3e_hugetlb_preload(struct vm_area_struct *vma, unsigned long ea,
local_irq_restore(flags);
}
+/*
+ * This is called at the end of handling a user page fault, when the
+ * fault has been handled by updating a PTE in the linux page tables.
+ *
+ * This must always be called with the pte lock held.
+ */
+void update_mmu_cache(struct vm_area_struct *vma, unsigned long address, pte_t *ptep)
+{
+ if (is_vm_hugetlb_page(vma))
+ book3e_hugetlb_preload(vma, address, *ptep);
+}
+
void flush_hugetlb_page(struct vm_area_struct *vma, unsigned long vmaddr)
{
struct hstate *hstate = hstate_file(vma->vm_file);
--
2.13.3
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH 3/5] powerpc/mm: move update_mmu_cache() into book3s hash utils.
2019-08-16 5:41 [PATCH 1/5] powerpc/mm: define empty update_mmu_cache() as static inline Christophe Leroy
2019-08-16 5:41 ` [PATCH 2/5] powerpc/mm: move FSL_BOOK3 version of update_mmu_cache() Christophe Leroy
@ 2019-08-16 5:41 ` Christophe Leroy
2019-08-16 5:41 ` [PATCH 4/5] powerpc/mm: Simplify update_mmu_cache() on BOOK3S32 Christophe Leroy
` (2 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Christophe Leroy @ 2019-08-16 5:41 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Paul Mackerras, Michael Ellerman
Cc: linuxppc-dev, linux-kernel
update_mmu_cache() is only for BOOK3S, and can be simplified
for BOOK3S32.
Move it out of mem.c into respective BOOK3S32 and BOOK3S64
files containing hash utils.
BOOK3S64 version of hash_preload() is only used locally,
declare it static.
Remove the radix_enabled() stuff in BOOK3S32 version.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
---
arch/powerpc/mm/book3s32/mmu.c | 45 +++++++++++++++++++++++++++++
arch/powerpc/mm/book3s64/hash_utils.c | 54 +++++++++++++++++++++++++++++++++--
arch/powerpc/mm/mem.c | 52 ---------------------------------
arch/powerpc/mm/mmu_decl.h | 7 ++---
4 files changed, 100 insertions(+), 58 deletions(-)
diff --git a/arch/powerpc/mm/book3s32/mmu.c b/arch/powerpc/mm/book3s32/mmu.c
index e249fbf6b9c3..3e3c4077cdb7 100644
--- a/arch/powerpc/mm/book3s32/mmu.c
+++ b/arch/powerpc/mm/book3s32/mmu.c
@@ -310,6 +310,51 @@ void hash_preload(struct mm_struct *mm, unsigned long ea,
}
/*
+ * This is called at the end of handling a user page fault, when the
+ * fault has been handled by updating a PTE in the linux page tables.
+ * We use it to preload an HPTE into the hash table corresponding to
+ * the updated linux PTE.
+ *
+ * This must always be called with the pte lock held.
+ */
+void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
+ pte_t *ptep)
+{
+ /*
+ * We don't need to worry about _PAGE_PRESENT here because we are
+ * called with either mm->page_table_lock held or ptl lock held
+ */
+ unsigned long trap;
+ bool is_exec;
+
+ /* We only want HPTEs for linux PTEs that have _PAGE_ACCESSED set */
+ if (!pte_young(*ptep) || address >= TASK_SIZE)
+ return;
+
+ /* We try to figure out if we are coming from an instruction
+ * access fault and pass that down to __hash_page so we avoid
+ * double-faulting on execution of fresh text. We have to test
+ * for regs NULL since init will get here first thing at boot
+ *
+ * We also avoid filling the hash if not coming from a fault
+ */
+
+ trap = current->thread.regs ? TRAP(current->thread.regs) : 0UL;
+ switch (trap) {
+ case 0x300:
+ is_exec = false;
+ break;
+ case 0x400:
+ is_exec = true;
+ break;
+ default:
+ return;
+ }
+
+ hash_preload(vma->vm_mm, address, is_exec, trap);
+}
+
+/*
* Initialize the hash table and patch the instructions in hashtable.S.
*/
void __init MMU_init_hw(void)
diff --git a/arch/powerpc/mm/book3s64/hash_utils.c b/arch/powerpc/mm/book3s64/hash_utils.c
index e6d471058597..fdad5bc07d79 100644
--- a/arch/powerpc/mm/book3s64/hash_utils.c
+++ b/arch/powerpc/mm/book3s64/hash_utils.c
@@ -1518,8 +1518,8 @@ static bool should_hash_preload(struct mm_struct *mm, unsigned long ea)
}
#endif
-void hash_preload(struct mm_struct *mm, unsigned long ea,
- bool is_exec, unsigned long trap)
+static void hash_preload(struct mm_struct *mm, unsigned long ea,
+ bool is_exec, unsigned long trap)
{
int hugepage_shift;
unsigned long vsid;
@@ -1599,6 +1599,56 @@ void hash_preload(struct mm_struct *mm, unsigned long ea,
local_irq_restore(flags);
}
+/*
+ * This is called at the end of handling a user page fault, when the
+ * fault has been handled by updating a PTE in the linux page tables.
+ * We use it to preload an HPTE into the hash table corresponding to
+ * the updated linux PTE.
+ *
+ * This must always be called with the pte lock held.
+ */
+void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
+ pte_t *ptep)
+{
+ /*
+ * We don't need to worry about _PAGE_PRESENT here because we are
+ * called with either mm->page_table_lock held or ptl lock held
+ */
+ unsigned long trap;
+ bool is_exec;
+
+ if (radix_enabled()) {
+ prefetch((void *)address);
+ return;
+ }
+
+ /* We only want HPTEs for linux PTEs that have _PAGE_ACCESSED set */
+ if (!pte_young(*ptep) || address >= TASK_SIZE)
+ return;
+
+ /* We try to figure out if we are coming from an instruction
+ * access fault and pass that down to __hash_page so we avoid
+ * double-faulting on execution of fresh text. We have to test
+ * for regs NULL since init will get here first thing at boot
+ *
+ * We also avoid filling the hash if not coming from a fault
+ */
+
+ trap = current->thread.regs ? TRAP(current->thread.regs) : 0UL;
+ switch (trap) {
+ case 0x300:
+ is_exec = false;
+ break;
+ case 0x400:
+ is_exec = true;
+ break;
+ default:
+ return;
+ }
+
+ hash_preload(vma->vm_mm, address, is_exec, trap);
+}
+
#ifdef CONFIG_PPC_MEM_KEYS
/*
* Return the protection key associated with the given address and the
diff --git a/arch/powerpc/mm/mem.c b/arch/powerpc/mm/mem.c
index c45d44538ddb..3e9e9a051c93 100644
--- a/arch/powerpc/mm/mem.c
+++ b/arch/powerpc/mm/mem.c
@@ -407,58 +407,6 @@ void flush_icache_user_range(struct vm_area_struct *vma, struct page *page,
EXPORT_SYMBOL(flush_icache_user_range);
/*
- * This is called at the end of handling a user page fault, when the
- * fault has been handled by updating a PTE in the linux page tables.
- * We use it to preload an HPTE into the hash table corresponding to
- * the updated linux PTE.
- *
- * This must always be called with the pte lock held.
- */
-#ifdef CONFIG_PPC_BOOK3S
-void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
- pte_t *ptep)
-{
- /*
- * We don't need to worry about _PAGE_PRESENT here because we are
- * called with either mm->page_table_lock held or ptl lock held
- */
- unsigned long trap;
- bool is_exec;
-
- if (radix_enabled()) {
- prefetch((void *)address);
- return;
- }
-
- /* We only want HPTEs for linux PTEs that have _PAGE_ACCESSED set */
- if (!pte_young(*ptep) || address >= TASK_SIZE)
- return;
-
- /* We try to figure out if we are coming from an instruction
- * access fault and pass that down to __hash_page so we avoid
- * double-faulting on execution of fresh text. We have to test
- * for regs NULL since init will get here first thing at boot
- *
- * We also avoid filling the hash if not coming from a fault
- */
-
- trap = current->thread.regs ? TRAP(current->thread.regs) : 0UL;
- switch (trap) {
- case 0x300:
- is_exec = false;
- break;
- case 0x400:
- is_exec = true;
- break;
- default:
- return;
- }
-
- hash_preload(vma->vm_mm, address, is_exec, trap);
-}
-#endif /* CONFIG_PPC_BOOK3S */
-
-/*
* System memory should not be in /proc/iomem but various tools expect it
* (eg kdump).
*/
diff --git a/arch/powerpc/mm/mmu_decl.h b/arch/powerpc/mm/mmu_decl.h
index 32c1a191c28a..9f325a7a09cb 100644
--- a/arch/powerpc/mm/mmu_decl.h
+++ b/arch/powerpc/mm/mmu_decl.h
@@ -82,10 +82,6 @@ static inline void print_system_hash_info(void) {}
#else /* CONFIG_PPC_MMU_NOHASH */
-extern void hash_preload(struct mm_struct *mm, unsigned long ea,
- bool is_exec, unsigned long trap);
-
-
extern void _tlbie(unsigned long address);
extern void _tlbia(void);
@@ -95,6 +91,9 @@ void print_system_hash_info(void);
#ifdef CONFIG_PPC32
+void hash_preload(struct mm_struct *mm, unsigned long ea,
+ bool is_exec, unsigned long trap);
+
extern void mapin_ram(void);
extern void setbat(int index, unsigned long virt, phys_addr_t phys,
unsigned int size, pgprot_t prot);
--
2.13.3
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH 4/5] powerpc/mm: Simplify update_mmu_cache() on BOOK3S32
2019-08-16 5:41 [PATCH 1/5] powerpc/mm: define empty update_mmu_cache() as static inline Christophe Leroy
2019-08-16 5:41 ` [PATCH 2/5] powerpc/mm: move FSL_BOOK3 version of update_mmu_cache() Christophe Leroy
2019-08-16 5:41 ` [PATCH 3/5] powerpc/mm: move update_mmu_cache() into book3s hash utils Christophe Leroy
@ 2019-08-16 5:41 ` Christophe Leroy
2019-08-16 5:41 ` [PATCH 5/5] powerpc/mm: ppc 603 doesn't need update_mmu_cache() Christophe Leroy
2019-08-22 13:09 ` [PATCH 1/5] powerpc/mm: define empty update_mmu_cache() as static inline Michael Ellerman
4 siblings, 0 replies; 6+ messages in thread
From: Christophe Leroy @ 2019-08-16 5:41 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Paul Mackerras, Michael Ellerman
Cc: linuxppc-dev, linux-kernel
On BOOK3S32, hash_preload() neither use is_exec nor trap,
so drop those parameters and simplify update_mmu_cached().
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
---
arch/powerpc/mm/book3s32/mmu.c | 29 +++++++----------------------
arch/powerpc/mm/mmu_decl.h | 3 +--
arch/powerpc/mm/pgtable_32.c | 2 +-
3 files changed, 9 insertions(+), 25 deletions(-)
diff --git a/arch/powerpc/mm/book3s32/mmu.c b/arch/powerpc/mm/book3s32/mmu.c
index 3e3c4077cdb7..3a62bf99f93f 100644
--- a/arch/powerpc/mm/book3s32/mmu.c
+++ b/arch/powerpc/mm/book3s32/mmu.c
@@ -297,8 +297,7 @@ void __init setbat(int index, unsigned long virt, phys_addr_t phys,
/*
* Preload a translation in the hash table
*/
-void hash_preload(struct mm_struct *mm, unsigned long ea,
- bool is_exec, unsigned long trap)
+void hash_preload(struct mm_struct *mm, unsigned long ea)
{
pmd_t *pmd;
@@ -324,34 +323,20 @@ void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
* We don't need to worry about _PAGE_PRESENT here because we are
* called with either mm->page_table_lock held or ptl lock held
*/
- unsigned long trap;
- bool is_exec;
/* We only want HPTEs for linux PTEs that have _PAGE_ACCESSED set */
if (!pte_young(*ptep) || address >= TASK_SIZE)
return;
- /* We try to figure out if we are coming from an instruction
- * access fault and pass that down to __hash_page so we avoid
- * double-faulting on execution of fresh text. We have to test
- * for regs NULL since init will get here first thing at boot
- *
- * We also avoid filling the hash if not coming from a fault
- */
+ /* We have to test for regs NULL since init will get here first thing at boot */
+ if (!current->thread.regs)
+ return;
- trap = current->thread.regs ? TRAP(current->thread.regs) : 0UL;
- switch (trap) {
- case 0x300:
- is_exec = false;
- break;
- case 0x400:
- is_exec = true;
- break;
- default:
+ /* We also avoid filling the hash if not coming from a fault */
+ if (TRAP(current->thread.regs) != 0x300 && TRAP(current->thread.regs) != 0x400)
return;
- }
- hash_preload(vma->vm_mm, address, is_exec, trap);
+ hash_preload(vma->vm_mm, address);
}
/*
diff --git a/arch/powerpc/mm/mmu_decl.h b/arch/powerpc/mm/mmu_decl.h
index 9f325a7a09cb..adbaf2167214 100644
--- a/arch/powerpc/mm/mmu_decl.h
+++ b/arch/powerpc/mm/mmu_decl.h
@@ -91,8 +91,7 @@ void print_system_hash_info(void);
#ifdef CONFIG_PPC32
-void hash_preload(struct mm_struct *mm, unsigned long ea,
- bool is_exec, unsigned long trap);
+void hash_preload(struct mm_struct *mm, unsigned long ea);
extern void mapin_ram(void);
extern void setbat(int index, unsigned long virt, phys_addr_t phys,
diff --git a/arch/powerpc/mm/pgtable_32.c b/arch/powerpc/mm/pgtable_32.c
index 35cb96cfc258..97f401a06fcc 100644
--- a/arch/powerpc/mm/pgtable_32.c
+++ b/arch/powerpc/mm/pgtable_32.c
@@ -252,7 +252,7 @@ static void __init __mapin_ram_chunk(unsigned long offset, unsigned long top)
map_kernel_page(v, p, ktext ? PAGE_KERNEL_TEXT : PAGE_KERNEL);
#ifdef CONFIG_PPC_BOOK3S_32
if (ktext)
- hash_preload(&init_mm, v, false, 0x300);
+ hash_preload(&init_mm, v);
#endif
v += PAGE_SIZE;
p += PAGE_SIZE;
--
2.13.3
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH 5/5] powerpc/mm: ppc 603 doesn't need update_mmu_cache()
2019-08-16 5:41 [PATCH 1/5] powerpc/mm: define empty update_mmu_cache() as static inline Christophe Leroy
` (2 preceding siblings ...)
2019-08-16 5:41 ` [PATCH 4/5] powerpc/mm: Simplify update_mmu_cache() on BOOK3S32 Christophe Leroy
@ 2019-08-16 5:41 ` Christophe Leroy
2019-08-22 13:09 ` [PATCH 1/5] powerpc/mm: define empty update_mmu_cache() as static inline Michael Ellerman
4 siblings, 0 replies; 6+ messages in thread
From: Christophe Leroy @ 2019-08-16 5:41 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Paul Mackerras, Michael Ellerman
Cc: linuxppc-dev, linux-kernel
On powerpc 603, there is no hash table so get out of
update_mmu_cache() early.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
---
arch/powerpc/mm/book3s32/mmu.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/arch/powerpc/mm/book3s32/mmu.c b/arch/powerpc/mm/book3s32/mmu.c
index 3a62bf99f93f..c20269be79ec 100644
--- a/arch/powerpc/mm/book3s32/mmu.c
+++ b/arch/powerpc/mm/book3s32/mmu.c
@@ -319,6 +319,8 @@ void hash_preload(struct mm_struct *mm, unsigned long ea)
void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
pte_t *ptep)
{
+ if (!mmu_has_feature(MMU_FTR_HPTE_TABLE))
+ return;
/*
* We don't need to worry about _PAGE_PRESENT here because we are
* called with either mm->page_table_lock held or ptl lock held
--
2.13.3
^ permalink raw reply related [flat|nested] 6+ messages in thread* Re: [PATCH 1/5] powerpc/mm: define empty update_mmu_cache() as static inline
2019-08-16 5:41 [PATCH 1/5] powerpc/mm: define empty update_mmu_cache() as static inline Christophe Leroy
` (3 preceding siblings ...)
2019-08-16 5:41 ` [PATCH 5/5] powerpc/mm: ppc 603 doesn't need update_mmu_cache() Christophe Leroy
@ 2019-08-22 13:09 ` Michael Ellerman
4 siblings, 0 replies; 6+ messages in thread
From: Michael Ellerman @ 2019-08-22 13:09 UTC (permalink / raw)
To: Christophe Leroy, Benjamin Herrenschmidt, Paul Mackerras
Cc: linuxppc-dev, linux-kernel
On Fri, 2019-08-16 at 05:41:40 UTC, Christophe Leroy wrote:
> Only BOOK3S and FSL_BOOK3E have a usefull update_mmu_cache().
>
> For the others, just define it static inline.
>
> In the meantime, simplify the FSL_BOOK3E related ifdef as
> book3e_hugetlb_preload() only exists when CONFIG_PPC_FSL_BOOK3E
> is selected.
>
> Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Series applied to powerpc next, thanks.
https://git.kernel.org/powerpc/c/d9642117914c9d3f800b3bacc19d7e388b04edb4
cheers
^ permalink raw reply [flat|nested] 6+ messages in thread