* [PATCH -V2] powerpc: thp: Fix crash on mremap
@ 2014-01-02 9:17 ` Aneesh Kumar K.V
0 siblings, 0 replies; 10+ messages in thread
From: Aneesh Kumar K.V @ 2014-01-02 9:17 UTC (permalink / raw)
To: benh, paulus, aarcange, kirill.shutemov
Cc: linuxppc-dev, linux-mm, Aneesh Kumar K.V
From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
This patch fix the below crash
NIP [c00000000004cee4] .__hash_page_thp+0x2a4/0x440
LR [c0000000000439ac] .hash_page+0x18c/0x5e0
...
Call Trace:
[c000000736103c40] [00001ffffb000000] 0x1ffffb000000(unreliable)
[437908.479693] [c000000736103d50] [c0000000000439ac] .hash_page+0x18c/0x5e0
[437908.479699] [c000000736103e30] [c00000000000924c] .do_hash_page+0x4c/0x58
On ppc64 we use the pgtable for storing the hpte slot information and
store address to the pgtable at a constant offset (PTRS_PER_PMD) from
pmd. On mremap, when we switch the pmd, we need to withdraw and deposit
the pgtable again, so that we find the pgtable at PTRS_PER_PMD offset
from new pmd.
We also want to move the withdraw and deposit before the set_pmd so
that, when page fault find the pmd as trans huge we can be sure that
pgtable can be located at the offset.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
---
Changes from V1:
* limit the withraw/deposit to only ppc64
arch/Kconfig | 3 +++
arch/powerpc/platforms/Kconfig.cputype | 1 +
include/linux/huge_mm.h | 6 ++++++
mm/huge_memory.c | 21 ++++++++++++---------
4 files changed, 22 insertions(+), 9 deletions(-)
diff --git a/arch/Kconfig b/arch/Kconfig
index f1cf895c040f..3759e70a649d 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -371,6 +371,9 @@ config HAVE_IRQ_TIME_ACCOUNTING
config HAVE_ARCH_TRANSPARENT_HUGEPAGE
bool
+config ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
+ bool
+
config HAVE_ARCH_SOFT_DIRTY
bool
diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype
index bca2465a9c34..5f83b4334e5f 100644
--- a/arch/powerpc/platforms/Kconfig.cputype
+++ b/arch/powerpc/platforms/Kconfig.cputype
@@ -71,6 +71,7 @@ config PPC_BOOK3S_64
select PPC_FPU
select PPC_HAVE_PMU_SUPPORT
select SYS_SUPPORTS_HUGETLBFS
+ select ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
select HAVE_ARCH_TRANSPARENT_HUGEPAGE if PPC_64K_PAGES
config PPC_BOOK3E_64
diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
index 91672e2deec3..836242a738a5 100644
--- a/include/linux/huge_mm.h
+++ b/include/linux/huge_mm.h
@@ -230,4 +230,10 @@ static inline int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_str
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
+#ifdef CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
+#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 1
+#else
+#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 0
+#endif
+
#endif /* _LINUX_HUGE_MM_H */
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 7de1bf85f683..32006b51d102 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -1505,19 +1505,22 @@ int move_huge_pmd(struct vm_area_struct *vma, struct vm_area_struct *new_vma,
spin_lock_nested(new_ptl, SINGLE_DEPTH_NESTING);
pmd = pmdp_get_and_clear(mm, old_addr, old_pmd);
VM_BUG_ON(!pmd_none(*new_pmd));
- set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
- if (new_ptl != old_ptl) {
+ /*
+ * Archs like ppc64 use pgtable to store per pmd
+ * specific information. So when we switch the pmd,
+ * we should also withdraw and deposit the pgtable
+ *
+ * With split pmd lock we also need to move preallocated
+ * PTE page table if new_pmd is on different PMD page table.
+ */
+ if (new_ptl != old_ptl || ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW) {
pgtable_t pgtable;
-
- /*
- * Move preallocated PTE page table if new_pmd is on
- * different PMD page table.
- */
pgtable = pgtable_trans_huge_withdraw(mm, old_pmd);
pgtable_trans_huge_deposit(mm, new_pmd, pgtable);
-
- spin_unlock(new_ptl);
}
+ set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
+ if (new_ptl != old_ptl)
+ spin_unlock(new_ptl);
spin_unlock(old_ptl);
}
out:
--
1.8.3.2
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related [flat|nested] 10+ messages in thread
* [PATCH -V2] powerpc: thp: Fix crash on mremap
@ 2014-01-02 9:17 ` Aneesh Kumar K.V
0 siblings, 0 replies; 10+ messages in thread
From: Aneesh Kumar K.V @ 2014-01-02 9:17 UTC (permalink / raw)
To: benh, paulus, aarcange, kirill.shutemov
Cc: linux-mm, linuxppc-dev, Aneesh Kumar K.V
From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
This patch fix the below crash
NIP [c00000000004cee4] .__hash_page_thp+0x2a4/0x440
LR [c0000000000439ac] .hash_page+0x18c/0x5e0
...
Call Trace:
[c000000736103c40] [00001ffffb000000] 0x1ffffb000000(unreliable)
[437908.479693] [c000000736103d50] [c0000000000439ac] .hash_page+0x18c/0x5e0
[437908.479699] [c000000736103e30] [c00000000000924c] .do_hash_page+0x4c/0x58
On ppc64 we use the pgtable for storing the hpte slot information and
store address to the pgtable at a constant offset (PTRS_PER_PMD) from
pmd. On mremap, when we switch the pmd, we need to withdraw and deposit
the pgtable again, so that we find the pgtable at PTRS_PER_PMD offset
from new pmd.
We also want to move the withdraw and deposit before the set_pmd so
that, when page fault find the pmd as trans huge we can be sure that
pgtable can be located at the offset.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
---
Changes from V1:
* limit the withraw/deposit to only ppc64
arch/Kconfig | 3 +++
arch/powerpc/platforms/Kconfig.cputype | 1 +
include/linux/huge_mm.h | 6 ++++++
mm/huge_memory.c | 21 ++++++++++++---------
4 files changed, 22 insertions(+), 9 deletions(-)
diff --git a/arch/Kconfig b/arch/Kconfig
index f1cf895c040f..3759e70a649d 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -371,6 +371,9 @@ config HAVE_IRQ_TIME_ACCOUNTING
config HAVE_ARCH_TRANSPARENT_HUGEPAGE
bool
+config ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
+ bool
+
config HAVE_ARCH_SOFT_DIRTY
bool
diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype
index bca2465a9c34..5f83b4334e5f 100644
--- a/arch/powerpc/platforms/Kconfig.cputype
+++ b/arch/powerpc/platforms/Kconfig.cputype
@@ -71,6 +71,7 @@ config PPC_BOOK3S_64
select PPC_FPU
select PPC_HAVE_PMU_SUPPORT
select SYS_SUPPORTS_HUGETLBFS
+ select ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
select HAVE_ARCH_TRANSPARENT_HUGEPAGE if PPC_64K_PAGES
config PPC_BOOK3E_64
diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
index 91672e2deec3..836242a738a5 100644
--- a/include/linux/huge_mm.h
+++ b/include/linux/huge_mm.h
@@ -230,4 +230,10 @@ static inline int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_str
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
+#ifdef CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
+#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 1
+#else
+#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 0
+#endif
+
#endif /* _LINUX_HUGE_MM_H */
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 7de1bf85f683..32006b51d102 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -1505,19 +1505,22 @@ int move_huge_pmd(struct vm_area_struct *vma, struct vm_area_struct *new_vma,
spin_lock_nested(new_ptl, SINGLE_DEPTH_NESTING);
pmd = pmdp_get_and_clear(mm, old_addr, old_pmd);
VM_BUG_ON(!pmd_none(*new_pmd));
- set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
- if (new_ptl != old_ptl) {
+ /*
+ * Archs like ppc64 use pgtable to store per pmd
+ * specific information. So when we switch the pmd,
+ * we should also withdraw and deposit the pgtable
+ *
+ * With split pmd lock we also need to move preallocated
+ * PTE page table if new_pmd is on different PMD page table.
+ */
+ if (new_ptl != old_ptl || ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW) {
pgtable_t pgtable;
-
- /*
- * Move preallocated PTE page table if new_pmd is on
- * different PMD page table.
- */
pgtable = pgtable_trans_huge_withdraw(mm, old_pmd);
pgtable_trans_huge_deposit(mm, new_pmd, pgtable);
-
- spin_unlock(new_ptl);
}
+ set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
+ if (new_ptl != old_ptl)
+ spin_unlock(new_ptl);
spin_unlock(old_ptl);
}
out:
--
1.8.3.2
^ permalink raw reply related [flat|nested] 10+ messages in thread
* RE: [PATCH -V2] powerpc: thp: Fix crash on mremap
2014-01-02 9:17 ` Aneesh Kumar K.V
@ 2014-01-02 9:41 ` Kirill A. Shutemov
-1 siblings, 0 replies; 10+ messages in thread
From: Kirill A. Shutemov @ 2014-01-02 9:41 UTC (permalink / raw)
To: Aneesh Kumar K.V
Cc: benh, paulus, aarcange, kirill.shutemov, linuxppc-dev, linux-mm
Aneesh Kumar K.V wrote:
> From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
>
> This patch fix the below crash
>
> NIP [c00000000004cee4] .__hash_page_thp+0x2a4/0x440
> LR [c0000000000439ac] .hash_page+0x18c/0x5e0
> ...
> Call Trace:
> [c000000736103c40] [00001ffffb000000] 0x1ffffb000000(unreliable)
> [437908.479693] [c000000736103d50] [c0000000000439ac] .hash_page+0x18c/0x5e0
> [437908.479699] [c000000736103e30] [c00000000000924c] .do_hash_page+0x4c/0x58
>
> On ppc64 we use the pgtable for storing the hpte slot information and
> store address to the pgtable at a constant offset (PTRS_PER_PMD) from
> pmd. On mremap, when we switch the pmd, we need to withdraw and deposit
> the pgtable again, so that we find the pgtable at PTRS_PER_PMD offset
> from new pmd.
>
> We also want to move the withdraw and deposit before the set_pmd so
> that, when page fault find the pmd as trans huge we can be sure that
> pgtable can be located at the offset.
>
> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
> ---
> Changes from V1:
> * limit the withraw/deposit to only ppc64
>
> arch/Kconfig | 3 +++
> arch/powerpc/platforms/Kconfig.cputype | 1 +
> include/linux/huge_mm.h | 6 ++++++
> mm/huge_memory.c | 21 ++++++++++++---------
> 4 files changed, 22 insertions(+), 9 deletions(-)
>
> diff --git a/arch/Kconfig b/arch/Kconfig
> index f1cf895c040f..3759e70a649d 100644
> --- a/arch/Kconfig
> +++ b/arch/Kconfig
> @@ -371,6 +371,9 @@ config HAVE_IRQ_TIME_ACCOUNTING
> config HAVE_ARCH_TRANSPARENT_HUGEPAGE
> bool
>
> +config ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
I don't like name of the option, but can't find any better... :(
> + bool
> +
> config HAVE_ARCH_SOFT_DIRTY
> bool
>
> diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype
> index bca2465a9c34..5f83b4334e5f 100644
> --- a/arch/powerpc/platforms/Kconfig.cputype
> +++ b/arch/powerpc/platforms/Kconfig.cputype
> @@ -71,6 +71,7 @@ config PPC_BOOK3S_64
> select PPC_FPU
> select PPC_HAVE_PMU_SUPPORT
> select SYS_SUPPORTS_HUGETLBFS
> + select ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
> select HAVE_ARCH_TRANSPARENT_HUGEPAGE if PPC_64K_PAGES
>
> config PPC_BOOK3E_64
> diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
> index 91672e2deec3..836242a738a5 100644
> --- a/include/linux/huge_mm.h
> +++ b/include/linux/huge_mm.h
> @@ -230,4 +230,10 @@ static inline int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_str
>
> #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
>
> +#ifdef CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
> +#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 1
> +#else
> +#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 0
> +#endif
> +
Just use config option directly:
if (new_ptl != old_ptl ||
IS_ENABLED(CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW))
...
Otherwise, looks good:
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> #endif /* _LINUX_HUGE_MM_H */
> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> index 7de1bf85f683..32006b51d102 100644
> --- a/mm/huge_memory.c
> +++ b/mm/huge_memory.c
> @@ -1505,19 +1505,22 @@ int move_huge_pmd(struct vm_area_struct *vma, struct vm_area_struct *new_vma,
> spin_lock_nested(new_ptl, SINGLE_DEPTH_NESTING);
> pmd = pmdp_get_and_clear(mm, old_addr, old_pmd);
> VM_BUG_ON(!pmd_none(*new_pmd));
> - set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
> - if (new_ptl != old_ptl) {
> + /*
> + * Archs like ppc64 use pgtable to store per pmd
> + * specific information. So when we switch the pmd,
> + * we should also withdraw and deposit the pgtable
> + *
> + * With split pmd lock we also need to move preallocated
> + * PTE page table if new_pmd is on different PMD page table.
> + */
> + if (new_ptl != old_ptl || ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW) {
> pgtable_t pgtable;
> -
> - /*
> - * Move preallocated PTE page table if new_pmd is on
> - * different PMD page table.
> - */
> pgtable = pgtable_trans_huge_withdraw(mm, old_pmd);
> pgtable_trans_huge_deposit(mm, new_pmd, pgtable);
> -
> - spin_unlock(new_ptl);
> }
> + set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
> + if (new_ptl != old_ptl)
> + spin_unlock(new_ptl);
> spin_unlock(old_ptl);
> }
> out:
> --
> 1.8.3.2
--
Kirill A. Shutemov
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 10+ messages in thread
* RE: [PATCH -V2] powerpc: thp: Fix crash on mremap
@ 2014-01-02 9:41 ` Kirill A. Shutemov
0 siblings, 0 replies; 10+ messages in thread
From: Kirill A. Shutemov @ 2014-01-02 9:41 UTC (permalink / raw)
To: Aneesh Kumar K.V
Cc: aarcange, linux-mm, paulus, Aneesh Kumar K.V, linuxppc-dev,
kirill.shutemov
Aneesh Kumar K.V wrote:
> From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
>
> This patch fix the below crash
>
> NIP [c00000000004cee4] .__hash_page_thp+0x2a4/0x440
> LR [c0000000000439ac] .hash_page+0x18c/0x5e0
> ...
> Call Trace:
> [c000000736103c40] [00001ffffb000000] 0x1ffffb000000(unreliable)
> [437908.479693] [c000000736103d50] [c0000000000439ac] .hash_page+0x18c/0x5e0
> [437908.479699] [c000000736103e30] [c00000000000924c] .do_hash_page+0x4c/0x58
>
> On ppc64 we use the pgtable for storing the hpte slot information and
> store address to the pgtable at a constant offset (PTRS_PER_PMD) from
> pmd. On mremap, when we switch the pmd, we need to withdraw and deposit
> the pgtable again, so that we find the pgtable at PTRS_PER_PMD offset
> from new pmd.
>
> We also want to move the withdraw and deposit before the set_pmd so
> that, when page fault find the pmd as trans huge we can be sure that
> pgtable can be located at the offset.
>
> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
> ---
> Changes from V1:
> * limit the withraw/deposit to only ppc64
>
> arch/Kconfig | 3 +++
> arch/powerpc/platforms/Kconfig.cputype | 1 +
> include/linux/huge_mm.h | 6 ++++++
> mm/huge_memory.c | 21 ++++++++++++---------
> 4 files changed, 22 insertions(+), 9 deletions(-)
>
> diff --git a/arch/Kconfig b/arch/Kconfig
> index f1cf895c040f..3759e70a649d 100644
> --- a/arch/Kconfig
> +++ b/arch/Kconfig
> @@ -371,6 +371,9 @@ config HAVE_IRQ_TIME_ACCOUNTING
> config HAVE_ARCH_TRANSPARENT_HUGEPAGE
> bool
>
> +config ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
I don't like name of the option, but can't find any better... :(
> + bool
> +
> config HAVE_ARCH_SOFT_DIRTY
> bool
>
> diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype
> index bca2465a9c34..5f83b4334e5f 100644
> --- a/arch/powerpc/platforms/Kconfig.cputype
> +++ b/arch/powerpc/platforms/Kconfig.cputype
> @@ -71,6 +71,7 @@ config PPC_BOOK3S_64
> select PPC_FPU
> select PPC_HAVE_PMU_SUPPORT
> select SYS_SUPPORTS_HUGETLBFS
> + select ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
> select HAVE_ARCH_TRANSPARENT_HUGEPAGE if PPC_64K_PAGES
>
> config PPC_BOOK3E_64
> diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
> index 91672e2deec3..836242a738a5 100644
> --- a/include/linux/huge_mm.h
> +++ b/include/linux/huge_mm.h
> @@ -230,4 +230,10 @@ static inline int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_str
>
> #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
>
> +#ifdef CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
> +#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 1
> +#else
> +#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 0
> +#endif
> +
Just use config option directly:
if (new_ptl != old_ptl ||
IS_ENABLED(CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW))
...
Otherwise, looks good:
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> #endif /* _LINUX_HUGE_MM_H */
> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> index 7de1bf85f683..32006b51d102 100644
> --- a/mm/huge_memory.c
> +++ b/mm/huge_memory.c
> @@ -1505,19 +1505,22 @@ int move_huge_pmd(struct vm_area_struct *vma, struct vm_area_struct *new_vma,
> spin_lock_nested(new_ptl, SINGLE_DEPTH_NESTING);
> pmd = pmdp_get_and_clear(mm, old_addr, old_pmd);
> VM_BUG_ON(!pmd_none(*new_pmd));
> - set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
> - if (new_ptl != old_ptl) {
> + /*
> + * Archs like ppc64 use pgtable to store per pmd
> + * specific information. So when we switch the pmd,
> + * we should also withdraw and deposit the pgtable
> + *
> + * With split pmd lock we also need to move preallocated
> + * PTE page table if new_pmd is on different PMD page table.
> + */
> + if (new_ptl != old_ptl || ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW) {
> pgtable_t pgtable;
> -
> - /*
> - * Move preallocated PTE page table if new_pmd is on
> - * different PMD page table.
> - */
> pgtable = pgtable_trans_huge_withdraw(mm, old_pmd);
> pgtable_trans_huge_deposit(mm, new_pmd, pgtable);
> -
> - spin_unlock(new_ptl);
> }
> + set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
> + if (new_ptl != old_ptl)
> + spin_unlock(new_ptl);
> spin_unlock(old_ptl);
> }
> out:
> --
> 1.8.3.2
--
Kirill A. Shutemov
^ permalink raw reply [flat|nested] 10+ messages in thread
* RE: [PATCH -V2] powerpc: thp: Fix crash on mremap
2014-01-02 9:41 ` Kirill A. Shutemov
@ 2014-01-02 10:52 ` Aneesh Kumar K.V
-1 siblings, 0 replies; 10+ messages in thread
From: Aneesh Kumar K.V @ 2014-01-02 10:52 UTC (permalink / raw)
To: Kirill A. Shutemov; +Cc: benh, paulus, aarcange, linuxppc-dev, linux-mm
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> writes:
> Aneesh Kumar K.V wrote:
>> From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
>>
>> This patch fix the below crash
>>
>> NIP [c00000000004cee4] .__hash_page_thp+0x2a4/0x440
>> LR [c0000000000439ac] .hash_page+0x18c/0x5e0
>> ...
>> Call Trace:
>> [c000000736103c40] [00001ffffb000000] 0x1ffffb000000(unreliable)
>> [437908.479693] [c000000736103d50] [c0000000000439ac] .hash_page+0x18c/0x5e0
>> [437908.479699] [c000000736103e30] [c00000000000924c] .do_hash_page+0x4c/0x58
>>
>> On ppc64 we use the pgtable for storing the hpte slot information and
>> store address to the pgtable at a constant offset (PTRS_PER_PMD) from
>> pmd. On mremap, when we switch the pmd, we need to withdraw and deposit
>> the pgtable again, so that we find the pgtable at PTRS_PER_PMD offset
>> from new pmd.
>>
>> We also want to move the withdraw and deposit before the set_pmd so
>> that, when page fault find the pmd as trans huge we can be sure that
>> pgtable can be located at the offset.
>>
>> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
>> ---
>> Changes from V1:
>> * limit the withraw/deposit to only ppc64
>>
>> arch/Kconfig | 3 +++
>> arch/powerpc/platforms/Kconfig.cputype | 1 +
>> include/linux/huge_mm.h | 6 ++++++
>> mm/huge_memory.c | 21 ++++++++++++---------
>> 4 files changed, 22 insertions(+), 9 deletions(-)
>>
>> diff --git a/arch/Kconfig b/arch/Kconfig
>> index f1cf895c040f..3759e70a649d 100644
>> --- a/arch/Kconfig
>> +++ b/arch/Kconfig
>> @@ -371,6 +371,9 @@ config HAVE_IRQ_TIME_ACCOUNTING
>> config HAVE_ARCH_TRANSPARENT_HUGEPAGE
>> bool
>>
>> +config ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
>
> I don't like name of the option, but can't find any better... :(
>
>> + bool
>> +
>> config HAVE_ARCH_SOFT_DIRTY
>> bool
>>
>> diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype
>> index bca2465a9c34..5f83b4334e5f 100644
>> --- a/arch/powerpc/platforms/Kconfig.cputype
>> +++ b/arch/powerpc/platforms/Kconfig.cputype
>> @@ -71,6 +71,7 @@ config PPC_BOOK3S_64
>> select PPC_FPU
>> select PPC_HAVE_PMU_SUPPORT
>> select SYS_SUPPORTS_HUGETLBFS
>> + select ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
>> select HAVE_ARCH_TRANSPARENT_HUGEPAGE if PPC_64K_PAGES
>>
>> config PPC_BOOK3E_64
>> diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
>> index 91672e2deec3..836242a738a5 100644
>> --- a/include/linux/huge_mm.h
>> +++ b/include/linux/huge_mm.h
>> @@ -230,4 +230,10 @@ static inline int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_str
>>
>> #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
>>
>> +#ifdef CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
>> +#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 1
>> +#else
>> +#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 0
>> +#endif
>> +
>
> Just use config option directly:
>
> if (new_ptl != old_ptl ||
> IS_ENABLED(CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW))
I didn't like that. I found the earlier one easier for reading.
If you and others strongly feel about this, I can redo the patch. Please let me know
> ...
>
>
> Otherwise, looks good:
>
> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
>
>> #endif /* _LINUX_HUGE_MM_H */
>> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
>> index 7de1bf85f683..32006b51d102 100644
>> --- a/mm/huge_memory.c
>> +++ b/mm/huge_memory.c
>> @@ -1505,19 +1505,22 @@ int move_huge_pmd(struct vm_area_struct *vma, struct vm_area_struct *new_vma,
>> spin_lock_nested(new_ptl, SINGLE_DEPTH_NESTING);
>> pmd = pmdp_get_and_clear(mm, old_addr, old_pmd);
>> VM_BUG_ON(!pmd_none(*new_pmd));
>> - set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
>> - if (new_ptl != old_ptl) {
>> + /*
>> + * Archs like ppc64 use pgtable to store per pmd
>> + * specific information. So when we switch the pmd,
>> + * we should also withdraw and deposit the pgtable
>> + *
>> + * With split pmd lock we also need to move preallocated
>> + * PTE page table if new_pmd is on different PMD page table.
>> + */
>> + if (new_ptl != old_ptl || ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW) {
>> pgtable_t pgtable;
>> -
>> - /*
>> - * Move preallocated PTE page table if new_pmd is on
>> - * different PMD page table.
>> - */
>> pgtable = pgtable_trans_huge_withdraw(mm, old_pmd);
>> pgtable_trans_huge_deposit(mm, new_pmd, pgtable);
>> -
>> - spin_unlock(new_ptl);
>> }
>> + set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
>> + if (new_ptl != old_ptl)
>> + spin_unlock(new_ptl);
>> spin_unlock(old_ptl);
>> }
>> out:
>> --
>> 1.8.3.2
>
> --
> Kirill A. Shutemov
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 10+ messages in thread
* RE: [PATCH -V2] powerpc: thp: Fix crash on mremap
@ 2014-01-02 10:52 ` Aneesh Kumar K.V
0 siblings, 0 replies; 10+ messages in thread
From: Aneesh Kumar K.V @ 2014-01-02 10:52 UTC (permalink / raw)
To: Kirill A. Shutemov
Cc: aarcange, linux-mm, paulus, linuxppc-dev, kirill.shutemov
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> writes:
> Aneesh Kumar K.V wrote:
>> From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
>>
>> This patch fix the below crash
>>
>> NIP [c00000000004cee4] .__hash_page_thp+0x2a4/0x440
>> LR [c0000000000439ac] .hash_page+0x18c/0x5e0
>> ...
>> Call Trace:
>> [c000000736103c40] [00001ffffb000000] 0x1ffffb000000(unreliable)
>> [437908.479693] [c000000736103d50] [c0000000000439ac] .hash_page+0x18c/0x5e0
>> [437908.479699] [c000000736103e30] [c00000000000924c] .do_hash_page+0x4c/0x58
>>
>> On ppc64 we use the pgtable for storing the hpte slot information and
>> store address to the pgtable at a constant offset (PTRS_PER_PMD) from
>> pmd. On mremap, when we switch the pmd, we need to withdraw and deposit
>> the pgtable again, so that we find the pgtable at PTRS_PER_PMD offset
>> from new pmd.
>>
>> We also want to move the withdraw and deposit before the set_pmd so
>> that, when page fault find the pmd as trans huge we can be sure that
>> pgtable can be located at the offset.
>>
>> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
>> ---
>> Changes from V1:
>> * limit the withraw/deposit to only ppc64
>>
>> arch/Kconfig | 3 +++
>> arch/powerpc/platforms/Kconfig.cputype | 1 +
>> include/linux/huge_mm.h | 6 ++++++
>> mm/huge_memory.c | 21 ++++++++++++---------
>> 4 files changed, 22 insertions(+), 9 deletions(-)
>>
>> diff --git a/arch/Kconfig b/arch/Kconfig
>> index f1cf895c040f..3759e70a649d 100644
>> --- a/arch/Kconfig
>> +++ b/arch/Kconfig
>> @@ -371,6 +371,9 @@ config HAVE_IRQ_TIME_ACCOUNTING
>> config HAVE_ARCH_TRANSPARENT_HUGEPAGE
>> bool
>>
>> +config ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
>
> I don't like name of the option, but can't find any better... :(
>
>> + bool
>> +
>> config HAVE_ARCH_SOFT_DIRTY
>> bool
>>
>> diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype
>> index bca2465a9c34..5f83b4334e5f 100644
>> --- a/arch/powerpc/platforms/Kconfig.cputype
>> +++ b/arch/powerpc/platforms/Kconfig.cputype
>> @@ -71,6 +71,7 @@ config PPC_BOOK3S_64
>> select PPC_FPU
>> select PPC_HAVE_PMU_SUPPORT
>> select SYS_SUPPORTS_HUGETLBFS
>> + select ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
>> select HAVE_ARCH_TRANSPARENT_HUGEPAGE if PPC_64K_PAGES
>>
>> config PPC_BOOK3E_64
>> diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
>> index 91672e2deec3..836242a738a5 100644
>> --- a/include/linux/huge_mm.h
>> +++ b/include/linux/huge_mm.h
>> @@ -230,4 +230,10 @@ static inline int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_str
>>
>> #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
>>
>> +#ifdef CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW
>> +#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 1
>> +#else
>> +#define ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW 0
>> +#endif
>> +
>
> Just use config option directly:
>
> if (new_ptl != old_ptl ||
> IS_ENABLED(CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW))
I didn't like that. I found the earlier one easier for reading.
If you and others strongly feel about this, I can redo the patch. Please let me know
> ...
>
>
> Otherwise, looks good:
>
> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
>
>> #endif /* _LINUX_HUGE_MM_H */
>> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
>> index 7de1bf85f683..32006b51d102 100644
>> --- a/mm/huge_memory.c
>> +++ b/mm/huge_memory.c
>> @@ -1505,19 +1505,22 @@ int move_huge_pmd(struct vm_area_struct *vma, struct vm_area_struct *new_vma,
>> spin_lock_nested(new_ptl, SINGLE_DEPTH_NESTING);
>> pmd = pmdp_get_and_clear(mm, old_addr, old_pmd);
>> VM_BUG_ON(!pmd_none(*new_pmd));
>> - set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
>> - if (new_ptl != old_ptl) {
>> + /*
>> + * Archs like ppc64 use pgtable to store per pmd
>> + * specific information. So when we switch the pmd,
>> + * we should also withdraw and deposit the pgtable
>> + *
>> + * With split pmd lock we also need to move preallocated
>> + * PTE page table if new_pmd is on different PMD page table.
>> + */
>> + if (new_ptl != old_ptl || ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW) {
>> pgtable_t pgtable;
>> -
>> - /*
>> - * Move preallocated PTE page table if new_pmd is on
>> - * different PMD page table.
>> - */
>> pgtable = pgtable_trans_huge_withdraw(mm, old_pmd);
>> pgtable_trans_huge_deposit(mm, new_pmd, pgtable);
>> -
>> - spin_unlock(new_ptl);
>> }
>> + set_pmd_at(mm, new_addr, new_pmd, pmd_mksoft_dirty(pmd));
>> + if (new_ptl != old_ptl)
>> + spin_unlock(new_ptl);
>> spin_unlock(old_ptl);
>> }
>> out:
>> --
>> 1.8.3.2
>
> --
> Kirill A. Shutemov
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH -V2] powerpc: thp: Fix crash on mremap
2014-01-02 10:52 ` Aneesh Kumar K.V
@ 2014-01-02 12:29 ` Benjamin Herrenschmidt
-1 siblings, 0 replies; 10+ messages in thread
From: Benjamin Herrenschmidt @ 2014-01-02 12:29 UTC (permalink / raw)
To: Aneesh Kumar K.V
Cc: Kirill A. Shutemov, paulus, aarcange, linuxppc-dev, linux-mm
On Thu, 2014-01-02 at 16:22 +0530, Aneesh Kumar K.V wrote:
> > Just use config option directly:
> >
> > if (new_ptl != old_ptl ||
> > IS_ENABLED(CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW))
>
>
> I didn't like that. I found the earlier one easier for reading.
> If you and others strongly feel about this, I can redo the patch.
> Please let me know
Yes, use IS_ENABLED, no need to have two indirections of #define's
Another option is to have
if (pmd_move_must_withdraw(new,old)) {
}
With in a generic header:
#ifndef pmd_move_must_withdraw
static inline bool pmd_move_must_withdraw(spinlock_t *new_ptl, ...)
{
return new_ptl != old_ptl;
}
#endif
And in powerpc:
static inline bool pmd_move_must_withdraw(spinlock_t *new_ptl, ...)
{
return true;
}
#define pmd_move_must_withdraw pmd_move_must_withdraw
Cheers,
Ben.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH -V2] powerpc: thp: Fix crash on mremap
@ 2014-01-02 12:29 ` Benjamin Herrenschmidt
0 siblings, 0 replies; 10+ messages in thread
From: Benjamin Herrenschmidt @ 2014-01-02 12:29 UTC (permalink / raw)
To: Aneesh Kumar K.V
Cc: aarcange, linuxppc-dev, paulus, Kirill A. Shutemov, linux-mm
On Thu, 2014-01-02 at 16:22 +0530, Aneesh Kumar K.V wrote:
> > Just use config option directly:
> >
> > if (new_ptl != old_ptl ||
> > IS_ENABLED(CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW))
>
>
> I didn't like that. I found the earlier one easier for reading.
> If you and others strongly feel about this, I can redo the patch.
> Please let me know
Yes, use IS_ENABLED, no need to have two indirections of #define's
Another option is to have
if (pmd_move_must_withdraw(new,old)) {
}
With in a generic header:
#ifndef pmd_move_must_withdraw
static inline bool pmd_move_must_withdraw(spinlock_t *new_ptl, ...)
{
return new_ptl != old_ptl;
}
#endif
And in powerpc:
static inline bool pmd_move_must_withdraw(spinlock_t *new_ptl, ...)
{
return true;
}
#define pmd_move_must_withdraw pmd_move_must_withdraw
Cheers,
Ben.
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH -V2] powerpc: thp: Fix crash on mremap
2014-01-02 12:29 ` Benjamin Herrenschmidt
@ 2014-01-03 5:28 ` Aneesh Kumar K.V
-1 siblings, 0 replies; 10+ messages in thread
From: Aneesh Kumar K.V @ 2014-01-03 5:28 UTC (permalink / raw)
To: Benjamin Herrenschmidt
Cc: Kirill A. Shutemov, paulus, aarcange, linuxppc-dev, linux-mm
Benjamin Herrenschmidt <benh@kernel.crashing.org> writes:
> On Thu, 2014-01-02 at 16:22 +0530, Aneesh Kumar K.V wrote:
>> > Just use config option directly:
>> >
>> > if (new_ptl != old_ptl ||
>> > IS_ENABLED(CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW))
>>
>>
>> I didn't like that. I found the earlier one easier for reading.
>> If you and others strongly feel about this, I can redo the patch.
>> Please let me know
>
> Yes, use IS_ENABLED, no need to have two indirections of #define's
>
> Another option is to have
>
> if (pmd_move_must_withdraw(new,old)) {
> }
>
> With in a generic header:
>
> #ifndef pmd_move_must_withdraw
> static inline bool pmd_move_must_withdraw(spinlock_t *new_ptl, ...)
> {
> return new_ptl != old_ptl;
> }
> #endif
>
> And in powerpc:
>
> static inline bool pmd_move_must_withdraw(spinlock_t *new_ptl, ...)
> {
> return true;
> }
> #define pmd_move_must_withdraw pmd_move_must_withdraw
This is better i guess. It is also in-line with rest of transparent
hugepage functions. I will do this.
-aneesh
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH -V2] powerpc: thp: Fix crash on mremap
@ 2014-01-03 5:28 ` Aneesh Kumar K.V
0 siblings, 0 replies; 10+ messages in thread
From: Aneesh Kumar K.V @ 2014-01-03 5:28 UTC (permalink / raw)
To: Benjamin Herrenschmidt
Cc: aarcange, linuxppc-dev, paulus, Kirill A. Shutemov, linux-mm
Benjamin Herrenschmidt <benh@kernel.crashing.org> writes:
> On Thu, 2014-01-02 at 16:22 +0530, Aneesh Kumar K.V wrote:
>> > Just use config option directly:
>> >
>> > if (new_ptl != old_ptl ||
>> > IS_ENABLED(CONFIG_ARCH_THP_MOVE_PMD_ALWAYS_WITHDRAW))
>>
>>
>> I didn't like that. I found the earlier one easier for reading.
>> If you and others strongly feel about this, I can redo the patch.
>> Please let me know
>
> Yes, use IS_ENABLED, no need to have two indirections of #define's
>
> Another option is to have
>
> if (pmd_move_must_withdraw(new,old)) {
> }
>
> With in a generic header:
>
> #ifndef pmd_move_must_withdraw
> static inline bool pmd_move_must_withdraw(spinlock_t *new_ptl, ...)
> {
> return new_ptl != old_ptl;
> }
> #endif
>
> And in powerpc:
>
> static inline bool pmd_move_must_withdraw(spinlock_t *new_ptl, ...)
> {
> return true;
> }
> #define pmd_move_must_withdraw pmd_move_must_withdraw
This is better i guess. It is also in-line with rest of transparent
hugepage functions. I will do this.
-aneesh
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2014-01-03 5:28 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-01-02 9:17 [PATCH -V2] powerpc: thp: Fix crash on mremap Aneesh Kumar K.V
2014-01-02 9:17 ` Aneesh Kumar K.V
2014-01-02 9:41 ` Kirill A. Shutemov
2014-01-02 9:41 ` Kirill A. Shutemov
2014-01-02 10:52 ` Aneesh Kumar K.V
2014-01-02 10:52 ` Aneesh Kumar K.V
2014-01-02 12:29 ` Benjamin Herrenschmidt
2014-01-02 12:29 ` Benjamin Herrenschmidt
2014-01-03 5:28 ` Aneesh Kumar K.V
2014-01-03 5:28 ` Aneesh Kumar K.V
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.