Hi Kevin,
On 2025/1/3 21:27, Kevin Brodsky wrote:
On 03/01/2025 10:35, Qi Zheng wrote:
On 2025/1/3 17:13, Qi Zheng wrote:
On 2025/1/3 16:02, Kevin Brodsky wrote:
On 03/01/2025 04:48, Qi Zheng wrote:
[...]
In __tlb_batch_free_encoded_pages(), we can indeed detect PageTable()
and call
On 2025/1/3 19:14, Peter Zijlstra wrote:
On Sat, Dec 28, 2024 at 11:26:22AM +0200, Mike Rapoport wrote:
On Mon, Dec 23, 2024 at 05:41:01PM +0800, Qi Zheng wrote:
Here we are explicitly dealing with struct page, and the following logic
semms strange:
tlb_remove_page_ptdesc((tlb
Cc: Catalin Marinas
Cc: Dave Hansen
Cc: Linus Walleij
Cc: Andy Lutomirski
Cc: Peter Zijlstra
Cc: "Mike Rapoport (IBM)"
Cc: Ryan Roberts
Cc: Thomas Gleixner
Cc: Will Deacon
Cc: Matthew Wilcox
Cc: Qi Zheng
Cc: linux-al...@vger.kernel.org
Cc: linux-a...@vger.kernel.org
Cc: linux-ar
On 2025/1/6 21:34, Alexander Gordeev wrote:
On Mon, Jan 06, 2025 at 07:05:16PM +0800, Qi Zheng wrote:
I understand that you want to sort p.._free_tlb() routines, but please
Yes, I thought it was a minor change, so I just did it.
do not move the code around or make a separate follow-up
On 2025/1/6 22:35, Alexander Gordeev wrote:
On Mon, Jan 06, 2025 at 09:34:55PM +0800, Qi Zheng wrote:
OK, will change the subject and description to:
s390: pgtable: also move pagetable_dtor() of PxD to pagetable_dtor_free()
To unify the PxD and PTE TLB free path, also move the
On 2025/1/7 17:57, Kevin Brodsky wrote:
On 06/01/2025 04:49, Qi Zheng wrote:
[...]
Once this is done, we should be able to replace all those confusing
calls to tlb_remove_page() on PTPs with tlb_remove_table() and remove
the explicit call to pagetable_dtor(). AIUI this is essentially what
On 2025/1/7 20:32, Andreas Larsson wrote:
On 2024-12-30 10:07, Qi Zheng wrote:
diff --git a/arch/sparc/include/asm/tlb_32.h b/arch/sparc/include/asm/tlb_32.h
index 5cd28a8793e39..910254867dfbd 100644
--- a/arch/sparc/include/asm/tlb_32.h
+++ b/arch/sparc/include/asm/tlb_32.h
@@ -2,6 +2,7
On 2025/1/7 19:58, Kevin Brodsky wrote:
On 07/01/2025 11:51, Qi Zheng wrote:
[...]
Author: Qi Zheng
Date: Fri Dec 13 17:13:48 2024 +0800
mm: pgtable: completely move pagetable_dtor() to generic
tlb_remove_table()
For the generic tlb_remove_table(), it is implemented in the
On 2025/1/6 18:34, Alexander Gordeev wrote:
On Mon, Dec 30, 2024 at 05:07:42PM +0800, Qi Zheng wrote:
The pagetable_p*_dtor() are exactly the same except for the handling of
ptlock. If we make ptlock_free() handle the case where ptdesc->ptl is
NULL and remove VM_BUG_ON_PAGE() f
On 2025/1/6 18:36, Alexander Gordeev wrote:
On Mon, Dec 30, 2024 at 05:07:47PM +0800, Qi Zheng wrote:
To unify the PxD and PTE TLB free path, also move the pagetable_dtor() of
PMD|PUD|P4D to __tlb_remove_table().
The above and Subject are still incorrect: pagetable_dtor() is
called from
On 2025/1/6 18:32, Alexander Gordeev wrote:
On Mon, Dec 30, 2024 at 05:07:41PM +0800, Qi Zheng wrote:
Like PMD and PTE level page table, also add statistics for PUD and P4D
page table.
...
diff --git a/arch/s390/include/asm/tlb.h b/arch/s390/include/asm/tlb.h
index e95b2c8081eb8
On 2025/1/6 20:44, Alexander Gordeev wrote:
On Mon, Jan 06, 2025 at 07:02:17PM +0800, Qi Zheng wrote:
On Mon, Dec 30, 2024 at 05:07:47PM +0800, Qi Zheng wrote:
To unify the PxD and PTE TLB free path, also move the pagetable_dtor() of
PMD|PUD|P4D to __tlb_remove_table().
The above and
On 2025/1/6 20:36, Alexander Gordeev wrote:
On Mon, Jan 06, 2025 at 06:55:58PM +0800, Qi Zheng wrote:
+static inline void pagetable_dtor(struct ptdesc *ptdesc)
+{
+ struct folio *folio = ptdesc_folio(ptdesc);
+
+ ptlock_free(ptdesc);
+ __folio_clear_pgtable(folio
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
By the way, move the comment above __tlb_remove_table() to
riscv_tlb_remove_ptdesc(), it will be more appropriate.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Reviewed-by:
s390 also calls generic pagetable_dtor_free().
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Reviewed-by: Kevin Brodsky
Acked-by: Alexander Gordeev # s390
---
arch/s390/mm/pgalloc.c| 6 --
include/asm-generic/pgalloc.h | 12
include/asm-generic/tlb.
d-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
Reviewed-by: Kevin Brodsky
---
Documentation/mm/split_page_table_lock.rst | 4 +-
arch/arm/include/asm/tlb.h | 4 +-
arch/arm64/include/asm/tlb.h | 8 ++--
arch/csky/include/asm/pgalloc.h| 2
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Reviewed-by: Kevin Brodsky
Cc: linux-arm-ker...@lists.infradead.org
---
arch/arm64/include/asm/tlb.h | 10 --
1 file ch
Like other levels of page tables, add statistics for P4D level page table.
Signed-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
Reviewed-by: Kevin Brodsky
---
arch/riscv/include/asm/pgalloc.h | 6 +-
arch/x86/mm/pgtable.c| 3 +++
include/asm-generic/pgalloc.h
Convert __tlb_remove_table() to use struct ptdesc, which will help to move
pagetable_dtor() to __tlb_remove_table().
And page tables shouldn't have swap cache, so use pagetable_free() instead
of free_page_and_swap_cache() to free page table pages.
Signed-off-by: Qi Zheng
---
arch/x86/in
Several architectures (arm, arm64, riscv and x86) define exactly the
same __tlb_remove_table(), just introduce generic __tlb_remove_table() to
eliminate these duplications.
The s390 __tlb_remove_table() is nearly the same, so also make s390
__tlb_remove_table() version generic.
Signed-off-by: Qi
more easily.
Signed-off-by: Qi Zheng
Suggested-by: Kevin Brodsky
Reviewed-by: Kevin Brodsky
---
arch/arm/include/asm/tlb.h | 4
include/asm-generic/tlb.h | 10 --
2 files changed, 8 insertions(+), 6 deletions(-)
diff --git a/arch/arm/include/asm/tlb.h b/arch/arm/include/asm/
The __tlb_remove_table_one() in x86 does not contain architecture-specific
content, so move it to the generic file.
Signed-off-by: Qi Zheng
Reviewed-by: Kevin Brodsky
---
arch/x86/include/asm/tlb.h | 19 ---
mm/mmu_gather.c| 20 ++--
2 files changed
Like PMD and PTE level page table, also add statistics for PUD and P4D
page table.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Reviewed-by: Kevin Brodsky
Cc: linux-s...@vger.kernel.org
---
arch/s390/include/asm/pgalloc.h | 29 +
arch/s390/include
Like other levels of page tables, also use mmu gather mechanism to free
p4d level page table.
Signed-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
Reviewed-by: Kevin Brodsky
Cc: linux-arm-ker...@lists.infradead.org
---
arch/arm64/include/asm/pgalloc.h | 1 -
arch/arm64/include/asm
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Reviewed-by: Kevin Brodsky
Cc: linux-arm-ker...@lists.infradead.org
---
arch/arm/include/asm/tlb.h | 9 ++---
1 file chang
Call pagetable_dtor() for PMD|PUD|P4D tables just before ptdesc is
freed - same as it is done for PTE tables. That allows consolidating
TLB free paths for all table types.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Reviewed-by: Kevin Brodsky
Cc: linux-s...@vger.kernel.org
Move pagetable_dtor() to __tlb_remove_table(), so that ptlock and page
table pages can be freed together (regardless of whether RCU is used).
This prevents the use-after-free problem where the ptlock is freed
immediately but the page table pages is freed later via RCU.
Signed-off-by: Qi Zheng
Hi Andrew,
On 2024/12/30 12:55, Andrew Morton wrote:
On Mon, 30 Dec 2024 11:12:00 +0800 Qi Zheng wrote:
For now struct ptdesc overlaps struct page, but the goal is to have them
separate and always operate on struct ptdesc when working with page tables.
OK, so tlb_remove_page_ptdesc() and
To unify the PxD and PTE TLB free path, also move the pagetable_dtor() of
PMD|PUD|P4D to __tlb_remove_table().
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: linux-s...@vger.kernel.org
---
arch/s390/include/asm/tlb.h | 3 ---
arch/s390/mm/pgalloc.c | 14
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: x...@kernel.org
---
arch/x86/include/asm/tlb.h | 17 ++---
arch/x86/kernel/paravirt.c | 1 +
arch/x86/mm/pgta
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: linux-arm-ker...@lists.infradead.org
---
arch/arm64/include/asm/tlb.h | 10 --
1 file changed, 4 insertions(+), 6 dele
loc: Provide generic p4d_{alloc_one,free}
Qi Zheng (13):
Revert "mm: pgtable: make ptlock be freed by RCU"
mm: pgtable: add statistics for P4D level page table
arm64: pgtable: use mmu gather to free p4d level page table
s390: pgtable: add statistics for PUD and P4D level page tab
generic version, so we remove it altogether.
This is consistent with the way arm64 and x86 handle this situation
(runtime check in p4d_free() only).
Signed-off-by: Kevin Brodsky
Acked-by: Dave Hansen
Signed-off-by: Qi Zheng
Acked-by: Palmer Dabbelt
---
arch/riscv/include/asm/pgalloc.h | 22
Like other levels of page tables, add statistics for P4D level page table.
Signed-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
---
arch/riscv/include/asm/pgalloc.h | 6 +-
arch/x86/mm/pgtable.c| 3 +++
include/asm-generic/pgalloc.h| 2 ++
include/linux/mm.h
Like other levels of page tables, also use mmu gather mechanism to free
p4d level page table.
Signed-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
Cc: linux-arm-ker...@lists.infradead.org
---
arch/arm64/include/asm/pgalloc.h | 1 -
arch/arm64/include/asm/tlb.h | 14
d-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
---
Documentation/mm/split_page_table_lock.rst | 4 +-
arch/arm/include/asm/tlb.h | 4 +-
arch/arm64/include/asm/tlb.h | 8 ++--
arch/csky/include/asm/pgalloc.h| 2 +-
arch/hexagon/include/asm/p
Hi Mike,
On 2024/12/28 17:26, Mike Rapoport wrote:
On Mon, Dec 23, 2024 at 05:41:01PM +0800, Qi Zheng wrote:
Here we are explicitly dealing with struct page, and the following logic
semms strange:
tlb_remove_page_ptdesc((tlb), (page_ptdesc(pte)));
tlb_remove_page_ptdesc
--> tlb_remove_p
This reverts commit 2f3443770437e49abc39af26962d293851cbab6d.
Signed-off-by: Qi Zheng
---
include/linux/mm.h | 2 +-
include/linux/mm_types.h | 9 +
mm/memory.c | 22 ++
3 files changed, 8 insertions(+), 25 deletions(-)
diff --git a/include
orted architectures, so we might as well include this check
in the generic implementation. No runtime check is required in
p4d_alloc_one() as the top-level p4d_alloc() already does the
required check.
Signed-off-by: Kevin Brodsky
Acked-by: Dave Hansen
Signed-off-by: Qi Zheng
---
arch/arm64/includ
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
By the way, move the comment above __tlb_remove_table() to
riscv_tlb_remove_ptdesc(), it will be more appropriate.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: lin
Several architectures (arm, arm64, riscv and x86) define exactly the
same __tlb_remove_table(), just introduce generic __tlb_remove_table() to
eliminate these duplications.
The s390 __tlb_remove_table() is nearly the same, so also make s390
__tlb_remove_table() version generic.
Signed-off-by: Qi
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: linux-arm-ker...@lists.infradead.org
---
arch/arm/include/asm/tlb.h | 9 -
1 file changed, 4 insertions(+), 5 dele
Like PMD and PTE level page table, also add statistics for PUD and P4D
page table.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: linux-s...@vger.kernel.org
---
arch/s390/include/asm/pgalloc.h | 29 +++---
arch/s390/include/asm/tlb.h | 37
The __tlb_remove_table_one() in x86 does not contain architecture-specific
content, so move it to the generic file.
Signed-off-by: Qi Zheng
---
arch/x86/include/asm/tlb.h | 19 ---
mm/mmu_gather.c| 20 ++--
2 files changed, 18 insertions(+), 21
s390 also calls generic pagetable_dtor_free().
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
---
arch/s390/mm/pgalloc.c| 6 --
include/asm-generic/pgalloc.h | 12
include/asm-generic/tlb.h | 3 +--
include/linux/mm.h| 6 ++
4 files c
: Kevin Brodsky
Acked-by: Dave Hansen
Signed-off-by: Qi Zheng
---
arch/riscv/include/asm/pgalloc.h | 22 --
1 file changed, 4 insertions(+), 18 deletions(-)
diff --git a/arch/riscv/include/asm/pgalloc.h
b/arch/riscv/include/asm/pgalloc.h
index f52264304f772..8ad0bbe838a24
On 2024/12/20 22:28, Kevin Brodsky wrote:
On 20/12/2024 15:16, Qi Zheng wrote:
Did I miss something?
My patch series is not only for cleanup, but also for fixes of
UAF issue [1], so is it possible to rebase your patch series onto
mine? I can post v3 ASAP.
I see, yours should be merged
On 2024/12/20 21:50, Kevin Brodsky wrote:
On 20/12/2024 12:46, Qi Zheng wrote:
Hi Kevin,
On 2024/12/20 18:49, Kevin Brodsky wrote:
[...]
Qi, shall we collaborate to make our series complementary? I believe my
series covers patch 2 and 4 of your series, but it goes further by
covering all
Hi Kevin,
On 2024/12/20 18:49, Kevin Brodsky wrote:
Hi Peter, Qi,
On 19/12/2024 18:19, Peter Zijlstra wrote:
On Thu, Dec 19, 2024 at 04:44:16PM +, Kevin Brodsky wrote:
Besides the ptlock management at PTE/PMD level, all the
pagetable_*_[cd]tor have the same implementation. Introduce commo
lloc_one
asm-generic: pgalloc: Provide generic p4d_{alloc_one,free}
Qi Zheng (15):
Revert "mm: pgtable: make ptlock be freed by RCU"
mm: pgtable: add statistics for P4D level page table
arm64: pgtable: use mmu gather to free p4d level page table
s390: pgtable: add statist
This reverts commit 2f3443770437e49abc39af26962d293851cbab6d.
Signed-off-by: Qi Zheng
---
include/linux/mm.h | 2 +-
include/linux/mm_types.h | 9 +
mm/memory.c | 22 ++
3 files changed, 8 insertions(+), 25 deletions(-)
diff --git a/include
generic version, so we remove it altogether.
This is consistent with the way arm64 and x86 handle this situation
(runtime check in p4d_free() only).
Signed-off-by: Kevin Brodsky
Acked-by: Dave Hansen
Signed-off-by: Qi Zheng
---
arch/riscv/include/asm/pgalloc.h | 22 --
1 file
Like PMD and PTE level page table, also add statistics for PUD and P4D
page table.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: linux-s...@vger.kernel.org
---
arch/s390/include/asm/pgalloc.h | 29 +++---
arch/s390/include/asm/tlb.h | 37
Like other levels of page tables, also use mmu gather mechanism to free
p4d level page table.
Signed-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
Cc: linux-arm-ker...@lists.infradead.org
---
arch/arm64/include/asm/pgalloc.h | 1 -
arch/arm64/include/asm/tlb.h | 14
To unify the PxD and PTE TLB free path, also move the pagetable_dtor() of
PMD|PUD|P4D to __tlb_remove_table().
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: linux-s...@vger.kernel.org
---
arch/s390/include/asm/tlb.h | 3 ---
arch/s390/mm/pgalloc.c | 14
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
By the way, move the comment above __tlb_remove_table() to
riscv_tlb_remove_ptdesc(), it will be more appropriate.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: lin
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: linux-arm-ker...@lists.infradead.org
---
arch/arm64/include/asm/tlb.h | 10 --
1 file changed, 4 insertions(+), 6 dele
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: x...@kernel.org
---
arch/x86/include/asm/tlb.h | 17 ++---
arch/x86/kernel/paravirt.c | 1 +
arch/x86/mm/pgta
Several architectures (arm, arm64, riscv and x86) define exactly the
same __tlb_remove_table(), just introduce generic __tlb_remove_table() to
eliminate these duplications.
The s390 __tlb_remove_table() is nearly the same, so also make s390
__tlb_remove_table() version generic.
Signed-off-by: Qi
The __tlb_remove_table_one() in x86 does not contain architecture-specific
content, so move it to the generic file.
Signed-off-by: Qi Zheng
---
arch/x86/include/asm/tlb.h | 19 ---
mm/mmu_gather.c| 20 ++--
2 files changed, 18 insertions(+), 21
tly.
Signed-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
---
arch/csky/include/asm/pgalloc.h | 2 +-
arch/hexagon/include/asm/pgalloc.h | 2 +-
arch/loongarch/include/asm/pgalloc.h | 2 +-
arch/m68k/include/asm/sun3_pgalloc.h | 2 +-
arch/mips/include/asm/pgalloc.h | 2 +-
arch/ni
Just like removing tlb_remove_page_ptdesc(), remove tlb_remove_ptdesc() as
well, and make callers call tlb_remove_table() directly.
Signed-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
---
arch/arm/include/asm/tlb.h | 8 ++--
arch/arm64/include/asm/tlb.h | 16
x27;s s390 also calls generic pagetable_dtor_free().
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
---
arch/s390/mm/pgalloc.c| 18 ++
include/asm-generic/pgalloc.h | 23 ---
include/asm-generic/tlb.h | 5 +
include/
orted architectures, so we might as well include this check
in the generic implementation. No runtime check is required in
p4d_alloc_one() as the top-level p4d_alloc() already does the
required check.
Signed-off-by: Kevin Brodsky
Acked-by: Dave Hansen
Signed-off-by: Qi Zheng
---
arch/arm64/includ
Like other levels of page tables, add statistics for P4D level page table.
Signed-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
---
arch/riscv/include/asm/pgalloc.h | 6 +-
arch/x86/mm/pgtable.c| 3 +++
include/asm-generic/pgalloc.h| 2 ++
include/linux/mm.h
d-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
---
Documentation/mm/split_page_table_lock.rst | 4 +-
arch/arm/include/asm/tlb.h | 4 +-
arch/arm64/include/asm/tlb.h | 8 ++--
arch/csky/include/asm/pgalloc.h| 2 +-
arch/hexagon/include/asm/p
have swap cache, so use pagetable_free() instead of
free_page_and_swap_cache() to free page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Cc: linux-arm-ker...@lists.infradead.org
---
arch/arm/include/asm/tlb.h | 9 -
1 file changed, 4 insertions(+), 5 dele
.
Signed-off-by: Qi Zheng
---
include/asm-generic/tlb.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/include/asm-generic/tlb.h b/include/asm-generic/tlb.h
index e402aef79c93e..10cdbe0e7adfe 100644
--- a/include/asm-generic/tlb.h
+++ b/include/asm-generic/tlb.h
@@ -228,10
/lkml/20250123042447.2259648-2-r...@surriel.com/
Qi Zheng (5):
mm: pgtable: make generic tlb_remove_table() use struct ptdesc
mm: pgtable: convert some architectures to use tlb_remove_ptdesc()
riscv: pgtable: unconditionally use tlb_remove_ptdesc()
x86: pgtable: unconditionally use tlb_remove_
If the CONFIG_MMU_GATHER_TABLE_FREE is disabled, the tlb_remove_table()
will fall back to pagetable_dtor() + tlb_remove_page(). So let's use
tlb_remove_table() unconditionally to free page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
---
arch/x86/kernel/parav
system.
So, in summary, we can use tlb_remove_ptdesc() to support fast gup in all
cases, and this interface is specifically designed for page table pages.
So let's use it unconditionally.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
---
arch/riscv/include/asm/pgalloc.h | 26
|table}() under !CONFIG_MMU_GATHER_TABLE_FREE, so
convert these architectures to use tlb_remove_ptdesc().
The ultimate goal is to make the architecture only use tlb_remove_ptdesc()
or tlb_remove_table() for page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
---
arch
ff-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
---
include/asm-generic/tlb.h | 6 --
1 file changed, 6 deletions(-)
diff --git a/include/asm-generic/tlb.h b/include/asm-generic/tlb.h
index 10cdbe0e7adfe..dec030cb12108 100644
--- a/include/asm-generic/tlb.h
+++ b/include/asm-generic/
On 2025/1/24 19:38, Peter Zijlstra wrote:
On Thu, Jan 23, 2025 at 09:26:17PM +0800, Qi Zheng wrote:
If the CONFIG_MMU_GATHER_TABLE_FREE is disabled, the tlb_remove_table()
will fall back to pagetable_dtor() + tlb_remove_page(). So let's use
tlb_remove_table() unconditionally to free
On 2025/1/24 19:47, Peter Zijlstra wrote:
On Thu, Jan 23, 2025 at 09:26:13PM +0800, Qi Zheng wrote:
Hi all,
As suggested by Peter Zijlstra below [1], this series aims to remove
tlb_remove_page_ptdesc().
: Fundamentally tlb_remove_page() is about removing *pages* as from a PTE,
: there
Hi Will,
On 2025/1/14 00:26, Will Deacon wrote:
On Wed, Jan 08, 2025 at 02:57:21PM +0800, Qi Zheng wrote:
Like other levels of page tables, also use mmu gather mechanism to free
p4d level page table.
Signed-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
Reviewed-by: Kevin Brodsky
On 2025/1/3 16:02, Kevin Brodsky wrote:
On 03/01/2025 04:48, Qi Zheng wrote:
Hi Kevin,
On 2025/1/3 00:53, Kevin Brodsky wrote:
On 30/12/2024 10:07, Qi Zheng wrote:
static inline void riscv_tlb_remove_ptdesc(struct mmu_gather *tlb,
void *pt)
{
- if (riscv_use_sbi_for_rfence
On 2025/1/3 17:13, Qi Zheng wrote:
On 2025/1/3 16:02, Kevin Brodsky wrote:
On 03/01/2025 04:48, Qi Zheng wrote:
Hi Kevin,
On 2025/1/3 00:53, Kevin Brodsky wrote:
On 30/12/2024 10:07, Qi Zheng wrote:
static inline void riscv_tlb_remove_ptdesc(struct mmu_gather *tlb,
void *pt
Hi Kevin,
On 2025/1/3 00:53, Kevin Brodsky wrote:
On 30/12/2024 10:07, Qi Zheng wrote:
static inline void riscv_tlb_remove_ptdesc(struct mmu_gather *tlb, void *pt)
{
- if (riscv_use_sbi_for_rfence())
+ if (riscv_use_sbi_for_rfence()) {
tlb_remove_ptdesc(tlb, pt
On 2025/1/3 00:53, Kevin Brodsky wrote:
On 30/12/2024 10:07, Qi Zheng wrote:
diff --git a/arch/riscv/include/asm/pgalloc.h b/arch/riscv/include/asm/pgalloc.h
index 551d614d3369c..3466fbe2e508d 100644
--- a/arch/riscv/include/asm/pgalloc.h
+++ b/arch/riscv/include/asm/pgalloc.h
@@ -108,8
On 2025/1/3 01:00, Kevin Brodsky wrote:
On 30/12/2024 10:07, Qi Zheng wrote:
Qi Zheng (13):
Revert "mm: pgtable: make ptlock be freed by RCU"
mm: pgtable: add statistics for P4D level page table
arm64: pgtable: use mmu gather to free p4d level page table
s390: pg
.
Comments and suggestions are welcome!
Thanks,
Qi
[1].
https://lore.kernel.org/all/20241211133433.gc12...@noisy.programming.kicks-ass.net/
[2]. https://lore.kernel.org/all/67548279.050a0220.a30f1.015b@google.com/
Kevin Brodsky (2):
riscv: mm: Skip pgtable level check in {pud,p4d}_alloc_one
asm-gen
This reverts commit 2f3443770437e49abc39af26962d293851cbab6d.
Signed-off-by: Qi Zheng
Reviewed-by: Kevin Brodsky
---
include/linux/mm.h | 2 +-
include/linux/mm_types.h | 9 +
mm/memory.c | 22 ++
3 files changed, 8 insertions(+), 25 deletions
generic version, so we remove it altogether.
This is consistent with the way arm64 and x86 handle this situation
(runtime check in p4d_free() only).
Signed-off-by: Kevin Brodsky
Signed-off-by: Qi Zheng
Acked-by: Dave Hansen
Reviewed-by: Alexandre Ghiti
---
arch/riscv/include/asm/pgalloc.h | 22
orted architectures, so we might as well include this check
in the generic implementation. No runtime check is required in
p4d_alloc_one() as the top-level p4d_alloc() already does the
required check.
Signed-off-by: Kevin Brodsky
Acked-by: Dave Hansen
Signed-off-by: Qi Zheng
---
arch/arm64/includ
On 2/27/25 9:21 PM, Geert Uytterhoeven wrote:
Hi Qi,
On Tue, 25 Feb 2025 at 04:46, Qi Zheng wrote:
Now, the nine architectures of csky, hexagon, loongarch, m68k, mips,
nios2, openrisc, sh and um do not select CONFIG_MMU_GATHER_RCU_TABLE_FREE,
and just call pagetable_dtor
On 3/3/25 9:45 AM, yunhui cui wrote:
Hi Qi,
[...]
This set of patches mainly refactors the remove page table function
interfaces. The comment "riscv_ipi_for_rfence" has been removed, as it
was no longer needed, and this patch handles that.
Additionally, whether riscv_use_sbi_for_rfen
|table}() under !CONFIG_MMU_GATHER_TABLE_FREE, so
convert these architectures to use tlb_remove_ptdesc().
The ultimate goal is to make the architecture only use tlb_remove_ptdesc()
or tlb_remove_table() for page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
Reviewed
On 3/4/25 7:53 AM, Andrew Morton wrote:
On Mon, 3 Mar 2025 15:26:03 +0800 Qi Zheng wrote:
Now, the nine architectures of csky, hexagon, loongarch, m68k, mips,
nios2, openrisc, sh and um do not select CONFIG_MMU_GATHER_RCU_TABLE_FREE,
and just call pagetable_dtor() + tlb_remove_page_ptdesc
On 3/4/25 12:08 PM, Andrew Morton wrote:
On Tue, 4 Mar 2025 10:31:07 +0800 Qi Zheng wrote:
On 3/4/25 7:53 AM, Andrew Morton wrote:
On Mon, 3 Mar 2025 15:26:03 +0800 Qi Zheng wrote:
Now, the nine architectures of csky, hexagon, loongarch, m68k, mips,
nios2, openrisc, sh and um do not
[1].
https://lore.kernel.org/linux-mm/20250103111457.gc22...@noisy.programming.kicks-ass.net/
Qi Zheng (6):
mm: pgtable: make generic tlb_remove_table() use struct ptdesc
mm: pgtable: change pt parameter of tlb_remove_ptdesc() to struct
ptdesc *
mm: pgtable: convert some architectures
|table}() under !CONFIG_MMU_GATHER_TABLE_FREE, so
convert these architectures to use tlb_remove_ptdesc().
The ultimate goal is to make the architecture only use tlb_remove_ptdesc()
or tlb_remove_table() for page table pages.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
---
arch
.
Signed-off-by: Qi Zheng
---
include/asm-generic/tlb.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/include/asm-generic/tlb.h b/include/asm-generic/tlb.h
index b35b36fa7aabf..54f579750c8e5 100644
--- a/include/asm-generic/tlb.h
+++ b/include/asm-generic/tlb.h
@@ -232,10
All callers of tlb_remove_ptdesc() pass it a pointer of struct ptdesc, so
let's change the pt parameter from void * to struct ptdesc * to perform
a type safety check.
Signed-off-by: Qi Zheng
Originally-by: Peter Zijlstra (Intel)
---
include/asm-generic/tlb.h | 2 +-
1 file changed, 1 inse
The x86 has already been converted to use struct ptdesc, so convert it to
use tlb_remove_ptdesc() instead of tlb_remove_table().
Signed-off-by: Qi Zheng
---
arch/x86/mm/pgtable.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/arch/x86/mm/pgtable.c b/arch/x86/mm
ff-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
---
include/asm-generic/tlb.h | 6 --
1 file changed, 6 deletions(-)
diff --git a/include/asm-generic/tlb.h b/include/asm-generic/tlb.h
index 18bf499ef8801..dd292c6d3ce88 100644
--- a/include/asm-generic/tlb.h
+++ b/include/asm-generic/
system.
So, in summary, we can use tlb_remove_ptdesc() to support fast gup in all
cases, and this interface is specifically designed for page table pages.
So let's use it unconditionally.
Signed-off-by: Qi Zheng
Suggested-by: Peter Zijlstra (Intel)
---
arch/riscv/include/asm/pgalloc.h | 26
On 2/26/25 上午10:35, Andrew Morton wrote:
On Tue, 25 Feb 2025 11:45:55 +0800 Qi Zheng wrote:
The x86 has already been converted to use struct ptdesc, so convert it to
use tlb_remove_ptdesc() instead of tlb_remove_table().
This is dependent upon Rik's a37259732a7dc ("x8
98 matches
Mail list logo
