All of lore.kernel.org
 help / color / mirror / Atom feed
From: Paul Walmsley <pjw@kernel.org>
To: Xu Lu <luxu.kernel@bytedance.com>
Cc: paul.walmsley@sifive.com, palmer@dabbelt.com,
	aou@eecs.berkeley.edu,  alex@ghiti.fr,
	linux-riscv@lists.infradead.org,  linux-kernel@vger.kernel.org,
	apw@canonical.com, joe@perches.com
Subject: Re: [PATCH RESEND 1/2] riscv: mm: Apply svinval in update_mmu_cache()
Date: Fri, 10 Jul 2026 18:04:22 -0600 (MDT)	[thread overview]
Message-ID: <dcd61aba-8622-b14d-6466-44a7f0dc0615@kernel.org> (raw)
In-Reply-To: <20250901114141.5438-2-luxu.kernel@bytedance.com>

Hi,

On Mon, 1 Sep 2025, Xu Lu wrote:

> Only flush tlb of the specified mm, and apply svinval if available.
> 
> Signed-off-by: Xu Lu <luxu.kernel@bytedance.com>

I wound up splitting this into two separate patches, rather than one 
patch, since there seem to be two orthogonal changes.  The first change
restricts the sfence.vma to a particular ASID (below).  The second change 
involves the Svinval path, and is sent in a subsequent message.
What do you think?


- Paul


From: Xu Lu <luxu.kernel@bytedance.com>
Date: Mon, 1 Sep 2025 19:41:40 +0800
Subject: [PATCH 1/2] riscv: mm: flush TLB by ASID in update_mmu_cache_range()

Only flush the TLB of the specified mm (via local_flush_tlb_page_asid)

Signed-off-by: Xu Lu <luxu.kernel@bytedance.com>
Link: https://patch.msgid.link/20250901114141.5438-2-luxu.kernel@bytedance.com
[pjw@kernel.org: split the non-Svinval code in the original into this patch; update description]
Signed-off-by: Paul Walmsley <pjw@kernel.org>
---
 arch/riscv/include/asm/pgtable.h  | 7 +++++--
 arch/riscv/include/asm/tlbflush.h | 5 +++++
 arch/riscv/mm/tlbflush.c          | 5 -----
 3 files changed, 10 insertions(+), 7 deletions(-)

diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
index 5d5756bda82e..755495a542cc 100644
--- a/arch/riscv/include/asm/pgtable.h
+++ b/arch/riscv/include/asm/pgtable.h
@@ -568,6 +568,8 @@ static inline void update_mmu_cache_range(struct vm_fault *vmf,
 		struct vm_area_struct *vma, unsigned long address,
 		pte_t *ptep, unsigned int nr)
 {
+	unsigned long asid;
+
 	/*
 	 * Svvptc guarantees that the new valid pte will be visible within
 	 * a bounded timeframe, so when the uarch does not cache invalid
@@ -583,10 +585,11 @@ static inline void update_mmu_cache_range(struct vm_fault *vmf,
 	 * Relying on flush_tlb_fix_spurious_fault would suffice, but
 	 * the extra traps reduce performance.  So, eagerly SFENCE.VMA.
 	 */
+	asid = get_mm_asid(vma->vm_mm);
 	while (nr--)
-		local_flush_tlb_page(address + nr * PAGE_SIZE);
-
+		local_flush_tlb_page_asid(address + nr * PAGE_SIZE, asid);
 }
+
 #define update_mmu_cache(vma, addr, ptep) \
 	update_mmu_cache_range(NULL, vma, addr, ptep, 1)
 
diff --git a/arch/riscv/include/asm/tlbflush.h b/arch/riscv/include/asm/tlbflush.h
index eed0abc40514..7c2cd5cc92d3 100644
--- a/arch/riscv/include/asm/tlbflush.h
+++ b/arch/riscv/include/asm/tlbflush.h
@@ -15,6 +15,11 @@
 #define FLUSH_TLB_NO_ASID       ((unsigned long)-1)
 
 #ifdef CONFIG_MMU
+static inline unsigned long get_mm_asid(struct mm_struct *mm)
+{
+	return mm ? cntx2asid(atomic_long_read(&mm->context.id)) : FLUSH_TLB_NO_ASID;
+}
+
 static inline void local_flush_tlb_all(void)
 {
 	__asm__ __volatile__ ("sfence.vma" : : : "memory");
diff --git a/arch/riscv/mm/tlbflush.c b/arch/riscv/mm/tlbflush.c
index 8404530ec00f..73c226f719c7 100644
--- a/arch/riscv/mm/tlbflush.c
+++ b/arch/riscv/mm/tlbflush.c
@@ -110,11 +110,6 @@ static void __ipi_flush_tlb_range_asid(void *info)
 	local_flush_tlb_range_asid(d->start, d->size, d->stride, d->asid);
 }
 
-static inline unsigned long get_mm_asid(struct mm_struct *mm)
-{
-	return mm ? cntx2asid(atomic_long_read(&mm->context.id)) : FLUSH_TLB_NO_ASID;
-}
-
 static void __flush_tlb_range(struct mm_struct *mm,
 			      const struct cpumask *cmask,
 			      unsigned long start, unsigned long size,
-- 
2.53.0




_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv

WARNING: multiple messages have this Message-ID (diff)
From: Paul Walmsley <pjw@kernel.org>
To: Xu Lu <luxu.kernel@bytedance.com>
Cc: paul.walmsley@sifive.com, palmer@dabbelt.com,
	aou@eecs.berkeley.edu,  alex@ghiti.fr,
	linux-riscv@lists.infradead.org,  linux-kernel@vger.kernel.org,
	apw@canonical.com, joe@perches.com
Subject: Re: [PATCH RESEND 1/2] riscv: mm: Apply svinval in update_mmu_cache()
Date: Fri, 10 Jul 2026 18:04:22 -0600 (MDT)	[thread overview]
Message-ID: <dcd61aba-8622-b14d-6466-44a7f0dc0615@kernel.org> (raw)
In-Reply-To: <20250901114141.5438-2-luxu.kernel@bytedance.com>

Hi,

On Mon, 1 Sep 2025, Xu Lu wrote:

> Only flush tlb of the specified mm, and apply svinval if available.
> 
> Signed-off-by: Xu Lu <luxu.kernel@bytedance.com>

I wound up splitting this into two separate patches, rather than one 
patch, since there seem to be two orthogonal changes.  The first change
restricts the sfence.vma to a particular ASID (below).  The second change 
involves the Svinval path, and is sent in a subsequent message.
What do you think?


- Paul


From: Xu Lu <luxu.kernel@bytedance.com>
Date: Mon, 1 Sep 2025 19:41:40 +0800
Subject: [PATCH 1/2] riscv: mm: flush TLB by ASID in update_mmu_cache_range()

Only flush the TLB of the specified mm (via local_flush_tlb_page_asid)

Signed-off-by: Xu Lu <luxu.kernel@bytedance.com>
Link: https://patch.msgid.link/20250901114141.5438-2-luxu.kernel@bytedance.com
[pjw@kernel.org: split the non-Svinval code in the original into this patch; update description]
Signed-off-by: Paul Walmsley <pjw@kernel.org>
---
 arch/riscv/include/asm/pgtable.h  | 7 +++++--
 arch/riscv/include/asm/tlbflush.h | 5 +++++
 arch/riscv/mm/tlbflush.c          | 5 -----
 3 files changed, 10 insertions(+), 7 deletions(-)

diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
index 5d5756bda82e..755495a542cc 100644
--- a/arch/riscv/include/asm/pgtable.h
+++ b/arch/riscv/include/asm/pgtable.h
@@ -568,6 +568,8 @@ static inline void update_mmu_cache_range(struct vm_fault *vmf,
 		struct vm_area_struct *vma, unsigned long address,
 		pte_t *ptep, unsigned int nr)
 {
+	unsigned long asid;
+
 	/*
 	 * Svvptc guarantees that the new valid pte will be visible within
 	 * a bounded timeframe, so when the uarch does not cache invalid
@@ -583,10 +585,11 @@ static inline void update_mmu_cache_range(struct vm_fault *vmf,
 	 * Relying on flush_tlb_fix_spurious_fault would suffice, but
 	 * the extra traps reduce performance.  So, eagerly SFENCE.VMA.
 	 */
+	asid = get_mm_asid(vma->vm_mm);
 	while (nr--)
-		local_flush_tlb_page(address + nr * PAGE_SIZE);
-
+		local_flush_tlb_page_asid(address + nr * PAGE_SIZE, asid);
 }
+
 #define update_mmu_cache(vma, addr, ptep) \
 	update_mmu_cache_range(NULL, vma, addr, ptep, 1)
 
diff --git a/arch/riscv/include/asm/tlbflush.h b/arch/riscv/include/asm/tlbflush.h
index eed0abc40514..7c2cd5cc92d3 100644
--- a/arch/riscv/include/asm/tlbflush.h
+++ b/arch/riscv/include/asm/tlbflush.h
@@ -15,6 +15,11 @@
 #define FLUSH_TLB_NO_ASID       ((unsigned long)-1)
 
 #ifdef CONFIG_MMU
+static inline unsigned long get_mm_asid(struct mm_struct *mm)
+{
+	return mm ? cntx2asid(atomic_long_read(&mm->context.id)) : FLUSH_TLB_NO_ASID;
+}
+
 static inline void local_flush_tlb_all(void)
 {
 	__asm__ __volatile__ ("sfence.vma" : : : "memory");
diff --git a/arch/riscv/mm/tlbflush.c b/arch/riscv/mm/tlbflush.c
index 8404530ec00f..73c226f719c7 100644
--- a/arch/riscv/mm/tlbflush.c
+++ b/arch/riscv/mm/tlbflush.c
@@ -110,11 +110,6 @@ static void __ipi_flush_tlb_range_asid(void *info)
 	local_flush_tlb_range_asid(d->start, d->size, d->stride, d->asid);
 }
 
-static inline unsigned long get_mm_asid(struct mm_struct *mm)
-{
-	return mm ? cntx2asid(atomic_long_read(&mm->context.id)) : FLUSH_TLB_NO_ASID;
-}
-
 static void __flush_tlb_range(struct mm_struct *mm,
 			      const struct cpumask *cmask,
 			      unsigned long start, unsigned long size,
-- 
2.53.0




  reply	other threads:[~2026-07-11  0:04 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-09-01 11:41 [PATCH RESEND 0/2] riscv: mm: Some optimizations for tlb flush Xu Lu
2025-09-01 11:41 ` Xu Lu
2025-09-01 11:41 ` [PATCH RESEND 1/2] riscv: mm: Apply svinval in update_mmu_cache() Xu Lu
2025-09-01 11:41   ` Xu Lu
2026-07-11  0:04   ` Paul Walmsley [this message]
2026-07-11  0:04     ` Paul Walmsley
2026-07-13  6:29     ` [External] " Xu Lu
2026-07-13  6:29       ` Xu Lu
2026-07-13 17:49       ` Paul Walmsley
2026-07-13 17:49         ` Paul Walmsley
2026-07-11  0:07   ` Paul Walmsley
2026-07-11  0:07     ` Paul Walmsley
2026-07-15  1:14   ` Klara Modin
2026-07-15  1:14     ` Klara Modin
2025-09-01 11:41 ` [PATCH RESEND 2/2] riscv: mm: Clear cpu in mm_cpumask after local_flush_tlb_all_asid Xu Lu
2025-09-01 11:41   ` Xu Lu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=dcd61aba-8622-b14d-6466-44a7f0dc0615@kernel.org \
    --to=pjw@kernel.org \
    --cc=alex@ghiti.fr \
    --cc=aou@eecs.berkeley.edu \
    --cc=apw@canonical.com \
    --cc=joe@perches.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=luxu.kernel@bytedance.com \
    --cc=palmer@dabbelt.com \
    --cc=paul.walmsley@sifive.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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.