All of lore.kernel.org
 help / color / mirror / Atom feed
From: Samuel Holland <samuel@sholland.org>
To: Palmer Dabbelt <palmer@dabbelt.com>,
	Alexandre Ghiti <alexghiti@rivosinc.com>,
	linux-riscv@lists.infradead.org
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Samuel Holland <samuel@sholland.org>
Subject: [PATCH 5/7] riscv: mm: Preserve global TLB entries when switching contexts
Date: Sat,  9 Sep 2023 15:16:33 -0500	[thread overview]
Message-ID: <20230909201727.10909-6-samuel@sholland.org> (raw)
In-Reply-To: <20230909201727.10909-1-samuel@sholland.org>

If the CPU does not support multiple ASIDs, all MM contexts use ASID 0.
In this case, it is still beneficial to flush the TLB by ASID, as the
single-ASID variant of the sfence.vma instruction preserves TLB entries
for global (kernel) pages.

This optimization is recommended by the RISC-V privileged specification:

  If the implementation does not provide ASIDs, or software chooses
  to always use ASID 0, then after every satp write, software should
  execute SFENCE.VMA with rs1=x0. In the common case that no global
  translations have been modified, rs2 should be set to a register
  other than x0 but which contains the value zero, so that global
  translations are not flushed.

It is not possible to apply this optimization when using the ASID
allocator, because that code must flush the TLB for all ASIDs at once
when incrementing the version number.

Signed-off-by: Samuel Holland <samuel@sholland.org>
---

 arch/riscv/mm/context.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/riscv/mm/context.c b/arch/riscv/mm/context.c
index 43a8bc2d5af4..3ca9b653df7d 100644
--- a/arch/riscv/mm/context.c
+++ b/arch/riscv/mm/context.c
@@ -200,7 +200,7 @@ static void set_mm_noasid(struct mm_struct *mm)
 {
 	/* Switch the page table and blindly nuke entire local TLB */
 	csr_write(CSR_SATP, virt_to_pfn(mm->pgd) | satp_mode);
-	local_flush_tlb_all();
+	local_flush_tlb_all_asid(0);
 }
 
 static inline void set_mm(struct mm_struct *prev,
-- 
2.41.0


WARNING: multiple messages have this Message-ID (diff)
From: Samuel Holland <samuel@sholland.org>
To: Palmer Dabbelt <palmer@dabbelt.com>,
	Alexandre Ghiti <alexghiti@rivosinc.com>,
	linux-riscv@lists.infradead.org
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Samuel Holland <samuel@sholland.org>
Subject: [PATCH 5/7] riscv: mm: Preserve global TLB entries when switching contexts
Date: Sat,  9 Sep 2023 15:16:33 -0500	[thread overview]
Message-ID: <20230909201727.10909-6-samuel@sholland.org> (raw)
In-Reply-To: <20230909201727.10909-1-samuel@sholland.org>

If the CPU does not support multiple ASIDs, all MM contexts use ASID 0.
In this case, it is still beneficial to flush the TLB by ASID, as the
single-ASID variant of the sfence.vma instruction preserves TLB entries
for global (kernel) pages.

This optimization is recommended by the RISC-V privileged specification:

  If the implementation does not provide ASIDs, or software chooses
  to always use ASID 0, then after every satp write, software should
  execute SFENCE.VMA with rs1=x0. In the common case that no global
  translations have been modified, rs2 should be set to a register
  other than x0 but which contains the value zero, so that global
  translations are not flushed.

It is not possible to apply this optimization when using the ASID
allocator, because that code must flush the TLB for all ASIDs at once
when incrementing the version number.

Signed-off-by: Samuel Holland <samuel@sholland.org>
---

 arch/riscv/mm/context.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/riscv/mm/context.c b/arch/riscv/mm/context.c
index 43a8bc2d5af4..3ca9b653df7d 100644
--- a/arch/riscv/mm/context.c
+++ b/arch/riscv/mm/context.c
@@ -200,7 +200,7 @@ static void set_mm_noasid(struct mm_struct *mm)
 {
 	/* Switch the page table and blindly nuke entire local TLB */
 	csr_write(CSR_SATP, virt_to_pfn(mm->pgd) | satp_mode);
-	local_flush_tlb_all();
+	local_flush_tlb_all_asid(0);
 }
 
 static inline void set_mm(struct mm_struct *prev,
-- 
2.41.0


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

  parent reply	other threads:[~2023-09-09 20:17 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-09 20:16 [PATCH 0/7] riscv: ASID-related and UP-related TLB flush enhancements Samuel Holland
2023-09-09 20:16 ` Samuel Holland
2023-09-09 20:16 ` [PATCH 1/7] riscv: Apply SiFive CIP-1200 workaround to single-ASID sfence.vma Samuel Holland
2023-09-09 20:16   ` Samuel Holland
2023-09-09 20:16 ` [PATCH 2/7] riscv: mm: Introduce cntx2asid/cntx2version helper macros Samuel Holland
2023-09-09 20:16   ` Samuel Holland
2023-09-09 20:16 ` [PATCH 3/7] riscv: mm: Use a fixed layout for the MM context ID Samuel Holland
2023-09-09 20:16   ` Samuel Holland
2023-09-09 20:16 ` [PATCH 4/7] riscv: mm: Make asid_bits a local variable Samuel Holland
2023-09-09 20:16   ` Samuel Holland
2023-09-09 20:16 ` Samuel Holland [this message]
2023-09-09 20:16   ` [PATCH 5/7] riscv: mm: Preserve global TLB entries when switching contexts Samuel Holland
2023-09-09 20:16 ` [PATCH 6/7] riscv: mm: Always flush a single MM context by ASID Samuel Holland
2023-09-09 20:16   ` Samuel Holland
2023-09-10 19:46   ` Conor Dooley
2023-09-10 19:46     ` Conor Dooley
2023-10-26 15:53     ` Palmer Dabbelt
2023-10-26 15:53       ` Palmer Dabbelt
2023-09-09 20:16 ` [PATCH 7/7] riscv: mm: Combine the SMP and non-SMP TLB flushing code Samuel Holland
2023-09-09 20:16   ` Samuel Holland
2023-09-09 23:02   ` kernel test robot
2023-09-09 23:02     ` kernel test robot
2023-09-11 22:08   ` kernel test robot
2023-09-11 22:08     ` kernel test robot
2023-09-12  2:03   ` kernel test robot
2023-09-12  2:03     ` kernel test robot

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=20230909201727.10909-6-samuel@sholland.org \
    --to=samuel@sholland.org \
    --cc=alexghiti@rivosinc.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=palmer@dabbelt.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.