linux-doc.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: guoren@kernel.org
To: paul.walmsley@sifive.com, anup@brainfault.org,
	peterz@infradead.org, mingo@redhat.com, will@kernel.org,
	palmer@rivosinc.com, longman@redhat.com, boqun.feng@gmail.com,
	tglx@linutronix.de, paulmck@kernel.org, rostedt@goodmis.org,
	rdunlap@infradead.org, catalin.marinas@arm.com,
	conor.dooley@microchip.com, xiaoguang.xing@sophgo.com,
	bjorn@rivosinc.com, alexghiti@rivosinc.com,
	keescook@chromium.org, greentime.hu@sifive.com,
	ajones@ventanamicro.com, jszhang@kernel.org, wefu@redhat.com,
	wuwei2016@iscas.ac.cn
Cc: linux-arch@vger.kernel.org, linux-riscv@lists.infradead.org,
	linux-doc@vger.kernel.org, kvm@vger.kernel.org,
	virtualization@lists.linux-foundation.org,
	linux-csky@vger.kernel.org, Guo Ren <guoren@linux.alibaba.com>,
	Guo Ren <guoren@kernel.org>
Subject: [PATCH V10 05/19] riscv: qspinlock: Introduce combo spinlock
Date: Wed,  2 Aug 2023 12:46:47 -0400	[thread overview]
Message-ID: <20230802164701.192791-6-guoren@kernel.org> (raw)
In-Reply-To: <20230802164701.192791-1-guoren@kernel.org>

From: Guo Ren <guoren@linux.alibaba.com>

Combo spinlock could support queued and ticket in one Linux Image and
select them during boot time via errata mechanism. Here is the func
size (Bytes) comparison table below:

TYPE			: COMBO | TICKET | QUEUED
arch_spin_lock		: 106	| 60     | 50
arch_spin_unlock	: 54    | 36     | 26
arch_spin_trylock	: 110   | 72     | 54
arch_spin_is_locked	: 48    | 34     | 20
arch_spin_is_contended	: 56    | 40     | 24
rch_spin_value_unlocked	: 48    | 34     | 24

One example of disassemble combo arch_spin_unlock:
   0xffffffff8000409c <+14>:    nop                # detour slot
   0xffffffff800040a0 <+18>:    fence   rw,w       # queued spinlock start
   0xffffffff800040a4 <+22>:    sb      zero,0(a4) # queued spinlock end
   0xffffffff800040a8 <+26>:    ld      s0,8(sp)
   0xffffffff800040aa <+28>:    addi    sp,sp,16
   0xffffffff800040ac <+30>:    ret
   0xffffffff800040ae <+32>:    lw      a5,0(a4)   # ticket spinlock start
   0xffffffff800040b0 <+34>:    sext.w  a5,a5
   0xffffffff800040b2 <+36>:    fence   rw,w
   0xffffffff800040b6 <+40>:    addiw   a5,a5,1
   0xffffffff800040b8 <+42>:    slli    a5,a5,0x30
   0xffffffff800040ba <+44>:    srli    a5,a5,0x30
   0xffffffff800040bc <+46>:    sh      a5,0(a4)   # ticket spinlock end
   0xffffffff800040c0 <+50>:    ld      s0,8(sp)
   0xffffffff800040c2 <+52>:    addi    sp,sp,16
   0xffffffff800040c4 <+54>:    ret

The qspinlock is smaller and faster than ticket-lock when all are in
fast-path, and combo spinlock could provide a compatible Linux Image
for different micro-arch design (weak/strict fwd guarantee) processors.

Signed-off-by: Guo Ren <guoren@kernel.org>
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
---
 arch/riscv/Kconfig                |  9 +++-
 arch/riscv/include/asm/hwcap.h    |  1 +
 arch/riscv/include/asm/spinlock.h | 87 ++++++++++++++++++++++++++++++-
 arch/riscv/kernel/cpufeature.c    | 10 ++++
 4 files changed, 104 insertions(+), 3 deletions(-)

diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
index e89a3bea3dc1..119e774a3dcf 100644
--- a/arch/riscv/Kconfig
+++ b/arch/riscv/Kconfig
@@ -440,7 +440,7 @@ config NODES_SHIFT
 
 choice
 	prompt "RISC-V spinlock type"
-	default RISCV_TICKET_SPINLOCKS
+	default RISCV_COMBO_SPINLOCKS
 
 config RISCV_TICKET_SPINLOCKS
 	bool "Using ticket spinlock"
@@ -452,6 +452,13 @@ config RISCV_QUEUED_SPINLOCKS
 	help
 	  Make sure your micro arch LL/SC has a strong forward progress guarantee.
 	  Otherwise, stay at ticket-lock.
+
+config RISCV_COMBO_SPINLOCKS
+	bool "Using combo spinlock"
+	depends on SMP && MMU
+	select ARCH_USE_QUEUED_SPINLOCKS
+	help
+	  Select queued spinlock or ticket-lock via errata.
 endchoice
 
 config RISCV_ALTERNATIVE
diff --git a/arch/riscv/include/asm/hwcap.h b/arch/riscv/include/asm/hwcap.h
index f041bfa7f6a0..08ae75a694c2 100644
--- a/arch/riscv/include/asm/hwcap.h
+++ b/arch/riscv/include/asm/hwcap.h
@@ -54,6 +54,7 @@
 #define RISCV_ISA_EXT_ZIFENCEI		41
 #define RISCV_ISA_EXT_ZIHPM		42
 
+#define RISCV_ISA_EXT_XTICKETLOCK	63
 #define RISCV_ISA_EXT_MAX		64
 #define RISCV_ISA_EXT_NAME_LEN_MAX	32
 
diff --git a/arch/riscv/include/asm/spinlock.h b/arch/riscv/include/asm/spinlock.h
index c644a92d4548..9eb3ad31e564 100644
--- a/arch/riscv/include/asm/spinlock.h
+++ b/arch/riscv/include/asm/spinlock.h
@@ -7,11 +7,94 @@
 #define _Q_PENDING_LOOPS	(1 << 9)
 #endif
 
+#ifdef CONFIG_RISCV_COMBO_SPINLOCKS
+#include <asm-generic/ticket_spinlock.h>
+
+#undef arch_spin_is_locked
+#undef arch_spin_is_contended
+#undef arch_spin_value_unlocked
+#undef arch_spin_lock
+#undef arch_spin_trylock
+#undef arch_spin_unlock
+
+#include <asm-generic/qspinlock.h>
+#include <asm/hwcap.h>
+
+#undef arch_spin_is_locked
+#undef arch_spin_is_contended
+#undef arch_spin_value_unlocked
+#undef arch_spin_lock
+#undef arch_spin_trylock
+#undef arch_spin_unlock
+
+#define COMBO_DETOUR				\
+	asm_volatile_goto(ALTERNATIVE(		\
+		"nop",				\
+		"j %l[ticket_spin_lock]",	\
+		0,				\
+		RISCV_ISA_EXT_XTICKETLOCK,	\
+		CONFIG_RISCV_COMBO_SPINLOCKS)	\
+		: : : : ticket_spin_lock);
+
+static __always_inline void arch_spin_lock(arch_spinlock_t *lock)
+{
+	COMBO_DETOUR
+	queued_spin_lock(lock);
+	return;
+ticket_spin_lock:
+	ticket_spin_lock(lock);
+}
+
+static __always_inline bool arch_spin_trylock(arch_spinlock_t *lock)
+{
+	COMBO_DETOUR
+	return queued_spin_trylock(lock);
+ticket_spin_lock:
+	return ticket_spin_trylock(lock);
+}
+
+static __always_inline void arch_spin_unlock(arch_spinlock_t *lock)
+{
+	COMBO_DETOUR
+	queued_spin_unlock(lock);
+	return;
+ticket_spin_lock:
+	ticket_spin_unlock(lock);
+}
+
+static __always_inline int arch_spin_value_unlocked(arch_spinlock_t lock)
+{
+	COMBO_DETOUR
+	return queued_spin_value_unlocked(lock);
+ticket_spin_lock:
+	return ticket_spin_value_unlocked(lock);
+}
+
+static __always_inline int arch_spin_is_locked(arch_spinlock_t *lock)
+{
+	COMBO_DETOUR
+	return queued_spin_is_locked(lock);
+ticket_spin_lock:
+	return ticket_spin_is_locked(lock);
+}
+
+static __always_inline int arch_spin_is_contended(arch_spinlock_t *lock)
+{
+	COMBO_DETOUR
+	return queued_spin_is_contended(lock);
+ticket_spin_lock:
+	return ticket_spin_is_contended(lock);
+}
+#else /* CONFIG_RISCV_COMBO_SPINLOCKS */
+
 #ifdef CONFIG_QUEUED_SPINLOCKS
 #include <asm/qspinlock.h>
-#include <asm/qrwlock.h>
 #else
-#include <asm-generic/spinlock.h>
+#include <asm-generic/ticket_spinlock.h>
 #endif
 
+#endif /* CONFIG_RISCV_COMBO_SPINLOCKS */
+
+#include <asm/qrwlock.h>
+
 #endif /* __ASM_RISCV_SPINLOCK_H */
diff --git a/arch/riscv/kernel/cpufeature.c b/arch/riscv/kernel/cpufeature.c
index bdcf460ea53d..e65b0e54152d 100644
--- a/arch/riscv/kernel/cpufeature.c
+++ b/arch/riscv/kernel/cpufeature.c
@@ -324,6 +324,16 @@ void __init riscv_fill_hwcap(void)
 		set_bit(RISCV_ISA_EXT_ZICSR, isainfo->isa);
 		set_bit(RISCV_ISA_EXT_ZIFENCEI, isainfo->isa);
 
+#ifdef CONFIG_RISCV_COMBO_SPINLOCKS
+		/*
+		 * The RISC-V Linux used queued spinlock at first; then, we used ticket_lock
+		 * as default or queued spinlock by choice. Because ticket_lock would dirty
+		 * spinlock value, the only way is to change from queued_spinlock to
+		 * ticket_spinlock, but can not be vice.
+		 */
+		set_bit(RISCV_ISA_EXT_XTICKETLOCK, isainfo->isa);
+#endif
+
 		/*
 		 * These ones were as they were part of the base ISA when the
 		 * port & dt-bindings were upstreamed, and so can be set
-- 
2.36.1


  parent reply	other threads:[~2023-08-02 16:49 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-02 16:46 [PATCH V10 00/19] riscv: Add Native/Paravirt/CNA qspinlock support guoren
2023-08-02 16:46 ` [PATCH V10 01/19] asm-generic: ticket-lock: Reuse arch_spinlock_t of qspinlock guoren
2023-08-02 16:46 ` [PATCH V10 02/19] asm-generic: ticket-lock: Move into ticket_spinlock.h guoren
2023-08-02 16:46 ` [PATCH V10 03/19] riscv: qspinlock: errata: Add ERRATA_THEAD_WRITE_ONCE fixup guoren
2023-08-02 16:46 ` [PATCH V10 04/19] riscv: qspinlock: Add basic queued_spinlock support guoren
2023-08-11 19:34   ` Waiman Long
2023-08-12  0:18     ` Guo Ren
2023-08-02 16:46 ` guoren [this message]
2023-08-11 19:51   ` [PATCH V10 05/19] riscv: qspinlock: Introduce combo spinlock Waiman Long
2023-08-12  0:22     ` Guo Ren
2023-08-02 16:46 ` [PATCH V10 06/19] riscv: qspinlock: Allow force qspinlock from the command line guoren
2023-08-02 16:46 ` [PATCH V10 07/19] riscv: qspinlock: errata: Introduce ERRATA_THEAD_QSPINLOCK guoren
2023-08-04  9:05   ` Conor Dooley
2023-08-04  9:53     ` Guo Ren
2023-08-04 10:06       ` Conor Dooley
2023-08-05  1:28         ` Guo Ren
2023-08-07  5:23   ` Stefan O'Rear
2023-08-08  2:12     ` Guo Ren
2023-09-13 18:54     ` Palmer Dabbelt
2023-09-13 19:32       ` Waiman Long
2023-09-14  3:31       ` Guo Ren
2023-08-02 16:46 ` [PATCH V10 08/19] riscv: qspinlock: Use new static key for controlling call of virt_spin_lock() guoren
2023-08-02 16:46 ` [PATCH V10 09/19] RISC-V: paravirt: pvqspinlock: Add paravirt qspinlock skeleton guoren
2023-08-02 16:46 ` [PATCH V10 10/19] RISC-V: paravirt: pvqspinlock: KVM: " guoren
2023-08-02 16:46 ` [PATCH V10 11/19] RISC-V: paravirt: pvqspinlock: KVM: Implement kvm_sbi_ext_pvlock_kick_cpu() guoren
2023-08-02 16:46 ` [PATCH V10 12/19] RISC-V: paravirt: pvqspinlock: Add nopvspin kernel parameter guoren
2023-08-02 16:46 ` [PATCH V10 13/19] RISC-V: paravirt: pvqspinlock: Remove unnecessary definitions of cmpxchg & xchg guoren
2023-08-02 16:46 ` [PATCH V10 14/19] RISC-V: paravirt: pvqspinlock: Add xchg8 & cmpxchg_small support guoren
2023-08-02 16:46 ` [PATCH V10 15/19] RISC-V: paravirt: pvqspinlock: Add SBI implementation guoren
2023-08-02 16:46 ` [PATCH V10 16/19] RISC-V: paravirt: pvqspinlock: Add kconfig entry guoren
2023-08-02 16:46 ` [PATCH V10 17/19] RISC-V: paravirt: pvqspinlock: Add trace point for pv_kick/wait guoren
2023-08-02 16:47 ` [PATCH V10 18/19] locking/qspinlock: Move pv_ops into x86 directory guoren
2023-08-11 20:42   ` Waiman Long
2023-08-12  0:24     ` Guo Ren
2023-08-12  0:47       ` Waiman Long
2023-08-02 16:47 ` [PATCH V10 19/19] locking/qspinlock: riscv: Add Compact NUMA-aware lock support guoren

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=20230802164701.192791-6-guoren@kernel.org \
    --to=guoren@kernel.org \
    --cc=ajones@ventanamicro.com \
    --cc=alexghiti@rivosinc.com \
    --cc=anup@brainfault.org \
    --cc=bjorn@rivosinc.com \
    --cc=boqun.feng@gmail.com \
    --cc=catalin.marinas@arm.com \
    --cc=conor.dooley@microchip.com \
    --cc=greentime.hu@sifive.com \
    --cc=guoren@linux.alibaba.com \
    --cc=jszhang@kernel.org \
    --cc=keescook@chromium.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-arch@vger.kernel.org \
    --cc=linux-csky@vger.kernel.org \
    --cc=linux-doc@vger.kernel.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=longman@redhat.com \
    --cc=mingo@redhat.com \
    --cc=palmer@rivosinc.com \
    --cc=paul.walmsley@sifive.com \
    --cc=paulmck@kernel.org \
    --cc=peterz@infradead.org \
    --cc=rdunlap@infradead.org \
    --cc=rostedt@goodmis.org \
    --cc=tglx@linutronix.de \
    --cc=virtualization@lists.linux-foundation.org \
    --cc=wefu@redhat.com \
    --cc=will@kernel.org \
    --cc=wuwei2016@iscas.ac.cn \
    --cc=xiaoguang.xing@sophgo.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).