From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx0a-001b2d01.pphosted.com (mx0b-001b2d01.pphosted.com [148.163.158.5]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by lists.ozlabs.org (Postfix) with ESMTPS id 3thtnp05V6zDwKr for ; Mon, 19 Dec 2016 19:07:57 +1100 (AEDT) Received: from pps.filterd (m0098417.ppops.net [127.0.0.1]) by mx0a-001b2d01.pphosted.com (8.16.0.17/8.16.0.17) with SMTP id uBJ83elc120343 for ; Mon, 19 Dec 2016 03:07:56 -0500 Received: from e23smtp07.au.ibm.com (e23smtp07.au.ibm.com [202.81.31.140]) by mx0a-001b2d01.pphosted.com with ESMTP id 27eaaru6x8-1 (version=TLSv1.2 cipher=AES256-SHA bits=256 verify=NOT) for ; Mon, 19 Dec 2016 03:07:55 -0500 Received: from localhost by e23smtp07.au.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Mon, 19 Dec 2016 18:07:53 +1000 Received: from d23relay06.au.ibm.com (d23relay06.au.ibm.com [9.185.63.219]) by d23dlp01.au.ibm.com (Postfix) with ESMTP id D505B2CE8057 for ; Mon, 19 Dec 2016 19:07:49 +1100 (EST) Received: from d23av05.au.ibm.com (d23av05.au.ibm.com [9.190.234.119]) by d23relay06.au.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id uBJ87nca9568340 for ; Mon, 19 Dec 2016 19:07:49 +1100 Received: from d23av05.au.ibm.com (localhost [127.0.0.1]) by d23av05.au.ibm.com (8.14.4/8.14.4/NCO v10.0 AVout) with ESMTP id uBJ87n7g028442 for ; Mon, 19 Dec 2016 19:07:49 +1100 From: Madhavan Srinivasan To: benh@kernel.crashing.org, mpe@ellerman.id.au Cc: anton@samba.org, paulus@samba.org, npiggin@gmail.com, linuxppc-dev@lists.ozlabs.org, Madhavan Srinivasan Subject: [PATCH v4 11/12] powerpc: rewrite local_t using soft_irq Date: Mon, 19 Dec 2016 13:37:07 +0530 In-Reply-To: <1482134828-18811-1-git-send-email-maddy@linux.vnet.ibm.com> References: <1482134828-18811-1-git-send-email-maddy@linux.vnet.ibm.com> Message-Id: <1482134828-18811-12-git-send-email-maddy@linux.vnet.ibm.com> List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Local atomic operations are fast and highly reentrant per CPU counters. Used for percpu variable updates. Local atomic operations only guarantee variable modification atomicity wrt the CPU which owns the data and these needs to be executed in a preemption safe way. Here is the design of this patch. Since local_* operations are only need to be atomic to interrupts (IIUC), we have two options. Either replay the "op" if interrupted or replay the interrupt after the "op". Initial patchset posted was based on implementing local_* operation based on CR5 which replay's the "op". Patchset had issues in case of rewinding the address pointor from an array. This make the slow path really slow. Since CR5 based implementation proposed using __ex_table to find the rewind addressr, this rasied concerns about size of __ex_table and vmlinux. https://lists.ozlabs.org/pipermail/linuxppc-dev/2014-December/123115.html But this patch uses, powerpc_local_irq_pmu_save to soft_disable interrupts (including PMIs). After finishing the "op", powerpc_local_irq_pmu_restore() called and correspondingly interrupts are replayed if any occured. patch re-write the current local_* functions to use arch_local_irq_disbale. Base flow for each function is { powerpc_local_irq_pmu_save(flags) load .. store powerpc_local_irq_pmu_restore(flags) } Reason for the approach is that, currently l[w/d]arx/st[w/d]cx. instruction pair is used for local_* operations, which are heavy on cycle count and they dont support a local variant. So to see whether the new implementation helps, used a modified version of Rusty's benchmark code on local_t. https://lkml.org/lkml/2008/12/16/450 Modifications to Rusty's benchmark code: - Executed only local_t test Here are the values with the patch. Time in ns per iteration Local_t Without Patch With Patch _inc 28 8 _add 28 8 _read 3 3 _add_return 28 7 Currently only asm/local.h has been rewritten, and also the entire change is tested only in PPC64 (pseries guest) and PPC64 host (LE) Reviewed-by: Nicholas Piggin Signed-off-by: Madhavan Srinivasan --- arch/powerpc/include/asm/local.h | 201 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 201 insertions(+) diff --git a/arch/powerpc/include/asm/local.h b/arch/powerpc/include/asm/local.h index b8da91363864..7d117c07b0b1 100644 --- a/arch/powerpc/include/asm/local.h +++ b/arch/powerpc/include/asm/local.h @@ -3,6 +3,9 @@ #include #include +#include + +#include typedef struct { @@ -14,6 +17,202 @@ typedef struct #define local_read(l) atomic_long_read(&(l)->a) #define local_set(l,i) atomic_long_set(&(l)->a, (i)) +#ifdef CONFIG_PPC64 + +static __inline__ void local_add(long i, local_t *l) +{ + long t; + unsigned long flags; + + powerpc_local_irq_pmu_save(flags); + __asm__ __volatile__( + PPC_LL" %0,0(%2)\n\ + add %0,%1,%0\n" + PPC_STL" %0,0(%2)\n" + : "=&r" (t) + : "r" (i), "r" (&(l->a.counter))); + powerpc_local_irq_pmu_restore(flags); +} + +static __inline__ void local_sub(long i, local_t *l) +{ + long t; + unsigned long flags; + + powerpc_local_irq_pmu_save(flags); + __asm__ __volatile__( + PPC_LL" %0,0(%2)\n\ + subf %0,%1,%0\n" + PPC_STL" %0,0(%2)\n" + : "=&r" (t) + : "r" (i), "r" (&(l->a.counter))); + powerpc_local_irq_pmu_restore(flags); +} + +static __inline__ long local_add_return(long a, local_t *l) +{ + long t; + unsigned long flags; + + powerpc_local_irq_pmu_save(flags); + __asm__ __volatile__( + PPC_LL" %0,0(%2)\n\ + add %0,%1,%0\n" + PPC_STL "%0,0(%2)\n" + : "=&r" (t) + : "r" (a), "r" (&(l->a.counter)) + : "memory"); + powerpc_local_irq_pmu_restore(flags); + + return t; +} + +#define local_add_negative(a, l) (local_add_return((a), (l)) < 0) + +static __inline__ long local_sub_return(long a, local_t *l) +{ + long t; + unsigned long flags; + + powerpc_local_irq_pmu_save(flags); + __asm__ __volatile__( + PPC_LL" %0,0(%2)\n\ + subf %0,%1,%0\n" + PPC_STL "%0,0(%2)\n" + : "=&r" (t) + : "r" (a), "r" (&(l->a.counter)) + : "memory"); + powerpc_local_irq_pmu_restore(flags); + + return t; +} + +static __inline__ long local_inc_return(local_t *l) +{ + long t; + unsigned long flags; + + powerpc_local_irq_pmu_save(flags); + __asm__ __volatile__( + PPC_LL" %0,0(%1)\n\ + addic %0,%0,1\n" + PPC_STL "%0,0(%1)\n" + : "=&r" (t) + : "r" (&(l->a.counter)) + : "xer", "memory"); + powerpc_local_irq_pmu_restore(flags); + + return t; +} + +/* + * local_inc_and_test - increment and test + * @l: pointer of type local_t + * + * Atomically increments @l by 1 + * and returns true if the result is zero, or false for all + * other cases. + */ +#define local_inc_and_test(l) (local_inc_return(l) == 0) + +static __inline__ long local_dec_return(local_t *l) +{ + long t; + unsigned long flags; + + powerpc_local_irq_pmu_save(flags); + __asm__ __volatile__( + PPC_LL" %0,0(%1)\n\ + addic %0,%0,-1\n" + PPC_STL "%0,0(%1)\n" + : "=&r" (t) + : "r" (&(l->a.counter)) + : "xer", "memory"); + powerpc_local_irq_pmu_restore(flags); + + return t; +} + +#define local_inc(l) local_inc_return(l) +#define local_dec(l) local_dec_return(l) + +#define local_cmpxchg(l, o, n) \ + (cmpxchg_local(&((l)->a.counter), (o), (n))) +#define local_xchg(l, n) (xchg_local(&((l)->a.counter), (n))) + +/** + * local_add_unless - add unless the number is a given value + * @l: pointer of type local_t + * @a: the amount to add to v... + * @u: ...unless v is equal to u. + * + * Atomically adds @a to @l, so long as it was not @u. + * Returns non-zero if @l was not @u, and zero otherwise. + */ +static __inline__ int local_add_unless(local_t *l, long a, long u) +{ + long t; + unsigned long flags; + + powerpc_local_irq_pmu_save(flags); + __asm__ __volatile__ ( + PPC_LL" %0,0(%1)\n\ + cmpw 0,%0,%3 \n\ + beq- 2f \n\ + add %0,%2,%0 \n" + PPC_STL" %0,0(%1) \n" +" subf %0,%2,%0 \n\ +2:" + : "=&r" (t) + : "r" (&(l->a.counter)), "r" (a), "r" (u) + : "cc", "memory"); + powerpc_local_irq_pmu_restore(flags); + + return t != u; +} + +#define local_inc_not_zero(l) local_add_unless((l), 1, 0) + +#define local_sub_and_test(a, l) (local_sub_return((a), (l)) == 0) +#define local_dec_and_test(l) (local_dec_return((l)) == 0) + +/* + * Atomically test *l and decrement if it is greater than 0. + * The function returns the old value of *l minus 1. + */ +static __inline__ long local_dec_if_positive(local_t *l) +{ + long t; + unsigned long flags; + + powerpc_local_irq_pmu_save(flags); + __asm__ __volatile__( + PPC_LL" %0,0(%1)\n\ + cmpwi %0,1\n\ + addi %0,%0,-1\n\ + blt- 2f\n" + PPC_STL "%0,0(%1)\n" + "\n\ +2:" : "=&b" (t) + : "r" (&(l->a.counter)) + : "cc", "memory"); + powerpc_local_irq_pmu_restore(flags); + + return t; +} + +/* Use these for per-cpu local_t variables: on some archs they are + * much more efficient than these naive implementations. Note they take + * a variable, not an address. + */ + +#define __local_inc(l) ((l)->a.counter++) +#define __local_dec(l) ((l)->a.counter++) +#define __local_add(i,l) ((l)->a.counter+=(i)) +#define __local_sub(i,l) ((l)->a.counter-=(i)) + +#else + #define local_add(i,l) atomic_long_add((i),(&(l)->a)) #define local_sub(i,l) atomic_long_sub((i),(&(l)->a)) #define local_inc(l) atomic_long_inc(&(l)->a) @@ -172,4 +371,6 @@ static __inline__ long local_dec_if_positive(local_t *l) #define __local_add(i,l) ((l)->a.counter+=(i)) #define __local_sub(i,l) ((l)->a.counter-=(i)) +#endif /* CONFIG_PPC64 */ + #endif /* _ARCH_POWERPC_LOCAL_H */ -- 2.7.4