From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Lameter Date: Mon, 26 Aug 2013 21:56:18 +0000 Subject: [guv v2 27/31] sh: Replace __get_cpu_var uses throughout the kernel Message-Id: <00000140bca09c50-26aada97-96a9-423b-936b-15a296842a76-000000@email.amazonses.com> List-Id: References: <20130826204351.725357339@linux.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Tejun Heo Cc: akpm@linuxfoundation.org, Paul Mundt , linux-sh@vger.kernel.org, linux-arch@vger.kernel.org, Steven Rostedt , linux-kernel@vger.kernel.org __get_cpu_var() is used for multiple purposes in the kernel source. One of them is address calculation via the form &__get_cpu_var(x). This calculates the address for the instance of the percpu variable of the current processor based on an offset. Others usage cases are for storing and retrieving data from the current processors percpu area. __get_cpu_var() can be used as an lvalue when writing data or on the right side of an assignment. __get_cpu_var() is defined as : #define __get_cpu_var(var) (*this_cpu_ptr(&(var))) __get_cpu_var() always only does a address determination. However, store and retrieve operations could use a segment prefix (or global register on other platforms) to avoid the address calculation. this_cpu_write() and this_cpu_read() can directly take an offset into a percpu area and use optimized assembly code to read and write per cpu variables. This patch converts __get_cpu_var into either and explicit address calculation using this_cpu_ptr() or into a use of this_cpu operations that use the offset. Thereby address calcualtions are avoided and less registers are used when code is generated. At the end of the patchset all uses of __get_cpu_var have been removed so the macro is removed too. The patchset includes passes over all arches as well. Once these operations are used throughout then specialized macros can be defined in non -x86 arches as well in order to optimize per cpu access by f.e. using a global register that may be set to the per cpu base. Transformations done to __get_cpu_var() 1. Determine the address of the percpu instance of the current processor. DEFINE_PER_CPU(int, y); int *x = &__get_cpu_var(y); Converts to int *x = this_cpu_ptr(&y); 2. Same as #1 but this time an array structure is involved. DEFINE_PER_CPU(int, y[20]); int *x = __get_cpu_var(y); Converts to int *x = this_cpu_ptr(y); 3. Retrieve the content of the current processors instance of a per cpu variable. DEFINE_PER_CPU(int, u); int x = __get_cpu_var(y) Converts to int x = __this_cpu_read(y); 4. Retrieve the content of a percpu struct DEFINE_PER_CPU(struct mystruct, y); struct mystruct x = __get_cpu_var(y); Converts to memcpy(this_cpu_ptr(&y), x, sizeof(x)); 5. Assignment to a per cpu variable DEFINE_PER_CPU(int, y) __get_cpu_var(y) = x; Converts to this_cpu_write(y, x); 6. Increment/Decrement etc of a per cpu variable DEFINE_PER_CPU(int, y); __get_cpu_var(y)++ Converts to this_cpu_inc(y) Index: linux/arch/sh/kernel/hw_breakpoint.c =================================--- linux.orig/arch/sh/kernel/hw_breakpoint.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/hw_breakpoint.c 2013-08-26 13:28:32.265604741 -0500 @@ -52,7 +52,7 @@ int arch_install_hw_breakpoint(struct pe int i; for (i = 0; i < sh_ubc->num_events; i++) { - struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); + struct perf_event **slot = this_cpu_ptr(&bp_per_reg[i]); if (!*slot) { *slot = bp; @@ -84,7 +84,7 @@ void arch_uninstall_hw_breakpoint(struct int i; for (i = 0; i < sh_ubc->num_events; i++) { - struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); + struct perf_event **slot = this_cpu_ptr(&bp_per_reg[i]); if (*slot = bp) { *slot = NULL; Index: linux/arch/sh/kernel/kprobes.c =================================--- linux.orig/arch/sh/kernel/kprobes.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/kprobes.c 2013-08-26 13:28:32.265604741 -0500 @@ -102,7 +102,7 @@ int __kprobes kprobe_handle_illslot(unsi void __kprobes arch_remove_kprobe(struct kprobe *p) { - struct kprobe *saved = &__get_cpu_var(saved_next_opcode); + struct kprobe *saved = this_cpu_ptr(&saved_next_opcode); if (saved->addr) { arch_disarm_kprobe(p); @@ -111,7 +111,7 @@ void __kprobes arch_remove_kprobe(struct saved->addr = NULL; saved->opcode = 0; - saved = &__get_cpu_var(saved_next_opcode2); + saved = this_cpu_ptr(&saved_next_opcode2); if (saved->addr) { arch_disarm_kprobe(saved); @@ -129,14 +129,14 @@ static void __kprobes save_previous_kpro static void __kprobes restore_previous_kprobe(struct kprobe_ctlblk *kcb) { - __get_cpu_var(current_kprobe) = kcb->prev_kprobe.kp; + __this_cpu_write(current_kprobe, kcb->prev_kprobe.kp); kcb->kprobe_status = kcb->prev_kprobe.status; } static void __kprobes set_current_kprobe(struct kprobe *p, struct pt_regs *regs, struct kprobe_ctlblk *kcb) { - __get_cpu_var(current_kprobe) = p; + __this_cpu_write(current_kprobe, p); } /* @@ -146,15 +146,15 @@ static void __kprobes set_current_kprobe */ static void __kprobes prepare_singlestep(struct kprobe *p, struct pt_regs *regs) { - __get_cpu_var(saved_current_opcode).addr = (kprobe_opcode_t *)regs->pc; + __this_cpu_write(saved_current_opcode.addr, (kprobe_opcode_t *)regs->pc); if (p != NULL) { struct kprobe *op1, *op2; arch_disarm_kprobe(p); - op1 = &__get_cpu_var(saved_next_opcode); - op2 = &__get_cpu_var(saved_next_opcode2); + op1 = this_cpu_ptr(&saved_next_opcode); + op2 = this_cpu_ptr(&saved_next_opcode2); if (OPCODE_JSR(p->opcode) || OPCODE_JMP(p->opcode)) { unsigned int reg_nr = ((p->opcode >> 8) & 0x000F); @@ -249,7 +249,7 @@ static int __kprobes kprobe_handler(stru kcb->kprobe_status = KPROBE_REENTER; return 1; } else { - p = __get_cpu_var(current_kprobe); + p = __this_cpu_read(current_kprobe); if (p->break_handler && p->break_handler(p, regs)) { goto ss_probe; } @@ -336,9 +336,9 @@ int __kprobes trampoline_probe_handler(s continue; if (ri->rp && ri->rp->handler) { - __get_cpu_var(current_kprobe) = &ri->rp->kp; + __this_cpu_write(current_kprobe, &ri->rp->kp); ri->rp->handler(ri, regs); - __get_cpu_var(current_kprobe) = NULL; + __this_cpu_write(current_kprobe, NULL); } orig_ret_address = (unsigned long)ri->ret_addr; @@ -383,19 +383,19 @@ static int __kprobes post_kprobe_handler cur->post_handler(cur, regs, 0); } - p = &__get_cpu_var(saved_next_opcode); + p = this_cpu_ptr(&saved_next_opcode); if (p->addr) { arch_disarm_kprobe(p); p->addr = NULL; p->opcode = 0; - addr = __get_cpu_var(saved_current_opcode).addr; - __get_cpu_var(saved_current_opcode).addr = NULL; + addr = __this_cpu_read(saved_current_opcode).addr; + __this_cpu_write(saved_current_opcode.addr, NULL); p = get_kprobe(addr); arch_arm_kprobe(p); - p = &__get_cpu_var(saved_next_opcode2); + p = this_cpu_ptr(&saved_next_opcode2); if (p->addr) { arch_disarm_kprobe(p); p->addr = NULL; @@ -511,7 +511,7 @@ int __kprobes kprobe_exceptions_notify(s if (kprobe_handler(args->regs)) { ret = NOTIFY_STOP; } else { - p = __get_cpu_var(current_kprobe); + p = __this_cpu_read(current_kprobe); if (p->break_handler && p->break_handler(p, args->regs)) ret = NOTIFY_STOP; Index: linux/arch/sh/kernel/localtimer.c =================================--- linux.orig/arch/sh/kernel/localtimer.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/localtimer.c 2013-08-26 13:28:32.265604741 -0500 @@ -32,7 +32,7 @@ static DEFINE_PER_CPU(struct clock_event */ void local_timer_interrupt(void) { - struct clock_event_device *clk = &__get_cpu_var(local_clockevent); + struct clock_event_device *clk = this_cpu_ptr(&local_clockevent); irq_enter(); clk->event_handler(clk); Index: linux/arch/sh/kernel/perf_event.c =================================--- linux.orig/arch/sh/kernel/perf_event.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/perf_event.c 2013-08-26 13:28:32.265604741 -0500 @@ -227,7 +227,7 @@ again: static void sh_pmu_stop(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; @@ -245,7 +245,7 @@ static void sh_pmu_stop(struct perf_even static void sh_pmu_start(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; @@ -262,7 +262,7 @@ static void sh_pmu_start(struct perf_eve static void sh_pmu_del(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); sh_pmu_stop(event, PERF_EF_UPDATE); __clear_bit(event->hw.idx, cpuc->used_mask); @@ -272,7 +272,7 @@ static void sh_pmu_del(struct perf_event static int sh_pmu_add(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; int ret = -EAGAIN; Index: linux/arch/sh/kernel/smp.c =================================--- linux.orig/arch/sh/kernel/smp.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/smp.c 2013-08-26 13:28:32.265604741 -0500 @@ -111,7 +111,7 @@ void play_dead_common(void) irq_ctx_exit(raw_smp_processor_id()); mb(); - __get_cpu_var(cpu_state) = CPU_DEAD; + __this_cpu_write(cpu_state, CPU_DEAD); local_irq_disable(); } From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752638Ab3HZV4X (ORCPT ); Mon, 26 Aug 2013 17:56:23 -0400 Received: from a9-42.smtp-out.amazonses.com ([54.240.9.42]:54949 "EHLO a9-42.smtp-out.amazonses.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752316Ab3HZV4T (ORCPT ); Mon, 26 Aug 2013 17:56:19 -0400 Message-ID: <00000140bca09c50-26aada97-96a9-423b-936b-15a296842a76-000000@email.amazonses.com> User-Agent: quilt/0.50-1 Date: Mon, 26 Aug 2013 21:56:18 +0000 From: Christoph Lameter To: Tejun Heo Cc: akpm@linuxfoundation.org, Paul Mundt , linux-sh@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: Steven Rostedt Cc: linux-kernel@vger.kernel.org Subject: [guv v2 27/31] sh: Replace __get_cpu_var uses throughout the kernel References: <20130826204351.725357339@linux.com> X-SES-Outgoing: 2013.08.26-54.240.9.42 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org __get_cpu_var() is used for multiple purposes in the kernel source. One of them is address calculation via the form &__get_cpu_var(x). This calculates the address for the instance of the percpu variable of the current processor based on an offset. Others usage cases are for storing and retrieving data from the current processors percpu area. __get_cpu_var() can be used as an lvalue when writing data or on the right side of an assignment. __get_cpu_var() is defined as : #define __get_cpu_var(var) (*this_cpu_ptr(&(var))) __get_cpu_var() always only does a address determination. However, store and retrieve operations could use a segment prefix (or global register on other platforms) to avoid the address calculation. this_cpu_write() and this_cpu_read() can directly take an offset into a percpu area and use optimized assembly code to read and write per cpu variables. This patch converts __get_cpu_var into either and explicit address calculation using this_cpu_ptr() or into a use of this_cpu operations that use the offset. Thereby address calcualtions are avoided and less registers are used when code is generated. At the end of the patchset all uses of __get_cpu_var have been removed so the macro is removed too. The patchset includes passes over all arches as well. Once these operations are used throughout then specialized macros can be defined in non -x86 arches as well in order to optimize per cpu access by f.e. using a global register that may be set to the per cpu base. Transformations done to __get_cpu_var() 1. Determine the address of the percpu instance of the current processor. DEFINE_PER_CPU(int, y); int *x = &__get_cpu_var(y); Converts to int *x = this_cpu_ptr(&y); 2. Same as #1 but this time an array structure is involved. DEFINE_PER_CPU(int, y[20]); int *x = __get_cpu_var(y); Converts to int *x = this_cpu_ptr(y); 3. Retrieve the content of the current processors instance of a per cpu variable. DEFINE_PER_CPU(int, u); int x = __get_cpu_var(y) Converts to int x = __this_cpu_read(y); 4. Retrieve the content of a percpu struct DEFINE_PER_CPU(struct mystruct, y); struct mystruct x = __get_cpu_var(y); Converts to memcpy(this_cpu_ptr(&y), x, sizeof(x)); 5. Assignment to a per cpu variable DEFINE_PER_CPU(int, y) __get_cpu_var(y) = x; Converts to this_cpu_write(y, x); 6. Increment/Decrement etc of a per cpu variable DEFINE_PER_CPU(int, y); __get_cpu_var(y)++ Converts to this_cpu_inc(y) Index: linux/arch/sh/kernel/hw_breakpoint.c =================================================================== --- linux.orig/arch/sh/kernel/hw_breakpoint.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/hw_breakpoint.c 2013-08-26 13:28:32.265604741 -0500 @@ -52,7 +52,7 @@ int arch_install_hw_breakpoint(struct pe int i; for (i = 0; i < sh_ubc->num_events; i++) { - struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); + struct perf_event **slot = this_cpu_ptr(&bp_per_reg[i]); if (!*slot) { *slot = bp; @@ -84,7 +84,7 @@ void arch_uninstall_hw_breakpoint(struct int i; for (i = 0; i < sh_ubc->num_events; i++) { - struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); + struct perf_event **slot = this_cpu_ptr(&bp_per_reg[i]); if (*slot == bp) { *slot = NULL; Index: linux/arch/sh/kernel/kprobes.c =================================================================== --- linux.orig/arch/sh/kernel/kprobes.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/kprobes.c 2013-08-26 13:28:32.265604741 -0500 @@ -102,7 +102,7 @@ int __kprobes kprobe_handle_illslot(unsi void __kprobes arch_remove_kprobe(struct kprobe *p) { - struct kprobe *saved = &__get_cpu_var(saved_next_opcode); + struct kprobe *saved = this_cpu_ptr(&saved_next_opcode); if (saved->addr) { arch_disarm_kprobe(p); @@ -111,7 +111,7 @@ void __kprobes arch_remove_kprobe(struct saved->addr = NULL; saved->opcode = 0; - saved = &__get_cpu_var(saved_next_opcode2); + saved = this_cpu_ptr(&saved_next_opcode2); if (saved->addr) { arch_disarm_kprobe(saved); @@ -129,14 +129,14 @@ static void __kprobes save_previous_kpro static void __kprobes restore_previous_kprobe(struct kprobe_ctlblk *kcb) { - __get_cpu_var(current_kprobe) = kcb->prev_kprobe.kp; + __this_cpu_write(current_kprobe, kcb->prev_kprobe.kp); kcb->kprobe_status = kcb->prev_kprobe.status; } static void __kprobes set_current_kprobe(struct kprobe *p, struct pt_regs *regs, struct kprobe_ctlblk *kcb) { - __get_cpu_var(current_kprobe) = p; + __this_cpu_write(current_kprobe, p); } /* @@ -146,15 +146,15 @@ static void __kprobes set_current_kprobe */ static void __kprobes prepare_singlestep(struct kprobe *p, struct pt_regs *regs) { - __get_cpu_var(saved_current_opcode).addr = (kprobe_opcode_t *)regs->pc; + __this_cpu_write(saved_current_opcode.addr, (kprobe_opcode_t *)regs->pc); if (p != NULL) { struct kprobe *op1, *op2; arch_disarm_kprobe(p); - op1 = &__get_cpu_var(saved_next_opcode); - op2 = &__get_cpu_var(saved_next_opcode2); + op1 = this_cpu_ptr(&saved_next_opcode); + op2 = this_cpu_ptr(&saved_next_opcode2); if (OPCODE_JSR(p->opcode) || OPCODE_JMP(p->opcode)) { unsigned int reg_nr = ((p->opcode >> 8) & 0x000F); @@ -249,7 +249,7 @@ static int __kprobes kprobe_handler(stru kcb->kprobe_status = KPROBE_REENTER; return 1; } else { - p = __get_cpu_var(current_kprobe); + p = __this_cpu_read(current_kprobe); if (p->break_handler && p->break_handler(p, regs)) { goto ss_probe; } @@ -336,9 +336,9 @@ int __kprobes trampoline_probe_handler(s continue; if (ri->rp && ri->rp->handler) { - __get_cpu_var(current_kprobe) = &ri->rp->kp; + __this_cpu_write(current_kprobe, &ri->rp->kp); ri->rp->handler(ri, regs); - __get_cpu_var(current_kprobe) = NULL; + __this_cpu_write(current_kprobe, NULL); } orig_ret_address = (unsigned long)ri->ret_addr; @@ -383,19 +383,19 @@ static int __kprobes post_kprobe_handler cur->post_handler(cur, regs, 0); } - p = &__get_cpu_var(saved_next_opcode); + p = this_cpu_ptr(&saved_next_opcode); if (p->addr) { arch_disarm_kprobe(p); p->addr = NULL; p->opcode = 0; - addr = __get_cpu_var(saved_current_opcode).addr; - __get_cpu_var(saved_current_opcode).addr = NULL; + addr = __this_cpu_read(saved_current_opcode).addr; + __this_cpu_write(saved_current_opcode.addr, NULL); p = get_kprobe(addr); arch_arm_kprobe(p); - p = &__get_cpu_var(saved_next_opcode2); + p = this_cpu_ptr(&saved_next_opcode2); if (p->addr) { arch_disarm_kprobe(p); p->addr = NULL; @@ -511,7 +511,7 @@ int __kprobes kprobe_exceptions_notify(s if (kprobe_handler(args->regs)) { ret = NOTIFY_STOP; } else { - p = __get_cpu_var(current_kprobe); + p = __this_cpu_read(current_kprobe); if (p->break_handler && p->break_handler(p, args->regs)) ret = NOTIFY_STOP; Index: linux/arch/sh/kernel/localtimer.c =================================================================== --- linux.orig/arch/sh/kernel/localtimer.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/localtimer.c 2013-08-26 13:28:32.265604741 -0500 @@ -32,7 +32,7 @@ static DEFINE_PER_CPU(struct clock_event */ void local_timer_interrupt(void) { - struct clock_event_device *clk = &__get_cpu_var(local_clockevent); + struct clock_event_device *clk = this_cpu_ptr(&local_clockevent); irq_enter(); clk->event_handler(clk); Index: linux/arch/sh/kernel/perf_event.c =================================================================== --- linux.orig/arch/sh/kernel/perf_event.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/perf_event.c 2013-08-26 13:28:32.265604741 -0500 @@ -227,7 +227,7 @@ again: static void sh_pmu_stop(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; @@ -245,7 +245,7 @@ static void sh_pmu_stop(struct perf_even static void sh_pmu_start(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; @@ -262,7 +262,7 @@ static void sh_pmu_start(struct perf_eve static void sh_pmu_del(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); sh_pmu_stop(event, PERF_EF_UPDATE); __clear_bit(event->hw.idx, cpuc->used_mask); @@ -272,7 +272,7 @@ static void sh_pmu_del(struct perf_event static int sh_pmu_add(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; int ret = -EAGAIN; Index: linux/arch/sh/kernel/smp.c =================================================================== --- linux.orig/arch/sh/kernel/smp.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/smp.c 2013-08-26 13:28:32.265604741 -0500 @@ -111,7 +111,7 @@ void play_dead_common(void) irq_ctx_exit(raw_smp_processor_id()); mb(); - __get_cpu_var(cpu_state) = CPU_DEAD; + __this_cpu_write(cpu_state, CPU_DEAD); local_irq_disable(); } From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Lameter Subject: [guv v2 27/31] sh: Replace __get_cpu_var uses throughout the kernel Date: Mon, 26 Aug 2013 21:56:18 +0000 Message-ID: <00000140bca09c50-26aada97-96a9-423b-936b-15a296842a76-000000@email.amazonses.com> References: <20130826204351.725357339@linux.com> Return-path: Received: from a9-42.smtp-out.amazonses.com ([54.240.9.42]:54949 "EHLO a9-42.smtp-out.amazonses.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752316Ab3HZV4T (ORCPT ); Mon, 26 Aug 2013 17:56:19 -0400 Sender: linux-arch-owner@vger.kernel.org List-ID: To: Tejun Heo Cc: akpm@linuxfoundation.org, Paul Mundt , linux-sh@vger.kernel.org, linux-arch@vger.kernel.org, Steven Rostedt , linux-kernel@vger.kernel.org __get_cpu_var() is used for multiple purposes in the kernel source. One of them is address calculation via the form &__get_cpu_var(x). This calculates the address for the instance of the percpu variable of the current processor based on an offset. Others usage cases are for storing and retrieving data from the current processors percpu area. __get_cpu_var() can be used as an lvalue when writing data or on the right side of an assignment. __get_cpu_var() is defined as : #define __get_cpu_var(var) (*this_cpu_ptr(&(var))) __get_cpu_var() always only does a address determination. However, store and retrieve operations could use a segment prefix (or global register on other platforms) to avoid the address calculation. this_cpu_write() and this_cpu_read() can directly take an offset into a percpu area and use optimized assembly code to read and write per cpu variables. This patch converts __get_cpu_var into either and explicit address calculation using this_cpu_ptr() or into a use of this_cpu operations that use the offset. Thereby address calcualtions are avoided and less registers are used when code is generated. At the end of the patchset all uses of __get_cpu_var have been removed so the macro is removed too. The patchset includes passes over all arches as well. Once these operations are used throughout then specialized macros can be defined in non -x86 arches as well in order to optimize per cpu access by f.e. using a global register that may be set to the per cpu base. Transformations done to __get_cpu_var() 1. Determine the address of the percpu instance of the current processor. DEFINE_PER_CPU(int, y); int *x = &__get_cpu_var(y); Converts to int *x = this_cpu_ptr(&y); 2. Same as #1 but this time an array structure is involved. DEFINE_PER_CPU(int, y[20]); int *x = __get_cpu_var(y); Converts to int *x = this_cpu_ptr(y); 3. Retrieve the content of the current processors instance of a per cpu variable. DEFINE_PER_CPU(int, u); int x = __get_cpu_var(y) Converts to int x = __this_cpu_read(y); 4. Retrieve the content of a percpu struct DEFINE_PER_CPU(struct mystruct, y); struct mystruct x = __get_cpu_var(y); Converts to memcpy(this_cpu_ptr(&y), x, sizeof(x)); 5. Assignment to a per cpu variable DEFINE_PER_CPU(int, y) __get_cpu_var(y) = x; Converts to this_cpu_write(y, x); 6. Increment/Decrement etc of a per cpu variable DEFINE_PER_CPU(int, y); __get_cpu_var(y)++ Converts to this_cpu_inc(y) Index: linux/arch/sh/kernel/hw_breakpoint.c =================================================================== --- linux.orig/arch/sh/kernel/hw_breakpoint.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/hw_breakpoint.c 2013-08-26 13:28:32.265604741 -0500 @@ -52,7 +52,7 @@ int arch_install_hw_breakpoint(struct pe int i; for (i = 0; i < sh_ubc->num_events; i++) { - struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); + struct perf_event **slot = this_cpu_ptr(&bp_per_reg[i]); if (!*slot) { *slot = bp; @@ -84,7 +84,7 @@ void arch_uninstall_hw_breakpoint(struct int i; for (i = 0; i < sh_ubc->num_events; i++) { - struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); + struct perf_event **slot = this_cpu_ptr(&bp_per_reg[i]); if (*slot == bp) { *slot = NULL; Index: linux/arch/sh/kernel/kprobes.c =================================================================== --- linux.orig/arch/sh/kernel/kprobes.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/kprobes.c 2013-08-26 13:28:32.265604741 -0500 @@ -102,7 +102,7 @@ int __kprobes kprobe_handle_illslot(unsi void __kprobes arch_remove_kprobe(struct kprobe *p) { - struct kprobe *saved = &__get_cpu_var(saved_next_opcode); + struct kprobe *saved = this_cpu_ptr(&saved_next_opcode); if (saved->addr) { arch_disarm_kprobe(p); @@ -111,7 +111,7 @@ void __kprobes arch_remove_kprobe(struct saved->addr = NULL; saved->opcode = 0; - saved = &__get_cpu_var(saved_next_opcode2); + saved = this_cpu_ptr(&saved_next_opcode2); if (saved->addr) { arch_disarm_kprobe(saved); @@ -129,14 +129,14 @@ static void __kprobes save_previous_kpro static void __kprobes restore_previous_kprobe(struct kprobe_ctlblk *kcb) { - __get_cpu_var(current_kprobe) = kcb->prev_kprobe.kp; + __this_cpu_write(current_kprobe, kcb->prev_kprobe.kp); kcb->kprobe_status = kcb->prev_kprobe.status; } static void __kprobes set_current_kprobe(struct kprobe *p, struct pt_regs *regs, struct kprobe_ctlblk *kcb) { - __get_cpu_var(current_kprobe) = p; + __this_cpu_write(current_kprobe, p); } /* @@ -146,15 +146,15 @@ static void __kprobes set_current_kprobe */ static void __kprobes prepare_singlestep(struct kprobe *p, struct pt_regs *regs) { - __get_cpu_var(saved_current_opcode).addr = (kprobe_opcode_t *)regs->pc; + __this_cpu_write(saved_current_opcode.addr, (kprobe_opcode_t *)regs->pc); if (p != NULL) { struct kprobe *op1, *op2; arch_disarm_kprobe(p); - op1 = &__get_cpu_var(saved_next_opcode); - op2 = &__get_cpu_var(saved_next_opcode2); + op1 = this_cpu_ptr(&saved_next_opcode); + op2 = this_cpu_ptr(&saved_next_opcode2); if (OPCODE_JSR(p->opcode) || OPCODE_JMP(p->opcode)) { unsigned int reg_nr = ((p->opcode >> 8) & 0x000F); @@ -249,7 +249,7 @@ static int __kprobes kprobe_handler(stru kcb->kprobe_status = KPROBE_REENTER; return 1; } else { - p = __get_cpu_var(current_kprobe); + p = __this_cpu_read(current_kprobe); if (p->break_handler && p->break_handler(p, regs)) { goto ss_probe; } @@ -336,9 +336,9 @@ int __kprobes trampoline_probe_handler(s continue; if (ri->rp && ri->rp->handler) { - __get_cpu_var(current_kprobe) = &ri->rp->kp; + __this_cpu_write(current_kprobe, &ri->rp->kp); ri->rp->handler(ri, regs); - __get_cpu_var(current_kprobe) = NULL; + __this_cpu_write(current_kprobe, NULL); } orig_ret_address = (unsigned long)ri->ret_addr; @@ -383,19 +383,19 @@ static int __kprobes post_kprobe_handler cur->post_handler(cur, regs, 0); } - p = &__get_cpu_var(saved_next_opcode); + p = this_cpu_ptr(&saved_next_opcode); if (p->addr) { arch_disarm_kprobe(p); p->addr = NULL; p->opcode = 0; - addr = __get_cpu_var(saved_current_opcode).addr; - __get_cpu_var(saved_current_opcode).addr = NULL; + addr = __this_cpu_read(saved_current_opcode).addr; + __this_cpu_write(saved_current_opcode.addr, NULL); p = get_kprobe(addr); arch_arm_kprobe(p); - p = &__get_cpu_var(saved_next_opcode2); + p = this_cpu_ptr(&saved_next_opcode2); if (p->addr) { arch_disarm_kprobe(p); p->addr = NULL; @@ -511,7 +511,7 @@ int __kprobes kprobe_exceptions_notify(s if (kprobe_handler(args->regs)) { ret = NOTIFY_STOP; } else { - p = __get_cpu_var(current_kprobe); + p = __this_cpu_read(current_kprobe); if (p->break_handler && p->break_handler(p, args->regs)) ret = NOTIFY_STOP; Index: linux/arch/sh/kernel/localtimer.c =================================================================== --- linux.orig/arch/sh/kernel/localtimer.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/localtimer.c 2013-08-26 13:28:32.265604741 -0500 @@ -32,7 +32,7 @@ static DEFINE_PER_CPU(struct clock_event */ void local_timer_interrupt(void) { - struct clock_event_device *clk = &__get_cpu_var(local_clockevent); + struct clock_event_device *clk = this_cpu_ptr(&local_clockevent); irq_enter(); clk->event_handler(clk); Index: linux/arch/sh/kernel/perf_event.c =================================================================== --- linux.orig/arch/sh/kernel/perf_event.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/perf_event.c 2013-08-26 13:28:32.265604741 -0500 @@ -227,7 +227,7 @@ again: static void sh_pmu_stop(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; @@ -245,7 +245,7 @@ static void sh_pmu_stop(struct perf_even static void sh_pmu_start(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; @@ -262,7 +262,7 @@ static void sh_pmu_start(struct perf_eve static void sh_pmu_del(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); sh_pmu_stop(event, PERF_EF_UPDATE); __clear_bit(event->hw.idx, cpuc->used_mask); @@ -272,7 +272,7 @@ static void sh_pmu_del(struct perf_event static int sh_pmu_add(struct perf_event *event, int flags) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; int ret = -EAGAIN; Index: linux/arch/sh/kernel/smp.c =================================================================== --- linux.orig/arch/sh/kernel/smp.c 2013-08-26 13:25:18.000000000 -0500 +++ linux/arch/sh/kernel/smp.c 2013-08-26 13:28:32.265604741 -0500 @@ -111,7 +111,7 @@ void play_dead_common(void) irq_ctx_exit(raw_smp_processor_id()); mb(); - __get_cpu_var(cpu_state) = CPU_DEAD; + __this_cpu_write(cpu_state, CPU_DEAD); local_irq_disable(); }