* [PATCH] gts and bfs fixes (v4). @ 2012-04-18 17:18 Konrad Rzeszutek Wilk 2012-04-18 17:18 ` [PATCH 1/2] ACPI: Convert wake_sleep_flags to a value instead of function (v4) Konrad Rzeszutek Wilk 2012-04-18 17:18 ` [PATCH 2/2] x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler Konrad Rzeszutek Wilk 0 siblings, 2 replies; 9+ messages in thread From: Konrad Rzeszutek Wilk @ 2012-04-18 17:18 UTC (permalink / raw) To: ming.m.lin, linux-kernel, linux-pm, lenb, linux-acpi, hpa, x86, rjw v2 [https://lkml.org/lkml/2012/4/9/474] - ditched the assembler version and used a C wrapper With git commit a2ef5c4fd44ce3922435139393b89f2cce47f576 "ACPI: Move module parameter gts and bfs to sleep.c" the sampling of gts/bfs module parameters flag by acpi_enter_sleep_state has been moved. It has been moved out of the inner ACPI code and out in the more x86 specific. This means that any caller of the exported acpi_enter_sleep_state function needs to provide two parameters now: sleep state and the bit flag whether to us _GTS|_BFS. All of the callers of acpi_enter_sleep_state did that, with the exception of the assembler version and as well any third-party callers of acpi_enter_sleep_state. These two patches expose 'wake_sleep_flags' as the product of the gts/bfs module parameter invocation. And also use said flag in the assembler 'do_suspend_lowlevel()' function. Konrad Rzeszutek Wilk (2): ACPI: Convert wake_sleep_flags to a value instead of function (v4) x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler. arch/x86/kernel/acpi/sleep.c | 4 +++ arch/x86/kernel/acpi/sleep.h | 4 +++ arch/x86/kernel/acpi/wakeup_32.S | 4 +-- arch/x86/kernel/acpi/wakeup_64.S | 4 +-- drivers/acpi/sleep.c | 52 ++++++++++++++++++++----------------- 5 files changed, 38 insertions(+), 30 deletions(-) ^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH 1/2] ACPI: Convert wake_sleep_flags to a value instead of function (v4) 2012-04-18 17:18 [PATCH] gts and bfs fixes (v4) Konrad Rzeszutek Wilk @ 2012-04-18 17:18 ` Konrad Rzeszutek Wilk 2012-04-18 19:46 ` Rafael J. Wysocki 2012-04-18 17:18 ` [PATCH 2/2] x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler Konrad Rzeszutek Wilk 1 sibling, 1 reply; 9+ messages in thread From: Konrad Rzeszutek Wilk @ 2012-04-18 17:18 UTC (permalink / raw) To: ming.m.lin, linux-kernel, linux-pm, lenb, linux-acpi, hpa, x86, rjw Cc: Konrad Rzeszutek Wilk With commit a2ef5c4fd44ce3922435139393b89f2cce47f576 "ACPI: Move module parameter gts and bfs to sleep.c" the wake_sleep_flags is required when calling acpi_enter_sleep_state, which means that if there are functions outside the sleep.c code they can't get get the wake_sleep_flags values. This converts the function in to a exported value and converts the module config operands to a function. Acked-by: Lin Ming <ming.m.lin@intel.com> [v2: Parameters can be turned on/off dynamically] [v3: unsigned char -> u8] [v4: val -> kp->arg] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> --- arch/x86/kernel/acpi/sleep.h | 2 + drivers/acpi/sleep.c | 52 ++++++++++++++++++++++------------------- 2 files changed, 30 insertions(+), 24 deletions(-) diff --git a/arch/x86/kernel/acpi/sleep.h b/arch/x86/kernel/acpi/sleep.h index 416d4be..fe5fdda 100644 --- a/arch/x86/kernel/acpi/sleep.h +++ b/arch/x86/kernel/acpi/sleep.h @@ -9,6 +9,8 @@ extern long saved_magic; extern int wakeup_pmode_return; +extern u8 wake_sleep_flags; + extern unsigned long acpi_copy_wakeup_routine(unsigned long); extern void wakeup_long64(void); diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c index 1d661b5..eb6fd23 100644 --- a/drivers/acpi/sleep.c +++ b/drivers/acpi/sleep.c @@ -28,23 +28,33 @@ #include "internal.h" #include "sleep.h" +u8 wake_sleep_flags = ACPI_NO_OPTIONAL_METHODS; static unsigned int gts, bfs; -module_param(gts, uint, 0644); -module_param(bfs, uint, 0644); -MODULE_PARM_DESC(gts, "Enable evaluation of _GTS on suspend."); -MODULE_PARM_DESC(bfs, "Enable evaluation of _BFS on resume".); - -static u8 wake_sleep_flags(void) +static int set_param_wake_flag(const char *val, struct kernel_param *kp) { - u8 flags = ACPI_NO_OPTIONAL_METHODS; + int ret = param_set_int(val, kp); - if (gts) - flags |= ACPI_EXECUTE_GTS; - if (bfs) - flags |= ACPI_EXECUTE_BFS; + if (ret) + return ret; - return flags; + if (kp->arg == (const char *)>s) { + if (gts) + wake_sleep_flags |= ACPI_EXECUTE_GTS; + else + wake_sleep_flags &= ~ACPI_EXECUTE_GTS; + } + if (kp->arg == (const char *)&bfs) { + if (bfs) + wake_sleep_flags |= ACPI_EXECUTE_BFS; + else + wake_sleep_flags &= ~ACPI_EXECUTE_BFS; + } + return ret; } +module_param_call(gts, set_param_wake_flag, param_get_int, >s, 0644); +module_param_call(bfs, set_param_wake_flag, param_get_int, &bfs, 0644); +MODULE_PARM_DESC(gts, "Enable evaluation of _GTS on suspend."); +MODULE_PARM_DESC(bfs, "Enable evaluation of _BFS on resume".); static u8 sleep_states[ACPI_S_STATE_COUNT]; @@ -263,7 +273,6 @@ static int acpi_suspend_enter(suspend_state_t pm_state) { acpi_status status = AE_OK; u32 acpi_state = acpi_target_sleep_state; - u8 flags = wake_sleep_flags(); int error; ACPI_FLUSH_CPU_CACHE(); @@ -271,7 +280,7 @@ static int acpi_suspend_enter(suspend_state_t pm_state) switch (acpi_state) { case ACPI_STATE_S1: barrier(); - status = acpi_enter_sleep_state(acpi_state, flags); + status = acpi_enter_sleep_state(acpi_state, wake_sleep_flags); break; case ACPI_STATE_S3: @@ -286,7 +295,7 @@ static int acpi_suspend_enter(suspend_state_t pm_state) acpi_write_bit_register(ACPI_BITREG_SCI_ENABLE, 1); /* Reprogram control registers and execute _BFS */ - acpi_leave_sleep_state_prep(acpi_state, flags); + acpi_leave_sleep_state_prep(acpi_state, wake_sleep_flags); /* ACPI 3.0 specs (P62) says that it's the responsibility * of the OSPM to clear the status bit [ implying that the @@ -550,30 +559,27 @@ static int acpi_hibernation_begin(void) static int acpi_hibernation_enter(void) { - u8 flags = wake_sleep_flags(); acpi_status status = AE_OK; ACPI_FLUSH_CPU_CACHE(); /* This shouldn't return. If it returns, we have a problem */ - status = acpi_enter_sleep_state(ACPI_STATE_S4, flags); + status = acpi_enter_sleep_state(ACPI_STATE_S4, wake_sleep_flags); /* Reprogram control registers and execute _BFS */ - acpi_leave_sleep_state_prep(ACPI_STATE_S4, flags); + acpi_leave_sleep_state_prep(ACPI_STATE_S4, wake_sleep_flags); return ACPI_SUCCESS(status) ? 0 : -EFAULT; } static void acpi_hibernation_leave(void) { - u8 flags = wake_sleep_flags(); - /* * If ACPI is not enabled by the BIOS and the boot kernel, we need to * enable it here. */ acpi_enable(); /* Reprogram control registers and execute _BFS */ - acpi_leave_sleep_state_prep(ACPI_STATE_S4, flags); + acpi_leave_sleep_state_prep(ACPI_STATE_S4, wake_sleep_flags); /* Check the hardware signature */ if (facs && s4_hardware_signature != facs->hardware_signature) { printk(KERN_EMERG "ACPI: Hardware changed while hibernated, " @@ -828,12 +834,10 @@ static void acpi_power_off_prepare(void) static void acpi_power_off(void) { - u8 flags = wake_sleep_flags(); - /* acpi_sleep_prepare(ACPI_STATE_S5) should have already been called */ printk(KERN_DEBUG "%s called\n", __func__); local_irq_disable(); - acpi_enter_sleep_state(ACPI_STATE_S5, flags); + acpi_enter_sleep_state(ACPI_STATE_S5, wake_sleep_flags); } /* -- 1.7.7.5 ^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH 1/2] ACPI: Convert wake_sleep_flags to a value instead of function (v4) 2012-04-18 17:18 ` [PATCH 1/2] ACPI: Convert wake_sleep_flags to a value instead of function (v4) Konrad Rzeszutek Wilk @ 2012-04-18 19:46 ` Rafael J. Wysocki 0 siblings, 0 replies; 9+ messages in thread From: Rafael J. Wysocki @ 2012-04-18 19:46 UTC (permalink / raw) To: Konrad Rzeszutek Wilk Cc: ming.m.lin, linux-kernel, linux-pm, lenb, linux-acpi, hpa, x86 On Wednesday, April 18, 2012, Konrad Rzeszutek Wilk wrote: > With commit a2ef5c4fd44ce3922435139393b89f2cce47f576 > "ACPI: Move module parameter gts and bfs to sleep.c" the Hmm. I don't recall seeing that one before it was applied. > wake_sleep_flags is required when calling acpi_enter_sleep_state, > which means that if there are functions outside the sleep.c code > they can't get get the wake_sleep_flags values. s/get// > This converts the function in to a exported value and converts > the module config operands to a function. > > Acked-by: Lin Ming <ming.m.lin@intel.com> > [v2: Parameters can be turned on/off dynamically] > [v3: unsigned char -> u8] > [v4: val -> kp->arg] > Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Rafael J. Wysocki <rjw@sisk.pl> Or I can take it, depending on what's preferred by Len. Thanks, Rafael > --- > arch/x86/kernel/acpi/sleep.h | 2 + > drivers/acpi/sleep.c | 52 ++++++++++++++++++++++------------------- > 2 files changed, 30 insertions(+), 24 deletions(-) > > diff --git a/arch/x86/kernel/acpi/sleep.h b/arch/x86/kernel/acpi/sleep.h > index 416d4be..fe5fdda 100644 > --- a/arch/x86/kernel/acpi/sleep.h > +++ b/arch/x86/kernel/acpi/sleep.h > @@ -9,6 +9,8 @@ extern long saved_magic; > > extern int wakeup_pmode_return; > > +extern u8 wake_sleep_flags; > + > extern unsigned long acpi_copy_wakeup_routine(unsigned long); > extern void wakeup_long64(void); > > diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c > index 1d661b5..eb6fd23 100644 > --- a/drivers/acpi/sleep.c > +++ b/drivers/acpi/sleep.c > @@ -28,23 +28,33 @@ > #include "internal.h" > #include "sleep.h" > > +u8 wake_sleep_flags = ACPI_NO_OPTIONAL_METHODS; > static unsigned int gts, bfs; > -module_param(gts, uint, 0644); > -module_param(bfs, uint, 0644); > -MODULE_PARM_DESC(gts, "Enable evaluation of _GTS on suspend."); > -MODULE_PARM_DESC(bfs, "Enable evaluation of _BFS on resume".); > - > -static u8 wake_sleep_flags(void) > +static int set_param_wake_flag(const char *val, struct kernel_param *kp) > { > - u8 flags = ACPI_NO_OPTIONAL_METHODS; > + int ret = param_set_int(val, kp); > > - if (gts) > - flags |= ACPI_EXECUTE_GTS; > - if (bfs) > - flags |= ACPI_EXECUTE_BFS; > + if (ret) > + return ret; > > - return flags; > + if (kp->arg == (const char *)>s) { > + if (gts) > + wake_sleep_flags |= ACPI_EXECUTE_GTS; > + else > + wake_sleep_flags &= ~ACPI_EXECUTE_GTS; > + } > + if (kp->arg == (const char *)&bfs) { > + if (bfs) > + wake_sleep_flags |= ACPI_EXECUTE_BFS; > + else > + wake_sleep_flags &= ~ACPI_EXECUTE_BFS; > + } > + return ret; > } > +module_param_call(gts, set_param_wake_flag, param_get_int, >s, 0644); > +module_param_call(bfs, set_param_wake_flag, param_get_int, &bfs, 0644); > +MODULE_PARM_DESC(gts, "Enable evaluation of _GTS on suspend."); > +MODULE_PARM_DESC(bfs, "Enable evaluation of _BFS on resume".); > > static u8 sleep_states[ACPI_S_STATE_COUNT]; > > @@ -263,7 +273,6 @@ static int acpi_suspend_enter(suspend_state_t pm_state) > { > acpi_status status = AE_OK; > u32 acpi_state = acpi_target_sleep_state; > - u8 flags = wake_sleep_flags(); > int error; > > ACPI_FLUSH_CPU_CACHE(); > @@ -271,7 +280,7 @@ static int acpi_suspend_enter(suspend_state_t pm_state) > switch (acpi_state) { > case ACPI_STATE_S1: > barrier(); > - status = acpi_enter_sleep_state(acpi_state, flags); > + status = acpi_enter_sleep_state(acpi_state, wake_sleep_flags); > break; > > case ACPI_STATE_S3: > @@ -286,7 +295,7 @@ static int acpi_suspend_enter(suspend_state_t pm_state) > acpi_write_bit_register(ACPI_BITREG_SCI_ENABLE, 1); > > /* Reprogram control registers and execute _BFS */ > - acpi_leave_sleep_state_prep(acpi_state, flags); > + acpi_leave_sleep_state_prep(acpi_state, wake_sleep_flags); > > /* ACPI 3.0 specs (P62) says that it's the responsibility > * of the OSPM to clear the status bit [ implying that the > @@ -550,30 +559,27 @@ static int acpi_hibernation_begin(void) > > static int acpi_hibernation_enter(void) > { > - u8 flags = wake_sleep_flags(); > acpi_status status = AE_OK; > > ACPI_FLUSH_CPU_CACHE(); > > /* This shouldn't return. If it returns, we have a problem */ > - status = acpi_enter_sleep_state(ACPI_STATE_S4, flags); > + status = acpi_enter_sleep_state(ACPI_STATE_S4, wake_sleep_flags); > /* Reprogram control registers and execute _BFS */ > - acpi_leave_sleep_state_prep(ACPI_STATE_S4, flags); > + acpi_leave_sleep_state_prep(ACPI_STATE_S4, wake_sleep_flags); > > return ACPI_SUCCESS(status) ? 0 : -EFAULT; > } > > static void acpi_hibernation_leave(void) > { > - u8 flags = wake_sleep_flags(); > - > /* > * If ACPI is not enabled by the BIOS and the boot kernel, we need to > * enable it here. > */ > acpi_enable(); > /* Reprogram control registers and execute _BFS */ > - acpi_leave_sleep_state_prep(ACPI_STATE_S4, flags); > + acpi_leave_sleep_state_prep(ACPI_STATE_S4, wake_sleep_flags); > /* Check the hardware signature */ > if (facs && s4_hardware_signature != facs->hardware_signature) { > printk(KERN_EMERG "ACPI: Hardware changed while hibernated, " > @@ -828,12 +834,10 @@ static void acpi_power_off_prepare(void) > > static void acpi_power_off(void) > { > - u8 flags = wake_sleep_flags(); > - > /* acpi_sleep_prepare(ACPI_STATE_S5) should have already been called */ > printk(KERN_DEBUG "%s called\n", __func__); > local_irq_disable(); > - acpi_enter_sleep_state(ACPI_STATE_S5, flags); > + acpi_enter_sleep_state(ACPI_STATE_S5, wake_sleep_flags); > } > > /* > ^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH 2/2] x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler. 2012-04-18 17:18 [PATCH] gts and bfs fixes (v4) Konrad Rzeszutek Wilk 2012-04-18 17:18 ` [PATCH 1/2] ACPI: Convert wake_sleep_flags to a value instead of function (v4) Konrad Rzeszutek Wilk @ 2012-04-18 17:18 ` Konrad Rzeszutek Wilk 2012-04-18 19:48 ` Rafael J. Wysocki 2012-04-18 20:49 ` H. Peter Anvin 1 sibling, 2 replies; 9+ messages in thread From: Konrad Rzeszutek Wilk @ 2012-04-18 17:18 UTC (permalink / raw) To: ming.m.lin, linux-kernel, linux-pm, lenb, linux-acpi, hpa, x86, rjw Cc: Konrad Rzeszutek Wilk With commit a2ef5c4fd44ce3922435139393b89f2cce47f576 "ACPI: Move module parameter gts and bfs to sleep.c" the wake_sleep_flags is required when calling acpi_enter_sleep_state. The assembler code in wakeup_*.S did not do that. One solution is to call it from assembler and stick the wake_sleep_flags on the stack (for 32-bit) or in %esi (for 64-bit). hpa and rafael both suggested however to create a wrapper function to call acpi_enter_sleep_state and call said wrapper function ("acpi_enter_s3") from assembler. For 32-bit, the acpi_enter_s3 ends up looking as so: push %ebp mov %esp,%ebp sub $0x8,%esp movzbl 0xc1809314,%eax [wake_sleep_flags] movl $0x3,(%esp) mov %eax,0x4(%esp) call 0xc12d1fa0 <acpi_enter_sleep_state> leave ret And 64-bit: movzbl 0x9afde1(%rip),%esi [wake_sleep_flags] push %rbp mov $0x3,%edi mov %rsp,%rbp callq 0xffffffff812e9800 <acpi_enter_sleep_state> leaveq retq Suggested-by: H. Peter Anvin <hpa@zytor.com> [v2: Remove extra assembler operations, per hpa review] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> --- arch/x86/kernel/acpi/sleep.c | 4 ++++ arch/x86/kernel/acpi/sleep.h | 2 ++ arch/x86/kernel/acpi/wakeup_32.S | 4 +--- arch/x86/kernel/acpi/wakeup_64.S | 4 +--- 4 files changed, 8 insertions(+), 6 deletions(-) diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c index 103b6ab..146a49c 100644 --- a/arch/x86/kernel/acpi/sleep.c +++ b/arch/x86/kernel/acpi/sleep.c @@ -24,6 +24,10 @@ unsigned long acpi_realmode_flags; static char temp_stack[4096]; #endif +asmlinkage void acpi_enter_s3(void) +{ + acpi_enter_sleep_state(3, wake_sleep_flags); +} /** * acpi_suspend_lowlevel - save kernel state * diff --git a/arch/x86/kernel/acpi/sleep.h b/arch/x86/kernel/acpi/sleep.h index fe5fdda..d68677a 100644 --- a/arch/x86/kernel/acpi/sleep.h +++ b/arch/x86/kernel/acpi/sleep.h @@ -3,6 +3,7 @@ */ #include <asm/trampoline.h> +#include <linux/linkage.h> extern unsigned long saved_video_mode; extern long saved_magic; @@ -10,6 +11,7 @@ extern long saved_magic; extern int wakeup_pmode_return; extern u8 wake_sleep_flags; +extern asmlinkage void acpi_enter_s3(void); extern unsigned long acpi_copy_wakeup_routine(unsigned long); extern void wakeup_long64(void); diff --git a/arch/x86/kernel/acpi/wakeup_32.S b/arch/x86/kernel/acpi/wakeup_32.S index 13ab720..7261083 100644 --- a/arch/x86/kernel/acpi/wakeup_32.S +++ b/arch/x86/kernel/acpi/wakeup_32.S @@ -74,9 +74,7 @@ restore_registers: ENTRY(do_suspend_lowlevel) call save_processor_state call save_registers - pushl $3 - call acpi_enter_sleep_state - addl $4, %esp + call acpi_enter_s3 # In case of S3 failure, we'll emerge here. Jump # to ret_point to recover diff --git a/arch/x86/kernel/acpi/wakeup_64.S b/arch/x86/kernel/acpi/wakeup_64.S index 8ea5164..014d1d2 100644 --- a/arch/x86/kernel/acpi/wakeup_64.S +++ b/arch/x86/kernel/acpi/wakeup_64.S @@ -71,9 +71,7 @@ ENTRY(do_suspend_lowlevel) movq %rsi, saved_rsi addq $8, %rsp - movl $3, %edi - xorl %eax, %eax - call acpi_enter_sleep_state + call acpi_enter_s3 /* in case something went wrong, restore the machine status and go on */ jmp resume_point -- 1.7.7.5 ^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH 2/2] x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler. 2012-04-18 17:18 ` [PATCH 2/2] x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler Konrad Rzeszutek Wilk @ 2012-04-18 19:48 ` Rafael J. Wysocki 2012-04-18 20:49 ` H. Peter Anvin 1 sibling, 0 replies; 9+ messages in thread From: Rafael J. Wysocki @ 2012-04-18 19:48 UTC (permalink / raw) To: Konrad Rzeszutek Wilk Cc: ming.m.lin, linux-kernel, linux-pm, lenb, linux-acpi, hpa, x86 On Wednesday, April 18, 2012, Konrad Rzeszutek Wilk wrote: > With commit a2ef5c4fd44ce3922435139393b89f2cce47f576 > "ACPI: Move module parameter gts and bfs to sleep.c" the > wake_sleep_flags is required when calling acpi_enter_sleep_state. > > The assembler code in wakeup_*.S did not do that. One solution > is to call it from assembler and stick the wake_sleep_flags on > the stack (for 32-bit) or in %esi (for 64-bit). hpa and rafael > both suggested however to create a wrapper function to call > acpi_enter_sleep_state and call said wrapper function > ("acpi_enter_s3") from assembler. > > For 32-bit, the acpi_enter_s3 ends up looking as so: > > push %ebp > mov %esp,%ebp > sub $0x8,%esp > movzbl 0xc1809314,%eax [wake_sleep_flags] > movl $0x3,(%esp) > mov %eax,0x4(%esp) > call 0xc12d1fa0 <acpi_enter_sleep_state> > leave > ret > > And 64-bit: > > movzbl 0x9afde1(%rip),%esi [wake_sleep_flags] > push %rbp > mov $0x3,%edi > mov %rsp,%rbp > callq 0xffffffff812e9800 <acpi_enter_sleep_state> > leaveq > retq > > Suggested-by: H. Peter Anvin <hpa@zytor.com> > [v2: Remove extra assembler operations, per hpa review] > Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Rafael J. Wysocki <rjw@sisk.pl> > --- > arch/x86/kernel/acpi/sleep.c | 4 ++++ > arch/x86/kernel/acpi/sleep.h | 2 ++ > arch/x86/kernel/acpi/wakeup_32.S | 4 +--- > arch/x86/kernel/acpi/wakeup_64.S | 4 +--- > 4 files changed, 8 insertions(+), 6 deletions(-) > > diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c > index 103b6ab..146a49c 100644 > --- a/arch/x86/kernel/acpi/sleep.c > +++ b/arch/x86/kernel/acpi/sleep.c > @@ -24,6 +24,10 @@ unsigned long acpi_realmode_flags; > static char temp_stack[4096]; > #endif > > +asmlinkage void acpi_enter_s3(void) > +{ > + acpi_enter_sleep_state(3, wake_sleep_flags); > +} > /** > * acpi_suspend_lowlevel - save kernel state > * > diff --git a/arch/x86/kernel/acpi/sleep.h b/arch/x86/kernel/acpi/sleep.h > index fe5fdda..d68677a 100644 > --- a/arch/x86/kernel/acpi/sleep.h > +++ b/arch/x86/kernel/acpi/sleep.h > @@ -3,6 +3,7 @@ > */ > > #include <asm/trampoline.h> > +#include <linux/linkage.h> > > extern unsigned long saved_video_mode; > extern long saved_magic; > @@ -10,6 +11,7 @@ extern long saved_magic; > extern int wakeup_pmode_return; > > extern u8 wake_sleep_flags; > +extern asmlinkage void acpi_enter_s3(void); > > extern unsigned long acpi_copy_wakeup_routine(unsigned long); > extern void wakeup_long64(void); > diff --git a/arch/x86/kernel/acpi/wakeup_32.S b/arch/x86/kernel/acpi/wakeup_32.S > index 13ab720..7261083 100644 > --- a/arch/x86/kernel/acpi/wakeup_32.S > +++ b/arch/x86/kernel/acpi/wakeup_32.S > @@ -74,9 +74,7 @@ restore_registers: > ENTRY(do_suspend_lowlevel) > call save_processor_state > call save_registers > - pushl $3 > - call acpi_enter_sleep_state > - addl $4, %esp > + call acpi_enter_s3 > > # In case of S3 failure, we'll emerge here. Jump > # to ret_point to recover > diff --git a/arch/x86/kernel/acpi/wakeup_64.S b/arch/x86/kernel/acpi/wakeup_64.S > index 8ea5164..014d1d2 100644 > --- a/arch/x86/kernel/acpi/wakeup_64.S > +++ b/arch/x86/kernel/acpi/wakeup_64.S > @@ -71,9 +71,7 @@ ENTRY(do_suspend_lowlevel) > movq %rsi, saved_rsi > > addq $8, %rsp > - movl $3, %edi > - xorl %eax, %eax > - call acpi_enter_sleep_state > + call acpi_enter_s3 > /* in case something went wrong, restore the machine status and go on */ > jmp resume_point > > ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 2/2] x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler. 2012-04-18 17:18 ` [PATCH 2/2] x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler Konrad Rzeszutek Wilk 2012-04-18 19:48 ` Rafael J. Wysocki @ 2012-04-18 20:49 ` H. Peter Anvin 2012-04-18 21:02 ` Rafael J. Wysocki 1 sibling, 1 reply; 9+ messages in thread From: H. Peter Anvin @ 2012-04-18 20:49 UTC (permalink / raw) To: Konrad Rzeszutek Wilk Cc: ming.m.lin, linux-kernel, linux-pm, lenb, linux-acpi, x86, rjw, Brown, Len On 04/18/2012 10:18 AM, Konrad Rzeszutek Wilk wrote: > > Suggested-by: H. Peter Anvin <hpa@zytor.com> > [v2: Remove extra assembler operations, per hpa review] > Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Reviewed-by: H. Peter Anvin <hpa@zytor.com> Len, your tree or mine? -hpa ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 2/2] x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler. 2012-04-18 20:49 ` H. Peter Anvin @ 2012-04-18 21:02 ` Rafael J. Wysocki 2012-05-08 6:35 ` Len Brown 0 siblings, 1 reply; 9+ messages in thread From: Rafael J. Wysocki @ 2012-04-18 21:02 UTC (permalink / raw) To: H. Peter Anvin Cc: Konrad Rzeszutek Wilk, ming.m.lin, linux-kernel, linux-pm, lenb, linux-acpi, x86, Brown, Len On Wednesday, April 18, 2012, H. Peter Anvin wrote: > On 04/18/2012 10:18 AM, Konrad Rzeszutek Wilk wrote: > > > > Suggested-by: H. Peter Anvin <hpa@zytor.com> > > [v2: Remove extra assembler operations, per hpa review] > > Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > > Reviewed-by: H. Peter Anvin <hpa@zytor.com> > > Len, your tree or mine? Well, I think you can take both, if that's not a problem. :-) Rafael ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 2/2] x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler. 2012-04-18 21:02 ` Rafael J. Wysocki @ 2012-05-08 6:35 ` Len Brown 0 siblings, 0 replies; 9+ messages in thread From: Len Brown @ 2012-05-08 6:35 UTC (permalink / raw) To: Rafael J. Wysocki Cc: H. Peter Anvin, Konrad Rzeszutek Wilk, ming.m.lin, linux-kernel, linux-pm, linux-acpi, x86, Brown, Len On 04/18/2012 05:02 PM, Rafael J. Wysocki wrote: > On Wednesday, April 18, 2012, H. Peter Anvin wrote: >> On 04/18/2012 10:18 AM, Konrad Rzeszutek Wilk wrote: >>> >>> Suggested-by: H. Peter Anvin <hpa@zytor.com> >>> [v2: Remove extra assembler operations, per hpa review] >>> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> >> >> Reviewed-by: H. Peter Anvin <hpa@zytor.com> >> >> Len, your tree or mine? I think it was a mistake to apply this. Instead we should have removed the code that it fixed. -Len ^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH] gts and bfs fixes (v3). @ 2012-04-17 18:29 Konrad Rzeszutek Wilk 2012-04-17 18:29 ` [PATCH 1/2] ACPI: Convert wake_sleep_flags to a value instead of function (v4) Konrad Rzeszutek Wilk 0 siblings, 1 reply; 9+ messages in thread From: Konrad Rzeszutek Wilk @ 2012-04-17 18:29 UTC (permalink / raw) To: linux-kernel, lenb, linux-acpi, ming.m.lin, hpa, rjw, x86 Since v2 [https://lkml.org/lkml/2012/4/9/474] - ditched the assembler version and used a C wrapper With git commit a2ef5c4fd44ce3922435139393b89f2cce47f576 "ACPI: Move module parameter gts and bfs to sleep.c" the sampling of gts/bfs module parameters flag by acpi_enter_sleep_state has been moved. It has been moved out of the inner ACPI code and out in the more x86 specific. This means that any caller of the exported acpi_enter_sleep_state function needs to provide two parameters now: sleep state and the bit flag whether to us _GTS|_BFS. All of the callers of acpi_enter_sleep_state did that, with the exception of the assembler version and as well any third-party callers of acpi_enter_sleep_state. These two patches expose 'wake_sleep_flags' as the product of the gts/bfs module parameter invocation. And also use said flag in the assembler 'do_suspend_lowlevel()' function. Konrad Rzeszutek Wilk (2): ACPI: Convert wake_sleep_flags to a value instead of function (v4) x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler. arch/x86/kernel/acpi/sleep.c | 4 +++ arch/x86/kernel/acpi/sleep.h | 4 +++ arch/x86/kernel/acpi/wakeup_32.S | 3 +- arch/x86/kernel/acpi/wakeup_64.S | 3 +- drivers/acpi/sleep.c | 52 ++++++++++++++++++++----------------- 5 files changed, 38 insertions(+), 28 deletions(-) ^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH 1/2] ACPI: Convert wake_sleep_flags to a value instead of function (v4) 2012-04-17 18:29 [PATCH] gts and bfs fixes (v3) Konrad Rzeszutek Wilk @ 2012-04-17 18:29 ` Konrad Rzeszutek Wilk 0 siblings, 0 replies; 9+ messages in thread From: Konrad Rzeszutek Wilk @ 2012-04-17 18:29 UTC (permalink / raw) To: linux-kernel, lenb, linux-acpi, ming.m.lin, hpa, rjw, x86 Cc: Konrad Rzeszutek Wilk With commit a2ef5c4fd44ce3922435139393b89f2cce47f576 "ACPI: Move module parameter gts and bfs to sleep.c" the wake_sleep_flags is required when calling acpi_enter_sleep_state, which means that if there are functions outside the sleep.c code they can't get get the wake_sleep_flags values. This converts the function in to a exported value and converts the module config operands to a function. Acked-by: Lin Ming <ming.m.lin@intel.com> [v2: Parameters can be turned on/off dynamically] [v3: unsigned char -> u8] [v4: val -> kp->arg] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> --- arch/x86/kernel/acpi/sleep.h | 2 + drivers/acpi/sleep.c | 52 ++++++++++++++++++++++------------------- 2 files changed, 30 insertions(+), 24 deletions(-) diff --git a/arch/x86/kernel/acpi/sleep.h b/arch/x86/kernel/acpi/sleep.h index 416d4be..fe5fdda 100644 --- a/arch/x86/kernel/acpi/sleep.h +++ b/arch/x86/kernel/acpi/sleep.h @@ -9,6 +9,8 @@ extern long saved_magic; extern int wakeup_pmode_return; +extern u8 wake_sleep_flags; + extern unsigned long acpi_copy_wakeup_routine(unsigned long); extern void wakeup_long64(void); diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c index 1d661b5..eb6fd23 100644 --- a/drivers/acpi/sleep.c +++ b/drivers/acpi/sleep.c @@ -28,23 +28,33 @@ #include "internal.h" #include "sleep.h" +u8 wake_sleep_flags = ACPI_NO_OPTIONAL_METHODS; static unsigned int gts, bfs; -module_param(gts, uint, 0644); -module_param(bfs, uint, 0644); -MODULE_PARM_DESC(gts, "Enable evaluation of _GTS on suspend."); -MODULE_PARM_DESC(bfs, "Enable evaluation of _BFS on resume".); - -static u8 wake_sleep_flags(void) +static int set_param_wake_flag(const char *val, struct kernel_param *kp) { - u8 flags = ACPI_NO_OPTIONAL_METHODS; + int ret = param_set_int(val, kp); - if (gts) - flags |= ACPI_EXECUTE_GTS; - if (bfs) - flags |= ACPI_EXECUTE_BFS; + if (ret) + return ret; - return flags; + if (kp->arg == (const char *)>s) { + if (gts) + wake_sleep_flags |= ACPI_EXECUTE_GTS; + else + wake_sleep_flags &= ~ACPI_EXECUTE_GTS; + } + if (kp->arg == (const char *)&bfs) { + if (bfs) + wake_sleep_flags |= ACPI_EXECUTE_BFS; + else + wake_sleep_flags &= ~ACPI_EXECUTE_BFS; + } + return ret; } +module_param_call(gts, set_param_wake_flag, param_get_int, >s, 0644); +module_param_call(bfs, set_param_wake_flag, param_get_int, &bfs, 0644); +MODULE_PARM_DESC(gts, "Enable evaluation of _GTS on suspend."); +MODULE_PARM_DESC(bfs, "Enable evaluation of _BFS on resume".); static u8 sleep_states[ACPI_S_STATE_COUNT]; @@ -263,7 +273,6 @@ static int acpi_suspend_enter(suspend_state_t pm_state) { acpi_status status = AE_OK; u32 acpi_state = acpi_target_sleep_state; - u8 flags = wake_sleep_flags(); int error; ACPI_FLUSH_CPU_CACHE(); @@ -271,7 +280,7 @@ static int acpi_suspend_enter(suspend_state_t pm_state) switch (acpi_state) { case ACPI_STATE_S1: barrier(); - status = acpi_enter_sleep_state(acpi_state, flags); + status = acpi_enter_sleep_state(acpi_state, wake_sleep_flags); break; case ACPI_STATE_S3: @@ -286,7 +295,7 @@ static int acpi_suspend_enter(suspend_state_t pm_state) acpi_write_bit_register(ACPI_BITREG_SCI_ENABLE, 1); /* Reprogram control registers and execute _BFS */ - acpi_leave_sleep_state_prep(acpi_state, flags); + acpi_leave_sleep_state_prep(acpi_state, wake_sleep_flags); /* ACPI 3.0 specs (P62) says that it's the responsibility * of the OSPM to clear the status bit [ implying that the @@ -550,30 +559,27 @@ static int acpi_hibernation_begin(void) static int acpi_hibernation_enter(void) { - u8 flags = wake_sleep_flags(); acpi_status status = AE_OK; ACPI_FLUSH_CPU_CACHE(); /* This shouldn't return. If it returns, we have a problem */ - status = acpi_enter_sleep_state(ACPI_STATE_S4, flags); + status = acpi_enter_sleep_state(ACPI_STATE_S4, wake_sleep_flags); /* Reprogram control registers and execute _BFS */ - acpi_leave_sleep_state_prep(ACPI_STATE_S4, flags); + acpi_leave_sleep_state_prep(ACPI_STATE_S4, wake_sleep_flags); return ACPI_SUCCESS(status) ? 0 : -EFAULT; } static void acpi_hibernation_leave(void) { - u8 flags = wake_sleep_flags(); - /* * If ACPI is not enabled by the BIOS and the boot kernel, we need to * enable it here. */ acpi_enable(); /* Reprogram control registers and execute _BFS */ - acpi_leave_sleep_state_prep(ACPI_STATE_S4, flags); + acpi_leave_sleep_state_prep(ACPI_STATE_S4, wake_sleep_flags); /* Check the hardware signature */ if (facs && s4_hardware_signature != facs->hardware_signature) { printk(KERN_EMERG "ACPI: Hardware changed while hibernated, " @@ -828,12 +834,10 @@ static void acpi_power_off_prepare(void) static void acpi_power_off(void) { - u8 flags = wake_sleep_flags(); - /* acpi_sleep_prepare(ACPI_STATE_S5) should have already been called */ printk(KERN_DEBUG "%s called\n", __func__); local_irq_disable(); - acpi_enter_sleep_state(ACPI_STATE_S5, flags); + acpi_enter_sleep_state(ACPI_STATE_S5, wake_sleep_flags); } /* -- 1.7.7.5 ^ permalink raw reply related [flat|nested] 9+ messages in thread
end of thread, other threads:[~2012-05-08 6:35 UTC | newest] Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2012-04-18 17:18 [PATCH] gts and bfs fixes (v4) Konrad Rzeszutek Wilk 2012-04-18 17:18 ` [PATCH 1/2] ACPI: Convert wake_sleep_flags to a value instead of function (v4) Konrad Rzeszutek Wilk 2012-04-18 19:46 ` Rafael J. Wysocki 2012-04-18 17:18 ` [PATCH 2/2] x86/acpi: Call acpi_enter_sleep_state via an asmlinkage C function from assembler Konrad Rzeszutek Wilk 2012-04-18 19:48 ` Rafael J. Wysocki 2012-04-18 20:49 ` H. Peter Anvin 2012-04-18 21:02 ` Rafael J. Wysocki 2012-05-08 6:35 ` Len Brown -- strict thread matches above, loose matches on Subject: below -- 2012-04-17 18:29 [PATCH] gts and bfs fixes (v3) Konrad Rzeszutek Wilk 2012-04-17 18:29 ` [PATCH 1/2] ACPI: Convert wake_sleep_flags to a value instead of function (v4) Konrad Rzeszutek Wilk
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).