| Message ID | 20220317141445.6498-1-mario.limonciello@amd.com (mailing list archive) |
|---|---|
| State | Handled Elsewhere, archived |
| Headers | show |
| Series | [v5,1/4] ACPI / x86: Add support for LPS0 callback handler | expand |
On Thu, Mar 17, 2022 at 3:15 PM Mario Limonciello <mario.limonciello@amd.com> wrote: > > Currenty the latest thing run during a suspend to idle attempt is > the LPS0 `prepare_late` callback and the earliest thing is the > `resume_early` callback. > > There is a desire for the `amd-pmc` driver to suspend later in the > suspend process (ideally the very last thing), so create a callback > that it or any other driver can hook into to do this. > > Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> This looks good to me, so: Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> and I don't have any conflicting changes queued up for 5.17, so from my POV it can go in via the x86 platform drivers tree along with the other patches in the series. > --- > changes from v4->v5: > * fix a check for handler->prepare > Reported by Dan Carpenter <dan.carpenter@oracle.com> > changes from v3->v4: > * drop use of mutex, use lock_system_sleep instead > * don't pass argument of context, driver will maintain this internally > * don't allow failing prepare stage > * don't allow unregistering if sleep_no_lps0 is set > changes from v2->v3: > * Check that callbacks exist before calling > changes from v1->v2: > * Change register/unregister arguments to be struct > > drivers/acpi/x86/s2idle.c | 40 +++++++++++++++++++++++++++++++++++++++ > include/linux/acpi.h | 10 +++++++++- > 2 files changed, 49 insertions(+), 1 deletion(-) > > diff --git a/drivers/acpi/x86/s2idle.c b/drivers/acpi/x86/s2idle.c > index abc06e7f89d8..031b20a547f9 100644 > --- a/drivers/acpi/x86/s2idle.c > +++ b/drivers/acpi/x86/s2idle.c > @@ -86,6 +86,8 @@ struct lpi_device_constraint_amd { > int min_dstate; > }; > > +static LIST_HEAD(lps0_s2idle_devops_head); > + > static struct lpi_constraints *lpi_constraints_table; > static int lpi_constraints_table_size; > static int rev_id; > @@ -444,6 +446,8 @@ static struct acpi_scan_handler lps0_handler = { > > int acpi_s2idle_prepare_late(void) > { > + struct acpi_s2idle_dev_ops *handler; > + > if (!lps0_device_handle || sleep_no_lps0) > return 0; > > @@ -474,14 +478,26 @@ int acpi_s2idle_prepare_late(void) > acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_ENTRY, > lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft); > } > + > + list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) { > + if (handler->prepare) > + handler->prepare(); > + } > + > return 0; > } > > void acpi_s2idle_restore_early(void) > { > + struct acpi_s2idle_dev_ops *handler; > + > if (!lps0_device_handle || sleep_no_lps0) > return; > > + list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) > + if (handler->restore) > + handler->restore(); > + > /* Modern standby exit */ > if (lps0_dsm_func_mask_microsoft > 0) > acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT, > @@ -524,4 +540,28 @@ void acpi_s2idle_setup(void) > s2idle_set_ops(&acpi_s2idle_ops_lps0); > } > > +int acpi_register_lps0_dev(struct acpi_s2idle_dev_ops *arg) > +{ > + if (!lps0_device_handle || sleep_no_lps0) > + return -ENODEV; > + > + lock_system_sleep(); > + list_add(&arg->list_node, &lps0_s2idle_devops_head); > + unlock_system_sleep(); > + > + return 0; > +} > +EXPORT_SYMBOL_GPL(acpi_register_lps0_dev); > + > +void acpi_unregister_lps0_dev(struct acpi_s2idle_dev_ops *arg) > +{ > + if (!lps0_device_handle || sleep_no_lps0) > + return; > + > + lock_system_sleep(); > + list_del(&arg->list_node); > + unlock_system_sleep(); > +} > +EXPORT_SYMBOL_GPL(acpi_unregister_lps0_dev); > + > #endif /* CONFIG_SUSPEND */ > diff --git a/include/linux/acpi.h b/include/linux/acpi.h > index 6274758648e3..47c16cdc8e0e 100644 > --- a/include/linux/acpi.h > +++ b/include/linux/acpi.h > @@ -1023,7 +1023,15 @@ void acpi_os_set_prepare_extended_sleep(int (*func)(u8 sleep_state, > > acpi_status acpi_os_prepare_extended_sleep(u8 sleep_state, > u32 val_a, u32 val_b); > - > +#ifdef CONFIG_X86 > +struct acpi_s2idle_dev_ops { > + struct list_head list_node; > + void (*prepare)(void); > + void (*restore)(void); > +}; > +int acpi_register_lps0_dev(struct acpi_s2idle_dev_ops *arg); > +void acpi_unregister_lps0_dev(struct acpi_s2idle_dev_ops *arg); > +#endif /* CONFIG_X86 */ > #ifndef CONFIG_IA64 > void arch_reserve_mem_area(acpi_physical_address addr, size_t size); > #else > -- > 2.34.1 >
[Public] > -----Original Message----- > From: Rafael J. Wysocki <rafael@kernel.org> > Sent: Thursday, March 17, 2022 11:46 > To: Limonciello, Mario <Mario.Limonciello@amd.com> > Cc: Hans de Goede <hdegoede@redhat.com>; Mark Gross > <mgross@linux.intel.com>; Rafael J . Wysocki <rjw@rjwysocki.net>; open > list:X86 PLATFORM DRIVERS <platform-driver-x86@vger.kernel.org>; ACPI > Devel Maling List <linux-acpi@vger.kernel.org>; S-k, Shyam-sundar <Shyam- > sundar.S-k@amd.com>; Goswami, Sanket <Sanket.Goswami@amd.com> > Subject: Re: [PATCH v5 1/4] ACPI / x86: Add support for LPS0 callback handler > > On Thu, Mar 17, 2022 at 3:15 PM Mario Limonciello > <mario.limonciello@amd.com> wrote: > > > > Currenty the latest thing run during a suspend to idle attempt is > > the LPS0 `prepare_late` callback and the earliest thing is the > > `resume_early` callback. > > > > There is a desire for the `amd-pmc` driver to suspend later in the > > suspend process (ideally the very last thing), so create a callback > > that it or any other driver can hook into to do this. > > > > Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> > > This looks good to me, so: > > Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> > Thanks! > and I don't have any conflicting changes queued up for 5.17, so from > my POV it can go in via the x86 platform drivers tree along with the > other patches in the series. Hans, FYI the series is based off of platform-x86/for-next also with changes to amd-pmc Rafael doesn't have in his tree, so I also agree it's better to go through x86 platform drivers. > > > --- > > changes from v4->v5: > > * fix a check for handler->prepare > > Reported by Dan Carpenter <dan.carpenter@oracle.com> > > changes from v3->v4: > > * drop use of mutex, use lock_system_sleep instead > > * don't pass argument of context, driver will maintain this internally > > * don't allow failing prepare stage > > * don't allow unregistering if sleep_no_lps0 is set > > changes from v2->v3: > > * Check that callbacks exist before calling > > changes from v1->v2: > > * Change register/unregister arguments to be struct > > > > drivers/acpi/x86/s2idle.c | 40 > +++++++++++++++++++++++++++++++++++++++ > > include/linux/acpi.h | 10 +++++++++- > > 2 files changed, 49 insertions(+), 1 deletion(-) > > > > diff --git a/drivers/acpi/x86/s2idle.c b/drivers/acpi/x86/s2idle.c > > index abc06e7f89d8..031b20a547f9 100644 > > --- a/drivers/acpi/x86/s2idle.c > > +++ b/drivers/acpi/x86/s2idle.c > > @@ -86,6 +86,8 @@ struct lpi_device_constraint_amd { > > int min_dstate; > > }; > > > > +static LIST_HEAD(lps0_s2idle_devops_head); > > + > > static struct lpi_constraints *lpi_constraints_table; > > static int lpi_constraints_table_size; > > static int rev_id; > > @@ -444,6 +446,8 @@ static struct acpi_scan_handler lps0_handler = { > > > > int acpi_s2idle_prepare_late(void) > > { > > + struct acpi_s2idle_dev_ops *handler; > > + > > if (!lps0_device_handle || sleep_no_lps0) > > return 0; > > > > @@ -474,14 +478,26 @@ int acpi_s2idle_prepare_late(void) > > acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_ENTRY, > > lps0_dsm_func_mask_microsoft, > lps0_dsm_guid_microsoft); > > } > > + > > + list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) { > > + if (handler->prepare) > > + handler->prepare(); > > + } > > + > > return 0; > > } > > > > void acpi_s2idle_restore_early(void) > > { > > + struct acpi_s2idle_dev_ops *handler; > > + > > if (!lps0_device_handle || sleep_no_lps0) > > return; > > > > + list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) > > + if (handler->restore) > > + handler->restore(); > > + > > /* Modern standby exit */ > > if (lps0_dsm_func_mask_microsoft > 0) > > acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT, > > @@ -524,4 +540,28 @@ void acpi_s2idle_setup(void) > > s2idle_set_ops(&acpi_s2idle_ops_lps0); > > } > > > > +int acpi_register_lps0_dev(struct acpi_s2idle_dev_ops *arg) > > +{ > > + if (!lps0_device_handle || sleep_no_lps0) > > + return -ENODEV; > > + > > + lock_system_sleep(); > > + list_add(&arg->list_node, &lps0_s2idle_devops_head); > > + unlock_system_sleep(); > > + > > + return 0; > > +} > > +EXPORT_SYMBOL_GPL(acpi_register_lps0_dev); > > + > > +void acpi_unregister_lps0_dev(struct acpi_s2idle_dev_ops *arg) > > +{ > > + if (!lps0_device_handle || sleep_no_lps0) > > + return; > > + > > + lock_system_sleep(); > > + list_del(&arg->list_node); > > + unlock_system_sleep(); > > +} > > +EXPORT_SYMBOL_GPL(acpi_unregister_lps0_dev); > > + > > #endif /* CONFIG_SUSPEND */ > > diff --git a/include/linux/acpi.h b/include/linux/acpi.h > > index 6274758648e3..47c16cdc8e0e 100644 > > --- a/include/linux/acpi.h > > +++ b/include/linux/acpi.h > > @@ -1023,7 +1023,15 @@ void acpi_os_set_prepare_extended_sleep(int > (*func)(u8 sleep_state, > > > > acpi_status acpi_os_prepare_extended_sleep(u8 sleep_state, > > u32 val_a, u32 val_b); > > - > > +#ifdef CONFIG_X86 > > +struct acpi_s2idle_dev_ops { > > + struct list_head list_node; > > + void (*prepare)(void); > > + void (*restore)(void); > > +}; > > +int acpi_register_lps0_dev(struct acpi_s2idle_dev_ops *arg); > > +void acpi_unregister_lps0_dev(struct acpi_s2idle_dev_ops *arg); > > +#endif /* CONFIG_X86 */ > > #ifndef CONFIG_IA64 > > void arch_reserve_mem_area(acpi_physical_address addr, size_t size); > > #else > > -- > > 2.34.1 > >
Hi All, On 3/17/22 17:49, Limonciello, Mario wrote: > [Public] > > > >> -----Original Message----- >> From: Rafael J. Wysocki <rafael@kernel.org> >> Sent: Thursday, March 17, 2022 11:46 >> To: Limonciello, Mario <Mario.Limonciello@amd.com> >> Cc: Hans de Goede <hdegoede@redhat.com>; Mark Gross >> <mgross@linux.intel.com>; Rafael J . Wysocki <rjw@rjwysocki.net>; open >> list:X86 PLATFORM DRIVERS <platform-driver-x86@vger.kernel.org>; ACPI >> Devel Maling List <linux-acpi@vger.kernel.org>; S-k, Shyam-sundar <Shyam- >> sundar.S-k@amd.com>; Goswami, Sanket <Sanket.Goswami@amd.com> >> Subject: Re: [PATCH v5 1/4] ACPI / x86: Add support for LPS0 callback handler >> >> On Thu, Mar 17, 2022 at 3:15 PM Mario Limonciello >> <mario.limonciello@amd.com> wrote: >>> >>> Currenty the latest thing run during a suspend to idle attempt is >>> the LPS0 `prepare_late` callback and the earliest thing is the >>> `resume_early` callback. >>> >>> There is a desire for the `amd-pmc` driver to suspend later in the >>> suspend process (ideally the very last thing), so create a callback >>> that it or any other driver can hook into to do this. >>> >>> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> >> >> This looks good to me, so: >> >> Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> >> > > Thanks! > >> and I don't have any conflicting changes queued up for 5.17, so from >> my POV it can go in via the x86 platform drivers tree along with the >> other patches in the series. > > Hans, > > FYI the series is based off of platform-x86/for-next also with changes to amd-pmc > Rafael doesn't have in his tree, so I also agree it's better to go through x86 platform > drivers. Ok, the entire series looks good to me: Reviewed-by: Hans de Goede <hdegoede@redhat.com> So I'll go and merge this into review-hans, do some compile tests and then push this directly to for-next to give it some minimum baking time there before the 5.18 merge-window opens. Regards, Hans > >> >>> --- >>> changes from v4->v5: >>> * fix a check for handler->prepare >>> Reported by Dan Carpenter <dan.carpenter@oracle.com> >>> changes from v3->v4: >>> * drop use of mutex, use lock_system_sleep instead >>> * don't pass argument of context, driver will maintain this internally >>> * don't allow failing prepare stage >>> * don't allow unregistering if sleep_no_lps0 is set >>> changes from v2->v3: >>> * Check that callbacks exist before calling >>> changes from v1->v2: >>> * Change register/unregister arguments to be struct >>> >>> drivers/acpi/x86/s2idle.c | 40 >> +++++++++++++++++++++++++++++++++++++++ >>> include/linux/acpi.h | 10 +++++++++- >>> 2 files changed, 49 insertions(+), 1 deletion(-) >>> >>> diff --git a/drivers/acpi/x86/s2idle.c b/drivers/acpi/x86/s2idle.c >>> index abc06e7f89d8..031b20a547f9 100644 >>> --- a/drivers/acpi/x86/s2idle.c >>> +++ b/drivers/acpi/x86/s2idle.c >>> @@ -86,6 +86,8 @@ struct lpi_device_constraint_amd { >>> int min_dstate; >>> }; >>> >>> +static LIST_HEAD(lps0_s2idle_devops_head); >>> + >>> static struct lpi_constraints *lpi_constraints_table; >>> static int lpi_constraints_table_size; >>> static int rev_id; >>> @@ -444,6 +446,8 @@ static struct acpi_scan_handler lps0_handler = { >>> >>> int acpi_s2idle_prepare_late(void) >>> { >>> + struct acpi_s2idle_dev_ops *handler; >>> + >>> if (!lps0_device_handle || sleep_no_lps0) >>> return 0; >>> >>> @@ -474,14 +478,26 @@ int acpi_s2idle_prepare_late(void) >>> acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_ENTRY, >>> lps0_dsm_func_mask_microsoft, >> lps0_dsm_guid_microsoft); >>> } >>> + >>> + list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) { >>> + if (handler->prepare) >>> + handler->prepare(); >>> + } >>> + >>> return 0; >>> } >>> >>> void acpi_s2idle_restore_early(void) >>> { >>> + struct acpi_s2idle_dev_ops *handler; >>> + >>> if (!lps0_device_handle || sleep_no_lps0) >>> return; >>> >>> + list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) >>> + if (handler->restore) >>> + handler->restore(); >>> + >>> /* Modern standby exit */ >>> if (lps0_dsm_func_mask_microsoft > 0) >>> acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT, >>> @@ -524,4 +540,28 @@ void acpi_s2idle_setup(void) >>> s2idle_set_ops(&acpi_s2idle_ops_lps0); >>> } >>> >>> +int acpi_register_lps0_dev(struct acpi_s2idle_dev_ops *arg) >>> +{ >>> + if (!lps0_device_handle || sleep_no_lps0) >>> + return -ENODEV; >>> + >>> + lock_system_sleep(); >>> + list_add(&arg->list_node, &lps0_s2idle_devops_head); >>> + unlock_system_sleep(); >>> + >>> + return 0; >>> +} >>> +EXPORT_SYMBOL_GPL(acpi_register_lps0_dev); >>> + >>> +void acpi_unregister_lps0_dev(struct acpi_s2idle_dev_ops *arg) >>> +{ >>> + if (!lps0_device_handle || sleep_no_lps0) >>> + return; >>> + >>> + lock_system_sleep(); >>> + list_del(&arg->list_node); >>> + unlock_system_sleep(); >>> +} >>> +EXPORT_SYMBOL_GPL(acpi_unregister_lps0_dev); >>> + >>> #endif /* CONFIG_SUSPEND */ >>> diff --git a/include/linux/acpi.h b/include/linux/acpi.h >>> index 6274758648e3..47c16cdc8e0e 100644 >>> --- a/include/linux/acpi.h >>> +++ b/include/linux/acpi.h >>> @@ -1023,7 +1023,15 @@ void acpi_os_set_prepare_extended_sleep(int >> (*func)(u8 sleep_state, >>> >>> acpi_status acpi_os_prepare_extended_sleep(u8 sleep_state, >>> u32 val_a, u32 val_b); >>> - >>> +#ifdef CONFIG_X86 >>> +struct acpi_s2idle_dev_ops { >>> + struct list_head list_node; >>> + void (*prepare)(void); >>> + void (*restore)(void); >>> +}; >>> +int acpi_register_lps0_dev(struct acpi_s2idle_dev_ops *arg); >>> +void acpi_unregister_lps0_dev(struct acpi_s2idle_dev_ops *arg); >>> +#endif /* CONFIG_X86 */ >>> #ifndef CONFIG_IA64 >>> void arch_reserve_mem_area(acpi_physical_address addr, size_t size); >>> #else >>> -- >>> 2.34.1 >>>
diff --git a/drivers/acpi/x86/s2idle.c b/drivers/acpi/x86/s2idle.c index abc06e7f89d8..031b20a547f9 100644 --- a/drivers/acpi/x86/s2idle.c +++ b/drivers/acpi/x86/s2idle.c @@ -86,6 +86,8 @@ struct lpi_device_constraint_amd { int min_dstate; }; +static LIST_HEAD(lps0_s2idle_devops_head); + static struct lpi_constraints *lpi_constraints_table; static int lpi_constraints_table_size; static int rev_id; @@ -444,6 +446,8 @@ static struct acpi_scan_handler lps0_handler = { int acpi_s2idle_prepare_late(void) { + struct acpi_s2idle_dev_ops *handler; + if (!lps0_device_handle || sleep_no_lps0) return 0; @@ -474,14 +478,26 @@ int acpi_s2idle_prepare_late(void) acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_ENTRY, lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft); } + + list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) { + if (handler->prepare) + handler->prepare(); + } + return 0; } void acpi_s2idle_restore_early(void) { + struct acpi_s2idle_dev_ops *handler; + if (!lps0_device_handle || sleep_no_lps0) return; + list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) + if (handler->restore) + handler->restore(); + /* Modern standby exit */ if (lps0_dsm_func_mask_microsoft > 0) acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT, @@ -524,4 +540,28 @@ void acpi_s2idle_setup(void) s2idle_set_ops(&acpi_s2idle_ops_lps0); } +int acpi_register_lps0_dev(struct acpi_s2idle_dev_ops *arg) +{ + if (!lps0_device_handle || sleep_no_lps0) + return -ENODEV; + + lock_system_sleep(); + list_add(&arg->list_node, &lps0_s2idle_devops_head); + unlock_system_sleep(); + + return 0; +} +EXPORT_SYMBOL_GPL(acpi_register_lps0_dev); + +void acpi_unregister_lps0_dev(struct acpi_s2idle_dev_ops *arg) +{ + if (!lps0_device_handle || sleep_no_lps0) + return; + + lock_system_sleep(); + list_del(&arg->list_node); + unlock_system_sleep(); +} +EXPORT_SYMBOL_GPL(acpi_unregister_lps0_dev); + #endif /* CONFIG_SUSPEND */ diff --git a/include/linux/acpi.h b/include/linux/acpi.h index 6274758648e3..47c16cdc8e0e 100644 --- a/include/linux/acpi.h +++ b/include/linux/acpi.h @@ -1023,7 +1023,15 @@ void acpi_os_set_prepare_extended_sleep(int (*func)(u8 sleep_state, acpi_status acpi_os_prepare_extended_sleep(u8 sleep_state, u32 val_a, u32 val_b); - +#ifdef CONFIG_X86 +struct acpi_s2idle_dev_ops { + struct list_head list_node; + void (*prepare)(void); + void (*restore)(void); +}; +int acpi_register_lps0_dev(struct acpi_s2idle_dev_ops *arg); +void acpi_unregister_lps0_dev(struct acpi_s2idle_dev_ops *arg); +#endif /* CONFIG_X86 */ #ifndef CONFIG_IA64 void arch_reserve_mem_area(acpi_physical_address addr, size_t size); #else
Currenty the latest thing run during a suspend to idle attempt is the LPS0 `prepare_late` callback and the earliest thing is the `resume_early` callback. There is a desire for the `amd-pmc` driver to suspend later in the suspend process (ideally the very last thing), so create a callback that it or any other driver can hook into to do this. Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> --- changes from v4->v5: * fix a check for handler->prepare Reported by Dan Carpenter <dan.carpenter@oracle.com> changes from v3->v4: * drop use of mutex, use lock_system_sleep instead * don't pass argument of context, driver will maintain this internally * don't allow failing prepare stage * don't allow unregistering if sleep_no_lps0 is set changes from v2->v3: * Check that callbacks exist before calling changes from v1->v2: * Change register/unregister arguments to be struct drivers/acpi/x86/s2idle.c | 40 +++++++++++++++++++++++++++++++++++++++ include/linux/acpi.h | 10 +++++++++- 2 files changed, 49 insertions(+), 1 deletion(-)