From 6c661da8add5168cee3b745594d66173189bc8a2 Mon Sep 17 00:00:00 2001 From: Scott B <28817345+foundObjects@users.noreply.github.com> Date: Thu, 5 Aug 2021 02:39:45 -0700 Subject: [PATCH] v5.13.8 s0ix patch 2021-08-05 Squashed commit of the following: commit 8f506eaa976e148acdc856704a655659cadd279e Author: Mario Limonciello Date: Wed Jul 7 09:16:47 2021 -0500 platform/x86: amd-pmc: Use return code on suspend Right now the driver will still return success even if the OS_HINT command failed to send to the SMU. In the rare event of a failure, the suspend should really be aborted here so that relevant logs can may be captured. Signed-off-by: Mario Limonciello Acked-by: Shyam Sundar S K commit fd065a1e8ee224a0c550662f3812e987c6d21fa6 Author: Mario Limonciello Date: Wed Jun 30 14:46:06 2021 -0500 ACPI: PM: Only mark EC GPE for wakeup on Intel systems When using s2idle on a variety of AMD notebook systems, they are experiencing spurious events that the EC or SMU are in the wrong state leading to a hard time waking up or higher than expected power consumption. These events only occur when the EC GPE is inadvertently set as a wakeup source. Originally the EC GPE was only set as a wakeup source when using the intel-vbtn or intel-hid drivers in commit 10a08fd65ec1 ("ACPI: PM: Set up EC GPE for system wakeup from drivers that need it") but during testing a reporter discovered that this was not enough for their ASUS Zenbook UX430UNR/i7-8550U to wakeup by lid event or keypress. Marking the EC GPE for wakeup universally resolved this for that reporter in commit b90ff3554aa3 ("ACPI: PM: s2idle: Always set up EC GPE for system wakeup"). However this behavior has lead to a number of problems: * On both Lenovo T14 and P14s the keyboard wakeup doesn't work, and sometimes the power button event doesn't work. * On HP 635 G7 detaching or attaching AC during suspend will cause the system not to wakeup * On Asus vivobook to prevent detaching AC causing resume problems * On Lenovo 14ARE05 to prevent detaching AC causing resume problems * On HP ENVY x360 to prevent detaching AC causing resume problems As there may be other Intel systems besides ASUS Zenbook UX430UNR/i7-8550U that don't use intel-vbtn or intel-hid avoid these problems by only universally marking the EC GPE wakesource on non-AMD systems. Link: https://patchwork.kernel.org/project/linux-pm/cover/5997740.FPbUVk04hV@kreacher/#22825489 Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1230 Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1629 Signed-off-by: Mario Limonciello Acked-by: Alex Deucher commit cd38229157b4ed96831763fed09a2ef1d67466b2 Author: Shyam Sundar S K Date: Tue Jun 29 14:18:03 2021 +0530 platform/x86: amd-pmc: Add new acpi id for future PMC controllers The upcoming PMC controller would have a newer acpi id, add that to the supported acpid device list. Signed-off-by: Shyam Sundar S K Reviewed-by: Hans de Goede commit 6cbf4ca877f9a7763eb7ceaec82884f7542b8584 Author: Shyam Sundar S K Date: Tue Jun 29 14:18:02 2021 +0530 platform/x86: amd-pmc: Add support for ACPI ID AMDI0006 Some newer BIOSes have added another ACPI ID for the uPEP device. SMU statistics behave identically on this device. Signed-off-by: Shyam Sundar S K Reviewed-by: Hans de Goede commit b08043dc2dc982e77f2a620f6ac64dcc77085796 Author: Shyam Sundar S K Date: Tue Jun 29 14:18:01 2021 +0530 amd-pmc: Add support for logging s0ix counters Even the FCH SSC registers provides certain level of information about the s0ix entry and exit times which comes handy when the SMU fails to report the statistics via the mailbox communication. This information is captured via a new debugfs file "s0ix_stats". A non-zero entry in this counters would mean that the system entered the s0ix state. If s0ix entry time and exit time don't change during suspend to idle, the silicon has not entered the deepest state. Signed-off-by: Shyam Sundar S K Reviewed-by: Hans de Goede commit abfb68f228563b457a9f7f2c8db8f9946f1c082d Author: Shyam Sundar S K Date: Tue Jun 29 14:18:00 2021 +0530 platform/x86: amd-pmc: Add support for logging SMU metrics SMU provides a way to dump the s0ix debug statistics in the form of a metrics table via a of set special mailbox commands. Add support to the driver which can send these commands to SMU and expose the information received via debugfs. The information contains the s0ix entry/exit, active time of each IP block etc. As a side note, SMU subsystem logging is not supported on Picasso based SoC's. Signed-off-by: Shyam Sundar S K Reviewed-by: Hans de Goede commit df0f8a8e29ae2d10f34d0f1ba879cd6df0004871 Author: Shyam Sundar S K Date: Tue Jun 29 14:17:59 2021 +0530 platform/x86: amd-pmc: call dump registers only once Currently amd_pmc_dump_registers() routine is being called at multiple places. The best to call it is after command submission to SMU. Signed-off-by: Shyam Sundar S K commit 0dd5c53f0d5b1a07b0473562a6be34005eb852a2 Author: Marcin Bachry Date: Tue Mar 16 15:28:51 2021 -0400 PCI: quirks: Quirk PCI d3hot delay for AMD xhci Renoir needs a similar delay. Signed-off-by: Marcin Bachry Signed-off-by: Alex Deucher commit df67da22f52dcde4ca3bba9609f9605565b3cfaf Author: Mario Limonciello Date: Thu Jun 17 11:42:12 2021 -0500 ACPI: PM: Adjust behavior for field problems on AMD systems Some AMD Systems with uPEP _HID AMD004/AMDI005 have an off by one bug in their function mask return. This means that they will call entrance but not exit for matching functions. Other AMD systems with this HID should use the Microsoft generic UUID. AMD systems with uPEP HID AMDI006 should be using the Microsoft method. Signed-off-by: Mario Limonciello commit c53aff424e31bcb6706d6c7b8a6c06f0bca05a19 Author: Pratik Vishwakarma Date: Thu Jun 17 11:42:11 2021 -0500 ACPI: PM: s2idle: Add support for new Microsoft UUID This adds supports for _DSM notifications to the Microsoft UUID described by Microsoft documentation for s2idle. Link: https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-firmware-notifications Co-developed-by: Mario Limonciello Signed-off-by: Mario Limonciello Signed-off-by: Pratik Vishwakarma commit 2019bfd57b6fe44cae0d356f54b948bfd707fa97 Author: Pratik Vishwakarma Date: Thu Jun 17 11:42:10 2021 -0500 ACPI: PM: s2idle: Add support for multiple func mask Required for follow-up patch adding new UUID needing new function mask. Signed-off-by: Pratik Vishwakarma commit d9c8d6d7d178fbfedddc1d1a9f27d189cb8a4df9 Author: Pratik Vishwakarma Date: Thu Jun 17 11:42:09 2021 -0500 ACPI: PM: s2idle: Refactor common code Refactor common code to prepare for upcoming changes. * Remove unused struct. * Print error before returning. * Frees ACPI obj if _DSM type is not as expected. * Treat lps0_dsm_func_mask as an integer rather than character * Remove extra out_obj * Move rev_id Co-developed-by: Mario Limonciello Signed-off-by: Mario Limonciello Signed-off-by: Pratik Vishwakarma commit 9f339765df127c0973766c4c282aab07999292a8 Author: Pratik Vishwakarma Date: Thu Jun 17 11:42:08 2021 -0500 ACPI: PM: s2idle: Use correct revision id AMD spec mentions only revision 0. With this change, device constraint list is populated properly. Signed-off-by: Pratik Vishwakarma commit 6e853c597c7786939ea18e03e161d4b098a5201c Author: Mario Limonciello Date: Wed Jun 9 13:40:18 2021 -0500 ACPI: Add quirks for AMD Renoir/Lucienne CPUs to force the D3 hint AMD systems from Renoir and Lucienne require that the NVME controller is put into D3 over a Modern Standby / suspend-to-idle cycle. This is "typically" accomplished using the `StorageD3Enable` property in the _DSD, but this property was introduced after many of these systems launched and most OEM systems don't have it in their BIOS. On AMD Renoir without these drives going into D3 over suspend-to-idle the resume will fail with the NVME controller being reset and a trace like this in the kernel logs: ``` [ 83.556118] nvme nvme0: I/O 161 QID 2 timeout, aborting [ 83.556178] nvme nvme0: I/O 162 QID 2 timeout, aborting [ 83.556187] nvme nvme0: I/O 163 QID 2 timeout, aborting [ 83.556196] nvme nvme0: I/O 164 QID 2 timeout, aborting [ 95.332114] nvme nvme0: I/O 25 QID 0 timeout, reset controller [ 95.332843] nvme nvme0: Abort status: 0x371 [ 95.332852] nvme nvme0: Abort status: 0x371 [ 95.332856] nvme nvme0: Abort status: 0x371 [ 95.332859] nvme nvme0: Abort status: 0x371 [ 95.332909] PM: dpm_run_callback(): pci_pm_resume+0x0/0xe0 returns -16 [ 95.332936] nvme 0000:03:00.0: PM: failed to resume async: error -16 ``` The Microsoft documentation for StorageD3Enable mentioned that Windows has a hardcoded allowlist for D3 support, which was used for these platforms. Introduce quirks to hardcode them for Linux as well. As this property is now "standardized", OEM systems using AMD Cezanne and newer APU's have adopted this property, and quirks like this should not be necessary. CC: Shyam-sundar S-k CC: Alexander Deucher CC: Prike Liang Link: https://docs.microsoft.com/en-us/windows-hardware/design/component-guidelines/power-management-for-storage-hardware-devices-intro Signed-off-by: Mario Limonciello Acked-by: Rafael J. Wysocki Tested-by: Julian Sikorski Signed-off-by: Christoph Hellwig commit fd82413988169aa648d1507298b925f73e5b0ce4 Author: Mario Limonciello Date: Wed Jun 9 13:40:17 2021 -0500 ACPI: Check StorageD3Enable _DSD property in ACPI code Although first implemented for NVME, this check may be usable by other drivers as well. Microsoft's specification explicitly mentions that is may be usable by SATA and AHCI devices. Google also indicates that they have used this with SDHCI in a downstream kernel tree that a user can plug a storage device into. Link: https://docs.microsoft.com/en-us/windows-hardware/design/component-guidelines/power-management-for-storage-hardware-devices-intro Suggested-by: Keith Busch CC: Shyam-sundar S-k CC: Alexander Deucher CC: Rafael J. Wysocki CC: Prike Liang Signed-off-by: Mario Limonciello Reviewed-by: Rafael J. Wysocki Signed-off-by: Christoph Hellwig --- drivers/acpi/device_pm.c | 32 ++++++ drivers/acpi/internal.h | 9 ++ drivers/acpi/x86/s2idle.c | 157 +++++++++++++++++-------- drivers/acpi/x86/utils.c | 25 ++++ drivers/nvme/host/pci.c | 28 +---- drivers/pci/quirks.c | 3 + drivers/platform/x86/amd-pmc.c | 204 ++++++++++++++++++++++++++++++--- include/linux/acpi.h | 5 + 8 files changed, 372 insertions(+), 91 deletions(-) diff --git a/drivers/acpi/device_pm.c b/drivers/acpi/device_pm.c index 9d2d3b9bb8b5..0cfdef2fc3ad 100644 --- a/drivers/acpi/device_pm.c +++ b/drivers/acpi/device_pm.c @@ -1338,4 +1338,36 @@ int acpi_dev_pm_attach(struct device *dev, bool power_on) return 1; } EXPORT_SYMBOL_GPL(acpi_dev_pm_attach); + +/** + * acpi_storage_d3 - Check if D3 should be used in the suspend path + * @dev: Device to check + * + * Return %true if the platform firmware wants @dev to be programmed + * into D3hot or D3cold (if supported) in the suspend path, or %false + * when there is no specific preference. On some platforms, if this + * hint is ignored, @dev may remain unresponsive after suspending the + * platform as a whole. + * + * Although the property has storage in the name it actually is + * applied to the PCIe slot and plugging in a non-storage device the + * same platform restrictions will likely apply. + */ +bool acpi_storage_d3(struct device *dev) +{ + struct acpi_device *adev = ACPI_COMPANION(dev); + u8 val; + + if (force_storage_d3()) + return true; + + if (!adev) + return false; + if (fwnode_property_read_u8(acpi_fwnode_handle(adev), "StorageD3Enable", + &val)) + return false; + return val == 1; +} +EXPORT_SYMBOL_GPL(acpi_storage_d3); + #endif /* CONFIG_PM */ diff --git a/drivers/acpi/internal.h b/drivers/acpi/internal.h index e21611c9a170..7ac01b03ba67 100644 --- a/drivers/acpi/internal.h +++ b/drivers/acpi/internal.h @@ -236,6 +236,15 @@ static inline int suspend_nvs_save(void) { return 0; } static inline void suspend_nvs_restore(void) {} #endif +#ifdef CONFIG_X86 +bool force_storage_d3(void); +#else +static inline bool force_storage_d3(void) +{ + return false; +} +#endif + /*-------------------------------------------------------------------------- Device properties -------------------------------------------------------------------------- */ diff --git a/drivers/acpi/x86/s2idle.c b/drivers/acpi/x86/s2idle.c index 2d7ddb8a8cb6..1c507804fb10 100644 --- a/drivers/acpi/x86/s2idle.c +++ b/drivers/acpi/x86/s2idle.c @@ -32,6 +32,9 @@ static const struct acpi_device_id lps0_device_ids[] = { {"", }, }; +/* Microsoft platform agnostic UUID */ +#define ACPI_LPS0_DSM_UUID_MICROSOFT "11e00d56-ce64-47ce-837b-1f898f9aa461" + #define ACPI_LPS0_DSM_UUID "c4eb40a0-6cd2-11e2-bcfd-0800200c9a66" #define ACPI_LPS0_GET_DEVICE_CONSTRAINTS 1 @@ -39,6 +42,8 @@ static const struct acpi_device_id lps0_device_ids[] = { #define ACPI_LPS0_SCREEN_ON 4 #define ACPI_LPS0_ENTRY 5 #define ACPI_LPS0_EXIT 6 +#define ACPI_LPS0_MS_ENTRY 7 +#define ACPI_LPS0_MS_EXIT 8 /* AMD */ #define ACPI_LPS0_DSM_UUID_AMD "e3f32452-febc-43ce-9039-932122d37721" @@ -49,7 +54,10 @@ static const struct acpi_device_id lps0_device_ids[] = { static acpi_handle lps0_device_handle; static guid_t lps0_dsm_guid; -static char lps0_dsm_func_mask; +static int lps0_dsm_func_mask; + +static guid_t lps0_dsm_guid_microsoft; +static int lps0_dsm_func_mask_microsoft; /* Device constraint entry structure */ struct lpi_device_info { @@ -70,15 +78,7 @@ struct lpi_constraints { int min_dstate; }; -/* AMD */ -/* Device constraint entry structure */ -struct lpi_device_info_amd { - int revision; - int count; - union acpi_object *package; -}; - -/* Constraint package structure */ +/* AMD Constraint package structure */ struct lpi_device_constraint_amd { char *name; int enabled; @@ -96,15 +96,15 @@ static void lpi_device_get_constraints_amd(void) int i, j, k; out_obj = acpi_evaluate_dsm_typed(lps0_device_handle, &lps0_dsm_guid, - 1, ACPI_LPS0_GET_DEVICE_CONSTRAINTS, + rev_id, ACPI_LPS0_GET_DEVICE_CONSTRAINTS, NULL, ACPI_TYPE_PACKAGE); - if (!out_obj) - return; - acpi_handle_debug(lps0_device_handle, "_DSM function 1 eval %s\n", out_obj ? "successful" : "failed"); + if (!out_obj) + return; + for (i = 0; i < out_obj->package.count; i++) { union acpi_object *package = &out_obj->package.elements[i]; @@ -317,14 +317,15 @@ static void lpi_check_constraints(void) } } -static void acpi_sleep_run_lps0_dsm(unsigned int func) +static void acpi_sleep_run_lps0_dsm(unsigned int func, unsigned int func_mask, guid_t dsm_guid) { union acpi_object *out_obj; - if (!(lps0_dsm_func_mask & (1 << func))) + if (!(func_mask & (1 << func))) return; - out_obj = acpi_evaluate_dsm(lps0_device_handle, &lps0_dsm_guid, rev_id, func, NULL); + out_obj = acpi_evaluate_dsm(lps0_device_handle, &dsm_guid, + rev_id, func, NULL); ACPI_FREE(out_obj); acpi_handle_debug(lps0_device_handle, "_DSM function %u evaluation %s\n", @@ -336,11 +337,33 @@ static bool acpi_s2idle_vendor_amd(void) return boot_cpu_data.x86_vendor == X86_VENDOR_AMD; } +static int validate_dsm(acpi_handle handle, const char *uuid, int rev, guid_t *dsm_guid) +{ + union acpi_object *obj; + int ret = -EINVAL; + + guid_parse(uuid, dsm_guid); + obj = acpi_evaluate_dsm(handle, dsm_guid, rev, 0, NULL); + + /* Check if the _DSM is present and as expected. */ + if (!obj || obj->type != ACPI_TYPE_BUFFER || obj->buffer.length == 0 || + obj->buffer.length > sizeof(u32)) { + acpi_handle_debug(handle, + "_DSM UUID %s rev %d function 0 evaluation failed\n", uuid, rev); + goto out; + } + + ret = *(int *)obj->buffer.pointer; + acpi_handle_debug(handle, "_DSM UUID %s rev %d function mask: 0x%x\n", uuid, rev, ret); + +out: + ACPI_FREE(obj); + return ret; +} + static int lps0_device_attach(struct acpi_device *adev, const struct acpi_device_id *not_used) { - union acpi_object *out_obj; - if (lps0_device_handle) return 0; @@ -348,28 +371,36 @@ static int lps0_device_attach(struct acpi_device *adev, return 0; if (acpi_s2idle_vendor_amd()) { - guid_parse(ACPI_LPS0_DSM_UUID_AMD, &lps0_dsm_guid); - out_obj = acpi_evaluate_dsm(adev->handle, &lps0_dsm_guid, 0, 0, NULL); + /* AMD0004, AMDI0005: + * - Should use rev_id 0x0 + * - function mask > 0x3: Should use AMD method, but has off by one bug + * - function mask = 0x3: Should use Microsoft method + * AMDI0006: + * - should use rev_id 0x0 + * - function mask = 0x3: Should use Microsoft method + */ + const char *hid = acpi_device_hid(adev); rev_id = 0; + lps0_dsm_func_mask = validate_dsm(adev->handle, + ACPI_LPS0_DSM_UUID_AMD, rev_id, &lps0_dsm_guid); + lps0_dsm_func_mask_microsoft = validate_dsm(adev->handle, + ACPI_LPS0_DSM_UUID_MICROSOFT, rev_id, + &lps0_dsm_guid_microsoft); + if (lps0_dsm_func_mask > 0x3 && (!strcmp(hid, "AMD0004") || + !strcmp(hid, "AMDI0005"))) { + lps0_dsm_func_mask = (lps0_dsm_func_mask << 1) | 0x1; + acpi_handle_debug(adev->handle, "_DSM UUID %s: Adjusted function mask: 0x%x\n", + ACPI_LPS0_DSM_UUID_AMD, lps0_dsm_func_mask); + } } else { - guid_parse(ACPI_LPS0_DSM_UUID, &lps0_dsm_guid); - out_obj = acpi_evaluate_dsm(adev->handle, &lps0_dsm_guid, 1, 0, NULL); rev_id = 1; + lps0_dsm_func_mask = validate_dsm(adev->handle, + ACPI_LPS0_DSM_UUID, rev_id, &lps0_dsm_guid); + lps0_dsm_func_mask_microsoft = -EINVAL; } - /* Check if the _DSM is present and as expected. */ - if (!out_obj || out_obj->type != ACPI_TYPE_BUFFER) { - acpi_handle_debug(adev->handle, - "_DSM function 0 evaluation failed\n"); - return 0; - } - - lps0_dsm_func_mask = *(char *)out_obj->buffer.pointer; - - ACPI_FREE(out_obj); - - acpi_handle_debug(adev->handle, "_DSM function mask: 0x%x\n", - lps0_dsm_func_mask); + if (lps0_dsm_func_mask < 0 && lps0_dsm_func_mask_microsoft < 0) + return 0; //function evaluation failed lps0_device_handle = adev->handle; @@ -386,11 +417,15 @@ static int lps0_device_attach(struct acpi_device *adev, mem_sleep_current = PM_SUSPEND_TO_IDLE; /* - * Some LPS0 systems, like ASUS Zenbook UX430UNR/i7-8550U, require the - * EC GPE to be enabled while suspended for certain wakeup devices to - * work, so mark it as wakeup-capable. + * Some Intel based LPS0 systems, like ASUS Zenbook UX430UNR/i7-8550U don't + * use intel-hid or intel-vbtn but require the EC GPE to be enabled while + * suspended for certain wakeup devices to work, so mark it as wakeup-capable. + * + * Only enable on !AMD as enabling this universally causes problems for a number + * of AMD based systems. */ - acpi_ec_mark_gpe_for_wake(); + if (!acpi_s2idle_vendor_amd()) + acpi_ec_mark_gpe_for_wake(); return 0; } @@ -408,12 +443,23 @@ int acpi_s2idle_prepare_late(void) if (pm_debug_messages_on) lpi_check_constraints(); - if (acpi_s2idle_vendor_amd()) { - acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_OFF_AMD); - acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY_AMD); + if (lps0_dsm_func_mask_microsoft > 0) { + acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_OFF, + lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft); + acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT, + lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft); + acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY, + lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft); + } else if (acpi_s2idle_vendor_amd()) { + acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_OFF_AMD, + lps0_dsm_func_mask, lps0_dsm_guid); + acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY_AMD, + lps0_dsm_func_mask, lps0_dsm_guid); } else { - acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_OFF); - acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY); + acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_OFF, + lps0_dsm_func_mask, lps0_dsm_guid); + acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY, + lps0_dsm_func_mask, lps0_dsm_guid); } return 0; @@ -424,12 +470,23 @@ void acpi_s2idle_restore_early(void) if (!lps0_device_handle || sleep_no_lps0) return; - if (acpi_s2idle_vendor_amd()) { - acpi_sleep_run_lps0_dsm(ACPI_LPS0_EXIT_AMD); - acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_ON_AMD); + if (lps0_dsm_func_mask_microsoft > 0) { + acpi_sleep_run_lps0_dsm(ACPI_LPS0_EXIT, + lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft); + acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_ENTRY, + lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft); + acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_ON, + lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft); + } else if (acpi_s2idle_vendor_amd()) { + acpi_sleep_run_lps0_dsm(ACPI_LPS0_EXIT_AMD, + lps0_dsm_func_mask, lps0_dsm_guid); + acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_ON_AMD, + lps0_dsm_func_mask, lps0_dsm_guid); } else { - acpi_sleep_run_lps0_dsm(ACPI_LPS0_EXIT); - acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_ON); + acpi_sleep_run_lps0_dsm(ACPI_LPS0_EXIT, + lps0_dsm_func_mask, lps0_dsm_guid); + acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_ON, + lps0_dsm_func_mask, lps0_dsm_guid); } } diff --git a/drivers/acpi/x86/utils.c b/drivers/acpi/x86/utils.c index bdc1ba00aee9..f22f23933063 100644 --- a/drivers/acpi/x86/utils.c +++ b/drivers/acpi/x86/utils.c @@ -135,3 +135,28 @@ bool acpi_device_always_present(struct acpi_device *adev) return ret; } + +/* + * AMD systems from Renoir and Lucienne *require* that the NVME controller + * is put into D3 over a Modern Standby / suspend-to-idle cycle. + * + * This is "typically" accomplished using the `StorageD3Enable` + * property in the _DSD that is checked via the `acpi_storage_d3` function + * but this property was introduced after many of these systems launched + * and most OEM systems don't have it in their BIOS. + * + * The Microsoft documentation for StorageD3Enable mentioned that Windows has + * a hardcoded allowlist for D3 support, which was used for these platforms. + * + * This allows quirking on Linux in a similar fashion. + */ +static const struct x86_cpu_id storage_d3_cpu_ids[] = { + X86_MATCH_VENDOR_FAM_MODEL(AMD, 23, 96, NULL), /* Renoir */ + X86_MATCH_VENDOR_FAM_MODEL(AMD, 23, 104, NULL), /* Lucienne */ + {} +}; + +bool force_storage_d3(void) +{ + return x86_match_cpu(storage_d3_cpu_ids); +} diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c index d963f25fc7ae..66455e2261d0 100644 --- a/drivers/nvme/host/pci.c +++ b/drivers/nvme/host/pci.c @@ -2880,32 +2880,6 @@ static unsigned long check_vendor_combination_bug(struct pci_dev *pdev) return 0; } -#ifdef CONFIG_ACPI -static bool nvme_acpi_storage_d3(struct pci_dev *dev) -{ - struct acpi_device *adev = ACPI_COMPANION(&dev->dev); - u8 val; - - /* - * Look for _DSD property specifying that the storage device on the port - * must use D3 to support deep platform power savings during - * suspend-to-idle. - */ - - if (!adev) - return false; - if (fwnode_property_read_u8(acpi_fwnode_handle(adev), "StorageD3Enable", - &val)) - return false; - return val == 1; -} -#else -static inline bool nvme_acpi_storage_d3(struct pci_dev *dev) -{ - return false; -} -#endif /* CONFIG_ACPI */ - static void nvme_async_probe(void *data, async_cookie_t cookie) { struct nvme_dev *dev = data; @@ -2955,7 +2929,7 @@ static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id) quirks |= check_vendor_combination_bug(pdev); - if (!noacpi && nvme_acpi_storage_d3(pdev)) { + if (!noacpi && acpi_storage_d3(&pdev->dev)) { /* * Some systems use a bios work around to ask for D3 on * platforms that support kernel managed suspend. diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c index 6d74386eadc2..446f8f7b0ee4 100644 --- a/drivers/pci/quirks.c +++ b/drivers/pci/quirks.c @@ -1900,6 +1900,9 @@ static void quirk_ryzen_xhci_d3hot(struct pci_dev *dev) } DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x15e0, quirk_ryzen_xhci_d3hot); DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x15e1, quirk_ryzen_xhci_d3hot); +/* Renoir XHCI requires longer delay when transitioning from D0 to + * D3hot */ +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x1639, quirk_ryzen_xhci_d3hot); #ifdef CONFIG_X86_IO_APIC static int dmi_disable_ioapicreroute(const struct dmi_system_id *d) diff --git a/drivers/platform/x86/amd-pmc.c b/drivers/platform/x86/amd-pmc.c index ca95c2a52e26..663a4ca0580d 100644 --- a/drivers/platform/x86/amd-pmc.c +++ b/drivers/platform/x86/amd-pmc.c @@ -46,26 +46,70 @@ #define AMD_PMC_RESULT_CMD_UNKNOWN 0xFE #define AMD_PMC_RESULT_FAILED 0xFF +/* FCH SSC Registers */ +#define FCH_S0I3_ENTRY_TIME_L_OFFSET 0x30 +#define FCH_S0I3_ENTRY_TIME_H_OFFSET 0x34 +#define FCH_S0I3_EXIT_TIME_L_OFFSET 0x38 +#define FCH_S0I3_EXIT_TIME_H_OFFSET 0x3C +#define FCH_SSC_MAPPING_SIZE 0x800 +#define FCH_BASE_PHY_ADDR_LOW 0xFED81100 +#define FCH_BASE_PHY_ADDR_HIGH 0x00000000 + +/* SMU Message Definations */ +#define SMU_MSG_GETSMUVERSION 0x02 +#define SMU_MSG_LOG_GETDRAM_ADDR_HI 0x04 +#define SMU_MSG_LOG_GETDRAM_ADDR_LO 0x05 +#define SMU_MSG_LOG_START 0x06 +#define SMU_MSG_LOG_RESET 0x07 +#define SMU_MSG_LOG_DUMP_DATA 0x08 +#define SMU_MSG_GET_SUP_CONSTRAINTS 0x09 /* List of supported CPU ids */ #define AMD_CPU_ID_RV 0x15D0 #define AMD_CPU_ID_RN 0x1630 #define AMD_CPU_ID_PCO AMD_CPU_ID_RV #define AMD_CPU_ID_CZN AMD_CPU_ID_RN +#define AMD_CPU_ID_YC 0x14B5 #define PMC_MSG_DELAY_MIN_US 100 #define RESPONSE_REGISTER_LOOP_MAX 200 +#define SOC_SUBSYSTEM_IP_MAX 12 +#define DELAY_MIN_US 2000 +#define DELAY_MAX_US 3000 enum amd_pmc_def { MSG_TEST = 0x01, MSG_OS_HINT_PCO, MSG_OS_HINT_RN, }; +struct amd_pmc_bit_map { + const char *name; + u32 bit_mask; +}; + +static const struct amd_pmc_bit_map soc15_ip_blk[] = { + {"DISPLAY", BIT(0)}, + {"CPU", BIT(1)}, + {"GFX", BIT(2)}, + {"VDD", BIT(3)}, + {"ACP", BIT(4)}, + {"VCN", BIT(5)}, + {"ISP", BIT(6)}, + {"NBIO", BIT(7)}, + {"DF", BIT(8)}, + {"USB0", BIT(9)}, + {"USB1", BIT(10)}, + {"LAPIC", BIT(11)}, + {} +}; + struct amd_pmc_dev { void __iomem *regbase; - void __iomem *smu_base; + void __iomem *smu_virt_addr; + void __iomem *fch_virt_addr; u32 base_addr; u32 cpu_id; + u32 active_ips; struct device *dev; struct mutex lock; /* generic mutex lock */ #if IS_ENABLED(CONFIG_DEBUG_FS) @@ -74,6 +118,7 @@ struct amd_pmc_dev { }; static struct amd_pmc_dev pmc; +static int amd_pmc_send_cmd(struct amd_pmc_dev *dev, bool set, u32 *data, u8 msg, bool ret); static inline u32 amd_pmc_reg_read(struct amd_pmc_dev *dev, int reg_offset) { @@ -85,13 +130,76 @@ static inline void amd_pmc_reg_write(struct amd_pmc_dev *dev, int reg_offset, u3 iowrite32(val, dev->regbase + reg_offset); } +struct smu_metrics { + u32 table_version; + u32 hint_count; + u32 s0i3_cyclecount; + u32 timein_s0i2; + u64 timeentering_s0i3_lastcapture; + u64 timeentering_s0i3_totaltime; + u64 timeto_resume_to_os_lastcapture; + u64 timeto_resume_to_os_totaltime; + u64 timein_s0i3_lastcapture; + u64 timein_s0i3_totaltime; + u64 timein_swdrips_lastcapture; + u64 timein_swdrips_totaltime; + u64 timecondition_notmet_lastcapture[SOC_SUBSYSTEM_IP_MAX]; + u64 timecondition_notmet_totaltime[SOC_SUBSYSTEM_IP_MAX]; +} __packed; + #ifdef CONFIG_DEBUG_FS static int smu_fw_info_show(struct seq_file *s, void *unused) { + struct amd_pmc_dev *dev = s->private; + struct smu_metrics table; + int idx; + + if (dev->cpu_id == AMD_CPU_ID_PCO) + return -EINVAL; + + memcpy_fromio(&table, dev->smu_virt_addr, sizeof(struct smu_metrics)); + + seq_puts(s, "\n=== SMU Statistics ===\n"); + seq_printf(s, "Table Version: %d\n", table.table_version); + seq_printf(s, "Hint Count: %d\n", table.hint_count); + seq_printf(s, "S0i3 Cycle Count: %d\n", table.s0i3_cyclecount); + seq_printf(s, "Time (in us) to S0i3: %lld\n", table.timeentering_s0i3_lastcapture); + seq_printf(s, "Time (in us) in S0i3: %lld\n", table.timein_s0i3_lastcapture); + + seq_puts(s, "\n=== Active time (in us) ===\n"); + for (idx = 0 ; idx < SOC_SUBSYSTEM_IP_MAX ; idx++) { + if (soc15_ip_blk[idx].bit_mask & dev->active_ips) + seq_printf(s, "%-8s : %lld\n", soc15_ip_blk[idx].name, + table.timecondition_notmet_lastcapture[idx]); + } + return 0; } DEFINE_SHOW_ATTRIBUTE(smu_fw_info); +static int s0ix_stats_show(struct seq_file *s, void *unused) +{ + struct amd_pmc_dev *dev = s->private; + u64 entry_time, exit_time, residency; + + entry_time = ioread32(dev->fch_virt_addr + FCH_S0I3_ENTRY_TIME_H_OFFSET); + entry_time = entry_time << 32 | ioread32(dev->fch_virt_addr + FCH_S0I3_ENTRY_TIME_L_OFFSET); + + exit_time = ioread32(dev->fch_virt_addr + FCH_S0I3_EXIT_TIME_H_OFFSET); + exit_time = exit_time << 32 | ioread32(dev->fch_virt_addr + FCH_S0I3_EXIT_TIME_L_OFFSET); + + /* It's in 48MHz. We need to convert it */ + residency = (exit_time - entry_time) / 48; + + seq_puts(s, "=== S0ix statistics ===\n"); + seq_printf(s, "S0ix Entry Time: %lld\n", entry_time); + seq_printf(s, "S0ix Exit Time: %lld\n", exit_time); + seq_printf(s, "Residency Time: %lld\n", residency); + + return 0; +} +DEFINE_SHOW_ATTRIBUTE(s0ix_stats); + static void amd_pmc_dbgfs_unregister(struct amd_pmc_dev *dev) { debugfs_remove_recursive(dev->dbgfs_dir); @@ -102,6 +210,8 @@ static void amd_pmc_dbgfs_register(struct amd_pmc_dev *dev) dev->dbgfs_dir = debugfs_create_dir("amd_pmc", NULL); debugfs_create_file("smu_fw_info", 0644, dev->dbgfs_dir, dev, &smu_fw_info_fops); + debugfs_create_file("s0ix_stats", 0644, dev->dbgfs_dir, dev, + &s0ix_stats_fops); } #else static inline void amd_pmc_dbgfs_register(struct amd_pmc_dev *dev) @@ -113,6 +223,32 @@ static inline void amd_pmc_dbgfs_unregister(struct amd_pmc_dev *dev) } #endif /* CONFIG_DEBUG_FS */ +static int amd_pmc_setup_smu_logging(struct amd_pmc_dev *dev) +{ + u32 phys_addr_low, phys_addr_hi; + u64 smu_phys_addr; + + if (dev->cpu_id == AMD_CPU_ID_PCO) + return -EINVAL; + + /* Get Active devices list from SMU */ + amd_pmc_send_cmd(dev, 0, &dev->active_ips, SMU_MSG_GET_SUP_CONSTRAINTS, 1); + + /* Get dram address */ + amd_pmc_send_cmd(dev, 0, &phys_addr_low, SMU_MSG_LOG_GETDRAM_ADDR_LO, 1); + amd_pmc_send_cmd(dev, 0, &phys_addr_hi, SMU_MSG_LOG_GETDRAM_ADDR_HI, 1); + smu_phys_addr = ((u64)phys_addr_hi << 32 | phys_addr_low); + + dev->smu_virt_addr = devm_ioremap(dev->dev, smu_phys_addr, sizeof(struct smu_metrics)); + if (!dev->smu_virt_addr) + return -ENOMEM; + + /* Start the logging */ + amd_pmc_send_cmd(dev, 0, NULL, SMU_MSG_LOG_START, 0); + + return 0; +} + static void amd_pmc_dump_registers(struct amd_pmc_dev *dev) { u32 value; @@ -127,10 +263,9 @@ static void amd_pmc_dump_registers(struct amd_pmc_dev *dev) dev_dbg(dev->dev, "AMD_PMC_REGISTER_MESSAGE:%x\n", value); } -static int amd_pmc_send_cmd(struct amd_pmc_dev *dev, bool set) +static int amd_pmc_send_cmd(struct amd_pmc_dev *dev, bool set, u32 *data, u8 msg, bool ret) { int rc; - u8 msg; u32 val; mutex_lock(&dev->lock); @@ -150,8 +285,8 @@ static int amd_pmc_send_cmd(struct amd_pmc_dev *dev, bool set) amd_pmc_reg_write(dev, AMD_PMC_REGISTER_ARGUMENT, set); /* Write message ID to message ID register */ - msg = (dev->cpu_id == AMD_CPU_ID_RN) ? MSG_OS_HINT_RN : MSG_OS_HINT_PCO; amd_pmc_reg_write(dev, AMD_PMC_REGISTER_MESSAGE, msg); + /* Wait until we get a valid response */ rc = readx_poll_timeout(ioread32, dev->regbase + AMD_PMC_REGISTER_RESPONSE, val, val != 0, PMC_MSG_DELAY_MIN_US, @@ -163,6 +298,11 @@ static int amd_pmc_send_cmd(struct amd_pmc_dev *dev, bool set) switch (val) { case AMD_PMC_RESULT_OK: + if (ret) { + /* PMFW may take longer time to return back the data */ + usleep_range(DELAY_MIN_US, 10 * DELAY_MAX_US); + *data = amd_pmc_reg_read(dev, AMD_PMC_REGISTER_ARGUMENT); + } break; case AMD_PMC_RESULT_CMD_REJECT_BUSY: dev_err(dev->dev, "SMU not ready. err: 0x%x\n", val); @@ -182,32 +322,54 @@ static int amd_pmc_send_cmd(struct amd_pmc_dev *dev, bool set) out_unlock: mutex_unlock(&dev->lock); + amd_pmc_dump_registers(dev); return rc; } +static int amd_pmc_get_os_hint(struct amd_pmc_dev *dev) +{ + switch (dev->cpu_id) { + case AMD_CPU_ID_PCO: + return MSG_OS_HINT_PCO; + case AMD_CPU_ID_RN: + case AMD_CPU_ID_YC: + return MSG_OS_HINT_RN; + } + return -EINVAL; +} + static int __maybe_unused amd_pmc_suspend(struct device *dev) { struct amd_pmc_dev *pdev = dev_get_drvdata(dev); int rc; + u8 msg; - rc = amd_pmc_send_cmd(pdev, 1); + /* Reset and Start SMU logging - to monitor the s0i3 stats */ + amd_pmc_send_cmd(pdev, 0, NULL, SMU_MSG_LOG_RESET, 0); + amd_pmc_send_cmd(pdev, 0, NULL, SMU_MSG_LOG_START, 0); + + msg = amd_pmc_get_os_hint(pdev); + rc = amd_pmc_send_cmd(pdev, 1, NULL, msg, 0); if (rc) dev_err(pdev->dev, "suspend failed\n"); - amd_pmc_dump_registers(pdev); - return 0; + return rc; } static int __maybe_unused amd_pmc_resume(struct device *dev) { struct amd_pmc_dev *pdev = dev_get_drvdata(dev); int rc; + u8 msg; - rc = amd_pmc_send_cmd(pdev, 0); + /* Let SMU know that we are looking for stats */ + amd_pmc_send_cmd(pdev, 0, NULL, SMU_MSG_LOG_DUMP_DATA, 0); + + msg = amd_pmc_get_os_hint(pdev); + rc = amd_pmc_send_cmd(pdev, 0, NULL, msg, 0); if (rc) dev_err(pdev->dev, "resume failed\n"); - amd_pmc_dump_registers(pdev); return 0; } @@ -216,6 +378,7 @@ static const struct dev_pm_ops amd_pmc_pm_ops = { }; static const struct pci_device_id pmc_pci_ids[] = { + { PCI_DEVICE(PCI_VENDOR_ID_AMD, AMD_CPU_ID_YC) }, { PCI_DEVICE(PCI_VENDOR_ID_AMD, AMD_CPU_ID_CZN) }, { PCI_DEVICE(PCI_VENDOR_ID_AMD, AMD_CPU_ID_RN) }, { PCI_DEVICE(PCI_VENDOR_ID_AMD, AMD_CPU_ID_PCO) }, @@ -227,9 +390,8 @@ static int amd_pmc_probe(struct platform_device *pdev) { struct amd_pmc_dev *dev = &pmc; struct pci_dev *rdev; - u32 base_addr_lo; - u32 base_addr_hi; - u64 base_addr; + u32 base_addr_lo, base_addr_hi; + u64 base_addr, fch_phys_addr; int err; u32 val; @@ -279,9 +441,21 @@ static int amd_pmc_probe(struct platform_device *pdev) if (!dev->regbase) return -ENOMEM; - amd_pmc_dump_registers(dev); - mutex_init(&dev->lock); + + /* Use FCH registers to get the S0ix stats */ + base_addr_lo = FCH_BASE_PHY_ADDR_LOW; + base_addr_hi = FCH_BASE_PHY_ADDR_HIGH; + fch_phys_addr = ((u64)base_addr_hi << 32 | base_addr_lo); + dev->fch_virt_addr = devm_ioremap(dev->dev, fch_phys_addr, FCH_SSC_MAPPING_SIZE); + if (!dev->fch_virt_addr) + return -ENOMEM; + + /* Use SMU to get the s0i3 debug stats */ + err = amd_pmc_setup_smu_logging(dev); + if (err) + dev_err(dev->dev, "SMU debugging info not supported on this platform\n"); + platform_set_drvdata(pdev, dev); amd_pmc_dbgfs_register(dev); return 0; @@ -298,6 +472,8 @@ static int amd_pmc_remove(struct platform_device *pdev) static const struct acpi_device_id amd_pmc_acpi_ids[] = { {"AMDI0005", 0}, + {"AMDI0006", 0}, + {"AMDI0007", 0}, {"AMD0004", 0}, { } }; diff --git a/include/linux/acpi.h b/include/linux/acpi.h index c60745f657e9..dd0dafd21e33 100644 --- a/include/linux/acpi.h +++ b/include/linux/acpi.h @@ -1004,6 +1004,7 @@ int acpi_dev_resume(struct device *dev); int acpi_subsys_runtime_suspend(struct device *dev); int acpi_subsys_runtime_resume(struct device *dev); int acpi_dev_pm_attach(struct device *dev, bool power_on); +bool acpi_storage_d3(struct device *dev); #else static inline int acpi_subsys_runtime_suspend(struct device *dev) { return 0; } static inline int acpi_subsys_runtime_resume(struct device *dev) { return 0; } @@ -1011,6 +1012,10 @@ static inline int acpi_dev_pm_attach(struct device *dev, bool power_on) { return 0; } +static inline bool acpi_storage_d3(struct device *dev) +{ + return false; +} #endif #if defined(CONFIG_ACPI) && defined(CONFIG_PM_SLEEP) -- 2.32.0