wireguard-linux/drivers/platform/x86/amd/pmc, branch stable

Merge tag 'platform-drivers-x86-v7.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86

2026-08-24T17:16:35Z

Pull x86 platform driver updates from Ilpo Järvinen "Highlights: Major refactoring effort: stop setting acpi_device_name/class() and pnp.device_class to facilitate their eventual removal Many rollback/remove path fixes (presumably mostly found by AI) Miscellaneous cleanups / refactoring / improvements amd/halo: - Add Halo RGB LED driver amd/hsmp: - Properly serialize probe, remove, and data paths - Add support for protocol v7 used by Family 1AH Model 80H - Fix error checking corner cases (largely from AI review) - Reject negative power cap amd/pmc: - Improve behavior on platforms that do not support STB - Add T14 Gen2 AMD (20XL) to s2idle quirk list amd/pmf: - Add ioctl interface to retrieve device metrics - Add support for new metrics tables used by Family 1AH Model 80H qcom-hamoa-ec (arm64): - Reject short responses asus-nb-wmi: - Support ProArt key on ASUS ProArt PX13 asus-armoury: - Gate PPT writes behind active fan curve - Add power limits for more models dell-wmi-base: - Fix handling of ultra performance key dell-wmi-sysman: - Don't hex dump attribute security buffer hp-bioscfg: - Various fixes - Improve reduced ACPI packages support (necessary for HP EliteBook 840 G2) lg-laptop: - Fix LED resource handling - Add support for events used in newer models - Fix keyboard backlight support on LG Gram 16T90SP hp-wmi: - Generalize thermal params to board params - Manage CPU and GPU PWM independently - Add GPU MUX switch support - Add Victus 15-fb0xxx support - Add OMEN MAX 16-ak0xxx, OMEN 16-n0xxx, OMEN 16-wd0xxx, OMEN 16-wf0xxx, and OMEN board ID 8D88 support - Add OMEN Transcend 16-u0xxx support huawei: - Add support for Fn-lock ACPI interface found on newer Huawei laptops such as MateBook 14 2024 ISST: - Improve input validation (many fixes) - Disallow SST-CP (core-power) feature if perf profile add fails lenovo/yb9-kbdock: - Add driver for Yoga Book 9 14IAH10 lenovo/ymc: - Extend hinge switch query to support Yoga 9 2-in-1 14IPH11 - Prevent loading on Yoga Book 9 14IAH10 to avoid duplicated input nodes msi-ec: - Add MSI Raider A18 HX A9WJG and MSI Katana GF76 11UEK support msi-wmi: - Add MSI Claw M-Center keys support oxpec: - Add support for OneXPlayer X2 Mini Pro redmi-wmi: - Report kbd backlight cycle, OEM preset power mode, and FnLock toggle events to userspace samsung-galaxybook: - Add Samsung Galaxy Book6 Pro support thinkpad_acpi: - Add USB-C Security support uniwill-laptop: - Add keyboard backlight, AC auto boot, and USB powershare support - Add MACHENIKE L16 Pro, AiStone X4SP4NAL, and Avell A60 MUV support - Make lightbar max brightness configurable and add support for LAPQC71A/B" * tag 'platform-drivers-x86-v7.3-1' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86: (152 commits) platform/x86: think-lmi: Fix current password length check platform/x86: redmi-wmi: report EC state change events MAINTAINERS: update Intel PMC Core maintainer contact platform/x86: oxpec: Add support for OneXPlayer X2 Mini Pro platform/x86: thinkpad_acpi: Fix fan speed reporting on Edge E330 platform/x86: msi-ec: Add MSI Katana GF76 11UEK EC firmware platform/x86: think-lmi: Fix certificate thumbprint sysfs output mlxbf-bootctl: fix the build error with FIELD_PREP() platform/x86: think-lmi: Free system certificate signatures platform/x86: ISST: Add a NULL check for sst_inst[] platform/x86: ISST: Return error during profile addition platform/x86: ISST: Just allow 2 bits for SST feature enable platform/x86: ISST: Use PP level enable mask platform/x86: ISST: Validate parameter for frequency and priority platform/x86: ISST: Validate parameter for core power state platform/x86: ISST: Validate max level for set feature platform/x86: ISST: Validate logical CPU id and clos id platform/x86: ISST: Validate level in perf mask ioctls platform/x86: ISST: Validate socket ID in clos_assoc ioctl platform/x86/amd/hsmp: Reject negative power cap writes in hwmon ...

platform/x86/amd/pmc: Do not fail probe when STB init fails

2026-07-27T11:42:25Z

STB (Spill to DRAM) is an optional debugging facility that is only enabled through the enable_stb module parameter. On some platforms the SMU refuses the S2D setup outright, and on long-running systems the large telemetry region can fail to ioremap. In either case amd_stb_s2d_init() returns an error and, because probe treated that as fatal, the entire PMC driver failed to load - silently disabling s0i3 support even though STB is only a debug aid. Downgrade the failure to a warning and continue probing so that s0i3 support via the LPS0 handler no longer depends on an optional debug feature. Since probe no longer aborts on this path, the LPS0 and debugfs unwinding added by the earlier fix in this series becomes unreachable and is removed. Reported-by: Francis De Brabandere Closes: https://bugzilla.kernel.org/show_bug.cgi?id=221759 Tested-by: Francis De Brabandere Signed-off-by: Mario Limonciello Link: https://patch.msgid.link/20260721181756.143084-7-mario.limonciello@amd.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

platform/x86/amd/pmc: Fix LPS0 and debugfs leaks when STB init fails

2026-07-27T11:42:24Z

amd_pmc_probe() registers the LPS0 s2idle handler with acpi_register_lps0_dev() and creates the driver's debugfs directory before calling amd_stb_s2d_init(), which is the last step in probe that can fail. When amd_stb_s2d_init() fails (for example the S2D telemetry region cannot be ioremapped on a long-running system, or the SMU rejects the S2D setup) the error path only calls pci_dev_put() and returns. This leaves amd_pmc_s2idle_dev_ops on the global lps0_s2idle_devops_head list and leaks the debugfs directory, while the devm-managed resources backing the handler are torn down. Reloading the module then walks the corrupted list in acpi_register_lps0_dev() and hits: list_add corruption. next->prev should be prev, but was NULL. kernel BUG at lib/list_debug.c:29! acpi_register_lps0_dev+0x44/0x80 amd_pmc_probe+0x224/0x380 [amd_pmc] platform_probe+0x67/0x90 Even without a reload, the stale registration means the next s2idle transition calls into torn-down driver state. Unwind the debugfs directory and the LPS0 registration on the amd_stb_s2d_init() error path. acpi_unregister_lps0_dev() is safe to call unconditionally here: it is guarded on the same conditions as acpi_register_lps0_dev(), which is exactly what amd_pmc_remove() already relies on. Reported-by: Francis De Brabandere Closes: https://bugzilla.kernel.org/show_bug.cgi?id=221759 Tested-by: Francis De Brabandere Fixes: 83ad6974dd3b ("platform/x86/amd/pmc: Move STB block into amd_pmc_s2d_init()") Cc: stable@vger.kernel.org Signed-off-by: Mario Limonciello Link: https://patch.msgid.link/20260721181756.143084-6-mario.limonciello@amd.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

platform/x86/amd/pmc: Only expose stb_read after telemetry buffer is mapped

2026-07-27T11:42:22Z

amd_stb_s2d_init() creates the v2 "stb_read" debugfs node before mapping the telemetry buffer into dev->stb_virt_addr, leaving a window during probe where a read faults on a NULL dev->stb_virt_addr in amd_stb_debugfs_open_v2()/amd_stb_handle_efr(). This becomes trivial to hit once a failed STB init no longer aborts probe (next patch), which leaves the node registered with a NULL buffer. Create it only after dev->stb_virt_addr is mapped. Signed-off-by: Mario Limonciello Link: https://patch.msgid.link/20260721181756.143084-5-mario.limonciello@amd.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

platform/x86/amd/pmc: Propagate SMU errors and validate S2D address

2026-07-27T11:42:20Z

amd_stb_s2d_init() discards the return value of several S2D SMU commands. When the SMU refuses a command (e.g. "SMU cmd failed. err: 0xff") the failure is only noticed indirectly - if at all - and reported as -EIO, masking the real error. More seriously, the S2D_PHYS_ADDR_LOW/HIGH return values are ignored, so on failure phys_addr_low/hi are left uninitialised and the assembled address is passed straight to devm_ioremap(). When the SMU leaves them at zero this maps physical address 0 and trips the ioremap-on-RAM warning: amd_pmc AMDI000B:00: SMU cmd failed. err: 0xff ioremap on RAM at 0x0000000000000000 - 0x0000000000ffffff WARNING: CPU: 13 PID: 4592 at arch/x86/mm/ioremap.c:... Check the return value of each SMU command and propagate it, and reject a zero physical address before calling devm_ioremap(). Reported-by: Francis De Brabandere Closes: https://bugzilla.kernel.org/show_bug.cgi?id=221759 Tested-by: Francis De Brabandere Fixes: 3d7d407dfb05 ("platform/x86: amd-pmc: Add support for AMD Spill to DRAM STB feature") Cc: stable@vger.kernel.org Signed-off-by: Mario Limonciello Link: https://patch.msgid.link/20260721181756.143084-4-mario.limonciello@amd.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

platform/x86/amd/pmc: Fix msg_port restoration in amd_stb_debugfs_open_v2()

2026-07-27T11:42:19Z

amd_stb_debugfs_open_v2() switches dev->msg_port to MSG_PORT_S2D to query S2D telemetry but only restores it to MSG_PORT_PMC on one path. The early return on the dump_custom_stb path (and the error/allocation returns) leave the port stuck on MSG_PORT_S2D, so subsequent SMU communication - including the s2idle prepare/restore handlers - is directed at the wrong mailbox. Consolidate the exit path through a single label so the message port is always restored, mirroring the fix in amd_stb_s2d_init(). Reported-by: sashiko.dev Link: https://sashiko.dev/#/patchset/20260717162023.956346-1-mario.limonciello%40amd.com Fixes: 2851f4f8ed4e ("platform/x86/amd/pmc: Define enum for S2D/PMC msg_port and add helper function") Cc: stable@vger.kernel.org Signed-off-by: Mario Limonciello Link: https://patch.msgid.link/20260721181756.143084-3-mario.limonciello@amd.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

platform/x86/amd/pmc: Restore msg_port on amd_stb_s2d_init() error paths

2026-07-27T11:42:16Z

dev->msg_port is switched to MSG_PORT_S2D before issuing the S2D SMU commands but is only restored to MSG_PORT_PMC on the success path. The early "return -EIO" and "return -ENOMEM" leave the port stuck on MSG_PORT_S2D, so all subsequent SMU communication - including the s2idle prepare/restore handlers - is directed at the wrong mailbox. Consolidate the exit path through a single label so the message port is always restored. Fixes: 3d7d407dfb05 ("platform/x86: amd-pmc: Add support for AMD Spill to DRAM STB feature") Cc: stable@vger.kernel.org Signed-off-by: Mario Limonciello Link: https://patch.msgid.link/20260721181756.143084-2-mario.limonciello@amd.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

platform/x86/amd/pmc: Add T14 Gen2 AMD (20XL) to s2idle quirk list

2026-07-21T15:28:19Z

The ThinkPad T14 Gen 2 AMD was sold under two machine types, 20XK and 20XL, but only 20XK has been in the s2idle quirk list since the quirk was introduced in commit 455cd867b85b ("platform/x86: thinkpad_acpi: Add a s2idle resume quirk for a number of laptops"). On 20XL machines the firmware SMI handler therefore still runs on the NVMe D3->D0 transition when exiting suspend-to-idle. With IOMMU translation enabled (the default), this intermittently stalls resume for ~10.25 seconds: seven devices across three root ports (nvme, both xhci_hcd functions, xhci-pci-renesas, snd_hda_intel x2, snd_rn_pci_acp3x) block in pci_pm_resume_noirq and are released simultaneously, consistent with all cores being held in SMM. Add the missing 20XL machine type, mirroring the existing 20XK entry. This is the same class of omission fixed by commit 9a469c6dfab3 ("platform/x86: thinkpad_acpi: Add missing T14s Gen1 type to s2idle quirk list"). Verified on a ThinkPad T14 Gen 2a (type 20XLS41C00, BIOS R1MET62W 1.32, kernel 6.12.95): with this entry the quirk message appears at probe and a 14-hour suspend resumes in ~1s with no noirq stalls, where previously even short suspends could hit the ~10s delay. A further two-day soak (short naps and two ~14-hour suspends) showed every resume clean. Link: https://bugzilla.kernel.org/show_bug.cgi?id=221755 Signed-off-by: Yap Weei Choong Tested-by: Yap Weei Choong Reviewed-by: Mario Limonciello (AMD) Link: https://patch.msgid.link/20260717060744.1252065-2-ywc8891@gmail.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

platform/x86: amd-pmc: Use correct IP block table for AMD 1Ah M80H SoC

2026-07-07T17:10:13Z

PMFW reports the S0i3 subsystem accounting per SoC, and the set of IP blocks and their bit ordering differ across SoC generations. Family 1Ah, Model 80h accounts for a distinct set of 19 IP blocks, which does not match the ordering in soc15_ip_blk[]. Commit 043af31c8d30 ("platform/x86/amd/pmc: Add PMC driver support for AMD 1Ah M80H SoC") wired amd_1ah_m80_cpu_info to soc15_ip_blk[], so M80H has been reporting incorrect S0i3 accounting via debugfs. Add soc15_ip_blk_v3[] with the correct ordering and point amd_1ah_m80_cpu_info at it. Fixes: 043af31c8d30 ("platform/x86/amd/pmc: Add PMC driver support for AMD 1Ah M80H SoC") Signed-off-by: Shyam Sundar S K Reviewed-by: Mario Limonciello (AMD) Link: https://patch.msgid.link/20260707122852.2066987-1-Shyam-sundar.S-k@amd.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

platform/x86/amd/pmc: Avoid logging "(null)" for DMI values

2026-07-03T13:24:28Z

dmi_get_system_info(...) can return NULL. Using that as %s arguments of dev_info() would log "(null)" (as part of a message like '... System Vendor: "(null)", Product Name: "(null)" ...'), which may be confusing for users. Use Elvis operator to print "(Unknown)" instead. Fixes: 428b9fd2dce5 ("platform/x86/amd/pmc: Add delay_suspend module parameter") Reported-by: kernel test robot Closes: https://lore.kernel.org/oe-kbuild-all/202606251540.Nr2BtaNu-lkp@intel.com/ Suggested-by: Ilpo Järvinen Signed-off-by: Daniel Gibson Link: https://patch.msgid.link/20260626220210.1761783-2-daniel@gibson.sh Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen