kernel/linux.git/drivers/phy/phy-core.c, branch v6.18.21

phy: use per-PHY lockdep keys

2025-06-15T15:49:08+00:00

If the PHY driver uses another PHY internally (e.g. in case of eUSB2, repeaters are represented as PHYs), then it would trigger the following lockdep splat because all PHYs use a single static lockdep key and thus lockdep can not identify whether there is a dependency or not and reports a false positive. Make PHY subsystem use dynamic lockdep keys, assigning each driver a separate key. This way lockdep can correctly identify dependency graph between mutexes. ============================================ WARNING: possible recursive locking detected 6.15.0-rc7-next-20250522-12896-g3932f283970c #3455 Not tainted -------------------------------------------- kworker/u51:0/78 is trying to acquire lock: ffff0008116554f0 (&phy->mutex){+.+.}-{4:4}, at: phy_init+0x4c/0x12c but task is already holding lock: ffff000813c10cf0 (&phy->mutex){+.+.}-{4:4}, at: phy_init+0x4c/0x12c other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&phy->mutex); lock(&phy->mutex); *** DEADLOCK *** May be due to missing lock nesting notation 4 locks held by kworker/u51:0/78: #0: ffff000800010948 ((wq_completion)events_unbound){+.+.}-{0:0}, at: process_one_work+0x18c/0x5ec #1: ffff80008036bdb0 (deferred_probe_work){+.+.}-{0:0}, at: process_one_work+0x1b4/0x5ec #2: ffff0008094ac8f8 (&dev->mutex){....}-{4:4}, at: __device_attach+0x38/0x188 #3: ffff000813c10cf0 (&phy->mutex){+.+.}-{4:4}, at: phy_init+0x4c/0x12c stack backtrace: CPU: 0 UID: 0 PID: 78 Comm: kworker/u51:0 Not tainted 6.15.0-rc7-next-20250522-12896-g3932f283970c #3455 PREEMPT Hardware name: Qualcomm CRD, BIOS 6.0.240904.BOOT.MXF.2.4-00528.1-HAMOA-1 09/ 4/2024 Workqueue: events_unbound deferred_probe_work_func Call trace: show_stack+0x18/0x24 (C) dump_stack_lvl+0x90/0xd0 dump_stack+0x18/0x24 print_deadlock_bug+0x258/0x348 __lock_acquire+0x10fc/0x1f84 lock_acquire+0x1c8/0x338 __mutex_lock+0xb8/0x59c mutex_lock_nested+0x24/0x30 phy_init+0x4c/0x12c snps_eusb2_hsphy_init+0x54/0x1a0 phy_init+0xe0/0x12c dwc3_core_init+0x450/0x10b4 dwc3_core_probe+0xce4/0x15fc dwc3_probe+0x64/0xb0 platform_probe+0x68/0xc4 really_probe+0xbc/0x298 __driver_probe_device+0x78/0x12c driver_probe_device+0x3c/0x160 __device_attach_driver+0xb8/0x138 bus_for_each_drv+0x84/0xe0 __device_attach+0x9c/0x188 device_initial_probe+0x14/0x20 bus_probe_device+0xac/0xb0 deferred_probe_work_func+0x8c/0xc8 process_one_work+0x208/0x5ec worker_thread+0x1c0/0x368 kthread+0x14c/0x20c ret_from_fork+0x10/0x20 Fixes: 3584f6392f09 ("phy: qcom: phy-qcom-snps-eusb2: Add support for eUSB2 repeater") Fixes: e2463559ff1d ("phy: amlogic: Add Amlogic AXG PCIE PHY Driver") Reviewed-by: Neil Armstrong Reviewed-by: Abel Vesa Reported-by: Johan Hovold Link: https://lore.kernel.org/lkml/ZnpoAVGJMG4Zu-Jw@hovoldconsulting.com/ Reviewed-by: Johan Hovold Tested-by: Johan Hovold Signed-off-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20250605-phy-subinit-v3-1-1e1e849e10cd@oss.qualcomm.com Signed-off-by: Vinod Koul

phy: core: Remove unused phy_pm_runtime_(allow|forbid)

2025-03-10T07:28:07+00:00

phy_pm_runtime_allow() and phy_pm_runtime_forbid() were added in 2013 as part of commit ff764963479a ("drivers: phy: add generic PHY framework") but have remained unused. Remove them. Fix up the (English) docs - I've left the Chinese translation. Signed-off-by: Dr. David Alan Gilbert Link: https://lore.kernel.org/r/20250306015408.277729-1-linux@treblig.org Signed-off-by: Vinod Koul

phy: core: don't require set_mode() callback for phy_get_mode() to work

2025-02-13T17:55:17+00:00

As reported by Damon Ding, the phy_get_mode() call doesn't work as expected unless the PHY driver has a .set_mode() call. This prompts PHY drivers to have empty stubs for .set_mode() for the sake of being able to get the mode. Make .set_mode() callback truly optional and update PHY's mode even if it there is none. Cc: Damon Ding Link: https://lore.kernel.org/r/96f8310f-93f1-4bcb-8637-137e1159ff83@rock-chips.com Tested-by: Damon Ding Signed-off-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20250209-phy-fix-set-moe-v2-1-76e248503856@linaro.org Signed-off-by: Vinod Koul

Merge tag 'phy-for-6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/phy/linux-phy

2025-01-29T22:32:38+00:00

Pull phy updates from Vinod Koul: "Lots of Qualcomm and Rockchip device support. New Support: - Qualcomm SAR2130P qmp usb, SAR2130P qmp pcie, QCS615 qusb2 and PCIe, IPQ5424 qmp pcie, IPQ5424 QUSB2 and USB3 PHY - Rockchip rk3576 combo phy support Updates: - Drop Shengyang for JH7110 maintainer - Freescale hdmi register calculation optimization - Rockchip pcie phy mutex and regmap updates" * tag 'phy-for-6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/phy/linux-phy: (37 commits) dt-bindings: phy: qcom,qmp-pcie: document the SM8350 two lanes PCIe PHY phy: rockchip: phy-rockchip-typec: Fix Copyright description dt-bindings: phy: qcom,ipq8074-qmp-pcie: Document the IPQ5424 QMP PCIe PHYs phy: qcom-qusb2: Add support for QCS615 dt-bindings: usb: qcom,dwc3: Add QCS615 to USB DWC3 bindings phy: core: Simplify API of_phy_simple_xlate() implementation phy: sun4i-usb: Remove unused of_gpio.h phy: HiSilicon: Don't use "proxy" headers phy: samsung-ufs: switch back to syscon_regmap_lookup_by_phandle() phy: qualcomm: qmp-pcie: add support for SAR2130P phy: qualcomm: qmp-pcie: define several new registers phy: qualcomm: qmp-pcie: split PCS_LANE1 region phy: qualcomm: qmp-combo: add support for SAR2130P dt-bindings: phy: qcom,sc8280xp-qmp-pcie-phy: Add SAR2130P compatible dt-bindings: phy: qcom,sc8280xp-qmp-usb43dp: Add SAR2130P compatible phy: freescale: fsl-samsung-hdmi: Clean up fld_tg_code calculation phy: freescale: fsl-samsung-hdmi: Stop searching when exact match is found phy: freescale: fsl-samsung-hdmi: Expand Integer divider range phy: rockchip-naneng-combo: add rk3576 support dt-bindings: phy: rockchip: add rk3576 compatible ...

phy: core: Simplify API of_phy_simple_xlate() implementation

2024-12-24T14:25:59+00:00

Simplify of_phy_simple_xlate() implementation by API class_find_device_by_of_node(). Also correct comments to mark its parameter @dev as unused instead of @args in passing. Signed-off-by: Zijun Hu Cc: Simon Horman Link: https://lore.kernel.org/r/20241213-phy_core_fix-v6-6-40ae28f5015a@quicinc.com Signed-off-by: Vinod Koul

phy: core: Fix an OF node refcount leakage in of_phy_provider_lookup()

2024-12-24T14:25:37+00:00

For macro for_each_child_of_node(parent, child), refcount of @child has been increased before entering its loop body, so normally needs to call of_node_put(@child) before returning from the loop body to avoid refcount leakage. of_phy_provider_lookup() has such usage but does not call of_node_put() before returning, so cause leakage of the OF node refcount. Fix by simply calling of_node_put() before returning from the loop body. The APIs affected by this issue are shown below since they indirectly invoke problematic of_phy_provider_lookup(). phy_get() of_phy_get() devm_phy_get() devm_of_phy_get() devm_of_phy_get_by_index() Fixes: 2a4c37016ca9 ("phy: core: Fix of_phy_provider_lookup to return PHY provider for sub node") Cc: stable@vger.kernel.org Reviewed-by: Johan Hovold Signed-off-by: Zijun Hu Link: https://lore.kernel.org/r/20241213-phy_core_fix-v6-5-40ae28f5015a@quicinc.com Signed-off-by: Vinod Koul

phy: core: Fix an OF node refcount leakage in _of_phy_get()

2024-12-24T14:25:37+00:00

_of_phy_get() will directly return when suffers of_device_is_compatible() error, but it forgets to decrease refcount of OF node @args.np before error return, the refcount was increased by previous of_parse_phandle_with_args() so causes the OF node's refcount leakage. Fix by decreasing the refcount via of_node_put() before the error return. Fixes: b7563e2796f8 ("phy: work around 'phys' references to usb-nop-xceiv devices") Cc: stable@vger.kernel.org Reviewed-by: Johan Hovold Signed-off-by: Zijun Hu Link: https://lore.kernel.org/r/20241213-phy_core_fix-v6-4-40ae28f5015a@quicinc.com Signed-off-by: Vinod Koul

phy: core: Fix that API devm_phy_destroy() fails to destroy the phy

2024-12-24T14:25:37+00:00

For devm_phy_destroy(), its comment says it needs to invoke phy_destroy() to destroy the phy, but it will not actually invoke the function since devres_destroy() does not call devm_phy_consume(), and the missing phy_destroy() call will cause that the phy fails to be destroyed. Fortunately, the faulty API has not been used by current kernel tree. Fix by using devres_release() instead of devres_destroy() within the API. Fixes: ff764963479a ("drivers: phy: add generic PHY framework") Reviewed-by: Johan Hovold Signed-off-by: Zijun Hu Link: https://lore.kernel.org/r/20241213-phy_core_fix-v6-3-40ae28f5015a@quicinc.com Signed-off-by: Vinod Koul

phy: core: Fix that API devm_of_phy_provider_unregister() fails to unregister the phy provider

2024-12-24T14:25:37+00:00

For devm_of_phy_provider_unregister(), its comment says it needs to invoke of_phy_provider_unregister() to unregister the phy provider, but it will not actually invoke the function since devres_destroy() does not call devm_phy_provider_release(), and the missing of_phy_provider_unregister() call will cause: - The phy provider fails to be unregistered. - Leak both memory and the OF node refcount. Fortunately, the faulty API has not been used by current kernel tree. Fix by using devres_release() instead of devres_destroy() within the API. Fixes: ff764963479a ("drivers: phy: add generic PHY framework") Reviewed-by: Johan Hovold Signed-off-by: Zijun Hu Link: https://lore.kernel.org/stable/20241213-phy_core_fix-v6-2-40ae28f5015a%40quicinc.com Link: https://lore.kernel.org/r/20241213-phy_core_fix-v6-2-40ae28f5015a@quicinc.com Signed-off-by: Vinod Koul

phy: core: Fix that API devm_phy_put() fails to release the phy

2024-12-24T14:25:37+00:00

For devm_phy_put(), its comment says it needs to invoke phy_put() to release the phy, but it will not actually invoke the function since devres_destroy() does not call devm_phy_release(), and the missing phy_put() call will cause: - The phy fails to be released. - devm_phy_put() can not fully undo what API devm_phy_get() does. - Leak refcount of both the module and device for below typical usage: devm_phy_get(); // or its variant ... err = do_something(); if (err) goto err_out; ... err_out: devm_phy_put(); // leak refcount here The file(s) affected by this issue are shown below since they have such typical usage. drivers/pci/controller/cadence/pcie-cadence.c drivers/net/ethernet/ti/am65-cpsw-nuss.c Fix by using devres_release() instead of devres_destroy() within the API. Fixes: ff764963479a ("drivers: phy: add generic PHY framework") Cc: stable@vger.kernel.org Cc: Lorenzo Pieralisi Cc: Krzysztof Wilczyński Cc: Bjorn Helgaas Cc: David S. Miller Cc: Eric Dumazet Cc: Jakub Kicinski Cc: Paolo Abeni Reviewed-by: Johan Hovold Signed-off-by: Zijun Hu Link: https://lore.kernel.org/r/20241213-phy_core_fix-v6-1-40ae28f5015a@quicinc.com Signed-off-by: Vinod Koul