kernel/git/stable/linux.git/drivers/usb/cdns3, branch linux-7.2.y

usb: cdnsp: fix wakeup from S3 after controller context loss

2026-09-11T09:50:45+00:00

commit eae6460f617382044c5afe5ef202f4d8b2c099b5 upstream. CDNSP controller may lose its runtime register programming across S3 suspend/resume, depending on SoC power domain configuration. After resume the operational and interrupter registers may contain reset values, which prevents the gadget side from recovering correctly and breaks wakeup from S3. Fix this by detecting whether the controller lost its register context after resume and handling both cases: - If context was lost (CFG_3XPORT_U1_PIPE_CLK_GATE_EN set or power lost): reset the controller and reprogram the state required for normal operation, including the command ring, DCBAA pointer, doorbell base, event ring, ERST base/size and event ring dequeue pointer. - If context was retained: restart the controller directly without reprogramming registers. Issue a wakeup if the link was in U3 before suspend. Move the basic controller register programming out of the one-time memory initialization path and make it reusable from the resume path. Also separate ring allocation from ring initialization so that rings can be reinitialized without reallocating DMA memory. Always perform the full suspend sequence regardless of the current link state. Previously, if the device was already in U3, the suspend callback returned early without stopping the controller, which could lead to commands being issued on a disabled slot during resume. Fixes: 3d82904559f4 ("usb: cdnsp: cdns3 Add main part of Cadence USBSSP DRD Driver") Cc: stable Signed-off-by: Pawel Laszczak Acked-by: Peter Chen Link: https://patch.msgid.link/20260820-suspend_resume_fix-v3-1-5a713098b977@cadence.com Signed-off-by: Greg Kroah-Hartman Signed-off-by: Greg Kroah-Hartman

usb: cdnsp: fix incorrect endian conversions for APB timeout register

2026-08-03T15:23:47+00:00

readl() already returns a CPU-endian value. Passing its return value to le32_to_cpu() is therefore redundant and causes an incorrect double byte swap on big-endian systems. Similarly, writel() expects a CPU-endian value, so passing the result of cpu_to_le32() is incorrect. Remove the unnecessary conversions and operate on the MMIO register value as a CPU-endian u32. Fixes: 241e2ce88e5a ("usb: cdnsp: Fix issue with resuming from L1") Suggested-by: Arnd Bergmann Cc: stable Signed-off-by: Pawel Laszczak Acked-by: Arnd Bergmann Link: https://patch.msgid.link/20260720-endian-fix-v1-v1-1-b5681fa1ea9f@cadence.com Signed-off-by: Greg Kroah-Hartman

usb: cdnsp: fix stream context array leak in cdnsp_alloc_stream_info()

2026-06-25T14:12:15+00:00

cdnsp_alloc_stream_info() allocates stream_info->stream_ctx_array with cdnsp_alloc_stream_ctx(). If a later stream ring allocation or stream mapping update fails, the error path frees the allocated stream rings and stream_rings array, but leaves stream_ctx_array allocated. Free the stream context array before falling through to the stream_rings cleanup path. Fixes: 3d82904559f4 ("usb: cdnsp: cdns3 Add main part of Cadence USBSSP DRD Driver") Cc: stable Signed-off-by: Haoxiang Li Acked-by: Peter Chen Link: https://patch.msgid.link/20260622052627.696373-1-haoxiang_li2024@163.com Signed-off-by: Greg Kroah-Hartman

Merge tag 'v7.1-rc6' into usb-next

2026-06-01T15:39:51+00:00

We need the USB and Thunderbolt fixes in here as well. Signed-off-by: Greg Kroah-Hartman

usb: cdnsp: Add support for device-only configuration

2026-05-22T09:36:05+00:00

This patch introduces support for the Cadence USBSSP (cdnsp) controller in hardware configurations where the Dual-Role Device (DRD) register block is not implemented or is inaccessible. In such cases, the driver cannot rely on the DRD logic to manage roles and must operate exclusively in a fixed peripheral/host mode. The change in BAR indexing (from BAR 2 to BAR 1) is a direct consequence of the 32-bit addressing used in this specific DRD-disabled hardware layout, compared to the 64-bit addressing used in DRD-enabled configurations. Tested on a PCI platform with a hardware configuration that lacks DRD support. Platform-side changes are included to support the PCI glue layer's property injection to handle this specific layout. Acked-by: Peter Chen Acked-by: Bjorn Helgaas Signed-off-by: Pawel Laszczak Reported-by: kernel test robot Closes: https://lore.kernel.org/oe-kbuild-all/202605141023.18vWXyw3-lkp@intel.com/ Link: Closes: https://lore.kernel.org/oe-kbuild-all/202605141023.18vWXyw3-lkp@intel.com/ Link: https://patch.msgid.link/20260521-no_drd_config_v9-v9-2-2512cef10104@cadence.com Signed-off-by: Greg Kroah-Hartman

usb: cdns3: plat: fix leaked role switch on core role initialization failure

2026-05-22T09:34:44+00:00

Calling usb_role_switch_unregister if core role initialization failure. Fixes: e4d7362dc9cd ("usb: cdns3: Add USBSSP platform driver support") Reported-by: sashiko-bot Closes: https://lore.kernel.org/linux-devicetree/agKaEePSFknhDBg2@nchen-desktop/T/#m21e1d9c1574eb127ce03c0c2a1a49002ce435b52 Signed-off-by: Peter Chen Link: https://patch.msgid.link/20260514010114.2436781-2-peter.chen@cixtech.com Signed-off-by: Greg Kroah-Hartman

usb: cdns3: gadget: fix request skipping after clearing halt

2026-05-21T14:56:29+00:00

According to the cdns3 datasheet, the EPRST (Endpoint Reset) command causes the DMA engine to reposition its internal pointer to the next Transfer Descriptor (TD) if it was already processing one. This issue is consistently observed during the ADB identification process on macOS hosts, where the host issues a Clear_Halt. Although commit 4bf2dd65135a ("usb: cdns3: gadget: toggle cycle bit before reset endpoint") attempted to avoid DMA advance by toggling the cycle bit, trace logs show that on certain hosts like macOS, the DMA pointer (EP_TRADDR) still shifts after EPRST: cdns3_ctrl_req: Clear Endpoint Feature(Halt ep1out) cdns3_doorbell_epx: ep1out, ep_trbaddr f9c04030 <-- Should be f9c04000 cdns3_gadget_giveback: ep1out: req: ... length: 16384/16384 As shown above, the DMA pointer jumped to the next TD, causing the controller to skip the initial TRBs of the request. This leads to data misalignment and ADB protocol hangs on macOS. Fix this by manually restoring the EP_TRADDR register to the starting physical address of the current request after the EPRST operation is complete. Fixes: 7733f6c32e36 ("usb: cdns3: Add Cadence USB3 DRD Driver") Cc: stable Cc: Peter Chen Signed-off-by: Yongchao Wu Acked-by: Peter Chen Link: https://patch.msgid.link/20260513160012.2547894-1-yongchao.wu@autochips.com Signed-off-by: Greg Kroah-Hartman

usb: cdns3: plat: fix unbalanced pm_runtime_forbid() call permanently leaks the runtime PM usage counter across bind/unbind cycles

2026-05-21T14:56:10+00:00

Call pm_runtime_allow(dev) conditionally at cdns3_plat_remove. Fixes: f738957277ba ("usb: cdns3: Split core.c into cdns3-plat and core.c file") Cc: stable Reported-by: sashiko-bot Closes: https://lore.kernel.org/linux-devicetree/agKaEePSFknhDBg2@nchen-desktop/T/#m21e1d9c1574eb127ce03c0c2a1a49002ce435b52 Signed-off-by: Peter Chen Link: https://patch.msgid.link/20260513085310.2217547-3-peter.chen@cixtech.com Signed-off-by: Greg Kroah-Hartman

usb: cdns3: plat: fix leaked usb2_phy initialization on usb3_phy acquisition failure

2026-05-21T14:56:06+00:00

Move usb2_phy initialization after usb3_phy acquisition. Fixes: f738957277ba ("usb: cdns3: Split core.c into cdns3-plat and core.c file") Cc: stable Reported-by: sashiko-bot Closes: https://lore.kernel.org/linux-devicetree/agKaEePSFknhDBg2@nchen-desktop/T/#m21e1d9c1574eb127ce03c0c2a1a49002ce435b52 Signed-off-by: Peter Chen Link: https://patch.msgid.link/20260513085310.2217547-2-peter.chen@cixtech.com Signed-off-by: Greg Kroah-Hartman

usb: cdns3: Add USBSSP platform driver support

2026-04-27T11:14:18+00:00

Expose Cadence USBSSP through the same platform path as USBSS, trim Kconfig and Makefile: one core loadable object plus separate glue .ko files. Single cdns.ko bundles core, DRD, the generic "cdns,usb3" platform driver in cdns3-plat.c, optional host.o, and optional gadget objects. Use CONFIG_USB_CDNS3_GADGET as a bool to compile gadget support into that module. Remove duplicate MODULE_* declarations from cdns3-plat.c now that it links into the same module. Kconfig: the generic platform driver is selected via CONFIG_USB_CDNS3. Move CONFIG_USB_CDNSP_PCI beside CONFIG_USB_CDNS3_PCI_WRAP under "Platform glue driver support". SoC glue entries (TI, i.MX, StarFive) depend only on CONFIG_USB_CDNS3. Tighten CONFIG_USB_CDNS_SUPPORT dependencies so the umbrella follows host or gadget when either is built as a module. Match host and gadget bools to the cdns.ko tristate with USB=USB_CDNS3 and USB_GADGET=USB_CDNS3 instead of comparing against USB_CDNS_SUPPORT. Link host.o when CONFIG_USB_CDNS3_HOST is enabled and use that symbol in host-export.h, removing the redundant CONFIG_USB_CDNS_HOST indirection. Export cdns_core_init_role and reorganize the function cdns_init, and controller version could be gotten before the gadget init function is decided per controller. Keep host_init / gadget_init callbacks in struct cdns, so core.c does not need direct linkage to host or gadget objects. Refactor cdnsp-pci.c into a thin PCI-to-platform wrapper. drivers/usb/Makefile: descend into drivers/usb/cdns3/ only when CONFIG_USB_CDNS_SUPPORT is enabled. Assisted-by: Cursor:claude-4.6-opus Suggested-by: Arnd Bergmann Signed-off-by: Peter Chen Reviewed-by: Arnd Bergmann Link: https://patch.msgid.link/20260421023459.506145-3-peter.chen@cixtech.com Signed-off-by: Greg Kroah-Hartman