commit 028e8ca7b2 upstream.
When the OTG USB port is used to power the SoC, configured as peripheral
and used in gadget mode, there is a disconnection about 6 seconds after the
gadget is configured and enumerated.
The problem was observed on a Radxa Rock Pi S board, which can only be
powered by the only USB-C connector. That connector is the only one usable
in gadget mode. This implies the USB cable is connected from before boot
and never disconnects while the kernel runs.
The problem happens because of the PHY driver code flow, summarized as:
* UDC start code (triggered via configfs at any time after boot)
-> phy_init
-> rockchip_usb2phy_init
-> schedule_delayed_work(otg_sm_work [A], 6 sec)
-> phy_power_on
-> rockchip_usb2phy_power_on
-> enable clock
-> rockchip_usb2phy_reset
* Now the gadget interface is up and running.
* 6 seconds later otg_sm_work starts [A]
-> rockchip_usb2phy_otg_sm_work():
if (B_IDLE state && VBUS present && ...):
schedule_delayed_work(&rport->chg_work [B], 0);
* immediately the chg_detect_work starts [B]
-> rockchip_chg_detect_work():
if chg_state is UNDEFINED:
if (!rport->suspended):
rockchip_usb2phy_power_off() <--- [X]
At [X], the PHY is powered off, causing a disconnection. This quickly
triggers a new connection and following re-enumeration, but any connection
that had been established during the 6 seconds is broken.
The code already checks for !rport->suspended (which, somewhat
counter-intuitively, means the PHY is powered on), so add a guard for VBUS
as well to avoid a disconnection when a cable is connected.
Fixes: 98898f3bc8 ("phy: rockchip-inno-usb2: support otg-port for rk3399")
Cc: stable@vger.kernel.org
Closes: https://lore.kernel.org/lkml/20250414185458.7767aabc@booty/
Signed-off-by: Louis Chauvet <louis.chauvet@bootlin.com>
Co-developed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Reviewed-by: Théo Lebrun <theo.lebrun@bootlin.com>
Link: https://patch.msgid.link/20251127-rk3308-fix-usb-gadget-phy-disconnect-v2-1-dac8a02cd2ca@bootlin.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit e07dea3de5 upstream.
The for_each_available_child_of_node() calls of_node_put() to
release child_np in each success loop. After breaking from the
loop with the child_np has been released, the code will jump to
the put_child label and will call the of_node_put() again if the
devm_request_threaded_irq() fails. These cause a double free bug.
Fix by returning directly to avoid the duplicate of_node_put().
Fixes: ed2b5a8e6b ("phy: phy-rockchip-inno-usb2: support muxed interrupts")
Cc: stable@vger.kernel.org
Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patch.msgid.link/20260109154626.2452034-1-vulab@iscas.ac.cn
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 7d8f725b79 upstream.
When the OTG USB port is used to power to SoC, configured as peripheral and
used in gadget mode, communication stops without notice about 6 seconds
after the gadget is configured and enumerated.
The problem was observed on a Radxa Rock Pi S board, which can only be
powered by the only USB-C connector. That connector is the only one usable
in gadget mode. This implies the USB cable is connected from before boot
and never disconnects while the kernel runs.
The related code flow in the PHY driver code can be summarized as:
* the first time chg_detect_work starts (6 seconds after gadget is
configured and enumerated)
-> rockchip_chg_detect_work():
if chg_state is UNDEFINED:
property_enable(base, &rphy->phy_cfg->chg_det.opmode, false); [Y]
* rockchip_chg_detect_work() changes state and re-triggers itself a few
times until it reaches the DETECTED state:
-> rockchip_chg_detect_work():
if chg_state is DETECTED:
property_enable(base, &rphy->phy_cfg->chg_det.opmode, true); [Z]
At [Y] all existing communications stop. E.g. using a CDC serial gadget,
the /dev/tty* devices are still present on both host and device, but no
data is transferred anymore. The later call with a 'true' argument at [Z]
does not restore it.
Due to the lack of documentation, what chg_det.opmode does exactly is not
clear, however by code inspection it seems reasonable that is disables
something needed to keep the communication working, and testing proves that
disabling these lines lets gadget mode keep working. So prevent changes to
chg_det.opmode when there is a cable connected (VBUS present).
Fixes: 98898f3bc8 ("phy: rockchip-inno-usb2: support otg-port for rk3399")
Cc: stable@vger.kernel.org
Closes: https://lore.kernel.org/lkml/20250414185458.7767aabc@booty/
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Reviewed-by: Théo Lebrun <theo.lebrun@bootlin.com>
Link: https://patch.msgid.link/20251127-rk3308-fix-usb-gadget-phy-disconnect-v2-2-dac8a02cd2ca@bootlin.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit fb21116099 ]
"family" is an enum, thus cast of pointer on 64-bit compile test with
clang W=1 causes:
phy-bcm-ns-usb3.c:206:17: error: cast to smaller integer type 'enum bcm_ns_family' from 'const void *' [-Werror,-Wvoid-pointer-to-enum-cast]
This was already fixed in commit bd6e74a2f0 ("phy: broadcom: ns-usb3:
fix Wvoid-pointer-to-enum-cast warning") but then got bad in commit
21bf6fc47a ("phy: Use device_get_match_data()").
Note that after various discussions the preferred cast is via "unsigned
long", not "uintptr_t".
Fixes: 21bf6fc47a ("phy: Use device_get_match_data()")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Link: https://patch.msgid.link/20251224115533.154162-2-krzysztof.kozlowski@oss.qualcomm.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 08aa19de72 ]
devm_pm_runtime_enable() can fail due to memory allocation. The current
code ignores its return value after calling pm_runtime_set_active(),
leaving the device in an inconsistent state if runtime PM initialization
fails.
Check the return value of devm_pm_runtime_enable() and return on
failure. Also move the declaration of 'ret' to the function scope
to support this check.
Fixes: ee8e41b504 ("phy: ti: phy-da8xx-usb: Add runtime PM support")
Suggested-by: Neil Armstrong <neil.armstrong@linaro.org>
Signed-off-by: Haotian Zhang <vulab@iscas.ac.cn>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patch.msgid.link/20251124105734.1027-1-vulab@iscas.ac.cn
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit cabd25b572 ]
The "index" variable is used as an index into the usbphyc->phys[] array
which has usbphyc->nphys elements. So if it is equal to usbphyc->nphys
then it is one element out of bounds. The "index" comes from the
device tree so it's data that we trust and it's unlikely to be wrong,
however it's obviously still worth fixing the bug. Change the > to >=.
Fixes: 94c358da3a ("phy: stm32: add support for STM32 USB PHY Controller (USBPHYC)")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Amelie Delaunay <amelie.delaunay@foss.st.com>
Link: https://patch.msgid.link/aTfHcMJK1wFVnvEe@stanley.mountain
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 1ca52c0983 ]
Enabling runtime PM before attaching the QPHY instance as driver data
can lead to a NULL pointer dereference in runtime PM callbacks that
expect valid driver data. There is a small window where the suspend
callback may run after PM runtime enabling and before runtime forbid.
This causes a sporadic crash during boot:
```
Unable to handle kernel NULL pointer dereference at virtual address 00000000000000a1
[...]
CPU: 0 UID: 0 PID: 11 Comm: kworker/0:1 Not tainted 6.16.7+ #116 PREEMPT
Workqueue: pm pm_runtime_work
pstate: 20000005 (nzCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : qusb2_phy_runtime_suspend+0x14/0x1e0 [phy_qcom_qusb2]
lr : pm_generic_runtime_suspend+0x2c/0x44
[...]
```
Attach the QPHY instance as driver data before enabling runtime PM to
prevent NULL pointer dereference in runtime PM callbacks.
Reorder pm_runtime_enable() and pm_runtime_forbid() to prevent a
short window where an unnecessary runtime suspend can occur.
Use the devres-managed version to ensure PM runtime is symmetrically
disabled during driver removal for proper cleanup.
Fixes: 891a96f65a ("phy: qcom-qusb2: Add support for runtime PM")
Signed-off-by: Loic Poulain <loic.poulain@oss.qualcomm.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Reviewed-by: Abel Vesa <abel.vesa@oss.qualcomm.com>
Link: https://patch.msgid.link/20251219085640.114473-1-loic.poulain@oss.qualcomm.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 72126e9623 ]
When making use of the clock provider functionality, the output clock
does normally match the TMDS character rate, which is what the PHY PLL
gets configured to.
However, this is only applicable for default color depth of 8 bpc. For
higher depths, the output clock is further divided by the hardware
according to the formula:
output_clock_rate = tmds_char_rate * 8 / bpc
Since the existence of the clock divider wasn't taken into account when
support for high bpc has been introduced, make the necessary adjustments
to report the correct clock rate.
Fixes: 9d0ec51d7c ("phy: rockchip: samsung-hdptx: Add high color depth management")
Reported-by: Andy Yan <andy.yan@rock-chips.com>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patch.msgid.link/20251028-phy-hdptx-fixes-v1-1-ecc642a59d94@collabora.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
The era of hand-rolled HIWORD_UPDATE macros is over, at least for those
drivers that use constant masks.
The Rockchip PCIe PHY driver, used on the RK3399, has its own definition
of HIWORD_UPDATE.
Remove it, and replace instances of it with hw_bitfield.h's
FIELD_PREP_WM16. To achieve this, some mask defines are reshuffled, as
FIELD_PREP_WM16 uses the mask as both the mask of bits to write and to
derive the shift amount from in order to shift the value.
In order to ensure that the mask is always a constant, the inst->index
shift is performed after the FIELD_PREP_WM16, as this is a runtime
value.
>From this, we gain compile-time error checking, and in my humble opinion
nicer code, as well as a single definition of this macro across the
entire codebase to aid in code comprehension.
Tested on a RK3399 ROCKPro64, where PCIe still works as expected when
accessing an NVMe drive.
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>
Do read-modify-write so that we re-use the characterized reset value as
specified in TRM [1] to program calibration wait time which defines number
of cycles to wait for after startup state machine is in bandgap enable
state.
This fixes PLL lock timeout error faced while using RPi DSI Panel on TI's
AM62L and J721E SoC since earlier calibration wait time was getting
overwritten to zero value thus failing the PLL to lockup and causing
timeout.
[1] AM62P TRM (Section 14.8.6.3.2.1.1 DPHY_TX_DPHYTX_CMN0_CMN_DIG_TBIT2):
Link: https://www.ti.com/lit/pdf/spruj83
Cc: stable@vger.kernel.org
Fixes: 7a343c8bf4 ("phy: Add Cadence D-PHY support")
Signed-off-by: Devarsh Thakkar <devarsht@ti.com>
Tested-by: Harikrishna Shenoy <h-shenoy@ti.com>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Link: https://lore.kernel.org/r/20250704125915.1224738-3-devarsht@ti.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
PLL lockup and O_CMN_READY assertion can only happen after common state
machine gets enabled by programming DPHY_CMN_SSM register, but driver was
polling them before the common state machine was enabled which is
incorrect. This is as per the DPHY initialization sequence as mentioned in
J721E TRM [1] at section "12.7.2.4.1.2.1 Start-up Sequence Timing Diagram".
It shows O_CMN_READY polling at the end after common configuration pin
setup where the common configuration pin setup step enables state machine
as referenced in "Table 12-1533. Common Configuration-Related Setup
mentions state machine"
To fix this :
- Add new function callbacks for polling on PLL lock and O_CMN_READY
assertion.
- As state machine and clocks get enabled in power_on callback only, move
the clock related programming part from configure callback to power_on
callback and poll for the PLL lockup and O_CMN_READY assertion after state
machine gets enabled.
- The configure callback only saves the PLL configuration received from the
client driver which will be applied later on in power_on callback.
- Add checks to ensure configure is called before power_on and state
machine is in disabled state before power_on callback is called.
- Disable state machine in power_off so that client driver can re-configure
the PLL by following up a power_off, configure, power_on sequence.
[1]: https://www.ti.com/lit/zip/spruil1
Cc: stable@vger.kernel.org
Fixes: 7a343c8bf4 ("phy: Add Cadence D-PHY support")
Signed-off-by: Devarsh Thakkar <devarsht@ti.com>
Tested-by: Harikrishna Shenoy <h-shenoy@ti.com>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Link: https://lore.kernel.org/r/20250704125915.1224738-2-devarsht@ti.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The era of hand-rolled HIWORD_UPDATE macros is over, at least for those
drivers that use constant masks.
Remove this driver's HIWORD_UPDATE macro, and replace all instances of
it with (hopefully) equivalent FIELD_PREP_WM16 instances. To do this, a
few of the defines are being adjusted, as FIELD_PREP_WM16 shifts up the
value for us. This gets rid of the icky update(mask, mask) shenanigans.
The benefit of using FIELD_PREP_WM16 is that it does more checking of
the input, hopefully catching errors. In practice, a shared definition
makes code more readable than several different flavours of the same
macro, and the shifted value helps as well.
I do not have the hardware that uses this particular driver, so it's
compile-tested only as far as my own testing goes.
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>
The era of hand-rolled HIWORD_UPDATE macros is over, at least for those
drivers that use constant masks.
phy-rockchip-samsung-dcphy is actually an exemplary example, where the
similarities to FIELD_PREP were spotted and the driver local macro has
the same semantics as the new FIELD_PREP_WM16 hw_bitfield.h macro.
Still, get rid of FIELD_PREP_HIWORD now that a shared implementation
exists, replacing the two instances of it with FIELD_PREP_WM16. This
gives us slightly better error checking; the value is now checked to fit
in 16 bits.
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>
The era of hand-rolled HIWORD_UPDATE macros is over, at least for those
drivers that use constant masks.
Replace the implementation of the rockchip eMMC PHY driver's
HIWORD_UPDATE macro with hw_bitfield.h's FIELD_PREP_WM16. This makes the
change more easily reviewable.
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Acked-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>
Commit 0cc22f5a86 ("phy: qcom: qmp-pcie: Add PHY register retention
support") added support for using the "no_csr" reset to skip configuration
of the PHY if the init sequence was already applied by the boot firmware.
The expectation is that the PHY is only turned on/off by using the "no_csr"
reset, instead of powering it down and re-programming it after a full
reset.
The boot firmware on X1E does not fully conform to this expectation: If the
PCIe3 link fails to come up (e.g. because no PCIe card is inserted), the
firmware powers down the PHY using the QPHY_PCS_POWER_DOWN_CONTROL
register. The QPHY_START_CTRL register is kept as-is, so the driver assumes
the PHY is already initialized and skips the configuration/power up
sequence. The PHY won't come up again without clearing the
QPHY_PCS_POWER_DOWN_CONTROL, so eventually initialization fails:
qcom-qmp-pcie-phy 1be0000.phy: phy initialization timed-out
phy phy-1be0000.phy.0: phy poweron failed --> -110
qcom-pcie 1bd0000.pcie: cannot initialize host
qcom-pcie 1bd0000.pcie: probe with driver qcom-pcie failed with error -110
This can be reliably reproduced on the X1E CRD, QCP and Devkit when no card
is inserted for PCIe3.
Fix this by checking the QPHY_PCS_POWER_DOWN_CONTROL register in addition
to QPHY_START_CTRL. If the PHY is powered down with the register, it
doesn't conform to the expectations for using the "no_csr" reset, so we
fully re-initialize with the normal reset sequence.
Also check the register more carefully to ensure all of the bits we expect
are actually set. A simple !!(readl()) is not enough, because the PHY might
be only partially set up with some of the expected bits set.
Cc: stable@vger.kernel.org
Fixes: 0cc22f5a86 ("phy: qcom: qmp-pcie: Add PHY register retention support")
Signed-off-by: Stephan Gerhold <stephan.gerhold@linaro.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250821-phy-qcom-qmp-pcie-nocsr-fix-v3-1-4898db0cc07c@linaro.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The existing ID detection logic returned false when both IDDIG and
VBUSVALID were set, which caused incorrect role determination in some
cases. The condition:
!(device && !vbus_valid)
did not properly reflect the intended relationship between IDDIG and
VBUSVALID signals.
Update the logic to:
return vbus_valid ? device : !device;
This ensures that when VBUS is valid, the role follows the IDDIG value,
and when VBUS is not valid, the role is inverted, matching the expected
OTG behavior.
Fixes: b725741f1c ("phy: renesas: rcar-gen3-usb2: Add support for RZ/T2H SoC")
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Link: https://lore.kernel.org/r/20250821155957.1088337-1-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Some SoCs are just validated with the TX delay enabled. With commit
ca13b249f2 ("net: ethernet: ti: am65-cpsw: fixup PHY mode for fixed
RGMII TX delay"), the network driver will patch the delay setting on the
fly assuming that the TX delay setting is fixed. In reality, the TX
delay is configurable and just skipped in the documentation. There are
bootloaders, which will disable the TX delay and this will lead to a
transmit path which doesn't add any delays at all.
Fix that by always writing the RGMII_ID setting and report an error for
unsupported RGMII delay modes.
This is safe to do and shouldn't break any boards in mainline because
the fixed delay is only introduced for gmii-sel compatibles which are
used together with the am65-cpsw-nuss driver and also contains the
commit above.
Fixes: ca13b249f2 ("net: ethernet: ti: am65-cpsw: fixup PHY mode for fixed RGMII TX delay")
Signed-off-by: Michael Walle <mwalle@kernel.org>
Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Link: https://lore.kernel.org/r/20250819065622.1019537-1-mwalle@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Add phy and pll regulator load voting support for all supported
platforms by introducing dedicated regulator bulk data arrays
with their load values.
This ensures stable operation and proper power management for these
platforms where regulators are shared between the QMP UFS PHY and
other IP blocks by setting appropriate regulator load currents during
PHY operations.
Signed-off-by: Nitin Rawat <quic_nitirawa@quicinc.com>
Acked-by: Manivannan Sadhasivam <mani@kernel.org>
Link: https://lore.kernel.org/r/20250830070353.2694-3-nitin.rawat@oss.qualcomm.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
On certain SoCs, power regulators are shared between the QMP UFS PHY
and other IP blocks. To ensure proper operation, the regulator
framework must be informed of the UFS PHY's load requirements.
This is essential because the regulator's operating mode—whether Low
Power or High Power—depends on the maximum expected load at any given
time, which the regulator driver needs to manage accordingly.
To support this, replace devm_regulator_bulk_get() with
devm_regulator_bulk_get_const() and inline the qmp_ufs_vreg_init()
function. additionally replace the array of regulator names with a
bulk regulator data structure, and utilize the init_load_uA field
provided by the regulator framework. This ensures that
regulator_set_load() is automatically invoked before the
first enable operation.
Reviewed-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Nitin Rawat <quic_nitirawa@quicinc.com>
Link: https://lore.kernel.org/r/20250830070353.2694-2-nitin.rawat@oss.qualcomm.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Add initial support for the Renesas RZ/T2H SoC to the R-Car Gen3 USB2 PHY
driver. The RZ/T2H SoC requires configuration of additional
hardware-specific bits for proper VBUS level control and OTG operation.
Introduce the `vblvl_ctrl` flag in the SoC-specific driver data to enable
handling of VBUS level selection logic using `VBCTRL.VBLVL` bits. This is
required for managing the VBUS status detection and drive logic based on
SoC-specific needs.
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://lore.kernel.org/r/20250808215209.3692744-5-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
