tegra_gpio_direction_input() and tegra_gpio_direction_output() already
program the GPIO controller direction registers directly. The additional
pinctrl_gpio_direction_input/output() calls do not add a Tegra pinctrl
operation, because the Tegra pinmux ops provide GPIO request/free
handling but no gpio_set_direction hook.
The extra call still enters the pinctrl core and takes pctldev->mutex.
Shared GPIO users can call the direction path while holding their
per-line spinlock, so this otherwise redundant pinctrl direction call can
sleep in an atomic context.
This was found by our static analysis tool and then confirmed by manual
review of tegra_gpio_probe(), the Tegra GPIO direction callbacks and the
Tegra pinctrl ops. The reviewed path has a default non-sleeping
struct gpio_chip while the direction callback still enters the pinctrl
mutex path.
A directed runtime validation kept the same non-sleeping chip registration
and drove:
gpio_shared_proxy_direction_output()
gpiod_direction_output_raw_commit()
tegra_gpio_direction_output()
pinctrl_gpio_direction_output()
Lockdep reported a sleep-in-atomic warning with the shared GPIO spinlock
held and pinctrl_get_device_gpio_range() plus tegra_gpio_direction_output()
on the stack.
Do not mark the whole chip as can_sleep to paper over this: can_sleep
describes whether get()/set() may sleep, and Tegra value access is MMIO.
Remove the redundant pinctrl direction calls and keep pinctrl involvement
in the existing request/free path.
Fixes: 11da905412 ("gpio: tegra: Fix offset of pinctrl calls")
Cc: stable@vger.kernel.org
Signed-off-by: Runyu Xiao <runyu.xiao@seu.edu.cn>
Link: https://patch.msgid.link/20260619152439.1239561-1-runyu.xiao@seu.edu.cn
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Add the device-tree binding documentation for the Rohm BM92TXX family
of USB Type-C and Power Delivery controllers. The device is controlled
over I2C and signals state changes through a dedicated interrupt line.
Document the rohm,* properties consumed by the driver: the optional
VCONN-enable GPIO, the DisplayPort alternate-mode toggles
(rohm,dp-disable, rohm,dp-alerts-enable, rohm,dp-signal-toggle-on-resume,
rohm,dp-lanes), the suspend/dock tunables (rohm,led-static-on-suspend,
rohm,dock-power-limit-disable) and the four
rohm,pd-{5,9,12,15}v-current-limit-ma per-PDO charging current limits.
A graph port is required to link the controller to a USB role switch
consumer.
Signed-off-by: Alexandre Hamamdjian <azkali.limited@gmail.com>
Add a driver for the Rohm Semiconductor BM92TXX family of USB Type-C
and Power Delivery controllers. The IC integrates an MCU that runs the
PD state machine; the host configures it and observes status over I2C
and reacts to a level-triggered ALERT interrupt.
The driver exposes the controller through extcon and a USB role switch,
manages the VBUS sink, optional VBUS source and battery-charger
regulators along with the VCONN-enable GPIO, and applies per-PDO
charging current limits (5 V, 9 V, 12 V, 15 V) sourced from device
tree. DisplayPort alternate-mode handling and dock LED behaviour are
configurable through rohm,* properties so the same driver can serve
boards that wire the part up differently. A debugfs interface under
bm92txx/ is provided for register dumps and low-level command access
when CONFIG_DEBUG_FS is enabled.
Signed-off-by: CTCaer <ctcaer@gmail.com>
Signed-off-by: Alexandre Hamamdjian <azkali.limited@gmail.com>
Tegra210b01 uses tegra186-emc driver as emc clock is controlled by BPMP.
However, this driver expects emc as subnode of mc. Adjust Tegra30 mc
probe routine to populate subnodes.
Sometimes CLKREQ# pin connection might be missing on a platform. Add DT
support to disable CLKREQ# control on PLLE in driver. Platform device tree
can use this DT property to disable CLKREQ# control on PLLE. In kernel-5.9,
upstream device tree property "supports-clkreq" is used in place of
kernel-4.14 custom device tree property "nvidia,disable-clock-request".
bug 200420606
Change-Id: I6ac44bd99d090e158fa10e5e9b62d63e52563c8c
Signed-off-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1786551
(cherry picked from commit 6870a94b48a4bbbe9df607802c7f7c1f61806b7b)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-5.9/+/2407880
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: Bitan Biswas <bbiswas@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
Programmed bits of register RP_VEND_XP_PAD_PWRDN are defined as follows,
L1(bit[0]):
This bit, when set, causes the analog pads to power down when we're in L1.
When clear, the LTSSM still enters L1 as required, but the analog pads
are left at full power.
DYNAMIC(bit[1]):
This bit, when set, causes unused analog pads to power down after
"Dynamic Link Width Re-negotiation" takes place. When clear, all analog
pads remain powered up, even those that are no longer being used after
"Dynamic Link Width Re-negotiation" down-sizes the link.
DISABLED(bit[2]):
This bit, when set, causes the analog pads to power down when we're in the
DISABLED_DOWN state. When clear, the LTSSM still enters DISBLED_DOWN as
required, but the analog pads are left at full power.
L1_CLKREQ(bit[15]):
This bit, when set, causes the analog pads to power down to
SLEEP_MODE_L1_REQ when clkreq signal is deasserted in L1.
(NV_PROJ__PCIE2_RP_VEND_XP_PAD_PWRDN_L1 needs to be set in order to power
down pad in L1.) Also, when clkreq is asserted in L1, ltssm will exit L1
immediately. When clear, the clkreq signal doesn't affect the power
management.
SLEEP_MODE_L1(bits[4:3]):
It defines the 2-bit sleep-mode coding in the pad when LTSSM is in the L1
state or DISABLED state.
SLEEP_MODE_DYNAMIC(bits[6:5]):
It defines the 2-bit sleep-mode coding in the pad when the lane shut down
due to dynamic link downsizing.
SLEEP_MODE_L1_CLKREQ(bits[17:16]):
It defines the 2-bit sleep-mode coding when clkreq is deasserted in L1.
The 2-bit sleep-mode encoding is defined as:
- L0 (0x0): Normal power-up mode
- L1 (0x1): L1 power-down mode; Tx common mode on, Rx e-idle detect on
- L1P (0x2): L1 power-down mode; Tx common mode off, Rx e-idle detect on
- L1PP (0x3): L1 power-down mode; Tx common mode off, Rx e-idle detect off
bug 200420606
Change-Id: I74aa9c1460d882084b1ecccb9c0168dd667c9d84
Signed-off-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1786547
(cherry picked from commit d7899c186f89ecf503812003b743adbae07c4cbe)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-5.9/+/2407873
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: Bitan Biswas <bbiswas@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
AFI module doesn't support WRAP transactions in Tegra. WRAP instructions
from CPU targeting PCIe BAR memory can cause data corruption. This can
happen if PCIe memory is marked as normal cacheable or GRE device.
Program mselect register to convert WRAP transactions to increment burst
for slave PCIe.
bug 200353330
bug 200420606
Change-Id: I867da85eb63332708bc294238e30ad99a71f345d
Signed-off-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1786561
(cherry picked from commit b345cfae7b7633a3a52e0f2a659baf44feee1466)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-5.9/+/2407869
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: Bitan Biswas <bbiswas@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
Few endpoints like Wi-Fi supports power on/off and to leverage that
root port must support hot-plug and hot-unplug. Tegra PCIe doesn't
support hot-plug and hot-unplug, however it supports endpoint power
on/off feature as follows,
- Power off sequence:
- Transition of PCIe link to L2
- Power off endpoint
- Leave root port in power up state with the link in L2
- Power on sequence:
- Power on endpoint
- Apply hot reset to get PCIe link up
PCIe client driver stops accessing PCIe endpoint config and BAR registers
after endpoint is powered off. However, software applications like x11
server or lspci can access endpoint config registers in which case
host controller raises "response decoding" errors. To avoid this scenario,
add PCIe link up check in config read and write callback functions before
accessing endpoint config registers.
Change-Id: I47653cb0b580c2b764cb0ad9c42deeb9c8823c58
Signed-off-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-4.14/+/2371131
(cherry picked from commit be664bfdd901138222a8ea8bb17160a2fa9de04a)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-5.9/+/2407866
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: Bitan Biswas <bbiswas@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
Tegra210b01 uses Tegra186 style emc management, as in it is
offloaded to BPMP. This driver already implements this, so no
reason to shoehorn in elsewhere.
Signed-off-by: Thomas Makin <halorocker89@gmail.com>
dsi-nx is a new blanket driver for panels on the Nintendo Switch.
The Switch boots Linux via the "hekate" custom bootloader, which
performs panel initialization and passes panel info to kernel.
This driver is based on panel-jdi-lpm062m326a by SwitchR.
Signed-off-by: Thomas Makin <halorocker89@gmail.com>
Add API to get min/max rail voltage configured from platform for
given rails. This will help to set the pad voltage based on platform
specific power tree.
bug 200083043
Change-Id: I054734ca2a25b5c830525e27ab22e43534f67351
Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Reviewed-on: http://git-master/r/712001
(cherry picked from commit c0d1de0da091132d037e1bb32f2ee09b5c00369b)
Reviewed-on: http://git-master/r/1119836
Signed-off-by: Thomas Makin <halorocker89@gmail.com>
Added PRE_SUBTREE_CHANGE and POST_SUBTREE_CHANGE that called before
and after rate change is propagated down the sub-tree rooted in clk.
This would allow children to be aware that next set rate operation
is triggered by downward rate propagation, rather than direct set rate
on a child clock.
Bug 200267979
Change-Id: I6f2acbeca7cead0ecd385dcb2e58ced32448e998
Signed-off-by: Alex Frid <afrid@nvidia.com>
Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
(cherry picked from commit 62f0d763fcb6c9d5bd9a373d22d802503ab0fdc5)
Reviewed-on: https://git-master.nvidia.com/r/1558143
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Timo Alho <talho@nvidia.com>
Signed-off-by: Thomas Makin <halorocker89@gmail.com>
Add initial device-tree support for NVIDIA Shield TV 2019 (a.k.a Sif)
based on the Tegra210B01 SoC with 2 GiB of LPDDR4 RAM.
Signed-off-by: Aaron Kling <webgeek1234@gmail.com>
tmakin
Laxman Dewangan
Maulik Shah
Runyu Xiao
Alexandre Hamamdjian
CTCaer
azkali
Alvin Park
Vidya Sagar
Manikanta Maddireddy
Alex Frid
Aaron Kling
Keir Fraser