limits_stress driver doesn't need to be probed at boot time.
It is a limits stress module and needs to be probed
on need basis for thermal stress testing.
Change-Id: Ib913a587905904672924988bf695224e1c52c11b
Signed-off-by: Nitesh Kumar <quic_nitekuma@quicinc.com>
Pitti is not DMA-coherent, so the support of an uncached system heap is
required for clients. Enable the uncached system heap for Pitti via the
config for it.
Change-Id: I7cbe19506f034b896acd771875cbefdfa4406b08
Signed-off-by: Jaskaran Singh <quic_jasksing@quicinc.com>
Enable the secure buffer driver for Pitti to allow memory buffers to be
secured.
Change-Id: I50893eb871d054efad458c7147ffc77d4fdcec6e
Signed-off-by: Jaskaran Singh <quic_jasksing@quicinc.com>
Enable the DMA-BUF system secure heaps for Pitti.
Change-Id: I3786a5b470c9266a7a2cbd04e0468b28addbdf8e
Signed-off-by: Jaskaran Singh <quic_jasksing@quicinc.com>
Enable bus config memlat, bwmon, pmu-lib, dcvs, cpucp and
cpucp log flag in pitti.
Change-Id: I42695692e17f036cc54b0597c8138dff018a0ef4
Signed-off-by: Vandana Gupta <quic_vandgupt@quicinc.com>
Enabling few additional kernel configs which are required
to meet the dependencies for enabling wlan on the target.
- Enable configs to compile wlan for Niobe target.
- Add required modules to compile wlan for Niobe target.
Change-Id: Id081ee21cbe1cb7751b127d3d0b735c1fc0f6084
CRs-Fixed: 3686653
Signed-off-by: Shailendra Singh <quic_shasing@quicinc.com>
This is a snapshot of ufs phy driver from msm-5.15 branch at
commit 69b9df0c3412 ("ufs: phy: Add phy wrapper file for sa8295").
Added phy-qcom-ufs-qmp-v4-lahaina.ko for makena to autogvm bazel list.
Change-Id: I59e0fe96930f0e1ac8bde67d03d44af4f7de6362
Signed-off-by: Ram Kumar Dwivedi <quic_rdwivedi@quicinc.com>
Signed-off-by: Abdul Salam <quic_asalam@quicinc.com>
Signed-off-by: Jayasri Bhattacharyya <quic_jbhattac@quicinc.com>
Added keepalive flag for PCI0 BCM to keep the pcie-anoc
related clocks ON for Hypervisor.
Change-Id: I9d77123f2c9abf94d11f96bd97eb6de83ef21d65
Signed-off-by: Raviteja Laggyshetty <quic_rlaggysh@quicinc.com>
Add wait mechanism for WCD usbss device to completely
come out of resume before perform register writes.
Change-Id: I0dff6f56ed7b685252e43e5db1e596ceb53a3a45
Signed-off-by: Phani Kumar Uppalapati <quic_phaniu@quicinc.com>
Signed-off-by: Prasad Kumpatla <quic_pkumpatl@quicinc.com>
Add cold temperature interrupt handling support in tsens driver.
Change-Id: I9a6bb653882bc59368396181cd9fb019f90ca90b
Signed-off-by: Nitesh Kumar <quic_nitekuma@quicinc.com>
On Hibernate entry, PMIC H/W context is lost and S/W context
is preserved. To make sure that H/W and S/W context are in sync
free the IRQs in freeze callback and in restore callback re-initialize
the H/W and re-register the IRQs to make sure that S/W and H/W context
are in sync. Similar change is required for Deep-Sleep use-case as well.
Add changes to support Deep-Sleep and Hibernate feature for pwrkey
driver.
Change-Id: I12abfb6a864f16e23676c2cd822ef804f6f0b952
Signed-off-by: Umang Chheda <quic_uchheda@quicinc.com>
Signed-off-by: Darshankumar Jagdishchandra Thakkar <quic_djagdish@quicinc.com>
Keep the eusb2 resources such as clocks and LDOs if EUD is in
debug mode. Avoid resetting and powering down the repeater,
since it can affect the EUD interrupts resulting in failure of
EUD spoof attach/detach.
Change-Id: Ib92d628127d07b57e404d87d3185863dd789f605
Signed-off-by: Prashanth K <quic_prashk@quicinc.com>
dwc->xhci in dwc3_msm_suspend
Consider a case if the config for xhci is not enabled,
and we are trying for role switch, in host mode will be
updated accordingly, but the xhci will be NULL. Fix this
by adding explicit NULL checks.
Change-Id: Ib7dea431b972f3f5ee179016a721591298340cb5
Signed-off-by: Manish Nagar <quic_mnagar@quicinc.com>
On SM6375/SM8150, it is observed that if the USB2PHYCFG SUSPHY bit is
set during clearing of RUN_STOP in device mode, then sometimes, the
DSTS Ctrl Dev halt bit it not being set soon and the timeout waiting
for it to be set is delaying run stop routing for more than 4 seconds
and during PIPO tests, this can result in causing next plugin
enumeration failure as the next plug-in's run stop setting doesn't go
through.
As per register dump observations, it is seen that UTMI clock is
suspended in fail case and as per SVE, it must be kept on during
this routine. So use the existing kretprobes to ensure UTMI clock
is kept on during run stop routine.
Ideally the wrapper takes care of setting and clearing this bit and
ensuring that it is cleared during runtime of core/gadget. So clear
this bit during run_stop(0) and leave it to msm_suspend for setting
it again.
Change-Id: I8878956c5dae82cc3566c8933407c5e10c0248dc
Signed-off-by: Krishna Kurapati <quic_kriskura@quicinc.com>
While exiting peripheral mode, it is better to explicitly check if we
are suspended or not instead of checking for return value. Whatever
be the reason, put sync on child is ideally supposed to suspend the
child and ensure we are good to suspend the wrapper. There is a chance
that put sync might return 0 although the disconnect interrupt didn't
come and still the child is not suspended. So check explicitly whether
the child is suspended or not before we try/skip force suspending it.
Change-Id: Ic9dcabc3f405f4bc4093bbb403410eedc0564361
Signed-off-by: Wesley Cheng <quic_wcheng@quicinc.com>
Signed-off-by: Krishna Kurapati <quic_kriskura@quicinc.com>
Add qfprom module in modules.list.msm.pitti so that the
module is loaded automatically from the first stage RAM
disk image during Linux kernel boot up.
It is needed for boot_config reg reading support from
nvmem cell to check if the boot device is ufs or eMMC.
Change-Id: I29caa56c5fb773077b220fae2cb024d5e4422f52
Signed-off-by: Manish Pandey <quic_mapa@quicinc.com>
Commit 21d22fcad404 ("soc: qcom: eud: Do USB spoof disconnect and
connect with enable/disable EUD") introduced a software WA for a
HW issue. While enabling EUD from sysfs, EUD driver was notifying
USB by sending extcon in the order 1->0->1. Thus disconnecting &
connecting USB, which in turn toggles UTMI_OTG_VBUS_VALID bit.
But this causes the USB2phy to suspend, eventually turning off
clocks/LDOs and resetting repeater (for eusb2). And during this
time EUD driver enables EUD interrupts by setting EUD_INTR_EN.
But this register write operation doesn't go through since the
phy/repeater resources are not active, and resulting in EUD
interrupts being inactive. During spoof attach/detach these
interrupts doesn't get fired causing spoof connect/disconnect
to fail.
----------------------------------------------------------
EUD driver dwc3-msm-core eusb2phy
-----------------------------------------------------------
[1]
extcon(0) stop_peripheral
clear vbus_valid
LPM sequence set_suspend
turn off clks/LDOs
repeater_reset
repeater_pwrdwn
-----------------------------------------------------------
[2]
enable_irq
enable_eud
-----------------------------------------------------------
[3]
extcon(1) start_peripheral
set vbus_valid phy_init
turn_on_clk
repeater_reset_init
-----------------------------------------------------------
Here step[2] fails since the eusb resources are not available.
Fix this by toggling the vbus_bit directly from vbus_notifier,
and returning afterwards. This would comply with the SW WA and
wouldn't cause the phy to suspend. We check for EXTCON_JIG for
this purpose, this will not be set if EUD is enabled from sysfs.
Change-Id: I9aa8d2e0d620f40dd9b3f477d258734d73997aa7
Signed-off-by: Prashanth K <quic_prashk@quicinc.com>
