On kernel startup, the cfg80211 core hint code will try to load
regulatory.db unconditionally, which can increase kernel boot
time on some platforms where no regulatory.db is available.
Enclose all the related code for regulatory.db manipulation
with a macro CONFIG_CORE_REGDB so any platform with no
regulatory.db will not be affected by this.
NOTE: it is expected that any platform with no regulatory.db
should have self-managed regulatory database supported
by wlan module.
Change-Id: I9fd96012a4c47c26be8c223fc4ee545dac8f640b
Signed-off-by: wadesong <wadesong@codeaurora.org>
[ Upstream commit e0f5a8e74be88f2476e58b25d3b49a9521bdc4ec ]
commit e96c6b8f212a ("memblock: report failures when memblock_can_resize
is not set") introduced the usage of panic, which is not defined in
memblock test.
Let's define it directly in panic.h to fix it.
Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
CC: Song Shuai <songshuaishuai@tinylab.org>
CC: Mike Rapoport <rppt@kernel.org>
Link: https://lore.kernel.org/r/20240402132701.29744-3-richard.weiyang@gmail.com
Signed-off-by: Mike Rapoport (IBM) <rppt@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Apply the same logic that Qualcomm uses on their 5.15 kernel_platform
build scripts, as seen in commit e10f5dc.
As a sidenote, passing DTC_FLAGS externally is not an option since it
overrides any other flags set by the kernel makefile.
It causes this issue with the Android build system:
out/soong/.intermediates/vendor/lineage/build/soong/generated_kernel_includes/ge
n/usr/include/asm/sigcontext.h:53:2: error: unknown type name '__uint128_t'
__uint128_t vregs[32];
^
Change-Id: Ia4cf9c0e2b6a32550ec6c4398139838add0b24a1
* This prevents inclusion of drivers/staging/greybus/tools/Android.mk
which will conflict in case we have more than 1 kernel tree in AOSP
source dir.
Change-Id: I335bca7b6d6463b1ffc673ab5367603347516e13
When slave is in sleep q2spi host transfer will fail with GSI
start sequence failure. In this scenario two events reported from
GSI channel error and TX dma completion event 0x22. Since GSI
reports these events separately if start sequence fail is not
processed before channel error gpi_q2spi_terminate_all is called twice
and leading to GSI failures.
Ensure Channel errors in GSI recover path are handled sequentially
only after START sequence fail is processed.
Change-Id: Ie85528b6354241153330c403ba026c5006d5c78e
Signed-off-by: Chandana Kishori Chiluveru <quic_cchiluve@quicinc.com>
For non QMI packets having data length less than eight bytes,
skb_copy_bits fails to copy packet data to log buffer.
So, Add condition to check data length. If the data length is less
than eight bytes, send actual data length to skb_copy_bits.
Change-Id: I2181016f224952d214a8f39fb06b47ace01dc51a
Signed-off-by: Pranav Mahesh Phansalkar <quic_pphansal@quicinc.com>
Add support for usb related defconfigs on neo.
Change-Id: I3b6270e4f2f543af77af2d6357aa7011074023b3
Signed-off-by: Shubham Chouhan <quic_schouhan@quicinc.com>
Currently, functions gsi_inst_clean & gsi_free_inst utilises
gsi_opts without any check, however there is a possibility
that the opts structure could become NULL. In such case, due
to lack of if checks can result in NULL pointer dereference.
Change-Id: I548690e2eee377b5292f258972ae7e38417f3085
Signed-off-by: Prashanth K <quic_prashk@quicinc.com>
Signed-off-by: Udipto Goswami <quic_ugoswami@quicinc.com>
After moving the flush_work() call after the usb_phy_notify_disconnect()
to address USB type C compliance issues (LFPS generated during host
teardown):
commit fb3c680116db ("usb: dwc3: dwc3-msm-core: Notify PHY disconnect
before doing flush_work")
During USB device PIPO, the notify PHY disconnect call to the QMP PHY
will cause the PHY to be powered down. As part of the stop host mode
routine, the DWC3 core has to be placed back into device/peripheral
mode. Some parts of the device mode initialization sequence, such as
the core soft reset, requires that the PIPE clk (or controller source
clock) be active, otherwise the core soft reset will time out.
To mitigate the side effect, temporarily switch to the UTMI as the
controller clock source, so that the PIPE clock can be powered off
without any consequences. Once the move to DWC3 gadget mode is
complete, re-enable the PIPE clock as the controller source. (after
flush_work() is complete).
Change-Id: I59de803d737581c0037348498b8447f872adb62f
Signed-off-by: Wesley Cheng <quic_wcheng@quicinc.com>