Brian Norris [Fri, 22 Apr 2022 17:08:53 +0000 (10:08 -0700)]
mmc: core: Set HS clock speed before sending HS CMD13
Way back in commit 4f25580fb84d ("mmc: core: changes frequency to
hs_max_dtr when selecting hs400es"), Rockchip engineers noticed that
some eMMC don't respond to SEND_STATUS commands very reliably if they're
still running at a low initial frequency. As mentioned in that commit,
JESD84-B51 P49 suggests a sequence in which the host:
1. sets HS_TIMING
2. bumps the clock ("<= 52 MHz")
3. sends further commands
It doesn't exactly require that we don't use a lower-than-52MHz
frequency, but in practice, these eMMC don't like it.
The aforementioned commit tried to get that right for HS400ES, although
it's unclear whether this ever truly worked as committed into mainline,
as other changes/refactoring adjusted the sequence in conflicting ways:
08573eaf1a70 ("mmc: mmc: do not use CMD13 to get status after speed mode
switch")
53e60650f74e ("mmc: core: Allow CMD13 polling when switching to HS mode
for mmc")
In any case, today we do step 3 before step 2. Let's fix that, and also
apply the same logic to HS200/400, where this eMMC has problems too.
Resolves errors like this seen when booting some RK3399 Gru/Scarlet
systems:
Ealier versions of this patch bumped to 200MHz/HS200 speeds too early,
which caused issues on, e.g., qcom-msm8974-fairphone-fp2. (Thanks for
the report Luca!) After a second look, it appears that aligns with
JESD84 / page 45 / table 28, so we need to keep to lower (HS / 52 MHz)
rates first.
Fixes: 08573eaf1a70 ("mmc: mmc: do not use CMD13 to get status after speed mode switch") Fixes: 53e60650f74e ("mmc: core: Allow CMD13 polling when switching to HS mode for mmc") Fixes: 4f25580fb84d ("mmc: core: changes frequency to hs_max_dtr when selecting hs400es") Cc: Shawn Lin <shawn.lin@rock-chips.com> Link: https://lore.kernel.org/linux-mmc/11962455.O9o76ZdvQC@g550jk/ Reported-by: Luca Weiss <luca@z3ntu.xyz> Signed-off-by: Brian Norris <briannorris@chromium.org> Tested-by: Luca Weiss <luca@z3ntu.xyz> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20220422100824.v4.1.I484f4ee35609f78b932bd50feed639c29e64997e@changeid Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The ON Semiconductor FSA4480 is a USB Type-C port multimedia switch with
support for analog audio headsets. It allows sharing a common USB Type-C
port to pass USB2.0 signal, analog audio, sideband use wires and analog
microphone signal.
Due to lacking upstream audio support for testing, the audio muxing is
left untouched, but implementation of muxing the SBU lines is provided
as a pair of Type-C mux and switch devices. This provides the necessary
support for enabling the DisplayPort altmode on devices with this
circuit.
The Fairchild/ON Semiconductor FSA4480 Analog Audio switch is used in
USB Type-C configurations for muxing analog audio onto the USB
connector, and as such used to control the SBU signals for altmodes such
as DisplayPort.
In the Qualcomm platforms the USB/DP PHY handles muxing and orientation
switching of the SuperSpeed lines, but the SBU lines needs to be
connected and switched by external (to the SoC) hardware.
It's therefor necessary to be able to have the TypeC controller operate
multiple TypeC muxes and switches. Use the newly introduced indirection
object to handle this, to avoid having to taint the TypeC controllers
with knowledge about the downstream hardware configuration.
The max number of devs per indirection is set to 3, which account for
being able to mux/switch the USB HS, SS and SBU lines, as per defined
defined in the usb-c-connector binding. This number could be grown if
need arrises at a later point in time.
device property: Use multi-connection matchers for single case
The newly introduced helpers for searching for matches in the case of
multiple connections can be resused by the single-connection case, so do
this to save some duplication.
device property: Add helper to match multiple connections
In some cases multiple connections with the same connection id
needs to be resolved from a fwnode graph.
One such example is when separate hardware is used for performing muxing
and/or orientation switching of the SuperSpeed and SBU lines in a USB
Type-C connector. In this case the connector needs to belong to a graph
with multiple matching remote endpoints, and the Type-C controller needs
to be able to resolve them both.
When role switch module probe late than ucsi module,
fwnode_usb_role_switch_get() will return -EPROBE_DEFER,
it is better to restart ucsi init work to find
it again every 100ms, total wait time is 10 second.
It also means change ucsi init work to delayed_work.
usb: typec: ucsi: add a common function ucsi_unregister_connectors()
In error path of ucsi_init(), it will unregister all valid ucsi connectors,
and similar operation also happen in ucsi_unregister(),
add a common function ucsi_unregister_connectors() for two places,
inside this function, if con->wq is NULL, it will break the loop,
if other kind of error happen after con->wq allocated,
ucsi/typec related API is safe to unregister.
Also in ucsi_init(), it allocate number of (ucsi->cap.num_connectors + 1)
connectors, there is one extra as the ending,
ucsi_unregister_connectors() is safe to unregister all ucsi connectors
according ucsi->cap.num_connectors,
remove the extra one connector to save memory.
Document RZ/G2ULSoC bindings. USBHS block is identical to one
found on RZ/A2 SoC. No driver changes are required as generic compatible
string "renesas,rza2-usbhs" will be used as a fallback.
While at it, make the resets property required on RZ/{G2L,G2UL,V2L} SoC.
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com> Link: https://lore.kernel.org/r/20220425140215.186797-1-biju.das.jz@bp.renesas.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Ilpo Järvinen [Mon, 25 Apr 2022 14:34:00 +0000 (17:34 +0300)]
serial: 8250: Handle UART without interrupt on TEMT
Add UART_CAP_NOTEMT for UARTs that lack interrupt on TEMT but want to
use em485. Em485 framework needs to ensure not only FIFO is empty but
also that tx shift register is empty.
This approach uses Uwe Kleine-König's suggestion on simply
using/incrementing stop_tx timer rather than adding another timer. When
UART_CAP_NOTEMT is set and THRE is present w/o TEMT, stop tx timer is
reused to wait for the emptying of the shift register.
This change does not add the UART_CAP_NOTEMT define as it already exist
but is currently no-op. See 7a107b2c6b81 (Revert "serial: 8250: Handle
UART without interrupt on TEMT using em485") for further details.
Vicente Bergas reported that RTS is deasserted roughly one bit too
early losing stop bit tx. To address this problem, stop_delay now
accounts for one extra bit using rough formula /7 (assumes worst-case
of 2+5 bits). I suspect this glitch had to do with when THRE is getting
asserted. If FIFO is emptied already during the tx of the stop bit,
perhaps it leads to HW asserting THRE early for the normal frame time
formula to work accurately.
Ilpo Järvinen [Mon, 25 Apr 2022 14:33:59 +0000 (17:33 +0300)]
serial: 8250: use THRE & __stop_tx also with DMA
8250 DMA tx complete path lacks calls to normal 8250 stop handling. It
does not use THRE to detect true completion of the tx and also doesn't
call __stop_tx. This leads to problems with em485 that needs to handle
RTS timing.
Instead of handling tx stop internally within 8250 dma code, enable
THRE when tx'able data runs out and tweak serial8250_handle_irq to call
only __stop_tx when uart is using DMA.
It also seems bit early to call serial8250_rpm_put_tx from there while
tx is still underway(?).
Ilpo Järvinen [Mon, 25 Apr 2022 14:33:58 +0000 (17:33 +0300)]
serial: Store character timing information to uart_port
Struct uart_port currently stores FIFO timeout. Having character timing
information readily available is useful. Even serial core itself
determines char_time from port->timeout using inverse calculation.
Store frame_time directly into uart_port. Character time is stored in
nanoseconds to have reasonable precision with high rates. To avoid
overflow, 64-bit math is necessary.
It might be possible to determine timeout from frame_time by
multiplying it with fifosize as needed but only part of the users seem
to be protected by a lock. Thus, this patch does not pursue storing
only frame_time in uart_port.
Phil Edworthy [Fri, 22 Apr 2022 18:06:15 +0000 (20:06 +0200)]
serial: 8250: dw: Improve RZN1 support
Renesas RZ/N1 SoC features a slightly modified DW UART.
On this SoC, the CPR register value is known but not synthetized in
hardware. We hence need to provide a CPR value in the platform
data. This version of the controller also relies on acting as flow
controller when using DMA, so we need to provide the
"is dma flow controller" quirk.
Phil Edworthy [Fri, 22 Apr 2022 18:06:14 +0000 (20:06 +0200)]
serial: 8250: dw: Add support for DMA flow controlling devices
DW based controllers like the one on Renesas RZ/N1 must be programmed as
flow controllers when using DMA.
* Table 11.45 of the system manual, "Flow Control Combinations", states
that using UART with DMA requires setting the DMA in the peripheral
flow controller mode regardless of the direction.
* Chapter 11.6.1.3 of the system manual, "Basic Interface Definitions",
explains that the burst size in the above case must be configured in
the peripheral's register DEST/SRC_BURST_SIZE.
Experiments shown that upon Rx timeout, the DMA transaction needed to be
manually cleared as well.
serial: 8250: dw: Move the IO accessors to 8250_dwlib.h
These accessors should be used instead of the regular readl/writel()
helpers. In order to use them also from 8250_dw.c in this directory,
move the helpers to 8250_dwlib.h
There is no functional change.
There is no need for declaring `struct uart_port` or even UPIO_MEM32BE
which both are already included in the 8250_dwlib.h header by 8250.h.
Suggested-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Link: https://lore.kernel.org/r/20220422180615.9098-8-miquel.raynal@bootlin.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
serial: 8250: dw: Introduce an rx_timeout variable in the IRQ path
In a next change we are going to need the same Rx timeout condition as
we already have in the IRQ handling code. Let's just create a boolean to
clarify what this operation does before reusing it.
serial: 8250: dma: Allow driver operations before starting DMA transfers
One situation where this could be used is when configuring the UART
controller to be the DMA flow controller. This is a typical case where
the driver might need to program a few more registers before starting a
DMA transfer. Provide the necessary infrastructure to support this
case.
Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Link: https://lore.kernel.org/r/20220422180615.9098-6-miquel.raynal@bootlin.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This offset is a good candidate to pdata's because it changes depending
on the vendor implementation. Let's move the usr_reg entry from regular
to pdata. This way we can drop initializing it at run time.
Let's also use a define for it instead of defining only the default
value.
Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Link: https://lore.kernel.org/r/20220422180615.9098-4-miquel.raynal@bootlin.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
virtio_net: fix wrong buf address calculation when using xdp
We received a report[1] of kernel crashes when Cilium is used in XDP
mode with virtio_net after updating to newer kernels. After
investigating the reason it turned out that when using mergeable bufs
with an XDP program which adjusts xdp.data or xdp.data_meta page_to_buf()
calculates the build_skb address wrong because the offset can become less
than the headroom so it gets the address of the previous page (-X bytes
depending on how lower offset is):
page_to_skb: page addr ffff9eb2923e2000 buf ffff9eb2923e1ffc offset 252 headroom 256
This is a pr_err() I added in the beginning of page_to_skb which clearly
shows offset that is less than headroom by adding 4 bytes of metadata
via an xdp prog. The calculations done are:
receive_mergeable():
headroom = VIRTIO_XDP_HEADROOM; // VIRTIO_XDP_HEADROOM == 256 bytes
offset = xdp.data - page_address(xdp_page) -
vi->hdr_len - metasize;
page_to_skb():
p = page_address(page) + offset;
...
buf = p - headroom;
Now buf goes -4 bytes from the page's starting address as can be seen
above which is set as skb->head and skb->data by build_skb later. Depending
on what's done with the skb (when it's freed most often) we get all kinds
of corruptions and BUG_ON() triggers in mm[2]. We have to recalculate
the new headroom after the xdp program has run, similar to how offset
and len are recalculated. Headroom is directly related to
data_hard_start, data and data_meta, so we use them to get the new size.
The result is correct (similar pr_err() in page_to_skb, one case of
xdp_page and one case of virtnet buf):
a) Case with 4 bytes of metadata
[ 115.949641] page_to_skb: page addr ffff8b4dcfad2000 offset 252 headroom 252
[ 121.084105] page_to_skb: page addr ffff8b4dcf018000 offset 20732 headroom 252
b) Case of pushing data +32 bytes
[ 153.181401] page_to_skb: page addr ffff8b4dd0c4d000 offset 288 headroom 288
[ 158.480421] page_to_skb: page addr ffff8b4dd00b0000 offset 24864 headroom 288
c) Case of pushing data -33 bytes
[ 835.906830] page_to_skb: page addr ffff8b4dd3270000 offset 223 headroom 223
[ 840.839910] page_to_skb: page addr ffff8b4dcdd68000 offset 12511 headroom 223
Offset and headroom are equal because offset points to the start of
reserved bytes for the virtio_net header which are at buf start +
headroom, while data points at buf start + vnet hdr size + headroom so
when data or data_meta are adjusted by the xdp prog both the headroom size
and the offset change equally. We can use data_hard_start to compute the
new headroom after the xdp prog (linearized / page start case, the
virtnet buf case is similar just with bigger base offset):
xdp.data_hard_start = page_address + vnet_hdr
xdp.data = page_address + vnet_hdr + headroom
new headroom after xdp prog = xdp.data - xdp.data_hard_start - metasize
compile: clang -O2 -g -Wall -target bpf -c xdp_pass.c -o xdp_pass.o
load on virtio_net: ip link set enp1s0 xdpdrv obj xdp_pass.o sec xdp_pass
CC: stable@vger.kernel.org CC: Jason Wang <jasowang@redhat.com> CC: Xuan Zhuo <xuanzhuo@linux.alibaba.com> CC: Daniel Borkmann <daniel@iogearbox.net> CC: "Michael S. Tsirkin" <mst@redhat.com> CC: virtualization@lists.linux-foundation.org Fixes: 8fb7da9e9907 ("virtio_net: get build_skb() buf by data ptr") Signed-off-by: Nikolay Aleksandrov <razor@blackwall.org> Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Michael S. Tsirkin <mst@redhat.com> Acked-by: Jason Wang <jasowang@redhat.com> Link: https://lore.kernel.org/r/20220425103703.3067292-1-razor@blackwall.org Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Changbin Du [Mon, 17 Jan 2022 15:43:00 +0000 (23:43 +0800)]
sysrq: do not omit current cpu when showing backtrace of all active CPUs
The backtrace of current CPU also should be printed as it is active. This
change add stack trace for current CPU and print a hint for idle CPU for
the generic workqueue based printing. (x86 already does this)
bus: mhi: ep: Add uevent support for module autoloading
Add uevent support to MHI endpoint bus so that the client drivers can be
autoloaded by udev when the MHI endpoint devices gets created. The client
drivers are expected to provide MODULE_DEVICE_TABLE with the MHI id_table
struct so that the alias can be exported.
The MHI endpoint reused the mhi_device_id structure of the MHI bus.
bus: mhi: ep: Add support for suspending and resuming channels
Add support for suspending and resuming the channels in MHI endpoint stack.
The channels will be moved to the suspended state during M3 state
transition and will be resumed during M0 transition.
bus: mhi: ep: Add support for queueing SKBs to the host
Add support for queueing SKBs to the host over the transfer ring of the
relevant channel. The mhi_ep_queue_skb() API will be used by the client
networking drivers to queue the SKBs to the host over MHI bus.
The host will add ring elements to the transfer ring periodically for
the device and the device will write SKBs to the ring elements. If a
single SKB doesn't fit in a ring element (TRE), it will be placed in
multiple ring elements and the overflow event will be sent for all ring
elements except the last one. For the last ring element, the EOT event
will be sent indicating the packet boundary.
bus: mhi: ep: Add support for processing channel rings
Add support for processing the channel rings from host. For the channel
ring associated with DL channel, the xfer callback will simply invoked.
For the case of UL channel, the ring elements will be read in a buffer
till the write pointer and later passed to the client driver using the
xfer callback.
The client drivers should provide the callbacks for both UL and DL
channels during registration.
bus: mhi: ep: Add support for reading from the host
Data transfer between host and the ep device happens over the transfer
ring associated with each bi-directional channel pair. Host defines the
transfer ring by allocating memory for it. The read and write pointer
addresses of the transfer ring are stored in the channel context.
Once host places the elements in the transfer ring, it increments the
write pointer and rings the channel doorbell. Device will receive the
doorbell interrupt and will process the transfer ring elements.
This commit adds support for reading the transfer ring elements from
the transfer ring till write pointer, incrementing the read pointer and
finally sending the completion event to the host through corresponding
event ring.
bus: mhi: ep: Add support for processing command rings
Add support for processing the command rings. Command ring is used by the
host to issue channel specific commands to the ep device. Following
commands are supported:
1. Start channel
2. Stop channel
3. Reset channel
Once the device receives the command doorbell interrupt from host, it
executes the command and generates a command completion event to the
host in the primary event ring.
bus: mhi: ep: Add support for handling SYS_ERR condition
Add support for handling SYS_ERR (System Error) condition in the MHI
endpoint stack. The SYS_ERR flag will be asserted by the endpoint device
when it detects an internal error. The host will then issue reset and
reinitializes MHI to recover from the error state.
Add support for handling MHI_RESET in MHI endpoint stack. MHI_RESET will
be issued by the host during shutdown and during error scenario so that
it can recover the endpoint device without restarting the whole device.
MHI_RESET handling involves resetting the internal MHI registers, data
structures, state machines, resetting all channels/rings and setting
MHICTRL.RESET bit to 0. Additionally the device will also move to READY
state if the reset was due to SYS_ERR.
bus: mhi: ep: Add support for powering down the MHI endpoint stack
Add support for MHI endpoint power_down that includes stopping all
available channels, destroying the channels, resetting the event and
transfer rings and freeing the host cache.
The stack will be powered down whenever the physical bus link goes down.
bus: mhi: ep: Add support for powering up the MHI endpoint stack
Add support for MHI endpoint power_up that includes initializing the MMIO
and rings, caching the host MHI registers, and setting the MHI state to M0.
After registering the MHI EP controller, the stack has to be powered up
for usage.
bus: mhi: ep: Add support for processing MHI endpoint interrupts
Add support for processing MHI endpoint interrupts such as control
interrupt, command interrupt and channel interrupt from the host.
The interrupts will be generated in the endpoint device whenever host
writes to the corresponding doorbell registers. The doorbell logic
is handled inside the hardware internally.
bus: mhi: ep: Add support for sending events to the host
Add support for sending the events to the host over MHI bus from the
endpoint. Following events are supported:
1. Transfer completion event
2. Command completion event
3. State change event
4. Execution Environment (EE) change event
An event is sent whenever an operation has been completed in the MHI EP
device. Event is sent using the MHI event ring and additionally the host
is notified using an IRQ if required.
Add support for managing the MHI ring. The MHI ring is a circular queue
of data structures used to pass the information between host and the
endpoint.
MHI support 3 types of rings:
1. Transfer ring
2. Event ring
3. Command ring
All rings reside inside the host memory and the MHI EP device maps it to
the device memory using blocks like PCIe iATU. The mapping is handled in
the MHI EP controller driver itself.
bus: mhi: ep: Add support for managing MMIO registers
Add support for managing the Memory Mapped Input Output (MMIO) registers
of the MHI bus. All MHI operations are carried out using the MMIO registers
by both host and the endpoint device.
The MMIO registers reside inside the endpoint device memory (fixed
location based on the platform) and the address is passed by the MHI EP
controller driver during its registration.
bus: mhi: ep: Add support for creating and destroying MHI EP devices
This commit adds support for creating and destroying MHI endpoint devices.
The MHI endpoint devices binds to the MHI endpoint channels and are used
to transfer data between MHI host and endpoint device.
There is a single MHI EP device for each channel pair. The devices will be
created when the corresponding channels has been started by the host and
will be destroyed during MHI EP power down and reset.
bus: mhi: ep: Add support for registering MHI endpoint client drivers
This commit adds support for registering MHI endpoint client drivers
with the MHI endpoint stack. MHI endpoint client drivers bind to one
or more MHI endpoint devices inorder to send and receive the upper-layer
protocol packets like IP packets, modem control messages, and
diagnostics messages over MHI bus.
bus: mhi: ep: Add support for registering MHI endpoint controllers
This commit adds support for registering MHI endpoint controller drivers
with the MHI endpoint stack. MHI endpoint controller drivers manage
the interaction with the host machines (such as x86). They are also the
MHI endpoint bus master in charge of managing the physical link between
the host and endpoint device. Eventhough the MHI spec is bus agnostic,
the current implementation is entirely based on PCIe bus.
The endpoint controller driver encloses all information about the
underlying physical bus like PCIe. The registration process involves
parsing the channel configuration and allocating an MHI EP device.
Channels used in the endpoint stack follows the perspective of the MHI
host stack. i.e.,
UL - From host to endpoint
DL - From endpoint to host
binder: Use memcpy_{to,from}_page() in binder_alloc_do_buffer_copy()
The use of kmap_atomic() is being deprecated in favor of kmap_local_page()
where it is feasible. Each call of kmap_atomic() in the kernel creates
a non-preemptible section and disable pagefaults. This could be a source
of unwanted latency, so kmap_local_page() should be preferred.
With kmap_local_page(), the mapping is per thread, CPU local and not
globally visible. Furthermore, the mapping can be acquired from any context
(including interrupts). binder_alloc_do_buffer_copy() is a function where
the use of kmap_local_page() in place of kmap_atomic() is correctly suited.
Use kmap_local_page() / kunmap_local() in place of kmap_atomic() /
kunmap_atomic() but, instead of open coding the mappings and call memcpy()
to and from the virtual addresses of the mapped pages, prefer the use of
the memcpy_{to,from}_page() wrappers (as suggested by Christophe
Jaillet).
Cc: Christophe JAILLET <christophe.jaillet@wanadoo.fr> Acked-by: Todd Kjos <tkjos@google.com> Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com> Link: https://lore.kernel.org/r/20220425175754.8180-4-fmdefrancesco@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
binder: Use kmap_local_page() in binder_alloc_copy_user_to_buffer()
The use of kmap() is being deprecated in favor of kmap_local_page()
where it is feasible. With kmap_local_page(), the mapping is per
thread, CPU local and not globally visible.
binder_alloc_copy_user_to_buffer() is a function where the use of
kmap_local_page() in place of kmap() is correctly suited because
the mapping is local to the thread.
Therefore, use kmap_local_page() / kunmap_local().
Cc: Christophe JAILLET <christophe.jaillet@wanadoo.fr> Acked-by: Todd Kjos <tkjos@google.com> Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com> Link: https://lore.kernel.org/r/20220425175754.8180-3-fmdefrancesco@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
binder: Use memset_page() in binder_alloc_clear_buf()
The use of kmap() is being deprecated in favor of kmap_local_page()
where it is feasible. With kmap_local_page(), the mapping is per
thread, CPU local and not globally visible.
binder_alloc_clear_buf() is a function where the use of kmap_local_page()
in place of kmap() is correctly suited because the mapping is local to the
thread.
Therefore, use kmap_local_page() / kunmap_local() but, instead of open
coding these two functions and adding a memset() of the virtual address
of the mapping, prefer memset_page().
Cc: Christophe JAILLET <christophe.jaillet@wanadoo.fr> Acked-by: Todd Kjos <tkjos@google.com> Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com> Link: https://lore.kernel.org/r/20220425175754.8180-2-fmdefrancesco@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
test_firmware: Error injection for firmware upload
Add error injection capability to the test_firmware module specifically
for firmware upload testing. Error injection instructions are transferred
as the first part of the firmware payload. The format of an error
injection string is similar to the error strings that may be read from
the error sysfs node.
To inject the error "programming:hw-error", one would use the error
injection string "inject:programming:hw-error" as the firmware payload:
The first part of the error string is the progress state of the upload at
the time of the error. The progress state would be one of the following:
"preparing", "transferring", or "programming". The second part of the
error string is one of the following: "hw-error", "timeout", "device-busy",
"invalid-file-size", "read-write-error", "flash-wearout", and "user-abort".
Note that all of the error strings except "user-abort" will fail without
delay. The "user-abort" error will cause the firmware upload to stall at
the requested progress state for up to 5 minutes to allow you to echo 1
to the cancel sysfs node. It is this cancellation that causes the
'user-abort" error. If the upload is not cancelled within the 5 minute
time period, then the upload will complete without an error.
test_firmware: Add test support for firmware upload
Add support for testing the firmware upload driver. There are four sysfs
nodes added:
upload_register: write-only
Write the name of the firmware device node to be created
upload_unregister: write-only
Write the name of the firmware device node to be destroyed
config_upload_name: read/write
Set the name to be used by upload_read
upload_read: read-only
Read back the data associated with the firmware device node named
in config_upload_name
You can create multiple, concurrent firmware device nodes for firmware
upload testing. Read firmware back and validate it using config_upload_name
and upload_red.
Example:
$ cd /sys/devices/virtual/misc/test_firmware
$ echo -n fw1 > upload_register
$ ls fw1
cancel data device error loading power remaining_size status
subsystem uevent
$ dd if=/dev/urandom of=/tmp/random-firmware.bin bs=512 count=4
4+0 records in
4+0 records out
2048 bytes (2.0 kB, 2.0 KiB) copied, 0.000131959 s, 15.5 MB/s
$ echo 1 > fw1/loading
$ cat /tmp/random-firmware.bin > fw1/data
$ echo 0 > fw1/loading
$ cat fw1/status
idle
$ cat fw1/error
$ echo -n fw1 > config_upload_name
$ cmp /tmp/random-firmware.bin upload_read
$ echo $?
0
$ echo -n fw1 > upload_unregister
firmware_loader: Add sysfs nodes to monitor fw_upload
Add additional sysfs nodes to monitor the transfer of firmware upload data
to the target device:
cancel: Write 1 to cancel the data transfer
error: Display error status for a failed firmware upload
remaining_size: Display the remaining amount of data to be transferred
status: Display the progress of the firmware upload
Extend the firmware subsystem to support a persistent sysfs interface that
userspace may use to initiate a firmware update. For example, FPGA based
PCIe cards load firmware and FPGA images from local FLASH when the card
boots. The images in FLASH may be updated with new images provided by the
user at his/her convenience.
A device driver may call firmware_upload_register() to expose persistent
"loading" and "data" sysfs files. These files are used in the same way as
the fallback sysfs "loading" and "data" files. When 0 is written to
"loading" to complete the write of firmware data, the data is transferred
to the lower-level driver using pre-registered call-back functions. The
data transfer is done in the context of a kernel worker thread.
firmware_loader: Split sysfs support from fallback
In preparation for sharing the "loading" and "data" sysfs nodes with the
new firmware upload support, split out sysfs functionality from fallback.c
and fallback.h into sysfs.c and sysfs.h. This includes the firmware
class driver code that is associated with the sysfs files and the
fw_fallback_config support for the timeout sysfs node.
CONFIG_FW_LOADER_SYSFS is created and is selected by
CONFIG_FW_LOADER_USER_HELPER in order to include sysfs.o in
firmware_class-objs.
This is mostly just a code reorganization. There are a few symbols that
change in scope, and these can be identified by looking at the header
file changes. A few white-space warnings from checkpatch are also
addressed in this patch.
net: dsa: mv88e6xxx: Fix port_hidden_wait to account for port_base_addr
The other port_hidden functions rely on the port_read/port_write
functions to access the hidden control port. These functions apply the
offset for port_base_addr where applicable. Update port_hidden_wait to
use the port_wait_bit so that port_base_addr offsets are accounted for
when waiting for the busy bit to change.
Without the offset the port_hidden_wait function would timeout on
devices that have a non-zero port_base_addr (e.g. MV88E6141), however
devices that have a zero port_base_addr would operate correctly (e.g.
MV88E6390).
Fixes: 609070133aff ("net: dsa: mv88e6xxx: update code operating on hidden registers") Signed-off-by: Nathan Rossi <nathan@nathanrossi.com> Reviewed-by: Marek Behún <kabel@kernel.org> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Link: https://lore.kernel.org/r/20220425070454.348584-1-nathan@nathanrossi.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The underlying option POLARFIRE_SOC_SYS_CTRL already supports
COMPILE_TEST so there is no need for this. What's more, if
we force this option on without the underlying option it fails
to build.
Reported-by: kernel test robot <lkp@intel.com> Reported-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Solomon Tan [Sat, 23 Apr 2022 13:01:50 +0000 (21:01 +0800)]
staging: rtl8192e: Remove u16 cast for u16 return value
Remove explicit u16 cast of the function Mk16 return value.
The variable hi does not need to be casted explicitly to u16 as it will
already be casted implicitly.
Remove the explicit u16 casts for u16 argument in the function call. This
function has declared its parameter to be of type u16, so the argument
given to that parameter will be automatically coerced to u16. Therefore,
an explicit cast is unnecessary.
Remove the explicit casts in assignment statements for u16 variables.
Because these variables are declared as u16, there is implicit
conversion to u16 during the assignment, so an explicit cast is
unnecessary.
Solomon Tan [Sat, 23 Apr 2022 13:01:46 +0000 (21:01 +0800)]
staging: rtl8192e: Remove unnecessary u8 cast in comparison
Since the macro VERSION_8190_BD is defined to be 0x3, and the structure
members card_8192_version and IC_Cut are both u8, an explicit u8 cast is
unnecessary.
Remove the unnecessary explicit u8 casts for u8 arguments in function
calls. These functions have declared their parameters to be of type u8,
so the arguments given to those parameters will be automatically coerced
to u8. Therefore, an explicit cast is unnecessary.
Remove the explicit casts in assignment statements for u8 variables.
Because these variables are declared as u8, there is implicit conversion
to u8 during the assignment, so an explicit cast is unnecessary.
staging: vc04_services: Re-add dependency on HAS_DMA to BCM2835_VCHIQ
Thanks to stubs for the NO_DMA=y case, drivers that use the DMA API can
be compile-tested on systems that do not support DMA. Hence the
dependency of BCM2835_VCHIQ on HAS_DMA was dropped to increase compile
coverage.
Unfortunately compilers became smarter, leading to new failures.
E.g. for a CONFIG_SUN3=y allmodconfig kernel with gcc 9.4.0:
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_2835_arm.c: In function ‘free_pagelist’:
arch/m68k/include/asm/string.h:72:25: warning: argument 2 null where non-null expected [-Wnonnull]
72 | #define memcpy(d, s, n) __builtin_memcpy(d, s, n)
| ^~~~~~~~~~~~~~~~~~~~~~~~~
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_2835_arm.c:614:4: note: in expansion of macro ‘memcpy’
614 | memcpy((char *)kmap(pages[0]) +
| ^~~~~~
arch/m68k/include/asm/string.h:72:25: note: in a call to built-in function ‘__builtin_memcpy’
72 | #define memcpy(d, s, n) __builtin_memcpy(d, s, n)
| ^~~~~~~~~~~~~~~~~~~~~~~~~
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_2835_arm.c:614:4: note: in expansion of macro ‘memcpy’
614 | memcpy((char *)kmap(pages[0]) +
| ^~~~~~
This happens because the compiler can trace back the source pointer to a
allocation by dma_alloc_attrs(), which always returns NULL if NO_DMA=y.
Avoid this reinstating the dependency of the BCM2835_VCHIQ symbol on
HAS_DMA, and by restricting the selection of BCM2835_VCHIQ.
Solomon Tan [Mon, 25 Apr 2022 05:28:03 +0000 (13:28 +0800)]
Revert "staging: r8188eu: use in-kernel ieee channel"
This reverts commit 0afaa121813ed602bd203759c339cb639493f8c2 as changing
rtw_ieee80211_channel to ieee80211_channel causes a memcpy bug as
reported in:
https://lore.kernel.org/linux-staging/67e2d10b-7f0f-9c5a-ce31-376b83ffba9e@gmail.com/
due to their size differences.
Fixes: 0afaa121813e ("staging: r8188eu: use in-kernel ieee channel") Reported-by: Michael Straube <straube.linux@gmail.com> Signed-off-by: Solomon Tan <wjsota@gmail.com> Link: https://lore.kernel.org/r/20220425052802.2419-1-wjsota@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Hangbin Liu [Mon, 25 Apr 2022 01:45:02 +0000 (09:45 +0800)]
net/af_packet: add VLAN support for AF_PACKET SOCK_RAW GSO
Currently, the kernel drops GSO VLAN tagged packet if it's created with
socket(AF_PACKET, SOCK_RAW, 0) plus virtio_net_hdr.
The reason is AF_PACKET doesn't adjust the skb network header if there is
a VLAN tag. Then after virtio_net_hdr_set_proto() called, the skb->protocol
will be set to ETH_P_IP/IPv6. And in later inet/ipv6_gso_segment() the skb
is dropped as network header position is invalid.
Let's handle VLAN packets by adjusting network header position in
packet_parse_headers(). The adjustment is safe and does not affect the
later xmit as tap device also did that.
In packet_snd(), packet_parse_headers() need to be moved before calling
virtio_net_hdr_set_proto(), so we can set correct skb->protocol and
network header first.
There is no need to update tpacket_snd() as it calls packet_parse_headers()
in tpacket_fill_skb(), which is already before calling virtio_net_hdr_*
functions.
skb->no_fcs setting is also moved upper to make all skb settings together
and keep consistency with function packet_sendmsg_spkt().
bug: Have __warn() prototype defined unconditionally
The __warn() prototype is declared in CONFIG_BUG scope but the function
definition in panic.c is unconditional. The IBT enablement started using
it unconditionally but a CONFIG_X86_KERNEL_IBT=y, CONFIG_BUG=n .config
will trigger a
arch/x86/kernel/traps.c: In function ‘__exc_control_protection’:
arch/x86/kernel/traps.c:249:17: error: implicit declaration of function \
‘__warn’; did you mean ‘pr_warn’? [-Werror=implicit-function-declaration]
Pull up the declarations so that they're unconditionally visible too.
arm64: dts: mediatek: Add device-tree for MT8195 Demo board
Add basic device-tree for the MT8195 Demo board. The
Demo board is made by MediaTek and has a MT8195 SoC,
associated with the MT6359 and MT6360 PMICs, and
the MT7921 connectivity chip.
The IOs available on that board are:
* 1 USB Type-C connector with DP aux mode support
* 1 USB Type-A connector
* 1 full size HDMI RX and 1 full size HDMI TX connector
* 1 uSD slot
* 40 pins header
* SPI interface header
* 1 M.2 slot
* 1 audio jack
* 1 micro-USB port for serial debug
* 2 connectors for DSI displays
* 3 connectors for CSI cameras
* 1 connector for a eDP panel
* 1 MMC storage
This commit adds basic support in order to be able to boot.
In mt8192 SoC, mmc driver dose not use the MSDC module to control
clock. It will read/write register to enable/disable clock. Also
there is no other device of mt8192 using MSDC controller.
We add mmc nodes for mt8192 SoC and remove the clock-controller in
dts for avoid a duplicate unit-address(11f60000) warning.
Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Reviewed-by: Nícolas F. R. A. Prado <nfraprado@collabora.com> Link: https://lore.kernel.org/r/20220407113703.26423-2-allen-kh.cheng@mediatek.com Signed-off-by: Matthias Brugger <matthias.bgg@gmail.com>
The driver data of MT8183 and MT8173 are different.
The value of has_gamma for MT8173 is true while the value of MT8183 is
false. Therefore, the compatible of disp_aal for MT8183 is not suitable
for the compatible for MT8173.
Hui Liu [Thu, 21 Apr 2022 01:26:23 +0000 (09:26 +0800)]
arm64: dts: mt6359: add PMIC MT6359 related nodes
MT6359 is the primary PMIC for MT8192.
Add PMIC MT6359 related node which is used for MT8192 platform.
Signed-off-by: Hui Liu <hui.liu@mediatek.com> Reviewed-by: Nícolas F. R. A. Prado <nfraprado@collabora.com> Tested-by: Nícolas F. R. A. Prado <nfraprado@collabora.com> Link: https://lore.kernel.org/r/20220421012623.3446-2-hui.liu@mediatek.com Signed-off-by: Matthias Brugger <matthias.bgg@gmail.com>
arm64: dts: mediatek: align thermal zone node names with dtschema
Align the name of thermal zone node to dtschema to fix warnings like:
arch/arm64/boot/dts/mediatek/mt8173-elm.dt.yaml:
thermal-zones: 'cpu_thermal' does not match any of the regexes: '^[a-zA-Z][a-zA-Z0-9\\-]{1,12}-thermal$', 'pinctrl-[0-9]+'
projects / people / ms / linux.git / log
