0c3db448bb
added the option to generate the waveforms to drive ws2812 and similar LEDs using different
hardware (e.g. RP2040 PIO instead of SPI).
For that new mode, the concept of CONFIG_WS2812_FREQUENCY is different. Instead of
the SPI frequency (commonly a few MHz), it is the frequency of the actual output waveform
(commonly 400 or 800 kHz).
There was an attempt to express the SPI frequency divided by 10, but it's not actually the
case either (it would be divided by 8).
I think it is clearer to explain in Kconfig what CONFIG_WS2812_FREQUENCY means for each mode
and go back to the previous behaviour for the original SPI mode (also to avoid breaking
out-of-tree boards).
The commit 664d45dcba updated the
behavior of the framebuffer's putarea callback enabling it to be
used to draw a particular area of the display. Previously, putarea
was only used to draw the entire area of the display. Any different
area was drawn, row by row, by putrun. Also, before checking for
which callback to call, the framebuffer driver adjusted the buffer
reference that was going to be used for calling the driver's callback
to point to the init of the relevant data. After that commit, the
framebuffer's buffer reference passed to the driver's putarea now
contains the data to draw the entire display. Unlike the previous
version of that implementation, only the putrun's callback buffer
was being referenced from the address that contains the data that
actually is being drawn.
This commit fixes it by adjusting the reference for the run buffer
passed to the putrun/putarea callback. It always starts from the
beginning of the relevant data that is actually being drawn. That is
necessary because lcddev (which uses the same LCD display driver
callbacks) actually don't allocate a buffer containing the data to
draw the whole display, so the same putarea implementation of the
LCD drivers would'n be able to work for both lcddev and framebuffer.
Also it's necessary to pass the stride argument to the LCD drivers
in order to enable them to do partial writes by calculating the
buffer offset while sending row-by-row. The stride is equal the
width multiplied by the bytes per pixel (may add some padding)
for framebuffer and is equal to the lenght of the row being drawn
(multiplied by the same BPP) for lcddev.
Why this approach?
Other possible approaches would be:
1) modify lcddev driver to translate received buffer data to a
buffer similar to the framebuffer. That wouldn't be efficient
considering memory allocation.
2) Create a new callback function. While possible, it'd be confusing
to create a different callback to draw the whole screen and another
to draw only an area of the screen. Also, these callbacks would
differ themselves only from the way the buffer is filled.
3) Simply reverting 664d45dcba would
break the usage of the putarea callback to draw an area of the
display, which would also be inefficient.
This approach is based on the Zephyr's implementation of the ST7789
driver: the buffer starts from the beginiing of the region that would
be drawn. The display device driver's putarea implementation should
check if the operation refers to a full screen/full row and implement
(if possible) a single operation to send the data to be drawn more
efficiently.
Finally, this approach requires that the drivers which implement
the putarea callback and expects the entire framebuffer buffer
to be modified. They don't need to calculate where the data begins
as the new buffer represents the data from the address that is
actually being drawn. This includes adjusting the LCD drivers
GC9A01 and ST7789 and the driver for APA102-based LED matrix display.
In file included from /home/archer/code/nuttx/n4/incubator-nuttx/drivers/rc/lirc_dev.c:29:
drivers/rc/lirc_dev.c: In function 'lirc_raw_event':
drivers/rc/lirc_dev.c:959:14: warning: format '%u' expects argument of type 'unsigned int',
but argument 3 has type 'uint32_t' {aka 'long unsigned int'} [-Wformat=]
959 | rcinfo("delivering %uus %d to lirc\n", ev.duration, ev.pulse ? 1 : 0);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~
| |
| uint32_t {aka long unsigned int}
drivers/rc/lirc_dev.c:959:27: note: format string is defined here
959 | rcinfo("delivering %uus %d to lirc\n", ev.duration, ev.pulse ? 1 : 0);
| ~^
| |
| unsigned int
| %lu
Signed-off-by: chao.an <anchao@xiaomi.com>
1. about interval:
If interval is not set, generation is increased by 1 along with
publish and copy, multi-copy is continuous.
If interval is set, pick proper samples from buffer based on
mainline/user generation, multi-copy is one-by-one.
2. about bufferpos:
user->bufferpos always point to next position to check.
data user last read
----------v--------------------------
| | | |
-------------------^-----------------
bufferpos
If buffer is full, bufferpos point to buffer.head
Examples:
If a buffer contains 4 samples, newest generatoin is 40.
-------------------------------------
|10 |20 |30 |40
------------------------------^------
|
if user's next generation is 42, notify user to copy No.40 sample,
because 42 is closer to 40 than 50.
-------------------------------------
|10 |20 |30 |40
----------------------------------^--
|
if user's next generation is 48, do not notify user,
because 48 is closer to 50, which is next mainline sample.
Signed-off-by: jihandong <jihandong@xiaomi.com>
Signed-off-by: Jiuzhu Dong <dongjiuzhu1@xiaomi.com>
This patch adds upper half driver for IPCC character driver. IPCC
is a Inter Processor Communication Controller.
Driver is still immature and not thoroughly tested, thus
CONFIG_EXPERIMENTAL is required.
Signed-off-by: Michał Łyszczek <michal.lyszczek@bofc.pl>
Certain variants of the W25QxxxJV Quad SPI NOR flash ICs have the Quad
SPI mode disabled after reset. As the w25qxxxjv driver is operating the
device in Quad SPI mode, make sure it gets enabled if required during
device initialization.
Signed-off-by: Michael Jung <michael.jung@secore.ly>
power/pm_procfs.c:282:27: warning: format specifies type 'unsigned long' but the argument has type 'time_t' (aka 'unsigned int') [-Wformat]
dom->wake[state].tv_sec,
^~~~~~~~~~~~~~~~~~~~~~~
make[2]: Entering directory '/home/ligd/platform/m2/audio/frameworks/media'
power/pm_procfs.c:283:27: warning: format specifies type 'unsigned long' but the argument has type 'unsigned int' [-Wformat]
100 * dom->wake[state].tv_sec / sum,
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
power/pm_procfs.c:284:27: warning: format specifies type 'unsigned long' but the argument has type 'time_t' (aka 'unsigned int') [-Wformat]
dom->sleep[state].tv_sec,
^~~~~~~~~~~~~~~~~~~~~~~~
power/pm_procfs.c:285:27: warning: format specifies type 'unsigned long' but the argument has type 'unsigned int' [-Wformat]
100 * dom->sleep[state].tv_sec / sum,
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
power/pm_procfs.c:286:27: warning: format specifies type 'unsigned long' but the argument has type 'time_t' (aka 'unsigned int') [-Wformat]
total.tv_sec,
^~~~~~~~~~~~
power/pm_procfs.c:287:27: warning: format specifies type 'unsigned long' but the argument has type 'unsigned int' [-Wformat]
100 * total.tv_sec / sum);
^~~~~~~~~~~~~~~~~~~~~~~~
Signed-off-by: ligd <liguiding1@xiaomi.com>
1.Allow multi users to access the same sensor device simultaneously.
2.Get real state of sensor device by cmd SNIOC_GET_STATE for users.
3.Get update state since last read by poll without timeout for users.
4.Sensor device will be activated when first user open and will close when
last user closed.
5.When multi users to access device, driver always set the minimum
sampling interval and latency to the sensor device and allow
downsampled for users above the minimum sampling interval.
6.The circbuffer will overwrite old data when buffer is full, so if users
don't read data soon, data will be lost, and the oldest data in circbuffer
are returned to the users.
7.Always read the last data in the circbuffer as initial value for new
users when the sensor device has not yet generated new data.
8.when user uses poll, if subscription interval is satisfied, the POLLIN
events is returned for each users.
9.When new user generate or the state of sensor device changed, the POLLPRI
will notify to all users.
10.Support multi advertisers to subscribe their own data as loop test.
Signed-off-by: Jiuzhu Dong <dongjiuzhu1@xiaomi.com>
1.cmd:SNIOC_REGISTER with struct sensor_reginfo_s to register user sensor
2.cmd:SNIOC_UNREGISTER to unregister user sensor
Signed-off-by: Jiuzhu Dong <dongjiuzhu1@xiaomi.com>
1.Simplify buffer opeations to avoid frequent resize for batch and no-batch mode.
2.When sensor event is first generated, the buffer is initialized.
3.Remove and merge batch_number to buffer_number when device support batch mode.
Signed-off-by: Jiuzhu Dong <dongjiuzhu1@xiaomi.com>
'putarea' function now has to select the appropriate values from the
whole frame buffer instead of the previous expectation of having just
specified rectangle passed to it.
This required extension of WRRAM handling function as we now are not
writing simply buffer but we want to skip some values and do that
multiple times. Having two implementation in this case is worst as this
function is actually called only twice in the whole code (thus making
dedicated function for every call if we would have two variants).
Calling `st7789_wrram` multiple times is not an option as command
ST7789_RAMWR resets the position and thus calling it multiple times just
overwrites the previous values.
I found an issue in the APA102 framebuffer that wasn't spot
before by fb test because this application draw symetric squares
Normally in a display each line starts at left and increase to the
right, however for this APA102 matrix the next line starts at the
end of previous line and move the opposite direction, forming a
zig-zag pattern.
If CONFIG_SMARTFS_MULTI_ROOT_DIRS is enabled and dev->partname is not empty,
dev->rwbuffer is an uninitialized variable and register_blockdriver is called.
This patch introduces a configuration option USBMSC_WRMULTIPLE,
which is used to store multiple blocks into a larger chunk that
then gets written via the mmcsd_writemultiple() (in case mmcsd
is used).
The bottleneck with the current implementation is the poor
performance due to short block writes. USBMSC_DRVR_WRITE()
always writes only one sector (with eMMC that's usually 512 bytes).
eMMC devices usually erase much larger regions with near constant
time (see Jedec JESD84-B51, Extended CSD register byte [225],
SUPER_PAGE_SIZE): 'This register defines one or multiple of
programmable boundary unit that is programmed at the same time.'
If USBMSC_WRMULTIPLE is defined, then USBMSC_NWRREQS is used to
allocate the write buffer size. We don't want this to be the
default behavior yet as this may reveal unseen bugs in usb drivers
due to the faster overall performance.
Sample configurations with measured performance:
- Without USBMSC_WRMULTIPLE: 470 Kb/s
- With USBMSC_WRMULTIPLE, CONFIG_USBMSC_NWRREQS=4: 1.1 Mb/s
(dd with bs=2k)
- With USBMSC_WRMULTIPLE, CONFIG_USBMSC_NWRREQS=16: 5.2 Mb/s
(dd with bs=8k)
No doubt, this feature alone may make the mass storage work 10
times faster than before with eMMC cards.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
Displays with resolution that is not multiplication of 8 had some issues in landscape mode (columns ware shifted).
Fixed function for copying bits.
Optimized memory reload for landscape mode.
only break out the continuous send if IP stack has no data to send to avoid
tx being constantly interrupted by RX
Signed-off-by: chao.an <anchao@xiaomi.com>
RX/TX shared free queue on bcmf implementation, if TX occupies the
free queue completely, RX will trigger read abort because it cannot
alloc buffer successfully from the shared free queue. This commit will
limit the sending entries of tx and prevent rx triggering abort
Signed-off-by: chao.an <anchao@xiaomi.com>
In order to support multiple LCD instances per board, add a pointer from
lcd_planeinfo_s to the lcd_dev_s which it belongs to. Also enhance the
putrun, getrun, putarea and getarea methods to pass through the
lcd_dev_s pointer to the respective device driver.
Port all LCD device drivers to this lcd_planeinfo_s extension.
Enhance SSD1306 driver to support multiple LCDs.
Signed-off-by: Michael Jung <michael.jung@secore.ly>
Use high priority queue if the bcmf daemon task has a higher priority
than HPWORK, which will bring better performance especially on devices
that focus on real-time of network.
Signed-off-by: chao.an <anchao@xiaomi.com>
1. Replace SCAN_RESULT_SIZE to SCAN_RESULT_ENTRIES
2. filter scan result with better rssi
3. Sort scan result by rssi
Signed-off-by: chao.an <anchao@xiaomi.com>
since oob irq is already setup in bcmf_bus_setup_interrupts, it waste
time and energy to do the same thing repeatly in the work thread.
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
1. Wake up tx if there is data alreay to sent after rx update
2. Try tx queue again after every data frame sent
iperf test:
```
TCP RX 4M->16M
TCP TX 2M->14M
UDP RX 18M->26M
UDP TX 2M->18M
```
Signed-off-by: chao.an <anchao@xiaomi.com>
added an option to load nvram from file system, which will help developers
easily to tuning the firmware parameters without compile firmware.
Signed-off-by: chao.an <anchao@xiaomi.com>
some SDIO device will return the larger frames out of bound
error log:
bcmf_sdpcm_readframe: Frame is too large, cancel 1544 1536
bcmf_sdpcm_readframe: Frame is too large, cancel 1544 1536
bcmf_sdpcm_readframe: Frame is too large, cancel 1544 1536
Signed-off-by: chao.an <anchao@xiaomi.com>
usbmsc.h uses the CONFIG_USBMSC_IFNOBASE configuration option,
but it cannot be set anywhere. Provide means to configure this
option which is useful when it's other than the default zero.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
