Recently, I found that RT-Thread cannot output floating-point numbers normally. When debugging, I found that the calculation results are correct, but the output numbers are not correct. The specific performance is that %f is not supported. The problem occurs in rt_kprintf() and ulog log output. .
The current solution after consulting the data is as follows:
rt_kprintf() cannot output floating point number problem solved
For details, please refer to the blog here: RT-thread rt_kprintf() function formats and outputs floating-point numbers
Modification method: Find the specific location of the rt_kprintf() function (under the /rt-thread/src/kservice.c file), and add the header file #include <stdio.h> at the top (remember to enable it in RT-Thread Settings libc support), and then in the specific implementation of this function, the original
length = rt_vsnprintf(rt_log_buf, sizeof(rt_log_buf) - 1, fmt, args);
change into
length = vsnprintf(rt_log_buf, sizeof(rt_log_buf) - 1, fmt, args);
Then the internal content becomes
void rt_kprintf(const char *fmt, ...)
{
va_list args;
rt_size_t length;
static char rt_log_buf[RT_CONSOLEBUF_SIZE];
va_start(args, fmt);
/* the return value of vsnprintf is the number of bytes that would be
* written to buffer had if the size of the buffer been sufficiently
* large excluding the terminating null byte. If the output string
* would be larger than the rt_log_buf, we have to adjust the output
* length. */
length = vsnprintf(rt_log_buf, sizeof(rt_log_buf) - 1, fmt, args);
if (length > RT_CONSOLEBUF_SIZE - 1)
length = RT_CONSOLEBUF_SIZE - 1;
#ifdef RT_USING_DEVICE
if (_console_device == RT_NULL)
{
rt_hw_console_output(rt_log_buf);
}
else
{
rt_uint16_t old_flag = _console_device->open_flag;
_console_device->open_flag |= RT_DEVICE_FLAG_STREAM;
rt_device_write(_console_device, 0, rt_log_buf, length);
_console_device->open_flag = old_flag;
}
#else
rt_hw_console_output(rt_log_buf);
#endif
va_end(args);
}
Recompile and then print out floating-point numbers.
ulog log floating point output
In the more options of the RT-Thread Settings configuration software package function, choose to turn on floating point support (tossing to death, I don’t know if I have to manually turn on this function)
after turning it on, you can use LOG_I() and other functions to perform floating point numbers normally. Output.