OXape reports: http://www.rt-thread.org/phpBB3/topic3939.html
          driver/ringbuffer: when data length is bigger than space length
          for rb, the read_index in rb is not update.
    When length > space, update the read_index, the bug can be fix, and
          works fine.

If there is no SHT$$INIT_ARRAY, calling
$Super$$__cpp_initialize__aeabi_() will fault. At least until Keil5.12
the problem still exists. So we have to initialize the C++ runtime our
own.

Currently only lpc43xx is supported.

Misc fix

As we are sending C continuously, there is a chance that the
sender(remote) receive an C after sending the first handshake package.
So the sender will interpret it as NAK and re-send the package. So we
just ignore it and proceed.

This saves 1KiB ram usage.

There is no list_mem and only list_memheap when RT_USING_MEMHEAP_AS_HEAP
is enabled.

System will crash when the romfs is erased. Add checks before using them
to avoid it.

sync

Move PNG_NO_WRITE_SUPPORTED outside of RT_USING_NEWLIB. #197

Frequent 'C' on the handshake will confuse some sender(lrzsz for
example).

Make fast path fast and slow path slow.

2. Fix data transfer size calculate error issue;

add some comments of lwip porting

The ptr is set in the following code. So remove the first one.

The digit in octal numbers should with in 0~7. Check on it in
token_proc_number. This issue is found by Clang.

Clang complains about "Null pointer argument in call to string
comparison function". So assert on it.

log_session_lvl is suitable for performance critical places where in
most cases, the log is turned off by level. If the session is const and
the level is greater than session->lvl, the whole function will be
optimized out.

Compiler may do more optimization when the parameter is const.

[USB]fix rt_usbd_set_config() BUG which set the USB device to configurat...

When MSH enabled, the prompt will change at the two shells. So we should
use dynamic prompt when FINSH_USING_MSH defined.

[msh] don't complain on empty command

Just branch out and no complain on command length == 0;

When DFS_USING_WORKDIR is enabled, msh will try to chdir if the command
is not found. For example, if there is folder named "SD", and there is
no `SD` command and `SD.mo` module, typing "SD" in msh will cd to the
folder.

In more detail, the execution order right now is:

1. built-in command
2. module
3. cd to path

If you have a folder named "ps", you have to run `cd ps` to change the
working dir.

Msh can execute module directly. So try to auto-complete the path when
user press TAB. There is a chance that the path is a module.

2.adjust some viable and micro name
3.some other small fix

This reverts commit 5aeedc4f3a, reversing
changes made to 3d80e414ff.

This merge breaks every uart driver in all BSPs. @bright-pan , is there
any way to get this done without breaking existing code?

Add hardware flow support.
Use new struct serial_ringbuffer.

Pass USB CV test verification;
Code cleanup;

This issue causes fstat & lseek not correct.

update CMSIS version to 3.20

fix msh_exec(char* cmd, rt_size_t length) bug

I have to mention that this bug is introduced by myself, in the last
commit <fd07f30f511335f506ae6557e300093d4dde421e> to dfs_nfs.c

fix dfs_unmount bug

more infomation:
http://www.rt-thread.org/phpBB3/topic2965.html

fix #211

empty is confused in dfs_filesystem_lookup.

clean code with pointer rather than index

decrease indentation to make cleaner code
fix mkfs bug when there is no mkfs implementation