Articles about cscope+vim+ctags+taglist

http://cscope.sourceforge.net/cscope_vim_tutorial.html
http://blog.verycd.com/yoyopub/showentry=22855
http://adrianhuang.blogspot.com/2007/09/trace-code-gvimctagscscope.html
http://blog.chhsu.org/2009/03/multi-projects-of-vim-using-cscope.html

• Install Ctags

$sudo apt-get install exuberant-ctags

• Install Cscope

$sudo apt-get install cscope

• Make Sure Your Vim Support Cscope

$ vim --version | grep cscope

• Download Cscope Plugin

$ mkdir ~/.vim/pluging
$cd ~/.vim/plugin
$ wget http://cscope.sourceforge.net/cscope_maps.vim

• Install Taglist

Download tagliast from http://vim-taglist.sourceforge.net/download.html.
Follow the installation steps from http://vim-taglist.sourceforge.net/installation.html
Add the following line into your .vimrc will help you to open taglist and switch between tag and editor window.

filetype on
nnoremap < silent > < f2 > :TlistToggle < cr >
nnoremap < silent > < f3 > :wincmd p < cr >

• Generate DB & test
  

#!/bin/bash
CSCOPE_FILE=cscope.out
if [ -n "$1" ]; then
   echo "Source code directory: " $1
   echo "Create file map : " $CSCOPE_FILE
   find $1 -name "*.h" -o -name "*.c" -o -name "*.cpp" > $CSCOPE_FILE
   cscope -bkq -i $CSCOPE_FILE
   # cscope -Rbkq
   ctags -R
else
   echo "Please key-in path of project"
fi

$cd source-code
$cscope

• Other IDEs

Other IDE includes Anjuta or Source Navigator. You can try it.

Memory remap

• ioremap

The function will map a range of I/O space to kernel space. That fix page table of some kernel memory to physical I/O memory. You can access the mapped kernel space as if you are accessing physical I/O memory.

• io_remap_pfn_range, remap_page_range

Remap address range in the kernel space to user space. User space process can access the mapped kernel address range by accessing user-space address.

Run level

debian的runlevel級別定義如下：
0 - Halt
1 - Single
2 - Full multi-user with display manager (GUI)
3 - Full multi-user with display manager (GUI)
4 - Full multi-user with display manager (GUI)
5 - Full multi-user with display manager (GUI)
6 - Reboot

而redhat的runlevel級別定義如下：
0 - Halt
1 - Single
2 - Not used/User definable
3 - Full multi-user NO display manager
4 - Not used/User definable
5 - Full multi-user with display manager (GUI)
6 - Reboot

You can modify your default run-level:
1. Ubuntu: edit /etc/event.d/rc-default
2. Redhat: edit /etc/inittab

Kernel scheduling

• Schedule

You can relinguish your control of CPU. Scheduler will run another task.

REF: schedule()

• Wait queue

When you can't satisfy user request, put them in wait queue. It's like a I/O wait queue in OS text book.

REF: init_wait_queue_head(), interruptible_sleep_on(), wake_up()

Note: You can also put yourself in the wait queue (when waiting for I/O) by manually:

DECLARE_WAIT_QUEUE(wait, current);
add_wait_queue(wait_q_head, &wait);
while(!condition){
    set_current_state(TASK_INERRUPTIBLE);
    schedule();
}
set_current_state(TASK_RUNNING);
remove_wait_queue(q, &wait);

• Kernel timer

You can trigger an kernel timer for a delay execution.

REF: init_timer(), del_timer(), add_timer()

• Task queue

You can use task queue for a delay execution. Very similar to kernel timer. Kernel maintain three pre-defined task queue: scheduler queue, tq_timer, tq_immediate

REF: INIT_TQUEUE(),queue_task(), schedule_task(), DECLARE_TASK_QUEUE(), run_task_queue()

• Tasklet

You can use a tasklet to schedule a interrupt work in bottom half. It is build on top of softirq

REF: DECLARE_TASKLET(), tasklet_init(), tasklet_schdule()