Re: disable interrput under ppc?

On Wednesday 05 September 2007, you wrote:

On Tue, 4 Sep 2007 13:24:33 -0500

Scott Wood [off-list ref] wrote:

quoted

On Tue, Sep 04, 2007 at 10:44:14AM +0800, Wang, Baojun wrote:

quoted

hi, list

  How can I disable interrput like `cli' in x86?

local_irq_disable().

quoted

I want the following code freeze the box but it doesn't, $MSR is
altered however the box is still alive, how can I freeze the box like
`cli' in x86? Thanks very much!

Disabling interrupts doesn't freeze the machine by itself -- it depends
on what you do afterwards, which you didn't show.

I need to check whether the interrupt have been disabled already, so I'm 
expected after local_irq_disable() ( __asm__ __volatile__ ("wrteei 
0": : :"memory) ), the machine is dead but it doesn't.

If you're using a BOOKE variant, there is a way to use the DBCR0
register to set debug wait enable or something to basically hang the
CPU until something external clears that bit.  But why you would want
to do that, I have no idea.

Sorry, I just want to check whether disable_irq/enable_irq is functioning.

josh

I'm not really getting confused about the following code:

  // Executing all pending events
  while ((event = get_next_set_bit_and_clear
          (pending, xm_root_domain -> events -> masked_events)) != -1) {
    // An event  pending when there  is not any  handler installed
    // catching it??? it must be an error
    //assert (xm_root_domain -> event_handler [event]);
    if (!xm_root_domain -> events -> event_handler [event]) continue;
      // Before executing an event handler, events' flag is disabled
      // and the event is masked

    printk(KERN_EMERG"sync_root_events: looping!\n");

    set_bit (xm_root_domain -> events -> masked_events, event);
    printk(KERN_EMERG"sync_root_events: looping 2!\n");
    disable_events_flag (xm_root_domain);
    printk(KERN_EMERG"sync_root_events: looping 3!\n");
    printk(KERN_EMERG"hw_irq_disabled?: %d\n", hw_irq_disabled());
    hw_enable_irq ();
    printk(KERN_EMERG"sync_root_events: looping 4!\n");
    // Here, the event handler is executed always
    // with the same conditions, that is, event flag disabled
    // and the executed event masked as well
    (*xm_root_domain -> events -> event_handler[event])
      (event, regs);
    printk(KERN_EMERG"sync_root_events: looping 5!\n");
    hw_disable_irq();
    printk(KERN_EMERG"sync_root_events: looping 6!\n");
    // Events flag is automatically enabled (iret emulation)
    enable_events_flag (xm_root_domain);
  }

hw_enable_irq() and hw_disable_irq() is exactly the same as 
local_irq_enable()/local_irq_disable(), but the output shows while running 
hw_eable_irq() (at that point irq is disabled), the rest of the code is 
blocked, and I don't know why:

output:
[  172.049381] sync_root_events: looping!
[  172.053159] sync_root_events: looping 2!
[  172.057094] sync_root_events: looping 3!
[  172.061027] hw_irq_disabled?: 1
[  172.997338] sync_root_events: looping!
[  173.001095] sync_root_events: looping 2!
[  173.005028] sync_root_events: looping 3!
[  173.008961] hw_irq_disabled?: 1


I'm sorry If I'm asking a bad question, I'm not familiar with ppc/kernel 
programming at this moment. Great thanks to Scott and josh for your reply. 

  Regards,
Wang
-- 
Wang, Baojun                                        Lanzhou University
Distributed & Embedded System Lab              http://dslab.lzu.edu.cn
School of Information Science and Engeneering        wangbj@lzu.edu.cn
Tianshui South Road 222. Lanzhou 730000                     .P.R.China
Tel:+86-931-8912025                                Fax:+86-931-8912022