RE: Non Preemptive Kernel.

["Dhiman, Gaurav" <Gaurav.Dhiman@ca.com>]

Thanks for the code snippet.
It really helped in understanding what exactly happens.
Can you please explain what we are doing in following code:

void update_process_times(int user_tick)
{
............
........
.......

  if (p->nice > 0)
   kstat.per_cpu_nice[cpu] += user_tick;
  else
   kstat.per_cpu_user[cpu] += user_tick;
  kstat.per_cpu_system[cpu] += system;
 } else if (local_bh_count(cpu) || local_irq_count(cpu) > 1)
  kstat.per_cpu_system[cpu] += system;

.......
.....
.......
}

I think this code is something related to SMP

Thanks & Regards,
Gaurav Dhiman


-----Original Message-----
From: mohanlal jangir [mailto:mohanlal@samsung.com] 
Sent: Friday, June 25, 2004 8:29 PM
To: Dhiman, Gaurav; Jan Hudec; sajay@temenos.com
Cc: kernelnewbies@nl.linux.org
Subject: Re: Non Preemptive Kernel.

Function do_timer is called during timer interrupt (from
do_timer_interrupt,
from timer_interrupt). This increments jiffies and calls
update_process_times. Function update_process_times checks for pid of
current process. I believe, if timer_interrupt is called from kernel,
current->pid will be zero, otherwise non-zero. If current->pid is zero,
current->need_resched is not set. Have a look at following code from
arch/i386/kernel/time.c

void do_timer(struct pt_regs *regs)
{
 (*(unsigned long *)&jiffies)++;
#ifndef CONFIG_SMP
 /* SMP process accounting uses the local APIC timer */

 update_process_times(user_mode(regs));
#endif
 mark_bh(TIMER_BH);
 if (TQ_ACTIVE(tq_timer))
  mark_bh(TQUEUE_BH);
}

/*
 * Called from the timer interrupt handler to charge one tick to the
current
 * process.  user_tick is 1 if the tick is user time, 0 for system.
 */
void update_process_times(int user_tick)
{
 struct task_struct *p = current;
 int cpu = smp_processor_id(), system = user_tick ^ 1;

 update_one_process(p, user_tick, system, cpu);
 if (p->pid) {
  if (--p->counter <= 0) {
   p->counter = 0;
   p->need_resched = 1;
  }
  if (p->nice > 0)
   kstat.per_cpu_nice[cpu] += user_tick;
  else
   kstat.per_cpu_user[cpu] += user_tick;
  kstat.per_cpu_system[cpu] += system;
 } else if (local_bh_count(cpu) || local_irq_count(cpu) > 1)
  kstat.per_cpu_system[cpu] += system;
}

Regards
Mohanlal

----- Original Message -----
From: "Dhiman, Gaurav" <Gaurav.Dhiman@ca.com>
To: "Jan Hudec" <bulb@ucw.cz>; <sajay@temenos.com>
Cc: <kernelnewbies@nl.linux.org>
Sent: Friday, June 25, 2004 5:20 PM
Subject: RE: Non Preemptive Kernel.



Can you tell us, how the retuirn code of interrupt identifies if the
control is returning to kernel space or user space. I mean how it comes
to know that interrupt occurred in kernel space or user space. What
exactly the returning code on any interrupt checks for this.

Regards,
Gaurav


-----Original Message-----
From: kernelnewbies-bounce@nl.linux.org
[mailto:kernelnewbies-bounce@nl.linux.org] On Behalf Of Jan Hudec
Sent: Friday, June 25, 2004 1:24 PM
To: sajay@temenos.com
Cc: kernelnewbies@nl.linux.org
Subject: Re: Non Preemptive Kernel.

On Fri, Jun 25, 2004 at 10:56:31 +0530, sajay@temenos.com wrote:
> Hi,
>
> I have a doubt in the Non Preemptive Kernel. In the case of a Non
> Preemptive Kernel, the Kernel Control Path cannot be preempted unless
it
> relinquishes the CPU Voluntarily. In this scenario, supposing a Timer
> interrupt occurs, how does the timer Interrupt handler find out that a

> Kernel Control Path is running and hence the process shouldn't be
> preempted. Or is this handled by the schedule() function.

It does not. IIRC schedule happens during return to userspace. Return to
userspace does not happen if the interrupt interrupted kernel code. And
note, that it really happens during EACH return to userspace (if
need_resched is set), so if interrupt causes process to wake up, it's
scheduled soon.

------------------------------------------------------------------------
-------
Jan 'Bulb' Hudec
<bulb@ucw.cz>

--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive:       http://mail.nl.linux.org/kernelnewbies/
FAQ:           http://kernelnewbies.org/faq/





--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive:       http://mail.nl.linux.org/kernelnewbies/
FAQ:           http://kernelnewbies.org/faq/