04-11-2006
dsberrf has a valid point, this would be a lengthy chapter in a book. Another point is that different kernels has different algorithms for scheduling. But briefly, the "clock" is a component that generates periodic interrupts. The interrupt routine does stuff when it is called by the clock. A major routine is called a certain number of times per second with 100 times per second being a common value. This major routine is called a "clock tick" (or "jiffy" on linux). A process' timeslice might be one tenth of a second so this would be represented as a 10. The timeslice is always some fixed number of "clock ticks". At each clock tick, it is decremented. When it hits zero, the scheduler will decide which process gets the cpu. If no other process wants the cpu, the process might get another timeslice.
There is a lot more to this, but I'm not going to write that lengthy chapter...
9 More Discussions You Might Find Interesting
1. UNIX for Advanced & Expert Users
Hi all,
we have been noticed that few of processes failed with no reason.
when I chcked the log messages, I have got the following error messages writen to ktlog
just a few minutes before!
When the kernel writes such warning meassages?
will the system sends/generates any signals like... (1 Reply)
Discussion started by: reddyb
1 Replies
2. Filesystems, Disks and Memory
Hi all,
I am running 7.3 Redhat Linux, I have Oracle database running on it and I have some problem with the memory. every time I startup the database, the memory was peak up to 630M of Ram out 640M ram on the entire box and I didn't specify that much of memory on my database. Oracle advise me... (2 Replies)
Discussion started by: lapnguyen
2 Replies
3. UNIX for Dummies Questions & Answers
Received the following on the weekend.
Panic:k_trap kernel mode trap type 0X0000000e cannot dump 163739 pages to dumpdev hd (1/41):space for only 48640 pages
Dump not complete
Safe to power off or press any key to reboot
Any ideas how to fix this one? It also happen a month ago. (3 Replies)
Discussion started by: jcoleman544
3 Replies
4. Linux
Hello!
Does anyone know howto reload your kernal without rebooting the machine?
I´ve updated the "/etc/security/limits.conf " file and need to make this active without rebooting the machine. this is Red Hat Advanced Server 2.1AS
Regards...
dOzY (3 Replies)
Discussion started by: dozy
3 Replies
5. UNIX for Dummies Questions & Answers
Hi friends,
i would like to know the kernal information (semaphores), which command do i use
Thanks (4 Replies)
Discussion started by: shaan_dmp
4 Replies
6. Solaris
hi,
can anybody tell me how to increase the parameters like project.max-shm-ids on solaris10.
i have used prctl, but got reset while server reboot.
thnks and regards
Ajay (1 Reply)
Discussion started by: ajaysahoo
1 Replies
7. SCO
I am trying to restore Unixware 7.1 from a backup using RestoreEdge which is from Microline version 2. The restore is on another machine using the same RAID controller and TBU. We are retiring the other machine.
Anyway, we get to disk #2 and it panics. Here is the Pic.
Can anyone tell me... (13 Replies)
Discussion started by: ccd1977
13 Replies
8. Solaris
Hi,
Can you please let me know about kernal parameters?
Where we can see that kernal parameters?
we are using System = SunOS 5.10.
Please let me know commands to see these kernal parameters file or files. (1 Reply)
Discussion started by: kancherla.sree
1 Replies
9. Linux
Please clarify on below two which one is the latest Kernal version.
Also clarify me , how to break and understand, which part meant for what?
kernel-2.6.18-348.6.1.el5.i686.rpm
kernel-2.6.18-308.11.1.el5.i686.rpm
Thanks (3 Replies)
Discussion started by: Siva SQL
3 Replies
LEARN ABOUT SUNOS
drv_usectohz
drv_usectohz(9F) Kernel Functions for Drivers drv_usectohz(9F)
NAME
drv_usectohz - convert microseconds to clock ticks
SYNOPSIS
#include <sys/types.h>
#include <sys/ddi.h>
clock_t drv_usectohz(clock_t microsecs);
INTERFACE LEVEL
Architecture independent level 1 (DDI/DKI).
PARAMETERS
microsecs The number of microseconds to convert.
DESCRIPTION
drv_usectohz() converts a length of time expressed in microseconds to a number of system clock ticks. The time arguments to timeout(9F)
and delay(9F) are expressed in clock ticks.
drv_usectohz() is a portable interface for drivers to make calls to timeout(9F) and delay(9F) and remain binary compatible should the
driver object file be used on a system with a different clock speed (a different number of ticks in a second).
RETURN VALUES
The value returned is the number of system clock ticks equivalent to the microsecs argument. No error value is returned. If the clock
tick equivalent to microsecs is too large to be represented as a clock_t, then the maximum clock_t value will be returned.
CONTEXT
drv_usectohz() can be called from user or interrupt context.
SEE ALSO
delay(9F), drv_hztousec(9F), timeout(9F)
Writing Device Drivers
SunOS 5.10 12 Nov 1992 drv_usectohz(9F)