101 - System Architecture

LPI – 101
System Architecture
Why Linux?
Open Source
 Many people have been able to adapt Linux in
different ways
 This caused wide range of new features and
versions customized to meet new specific needs

Why Linux?
Multiuser and Multitasking
 Linux is designed to be multi-user operating
system.
 Each user can run multiple tasks at the same time,
without affecting each other.

Why Linux?
Tools
 Large number of tools available for a variety of
jobs

Why Linux?
Portability
 Linux Operating system is programmed to be
portable across different hardware.


Why Linux?
Other reasons ...
 Reliability, security and speed
 Critical businesses, database servers, data centers,
all run on Linux.

 Amazon, Google, Wikipedia are few
 Mature Operating System, long time
 Best choices for Developing Countries as there are
no or low cost associated

 It is used as a model for teaching Operating
systems “having the source code makes learning a

great experience”

101: system Architecuture
Key Knowledge Areas
 Enable and disable integrated devices
 Set the correct hardware ID for different devices,
the boot device

 Difference between coldplug, and hotplug
 Tools and utilities to list various hardware
information
 Tools and utilities to manipulate USB devices
 Conceptual understanding of sysfs, udev, hald, dbus
8
1. Determine and configure Hardware settings
Interrupt Requests
 IO addresses are regions of memory mapped to devices where the
CPU can write information of devices

 The limited IRQ lines become a problem as some devices were
hardwired to use specific IRQ.

 Plug and play have solved this limitation by sharing interrupt
lines to devices.
 Now with Intel’s Advanced Programmable Interrupt Controller
(APIC) architecture, there are now usually 24 interrupt lines

available that can accommodate upto 255 devices
9
1. Determine and configure Hardware settings ...
Interrupt Requests … contd.
 /proc/interrupts – directory that lists allocation of IRQs to
devices in Linux System

 /proc – It is a virtual file system, kernel sets up this file
system
 to export information about the running kernel user processes,
& hardware devices that have been configured
 Exist only in memory & doesnot represent any true physical
files
10
Input / Output Addresses.
 CPU and peripheral devices, communications make use of IO
addresses
 Two types of methods 1)Memory Mapped IO
 2) Port Mapped IO
 MMIO – regions of memory reserved for communication
 PMIO – CPU has separate set of instructions for performing IO
with devices having a separate address space

#less /proc/ioports – reveals allocation of IO ports
#less /proc/iomem – reveals allocation of IO memory
11
DMA Addresses.
 Enahancement /optimisation that allows devices to read/write
directly to memory w/o CPU involvement

#less /proc/dma – shows allocation of DMA addresses
 For most devices it is automatically configured by the OS
 PATA devices may not be automatically configured
#hdparm –d1 /dev/sda – enabling DMA address

#hdparm –d /dev/sda – see current setting on hard disk
12
Configuring IRQs, IO Ports/Addresses
 Configuration for device IRQ, IO addresses, DMA channels and
memory regions happens automatically with PnPs

 Via system BIOS, Linux Kernel & device drivers resource
conflicts are resolved for PnPs

 Linux provides utilities to query devices to find resources they
use,
 lspci, setpci } utilities that use information from /proc
#lspci -vv : provides verbose information on any devices using PCI

bus
#setpci : for configuring and querying pci devices
13
Linux Device Management Overview
 Device driver – a kernel space program that allow applications to
interact with hardware.

 Kernel space – privileged code having full access to hardware and runs
in ring 0
 User space applications – run in less privileged mode ring 3, and
cannot access hardware directly, calls kernel space in order to access

hardware on their behalf.
 Linux exports information about devices to /sys, virtual FS when PnP
device is inserted/removed /sys is updated, and udev daemon takes care

of this.
 udev also creates relevant entries under /dev FS
 D-Bus daemon then receive and post notification to Desktop

14
Linux Device Management Overview
15
Mass Storage Devices … contd
 Originally x86 machines hard disks only supported a max. of 4
partitions, i.e. partition table holds only 4 entry

 If more partition required, mark one of 4 partitions as extended
partition in order to create as many logical partitions as we want.

 MBR – is a special type of boot sector, at the beginning of
partitioned computer
 Boot sector – is region of hard disk that contains machine code to
be loaded to RAM, allow boot process of a computer to load

programs.
16
Mass Storage Devices
 PATA – Parallel Advanced Technology Attachment
 Linux identifies it as /dev/hd[a-z]
 /dev/hda – master drive on primary connector
 /dev/hdc – master drive on secondary connector

 /dev/hda1 – first primary partition on hda
 /dev/hda2 – second primary partition on hda
 Hard disks have their data accessible randomly and are called block
devices
 Linux kernels’ device-mapper framework allows for creation of virtual
block devices on which FS is created.

 Virtual Block Devices allow creation of encrypted partition as well as
RAID, w/o need for special hardware

101: system Architecture
1. Determine and configure Hardware setting …
 http://wiki.osdev.org/MBR_(x86)#Building_a_Custom_MB
R_Bootstrap_program
18
USB Support
 Hot plug interface
 USB support is present from Linux kernel 2.2.7
 #lsusb – command to list the USB devices
Device Drivers
 Kernel uses drivers to export device info to /sys file system
 In most Linux distributions, drivers are provided as part of the kernel, can
be loaded/unloaded dynamically without the need to recompile the Kernel.

 #lsmod – lists currently installed modules
 #modinfo – query a module for dependency, author
 #insmod – install a module, by providing path to driver
 #modprob – installs a module via module name, handles deps
 #rmmod – Remove
19
1. Boot the System
Introduction
 BIOS, CPU goes to address of BIOS and loads it
 BIOS loads MBR of Boot device

 MBR contains bootstrap program, written in Assembly
 MBR has partition table portion
 MBR format
 The partition has 4 entries, order doesn’t matter
 There can be one active partition at most
 Partition table (PT) conforms to a standard layout, that is
independent of the Operating System

 Each PT is 16-byte long, max of 4 entries
20
1. Boot the System … contd
Dual Booting
 A huge improvement over the standard MBR
 Standard x86 boot sequence only boot the MBR from the
first disk found during disk detection phase, as well as

allow a single active partition on that disk
 The standard MBR is replaced with dual booting MBR.
 Now, one MBR can load and run a different MBR of
different drive, “Chain Loading”

22
Boot process
1. BIOS loads MBR, the first stage boot loader
2. First boot loader loads second stage boot loader, this stage gives
flexibility in selecting the Kernel
3. Kernel takes over the control
 The kernel starts and configures the CPU type, interrupts …
 Kernel loads initial RAM Disk (initrd), way to access system
hardware where root partition is located.

 as root partition is loaded, initrd swaps out
4. Kernel becomes fully operational, then loads /sbin/init, which sets
up user space, login shell, graphical logins …

23
2nd stage boot loader,
 Gives a lot of options that help us override some
parameters
e.g. init – overrides the process that is run by the kernel after
it has finished loading,
“init=/bin/bash” – is used to by pass login prompt in
case password is forgotten.
24
Init Process overview
 After loading the kernel, init process is started,
 Init is responsible for checking and mounting file systems,
and starting up configured services.

 Init starts services after entering default run level
e.g. default run level looks like this (in /etc/inittab)

id : 3 : initdefault:
 Where, id= a 4 character name, 3=run level,
initdefault=action
25
Actions
 wait – init will wait till process starting is completed
 respawn – init restarts even when process terminates, usual used
for getty processes

 ** getty, short for "get teletype", is a Unix program running on a
host computer that manages physical or virtual terminals (TTYs).

When it detects a connection, it prompts for a username and runs
the 'login' program to authenticate the user. It is usually called by
init.
 Ctrl-Alt-Del – traps key combination
 powerfail, pwerokawait – used to respond to notifications to UPS

26
The /etc/inittab
 You can configure your system to easily capture the Alt-Ctrl-Del
key, by adding a line to /etc/inittab
Inittab taken from fedora

27
The /etc/inittab … contd
28
The /etc/inittab .. contd
As usual, lines starting with # are comments. Other lines have several
fields with the following format:
 id:runlevels:action:process
 id is a unique identifier of one to four characters. Older versions
limited this to two characters, so you will often see only two characters
used.
 runlevels lists the runlevels for which the action for this id should be
taken. If no runlevels are listed, do the action for all runlevels.

 action describes which of several possible actions should be taken
 process tells which process, if any, should be run when the action on
this line is performed.

29
The /etc/inittab .. contd
 Initialization scripts
 In this example, init will run the /etc/rc.d/rc script (or

command) with the parameter of 5 whenever runlevel 5 is
entered.
 init will wait until this command completes before doing
anything else.
30
Troubleshooting Boot Process
dmesg
 Contents of ring buffer (memory location) can be read with dmesg tool
 It can also tell us screen outputs during bootup
/var/log/syslog or /var/log/messages
 These two files are additional locations where we have to look at to
examine what happened during boot process

Failure to boot
 “Root device cannot be found, ctrl + D to perform maintenance”
 “Kernel Panic” – Incorrectly configured initrd image or missing root
device
 Maintain through Live CD, with out mounting Root file system
31
3. Runlevels, Rebooting, Shutting Down System
Introduction
 Linux has 7 run-levels [0 – 6]
 Run-levels are assigned to the following configurations
 1) single user mode
 2) multi-user, text mode, no servers
 3) multi-user, text mode, servers
 4) multi-user, GUI mode, no servers
 5) multi-user, GUI mode, servers.
32
Introduction … contd
 Run levels on Red hat server
 # Default runlevel. The runlevels used by RHS are:
 # 0 - halt (Do NOT set initdefault to this)
 # 1 - Single user mode
 # 2 - Multiuser, without NFS (The same as 3, if you do not have
networking)
 # 3 - Full multiuser mode
 # 4 - unused
 # 5 - X11
33
Introduction … contd
 Run levels specific scripts are stored in /etc/init.d/rc?.d
or /etc/rc?.d/, where ? Represents run levels [0 – 6]

 Original scripts stored in /etc/rc.d, /etc/init.d or
/etc/rc.d/init.d
 Symbolic links: must have S – for start at run level entry
K – killed at run level

entry
34
Debian based systems
#update-rc.d disable 3 apache2
#update-rc.d enable 3 apache2

RPM based systems Remove apache service from rc3
#chkconfig –del httpd
#chkconfig –add httpd Add service: apache2 to rc3
Or
#chkconfig –level 3 httpd on
35
Changing run levels
 Single-user mode
 Linux is inherently a multiuser system.
 However, there are times when that can be a problem, such as when you need
to recover a major filesystem or database, or install and test some new

hardware.
 Runlevel 1, or single-user mode, is your answer for these situations. As it start
in a shell with only a minimal system.

 Usually there will be no networking and no (or very few) daemons running. On
some systems, you must authenticate by logging in, but on others you go
straight into a shell prompt as root.
 Single-user mode can be a lifesaver, but you can also destroy your system, so
always be careful whenever you are running with root authority. Reboot to
normal multiuser mode as soon as you are done.
36
Changing run levels
 Single-user mode
 As with switching to regular multiuser runlevels, you can also switch to
single-user mode using telinit 1

#telinit 1
Clean shutdown
 telinit or init to stop multiuser activity and switch to single-user mode, this
can be rather abrupt and cause users to lose work and processes to
terminate abnormally.
 The preferred method to shut down or reboot the system is to use the
shutdown command,
 Shutdown first sends a warning message to all logged-in users and blocks
any further logins. It then signals init to switch runlevels.

37
Clean shutdown
 The init process then sends all running processes a SIGTERM signal,
(process termination signal)

 giving them a chance to save data or otherwise properly terminate.
After 5 seconds, or another delay if specified, init sends a SIGKILL

signal to forcibly end each remaining process.
 By default, shutdown switches to runlevel 1 (single-user mode).
 You may specify the -h option to halt the system, or the -r option to
reboot.
 A standard message is issued in addition to any message you specify.
The time may be specified as an absolute time in hh:mm format

38
Clean shutdown
Thus,
#shutdown –h } just shutdown the system
#shutdown –h +10 } shutdown in 10 minutes
#shutdown –r 14:30 } reboot at 14:30
#shutdown –c } cancel any schedules shutdown process
39
Halt, reboot, and poweroff
 The halt command halts the system, system temporary suspension
 The poweroff command is a symbolic link to the halt command,
which halts the system and then attempts to power it off.

 The reboot command is another symbolic link to the halt
command, which halts the system and then reboots it.

 For additional options that you may use with these commands, as
well as more detailed information on their operation, consult the

man page.
40
Starting | stopping services
Debian based
#/etc/init.d/apache2 stop|start|restart|reload
RPM based
#service httpd start|stop|reload

101 - System Architecture

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

101 - System Architecture

Uploaded by

Copyright:

Available Formats

LPI – 101

versions customized to meet new specific needs

without affecting each other.

portable across different hardware.

all run on Linux.

no or low cost associated

systems “having the source code makes learning a

the boot device

CPU can write information of devices

hardwired to use specific IRQ.

(APIC) architecture, there are now usually 24 interrupt lines

devices in Linux System

with devices having a separate address space

directly to memory w/o CPU involvement

#hdparm –d1 /dev/sda – enabling DMA address

memory regions happens automatically with PnPs

conflicts are resolved for PnPs

#lspci -vv : provides verbose information on any devices using PCI

interact with hardware.

cannot access hardware directly, calls kernel space in order to access

device is inserted/removed /sys is updated, and udev daemon takes care

 D-Bus daemon then receive and post notification to Desktop

partitions, i.e. partition table holds only 4 entry

partition in order to create as many logical partitions as we want.

be loaded to RAM, allow boot process of a computer to load

 Linux identifies it as /dev/hd[a-z]

 /dev/hda – master drive on primary connector

 /dev/hdc – master drive on secondary connector

block devices on which FS is created.

RAID, w/o need for special hardware

be loaded/unloaded dynamically without the need to recompile the Kernel.

 BIOS loads MBR of Boot device

independent of the Operating System

first disk found during disk detection phase, as well as

different drive, “Chain Loading”

hardware where root partition is located.

4. Kernel becomes fully operational, then loads /sbin/init, which sets

up user space, login shell, graphical logins …

and starting up configured services.

e.g. default run level looks like this (in /etc/inittab)

for getty processes

host computer that manages physical or virtual terminals (TTYs).

 powerfail, pwerokawait – used to respond to notifications to UPS

key, by adding a line to /etc/inittab

Inittab taken from fedora

 id is a unique identifier of one to four characters. Older versions

taken. If no runlevels are listed, do the action for all runlevels.

this line is performed.

 In this example, init will run the /etc/rc.d/rc script (or

examine what happened during boot process

or /etc/rc?.d/, where ? Represents run levels [0 – 6]

K – killed at run level

#update-rc.d enable 3 apache2

to recover a major filesystem or database, or install and test some new

in a shell with only a minimal system.

single-user mode using telinit 1

any further logins. It then signals init to switch runlevels.

(process termination signal)

After 5 seconds, or another delay if specified, init sends a SIGKILL

The time may be specified as an absolute time in hh:mm format

which halts the system and then attempts to power it off.

command, which halts the system and then reboots it.

well as more detailed information on their operation, consult the

You might also like