Using Linux in a Windows world

Interacting with Windows resources after migrating to Linux

Linux is the operating system of choice for most servers on the Internet and is growing in popularity as a desktop operating system. However, simply migrating to Linux on your desktop doesn't guarantee that you can still interact with all of the enterprise resources that your job or interests may require. This article discusses additional configuration tasks that you may need to undertake to interact with legacy Microsoft Windows files and file servers and use new network hardware from your desktop Linux system.

Share:

William von Hagen, Systems Administrator, Writer, WordSmiths

William von Hagen has been a writer and UNIX systems administrator for more than 20 years and a Linux advocate since 1993. Bill is the author or co-author of books on subjects such as Ubuntu Linux, Xen virtualization, the GNU Compiler Collection (GCC), SUSE Linux, Mac OS X, Linux file systems, and SGML. He has also written numerous articles for the Linux and Mac OS X publications and websites.


developerWorks Contributing author
        level

03 April 2013

Also available in Chinese Russian Japanese

Every year, Linux journalists and fans openly ponder whether the new year will be the year of the "Linux desktop." Although it's easier than ever to buy a new machine that runs Linux out of the box or migrate an existing Windows system to Linux, simply installing and booting Linux isn't the final step in making it your daily driver operating system. You may have switched to using Linux on your desktop, but the environment in which you use your new Linux system may not be as modern. Most enterprise, small-to-medium business, and small office/home office environments have Windows-related requirements you must consider. You may also need to use hardware with your Linux system that is only actively supported on a Windows platforms. Luckily, a variety of solutions to these issues exist.

Accessing shared Windows file systems

Access to shared Windows file systems is probably the most common business requirement for Linux system users. Many businesses still use Windows as the core of the infrastructure to support enterprise mail and calendaring systems, such as Microsoft Exchange Server, and for simple file and data sharing through exported file systems, generically known as Windows shares.

Linux provides multiple ways to access Windows shares. The mechanism you choose depends on the frequency with which you need to access data on a Windows share and whether the Linux system you are configuring is a personal, single-user system or one that multiple users share.

Accessing Windows shares from the command line

The Samba Project (see Resources) is an open source project that enables Linux, UNIX, and other UNIX-like operating systems to interoperate with Windows. Samba is available in the repositories for most Linux distributions and is generally installed and configured to enable a Linux system to join and interoperate with an existing Windows workgroup or domain, often serving as a Common Internet File System (CIFS) file server that Windows systems can use. Samba 4.x installations also let a Samba server act as an Active Directory Domain Services controller and integrated Domain Name System for a Windows domain.

Although a Samba server provides an excellent mechanism for making Linux resources available from Windows systems and enabling Linux systems to print to Windows printers, installing and configuring a Samba server is overkill if all you are interested in is occasionally retrieving files from a Windows share. The Samba client package (samba-client) contains several utilities designed for such occasional interaction, including the smbclient application, which provides an interactive command-line interface (CLI) to remote Windows shares.

If you're not sure of the names of the shares available on a remote Windows server, you can use the smbclient application to query a server and list the resources available on it by typing the -L option followed by the name or IP address of the server, as shown in Listing 1.

Listing 1. Using smbclient to list Windows server resources
$ smbclient -L win2008server
Enter wvh's password:
Domain=[WORDSMITHS]
OS=[Windows (R) Small Business Server 2008 6001 Service Pack 1]
Server=[Windows (R) Small Business Server 2008 6.0]

Sharename        Type      Comment
---------        ----      -------
Address           Disk      "Access to address objects"
ADMIN$            Disk      Remote Admin
Brother HL-2070N  Printer   Brother HL-2070N
C$                Disk      Default share
ExchangeOAB       Disk      OAB Distribution share
IPC$              IPC       Remote IPC
NETLOGON          Disk      Logon server share
print$            Disk      Printer Drivers
Public            Disk
RedirectedFolders Disk
Resources$        Disk      "Event logging files"
SYSVOL            Disk      Logon server share
[...]

As shown, the smbclient application uses the current Linux user as the user to authenticate as on the remote Windows server and prompts for that user's password to access the remote server. If your login is different on the Windows server, you can use the -U option to specify the login to use.

After you have identified the name of the share you want to connect to, you can use the smbclient application to connect to a specific share to examine or retrieve files, as shown in Listing 2.

Listing 2. Using smbclient to retrieve files from a Windows share
$ smbclient //win2008server/Public
Enter wvh's password:
Domain=[WORDSMITHS]
OS=[Windows (R) Small Business Server 2008 6001 Service Pack 1]
Server=[Windows (R) Small Business Server 2008 6.0]
smb: \> cd Software\Linux
smb: \Software\Linux\> ls
  .                            D        0  Sun Jan 27 23:40:42 2013
  ..                           D        0  Sun Jan 27 23:40:42 2013
  acl82express-linux-x86.bz2   A 44960643  Sun Jan 27 23:40:24 2013
  alien_8.87.tar.gz            A    64336  Sun Jan 27 23:39:24 2013
  crossover-12.1.0-1.i386.rpm  A 47233108  Sun Jan 27 23:39:14 2013
  [...]
		59997 blocks of size 1048576. 32474 blocks available
smb: \Software\Linux\> get softmaker-office-2012-674.x86_64.rpm
getting file \Software\Linux\softmaker-office-2012-674.x86_64.rpm \
        of size 147058656 as softmaker-office-2012-674.x86_64.rpm \
        (16345.5 KiloBytes/sec) (average 16345.5 KiloBytes/sec)
smb: \Software\Linux\> quit

The example in Listing 2 shows that smbclient provides a CLI to the remote server. This interface supports familiar Linux commands such as cd (for moving around in the directories on the remote server) and ls (for listing the contents of a given directory). When you have located a file that you want to retrieve, you can use the get command to retrieve that file, storing it in the directory from which you executed smbclient. To store a retrieved file in another directory, you can use the lcd (local change directory) command to change the directory that smbclient views as the current directory.

Using the smbclient application is fine if you need only occasional access to one or more files in a Windows share and don't care about using a file manager such as Gigolo, GNOME Nautilus, KDE Dolphin, Konqueror, or Thunar to graphically browse a Windows share. The smbclient application makes a remote share available within the context of that application, not to your system as a whole. To browse a Windows share and make it available to the applications of your choice, you will have to mount that share on your system in the same way local file systems are mounted.

Permanently mounting Windows shares

You can manually mount a Windows share on a directory on your system by using the mount command as the root user or by using the sudo application, as shown in Listing 3. The mount command's -t option and its cifs argument specify the type of file system you are mounting and, in this case, causes the generic mount command to invoke the /bin/mount.cifs command. Use your system's package-management commands to install the cifs-utils package, which is part of the Samba suite.

Listing 3. Mounting and listing a manually mounted Windows share
# mkdir /mnt/PUBLIC
# mount -t cifs //winserver2008/Public /mnt/PUBLIC
Password:
# ls -l /mnt/PUBLIC
total 1
drwxr-xr-x. 1 root root   0 Jan 27 16:19 Desktop
-rwxr-xr-x. 1 root root 174 Jan 19  2008 desktop.ini
drwxr-xr-x. 1 root root   0 Jan 19  2008 Documents
drwxr-xr-x. 1 root root   0 Jan 27 23:38 Downloads
drwxr-xr-x. 1 root root   0 Jan 19  2008 Favorites
drwxr-xr-x. 1 root root   0 Jan 19  2008 Music
drwxr-xr-x. 1 root root   0 Jan 19  2008 Pictures
drwxr-xr-x. 1 root root   0 Jan 27 23:37 Software
drwxr-xr-x. 1 root root   0 Jan 19  2008 Videos

The example mount command in Listing 3 makes the remote share available through the /mnt/PUBLIC directory, but has a few shortcomings:

  • You cannot simply add this command to your /etc/fstab file to automatically mount the share because you are prompted for a password after executing the mount command. Unless you identify a specific user to authenticate as, you must supply the remote Windows password for the user you executed the su or sudo commands as.
  • Executing the mount command as the root user mounts the share as being owned by the root user. Although mounting a share in this fashion is fine if you only want to read files in that share, you can only write to the mounted share as the root user, which is usually inconvenient.

To resolve these problems, you can specify special, CIFS-related options when executing the mount command to mount a CIFS share. These options are:

  • username=<login> - Specifies the user that you will use to authenticate to the remote share. You can abbreviate this option to user=.
  • password=<login> - Specifies the password for the user who will authenticate to the remote share (the user you specified as the user that you will use).
  • uid=<UID> - Specifies the user ID (UID) of the user who will own the files and directories in the mounted share. If you are using password file authentication on your Linux desktop, find your UID by using the grep command to search for your login in the system's password file, as shown in the following example.
    $ grep wvh /etc/passwd
    wvh:x:500:500:William von Hagen:/home/wvh:/bin/bash

    The UID is the third field in a password file entry — in this case, efficiently500.

  • gid=<GID> - Specifies the group ID (GID) of the user who will own the files and directories in the mounted share. If you are using password file authentication on your Linux desktop, find your GID by using the grep command to search for your login in the system's password file. The GID is the fourth field in a password file entry, as shown in the example in the previous bullet.

An example mount command using these options, broken across two lines for formatting purposes, is as follows:

# mount -t cifs -o username=wvh,password=MYPASSWORD,uid=500,gid=500 \
     //win2008server/Public /mnt/PUBLIC

You can use a command like this one from the CLI or add a similar entry to your /etc/fstab file to automatically mount the file system whenever you boot your system. The following is an equivalent /etc/fstab file entry to the previous mount:

//win2008server/Public /mnt/PUBLIC cifs \
    username=wvh,password=MYPASSWORD,uid=500,gid=500 0 0

This command would appear as a single line in your /etc/fstab file and is broken across lines in this example for formatting purposes. Unfortunately, this command has some obvious security problems because you must expose your password in the /etc/fstab entry and the file /etc/fstab is typically readable by every user on the system. To resolve this problem, put your user name and password in a text file that you identify using the credentials mount option. For example:

//win2008server/Public /mnt/PUBLIC cifs \
    credentials=/etc/cred.wvh,uid=500,gid=500 0 0

The format of a credentials file is as follows:

username=wvh
password=MYPASSWORD

You can then make the /etc/cred.wvh file readable only by the root user (as whom the mount command must be executed) using the following command:

# chmod 600 /etc/cred.wvh

Using the credentials mount option provides a nice balance between having to specify a user name and password each time you mount a file system and not exposing that information to everyone who has access to your system.

Mounting Windows shares on demand using sudo

If you only want to occasionally mount a share, you could add the noauto option to the list of mount options shown in the examples in the previous section. This option prevents the specified Windows share from being mounted automatically each time you restart your system. You would then have to mount the share from the command line whenever you wanted to access files in that share.

You must always execute the mount command as the root user, which usually means that anyone who wants to mount that share must know the root password on the system. If you do not want to automatically mount the share each time you restart your system and want to limit the people who need to know the root password (or simply limit the number of times you have to use that command), you can enable a regular user to mount the share using the sudo command. To do so, use the visudo command to edit the /etc/sudoers file, adding an entry like the following to the end of the /etc/sudoers file.

wvh  ALL=/bin/mount /mnt/PUBLIC, /bin/umount /mnt/PUBLIC

This entry enables the user wvh to execute the /bin/mount /mnt/PUBLIC and /bin/umount /mnt/PUBLIC commands using the sudo command but does not enable that user to use the mount or umount commands to mount or unmount any other file systems. If you want all users on a given system to be able to mount and umount the /mnt/PUBLIC share, you could replace the user name wvh with the generic %users expression. In that case, make sure that the files that users might want to update are writable by the group identified by the anongid identified in the associated /etc/fstab entry and that all users belong to this group.

Note: Another approach to mounting Windows shares is to use one of the File system in User Space (FUSE) packages that support CIFS file systems. The smbnetfs and fusesmb packages are fairly popular but have not been updated in a while and require substantial configuration to make them work properly. (See Resources for links to these projects.)


Working with Windows text files on Linux systems

Most people think of text files as a standard type of file that is the same on every computer system and can therefore be safely used for configuration files, quick notes that you can read on any operating system, and so on. Unfortunately, that's not the way it is. Although the same character sets are used for text files on Windows and Linux systems, the characters used to identify the end of a line of text differ. Windows and DOS systems mark the end of a line of text with two characters: a carriage return (Ctrl-M) and a line feed (Ctrl-J). Linux and all UNIX-like systems mark the end of a line of text with a single character: the line feed character. If you use a Linux utility to open a text file that was created on a Windows system, you will see a Ctrl-M character at the end of each line, often displayed as ^M or \r. This extra character can confuse Linux utilities that need to read such files.

Linux provides two commands that enable you to automatically change the end-of-line characters in a text file. The dos2unix command, which is available for most Linux distributions in a package by the same name, automatically converts text files that were created in Windows (or DOS) to use the Linux and UNIX conventions to mark the end of a line of text, stripping out all of the Ctrl-M characters. An equivalent program, unix2dos, is available in a package of the same name to convert Linux and UNIX text files to Windows-format text files.

If you are a fan of the Emacs text editor, you can add the Emacs commands shown in Listing 4 to your Emacs configuration file (~/.emacs). You will then be able to execute the dos-file, mac-file, and unix-file commands to convert the contexts of a text buffer to the text file format specified in the name of the command.

Listing 4. Emacs commands for changing text file types
;
; Functions for changing buffer modes
;
(defun dos-file ()
   "Change the current buffer to Latin 1 with DOS line-ends."
   (interactive)
   (set-buffer-file-coding-system 'iso-latin-1-dos t))

(defun mac-file ()
   "Change the current buffer to Latin 1 with Mac line-ends."
   (interactive)
   (set-buffer-file-coding-system 'iso-latin-1-mac t))

(defun unix-file ()
   "Change the current buffer to Latin 1 with Unix line-ends."
   (interactive)
   (set-buffer-file-coding-system 'iso-latin-1-unix t))

Note: The sample Emacs commands shown in Listing 4 assume that you are using the Latin 1 character set. If you are using another character set in your text files, you will need to modify these commands to reflect that character set.


Using network devices that were designed for Windows

Although Linux provides the software that most people need to use on their computers, keeping up with the latest network hardware is a problem for any operating system. Modern Linux distributions include device drivers for an impressive range of network hardware, but Linux drivers for new network protocols and related hardware can lag behind, especially in the wireless space. If your Linux machine cannot recognize or use a new network device when the system boots, the drivers for that device are not available or some other resource that the device drivers require, such as hardware-specific firmware, was not available. In such cases, your friendly neighborhood search engine should always be your first resort: Someone else with the same newly released hardware may already have gotten it working on a Linux system. (See Resources for links to multiple sites that provide information about hardware that Linux supports and people's experiences with that hardware.)

If you can't find built-in Linux drivers for a newly released piece of network hardware, most new hardware ships with a CD that contains special drivers that enable supported operating systems to efficiently use that hardware. Unfortunately, these driver CDs rarely include drivers that are specifically designed for Linux systems. In some cases, Linux drivers for specific hardware are available from the hardware vendor's website. If this is the case, the vendor usually provides installation and configuration instructions as well. Many hardware vendors also provide online forums on which you can ask questions and discuss any installation or configuration problems you're having.

If the Linux distribution that you're using or experimenting with doesn't include support for a specific piece of network hardware that is in your desktop or laptop computer and the vendor does not provide Linux drivers, all is not lost. The NDISwrapper package enables you to use most Windows XP network device drivers on your Linux system. The Network Driver Interface Specification (NDIS) is a standard application programming interface for network interface cards (NICs). The NDISwrapper package does not support NDIS, Version 6, which is the driver format that Windows 8, Windows 7, and Windows Vista use, so you will only be able to use the NDISwrapper package with Windows XP drivers. If you are using a 64-bit Linux system, you can only use 64-bit Windows XP drivers; 32-bit Windows XP drivers will not work on a 64-bit system.

After installing the ndiswrapper-utils package (which may also install the ndiswrapper-common package on some Linux distributions) on your system, you can install and use a Windows XP NDIS driver by performing the following steps:

  1. Check whether any Linux driver for the same hardware is already loaded, and, if so, unload it using the rmmod DRIVER command.
  2. Locate the .inf and .sys files for the driver you want to use.

    If the Windows XP drivers for your hardware are provided in a self-extracting .exe file, you can usually extract them using the Linux unzip utility. If they are provided in a .cab file, you can extract them using the Linux cabextract utility. If they are provided in InstallShield files, you can extract them using the unshield utility. (All of these utilities are in packages by the same name in the repositories for most Linux distributions.) Worst case, you can install the drivers on a Windows machine, and then copy them to your Linux system.

  3. Install the driver using a command like the following as the root user or by using the sudo command:
    ndiswrapper -i DRIVER.inf
  4. Verify that the Windows driver is installed correctly by running the following command as the root user or by using the sudo command:
    ndiswrapper -l
  5. Load the NDISwrapper loadable kernel module by running the following command as the root user or by using the sudo command:
    modprobe ndiswrapper
  6. Create a module configuration file alias for your network interface by running the following command as the root user or by using the sudo command:
    ndiswrapper -m

    This entry will look something like the following:

    alias wlan0 ndiswrapper

    Make sure the name of the Ethernet interface with which the ndiswrapper kernel module is associated matches the interface that you want to use it with, as shown in the output of a command like ifconfig -a.

At this point, your new network hardware should be visible from your Linux system, and you should be able to use standard Linux network configuration utilities such as /sbin/ifconfig or configure and use the associated NIC.

If you want to use the ndiswrapper kernel module and a Windows driver for hardware that also has a native Linux driver, you will want to create a file called something like /etc/modprobe.d/blacklist-DRIVER.conf that contains an entry like blacklist DRIVER to ensure that the Linux driver is not loaded the next time you restart your system.

If you prefer the simplicity of a graphical application, multiple options are available for the NDISwrapper package. Figure 1 shows the ndisgtk application, which you can install on your system from the package of the same name.

Figure 1. Adding a Windows driver using ndisgtk
Image showing how to add a Windows driver using ndisgtk. Buttons include 'Install New Driver,' 'Remove Driver,' 'Configure Network,' and 'Close.'

The NDISwrapper package is both useful and impressive. Even if Linux provides native drivers for your hardware, you may find that using the Windows drivers can provide better performance than using the built-in Linux equivalents.


Conclusion

Linux is a powerful operating system that is free and — nowadays — also easy to use. Using Linux on your desktop or portable computer can provide substantially better performance than operating systems such as Windows and is also a great way to give new life to older computers. However, interoperating with Windows-based servers, certain types of Windows files, or Windows-oriented network hardware is still a requirement in many business and home environments. Linux provides a wide selection of software and techniques that enable you to interoperate with such systems and hardware, giving you the best of both worlds from your desktop Linux system.

Resources

Learn

Get products and technologies

  • Several graphical file managers exist for Linux and UNIX environments. Examples mentioned in this article are:
    • Gigolo. This graphical application simplifies managing and browsing local and remote file systems and was originally developed for the Xfce desktop environment. Based on the GIMP Toolkit (GTK), Gigolo can be used on all Linux distributions that use the GNOME desktop or have otherwise installed GTK.
    • Nautilus. This is the default graphical file manager on computer systems that use the GNOME desktop.
    • Dolphin. This is the default graphical file manager on Linux systems that use the KDE desktop.
    • Konqueror. This is an extremely powerful graphical file manager for Linux systems that use the KDE desktop. Konqueror was the default KDE file manager for many years and supports many capabilities that are not provided in other file managers for the KDE desktop.
    • Thunar. This graphical file manager for Linux systems was originally developed for the Xfce desktop environment.
  • dos2unix and unix2dos are applications that convert text files between the formats used on Linux and UNIX, DOS, and Macintosh systems.
  • Emacs is the one true text editor for modern computer systems.
  • Ndisgtk is a graphical application for interacting with the command-line NDISwrapper application.
  • FUSE implementations that enable users to mount Windows shares without requiring root privileges are the fusesmb and smbnetfs projects. Both of these require the libsmbclient library, which is included in the samba-client package. Neither of these have been updated for a while, which either indicates that they "just work" or that they are no longer actively used.
  • See the NDISwrapper Project page on SourceForge to get the latest version of NDISwrapper or detailed information about using the NDISwrapper package.
  • See the Mono Project site for information about using Mono to run Microsoft .NET code on a Linux system.
  • The Wine Project offers software that lets you run Windows software on your Linux computer. To see whether the software you're interested will run under Wine, check out the Wine AppDB.
  • See CodeWeavers for a commercially supported version of Wine. See their What Runs? page for information about Windows software that is known to work with CrossOver.
  • Consider running Windows in a virtual machine using VirtualBox.

Discuss

  • Get involved in the My developerWorks community. Connect with other developerWorks users while exploring the developer-driven blogs, forums, groups, and wikis.

Comments

developerWorks: Sign in

Required fields are indicated with an asterisk (*).


Need an IBM ID?
Forgot your IBM ID?


Forgot your password?
Change your password

By clicking Submit, you agree to the developerWorks terms of use.

 


The first time you sign into developerWorks, a profile is created for you. Information in your profile (your name, country/region, and company name) is displayed to the public and will accompany any content you post, unless you opt to hide your company name. You may update your IBM account at any time.

All information submitted is secure.

Choose your display name



The first time you sign in to developerWorks, a profile is created for you, so you need to choose a display name. Your display name accompanies the content you post on developerWorks.

Please choose a display name between 3-31 characters. Your display name must be unique in the developerWorks community and should not be your email address for privacy reasons.

Required fields are indicated with an asterisk (*).

(Must be between 3 – 31 characters.)

By clicking Submit, you agree to the developerWorks terms of use.

 


All information submitted is secure.

Dig deeper into Linux on developerWorks


static.content.url=http://www.ibm.com/developerworks/js/artrating/
SITE_ID=1
Zone=Linux
ArticleID=863545
ArticleTitle=Using Linux in a Windows world
publish-date=04032013