The Linux Process Manager [Paperback]

John O'Gorman (Author)

Book Description

Publication Date: February 13, 2003 | ISBN-10: 0470847719 | ISBN-13: 978-0470847718 | Edition: 1

• An ideal resource to understand the internals, line by line
• Looks at architecture dependent code for PC version
• Addresses process creation, timers, software and hardware interrupts, signals, tracing processes, and the virtual 8086 mode

Editorial Reviews

Review

"…lives up to Wiley’s ideal of providing timely, practical, and reliable texts…such an interesting read…a classic is born…" (Linux Format, August 2003)

“…there is no other book I would recommend over this one for the Linux programmer wanting to learn about this particular topic…” (M2 Best Books, October 03)

From the Back Cover

The Linux Process Manager is a line-by-line explanation of the subset of the Linux source code kernel which deals with process management, covering:

• process creation, scheduling and termination
• synchronisation, including locks and semaphores
• interrupts (hardware and software)
• signal handling

Complete in every way, this book describes every function and macro. All data structures used by the process manager are dealt with and their role in the overall picture explained along with the accompanying raw code.

This indispensable guide to the kernel subset is aimed at readers with some previous knowledge of operating system structure and terminology, as well as a minimum knowledge of the C programming language. The comprehensive index which lists every identifier, makes in-depth information instantly available to programmers, software engineers, system administrators and anyone who wants to modify the kernel, or wants to know about the implications of kernel configuration variables.

See all Editorial Reviews

Product Details

• Paperback: 640 pages
• Publisher: Wiley; 1 edition (February 13, 2003)
• Language: English
• ISBN-10: 0470847719
• ISBN-13: 978-0470847718
• Product Dimensions: 9.3 x 7.5 x 1.8 inches
• Shipping Weight: 3.4 pounds (View shipping rates and policies)
• Average Customer Review: 5.0 out of 5 stars  See all reviews (1 customer review)
• Amazon Best Sellers Rank: #2,028,434 in Books (See Top 100 in Books)

More About the Author

Discover books, learn about writers, read author blogs, and more.

Customer Reviews

Most Helpful Customer Reviews

2 of 3 people found the following review helpful:

5.0 out of 5 stars Very solid Linux commentary, November 11, 2009

By 

Yong Zhi (TX USA) - See all my reviews
(REAL NAME)   

Amazon Verified Purchase

This review is from: The Linux Process Manager (Paperback)

This is a surprisingly good book written by Irish Linux pioneer, it's basically a commentary, it comments every line of the source code literarily, the writing style s not entertaining, but not dry either, it's solid, quality stuff. Some thing goes like this:

[1] void __init fork_init(unsigned long mempages)

[2] max_threads = mempages /(THREAD_SIZE/PAGE_SIZE)/2;
76
[3] init_task.rlim[RLIMIT_NPROC].rlim_cur = max_threads/2;
[3] init_task.rlim[RLIMIT_NPROC].rlim_max = max_threads/2;
79 }

Figure 8.2: Initialisation routine
[1] the parameter passed is the total number of page frames present in the system.
[2] the literal constant THREAD_SIZE is defined in <asm-i386/processor.h> as 2 * PAGE_SIZE. It is the amount of kernel memory allocated to a process - thetask_struct plus the kernel stack. So THREAD_SIZE/PAGE_SIZE is the size of this in pages (2). Then mempages /(THREAD_SIZE/PAGE_SIZE) is the number of processes that could be created using the whole memory. As each process needs 2 pages, this is half the number of physical pages available. To allow room for code and data as well, the default maximum number of threads is set to half of this again, or a quarter of the number of physical pages.
[3] [3] the limits for the init_task are set at half of this again, or a maximum of one eighth of the number of physical pages.

This is Tony Cahill's comment about the author's background:

"His work involved constructing an interpreter for a proprietary language designed for developing and delivering Computer Aided Learning material, which generated Intermediate (what we would now call Byte) Code. John's research was essentially an exercise in systematic reverse engineering, and demonstrated that it was relatively easy to develop alternative interpreters, which could deliver the same lessons on other hardware, such as PCs."

The Irish Linux pioneer died couple years ago.