You're receiving this e-mail as a friend, client,
or associate of Netrino, LLC.
Please confirm your interest in receiving this
newsletter. You may unsubscribe if you no longer wish to receive our emails. |
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Firmware Update
- October 30, 2009
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In this issue:
-- Is Linux Too Bloated for Embedded Systems Use?
-- Bitmasks vs. Bitfields: Which is Better?
-- Register Now for Final Hands-on Training of 2009
-- Industry News and Other Fun Stuff -- Bring Embedded Software Boot Camp to Your City
Firmware Update is a free newsletter containing analysis, insight, and commentary on firmware architecture and process by embedded software expert Michael Barr. Firmware Update is Copyright 2009 by Netrino, LLC, but may be reprinted or forwarded for interested colleagues. |
|
Is Linux Too Bloated for Embedded Systems Use? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Linus Torvalds was quoted last month saying
the
Linux kernel he created is "huge and scary
now"
rather than the "streamlined, hyper-efficient
kernel I
envisioned." Or as Jack Ganssle more
succinctly put
it in a recent column at Embedded.com, "Linux
Sucks".
Never one to shy away from a fight, embedded Linux specialist Bill Gatliff comes to the open source software's defense. Gatliff says if you can look past the 11 million lines of kernel source code and list of its 5,000 contributors, you'll find that the "performance of the Linux kernel itself has gotten measurably better with each kernel release." |
|
Bitmasks vs. Bitfields: Which is Better? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Dan S.: I enjoy your newsletters, there are
very interesting links and concepts contained
therein. However, I have personal heart-burn
over your recent promotion of the free binary
literal macros.
The main thing I have a problem with is the bit-shift operations in-line. I have 14+ years embedded experience, and I work with bit- mapped data structures daily. I have always vastly preferred to implement bit-mapped data structures using C's bitfields, inside structs. I know that this data structure methodology is viewed as less "portable" in C, given the compiler's right to choose the LSB within the register. However, I think that the code is much easier to write, read, and debug using bit-mapped data structures. Can you explain to me why the bit-shift methodology seems to be the "industry standard"? What else does it have going for it that I don't understand? Michael Barr: Bitfields and bitmasks (i.e., hex literals whether named or not) have their individual pros and cons. Specifically, bitfields (once layed out) are easier to use properly time after time without coding errors. Bitmasks, however, are usually faster for multi-field register initializations (e.g., my_reg = FIELD1_8 | FIELD2_N | FIELD3_1;) and never slower than bitfields for any specific field access. Unfortunately, I don't think that picking one vs. the other is winnable across all situations or arguments in embedded programming. |
|
Register Now for Final Hands-on Training of 2009 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Precious few early registration days and fewer
seats
remain for the final Embedded Software Boot
Camp
of 2009. This hands-on training will be in
Maryland
the week of November 16-20 and I'll be the
instructor. I hope to meet you there.
|
|
Industry News and Other Fun Stuff ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Open Virtual Platforms offers free simulation
models of ARM
processor cores. http://bit.ly/2TiW8C
IEEE-USA announces training for engineers interested in gaining political office. Details: http://bit.ly/ToELg Why do FPGA startups keep failing? Olivier Coudert explains: http://bit.ly/1spfgk Users of static analysis tools often fail to fix detected defects. Study: http://bit.ly/4rSs0n New Electronics interviews Green Hills VP Chris Smith about the recent Embedded Systems Conference UK. http://bit.ly/2Y3W4F Hey, You, Engineer: Want success? Build crappier products! http://bit.ly/1mhxVR Intel Processors Then, Now, & Tomorrow. Busicom calculator, Microma digital watch, and more. Photos: http://bit.ly/3yv04n Excellent article by Jon Titus on whether the 80x86 CPU architecture will survive. http://bit.ly/1YYQ2C Engineers Without Borders are changing the world. Can you help? http://bit.ly/2v0NBf Group aims to create a national science fair (i.e., Woodstock of engineering) in Washington DC in the fall of 2010. http://bit.ly/QHwk0 My recent webinar Keeping Bugs Out of Embedded Software with Coding Standards is now available anytime on-demand. http://bit.ly/2C1JCf GE's Jeff Immelt shows off "stethoscope of the 21st century". Ipod-sized ultrasound: http://bit.ly/2Y4vS3 Open source e-voting system code begins to be ready for public review. http://bit.ly/1OGgLo Volunteers wanted for (simulated) Mars mission. http://bit.ly/2HvOaS Sign up now for the free EE Times "Connected Devices" virtual conference, which I'll co-chair. http://bit.ly/2on2if Alpine skiing robot. Videos: http://tw2t.com/e8 Automotive semiconductor demand begins slow recovery. http://bit.ly/15Sx2n |
|
Bring Embedded Software Boot Camp to Your City ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
We plan to take the popular Embedded
Software
Boot Camp "on the road" in 2010, to as many
cities and countries as possible. If you might
be
interested in attending one of these, please
help
us choose the cities and dates. It will only
take a
few minutes of your time to answer our 5
survey
questions.
|
|
Quick Links... ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
|
Contact Us... ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
email:
mbarr@netrino.com
phone:
866.78.EMBED
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|