Alan Cox released Linux kernel 2.4.13-ac7 and it is available here. Read on for the changelog.

*
* Handle with care. This has the IDE driver update
*

2.4.13-ac7
o Add support for one shot interrupt urbs in usbdevfs (Maksim Krasnyanskiy)
o Add Aiptek docs (Chris Atenasio)

o Fix HID docs path (Seth Arnold)

o ir-usb driver fixes (Dag Brattli, Greg Kroah-Hartmann)

o Support IRQ sharing on PCI IO8 boards (Patrick van de Lageweg)

o Fix olympic panic if mtu is set very low (Mike Phillips)
o LDM database reading fix (Richard Russon)

o Fix LSR handling bug in serial.c (Russell King)

o Fix problems with bridges over cipe (Lennert Buytenhek)
o Fix oops when an NFS server hands back an invalid/overlong response (Trond Myklebust)
o Fix missing error path unlock in ipconfig (Neil Brown)

o Fix a firestream ioctl return code (Patrick van de Lageweg)

o Core dump deadlock fixes (Manfred Spraul)
o UFS byteswap cleanup (Christoph Hellwig)

o Add another 440GX box to the hall of shame (Arjan van de Ven)

o Support multiple redboot images (David Woodhouse)
o Fix performance and misdetect problems in eepro driver (Chris)
o eepro100 ethtool support (Dell)
o Fix cache reporting (Rik van Riel)

o PCI hacks to make Intel alder boards boot (James Bottomley)

o APM tweaks (Thomas Hood)
o LVM updates (Joe Thornber)

o Intel 845 AGP support (GOTO Masanori)

o Remove accidentally added hook for a binary only module (Alan Cox)
o Added pci generic state save/restore (Only enabled by default on known problem boxes) (Arjan van de Ven)
o CRIS architecture updates (Bjorn Wesen)

o Free lingering swap cache on exit() (Rik van Riel, Marcelo Tosatti)

o Support HT processors microcode uploading (Tigran Aivazian)

o Reiserfs consts/formats/cleanup (Nikita Danilov)
o Scsi generic update, use kiovec_sz smartly (Doug Gilbert)

o PnPBIOS update (Thomas Hood)

o Fix T23 suspend oops in i810 audio (Manoj Kasichainula)

Popular Mechanics discusses the various wars over mp3s and how businesses are looking to outdo eachother in the quality of their mp3 players.

From PC World: “A controversial antipiracy technology built into Microsoft’s Windows XP operating system has been cracked, a security firm reports.”

Slashdot readers discuss an announcement about the 2.5 kernel tree, which includes Alan Cox’s decision that he will passed maintenance of 2.4 to Marcelo Tosatti.

Jeremy Andrews writes, “This week KernelTrap interviews John Levon, the author of OProfile and a contributer to KernelNewbies. He offers much insight into both of these projects, as well as reflecting on Linux in general. OProfile is a statistical x86 profiling system for the 2.4 Linux kernel, useful in understanding what percentage of the CPU is being utilized by different processes, including those in kernel space and those in user space. KernelNewbies is an excellent resource for people looking to understand the Linux kernel, comprised of a web page, an IRC channel, and a mailing list. The interview is here.”

– By Adam Cutchin –
Hardware vendors really don’t care much about Linux. It’s nothing personal. If Satan himself handed Michael Dell a copy of Windows 666, Mr. Dell would no doubt install it happily if he felt it would move machines out the door. His responsibilities are to his stockholders, period. People asking for Linux desktop computers are a statistical anomaly, and that in itself could easily be enough to justify their disinterest. But I think there are other contributing factors.

The PC industry is having some of its worst times in over a decade, and for good reason. CPUs are gaining performance faster than software can be written to exploit their power. Hard drives and memory are cheap enough that a $200 upgrade will give you more of both than just about any user can realistically use. Gnome is pretty resource-hungry, yet my 2.5-year-old computer serves me well while running a Gnome desktop. Who truly needs a new computer to run their existing software these days?

So what would make life easier on the hardware vendors? One gets the impression Microsoft asked the same question, and as an answer offered Windows XP.

XP is much less likely to work with existing hardware than any recent Microsoft OS. It complains loudly about drivers not “signed” by Microsoft. Microsoft discourages its installation on machines more that a year and a half old. Linux bends over backwards to work on any system and support the most obscure of hardware, even if that meant running with no graphical environment (I’ve never seen it come to that — X has great hardware support), whereas Microsoft places much higher demands on your hardware being up to date. Why shouldn’t the hardware industry support an effort resulting in more upgraded hardware?

XP actively discourages users from upgrading their own hardware after the fact. If XP determines that your system has changed significantly, it will require you to reactivate it. For users without an Internet connection, this activation process is astoundingly inconvenient. XP makes a 50-digit hash code representing your current system state. You have to call Microsoft, read them the entire 50-digit number, and then have to enter a 42-digit number for XP to become active. You have to exchange 92(!) digits via telephone with another human being to use your $200 (to be fair, the home version upgrade of XP is $99) operating system.

Users with Internet connectivity will find the process a bit easier, but then again users with Internet connectivity will have XP try to sign them up with Passport five times before it finally stops asking. Hardware vendors can get around that by having XP key itself off of the BIOS serial number, ensuring you will never upgrade your motherboard. (Or if you do, it will cost you the price of a new OS on top of the hardware cost.)

Mandrake, Red Hat, or whomever, can offer a great OS for a great price,
but they can’t offer the same sort of reciprocity the Evil Empire is
offering. The American consumer base has become so complacent that we
tolerate being insulted by Microsoft for performing simple tasks like
upgrading hardware, for not getting a Passport account or for playing MP3s (to
name just a few tasks). As long as we demonstrate a willingness to buy into
the monopoly, it’s just not in the vendors’ best interests to offer a
non-Microsoft OS on a desktop computer.

Editor’s note: This weekend community opinion feature is for readers of NewsForge and Linux.com to express their opinions about issues related to Linux or Open Source. We do not pay for these opinion pieces, but it’s a chance for you to get your opinion out, front and center, to thousands of other readers. Email editors@linux.com to pitch us an idea for an opinion piece you’d like to do.

theHighlander writes “Today, 90% of the microprocessors shipped are used in embedded systems and an average number of microprocessors found at homes exceed 50. Studies estimate the latter figure to reach 200 in 3 to 4 years. “Embedded Internet”, a marriage of embedded systems and Internet technology, are the two highly successful miracle technologies that will be responsible for the increase. Here are the factors on how itll happen…”

ZDNet reports that Sun is considering suing Microsoft to prevent the software giant from continuing to harm its competitors, a problem which Sun claims is not resolved by the anti-trust settlement.

PC World reports: “Intel will demonstrate the first 0.13-micron Pentium 4 processors at the upcoming Comdex computer industry trade show in Las Vegas,” which “lays the groundwork for the Pentium 4 chip family to attian speeds as high as 3 GHz by the end of next year.”

Timothy R. Butler writes, “As an IT worker, I’m sure that you recognize the USA v. Microsoft case as something that could profoundly effect the direction IT will move in. However, after today’s settlement announcement, many pundits suggested that the proposed end accomplished little at all. I beg to differ. See what I mean here.”