Many of my traditional blog post live on this site, but a great majority of my social-style posts can be found on my much-busier microbloging site at updates.passthejoe.net. It's busier because my BlogPoster "microblogging" script generates short, Twitter-style posts from the Linux or Windows (or anywhere you can run Ruby with too many Gems) command line, uploads them to the web server and send them out on my Twitter and Mastodon feeds.
I used to post to this blog via scripts and Unix/Linux utilities (curl and Unison) that helped me mirror the files locally and on the server. Since this site recently moved hosts, none of that is set up. I'm just using SFTP and SSH to write posts and manage the site.
Disqus comments are not live just yet because I'm not sure about what I'm going to do for the domain on this site. I'll probably restore the old domain at first just to have some continuity, but for now I like using the "free" domain from this site's new host, NearlyFreeSpeech.net.
(This is what the Fedora 19 AMD Catalyst packages look like on my Fedora 20 system when seen in Yumex. Click the image above for a bigger version.)
Who better to tell you how to find and install an RPM package for the AMD Catalyst driver in Fedora 20 than the very person who ophaned the driver for that very release?
That's right, Leigh Scott, who had every right to drop the packaging of the AMD Catalyst driver in RPM Fusion for Fedora 20, is still maintaining it for Fedora 19.
He has an easy recipe for using the F19 driver on F20 systems. I can confirm that his method works. As is, this RPM of the Catalyst driver does not work with GNOME 3 (due to previously mentioned Wayland code that GNOME is now including). It does work with Xfce and KDE (and everything else that isn't GNOME 3, I presume).
Here are the instructions, originally from Leigh's post on the Fedora Forums, with my annotations:
First, make a directory and cd into it. Leigh suggests calling it 'catalyst':
$ mkdir catalyst
$ cd catalyst
Grab the needed Fedora 19 packages with yumdownloader:
$ yumdownloader --releasever 19 xorg-x11-drv-catalyst-libs.i686 akmod-catalyst.$(uname -m) xorg-x11-drv-catalyst.$(uname -m) xorg-x11-drv-catalyst-libs.$(uname -m)
Use yum to install the packages (shown here using sudo, though you can also su to root if you wish):
$ sudo yum --nogpgcheck install *.rpm
After this installation, I rebooted and had a working Catalyst/fglrx driver on my system. As I said above, it doesn't work with GNOME 3, but neither did the upstream AMD package before it stopped working altogether with the 3.14 kernel.
Configuration note: I did NOT need to do this, but if you have problems, you might want to use the aticonfig utility as suggested here:
$ sudo aticonfig --initial
Again, I did NOT need to do this.
Also, I'm not sure if these Fedora 19 packages will be updated with when I run Yumex or yum update. I do know that it's a good idea to keep an eye on the latest packages in RPM Fusion (in this case the non-free F19 updates repository) to make sure that you don't install any kernels before a new Catalyst is ready for them.
I will update this post when I have more information on how long this fix continues to work.
I'd like to thank Leigh both for his work on AMD Catalyst in RPM Fusion until now as well as for this temporary Fedora 20 fix.
At the same time, I once again call attention to how the lack of an RPM package of AMD Catalyst for Fedora takes away choice and functionality from the distribution and its users.
As much as I love Fedora and its community, if you have a newish AMD-running computer, I really can't recommend Fedora because of this continuing problem. Sure, the open Radeon driver for AMD graphics chips/cards is better than ever, but I can't get suspend/resume with it. Once that starts working for me, I'll shut up.
So what do you do if you need AMD Catalyst? Distributions that haven't fallen into this rabbit hole include Debian, Ubuntu, and every single other one I can think of.
I'll ride this fix as long as I can, but you can bet I'm thinking of where I can go in terms of a new Linux distribution in order to have my choice of video driver.
Since Phoronix is the best source of news on Linux graphics hardware (and probably Linux hardware and benchmarking, too), I decided to e-mail the site's Michael Larabel and see what he thought about the fact that there has never been an RPM-packaged AMD Catalyst driver for Fedora 20, and at the moment even the upstream AMD installer won't work with the 3.14.x Linux kernel.
I'd like to thank Michael for turning that e-mail into this article: AMD Catalyst On Fedora 20 Is Left In An Awkward State.
The next day, he followed it up with How-To Install AMD Catalyst 14.4 On Fedora 20 With Linux 3.14.
I'd like to thank Michael for this, and for all the day-to-day reporting he does on Linux (and often BSD) in regard to drivers and hardware.
In a forum post for the last article, I found a different, better way to get AMD Catalyst/fglrx working in Fedora 20.
Ironically (or perhaps coincidentally), this tip comes from the guy who orphaned the Catalyst packages in RPM Fusion for Fedora 20 but still maintains them for Fedora 19 (and it involves using the F19 packages in F20).
I have to do a few more tests of this method before I detail it in another post, but first I'd like thank Michael again for his posts, and Leigh Scott right now for this too-easy way of getting the Catalyst driver working again.
Final word: I don't blame Leigh, per se, for dropping the Catalyst package in RPM Fusion. It's every maintainer's right to quit whenever they want. I'm just continually stunned and saddened that in the many months since Leigh made this decision, nobody else has stepped in to fill this crucial gap in the Fedora/RPM Fusion/Catalyst world.
I think a Mesa update broke my AMD Catalyst driver's 3D hardware acceleration -- here's how I fixed it
I'm only speculating as to what caused 3D acceleration to stop working on my Fedora 20 system using the upstream-installed AMD Catalyst driver.
But I'm pretty sure it is the new Mesa packages that rolled into Fedora a day or so ago.
Even if Mesa isn't the culprit, apps that require 3D hardware acceleration are either throwing warnings about the lack of this particular feature, or just crashing immediately.
Running glxinfo in a terminal gave me the following message:
direct rendering: No (If you want to find out why, try setting LIBGL_DEBUG=verbose)
I suspected that the AMD Catalyst video driver, which I'm installing with AMD own .run installer because there is no Fedora 20 package for it in RPM Fusion, was somehow broken.
When I have problems with the proprietary video driver, I usually uninstall Catalyst, check whatever's broken while running the open Radeon driver and then reinstall Catalyst and check again.
Except this time Catalyst wouldn't uninstall. The error message I received said something about the configuration being changed.
Catalyst wouldn't reinstall, either.
The script output suggested that using --force would overcome the errors in either case -- uninstalling or reinstalling.
So I decided to reinstall AMD Catalyst over the current installation.
Since I was already running the latest Catalyst driver from AMD, I had previously downloaded, unzipped it and installed it, and the .run file was already on my system for the reinstall.
I did this as root:
# ./amd-driver-installer-14.10-x86.x86_64.run --force
Catalyst reinstalled with no trouble, I rebooted, and 3D hardware acceleration was back.
Kernel and other packages I have in Fedora that allow me to install and run AMD's upstream Catalyst driver
You need more than just kernel packages to successfully install the upstream AMD Catalyst driver in Fedora, and you might not need every last one of these packages. But it couldn't hurt to have:
kernel-devel
kernel-headers
kernel-modules-extras
kernel-tools
kernel-tools-libs
Other packages that you need or are helpful include:
dkms
gcc
binutils
make
Then you can go to the AMD Catalyst site for Linux, download a .zip file, unpack it and use the resulting .run file to install the driver.
I wrote up more detailed instructions on how to install the driver in January. Those instructions are probably due for an update. I'll do that soon -- maybe when AMD updates the driver for the 3.14 Linux kernel.
The new Firefox, version 29, brings a whole new look to Mozilla's web browser.
I hope it brings a lot of other new things, too. I pretty much run Firefox exclusively in Linux, and I'd love to do the same at my day job, in Windows 7, where I use Google Chrome for the most part. In my day job, I have a whole lot of tabs open, and Chrome seems to handle it better. I would welcome a more robust Firefox in this regard.
Better or not, Firefox 29 is now in Fedora. The image above comes from the Fedora Magazine post announcing the update, which already flowed onto my installation via the Yumex package manager.
It happened a day later than it should have, meaning Fedora got spanked by Debian, but the Fedora 20 patch for the OpenSSL 'Heartbleed' bug did roll onto my system today.
I would have liked Fedora to be ahead of Debian rather than behind it, but a day's delay isn't a deal-breaker. And I could have installed the OpenSSL update from Koji early if this were a server installation.
Overall, the free-software community's response to the 'Heartbleed' bug shows the power of open development and how these projects and products are stronger through transparency and sharing.
I was ready to give up. But what's great about Fedora is if something's broken, sometimes waiting is all you need to do.
Your problem will be resolved somewhere upstream. And Fedora gets newness from upstream faster than almost anyone (Arch notwithstanding).
So I was able to print to the HP LaserJet 1020 from Fedora 18 and 19 but not Fedora 20.
It has much, much more to do with the HP LaserJet 1020 printer than it does with any part of the Linux operating system.
I lasted four days this time. After I couldn't log in one morning after rebooting Fedora 20 under AMD Catalyst, I pulled the proprietary driver, leaving the open Radeon driver to run the graphics on my HP Pavilion g6-2210us laptop.
With every new kernel, Radeon gets better. I'd say the performance differences between Catalyst and Radeon on this hardware are small enough that I'd be happy to stick with Radeon and leave Catalyst upgrade trouble behind (mostly because THERE IS NO CATALYST PACKAGE FOR FEDORA 20, THOUGH NOBODY SEEMS TO CARE).
Once again, I did some updates on my Fedora 20 system. And after happily suspending and resuming the laptop for days, I crashed in the OpenShot video editor and had to do a hard reboot.
Except that I never got to the login screen. Just like the last time this happened, I suspected that the Catalyst driver I downloaded and installed from AMD's .run package was not playing well with the latest kernel from Fedora.
The LWN security updates posted today include the GnuTLS updates for Fedora and OpenSUSE. Debian and Ubuntu pushed out their patch a couple days earlier.
It's a pretty big bug that is being closed. Says Tomas Hoger in the bug report:
It was discovered that GnuTLS X.509 certificate verification code failed to properly handle certain errors that can occur during the certificate verification. When such errors are encountered, GnuTLS would report successful verification of the certificate, even though verification should end with failure. A specially-crafted certificate can be accepted by GnuTLS as valid even if it wasn't issued by any trusted Certificate Authority. This can be used to perform man-in-the-middle attacks against applications using GnuTLS.
This has been all over the Internet the last week or so.
Selena Larson of Readwrite.com writes:
A variety of Linux distributions are vulnerable to hacks because of a bug that allows people to bypass security protocols to intercept and disseminate encrypted information. A member of the Red Hat security team discovered a bug in the GnuTLS library that allows hackers to easily circumvent the Transport Layer Security (TLS) and secure sockets layer (SSL).
The vulnerability affects the certificate verification, meaning secure connections that are supposedly going through as secure, are not. Someone could compromise a secure connection by using a “man-in-the-middle” attack, acting as the server to intercept traffic, financial transactions or secure information.