howdy, first time caller... ( new Kali user)



Thanks for these scripts by the way, they are great and one heck of a frustration saver.

according to the log, the driver package downloads just fine
then the fun seems to happen:

Function: pre_extract_clean_setup - Utility: Start
Args: amd-catalyst-13.4-linux-x86.x86_64
Function: pre_extract_clean_setup - Utility: End
./: ./: is a directory
ERROR: (196) The Driver Package: failed to extract and has been deleted.


to try and fix this, I extracted the package using the zip utility on the amd*zip file in /usr/src/sgfxi-downloads directory.

I then ran the *.run file from that location, and it would appear to have run just fine...

so either I am a bone head, or something is up.
Back to top

something is up, amd has just, again, changed something in both their file naming and packaging junk.

To say that trying to support amd is an exercise in frustration and futility is a massive understatement, their total and utter lack of internal discipline, which manifests in these types of failures, is beyond belief, my current theory is that the company has dumped all their already poor programmers and replaced them with interns, who are just doing random stuff.

I have not had time to track down what they did this time, but I do find this increasingly ridiculous, while nvidia requires a fix or update to the file naming/install logic maybe once every few years, then remains predictable and stable for many years after, amd changes things quite literally every few months at times.

I cannot tell you how many times I have wanted to just dump amd support in sgfxi because I don't get paid to do it, don't use amd/fglrx for any reason ever, and gain zero joy from re-engineering sgfxi to handle the latest random bs amd change.

They've also now randomly renamed their driver package by the way for the beta series, which I just saw last week when something else with amd failed in my backend.

It's gotten beyond pathetic at this point, this company will clearly never ever put out good quality or consistent methods. I will have to at some point decide if it's worth my time to keep playing with these losers or to finally just say, fine, I give up, I cannot keep up with amd's incompetence.
Back to top

Thanks for looking at my post. I have skimmed the forum, and the -h for sgfxi, and would like to know if it is possible to download manually, to the correct directory, unpack the .zip, again manually, run the *.run file and then run sgfxi with an option to tell the script that the driver package has been installed, now on with the show.

I would be more than happy to test with/for you and I have to say, after looking at your script, it is a thing of beauty, just really friggin long

sent some donation your way as a thanks for your hard work, attention to detail, and support
Back to top

this issue I have located the cause, amd actually physically changed the existing 13.4 package contents, which is kind of amazing, that's bad even by amd standards.

I have that corrected now but not on the live sgfxi, but am also going to try to patch for their latest randomized file naming revealed in the beta file name changes before uploading the new sgfxi.

Basically I have to code in as default the expectation that names will change at total random, so sgfxi has to default to, for amd, a range of possible file names, and possible contents of files and zip files, instead of assuming, that is, professional competence on amd's part, which was already a joke, sgfxi will now assume complete incompetence.
Back to top

please note, because fixing the beta section is much harder than I expected, I just uploaded the fixed regular catalyst installer section in sgfxi.
Back to top

damn, awesome, so smxi -U or just sgfxi -U?
Back to top

sgfxi always updates itself automatically, so: sgfxi

I have now also debugged the newly altered beta driver file stuff as well.

AMD is making it nearly impossible to support them with their near psychotic rate of random file name/ download url/path changes, they this time actually changed the already uploaded 13.4 to 13-4, and changed the beta file name syntax again back to 13-6 while dropping, maybe, the numbered beta version in the file name.

AMD is a company to avoid for your video needs if you run linux, I have never seen this fast a rate of stupid random changes coming from them, sgfxi is now increasingly internally expecting only randomness from amd, but randomness is much harder to handle than non randomness.

I view amd linux group as an example of how not to do things right.
Back to top

amd, again, changed file names/paths literally as I was doing this debugging, forcing yet another round of testing.

It looks to me like their linux group is having a psychotic episode, this is the most random and poorly executed set of file name/syntax changes for their drivers I've ever seen, clearly nobody is doing any planning in this group, I cannot urge anyone reading this strongly enough: AVOID this company if you need any high end video drivers for your linux system, just run as quickly as you can, I have never seen them be this bad in my life so far, and that's saying a lot because they have been very bad.

Clearly the layoffs in their linux group are starting to have a real impact, do NOT repeat NOT buy amd video cards if you need the non free driver to work consistently over the next years.

Their beta 3.6 is now installable via sgfxi as well.
Back to top

spoke too soon, driver is installed, but X says tuff $hit!

from Xorg.log

fglrx: No matching device section for instance 9BusID PCI:0@1:0:1) found

when I try and start X by hand using startx, it bails on me and saying

(WW) fglrx: No Matching Device section for instance (BusID PCI:0@1:0:1 found ((lspci says that the radeon card ( on this hp envy 14 laptop) is at 01:00.0)
Back to top

will try the beta drivers...

does it matter that this is a laptop with both intel and ati graphics on it, and they are theoretically switchable do anything for/against me in this case?

Dan

** just another tidbit from the install of the beta driver,

Removing module drm . . . . . . failed


whiskey tango foxtrot over? thoughts?
Back to top