So, here's my 'inxi -bxx' output:


However when I run sgfxi I run into many issues... such as it tells me to select Number 1
(Actually it tells me to Enter 1, however, 1 is preselected, and this results in building ubuntu packages??? Though the help says this is -not- the default method after passing option --help??? and where my confusion begins...)

So I do, and press enter, the requisite four or five times (didn't count) afterwards... until it asks me whether or not I want to install it
(Mind you... I'm following the scripts instructions, and it is a -NICE- script...)

Selecting Yes -or- No, the script reports success, however It doesn't seem to succeed -without- (selecting yes) installing, as X is loaded in what appears to be a fallback FB mode? and selecting yes, clearly installs the packages... which if I am not mistaken, AMD's make script list DKMS as a dependency? So you can see my issue as it reinstalls dkms, and tries to build for Liquorix, and fails yet again.

-I'm- sure (though I am a fifteen year beginner at linux[Not my age, rather experience]) I should be selecting direct install method since I am using an xorg.conf...
(Though I could be wrong, I am following the scripts instructions mind you... though there is a lot of confusion, and I haven't tried going against doctor's orders yet.)

I -HAVE- manually/directly installed both ATI & NVIDIA back in the day, but since everything is so automagical these days typically, I have not dove deep into any of this in almost 9 years.

Graciously awaiting some sort of insight,
-- Jase
P.S. Thanks for the great stuff. Just really want to get this figured out. I do have a fallback installation so I am not a$$ed out while waiting for your fantastical help.
P.P.S. I am running this script against the generic kernel, since Network doesn't seem to start, since X is not run during startup of Liquorix? Though I assume, I could always build for that kernel using some of the options I noticed in --help?
P.P.P.S. I have tried both purging DKMS myself, and letting the script try and remove it (seen something about it trying to remove it at least in the output)... I have tried running this script about 10 times at this point I think.
Back to top

I honestly wish I knew what you were talking about. I really do.

I mean, you posted the inxi -bxxx as we ask, but I still have no idea what you are talking about.

sgfxi runs like this, out of X: sgfxi

that's it, nothing more, unless you want beta nvidia, then you'd do: sgfxi -B

but there hasn't been an amd beta driver for ages, so that's irrelevant.

I really can't say, you may be confusing the fglrx catalyst installer asking you things with sgfxi, but I really cannot say, I'm sorry.

Now note: you have my sincere and deepest condolences for having to try to run amd catalyst fglrx drivers on linux, again, you have my deepest sympathy.

But honestly all I do in sgfxi is make sure the ongoing and routine random changes amd does to the fglrx packaging and download locations is smoothed over. What fglrx / amd driver itself does once it hits your system, well, all I can do is repeat my sincere and heartfelt sympathy for your plight.

I would note one thing: amd drops support for their non free catalyst drivers VERY fast, so don't assume you can run those on new systems over time, that's very unlikely, so get used to radeon drivers, or their coming kernel thing they are doing, which is sure to break almost all existing older systems on its release.

sgfxi doesn't ask you for anything unless you have multiple graphics cards installed. Unless something else is going on.

catalyst install issues or bugs I really can't comment about.

I see however from your inxi output that you do in fact have fglrx installed and apparently running. If you hadn't run inxi as root we'd see more gfx data though.

If you tried the -d distro driver option in sgfxi, well, that's a mistake, sgfxi has almost no ongoing ubuntu support in terms of my wasting my time tracking the way ubuntu keeps changing the non free video driver names, so I would strongly recommend against using that in ubuntu. Debian it should be fine.
Back to top

It appears in my attempt to bring clarity to my situation, I made it less clear as to what my true issue is, and muddied the situation.

I -DID- just run sgfxi, as is, after running the following commands.


The initial prompt (after script update check, and initial download) tells me to just hit enter through installation, and tells me if asked, to choose method 1.
Method 1 is "Build distro specific packages (this isn't verbatim, I am paraphrasing for most of this)" not "Direct install".

*I AM* assuming method 1 is what I would need to pass the -d switch for, but as I haven't passed -d, *I AM* assuming, in actuality (and contrary to what your script tells me) *I* should select method 2, for direct install.

I know AMD Hardware/Catalyst drivers are pretty bollocks, but this is a second hand budget PC that a friend gave me.

My issue is: trying to use SGFXI, as is, tells me to do this one thing, which seems to conflict with the --help switch's output (--help tells me default method of the script is to handle direct install).

Again, the help switch explains that -d is the switch for distro build option... which from what I gather, is not the default method (direct install is the default method according to what *I* can gather from --help)...

HOWEVER, Catalyst insists "Build Distro Specific Packages" is method 1 (The one your script tells me to select, however this conflicts with direct install being the default method for handling, according to your --help)

But again, assuming from the --help switch, what I really want is direct install method, or in other words option 2 in the catalyst installer.

My implication is simple: There is outdated information being printed out upon initiating the Catalyst portion of the script, in *MY* observed opinion.

*I* was trying to follow instructions in the script, as I mentioned, however no fruit came to bear.

FLGRX is installed, because I selected method 1, and it built the packages, as THAT IS THE DEFAULT STEP YOUR SCRIPT INFORMS ME TO TAKE

And also after building the packages (That I am sure I really shouldn't be) I tried both options of letting the catalyst executable install it, and not installing it...

After which: (BOTH WAYS) SGFXI reports success of installing the driver.

If I opt to let the Catalyst Installer, install aforementioned built packages, I get what appears to be a fallback X almost (no modes set? [perhaps because I never bothered to run Catalyst Configuration since I knew something was occurring that shouldn't and was trying to remedy it before progressing])

Selecting not to install the built packages yields no X at all.

Again, with either option chosen at the end of the Catalyst Installer, SGFXI reports Success.

What this tells me, as someone who is well read and pretty able at comprehending instructions, even poor ones; is that two things are occurring:

#1> I'm following old oudated instructions[As provided by SGFXI], and having partial success, because while SGFXI is not truly succeeding, the Catalyst Installer still is: at building and installing the .deb packages (Again, it seems this is not what SGFXI intends for me to do)

#2> Although the SGFXI script is reporting success, it truly hasn't succeeded, because I didn't do the direct install method, and from my understanding, if I am selecting option 1 during initiation [LIKE THE SCRIPT SAYS TO] I --SHOULD-- be passing -d with the script, which isn't what I want anyways. I -want- the direct install method, and that was going to be my next shot.

Anyways, I have tried to be as clerically efficient as possible while still being able to be understood. It was hard enough to write outside my style of communication twice, I hope I do not have to reclass this information again :) It is hard to figure out how to be as clear as possible in this situation, because we are looking at it from two different standpoints.
Back to top

When trying to run it =SGFXI= while running Liquorix, I receive this error.

Headers meta-package and headers -ARE- installed.

I guess technically this is in the wrong forum too... because the problem pertains to SGFXI rather than Liquorix itself, if you could deign to move it?
Back to top

Oh, now I see.

I don't actively follow catalyst, since one of the signs of amd's ongoing self destrution is their total inability to actually be consistent release to release, in any way, shape, or form.

So it looks like I had some older amd instructions there, which they then, in a totally typical style, randomly changed, reversed, for no reason at all, but that's honestly what AMD does almost every single new driver release.

I'll check sgfxi's instructions,, and update them, since AMD will probably reverse this stuff every other release, I'll have sgfxi tell the users to pick the direct install option, and I won't try to guess what number AMD uses today, on today's release, which will be different from what they say on tomorrow's release, or the next one.
Back to top

The direct install method should have installed the headers by default, so that is maybe an ubuntu glitch, 10 to 1 ubuntu may have changed some package name, only a look at the sgfxi log would show me that.
Back to top

I cannot find any trace of that message in sgfxi, it's a long program, so it doesn't mean it's not there.

Note: output from sgfxi is colored, ie, not white or black or whatever.

If you have a black console screen, the system output is white, if it's white screen, it's black by default. system and catalyst output will be distinctly non sgfxi - like, so don't mistake the two.

You'll have to provide me with more information, like the EXACT text you see sgfxi showing, so I can search sgfxi for it and fix it.

But I can't find it. Is it in smxi by chance? No, it can't be, because you are running ubuntu, and smxi doesn't support ubuntu, so it has to be sgfxi.

Also post: /var/log/sgfxi/sgfxi.log and maybe it will have a hint there.

Best to paste that using pastebin or some other paste site, then paste the url, those log files clutter up the forums and usually don't provide much useful info.

I thought I would be able to find the text you described by looking, and I don't, so you'll have to be more specific.

It certainly sounds like an error I'd make, putting in a specific adjustment for a catalyst issue, then amd changing that at random to be something else, that's their MO, modus operandi, it's one of the dead giveaways that their group is led by, and has been led by, incompetents for a long, long time, though now they are basically getting ready to just shut down their linux group as much as possible, it's already been massively trimmed, and I'm sure they want it gone, which is probably why they are trying to shove as much of the logic into some open source software layer in the kernel as possible before they really minimize their linux group even more.
Back to top

I will be sure to get those to you. I am happy to help in anyway to improve this tool. I will try and show exactly what I mean physically.

Give me some time and I will provide this.
Back to top