LinuxQuestions.org
Help answer threads with 0 replies.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware
User Name
Password
Slackware This Forum is for the discussion of Slackware Linux.

Notices


Reply
  Search this Thread
Old 11-07-2019, 09:46 AM   #91
baumei
Member
 
Registered: Feb 2019
Location: USA; North Carolina
Distribution: Slackware!
Posts: 59

Rep: Reputation: 19

Hi Twigster,

What is the method you found for reliably getting the trouble to occur?

Back when I was trying to figure out a fix or work-around for the trouble I could not well test my attempts, because I did not know of any reliable way to initiate the trouble. As far as I could tell, the trouble had a random component --> so, I would make a change, and then continue using the computers. If after a while the trouble had not showed up, then I guessed what I had done had an effect, but I was very uncertain of this...

Of the computers my family owns which have the trouble, one of them has been running Slackware long enough to have had 13.37 and 14.1 before 14.2 was installed in 2016/Oct. According to my recollection, neither 13.37 nor 14.1 had the trouble --- if this recollection is correct, and if 14.1 is still being updated/patched, then running 14.1 may be a solution for your Dell D620. In any case, a comparison between 14.1 and 14.2 may provide a hint as to what causes the trouble.
 
Old 11-08-2019, 04:53 AM   #92
Twigster
Member
 
Registered: Oct 2019
Location: France
Distribution: Slackware64 14.2
Posts: 34

Original Poster
Rep: Reputation: Disabled
REDACTED

Last edited by Twigster; 11-08-2019 at 05:40 AM.
 
Old 11-08-2019, 05:40 AM   #93
Twigster
Member
 
Registered: Oct 2019
Location: France
Distribution: Slackware64 14.2
Posts: 34

Original Poster
Rep: Reputation: Disabled
Hello,

my way is to open this pdf from this page in firefox 68 ESR, then select 'open with Okular'; then Okular opens, and I randomly drag around the document with the mouse and it triggers the hang.

What kind of comparison between 14.1 and 14.2 do you suggest? I do not know anything about how the distribution is made, and it looks like finding a needle in a haystack.

Besides, I managed to compile a 5.3.8 kernel. Not only does it still have the problem, but currently the boot process has an issue : after prompting 'starting X server' and 'login:' XDM takes more than 1 min to start.

I provide the config file for 5.3.8 (here is 4.19 from slackware for comparison)

And dmesg


/var/log/messages section (just to check the hours against X start time)

Xorg log

Now, I don't know why XDM wouldnt start immediately like it does with 4.4.190 and 4.19.81, and I don't know what else I could try to change on my system.

Last edited by Twigster; 11-08-2019 at 05:59 AM.
 
Old 11-08-2019, 10:31 AM   #94
enorbet
Senior Member
 
Registered: Jun 2003
Location: Virginia
Distribution: Slackware = Main OpSys for decades while testing others to keep up
Posts: 2,483

Rep: Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528
Hello Twigster. Please forgive me for not searching through what is now for me 6 full pages of posts but if this is a repeat then let it just be a double-check. The latest BIOS firmware I see for your Dell is A10 from 2008. Is that the version you are running?

I ask this because ACPI from that era was in transition and is thankfully since fixed in virtually every BIOS implementation since, but it was and is still a pretty important obstacle for many if not most computing devices before the fix. Additionally Dell, being a company whose market while certainly including Enterprise and other serious uses, also had massive appeal among the general public ( remember "Dude! Your getting a DELL!" on TV commercials?). It is common in digital devices aimed at the more "low tech" to limit access to anything that would create "callbacks" and Dell is no exception. Their BIOS firmware hides a multitude of options common in more generic and self-build units. For this reason the best you can hope for, outside of using something like "modbin" to unhide options (and even that is limited on Dell firmware) is to get and try the latest and greatest and hope it contains the fix you need.

If you've got A10 now, I'm left with the only suggestion I can presently see - Go back to a kernel that worked with your firmware. OTOH if this is just a fun experiemental challenge for you then by all means keep slogging away. I can relate. If OTOH you need a fully functional laptop at some point your time would have bought you a better/newer one absent what seems to be tied into ACPI issues.
 
Old 11-08-2019, 10:56 AM   #95
Twigster
Member
 
Registered: Oct 2019
Location: France
Distribution: Slackware64 14.2
Posts: 34

Original Poster
Rep: Reputation: Disabled
Hello
Quote:
The latest BIOS firmware I see for your Dell is A10 from 2008. Is that the version you are running?
My laptop ran A08 before I created this thread then I made the update to A10 based on a suggestion by abga (I think)

Quote:
It is common in digital devices aimed at the more "low tech" to limit access to anything that would create "callbacks" and Dell is no exception. Their BIOS firmware hides a multitude of options common in more generic and self-build units.
One weird thing about the BIOS is that it says memory installed: 4GB and memory available: 3GB and I didn't find a way to get my 4GB to be usable by the OS. But that's not the point of this thread.


Quote:
Go back to a kernel that worked with your firmware
All the kernels I've used so far worked fine, except for this annoying but not deal breaking issue.

If I keep receiving tech support I might as well continue to investigate as I find this interesting and already learned a lot.

Last edited by Twigster; 11-08-2019 at 11:05 AM.
 
Old 11-08-2019, 02:29 PM   #96
enorbet
Senior Member
 
Registered: Jun 2003
Location: Virginia
Distribution: Slackware = Main OpSys for decades while testing others to keep up
Posts: 2,483

Rep: Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528Reputation: 2528
Quote:
Originally Posted by Twigster View Post
Hello
My laptop ran A08 before I created this thread then I made the update to A10 based on a suggestion by abga (I think)
Good. That was the wise choice especially with Dell.

Quote:
Originally Posted by Twigster View Post
One weird thing about the BIOS is that it says memory installed: 4GB and memory available: 3GB and I didn't find a way to get my 4GB to be usable by the OS. But that's not the point of this thread.
That laptop shipped with Windows XP Pro which even if it was the 64bit version wouldn't support all of your 4GB due to licensing issues that severely limited an otherwise decent (for microsoft) OpSys. I don't know if the BIOS was also limited because of knowing XP couldn't use it anyway, but surely you can see that when Linux boots whether 64 bit or PAE enabled 32bit. Apparently there were some versions of 64 bit XP designed for server use that could but of course that would not include a laptop. XP was a bit weird about RAM. IIRC though even 32 bit XP could address 3.5GB if PAE was enabled in boot.ini so is it possible your graphics is "sharing" 1GB RAM? Here again that may be hidden option in BIOS but visible through some operating systems.

Quote:
Originally Posted by Twigster View Post
All the kernels I've used so far worked fine, except for this annoying but not deal breaking issue.
Ahhh. I misunderstood and thought one worked properly.

Quote:
Originally Posted by Twigster View Post
If I keep receiving tech support I might as well continue to investigate as I find this interesting and already learned a lot.
Understood which is exactly why I wrote "I can relate". The Crash and Burn the Recover cycle is a sometimes painful but often also joyful and highly efficient way to attack a learning curve, revealing serious deep level gems. Best Wishes!
 
Old 11-08-2019, 11:11 PM   #97
abga
Senior Member
 
Registered: Jul 2017
Location: EU
Distribution: Slackware
Posts: 1,275

Rep: Reputation: 673Reputation: 673Reputation: 673Reputation: 673Reputation: 673Reputation: 673
@ Twigster
Quote:
If I keep receiving tech support I might as well continue to investigate as I find this interesting and already learned a lot.
From my PoV the tech support you need is not that much Slackware (Linux Distro) related anymore, but kernel devs territory(competency). I wish I could help you more, unfortunately I mentioned already a few times in this thread that I ran out of both ideas and knowledge. You should realize that the tech info you are looking after is not always to be found in docs but inside the kernel source code and in "the heads" of the devs that develop and maintain that code. These devs will see your bug report @ kernel.org and could provide competent advice/fixes.

- in post #46 I advised you already to get in touch (register a bug) with the kernel devs and provide them with the content of your /sys/firmware/acpi/tables/ I looked at those files myself and found them gibberish...

- in post #74 a link to an early discussion about the ACPI development and support state was provided, hinting that at the time your laptop was actual, ACPI was a mess. Also meaning that the systems at that time weren't properly supported by the kernel and that better support started afterwards, with newer systems - yours might have been left out - see enorbet points on ACPI from post #94

- in post #90 you asked some questions - the EDIT sections.
First EDIT: yes, just for inspiration or diff, but don't substitute. Actually the kernel config&make helper scripts should fix it - ask you to reconfigure it, but I never tried such a process and cannot guarantee a success
EDIT 2: it could help the kernel devs if they ask you for some extra info - see https://lwn.net/Articles/724562/
EDIT 3: https://www.kernel.org/doc/Documenta...freq-stats.txt (bottom of the page)
 
Old 11-09-2019, 12:02 AM   #98
baumei
Member
 
Registered: Feb 2019
Location: USA; North Carolina
Distribution: Slackware!
Posts: 59

Rep: Reputation: 19
Hi Twigster,

Earlier today, right after I read your reply (post #93), the trouble occurred on my little laptop, and it locked up.

I had been thinking about what you had written, and I happened to notice what had just been occurring on my laptop --> I had been moving the mouse pointer across my web-browser, one of mouse-over actions highlighted a button, and a split second later the laptop locked up. Then, I realized that a few decades ago I had seen this sort of thing.

A long time ago video cards and whatnot which one might put in a computer had jumpers for setting the 'interrupt' to be used. If one set a new card to use an interrupt which was already in use in the computer, and installed this new card, then one would get seemingly random errors which were generally severe (such as a crash).

For instance, if the mouse card and video card used the same interrupt, then the computer would likely crash when the mouse was moving while the screen was being redrawn. (Of course, the interrupt conflict could involve some card other than the two already mentioned.)

According to my understanding, with modern computers, at least some interrupts are issued during boot-up by the ACPI software.

Also, according to my recollection, in my stash of old computer parts I have a video card or network card which has an error: it uses an undocumented interrupt in addition to the one set by the jumper (somebody figured this out, and put the information on the Internet). The card and the computer both worked fine when there was no interrupt conflict.
 
Old 11-09-2019, 11:00 AM   #99
Twigster
Member
 
Registered: Oct 2019
Location: France
Distribution: Slackware64 14.2
Posts: 34

Original Poster
Rep: Reputation: Disabled
@abga :
I understand I might have to go ask elsewhere, either on a BIOS forum for D620 hidden options, or at kernel.org. Should I mark this thread as done, or come back if/when I get some answers?

According to cpufreq-stats.txt, I would have to disable CONFIG_CPU_FREQ_STAT in the configurator before being able to disable CONFIG_CPU_FREQ?

@baumei :
Because I could reliably reproduce the problem I tried stracing okular with abga. But AFAIK strace is about program system calls rather than hardware interrupts. Tracing those look like another complicated endeavour that I do not wish to embark upon unless some knowledgeable kernel person needs the info to fix the bug.
 
Old 11-09-2019, 11:29 PM   #100
baumei
Member
 
Registered: Feb 2019
Location: USA; North Carolina
Distribution: Slackware!
Posts: 59

Rep: Reputation: 19
Hi Twigster,

It may be that we currently see no approach to getting a solution, however the problem is not solved, and so I would NOT mark the question "solved" --- because if it is marked "solved" then someday someone having the trouble will come to this thread and read it, and be disgusted that there is no solution in the thread marked "solved".

In my opinion, one of the next steps is to figure where to actually make the bug report. A while back I had a clear cut bug (load one kernel, and the computer had no display; load the other kernel, and the display worked fine, and it was repeatable). I spent some time looking for where to make the report, and I could find no such place or method. So, if you do find such a place or method, then please let me know. If you do not want to report this trouble, then I will.
 
Old 11-10-2019, 12:06 AM   #101
abga
Senior Member
 
Registered: Jul 2017
Location: EU
Distribution: Slackware
Posts: 1,275

Rep: Reputation: 673Reputation: 673Reputation: 673Reputation: 673Reputation: 673Reputation: 673
@ Twigster
I doubt you'll find a BIOS forum or one focusing solely on the rather antique D620. I'd advise to direct your efforts at the kernel devs first and provide them with your acpi tables & point them to this thread for the background.
Regarding putting the thread on RESOLVED, well, it's up to you, even if it's not something Slackware (as a distro) can help you with, your problem isn't solved yet. In my previous post I didn't mean to push you away from this forum, but explain that your issue should be discussed with the kernel folks. I guess it's beneficial for the Slackware community if you keep updating this thread with your findings & hopefully a resolution. Nevertheless, feel free to continue asking for "tech support" if your requests are Slackware (& Linux) related.

CONFIG_CPU_FREQ_STAT appears to be depending on CONFIG_CPU_FREQ, so yes, you need to disable it if you also want to disable CONFIG_CPU_FREQ. The kernel make file(s) will check it & prompt if you make such mistakes with dependent options. However, I don't understand how you could enable CONFIG_CPU_FREQ_STAT and not have CONFIG_CPU_FREQ already enabled. I'm only using the text based (ncurses) "make menuconfig" and never got into such a situation - able to enable a depending option and not enable the parent. If this is the case in "make xconfig", then I'm happy I never used it in almost 25 years of "Linuxing" ...



@ baumei
Usually the kernel handles IRQ issues, try booting with the kernel parameter: irqpoll
on the system that you believe it's unstable because of the IRQ sharing(conflict?).
https://www.kernel.org/doc/html/v4.1...arameters.html
Also worth reading:
https://www.tldp.org/HOWTO/Plug-and-Play-HOWTO-10.html
 
Old 11-12-2019, 01:01 AM   #102
baumei
Member
 
Registered: Feb 2019
Location: USA; North Carolina
Distribution: Slackware!
Posts: 59

Rep: Reputation: 19
Hi "abga",

When one sees entries in /var/log/messages such as the following
Code:
Nov 11 19:26:22 darkstar kernel: [    0.613767] ACPI: PCI Interrupt Link [LNKA] (IRQs 3 4 5 6 7 *10 11 12 14 15)
Nov 11 19:26:22 darkstar kernel: [    0.614075] ACPI: PCI Interrupt Link [LNKB] (IRQs 3 4 5 6 7 10 *11 12 14 15)
Nov 11 19:26:22 darkstar kernel: [    0.614378] ACPI: PCI Interrupt Link [LNKC] (IRQs 3 4 5 6 7 10 11 12 14 *15)
Nov 11 19:26:22 darkstar kernel: [    0.614685] ACPI: PCI Interrupt Link [LNKD] (IRQs *3 4 5 6 7 10 11 12 14 15)
Nov 11 19:26:22 darkstar kernel: [    0.615025] ACPI: PCI Interrupt Link [LNKE] (IRQs 3 4 5 6 7 10 11 12 14 15) *0, disabled.
does this mean the ACPI software is setting IRQs, or reporting what it has discovered? I had guessed it was setting them, but I do not know.

Hi Twigster,

Have you tested 14.1 on your D620? If so, does it have the crashing problem?

You asked, "What kind of comparison between 14.1 and 14.2 do you suggest?" If the trouble does not exist for 14.1, then I would start looking at what is different between the 14.1 and 14.2, and use this information to decide on a next step.

I think modules are, or can be, written independently of the kernel; and can be updated/patched, and then compiled to work with the kernel. If so, then perhaps it would be interesting to see what version of the i915 module is used in 14.1, and then see about downgrading 14.2 to that module version.
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
Problems with HP 3500c scanner and Dell D620 laptop runing Ubuntu 9.10 wa3fkg Linux - Laptop and Netbook 1 12-03-2009 10:37 AM
SuSE on Dell D620 laptop cvzyl SUSE / openSUSE 4 04-08-2008 07:53 AM
FC5 and Laptop Dell Latitude D620 zillah Linux - General 3 11-07-2006 12:51 AM
Dell d620 graphics support bagpussnz Linux - Laptop and Netbook 2 05-10-2006 04:04 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware

All times are GMT -5. The time now is 07:47 PM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration