[SOLVED] Screen Goes Black After 14.2 Update

poetgrant · 10-25-2019, 04:51 AM

Yes, I have scouted the threads for an answer, and no none of them helped.

I recently bought an ASUS Eee PC 1005HAB. I downloaded the 32bit Slackware image and installed. Everything was going great. I configured everything and setup DWM just to my liking. Then I decided to connect to one of the mirrors and run slackpkg update and upgrade-all just in case. I have done this hundreds of times on my main PC without issues.

However, I have reinstalled Slackware 8 times now and tried updating again with no luck. When I reboot after an update, it starts loading the kernel, the screen flickers when loading the video driver and everything gets small, then it begins loading the rest of Linux. Then bam, screen goes Black and nothing. I have tried blind typing my info then started like another thread suggested, but it does nothing.

Can someone explain what's going on to me? I have just reinstalled again and am not planning on updating until I figure out what the heck is going wrong.

Lysander666 · 10-25-2019, 05:15 AM

I'm sorry, my response is not going to be that useful. All I can say is that I'm one of the only others here I know of who runs Slackware [occasionally] on a 32bit Atom netbook.

The user experience is not that great, no matter what WM/DE you're using, unless you're going to be using the terminal mostly: web browsing is slow and video streaming impossible on some sites.

I can only imagine one of the newer video drivers is not compatible with your hardware. It may be worth reinstalling and not updating, or at least not updating the video drivers. Did you also update the kernel or not?

hazel · 10-25-2019, 05:56 AM

Are you using a proprietary video driver?

Does the blackout only affect X or are virtual consoles blank as well? Have you tried booting to level 3 instead of 4?

Also try putting the caps lock down and see if the indicator lights up. If it does, you have a working console and only the video is kaput.

SpacePlod · 10-25-2019, 06:27 AM

Just a thought, but have you tried ssh'ing into the box once the screen goes black? Perhaps blacklist the kernel, update again, and if the screen goes black try logging in via another computer over ssh to check dmesg and/or X logs.

0XBF · 10-25-2019, 06:46 AM

I have an old Asus Eee 1000H that runs 14.2, although I just use it to run mocp from runlevel 3 (since running X is so slow). Its only purpose is to play music from my server to my stereo so I haven't updated it in about 3 years. When I get home from work today I will try giving it an update and report back to you.

In the meanwhile try figuring out its local IP and ssh'ing to it like SpacePlod says. It sounds like its crapping out during the init process. Maybe you can retrieve a log from it if you can get in remotely.

poetgrant · 10-25-2019, 08:13 AM

Oh gosh, I can't believe I didn't think about SSH! That is one of my favorite things about computing! I will try it when I get home from work. Thank you for all the quick responses! I expected this to be one of those threads that people ignore because it is a boring problem. Hehehe

enorbet · 10-25-2019, 09:21 AM

If you are using LILO it might help to specify "vga = normal" and add "append = 'nomodeset'" so you insure a simple vesa-like display independent of any specific driver until you get a proper terminal login to be able to work on the system. SSH, of course, is a great solution, but if you run into any difficulty, try the above.

NOTE" The actual addon line looks like this in /etc/lilo.conf.

Code:

image = vmlinuz-foo
root = /dev/foo
label = Foo
append = "nomodeset"
read-only

allend · 10-25-2019, 09:48 AM

Quote:

All I can say is that I'm one of the only others here I know of who runs Slackware [occasionally] on a 32bit Atom netbook.

@Lysander666 - For me, just a bit more than occasionally.

@poetgrant - On my ancient eMachines eM250 I am running Slackware-current.
I boot with lilo, and have this in my /etc/lilo.conf

Code:

append = " vt.default_utf8=1 resume=/dev/sda6 resume_offset=0 video=intelfb:mode=1024x600-32@60 stack_guard_gap=6 nokaslr"

/dev/sda6 is my swap partition used for suspend to disk
video=intelfb:mode=1024x600-32@60 is what may help with your issue. Like your PC, this also has Intel GMA 950 graphics.
stack_guard_gap=6 is included to get around an issue with libreoffice
nokaslr is included as an option to get around an issue with resume from suspend on 32 bit architecture.
Thanks again to @Lysander666 for pointing me to this fix.

urbanwks · 10-25-2019, 10:04 AM

This is probably a long shot, but this happened to me once long ago when I ran 'update' and 'upgrade-all' but didn't run 'install-new'... something new to be installed was a dependency for something else.

Again, long shot especially compared to the other responses, but you didn't mention 'install-new', so might be something to consider.

Lysander666 · 10-25-2019, 10:09 AM

As far as I'm aware, the install-new option is only for -current?

allend · 10-25-2019, 10:23 AM

It is rare for new packages to be added to a stable release, but it has happened. For example, in 14.1

Quote:

Mon Oct 28 22:08:47 UTC 2013
...
kdei/calligra-l10n-ca@valencia-2.7.4-noarch-1.txz: Added.

abga · 10-25-2019, 11:00 AM

@poetgrant

I own an Acer Atom N270 which looks to have the same HW conf like your ASUS Eee PC 1005HAB.
I compared mine to this:
https://www.cnet.com/products/asus-eee-pc-1005ha/specs/

I'm running Slackware 14.2 - 32bit on it and it's updated to the last stand. Never experienced any hiccups (it just doesn't want to die) on this system and lately I'm also running my own 5.3.1 kernel compilation.
Some details:
https://en.wikipedia.org/wiki/Intel_GMA_950#GMA_950

- VGA controller lspci output:

Code:

00:02.0 VGA compatible controller: Intel Corporation Mobile 945GSE Express Integrated Graphics Controller (rev 03)
00:02.1 Display controller: Intel Corporation Mobile 945GM/GMS/GME, 943/940GML Express Integrated Graphics Controller (rev 03)

- lilo is set on vga = normal

- kernel version:
# uname -r
4.4.190-smp

- graphic driver related modules:

Code:

# lsmod | grep 915
i915                 1130183  1
drm_kms_helper        109599  1 i915
drm                   278702  3 i915,drm_kms_helper
intel_gtt              11990  3 i915,intel_agp
i2c_algo_bit            5382  1 i915
i2c_core               41937  8 drm,i915,i2c_i801,i2c_dev,drm_kms_helper,i2c_algo_bit,v4l2_common,videodev
fjes                   19156  0
button                  4787  1 i915
video                  27740  1 i915

- graphic driver kernel boot log (dmesg)

Code:

[    5.951007] [drm:intel_cpu_fifo_underrun_irq_handler [i915]] *ERROR* CPU pipe A FIFO underrun
[    5.951049] [drm:intel_cpu_fifo_underrun_irq_handler [i915]] *ERROR* CPU pipe B FIFO underrun
[    5.995279] [drm] Initialized i915 1.6.0 20151010 for 0000:00:02.0 on minor 0
[    6.824804] i915 0000:00:02.0: fb0: inteldrmfb frame buffer device
[    7.362083] [drm] GMBUS [i915 gmbus dpd] timed out, falling back to bit banging on pin 6
[    7.436076] [drm] GMBUS [i915 gmbus dpc] timed out, falling back to bit banging on pin 4
...
[   78.700389] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in gmch_pfit.lvds_border_bits (expected 0x00008000, found 0x00000000)
[   78.700485] WARNING: CPU: 0 PID: 1350 at drivers/gpu/drm/i915/intel_display.c:12826 intel_atomic_commit+0xaf9/0x12f0 [i915]()
[   78.700493] Modules linked in: tun xt_conntrack xt_tcpudp nf_log_ipv4 nf_log_common xt_string xt_pkttype xt_owner xt_multiport xt_iprange xt_DSCP xt_recent xt_state xt_limit xt_LOG iptable_nat nf_nat_ipv4 nf_nat iptable_mangle iptable_filter nf_conntrack_ipv4 nf_defrag_ipv4 nf_conntrack ip_tables x_tables fuse uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_v4l2 videobuf2_core v4l2_common videodev i2c_dev sparse_keymap acerhdf ath5k ath sdhci_pci mac80211 i915 coretemp hwmon drm_kms_helper evdev lpc_ich jmb38x_ms sdhci psmouse serio_raw memstick drm cfg80211 rfkill mmc_core i2c_i801 r8169 snd_hda_codec_realtek snd_hda_codec_generic mii uhci_hcd ehci_pci intel_agp intel_gtt agpgart snd_hda_intel snd_hda_codec i2c_algo_bit fb_sys_fops snd_hda_core syscopyarea sysfillrect sysimgblt i2c_core ehci_hcd
[   78.700790]  [<f9533c59>] ? intel_atomic_commit+0xaf9/0x12f0 [i915]
[   78.700872]  [<f9533c59>] ? intel_atomic_commit+0xaf9/0x12f0 [i915]
[   78.700960]  [<f9533c59>] intel_atomic_commit+0xaf9/0x12f0 [i915]
[   78.701046]  [<f9533160>] ? modeset_get_crtc_power_domains+0x180/0x180 [i915]

- and here is the 5.3.1 graphic driver kernel boot log (dmesg)

Code:

[    0.924486] i915 0000:00:02.0: vgaarb: deactivate vga console
[    0.926498] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
[    0.947220] [drm] Initialized i915 1.6.0 20190619 for 0000:00:02.0 on minor 0
[    0.988283] fbcon: i915drmfb (fb0) is primary device
[    1.928965] i915 0000:00:02.0: fb0: i915drmfb frame buffer device

poetgrant · 10-25-2019, 11:35 AM

So with all of this talk of LILO, when it tells me to run LiLo and I do, it also allows me what to do with the conf files. On my main rig, I have always just said to overwrite the old ones. Is that a mistake? And if I customize my LiLo conf, then I probably don't want to do that right?

poetgrant · 10-25-2019, 11:38 AM

Quote:

Originally Posted by hazel

Are you using a proprietary video driver?

Does the blackout only affect X or are virtual consoles blank as well? Have you tried booting to level 3 instead of 4?

Also try putting the caps lock down and see if the indicator lights up. If it does, you have a working console and only the video is kaput.

LiLo is almost finished loading Linux, then it just goes blank. No X, no TTY, the screen seems to just turn off. And the caps button does turn the light on, so when I get home I'll try to SSH into my machine.

Lysander666 · 10-25-2019, 12:16 PM

Quote:

Originally Posted by poetgrant

So with all of this talk of LILO, when it tells me to run LiLo and I do, it also allows me what to do with the conf files. On my main rig, I have always just said to overwrite the old ones. Is that a mistake? And if I customize my LiLo conf, then I probably don't want to do that right?

Overwriting configuration files can be problematic if you have configured those files for your machine[s] and needs. For instance, I have configured slackpkgplus, my networks.new file and my pulse audio config files. If I overwrote those with the newer versions, I would lose my changes, hence I select K for those after running slackpkg upgrade-all. However, and someone can correct me if I'm wrong, but I don't think lilo.conf is ever updated, PV has said as much and is the unofficial maintainer for that very reason - it doesn't need any improvements or fixes.