Headless Squeeze System
14/04/2018 15:30
Hello.

I built a little wireless router using a Dell GX-260 PC runningDebian Lenny, but I tracked down a few issues to the kernel and acouple of other packages in Lenny, so I decided to upgrade the routerto Squeeze. The bugs are gone, now, but now there is an issue withthe system that wasn't there under Lenny. Specifically, I can't runthe router headless. If I disconnect the monitor and reboot therouter, the system hangs and locks up tight. I believe I see where theproblem occurs. This is from the syslog: Feb 7 01:26:25 Cricket kernel: [ 32.794563] [drm] Initialized drm1.1.0 20060810Feb 7 01:26:25 Cricket kernel: [ 32.843296] i915 0000:00:02.0: PCIINT A -> GSI 16 (level, low) -> IRQ 16Feb 7 01:26:25 Cricket kernel: [ 32.843309] i915 0000:00:02.0: setting latency timer to 64Feb 7 01:26:25 Cricket kernel: [ 32.867580] [drm] set up 7M of stolenspaceFeb 7 01:26:25 Cricket kernel: [ 33.016606] [drm] DAC-5: set mode640x480 0Feb 7 01:26:25 Cricket kernel: [ 33.436394] ------------[ cuthere ]------------Feb 7 01:26:25 Cricket kernel: [ 33.436409] WARNING: at /tmp/buildd/linux-2.6-2.6.32/debian/build/source_i386_none/drivers/gpu/drm/drm_crtc_helper.c:1035drm_helper_initial_config+0x3a/0x53 [drm_kms_helper]()Feb 7 01:26:25 Cricket kernel: [ 33.436417] Hardware name: OptiPlexGX260 Feb 7 01:26:25 Cricket kernel: [ 33.436421] No connectors reportedconnected with modesFeb 7 01:26:25 Cricket kernel: [ 33.436425] Modules linked in: i91(+) drm_kms_helper drm i2c_algo_bit video output speedstep_libcpufreq_userspace cpufreq_conservative cpufreq_powersave cpufreq_statsiptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4ip_tables x_tables binfmt_misc loop sg sr_mod sd_mod crc_t10difsnd_intel8x0 snd_ac97_codec ac97_bus snd_pcm_oss snd_mixer_oss snd_pcmsnd_seq_midi snd_rawmidi snd_seq_midi_event snd_seq snd_timersnd_seq_device snd shpchp soundcore parport_pc i2c_i801 parport dcdbassnd_page_alloc pci_hotplug i2c_core psmouse processor rng_core pcspkrevdev serio_raw ext3 jbd mbcache usb_storage ide_cd_mod cdromide_gd_mod ata_generic libata scsi_mod ide_pci_generic uhci_hcd piixintel_agp floppy e1000 ehci_hcd button ide_core usbcore nls_baseagpgart thermal fan thermal_sys [last unloaded: scsi_wait_scan]Feb 7 01:26:25 Cricket kernel: [ 33.436543] Pid: 1446, comm: modprobeNot tainted 2.6.32-trunk-686 #1Feb 7 01:26:25 Cricket kernel: [ 33.436548] Call Trace: Feb 7 01:26:25 Cricket kernel: [ 33.436562] [<c1030ba1>] ?warn_slowpath_common+0x5e/0x8aFeb 7 01:26:25 Cricket kernel: [ 33.436594] [<e0db00c9>] ?i915_vga_set_decode+0x0/0x18 [i915]Feb 7 01:26:25 Cricket kernel: [ 33.436602] [<c1030bff>] ?warn_slowpath_fmt+0x26/0x2aFeb 7 01:26:25 Cricket kernel: [ 33.436610] [<e07a7c80>] ?drm_helper_initial_config+0x3a/0x53 [drm_kms_helper]Feb 7 01:26:25 Cricket kernel: [ 33.436637] [<e0db0ef9>] ?i915_driver_load+0xe18/0xed2 [i915]Feb 7 01:26:25 Cricket kernel: [ 33.436653] [<e080ce48>] ?drm_get_dev+0x2d3/0x397 [drm]Feb 7 01:26:25 Cricket kernel: [ 33.436663] [<c113fa5d>] ?local_pci_probe+0xb/0xcFeb 7 01:26:25 Cricket kernel: [ 33.436670] [<c11403a7>] ?pci_device_probe+0x41/0x63Feb 7 01:26:25 Cricket kernel: [ 33.436681] [<c11a1cc6>] ?driver_probe_device+0x8a/0x11eFeb 7 01:26:25 Cricket kernel: [ 33.436689] [<c11a1d9a>] ?__driver_attach+0x40/0x5bFeb 7 01:26:25 Cricket kernel: [ 33.436696] [<c11a1709>] ?bus_for_each_dev+0x37/0x5fFeb 7 01:26:25 Cricket kernel: [ 33.436703] [<c11a1b99>] ?driver_attach+0x11/0x13Feb 7 01:26:25 Cricket kernel: [ 33.436711] [<c11a1d5a>] ?__driver_attach+0x0/0x5bFeb 7 01:26:25 Cricket kernel: [ 33.436718] [<c11a11d1>] ?bus_add_driver+0x99/0x1c5Feb 7 01:26:25 Cricket kernel: [ 33.436726] [<c11a1fcb>] ?driver_register+0x87/0xe0Feb 7 01:26:25 Cricket kernel: [ 33.436733] [<c1140578>] ?__pci_register_driver+0x33/0x89Feb 7 01:26:25 Cricket kernel: [ 33.436758] [<e0973000>] ?i915_init+0x0/0x48 [i915]Feb 7 01:26:25 Cricket kernel: [ 33.436766] [<c100113e>] ?do_one_initcall+0x55/0x155Feb 7 01:26:25 Cricket kernel: [ 33.436774] [<c10579f6>] ?sys_init_module+0xa7/0x1d7Feb 7 01:26:25 Cricket kernel: [ 33.436783] [<c10030fb>] ?sysenter_do_call+0x12/0x28Feb 7 01:26:25 Cricket kernel: [ 33.436789] ---[ end trace9cbaaba06076f10a ]---Feb 7 01:26:25 Cricket kernel: [ 33.440142] [drm] Initialized i9151.6.0 20080730 for 0000:00:02.0 on minor 0Feb 7 01:26:26 Cricket kernel: [ 33.548859] [drm] DAC-5: set mode640x480 0Feb 7 01:26:26 Cricket avahi-daemon[1286]: Joining mDNS multicast groupon interface eth0.IPv6 with address fe80::208:74ff:fe2a:4782.

Feb 7 01:26:26 Cricket avahi-daemon[1286]: New relevant interfaceeth0.IPv6 for mDNS.

Feb 7 01:26:26 Cricket avahi-daemon[1286]: Registering new addressrecord for fe80::208:74ff:fe2a:4782 on eth0.*.

Feb 7 01:26:26 Cricket kernel: [ 34.043246] [drm] DAC-5: set mode640x480 0Feb 7 01:33:22 Cricket kernel: imklog 4.4.2, log source = /proc/kmsgstarted.

Feb 7 01:33:22 Cricket rsyslogd: [origin software=rsyslogdswVersion=4.4.2 x-pid=1063 x-info=http://www.rsyslog.com](re)startFeb 7 01:33:22 Cricket kernel: [ 0.000000] Initializing cgroupsubsys cpuset As you can see, the system hangs shortly after the warning is issued. I don't see this warning in the log if the monitor is attached. Looking for what I think is the offending kernel module I get: Cricket:/var/log# lsmod | grep drmdrm_kms_helper 17183 1 i915drm 107507 3 i915,drm_kms_helperi2c_core 12612 4 i915,drm,i2c_algo_bit,i2c_i801agpgart 19516 2 drm,intel_agp Should I simply remove the modules drm_kms_helper and drm? What abouti2c_core and agpart? The modules are in use, so modprobe won't removethem, will it? This system is sitting 60 miles away from me, so if Ireboot the shelf and it doesn't come up, I've got a problem. I need tomake sure whatever I do won't result in an irretrievably hung system.


Source is Usenet: alt.os.linux.debian
Sign in to add a comment

Answer score: 5
14/04/2018 15:30 - I demand that lrhorer may or may not have written...

[snip] Not immediately obvious from what you posted; had your UA not hard-wrappedeverything, I might have been inclined to examine that more closely...

[snip][snip] Either would do; the file won't be altered or restored without either firstpurging the package or making use of dpkg's --force-confmiss option.

I don't think so.

Fair enough, but you did mention a monitor; I don't find it unreasonable toassume that it's possible that the monitor may be there permanently.

[snip reasons why it won't be] If you're not normally making use of its graphics hardware then you don'tneed that or, for that matter, an X server. (X client libs etc. are aseparate matter, of course.) You could load vesafb instead of intelfb or i915, though I understand thatcurrent grub2 may cause use of efifb.

[snip] Removing whichever X login manager is installed will do that.

It'll be upgraded, I expect: security fixes, for a start. Also remember thatsqueeze isn't yet released; freeze is planned for March but looks very muchlike it won't happen /quite/ that soon...


Source is Usenet: alt.os.linux.debian
Sign in to add a comment

Answer score: 5
14/04/2018 15:30 - Yeah, really. A normal boot sequence takes about 100 seconds. Withthe monitor attached, there are a couple of dozen additional lines fromkernel modules in syslog before one can see all sorts of real-timeprocesses (ntp, dnsmasq, etc) reporting various normal activity. Inthe snippet I provided, the last response from anything was: Feb 7 01:26:26 Cricket kernel: [ 34.043246] [drm] DAC-5: set mode640x480 0 thirty four seconds into the boot sequence, then nothing whatsoeveruntil I had to hit the big red switch and then the system startedrebooting at 01:33:22.

Based on your suggestion, I did a little research using Google. Onearticle suggested I could merely comment out the options i915modeset=1 line in the file. This would seem to be a better solutionthan removing the file entirely, if it will work. I take it KMS isn'tloaded in initramfs? This is a *HEADLESS* system. It's an embedded firewall router.

1. It must have no keyboard, monitor, or mouse. The only time it willhave any of these attached is if the router fails to boot and needs tobe brought back to me for maintenance.

2. It would be absolutely boneheaded to purchase a monitor (even a $150one) which will never be used just so the system will boot on thosevery rare occasions when it needs to be brought down.

3. This little router sits in a full stereo cabinet. There is no placefor a monitor. In order to get the router to boot, the monitor issitting on the floor. My sister doesn't want a monitor sitting on thefloor in her livingroom indefinitely. I don't blame her. I also needmy monitor back. An $800 monitor sitting on my sister's living roomfloor is not a good use of my money.

Installed? Yes. I'm not sure it needs to be. In use? I don't know. Note I don't care in the least how fast or slow Gnome or KDE load, orwhether the screen resolution is optimal or not.

All I need is ssh / telnet / ftp access to the unit over the VPN I haveset up. It does need to boot to the console when the system is broughtback for maintenance (because the network won't come up). Otherwise,it needs no real monitor support. It would be perfectly fine for thesystem to boot to a login prompt, rather than a GUI. This is anembedded application, and I am not sure it will ever be upgraded, buteven if it is, it still must boot without the presence of a monitor.


Source is Usenet: alt.os.linux.debian
Sign in to add a comment

Answer score: 5
14/04/2018 15:30 - I demand that lrhorer may or may not have written...

[snip log; too badly wrapped] Not really...

You need do nothing about them.

You could try removing /etc/modprobe.d/i915-kms.conf; or you could leave themonitor attached but cause it to be switched off via DPMS after a certaintime.

However, if xserver-xorg-video-intel is installed and in use, note thatupgrading that to 2.10 (not yet in Debian) will require that KMS is enabled.


Source is Usenet: alt.os.linux.debian
Sign in to add a comment

eDiscover
Helpforce eDiscover provides technical articles updated each dayHelpforce eDiscover RSS feed contains the latest technical articles in RSS
Click the logo to go back to the main page
Search eDiscover
  
Categories

Click an icon to go to that category

Helpforce eDiscover contains articles about Microsoft Windows Helpforce eDiscover contains articles about Apple products and MacOS Helpforce eDiscover contains articles about Linux and POSIX operating systems Helpforce eDiscover contains articles about Helpforce Helpforce has a large variety of technical information and articles for you to read Helpforce eDiscover contains articles about databases, MYSQL, SQL Server Oracle Helpforce eDiscover contains articles about Java, JVM and the JRE Helpforce eDiscover contains articles about the QNX operating system Helpforce eDiscover contains articles about Oracle Solaris and Open Solaris Helpforce eDiscover contains articles about RISC OS, Acorn and the BBC Micro Helpforce eDiscover contains articles about Amiga and AmigaOS

Type your comment into the box below