- Grml - lists.mur.at

[Grml] slow start of xserver , very poor responsiveness with nohwcap
by Erich Minderlein 25 Nov '06

25 Nov '06

Hello see also my contributions @ http://lists.mur.at/pipermail/grml/ where I followed before on this subject Hardware : Amilo Pro Notebook V2030 Inter celeron , 1,5 GHz, 256 MByte RAM, VGA Via <6>ACPI: Video Device [VGA] (multi-head: yes rom: no post: no) -- CGA/Mono Video supported ACPI supported USB Legacy supported AGP supported -- lspci -vvs 0000:01:00.0 >> 01:00.0 VGA compatible controller: VIA Technologies, Inc. UniChrome Pro IGP (rev 01) (prog-if 00 [VGA]) Subsystem: Fujitsu Siemens Computer GmbH Unknown device 109b Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 10 Memory at f0000000 (32-bit, prefetchable) [size=64M] Memory at d1000000 (32-bit, non-prefetchable) [size=16M] Capabilities: [60] Power Management version 2 Capabilities: [70] AGP version 2.0 I was wondering why this beast was slower in starting X than my (with display mechanically broken Toshiba Portege 3110CT) Subnotebook with PentiumII 300Mhz, 192 MByte and a 7 GB HDD. from the last millenium. (I use it as xterminal right now!) sw: a system mixed from unstable and the grml distro kernel 2.6.17-grml I did init 5 (for Xserver ) ps ax and checked PID for the gdm processes strace -f ... PID (2*) ALT F7 login in with my $USER $PASSWORD CTRL ALT F1 init 2 I found that 427 accesses for /etc/ld.so.nohwcap were happening, and no file there. touch /etc/ld.so.nohwcap and it maked 822 accesses, but slightly faster. I think this is a performance killer. Mailingliste grml recommended to report it to debian-x, so here we go. Please feel free to ask more, also that I did deinstall this Xorg server now. regards -- Erich Minderlein <erminderlein(a)locoware.de> MARC Mannheim Rail Commerce GmbH

1 0

amilopro slow : solution (part if it)
by Erich Minderlein 25 Nov '06

25 Nov '06

Hello during a login to the x-server the strace reveals that 472 file accesses to /etc/ld.so.nohwcap result with -1, because the file is nowhere and also not there. touch /etc/ld.so.nohwcap helps partly, it would even be better, if the file is not a million times asked for. best regards -- Erich Minderlein <erminderlein(a)locoware.de> MARC Mannheim Rail Commerce GmbH

3 2

kernel 2.6.17-grml & software suspend to disk / swap
by Erich Minderlein 24 Nov '06

24 Nov '06

Hello What is provided in the kernel to do softwaresuspend ? I would like to avoid to compile a new kernel, what will you recommand to get softwaresuspend aka hibernate working? regards -- Erich Minderlein <erminderlein(a)locoware.de> MARC Mannheim Rail Commerce GmbH

2 1

[Fwd: Re: startx minor bug : argument -listen instaed of -nolisten]
by Erich Minderlein 23 Nov '06

23 Nov '06

-------- Weitergeleitete Nachricht -------- Von: Julien Cristau <jcristau(a)liafa.jussieu.fr> An: Erich Minderlein <erminderlein(a)locoware.de> Kopie: Debian X Strike Force <debian-x(a)lists.debian.org> Betreff: Re: startx minor bug : argument -listen instaed of -nolisten Datum: Thu, 23 Nov 2006 15:34:14 +0100 On Thu, Nov 23, 2006 at 14:59:40 +0100, Erich Minderlein wrote: > Hello > I don't know if I am up to date : > I see this bug and it did not go away in an upgrade, > so i fixed it from hand : > however if it comes from some debian package , it's reason must be found > $ GET http://necrotic.deadbeast.net/svn/xorg-x11/tags/app/xbase-clients/xbase-cli… #!/bin/sh # $Id: xserverrc 189 2005-06-11 00:04:27Z branden $ exec /usr/bin/X11/X -dpi 100 -nolisten tcp --- You've probably modified this file yourself, because as far as I know it's been this way since a long time in the debian packages. Cheers, Julien

1 0

xinit , xinitrc, startx
by Erich Minderlein 23 Nov '06

23 Nov '06

Hallo when using the binaries from sarge I have dpkg -l xbase-clients GewÃ¼nscht=Unbekannt/Installieren/R=Entfernen/P=SÃ¤ubern/Halten | Status=Nicht/Installiert/Config/U=Entpackt/Fehlgeschl. Konf./Halb install. |/ Fehler?=(keiner)/Halten/R=Neuinst. notw/X=beides (Status, Fehler: GROÃŸ=schlecht) ||/ Name Version Beschreibung +++-================-=========================-====================== ii xbase-clients 4.3.0.dfsg.1-14sarge2 miscellaneous X clients dpkg -S xserverrc xbase-clients: /etc/X11/xinit/xserverrc grep listen /etc/X11/xinit/xserverrc exec /usr/bin/X11/X -dpi 100 -nolisten tcp ==============================^ when using the binaries from testing/unstable on a grml system (amilopro) I have dpkg -l xbase-clients GewÃ¼nscht=Unbekannt/Installieren/R=Entfernen/P=SÃ¤ubern/Halten | Status=Nicht/Installiert/Config/U=Entpackt/Fehlgeschl. Konf./Halb install. |/ Fehler?=(kein)/Halten/R=Neuinst notw/X=beide (Status, Fehler: GROSS=schlecht) ||/ Name Version Beschreibung +++-=================-====================-================ ii xbase-clients 7.1.ds-3 miscellaneous X clients dpkg -S xserverrc xbase-clients: /etc/X11/xinit/xserverrc grep listen /etc/X11/xinit/xserverrc exec /usr/bin/X11/X -dpi 100 -listen tcp ==============================^ when I do /usr/bin/X11/X the xserver starts. when I do startx the x server does not start, it does not accept the argument "-listen" . ( as planned, but why what or who changed the script ? grml or package maintainer ? and who must fix it? ) regards -- Erich Minderlein <erminderlein(a)locoware.de> MARC Mannheim Rail Commerce GmbH

2 1

AMILOPRO V2030 sluggish : cpu frequency modulation slow speedstep-centrino vs. p4-clockmod
by Erich Minderlein 22 Nov '06

22 Nov '06

Hello Frequency is increasing to the limit of 1500 MHz and keyboard response is much better during the test, then decreasing again. Also with other processes frequency scaling by userspace is varying. Thank you, with this inquiry I found out, that the problem is with another issue, (X.org), this will not bother you and I follow up there. Thank you again > date: Tue, 21 Nov 2006 21:46:25 +0100 > From: Michael Prokop <mika(a)grml.org> > Subject: Re: [Grml] AMILOPRO V2030 sluggish : cpu frequency modulation > slow speedstep-centrino vs. p4-clockmod > To: grml(a)mur.at > Message-ID: <2006-11-21T21-30-36(a)devnull.michael-prokop.at> > Content-Type: text/plain; charset="iso-8859-1" > > * Erich Minderlein <erminderlein(a)locoware.de> [20061121 21:29]: > > > cpufrequtils 002: cpufreq-info (C) Dominik Brodowski 2004-2006 > [...] > > Hardwarebedingte Grenzen der Taktfrequenz: 188 MHz - 1.50 GHz > > mgliche Taktfrequenzen: 188 MHz, 375 MHz, 563 MHz, 750 MHz, 938 MHz, > > 1.13 GHz, 1.31 GHz, 1.50 GHz > > mgliche Regler: userspace, performance > > momentane Taktik: die Frequenz soll innerhalb 188 MHz und 1.50 GHz. > > liegen. Der Regler "userspace" kann frei > > entscheiden, > > welche Taktfrequenz innerhalb dieser Grenze > > verwendet wird. > > momentane Taktfrequenz ist 188 MHz (verifiziert durch Nachfrage bei > > der Hardware). > > That looks OK for me - basically. Are powernowd or cpudyn running, > right? > > > this is not what we want ! > > so the change of frequency is not allowed as I understand > > No, your system should adjust cpu frequency according to the needs. > Services like powernowd can handle that for you. You can adjust the > limits manually as well, see > http://acpi.sourceforge.net/documentation/processor.html > > If you do not want to use frequency scaling at all you just have to > unload the cpu-frequency modules of course. > > But let's check whether your system really adjusts frequency > according to your needs. So just run cpuburn-in and take a look at > the current processor frequency (running cpu-screen, > cpufreq-info,... or just 'cat > /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq'). Is it still > at 188MHz is it increasing? > > Whover choose the smp for a single cpu laptop or is this double core ? > > Hm? > > The SMP-alternatives feature of recent kernels allows use of SMP > enabled kernels on unicore systems, nothing to care about. > thanks understood this, now one kernel variant less

1 0

AMILOPRO V2030 sluggish : cpu frequency modulation slow speedstep-centrino vs. p4-clockmod
by Erich Minderlein 21 Nov '06

21 Nov '06

Hello oder Hallo cpufrequtils 002: cpufreq-info (C) Dominik Brodowski 2004-2006 Bitte melden Sie Fehler an linux(a)brodo.de. analysiere CPU 0: Treiber: p4-clockmod Folgende CPUs kÃ¶nnen nur gleichzeitig ihre Frequenz variieren: 0 Hardwarebedingte Grenzen der Taktfrequenz: 188 MHz - 1.50 GHz mÃ¶gliche Taktfrequenzen: 188 MHz, 375 MHz, 563 MHz, 750 MHz, 938 MHz, 1.13 GHz, 1.31 GHz, 1.50 GHz mÃ¶gliche Regler: userspace, performance momentane Taktik: die Frequenz soll innerhalb 188 MHz und 1.50 GHz. liegen. Der Regler "userspace" kann frei entscheiden, welche Taktfrequenz innerhalb dieser Grenze verwendet wird. momentane Taktfrequenz ist 188 MHz (verifiziert durch Nachfrage bei der Hardware). this is not what we want ! so the change of frequency is not allowed as I understand uname -a : Linux amilopro 2.6.17-grml #1 SMP PREEMPT Tue Jul 25 10:57:35 CEST 2006 i686 GNU/Linux Whover choose the smp for a single cpu lap top oris this double core ? regards -- Erich Minderlein <erminderlein(a)locoware.de> MARC Mannheim Rail Commerce GmbH

2 1

Terminal autologin
by Shehu Dikko 21 Nov '06

21 Nov '06

Where is the tty autologin stuff configured? I'd like to switch things around but can't locate the config files. regards s h e h u --

2 2

JFYI: mail policy /spamfilters of mailinglist server
by Michael Prokop 20 Nov '06

20 Nov '06

Hoi, just to inform you about the mail policy of the grml mailinglist(s), thanksfully hosted at mur.at: Since 14th of November 2006 a new mail policy regarding the spamfilters is active, see http://mur.at/support/mail/policy-1?set_language=en for details. If you don't know what to do right now just forget my mail. ;) But if you *might* notice any problems with sending mails to the grml mailinglist please let me know. Just send me a personal mail (off the list) providing bounce details/logs so we (mainly Jogi of NOC @ mur.at) can investigate. JFYI && regards, -mika- -- http://grml.org/ # Linux for texttool-users and sysadmins http://wiki.grml.org/ # share your knowledge http://grml.supersized.org/ # the grml development weblog #grml @ irc.freenode.org # meet us on irc

1 0

AMILOPRO V2030 sluggish : cpu frequency modulation slow ? speedstep-centrino vs. p4-clockmod
by Erich Minderlein 20 Nov '06

20 Nov '06

Hi Google is responding poorly with this problem. My AMILOPRO V2030 with 1.5 GHertz centrino gives the following output: /var/log/kern.log: Nov 20 16:33:49 amilopro kernel: p4-clockmod: Warning: Pentium M detected. The speedstep_centrino module offers voltage scaling in addition of frequency scaling. You should use that instead of p4-clockmod, if possible. Nov 20 16:33:49 amilopro kernel: p4-clockmod: P4/Xeon(TM) CPU On-Demand Clock Modulation available /proc/cpuinfo: processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : Intel(R) Celeron(R) M processor 1.50GHz stepping : 8 cpu MHz : 187.500 cache size : 1024 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss tm pbe up bogomips : 3019.16 I have put p4-clockmod in /etc/modprobe/blacklist , but output of kern.log remains the same after init 6 locate speedstep-centrino shows the driver is there. What else can I doo to use correct driver and crank cpu speed up after keypressed ? best regards -- Erich Minderlein <erminderlein(a)locoware.de> MARC Mannheim Rail Commerce GmbH

2 1