more on hardware clock

more on hardware clock

[Charles Hallenbeck]

When I do "strace hwclock --systohc" I get a continuous output of
trace messages that have to be interrupted eventually with
control-C. I have no idea what they mean, except the fact that
they continuous means that the clock utility is not hanging, it
is looping.

I give up!! Back to the abacus!


-- 
The Moon is Waxing Gibbous (81% of Full)
So visit me at http://www.valstar.net/~hallenbeck

Re: more on hardware clock

[Ari Moisio]

Hi!

Charles Hallenbeck 16.10.02:

>When I do "strace hwclock --systohc" I get a continuous output of
>trace messages that have to be interrupted eventually with
>control-C. I have no idea what they mean, except the fact that
>they continuous means that the clock utility is not hanging, it
>is looping.

  Quite strange. What kind of system calls they are?

  Btw: do you have other clock sync daemons running, ntp or similar?

-- 
Mr. Ari Moisio, Niittykatu 5, 41160 Tikkakoski
gsm: +358-40-5055239, icq: 164241561

Re: more on hardware clock

[Charles Hallenbeck]

Hi Ari,

No other time utilities are running. I have added a "netdate"
command to my ip-up script to get the current time from the
internet, but it only runs once and seems okay. Itt has been
there for over a year now.

I just ran "strace hwclock -w" again redirecting output to a
file, and after waiting for 15 seconds I hit control-C and the
resulting file was 25 MB long. I em inserting the last 21 lines
of that file in the hope that you or someone will be able to spot
something familiar. I am way over my head on this one.

----------
send(3, "<82>Oct 16 12:24:52 libsafe.so[3"..., 52, 0) = 52
rt_sigaction(SIGPIPE, {SIG_DFL}, NULL, 8) = 0
time([1034785492])                      = 1034785492
getpid()                                = 399
rt_sigaction(SIGPIPE, {0x400fa78c, [], 0x4000000}, {SIG_DFL}, 8) = 0
send(3, "<82>Oct 16 12:24:52 libsafe.so[3"..., 52, 0) = 52
rt_sigaction(SIGPIPE, {SIG_DFL}, NULL, 8) = 0
time([1034785492])                      = 1034785492
getpid()                                = 399
rt_sigaction(SIGPIPE, {0x400fa78c, [], 0x4000000}, {SIG_DFL}, 8) = 0
send(3, "<82>Oct 16 12:24:52 libsafe.so[3"..., 52, 0) = 52
rt_sigaction(SIGPIPE, {SIG_DFL}, NULL, 8) = 0
time([1034785492])                      = 1034785492
getpid()                                = 399
rt_sigaction(SIGPIPE, {0x400fa78c, [], 0x4000000}, {SIG_DFL}, 8) = 0
send(3, "<82>Oct 16 12:24:52 libsafe.so[3"..., 52, 0) = 52
rt_sigaction(SIGPIPE, {SIG_DFL}, NULL, 8) = 0
time([1034785492])                      = 1034785492
getpid()                                = 399
rt_sigaction(SIGPIPE, {0x400fa78c, [], 0x4000000}, {SIG_DFL}, 8) = 0
send(3, "<82>Oct 16 12:24:52 libsafe.so[3"..., 52, 0) = 52
----------

Thanks for your help and suggestions

Chuck

-- 
The Moon is Waxing Gibbous (82% of Full)
So visit me at http://www.valstar.net/~hallenbeck

Re: more on hardware clock

[Charles Hallenbeck]

Ari!!!

Looking more closely at this output I see that "libsafe" is still
plaguing me. I removed that package and eliminated the
environment variable that sets it up and have rebooted since
then, but the thing is still executing!!! I will hunt it down and
destroy it!!!


-- 
The Moon is Waxing Gibbous (82% of Full)
So visit me at http://www.valstar.net/~hallenbeck

Re: more on hardware clock

[Ari Moisio]

Hi!

Charles Hallenbeck 16.10.02:

>Looking more closely at this output I see that "libsafe" is still
>plaguing me. I removed that package and eliminated the
>environment variable that sets it up and have rebooted since
>then, but the thing is still executing!!! I will hunt it down and
>destroy it!!!

  Interesting... i visited google and found following url

http://bugs.debian.org/cgi-bin/bugreport.cgi?archive=no&bug=136946

  so libsafe really is the problem. Not sure if given instructions will
help.


>
>

-- 
Mr. Ari Moisio, Niittykatu 5, 41160 Tikkakoski
gsm: +358-40-5055239, icq: 164241561

Re: more on hardware clock

[Kerry Hoath]

Ok; we have managed to determine that clock under Linux
fails; can you set the clock from a dos boot disk
and does it stay set?
I have  never seen hwclock fail in all the 9 years
I have been running Linux. It was just called clock back then.
On Wed, Oct 16, 2002 at 11:34:38AM -0400, Charles Hallenbeck wrote:
> When I do "strace hwclock --systohc" I get a continuous output of
> trace messages that have to be interrupted eventually with
> control-C. I have no idea what they mean, except the fact that
> they continuous means that the clock utility is not hanging, it
> is looping.
> 
> I give up!! Back to the abacus!
> 
> 
> -- 
> The Moon is Waxing Gibbous (81% of Full)
> So visit me at http://www.valstar.net/~hallenbeck
> 
> 
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
> 

-- 
Kerry Hoath:  kerry@gotss.net kerry@gotss.eu.org or  kerry@gotss.spice.net.au
ICQ: 8226547 msn: kerry@gotss.net Yahoo: kerryhoath@yahoo.com.au

Re: more on hardware clock

[Charles Hallenbeck]

Hi Ari and all,

I really blew it yesterday when I hastily deleted the residual
libsafe pieces that the Slackware "removepkg" did not get rid of.
It turns out that the damn thing had modified the linux dynamic
loader to invoke libsafe every time it was executed, and without
the libsafe component I threw away, the dynamic loader was
broken. Try doing ANYTHING on a Linux system with a broken
dynamic loader!!!

Of course I made it worse in my panic and anger, and ended up
having to reinstall my root partition with a new base system and
then scrambling to get my /usr and /home partitions back on
board.

I guess it pays to be more cautious, but I was really on a roll
with putting new stuff on this box and added libsafe hastily.

I appreciate everybody's help with this.

Chuck


-- 
The Moon is Waxing Gibbous (89% of Full)
So visit me at http://www.valstar.net/~hallenbeck

Re: more on hardware clock

[Steve Holmes]

Tell me, just what the hell is libsafe and what is it good for?  Based
on Chuck's experience, I don't think I will be doin' libsafe any time soon:).

On Thu, Oct 17, 2002 at 03:59:57PM -0400, Charles Hallenbeck wrote:
> Hi Ari and all,
> 
> I really blew it yesterday when I hastily deleted the residual
> libsafe pieces that the Slackware "removepkg" did not get rid of.
> It turns out that the damn thing had modified the linux dynamic
> loader to invoke libsafe every time it was executed, and without
> the libsafe component I threw away, the dynamic loader was
> broken. Try doing ANYTHING on a Linux system with a broken
> dynamic loader!!!
> 
> Of course I made it worse in my panic and anger, and ended up
> having to reinstall my root partition with a new base system and
> then scrambling to get my /usr and /home partitions back on
> board.
> 
> I guess it pays to be more cautious, but I was really on a roll
> with putting new stuff on this box and added libsafe hastily.
> 
> I appreciate everybody's help with this.
> 
> Chuck
> 
> 
> -- 
> The Moon is Waxing Gibbous (89% of Full)
> So visit me at http://www.valstar.net/~hallenbeck
> 
> 
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup

-- 
Please avoid sending me Word or PowerPoint attachments.
   See http://www.fsf.org/philosophy/no-word-attachments.html

Re: more on hardware clock

[Charles Hallenbeck]

Hi Steve,

On Thu, 17 Oct 2002, Steve Holmes wrote:

> Tell me, just what the hell is libsafe and what is it good for?  Based
> on Chuck's experience, I don't think I will be doin' libsafe any time soon:).

Libsafe is an evil device created by Osama Bin Ladin which is
designed to bring down the most stable systems in the Western
World!!! I am reporting it to John Ashcroft in the morning.

You will find it in the "extra" directory of Slackware 8.1 and
the description is actually very intriguing. It seems to do the
same thing for ld-linux.so that vsound does for /dev/dsp -- it
intercepts its invocations and attempts to plug a number of
common security holes in lots of software related to buffer
overruns and such. My warmest thanks to Jude who called it to our
attention a week or so ago!!!

Hey Jude, have you tried shutting your computer down yet? <smile>

Chuck


-- 
The Moon is Waxing Gibbous (90% of Full)
So visit me at http://www.valstar.net/~hallenbeck

Re: more on hardware clock

[Gregory Nowak]

On Thu, Oct 17, 2002 at 07:23:52PM -0400, Charles Hallenbeck wrote:
> the description is actually very intriguing. It seems to do the
> same thing for ld-linux.so that vsound does for /dev/dsp -- it
> intercepts its invocations and attempts to plug a number of
> common security holes in lots of software related to buffer
> overruns and such. 

Hmm, and it hung your system when running clock, right?

This fact along with your description of what it does above smells very much of Macroslop.

Owo, the Redmond boys have infiltrated the free software movement, help us all! <smile>

Greg

suggestion for disastery recovery (was Re: more on hardware clock)

[Adam Myrow]

For those situations where you've installed something that messed your
whole system up, or if you scramble your partition table and have to start
over, I have a little utility to suggest.  http://dump.sourceforge.net
contains Ext2 dump and restore.  This is a simple backup utility based on
the dump and restore that first appeared in BSD and later got renamed to
ufsdump and ufsrestore in Solaris.  Basically, it makes backup and
restoration fairly simple.  It was originally designed for tapes, but can
write to files just as easily.  It's a bit tricky to install, because it
wants to set the owner of the man pages to "man" which seldom exists.
Look at the configure options carefully.  What I did is to build the
package in a statically linked form.  Then, I put a copy of this
statically linked restore on a floppy.  That way, I could get to it from a
rescue disk and it would work no matter what libraries are on the rescue
disk.  It lets you do a full restore or an interactive restore where you
can type "ls" and "cd" just like in the shell to navigate the backup and
select files.  One word of warning, it is slow about restoring single
files from the middle of a backup since it's designed for tapes, but there
are options to help with this.  However, a full or incremental backup and
restore are very quick.  It can take a 2GB partition and can back it up in
about 10 minutes uncompressed on my system.  It takes just under 2 hours
with compression, but it packs that 2GB into about 790MB with the highest
compression enabled.  Give it a look.  I wish Slackware included it by
default, but it doesn't.

Re: more on hardware clock

[Charles Hallenbeck]

Hi Greg,
On Thu, 17 Oct 2002, Gregory Nowak wrote:

>
> Hmm, and it hung your system when running clock, right?

Yes. Of course you normally don't run hwclock very often, but the
Slackware shutdown procedure attempts to write the system time to
the hardware clock, and so your shutdown is stopped in its
tracks.

-- 
The Moon is Waxing Gibbous (90% of Full)
So visit me at http://www.valstar.net/~hallenbeck

Re: suggestion for disastery recovery (was Re: more on hardware clock)

[Charles Hallenbeck]

Thanks, Adam,

It might be slow pulling a single file out of its context, but it
cannot be slower than the way I did it!

However -- was there something about clouds and silver linings? I
had to patch speakup into a clean kernel source tree and begin
with "make config" from scratch, and guess what?

alsa RC3 just compiled without error.

Chuck

On Thu, 17 Oct 2002, Adam Myrow wrote:

> For those situations where you've installed something that messed your
> whole system up, or if you scramble your partition table and have to start
> over, I have a little utility to suggest.  http://dump.sourceforge.net
> contains Ext2 dump and restore.  This is a simple backup utility based on
> the dump and restore that first appeared in BSD and later got renamed to
> ufsdump and ufsrestore in Solaris.  Basically, it makes backup and
> restoration fairly simple.  It was originally designed for tapes, but can
> write to files just as easily.  It's a bit tricky to install, because it
> wants to set the owner of the man pages to "man" which seldom exists.
> Look at the configure options carefully.  What I did is to build the
> package in a statically linked form.  Then, I put a copy of this
> statically linked restore on a floppy.  That way, I could get to it from a
> rescue disk and it would work no matter what libraries are on the rescue
> disk.  It lets you do a full restore or an interactive restore where you
> can type "ls" and "cd" just like in the shell to navigate the backup and
> select files.  One word of warning, it is slow about restoring single
> files from the middle of a backup since it's designed for tapes, but there
> are options to help with this.  However, a full or incremental backup and
> restore are very quick.  It can take a 2GB partition and can back it up in
> about 10 minutes uncompressed on my system.  It takes just under 2 hours
> with compression, but it packs that 2GB into about 790MB with the highest
> compression enabled.  Give it a look.  I wish Slackware included it by
> default, but it doesn't.
>
>
>
>
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
>

-- 
The Moon is Waxing Gibbous (90% of Full)
So visit me at http://www.valstar.net/~hallenbeck

Re: more on hardware clock

[Igor Gueths]

You could always remove this by editing some of the sutdown scripts that
are run from init.d/rc.d I believe on slackware. I had to do this when
Woody played with it (upgrade process). I have no idea why it decided to
play wit shutdown, but in any case I had to edit it back to a working
state.

Microsoft dialogue
This company has performed an illegal operation and will be shutdown. If this problem persists, delete Winblows and install Linux. Close button

On Thu, 17 Oct 2002, Charles Hallenbeck wrote:

>
> Hi Greg,
> On Thu, 17 Oct 2002, Gregory Nowak wrote:
>
> >
> > Hmm, and it hung your system when running clock, right?
>
> Yes. Of course you normally don't run hwclock very often, but the
> Slackware shutdown procedure attempts to write the system time to
> the hardware clock, and so your shutdown is stopped in its
> tracks.
>
> --
> The Moon is Waxing Gibbous (90% of Full)
> So visit me at http://www.valstar.net/~hallenbeck
>
>
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
>

Re: more on hardware clock

[Jude DaShiell]

Unfortunately, I did install libsafe.  What it is supposed to protect
against is frame smashing attacks on your system that viruses can do.



On Thu, 17 Oct 2002, Steve Holmes wrote:

> Tell me, just what the hell is libsafe and what is it good for?  Based
> on Chuck's experience, I don't think I will be doin' libsafe any time soon:).
>
> On Thu, Oct 17, 2002 at 03:59:57PM -0400, Charles Hallenbeck wrote:
> > Hi Ari and all,
> >
> > I really blew it yesterday when I hastily deleted the residual
> > libsafe pieces that the Slackware "removepkg" did not get rid of.
> > It turns out that the damn thing had modified the linux dynamic
> > loader to invoke libsafe every time it was executed, and without
> > the libsafe component I threw away, the dynamic loader was
> > broken. Try doing ANYTHING on a Linux system with a broken
> > dynamic loader!!!
> >
> > Of course I made it worse in my panic and anger, and ended up
> > having to reinstall my root partition with a new base system and
> > then scrambling to get my /usr and /home partitions back on
> > board.
> >
> > I guess it pays to be more cautious, but I was really on a roll
> > with putting new stuff on this box and added libsafe hastily.
> >
> > I appreciate everybody's help with this.
> >
> > Chuck
> >
> >
> > --
> > The Moon is Waxing Gibbous (89% of Full)
> > So visit me at http://www.valstar.net/~hallenbeck
> >
> >
> > _______________________________________________
> > Speakup mailing list
> > Speakup@braille.uwo.ca
> > http://speech.braille.uwo.ca/mailman/listinfo/speakup
>
> --
> Please avoid sending me Word or PowerPoint attachments.
>    See http://www.fsf.org/philosophy/no-word-attachments.html
>
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
>

Re: more on hardware clock

[Jude DaShiell]

Hi Chuck, yes I've shut down successfully several times.  Interesting
since I'm using slackware 8.0.0 not 8.1.  It may be a fix to 8.0.0 that
made slackware 8.1 actually broke libsafe.



On Thu, 17 Oct 2002, Charles Hallenbeck wrote:

> Hi Steve,
>
> On Thu, 17 Oct 2002, Steve Holmes wrote:
>
> > Tell me, just what the hell is libsafe and what is it good for?  Based
> > on Chuck's experience, I don't think I will be doin' libsafe any time soon:).
>
> Libsafe is an evil device created by Osama Bin Ladin which is
> designed to bring down the most stable systems in the Western
> World!!! I am reporting it to John Ashcroft in the morning.
>
> You will find it in the "extra" directory of Slackware 8.1 and
> the description is actually very intriguing. It seems to do the
> same thing for ld-linux.so that vsound does for /dev/dsp -- it
> intercepts its invocations and attempts to plug a number of
> common security holes in lots of software related to buffer
> overruns and such. My warmest thanks to Jude who called it to our
> attention a week or so ago!!!
>
> Hey Jude, have you tried shutting your computer down yet? <smile>
>
> Chuck
>
>
> --
> The Moon is Waxing Gibbous (90% of Full)
> So visit me at http://www.valstar.net/~hallenbeck
>
>
> _______________________________________________
> Speakup mailing list
> Speakup@braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
>

Re: suggestion for disastery recovery (was Re: more on hardware clock)

[Jude DaShiell]

make dep blew up when I tried building a kernel.

Re: more on hardware clock

[Charles Hallenbeck]

Hi Jude,
That is most interesting. It certainly caused a problem for me
here on this 8.1 distro, and until I got rid of it I was obliged
to shutdown without unmounting my partitions first due to the
hwclock program hanging.

Chuck

On Sun, 20 Oct 2002, Jude DaShiell wrote:

> Hi Chuck, yes I've shut down successfully several times.  Interesting
> since I'm using slackware 8.0.0 not 8.1.  It may be a fix to 8.0.0 that
> made slackware 8.1 actually broke libsafe.



-- 
The Moon is Waxing Gibbous (99% of Full)
So visit me at http://www.valstar.net/~hallenbeck

Re: more on hardware clock

[Toby Fisher]

On Sun, 20 Oct 2002, Charles Hallenbeck wrote:

> Hi Jude,
> That is most interesting. It certainly caused a problem for me
> here on this 8.1 distro, and until I got rid of it I was obliged
> to shutdown without unmounting my partitions first due to the
> hwclock program hanging.

You could have got around this by compiling y our kernel with sysrq
support, (see /path/to/linux-sources/Documentation/sysrq.txt).  Nice
stuff, if you know nobody else will be accessing your consol, as you don't
even need to be logged in to use it.

HTH

-- 
Toby Fisher	Email: toby@gw0ucu.plus.com
Tel.: +44(0)1480 417272	Mobile: +44(0)7974 363239
ICQ: #61744808
   Please avoid sending me Word or PowerPoint attachments.
   See http://www.fsf.org/philosophy/no-word-attachments.html