Linux-Development-Sys Digest #940, Volume #7 Thu, 8 Jun 00 11:13:17 EDT
Contents:
Re: RedHat 6.2 Autofs Broken (Bob Tennent)
what is "gtcd"? ("Brian")
Re: exit()/pthreads bug on Linux PPC (Peter Burka)
Re: exit()/pthreads bug on Linux PPC (Peter Burka)
Re: finding filename from file descriptor (Mario Klebsch)
Re: finding filename from file descriptor (Josef Moellers)
Re: [Q] Hook system call (Chetan Ahuja)
Re: what is "gtcd"? (Robert Hampf)
syslogd no longer work , why ? ("�Ѽ���")
Journaling File System (Herve Gautier)
Re: Journaling File System (Anes Lihovac)
Re: Journaling File System ("Amit Ghosh")
Re: Where to get Java Compiler (Timothy Murphy)
Re: finding filename from file descriptor (John Reiser)
Re: Journaling File System (Aki M Laukkanen)
getting kernel panic messages ? (Tom)
Re: Sound Card (Toby Haynes)
sprintf in kernel space (Daniel Krippner)
Programe Interface For TFTP (Colin Ford)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Bob Tennent)
Crossposted-To: comp.os.linux.portable
Subject: Re: RedHat 6.2 Autofs Broken
Date: 8 Jun 2000 01:50:17 GMT
Reply-To: [EMAIL PROTECTED]
On Wed, 07 Jun 2000 17:49:11 -0700, David Highley wrote:
>> >The auto.master file has this entry:
>> >/home auto.home
>> Try
>>
>> /home /etc/auto.home
>>
>Not if it is being defined by NIS.
The man page isn't great, but I would guess you need to use
/home yp:auto.home
Bob T.
------------------------------
From: "Brian" <[EMAIL PROTECTED]>
Subject: what is "gtcd"?
Date: 8 Jun 2000 02:33:58 GMT
Hi,
Sometimes my linux (Red Hat 6.1) will pop up a warning message like this:
Warning:
gtcd--sm-client-id 117f0000001000096024323800000005750019
No response to the save Yourself command.
The program may be slow, stopped or broken.
You may wait for it to respond or remove.
I'm trying to do some SCSI driver programming and I suspect this message has
something to do with my SCSI CDROM, because it comes up when I attach the
CDROM to the SCSI card. Could anyone tell me more about "gtcd"? Thanks in
advance.
Brian
------------------------------
From: Peter Burka <[EMAIL PROTECTED]>
Subject: Re: exit()/pthreads bug on Linux PPC
Date: Thu, 08 Jun 2000 03:43:19 GMT
Kaz Kylheku wrote:
>
> On Thu, 08 Jun 2000 00:18:06 GMT, Peter Burka <[EMAIL PROTECTED]> wrote:
> >I'm porting a program to Linux PPC (Yellow Dog 2.1) and
> >I've encountered a strange bug regarding exit(). The
> >same program works fine under Linux 386 and Linux ARM,
> >and on other PPC platforms (Neutrino and AIX).
> >
> >The program does the following:
> >
> >- create several pthreads
> >- one of the pthreads blocks on a mutex
> >- one of the pthreads waits on a condvar
> >- the main thread calls exit()
>
> Sounds like this program is small enough to include in a posting,
> hint hint.
You'd think so, wouldn't you? :)
I tried building a simple program which demonstrates the same problem,
but was unable to do so. I've simplified what the program is doing
significantly in my description -- it's actually several hundred
thousand lines of code (most of which aren't being run in this
particular test case).
The problem is probably timing dependent, since adding printfs seems
to change the frequency of hangs.
The test case doesn't always hang, but it usually does. It also runs
just fine on the other 2500+ test cases I have. But in one particular
test case it hangs. There's nothing really special about the test
which hangs it. It's a very mundane test which doesn't do anything
unusual. And this test case runs fine on about 20 other platforms,
including Linux on other chips, and other (non-Linux) OSes on PPC.
>
> >I put a printf right before the call to exit(), and it
> >is being reached. However the program doesn't always
> >exit. It will usually (about 90% of the time) hang at
> >this point.
> >
> >Running it under gdb shows that the main thread has
> >terminated, but that the two blocked threads are still
> >hanging around:
>
> This should not happen; the main thread ought to block until the other two
> threads terminate (that is LinuxThreads behavior not required by POSIX).
>
> When your main thread calls exit(), the exit() function will eventually call a
> registered handler within LinuxThreads. This handler spits a special process
> exit request into the thread manager's pipe and waits for a reply. The manager
> kills all the other threads and then unblocks the requesting thread which
> then goes on to complete the process termination. (Disclaimer: I haven't
> looked at that code in a few weeks, but looks like that's about to change.)
>
> So it is very odd that the exiting thread should disappear, but the other
> threads are still running. It could be a kernel problem.
That was my guess. Once I determined that the program wasn't
deadlocked, and that exit was being reached I was pretty sure it
was a kernel problem.
It just occurred to me, though, that I do install an atexit hook.
I'll try taking that out and see what happens.
>
> >The system is a Mac G3 with a Yellow Dog 2.1 distro; glibc 2.1.3 (??
> >I thought the latest version was 2.1.2). I also tried this on a
>
> No, 2.1.3 came out in March or so. It has many LinuxThreads
> related fixes.
>
> >slightly older RedHat(?) distro using glibc 2.1.1 and encountered
> >exactly the same behaviour.
>
> Hmmm. That could be relevant information, good that you tried.
>
> As far as glibc-2.1.3 goes, we know that this bug happens on one PPC
> installation of a particular distro and that you can't repro it on other PPC
> installations (also Yellow Dog 2.1?). So nothing can be ruled out, including a
> corrupt installation or installation media.
Nope. I CAN reproduce it. I've tried two different G3s with two
different distros and it happens with about the same consistency
on both boxes.
>
> Nevertheless, you might want to submit a report with glibcbug.
>
How does one go about doing that?
thx/peter
------------------------------
From: Peter Burka <[EMAIL PROTECTED]>
Subject: Re: exit()/pthreads bug on Linux PPC
Date: Thu, 08 Jun 2000 03:53:53 GMT
Kaz Kylheku wrote:
>
> On Thu, 08 Jun 2000 00:18:06 GMT, Peter Burka <[EMAIL PROTECTED]> wrote:
> >I'm porting a program to Linux PPC (Yellow Dog 2.1) and
> >I've encountered a strange bug regarding exit(). The
> >same program works fine under Linux 386 and Linux ARM,
> >and on other PPC platforms (Neutrino and AIX).
> >
> >The program does the following:
> >
> >- create several pthreads
> >- one of the pthreads blocks on a mutex
> >- one of the pthreads waits on a condvar
> >- the main thread calls exit()
>
> Sounds like this program is small enough to include in a posting,
> hint hint.
You'd think so, wouldn't you? :)
I tried building a simple program which demonstrates the same problem,
but was unable to do so. I've simplified what the program is doing
significantly in my description -- it's actually several hundred
thousand lines of code (most of which aren't being run in this
particular test case).
The problem is probably timing dependent, since adding printfs seems
to change the frequency of hangs.
The test case doesn't always hang, but it usually does. It also runs
just fine on the other 2500+ test cases I have. But in one particular
test case it hangs. There's nothing really special about the test
which hangs it. It's a very mundane test which doesn't do anything
unusual. And this test case runs fine on about 20 other platforms,
including Linux on other chips, and other (non-Linux) OSes on PPC.
>
> >I put a printf right before the call to exit(), and it
> >is being reached. However the program doesn't always
> >exit. It will usually (about 90% of the time) hang at
> >this point.
> >
> >Running it under gdb shows that the main thread has
> >terminated, but that the two blocked threads are still
> >hanging around:
>
> This should not happen; the main thread ought to block until the other two
> threads terminate (that is LinuxThreads behavior not required by POSIX).
>
> When your main thread calls exit(), the exit() function will eventually call a
> registered handler within LinuxThreads. This handler spits a special process
> exit request into the thread manager's pipe and waits for a reply. The manager
> kills all the other threads and then unblocks the requesting thread which
> then goes on to complete the process termination. (Disclaimer: I haven't
> looked at that code in a few weeks, but looks like that's about to change.)
>
> So it is very odd that the exiting thread should disappear, but the other
> threads are still running. It could be a kernel problem.
That was my guess. Once I determined that the program wasn't
deadlocked, and that exit was being reached I was pretty sure it
was a kernel problem.
It just occurred to me, though, that I do install an atexit hook.
I'll try taking that out and see what happens.
>
> >The system is a Mac G3 with a Yellow Dog 2.1 distro; glibc 2.1.3 (??
> >I thought the latest version was 2.1.2). I also tried this on a
>
> No, 2.1.3 came out in March or so. It has many LinuxThreads
> related fixes.
>
> >slightly older RedHat(?) distro using glibc 2.1.1 and encountered
> >exactly the same behaviour.
>
> Hmmm. That could be relevant information, good that you tried.
>
> As far as glibc-2.1.3 goes, we know that this bug happens on one PPC
> installation of a particular distro and that you can't repro it on other PPC
> installations (also Yellow Dog 2.1?). So nothing can be ruled out, including a
> corrupt installation or installation media.
Nope. I CAN reproduce it. I've tried two different G3s with two
different distros and it happens with about the same consistency
on both boxes.
>
> Nevertheless, you might want to submit a report with glibcbug.
>
How does one go about doing that?
thx/peter
------------------------------
From: [EMAIL PROTECTED] (Mario Klebsch)
Subject: Re: finding filename from file descriptor
Date: Thu, 8 Jun 2000 00:59:24 +0200
[EMAIL PROTECTED] (Alexander Viro) writes:
>Let one of them create an AF_UNIX socket and bind it to some name.
>Another should connect() to that name and then they can pass opened
>files to each other - sendmsg() with SCM_RIGHTS cookie will do nicely.
Sending rights over a connected UNIX domain socket? Connected sockets
are STREAM sockets. This did not work on Linux, the last time I tried
it. When I used UNIX domain DGRAM socket instead, it worked.
73, Mario
--
Mario Klebsch [EMAIL PROTECTED]
------------------------------
From: Josef Moellers <[EMAIL PROTECTED]>
Subject: Re: finding filename from file descriptor
Date: Thu, 08 Jun 2000 08:19:06 +0200
Mario Klebsch wrote:
> =
> [EMAIL PROTECTED] (Alexander Viro) writes:
> =
> >Let one of them create an AF_UNIX socket and bind it to some name.
> >Another should connect() to that name and then they can pass opened
> >files to each other - sendmsg() with SCM_RIGHTS cookie will do nicely.=
> =
> Sending rights over a connected UNIX domain socket? Connected sockets
> are STREAM sockets. This did not work on Linux, the last time I tried
> it. When I used UNIX domain DGRAM socket instead, it worked.
Why so overly complicated?
Just open /proc/<oldprocess>/fd/<filedescriptor>!
-- =
Josef M=F6llers
Fujitsu Siemens Computers
SHV Server DS 1
------------------------------
From: Chetan Ahuja <[EMAIL PROTECTED]>
Subject: Re: [Q] Hook system call
Date: 8 Jun 2000 07:37:47 GMT
You don't have to modify the kernel for that. Just use the already
built in ptrace system call. In fact there's a userland program
available that does almost what you want without you having to do
ANY programming. Do a "man strace" on your system.
usenet.seri.re.kr <[EMAIL PROTECTED]> spoke thusly:
> I want to modify kernel source. Is it possible?
> I wanna trace system call.
> Whenever system call is invoked, I want to know who invoke system call, what
> system call is invoked, when system call is invoked and etc................
> ...I'll add new system call.
> After my new system call invoked,
> Whenener system cal l invoked, I want to get a file(something other) that
> contain invoked system call's information.
> Is that possible.
> If you know solution, please tell me.
> thanks for reply.
> Jongyun, Jeong........from Korea.
------------------------------
From: [EMAIL PROTECTED] (Robert Hampf)
Subject: Re: what is "gtcd"?
Date: Thu, 8 Jun 2000 11:25:24 +0300
Brian <[EMAIL PROTECTED]> h�lt �essu fram:
:
: I'm trying to do some SCSI driver programming and I suspect this message has
: something to do with my SCSI CDROM, because it comes up when I attach the
: CDROM to the SCSI card. Could anyone tell me more about "gtcd"? Thanks in
: advance.
gtcd is the Gnome CD player. In some configurations in RH it starts
automatically if you insert an audio CD into your CD-ROM.
rh
------------------------------
From: "�Ѽ���" <[EMAIL PROTECTED]>
Subject: syslogd no longer work , why ?
Date: Thu, 8 Jun 2000 18:26:46 +0900
hi!
the log file in my linux system are no longer recorded but it rotated. i
don't know why
/var/log/secure, /var/log/maillog .. these logfile can't be writed.
"df" shows there enough space for /dev/log , and syslogd still running
is there anybody who know this ?
thanks
------------------------------
From: Herve Gautier <[EMAIL PROTECTED]>
Subject: Journaling File System
Date: Thu, 08 Jun 2000 11:22:39 +0200
Reply-To: [EMAIL PROTECTED]
Hi there !
Anybody have already install a JFS ? (ReiserFS, XFS, ...)
Can you give some instructions and advices ?
--
GAUTIER Herve (R.V)
professional email : [EMAIL PROTECTED]
------------------------------
From: Anes Lihovac <[EMAIL PROTECTED]>
Subject: Re: Journaling File System
Date: Thu, 08 Jun 2000 14:47:56 +0200
Yes I have ReiserFS installed. It is very sexy. After a crash the
system is up in a second !
The SuSE Linux 6.4 supports it by install, Mandrake 7.1 also (I think)!
XFS would be nice, but they havn't get released it yet I think (the
stable/final
Version) !
JFS -- ? I don't know.
The strange thing I discovered that eg:
This was when I worked with the Nvidia GLX Beta drivers for my video
card (they used to be very
crappy )
When I extract a source package, configure and compiled it,than started
it and crashes, the whole Directory
wasn't there any more after the reboot!
The Rest is fine. I didn't compared the performance speed between Ext2
and ReiserFS.
Greetings
------------------------------
From: "Amit Ghosh" <[EMAIL PROTECTED]>
Subject: Re: Journaling File System
Date: Thu, 08 Jun 2000 14:58:43 +0100
In article <[EMAIL PROTECTED]>, Herve Gautier
<[EMAIL PROTECTED]> wrote:
>
> Hi there !
>
> Anybody have already install a JFS ? (ReiserFS, XFS, ...) Can you give
> some instructions and advices ?
I installed ReiserFS (AFAIK the others aren't stable yet, so *don't*
use them on a production machine). Did it once with an existing ext2
installation with some problems due to my own imperfection (why should
one read documentation? Well, now I know! :) So be sure to read the
manuals twice or you will have problems with booting). Currently I'm
working on SuSE 6.4 that comes preconfigured. While installing I only
had to choose ReiserFS over ext2. That's it! BTW I'm *very* happy with
ReiserFS.
Where are your problems?
--
Amit Ghosh
mailto:[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Timothy Murphy)
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.misc
Subject: Re: Where to get Java Compiler
Date: 8 Jun 2000 14:30:29 +0100
Bob Koss <[EMAIL PROTECTED]> writes:
>Does anyone have VisualAge Java 3.0 for Linux ?? I've managed to
>download the winblows version (and it's impressive!), but I can't for
>the life of me navigate ibm's website to get the Linux version.
I think if you start at http://www.software.ibm.com/ and follow the links
it is fairly straightforward.
Nb it is very large (37MB).
--
Timothy Murphy
e-mail: [EMAIL PROTECTED]
tel: 086-233 6090
s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland
------------------------------
From: John Reiser <[EMAIL PROTECTED]>
Subject: Re: finding filename from file descriptor
Date: Thu, 08 Jun 2000 07:20:19 -0700
Alexander Viro wrote:
>
> In article <[EMAIL PROTECTED]>,
> John Reiser <[EMAIL PROTECTED]> wrote:
> >When the /proc filesystem is installed, then each open file descriptor
> >has a symbolic link named /proc/self/fd/j where 'j' is the decimal
> >representation of the file descriptor number; and this is also accessible
> >as /proc/pid/fd/j where 'pid' is the decimal representation of the
> >processs ID number. In some common cases, this knowledge can be exploited
> >to provide what many people want, but probably not with the speed that
> >some might expect. See the utility 'lsof' ("ls open files").
>
> That still does not address the issue with unlinked files. Try and you will
> see. Open a file, unlink it and look into proc/<pid>/fd. Think a bit and you'll
> see why it is so.
>
> Again, "name of the opened file" is a nonsense on UNIX. Why do you need it,
> BTW?
You need a lesson in quantifiers. Read what I posted. Note the use of
"some" and "many". I did not promise anything about unlinked files.
The knowledge of one or more pathnames associated with an open file
descriptor can be useful for documentation, audit or debugging.
"/tmp is accumulating zero-length files. Is some application crashing
before unlink?" "Here's a 100MB file that is growing by 100 bytes per
second. Which process is responsible?" "The detailed analysis for
August is producing bad data. Which of the twelve processes that I
started (one for each month of last year) should I attach with gdb?"
Exclude /proc for the moment. The mapping from the set of open inodes
to the set of pathnames in the filesystem is a [time-varying] partially-
defined relation. For each open inode, there may be zero, one, or more
than one pathnames that resolve to it. In practice, many of the common
cases that developers and/or users care about, such as redirection of
stdin, stdout, or stderr to a regular file (not a pipe) have exactly
one associated pathname. Most shells hide the name, but knowing or
discovering it can provide peace of mind, and remove some of the mystery
surrounding "What is my machine doing?"
--
John Reiser, [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Aki M Laukkanen)
Subject: Re: Journaling File System
Date: 8 Jun 2000 14:16:05 GMT
In article <[EMAIL PROTECTED]>, Anes Lihovac wrote:
>When I extract a source package, configure and compiled it,than started
>it and crashes, the whole Directory
>wasn't there any more after the reboot!
Reiserfs only does meta-data journaling. Which program crashes the
whole system?
--
D.
------------------------------
From: Tom <[EMAIL PROTECTED]>
Subject: getting kernel panic messages ?
Date: Thu, 08 Jun 2000 16:31:38 +0200
Hi,
I=B4m writing a network device driver (the driver works well under 2.0.x)=
for 2.2.x using the AMCC S5933 chip.
When the drivers open function get called (by ifconfig) the machine dies
(It has something to do with the PCI_READ_COMPLETE enabling). =
There is a lot of information on the debug console, but I can=B4t read
them all (it scrolls to fast). =
Any idea how I can read the debug output on a second machine using a
serial cable ?
=
any help will be appreciated,
Tom
------------------------------
From: Toby Haynes <[EMAIL PROTECTED]>
Subject: Re: Sound Card
Date: 08 Jun 2000 10:26:46 -0400
!! "Bo" == Bo <- Sun <[EMAIL PROTECTED]>> writes:
Bo> hi: I am using Turbo linux on my machine, and I now want to play
Bo> .mp3.
Bo> I first use xmms, but get the error messages that the sound card
Bo> driver is not compiled into the kernel. and some other kinds of
Bo> error messages. I installed linux from CD-ROM, so I don't know
Bo> how to config sound card.
Some information on what sound card you have is necessary before
you'll get any decent help with this! There are lots of cards, lots of
drivers and lots of potential problems ...
Cheers,
Toby Haynes
--
Toby Haynes
The views and opinions expressed in this message are my own, and do
not necessarily reflect those of IBM Canada.
------------------------------
Date: Thu, 08 Jun 2000 16:50:15 +0200
From: Daniel Krippner <[EMAIL PROTECTED]>
Subject: sprintf in kernel space
Hi everyone
this is a quickie: I need a sprintf in kernel space - what do I do since
I cannot #include <stdio.h> ?
thanks
Daniel
--
[EMAIL PROTECTED]
Student of Communication Engineering/Computer Networking
Furtwangen University of Applied Sciences
------------------------------
From: Colin Ford <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Programe Interface For TFTP
Date: Thu, 08 Jun 2000 16:02:11 +0100
Hi All,
Any one know how I can programmatically use a linux
tftp client to go get some data? Know any good API's?
Cheers,
Col.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.development.system) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Development-System Digest
******************************
