neral
direction of the specifics of the problem (if they exist, if not, I may
try to create a simpler test case then java)?
I tried a few searches, but nothing matching what I remembered came up.
--
David E. Cross
___
freebsd-hackers@freebsd.org mailing
I am looking at a project that will require a user based process to interact
with the system as if it were a filesystem. The traditional way I have seen
this done is as the system NFS mounting itself (ala AMD). I would really like
a more clean approach to this. What I am interested in is a 'Use
> :
> :Look into the portal filesystem. This is what you want :)
> :
> : Brian Fundakowski Feldman _ __ ___ ___ ___ ___
> : gr...@freebsd.org _ __ ___ | _ ) __| \
>
> Actually, it isn't quite. All the portal filesystem will allow you
> to do is pass back
I thought now would be a good time to chime in on some of my wild schemes...
The reason I am interested in 'userfs' is to enable me to write a version
of 'nsd'. Those of you familiar with Irix will recognize it. For others,
what it does is to present the name-space on a machine as filespace.
Th
> Mike Smith wrote:
> > On Mon, Jul 19, 1999 at 06:13:51PM +0200, Dag-Erling Smorgrav wrote:
> > > Oscar Bonilla writes:
> > > > the idea is to have an entry in the /etc/passwd enabling LDAP lookups.
> > > > the Entry would be of the form
> > > >
> > > > ldap:*:389:389:o=My Organization, c=BR:uid
> > Lovely. Sounds like a much better way to do the Solaris/Linux (and
> > NetBSD?) /etc/nsswitch.conf stuff. On Solaris at least, this is
> > implemented using masses of weird shared objects...
>
>The plan for NetBSD is that things will also be handled with dynamic
>modules, but those dynamic mo
> Couldn't we do this with /etc/auth.conf? What's the real purpose of this
> file? From the man page: "auth.conf contains various attributes important to
> the authentication code, most notably kerberos(5) for the time being."
> Isn't this what PAM is about? authentication? or does auth.conf cover
Yes, I am still working on it, don't despair ;)
This is the case of project creep... I am now working on the 'isw*()'
functions, and I have a couple of questions regarding locale support in
FreeBSD. Namely, how the heck do I get access to the database? I see
that the LC_* databases have all the
We had a similiar problem here. We had meant to submit-pr it but forgot.
In our case it was because inetd had only the amanda line in it (inetd was
not responsible for any other services. Our guess was that it is an off by
one error in inted somewhere, but we never traced it down further. Our
w
I have a program (part of CDE)... we will call it 'foo',
"foo" has library dependancies: libtt.so, libX11.so, libXt.so, libXext.so, and
libwcs.so(this last one is mine).
libtt.so depends on iswalpha() and iswspace() (which are defined in libwcs.so)
If I link with all of those I get an error th
Nope, that is all that we had time to track down. We were fighting NFS panics
arround the same time, stuff got lost in the shuffle :)
--
David Cross | email: cro...@cs.rpi.edu
Systems Administrator/Research Programmer | Web: http://www.cs.rpi.edu/~crossd
Rensselaer
> > I have a program (part of CDE)... we will call it 'foo',
> >
> > "foo" has library dependancies: libtt.so, libX11.so, libXt.so, libXext.so,
> > and
> > libwcs.so(this last one is mine).
> >
> > libtt.so depends on iswalpha() and iswspace() (which are defined in
> > libwcs.so)
> >
> > If I
I updated a system to -CURRENT last night and got a panic with alot of
messages about UDMA failing (I don't have the exact messages, I can get
them if need be). I backed down the wdc0/wdc1 controller flags from
0xa0ffa0ff to 0x0 and everything is happy. I figured its -CURRENT, and that
is to be e
> I was in the UDMA code yesterday
> (but mostly in the CYRIX code... (changes elsewhere should have been
> mostly cosmetic).
>
>
> can you get the exact error message?
>
> julian
I got it..., I happened to be working on something else at the time
and I let it sit unattended for awhile.. it
Since I am planning on writing userfs in order to impliment 'nsd' (and
some other ideas I have hatching too :). I need to know how filesystem
accesses work. Can they be queued up, and responded to out of order?
For example... I have a request come in (via the filesystem), that request
is going t
I have 2 NFS servers. One is primarily read-only, the other read-write, they
service the same clients (the read-only services more). They are (were) of
the same build. I have a problem on the read/write server where it chews
through mbuf clusters (it goes through about 3k in a day). Especially
Well, I just -STABLED the server to see if it fixed it, but I was certainly
running out. the server had only 3000-ish mbuf chains, and it would go through
them all in a day.
--
David Cross | email: cro...@cs.rpi.edu
Systems Administrator/Research Programmer | Web: h
Ok, here are some real stats
"w" is the read-only machine, it services everything that "s" (the
read-write machine) does... in fact it services more.
*w crossd $ strings -a /kernel | grep \^___maxusers
___maxusers 96
*w crossd $ uname -a
FreeBSD w.cs.rpi.edu 3.2-STABLE FreeBSD 3.2-STABLE
Well, backing out now is not really an option... But given my past history
with NFS, and knowledge of this site I think I have a fair idea where the
leak is... I think it is in the nfsv3 "commit" handler.
Why do I think this? Simple, this problem started when a user started running
a large jo
Well, it doesn't appear to be commit() :(.
Any-who, is there a way I can get a look at the raw mbuf/mbuf-clusters?
I have a feeling that seeing the data in them would speak volumes of
information. Preferably a way to see them without DDB/panic would be ideal.
--
David Cross
I found it... our favorite function... nfsrv_create()!!! :)
The problem was/is a create of an already existing file (with O_EXCL|O_CREATE,
I would bet, but I don't have anyway to tell) returns *nothing* to the sender.
The last time I had this problem it was because nfsrv_create() was not clearing
PS: I was down to only 3k mbuf-clusters free on the server, so I 'rm'-ed
the troublesome file and the create went through and no more mbuf-leaking.
On the downside, I cannot reproduce this problem any longer with any
reliability.
--
David Cross | email: cro...@cs.rpi
> Hmmm. Interesting. An EEXIST error occuring at that point for an
> NFSV3 mount will execute the correct nfsm_reply(), but since it is
> NFSV3 the nfsm_reply() macro will not jump to a return(0) ... when
> it finishes constructing the reply it falls through instead.
>
> In
I am wading through the portalfs and nullfs source, but I am desperately
lost. I would love to be able to find out who would be willing to help out
with questions. I feel I would be spamming far too many people by just sending
to -hackers. Some of the topics I am curious about are general fs-sty
Here is a pro vote for enabling BPF in GENERIC:
It will let us use a dhcp client in the install programs, this is of tremendous
use to many people as DHCP starts to become much more popular. I cannot
net install a machine at home since that is on a DHCP cable modem service.
Also, if root is comp
A friend writing some portable network tunneling software ran into an
interesting thing... when you specify "IP_HDRINCL" with SOCK_RAW, and
IPPROTO_RAW you need to construct the outgoing packet in host byte order.
This seems wonderfully inconsistent with all of the other socket based
networking
I am attempting to get FreeBSD 3.2 and/or 4.0 to go on a TP 360c. The
problem I am having is that the keyboard works all the way up to sysinstall.
I can use the keyboard in the visual kernel config/etc. I searched and found
under 2.2 they suggested setting flags 0x10 on syscons. 0x10 isn't docu
> I am attempting to get FreeBSD 3.2 and/or 4.0 to go on a TP 360c. The
> problem I am having is that the keyboard works all the way up to sysinstall.
> I can use the keyboard in the visual kernel config/etc. I searched and found
> under 2.2 they suggested setting flags 0x10 on syscons. 0x10 is
> You are quite right that the code in question was just moved from sc
> to atkbd and there is essentially no difference between the two
> versions.
>
> This is the first time that I hear the flag 0x10 for sc works in 2.X,
> but the flag 0x4 for atkbd does not in 3.1 or later :-( I think
> I hear
I offered (to Theo T'So) before our (Computer Science Department at RPI)
resources to setup a RO CVS repo for Kerberos V. He accepted out offer
but things stagnated after that on setting up the details. My fault mostly
for not taking the tourch that has been passed. I am [now] offering
again, an
I am terribly sorry. I had 2 messages about kerboers5 come in at the same
time (one from -hackers, one from mit), I replied to to wrong one.
--
David Cross | email: cro...@cs.rpi.edu
Systems Administrator/Research Programmer | Web: http://www.cs.rpi.edu/~crossd
Ren
I have been attempting to track down why cdrom boots will not work with
/boot/loader, but do just fine with the boot-block. I have come to the
following wild speculation, and stab in the dark. /boot/loader uses some
int13 stuff, which I found while reading in the boot0inst man page may cause
tro
I am trying to write a very kludgey/monolithic driver for a CardBus ethernet
adapter. I have run into a bit of a stumbling block on some issues. One such
issue is the attach (I need to map some registers of the adapter into memory
space so I can read/write values.). Anyway if someone could expla
I have been writing a nasty kludge to treat a CardBus bridge as a standard
PCI bridge (with static config) . I have
it to the point where I can (after the system is booted) 'pciconf -r
pci5:0:0 0' and get scan information (neat, huh :). Welll, I thought it would
then just be a simple matter of '
I am modifying the tulip device driver to support this xircom card. I have it
almost entirely working, *except* that it goes into infinite re-neogitiate
loops. The card probes correctly at bootup, but any attempt to change
information via ifconfig ("ifconfig de0 inet ..." and "ifconfig de0 up",
Well, it has been a long time since I have needed to write an email with that
tagline. Our primary NFS server had been up for almost 2 months with no
panics. We did need to reboot it for a network change, but it was up for 28
days at that point. Anyway here are the details:
dev = 0x20014, b
Our ftp server crashed early this morning with what appears to be a softupdates
error:
> Sep 13 09:56:19 stumble /kernel: pid 41477 (perl), uid 0 on
> /exports/share3/ftp/.2: file system full
>
> panic: softdep_write_inodeblock: indirect pointer #0 mismatch 0 != 15597568
> syncing disks... panic
We have a very hetergenous environment here (even among the FreeBSD boxes).
Each PC tends to be just a little bit different. This expecially causes
problems since we wish to have XDM on each machine on boot and have X
on a NFS partition. TO alleviate this we invented a simple Perl script
to repla
> Softupdates has known bugs relating to filesystem full conditions which
> I believe Kirk is working on. There isn't much you can do until then
> other then either disable softupdates or work to avoid the disk-full
> condition. The panic does not occur very frequently so working
> Umm, you can edit /usr/X11R6/lib/X11/xdm/Xservers to configure xdm to
> run say /usr/config/X (which would be stored on the local machiens hard
> drive) instead of /usr/X11R6/bin/X. This is a much simpler solution.
> :) (Just symlink /usr/config/X to /usr/X11R6/bin/XF86_Whatever.)
Simpler? It
> It's been noted on several occasions that with large (> 256MB) of RAM, one
> has to be "careful" with the configuration (NMBCLUSTERS, MAXUSERS) to
> prevent the box from falling over every few days due to kvm problems.
>
> Can somebody be more specific? I'm just about to order a really, really
> > > has to be "careful" with the configuration (NMBCLUSTERS, MAXUSERS) to
> > > prevent the box from falling over every few days due to kvm problems.
> > >
> > > Can somebody be more specific? I'm just about to order a really, really
> > > expensive machine and I want to be sure I can get it to
> > > > It seems there's a need, and the possibility. Would somebody like to
> > > > suggest a syntax?
> > >
> > > The precedent would be the socket ioctls SIOCGIFHWADDR and
> > > SIOCSIFHWADDR. The Linux emulator suppors the get-only version
> > > already.
> >It's already been mentioned tha
I dug through the archives and found peopel with similiar problems to
what I am experiencing, but I didn't find any answers that have worked
for me. Here are the problem I am having:
1: The built-in SCSI ROM is v2.01, there was mention of BIOS 1008 including
2.11. I applied the 1008 flash a
First, I would like to take this opportunity the thank Matt Dillon for
his excellent work with NFS/TCP. Wow, way to go :)
Now on to the real problem :)
One of our users way able to reliably crash an NFS server 3 times today.
I have since copied his program and have reliably crashed a seperate an
> > One of our users way able to reliably crash an NFS server 3 times today.
> > I have since copied his program and have reliably crashed a seperate and
> > unloaded machine with the exact same panic, "lockmgr: locking against
> > myself". I check the recent DG patches that went in after -RELEASE
> > gdb -k kernel.1 vmcore.1
> > GNU gdb 4.18
> > Copyright 1998 Free Software Foundation, Inc.
> > GDB is free software, covered by the GNU General Public License, and you are
> > welcome to change it and/or distribute copies of it under certain
> > conditions.
> > Type "show copying" to see the
> Another possibly re: debugging. If you compile up a kernel with
> options DDB and options BREAK_TO_DEBUGGER, the kernel will break into
> DDB when the panic occurs. You can then issue a 'trace' command to get
> a backtrace.
>
> This may be good enough to determine what the
I have done simple drivers before. I would be interested in working with you
on this (it would benefit me as well). If you couild provide a web site with
more information that would help too.
--
David Cross | email: cro...@cs.rpi.edu
Systems Administrator/Research
> Using my FreeBSD CD-ROMs, I've been able to go into fixit mode and mount
> the root filesystem of the drive, but I'm not sure where to go from there.
> How can I figure out what my old disklabel was? Is there some way I can
> search the raw disk for the locations of the file systems?
>
> Any hel
I have just installed 3.2-19990521-STABLE from releng3.freebsd.org. I
compiled a debugging kernel, forced a panic with DDB, and created a crashdump,
on bootup it saved it to /var/crash/[kernel|vmcore].0. I then ran a
'gdb -k kernel.0 vmcore.0' and received the happy fun 'gd_curpcb' symbol
not fou
Is there anything I can do on my end (we have a complete CVS repository,
synced every 4 hours. I would like to use the *exact* same gdb that we
compiled the world from if possible, I am affraid of using a more recent
gdb will result in not being able to read the core (I am unable to use
a 3.1-STAB
Here it is:
(I am not sure that the source didn't get updated since the kernel was
compiled, so the line numbers may be meaningless. The second backtrace is
from and earlier kernel and has 0 line number information in it.
IdlePTD 2985984
initial pcb at 264eac
panicstr: lockmgr: locking against my
I am trying to trace down the cause of the recursive lock and I stumbled upon
this:
(kgdb) bt
#0 boot (howto=256) at ../../kern/kern_shutdown.c:285
#1 0xc014b3f4 in at_shutdown (
function=0xc0234aca
<__set_sysuninit_set_sym_M_KTRACE_uninit_sys_uninit+154>, arg=0x10002,
queue=-951064448) at
> I don't think that this dump is useful for debugging this problem. Perhaps,
> if
> you compile the kernel with DEBUG_LOCKS, you will get more useful info.
>
> Dima
I checked through the source for DEBUG_LOCKS, it doesn't appear to do anything
other than to printout information information
Well, I have a debuging kernel compiled with DEBUG_LOCKS turned on, below is
the result. This was cvsup-ed and built about May 27, 14:00 EDT.
gdb -k /usr/src/sys/compile/STAGGER/kernel.debug vmcore.10 (yes, that's right,
#10 in a little over one week)
GNU gdb 4.18
Copyright 1998 Free Software Fo
Here is a backtrace from our latest. Does anyone have any ideas to try. Is
there any way to loock at who has the other lock on the file? (Yes, I know
it is the kernel who has it, but it is requested on behalf of a NFSd, no?)
(kgdb) bt
#0 boot (howto=256) at ../../kern/kern_shutdown.c:285
#1 0
Our home directory NFS server went down again today, "same bat-panic".
This time it went down on ".Maillock" (usually it goes down on a netscape
cache file or .Xauthorit-c. Piecing some more together I modified my old
"crash_patoot.c" file (which didn't cause any problems), to the new and
improve
Just need to make a small amendment to the instructions for the previous
prgram. I am unable to cause a panic by running it on just one machine, I
need to have it run on 2 machines accessing the same file.
client1% ./crash_patoot cp1
client2% ./crash_patoot cp1
--
David Cross
I had the hunch that the problem I am dealing with related to the unlink
portion of NFS... So I have simplified the code down to this tiny snipet which
will reliably crash the system (I left it running by accident and it brought
my test machine down 3 times before I remembered to kill it :). This
Yes, I am *very* certain that it is the server that is crashing. The
test server in question has no user login-able accounts at all, and is only
running nfsd/mountd/portmap.
--
David Cross | email: cro...@cs.rpi.edu
Systems Administrator/Research Programmer | Web: h
> > fd=open(argv[1], O_CREAT, 600);
>
> Since this opens the file so that it cannot be written to, not
> to mention the really weird mode it will get if it's created by
> that open(), the rest of the thing doesn't deserve to work.
>
> Generally speaking, it's a good idea to make s
We have similiar restraints for a certain number of our machines, we have
solved this problem by using FreeBSD's built in firewall
(just add 'options IPFIREWALL' to your kernel config script). Here is
a *very* simple firewall config to do some such restrictions):
You may note that there are mutlip
Just doing some performance testing and I noticed something rather
disturbing
Here is the test program:
int main (void)
{
int count=0;
for(count=0;count <1000;++count)
getppid();
return 0;
}
The time on linux for this program is ~5 seconds (linux "time" re
Oops, here is some additional information from my system:
bash-2.02$ cat sc.c
int main (void)
{
int count=0;
for(count=0;count <1000;++count)
getppid();
return 0;
}
bash-2.02$ cc -o sc sc.c
bash-2.02$ uptime
11:19AM up 3 hrs, 4 users, load averages: 0.01, 0.01
Yes, it has happened again (twice in fact).
I am desperately trying to find the source of this. So far my conclusions
have led me to a race in unlink and NFS somewhere (still have no clue where).
And it is only from Sun clients to date. Also, this started happening in
ernest arround when we put
Yes, I have determined (just today) that the PANIC is only Solaris, and only
with NFSv3 (It may be posssible with NFSv2, but my program doesn't do it
as quickly.). I have a NFS traffic dump of a mere 19K of all nfs traffic to
the machine before the panic. Also, it does NOT ALWAYS cause a panic.
Update, even smaller... 6.5K file, patoot.2 now exists in the same location.
--
David Cross | email: cro...@cs.rpi.edu
Systems Administrator/Research Programmer | Web: http://www.cs.rpi.edu/~crossd
Rensselaer Polytechnic Institute, | Ph: 518.276.2860
Ok, I am hot on the trail... I have found a comonality(sp?) between at least
2 of the Panics. (the 2 I listed)...
it is as follows:
request: create cp1
request: create cp1
reply: ok
reply: error, file exists
request: lookup
request: lookup
(never any response to those.)
I am guessting that: th
I certainly hope they are working under SMP I am running
a 4-way Pentium-III xeon box using vinum and softupdates. so far it has
been a champ.
--
David Cross | email: cro...@cs.rpi.edu
Systems Administrator/Research Programmer | Web: http://www.cs.rpi.edu/~cross
Ok, now that I have hopefully gotten the criticial people's attention, I
will proceed with the details:
1: My test program can only reproduce this with NFSv3/UDP from a recently
patched Solaris system to a FreeBSD server. I have not tested older
Solaris patches, I suspect that they will not caus
I think I found the problem, and I have a pseudo-fix... (the machine nolonger
crashes).
This is a bleeding edge development, I have not had time to refine this code
any. The problem is that nfs_create for NFSv3 does not release the lock
for "vp" with a vput() before it exits. My crude "patch" fo
*** nfs_serv.c Tue Jun 8 15:53:11 1999
--- /cs/crossd/nfs_serv.c Mon Jun 14 16:05:45 1999
***
*** 1343,1348
--- 1343,1349
fhandle_t *fhp;
u_quad_t frev, tempsize;
u_char cverf[NFSX_V3CREATEVERF];
+ int eexistdebug=0;
#ifndef nolint
I have been looking at the code for UMAPfs... I am trying to understand
conceptually why it is so unstable... It looks straightforward enough as
simply passing the calls it receives on to the FS below it, almost like it
didn't exist at all. Why does this cause problems? Isn't the only differenc
Sorry about that everyone, I 'repl'ied to the wrong message.
> Ack, you may have opened up a can of worms here. I don't even think
> that nfs_namei() does the right thing when it returns an error... it
> doesn't look like it clears the ndp->ni_vp either in some error cases.
Who, me?
>> I have been looking at the code for UMAPfs... I am trying to understand
>> conceptually why it is so unstable...
>
>You're looking in the wrong place. It's unstable because of
>infrastructure problems which require fairly substantial amounts of
>work to correct.
>
>DES
I guess that is what I a
> :Umm, okay but I'm a little confused about how the zfree I'm adding to
> :nfs_nget falls under this. Am I being really stupid here?
>
> it's unrelated. I was starting a new thread.
>
> I have finished fixing up nfs_serv.c and am now testing it. Most of
> the procedures require
> The differences between -current and -stable for nfs_serv.c and nfs_subs.c
> are relatively minor. Once we've life tested the hell out of it in
> current, it should be easy to MFC into stable. Maybe 3 weeks total.
Hmm... that is a bit long for us... 3 weeks, 21 days at 1.7 day/pan
I have a drive that is rated at ~16 Meg/second, and indeed it delivers on the
order of 15+ Meg/second. If I use Vinum to create a concatinated device
of 2 such units performance drops to 2.5 Meg/sec. This seems like a
drastic drop in performance. Any ideas what I am doin incorrectly?
--
David
> > I have a drive that is rated at ~16 Meg/second, and indeed it delivers on
> > the
> > order of 15+ Meg/second. If I use Vinum to create a concatinated device
> > of 2 such units performance drops to 2.5 Meg/sec. This seems like a
> > drastic drop in performance.
>
> Indeed, if you're compa
I certainly hope you have applied the recent NFS patches. That should solve
your problem.
--
David Cross | email: cro...@cs.rpi.edu
Systems Administrator/Research Programmer | Web: http://www.cs.rpi.edu/~crossd
Rensselaer Polytechnic Institute, | Ph: 518.27
> A simple start would be to explicitly put a macro or call in each
> syscall to push down the lock. That way people can move that
> macro farther and farther down in the syscall code path, hopefully
> removing it entirely in some cases. I think having the call at
> the beginning of each syscall
There is a small by critical error in the latest patches which causes the
server to never transmit a response packet back to the client in certain
conditions on a nfs create RPC. Below is the updated NFS3 patch. If
jullian could take this for review and place it at the "official" unoffical
URL i
> Miguel Gilly wrote:
> >
> > Bonsai Studio: Web Design and More
> > http://www.bonsai-studio.com
> > Content-type: text/plain; charset="US-ASCII"
> > Content-transfer-encoding: 7bit
> >
> > Dear Sirs,
> >
> > I would find it extremely helpful if FreeBSD could offer redundant
> > clustering capa
The error message in the subject (atapi 1.1: unknown phase) has plagued me
for some time... everything still works, it just displays that error on
the first access to the disk... untill today. Today I am trying to install
FreeBSD 3.2 (19990630) from CDROM. It hangs on probing devices (likely
acd
Hmm... perhaps if Anthony is willing we can use his experience to help us
further document the procedure for writing a FreeBSD PCI device driver?
--
David Cross | email: cro...@cs.rpi.edu
Systems Administrator/Research Programmer | Web: http://www.cs.rpi.edu/~crossd
I am helping a freind install FreeBSD on his machine
(it is running 4.0-CURRENT now). everything works flawlessly, except his
OEM BrookTree 848 based soundcard. The card itself is transplanted from
his gateway machine (where it also had the same problems). Here are some
specifics:
(summary)
Ma
> > No, wait, I got that wrong I think.
> >
> > Oh yah, I remember now. Hmm. How odd. I came across a case where
> > read() could return -1 and not set errno properly if errno
> > was already set, but a perusal of the kernel code seems to indicate
> > that this can't happen.
I am writing a simple, I/O only device driver (no lectures about /dev/io
please ;). It has not PnP abilities, and I have run into the following
problem with bus_set_resource():
static int das1400adc_isa_probe(device_t dev)
{
struct das1400adc_softc *sc = device_get_softc(dev);
Thank you...
After a couple of hours, Jon Chen and I have figured out most of what you
just said :P :)
How would one use hints with a kld?
--
David Cross | email: [EMAIL PROTECTED]
Lab Director | Rm: 308 Lally Hall
Rensselaer Polytech
> > Thank you...
> >
> > After a couple of hours, Jon Chen and I have figured out most of what you
> > just said :P :)
> >
> > How would one use hints with a kld?
>
> Badly. 8( You can only really set them with the loader right now.
>
> There are a couple of kernel datastores that need some
> > Thank you...
> >
> > After a couple of hours, Jon Chen and I have figured out most of what you
> > just said :P :)
> >
> > How would one use hints with a kld?
>
> Badly. 8( You can only really set them with the loader right now.
>
> There are a couple of kernel datastores that need some
I need to use multiple malloc disks for a custom net-boot image I am working
on. The problem is that whenever I access /dev/md1 from the disk it gives
me a 'device not configured' error. I originally thought that this was an
error in how a preloaded image interfaced with the system, but I also g
I recently tried (for the first time) to get gif running under FreeBSD
4.3-BETA (cvsup-ed yesterday). I noticed the following:
gifconfig gif0 inet 10.1.1.1 10.1.2.1
ifconfig gif0 192.168.1.1 192.168.1.2 netmask 0xff00
and then I 'ping 192.168.1.1' it will try to route the packet instead of
Well, I am able to reproduce the crash pretty reliably, I don't know what is
causing it yet, I just kill all the other ypservs on a subnet except for this
one and it crashes about once every 5 minutes. I have some questions/theories
that I'd like to bounce off of people:
1) In the yp_all functi
I have trace the problem in ypserv down to the RPC dispatch routines..
I am digging further and I hope to have it found and eliminated today
(in time for -RELEASE ;)
If anyone has any idea how it could be tripping up here, please let me
know. My 2 guesses are a corrupted svc_callback entry (no
I have found _a_ bug in ypserv (I think I may be stumbling over multiple
different bugs, but this one is very reproducable).
It is dying in the yp_testflags routine, in the for loop that goes through
the CIRCLEQ. The loop dies with qptr pointing to a struct that is all NULL
(my reading of CIRCLE
Ok... I am coming to the conclusion that there is some sort of kernel
issue that is causing this problem. Here is what I have done and discovered
to date (this is all with 4.3-RC2 FWIW):
At some point the 'qhead' CIRCLEQ structure in yp_dblookup.c gets corrupted.
This is declared as a static, an
After some more intensive debugging, and a leap of faith, I _think_ I have
the problem licked, but I would appreciate some more brains to examine the
logic.
The original cause of ypserv's problems was the sharing of DBPs between
the parent and child. The resolution to this was to close all of
The ypserv bug (the one where ypserv randomly stops responding or
just seg-faults) is still very much alive. I had to restart it
about 11 times in the course of 20 minutes this morning. That's
the bad news, the good news is that I started it each time with
'ktrace -i'.
Going back a bit, Matt
1 - 100 of 218 matches
Mail list logo
