I've occasionally seen systems wedged in a similar state. I reported
my sighting of this on May 24th. Haven't seen it since.
The one bit of useful info I've learned since my report was that from
a talk with the program's author, I suspect the object in question may
have been created with
quite well (SimOS, Netscape, Mathematica,
Matlab, S-Plus, emacs, etc all run), but I'd really like somebody to
look it over before I commit it. I've left it at
http://www.freebsd.org/~gallatin/osf1.tar.gz
------
Andrew Gallatin,
--
Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin
Duke University Email: [EMAIL PROTECTED]
Department of Computer Science Phone: (919) 660-6590
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" i
Doug Rabson writes:
On Thu, 19 Aug 1999, Andrew Gallatin wrote:
I do most of my development on alphas I just turned some local code
into a loadable kernel module. It works fine when compiled into the
kernel statically, but fails miserably when loaded into an alpha
kernel
}
len = tlen = strlen (member);
Anybody interested in comitting this? I passed it by the person who
committed 1.13 of arch.c was ignored. I don't know make well
enough to feel comfortable committing this myself.
Drew
-------
being set to "/".
There's a PR about this now. The smaller fix in the PR seems to work.
Ah yes, in bin/14167. The fix is much tighter appears to work.
------
Andrew Gallatin, Sr Systems Progra
Julian Elischer writes:
that fix was applied by me about 5 minutes ago..
Thank you!
Drew
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
Matthew Jacob writes:
UDP. Local network. Very puzzling.
Oh well. So much for a shot in the dark..
--
Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin
Duke University
? Should it be returning some sort of error?
Thanks,
Drew
------
Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin
Duke University Email: [EMAIL PROTECTED]
Department of Computer Science P
(nfs/nfs_socket.c:110) or nfs_readrpc
(nfs/nfs_vnops.c:1093). These are both calls to nfs macros that would
be a lot easier to debug if they weren't macros ;-)
Thanks,
Drew
--
Andrew Gallatin, Sr Systems Programmer http
in the
code.
-Matt
I'm all in favor of having all the developers have alphas so these
things get caught early ;-)
Cheers,
Drew
--
Andrew Gallatin, Sr Systems Programmer
Matthew Dillon writes:
Well, there was a bug in nfsrv_create() which caused the server to
not reply to an NFS packet. This led to a general revamping of the
server side code which may have fixed other rpc's at the same time.
Whether fixing that bug solves the problem
--
Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin
Duke University Email: [EMAIL PROTECTED]
Department of Computer Science Phone: (919) 660-6590
To Unsubscribe: send mail to [EMAIL PROTECTED
not trying to contribute to the FUD. I'm just not sure I
understand how what you are proposing will affect me.
Thanks,
Drew
--
Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin
Duke University
of any x86 hardware I've seen which is in the same price range.
I haven't played with a rambus based machines yet though..
Drew
--
Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin
Duke University
, their symbols will not be present in the kernel.
This is a good time to rebuild your modules anyway, as pal.s will be
going away shortly.
Andrew Gallatin writes:
gallatin1999/11/29 12:31:46 PST
Modified files:
sys/alpha/includeipl.h
sys/alpha/alpha ipl_funcs.c
Log
--
Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin
Duke University Email: [EMAIL PROTECTED]
Department of Computer Science Phone: (919) 660-6590
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" i
David O'Brien writes:
On Thu, Oct 24, 2002 at 06:39:15PM -0400, Jake Burkholder wrote:
You can also get various new machines on sun.com for around $1000 USD,
IIRC a 500mhz blade 100 does a buildworld in around 2-3 hours.
A $1000 (new) 500 MHz blade running GENERIC (minus WITNESS)
Gordon Tetlow writes:
On Wed, Oct 30, 2002 at 11:50:45AM -0800, Tim Kientzle wrote:
I find the standard arguments used by RCng quite
awkward. In particular, especially for people who
have worked with SysV-style init scripts, it's
rather surprising that /etc/rc.d/nfsd stop does
Joel M. Baldwin writes:
...
don't think this is related to the X FP problem ( although I am
running X11 ). There have been many times when I'd walk in the
...
options DDB #Enable the kernel debugger
...
Its likley you are panic'ing in X, and the system is
DDB is smart enough (if you are using sc(4)) to notice that the display
is in graphics mode and won't drop into a debugger prompt if you panic.
Coolness. Must be something new..
Drew
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the
Tom Ponsford writes:
Hi All,
I'm getting a machine check on my new-toy: an Alphaserver 2100A 4/275 RM.
It boots and runs 4.6.2 fine with a single cpu. But after I supped the
current 5.0 sources from
10/29/2002 17:39 and did a buildworld and a Generic buildkernel, (plus
all
Try backing out 1.544 of src/sys/i386/i386/machdep.c You'll need to do
it as a reversed patch or by hand, as there are some unrelated signal
handling things in 1.545 which you'll really need.
Drew
Michael G. Petry writes:
I'm noticing the same behavior on a PPro system I have and am
Harti Brandt writes:
On Mon, 11 Nov 2002, TOMITA Yoshinori wrote:
This is probably not a bug, but a feature. You are not expected to access
a variable through a pointer to a non-compatible type. int and short are
not compatible. (see your ISO C standard on this topic).
Try to use
Robert Watson writes:
On Thu, 21 Nov 2002, Luigi Rizzo wrote:
...
The label is 5 ints, the pkthdr a total of 11 ints (and m_hdr takes
another 6, for a total of 136 bytes of header info on 64-bit
architectures).
Of the pkthdr, only 3 fields (rcvif, len, tags) are of really
Julian Elischer writes:
On Sat, 23 Nov 2002, Andrew Gallatin wrote:
As you eloquently state, there are a number of tradeoffs involved. On
a 64-bit platform, 99% of users are paying 40 bytes/pkt for something
that they will never use. On x86, 99.99% of users are paying 20
Bosko Milekic writes:
On Sun, Nov 24, 2002 at 04:23:33PM -0500, Andrew Gallatin wrote:
If we're going to nitpick the mbuf system, a much, much worse problem
is that you cannot allocate an mbuf chain w/o holding Giant, which
stems from the mbuf system eventually calling kmem_malloc
Luigi Rizzo writes:
The mbuf bloat has two aspects -- first it does have some cost to
initialize and reset all these extra fields (and it is bug prone --
witness is the missing cleanup in m_getcl(), because m-tags were
introduced after m_getcl() and probably it was forgotten); second,
a
Bosko Milekic writes:
Firstly, it should be noted that the behavior of calling kmem_malloc()
when its caches are empty is an old property that has been carried
over from the original allocator - in other words, it is not something
that I arbitrarily introduced.
Certainly. I'm
Bosko Milekic writes:
Well, first of all, I never call kmem_malloc() with any locks held so
this argument about grabbing Giant while other locks are held is not
applicable in my case.
Well, not for the mbuf allocator itself, but for its consumers.
Given that you call the
Bosko Milekic writes:
back. We'll also need a kproc that can wake up every now and then to
expand the pool if allocations at interrupt time failed. Or do you
already have a mechanism for that?
The intended mechanism is the kproc and when the allocator was first
designed
Robert Watson writes:
Andrew,
Thanks for your patience as I finished some research and experimentation
regarding the options there. Some more details below.
On Sat, 23 Nov 2002, Andrew Gallatin wrote:
On the contrary, I think that if anything is going to be done, it must
Terry Lambert writes:
Andrew Gallatin wrote:
What I (as a 3rd party driver author working in a GNUish
...
How is one supposed to build a 3rd party module these days?
One is not. The vendor supplies only a binary.
Damn it Terry, I AM the vendor. Somtimes I wonder if you even
Terry Lambert writes:
Andrew Gallatin wrote:
Terry Lambert writes:
Andrew Gallatin wrote:
What I (as a 3rd party driver author working in a GNUish
This is how I do it.
...
How is one supposed to build a 3rd party module these days?
How are you supposed
Kris Kennaway writes:
I'm getting this on several of my alphas. Any ideas? The traceback
and panic message is weird.
gdb sucks bigtime on alpha. The weirdness is just gdb throwing its
hands up in the air because it doesn't understand the trapframe and
can't grok anymore of the stack than
Brad Knowles writes:
At 3:32 PM -0700 2002/12/02, Cliff L. Biffle wrote:
One thing I've used in the past that improves Realtek throughput is forcing
the media type and duplex setting on both ends of the connection. Autodetect
in the 8139s seems to be unreliable at times.
Hi,
I've been hacking on DMA windows on alpha. While I've been getting
things wrong, I've been provoking a panic in the ata driver because
it cannot do DMA when attaching a disk.
The panic is triggered at the end of ad_attach() by the raid probe
causing a transfer to fail and the disk to be
Kris Kennaway writes:
I got this on one of the gohan machines overnight. These machines
have failing disks -- I get a lot of hard read errors, but the
INVARIANTS panic could better be replaced by something else.
I reported another instance of this to sos a few weeks ago but didn't
I just upgraded my UP1000 from 4.7-stable to 5.0.
Only weird thing left is the console seems to drop a lot of characters
just after syscons takes over.
I see this on console:
...
ppc0: interrupting at ISA irq 7
sc0: System console on isa0
sc0: VGA 16 virtual consoles, flags=0x0
Wilko Bulte writes:
On Tue, Dec 10, 2002 at 11:41:35PM +0100, Wilko Bulte wrote:
Well, I think my DS10 does not have that problem as in:
..
50 18 10 00 00 00 03 00 02 67 60 4f 50 83 55 81
..
Try booting without verbose..
Drew
To Unsubscribe: send mail to [EMAIL PROTECTED]
with
Wilko Bulte writes:
On Wed, Dec 11, 2002 at 02:22:12PM -0500, Andrew Gallatin wrote:
Wilko Bulte writes:
On Tue, Dec 10, 2002 at 11:41:35PM +0100, Wilko Bulte wrote:
Well, I think my DS10 does not have that problem as in:
..
50 18 10 00 00 00 03 00 02 67 60 4f 50
Ugh. Since it may call kmem_malloc(), UMA must hold Giant.
This is the same problem the mbuf system has, and its what's
keeping network device drivers under Giant in 5.0.
Both subsytems should probably have GIANT_REQUIRED at all entry
points so as to catch locking problems like this earlier.
This was on an x86, 5.0-current as of roughly 8am EST today.
Machine panic'ed when it finished /usr and moved on to /var on
a manually invoked fsck -B (I'd hit ^C to abort multi-user startup
during fsck and was surprised to see the system continue on.. ;)
Drew
panic: ffs_blkfree: freeing free
Hi,
I've got an Intel D845EBG2. I have 2 ACPI related problems I'm
looking for advice on:
- ATI Rage 128 Pro TF
If I suspend to S3 from text mode on console, the screen remains blank
after resume. The system is otherwise functional, and can be logged
into remotely.
If I start X after
Kyunghwan Kim writes:
Is it okay to add INTR_MPSAFE for all INTR_TYPE_NET drivers?
NO!
mbuf and bpf routines are all mp-safe, so it seems that
it is safe to make network device drivers out of Giant lock.
Or is there any unresolved related issues?
Yes, the mbuf allocator must
Kyunghwan Kim writes:
On Wed, Dec 18, 2002 at 04:53:00AM +0900, Kyunghwan Kim wrote:
On Tue, Dec 17, 2002 at 02:31:31PM -0500, Andrew Gallatin wrote:
mbuf and bpf routines are all mp-safe, so it seems that
it is safe to make network device drivers out of Giant lock
Bosko Milekic writes:
On Mon, Nov 25, 2002 at 08:13:46PM -0500, Andrew Gallatin wrote:
It is not out of date. The code means:
If you've given me a counter then I'll use it otherwise I'll try to
allocate one with malloc().
Ah, duh. Thanks. I'd better start
Bosko Milekic writes:
On Thu, Jan 02, 2003 at 01:53:53PM -0500, Andrew Gallatin wrote:
I'm just tuning up my driver now to catch up to the recent interface
changes. While there, I went to add a ref count for my driver managed
M_EXT clusters. However, m_extadd() does not take
Bosko Milekic writes:
Yeah, this looks like the least-intrusive way to do it. I'm okay with
the patch. I like the idea of using an EXT-type flag to mark the data
buffer types using this method. Thanks.
Thanks.. Committed.
P.S.: Try not to use MEXTADD, if possible. Use
M. Warner Losh writes:
..
However in if_slowtimo we have:
if_slowtimo(arg)
{
... IFNET_RLOCK();
... if (ifp-if_watchdog)
(*ifp-if_watchdog)(ifp);
... IFNET_RUNLOCK();
}
and dc_watchdog does a DC_LOCK/UNLOCK pair). This is a Lock Order
M. Warner Losh writes:
In message: [EMAIL PROTECTED]
Andrew Gallatin [EMAIL PROTECTED] writes:
: The IFNET_RLOCK() called in if_slowtimo() is a global lock for the
: list of ifnet structs to ensure that no devices are removed or added
: while something may be using
Are kernel modules pessimized in any way with respect to using
mutexes as compared to statically compiled kernel code?
I seem to remember some discussion a year or more ago indicating that
they would be, but I'm not seeing it in the code.
Thanks,
Drew
To Unsubscribe: send mail to [EMAIL
John Baldwin writes:
On 07-Jan-2003 Andrew Gallatin wrote:
Are kernel modules pessimized in any way with respect to using
mutexes as compared to statically compiled kernel code?
I seem to remember some discussion a year or more ago indicating that
they would be, but I'm
I'm just wondering -- when is RC3 expected? ? A lot has happened
since the 2/14 snap..
Thanks,
Drew
--
Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin
Duke University
'
Drew
--
Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin
Duke University Email: [EMAIL PROTECTED]
Department of Computer Science Phone: (919) 660-6590
To Unsubscribe
ime to MFC Jonathan Lemon's checksum offloading code.
Doing this would require changing MSIZE to 256, which in turn would
require recompiling any module using mbufs (all NICs, network
filesystems, etc).
Cheers,
Drew
------
Andrew Ga
Adam Migus writes:
Folks,
While doing some performance analysis (doing make -j5 buildkernel)
on a set of 14 kernels I've hit one using the SCHED_ULE scheduler
that hangs. It happens every time but not necessarily in the same
place in the make.
...
The hardware is a dual Xeon
Craig Rodrigues writes:
On Thu, Aug 14, 2003 at 08:17:33PM -0400, Andrew Gallatin wrote:
You have machdep.hlt_logical_cpus: 1 in your sysctl output. [BTW,
lots of people read this mail via the web archives at
http://docs.freebsd.org/cgi/getmsg.cgi?fetch=1073654+0+current/freebsd
John Baldwin writes:
No, generic modules would always work with all kernels except for
exceptional cases like PAE (unavoidable, really), and MUTEX_PROFILING
(this is a debugging thing, so ISV's wouldn't need to ship modules
with that turned on). All this would add is the ability to
I see an ICE building the math/R-letter port on -current (x86) from
late last week.
% gcc -v
Using built-in specs.
Configured with: FreeBSD/i386 system compiler
Thread model: posix
gcc version 3.3.1 [FreeBSD] 20030711 (prerelease)
cc -I../../src/extra/pcre -I. -I../../src/include
Tinderbox writes:
cc -O -pipe -mcpu=ev4 -mtune=ev5 -mieee -DIN_GCC -DHAVE_CONFIG_H -DPREFIX=\/usr\
-I/vol/vol0/users/des/tinderbox/CURRENT/alpha/alpha/obj/alpha/vol/vol0/users/des/tinderbox/CURRENT/alpha/alpha/src/gnu/usr.bin/cc/f771/../cc_tools
I'm running a ~2 week old -current, and when debugging some hacks of
mine, I noticed that gdb -k seems to be missing at least one frame
in its stack. Eg, ddb shows:
panic: page fault
cpuid = 0; lapic.id =
Stack backtrace:
backtrace(c0380cf2,0,c036e757,d96d4bd4,100) at backtrace+0x17
Sam Leffler writes:
It reliably locks up for me when you break into a running system; set a
breakpoint; and then continue. Machine is UP+HTT. Haven't tried other
machines.
Perhaps related, perhaps a red-herring: With a single P4 + HTT, +
SMP kernel, if I break into the ddb debugger
Hi,
I recently decided to update my alpha UP1000 to today's current from a
mid-July build. However, UDMA33 did not work on a hard disk attached
to the built-in Acer Aladdin controller (verbose dmesg appended).
I think I have narrowed the problem down to this line of code:
Soren Schmidt writes:
Yep, thats close, I have a patch out for testing that looks semilar,
if you can confirm it works, I'll commit it asap:
And yes pointy hat to me :)
This fixes my box with the acer aladdin chip.
Thanks!
Drew
___
Poul-Henning Kamp writes:
As soon as these uses of cloning code has been removed, I will move
the floppy and CD drivers under GEOM, paving the way for the
significant changes to the buf/VM system which some of you have
already heard rumours about. (more will emerge after BSDcon'03)
David Malone writes:
This may be my fault, as I made some changes recently that assumed that
the mbuf allocator grabbed giant when needed. I'll check the code path
you've mentioned to see if it grabs giant now, but I suspect that I just
need to move the giant grabbing back where
Alan L. Cox writes:
Thanks for letting me know. This is another false positive: Witness
can't distinguish the lock on the object being destroyed from the lock
on the object used by UMA because their labels are the same. They will
never, however, be the same object. So, deadlock isn't a
Jeff,
On an SMP box I have, which is really a p4 box with one physical
CPU, and 2 HTT cores, I've seen some strange behaviour with ULE.
With ULE enabled, I've see jobs wedge for no apparent reason.
Some examples are fsck, dhclient and gcc.
Here's an example of fsck after it stopped responding:
Jeff Roberson writes:
On Fri, 27 Jun 2003, Andrew Gallatin wrote:
Jeff,
On an SMP box I have, which is really a p4 box with one physical
CPU, and 2 HTT cores, I've seen some strange behaviour with ULE.
With ULE enabled, I've see jobs wedge for no apparent reason.
Some
Jeff Roberson writes:
Can you call kseq_print(0) and kseq_print(1) from ddb?
I found a different problem which is nearly as interesting.
Note that ps thinks sysctl is on cpu 255...
Boot hangs here:
cd0: Attempt to query device size failed: NOT READY, Medium not
present
SMP: AP CPU #1
Jeff Roberson writes:
On Fri, 27 Jun 2003, John Baldwin wrote:
On 27-Jun-2003 Andrew Gallatin wrote:
Jeff Roberson writes:
Can you call kseq_print(0) and kseq_print(1) from ddb?
I found a different problem which is nearly as interesting.
Note
Kris Kennaway writes:
On Thu, Jul 31, 2003 at 01:48:59AM -0700, Kris Kennaway wrote:
I upgraded the alpha package machines tonight, and one of them fell
over shortly after taking load, with the following:
..
Two more panics on alpha:
panic: vm_fault: fault on nofault entry,
Kris Kennaway writes:
On Fri, Aug 01, 2003 at 09:00:44AM -0400, Andrew Gallatin wrote:
The crashdump might actually be useful here. You'd have only the
trap() and vm_fault() frames, but at least you'd have information
about the state of the vm system.
Two crashdumps coming up
John Baldwin writes:
On 14-Aug-2003 Ruslan Ermilov wrote:
On Thu, Aug 14, 2003 at 02:10:19AM -0600, Scott Long wrote:
Luoqi Chen wrote:
[...]
On the other hand, all modules should create all the opt_*.h files
it needs when built individually. Add opt_ddb.h to nullfs's Makefile
John Baldwin writes:
On 14-Aug-2003 Andrew Gallatin wrote:
John Baldwin writes:
On 14-Aug-2003 Ruslan Ermilov wrote:
On Thu, Aug 14, 2003 at 02:10:19AM -0600, Scott Long wrote:
Luoqi Chen wrote:
[...]
On the other hand, all modules should create all
[EMAIL PROTECTED] writes:
In message [EMAIL PROTECTED], Kris Kennaway writes:
--ALfTUftag+2gvp1h
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
The tool costs $699 and Intel said it will be available in February.
That would make it somewhat difficult,
Kris Kennaway writes:
I just got this on one of the axp machines [*]:
unexpected machine check:
mces= 0x1
vector = 0x670
670 is a cpu machine check -- thats most likely an uncorretable
memory parity error or some other (intermittent) hardware failure
caused by
Mike Tibor writes:
On Thu, 16 Jan 2003, Trevor Johnson wrote:
I just got a similar crash:
unexpected machine check:
mces= 0x1
vector = 0x670
I believe a 670 machine check can also result from a read of a
non-existent I/O space. I'm not a
Mike Tibor writes:
On Thu, 16 Jan 2003, Andrew Gallatin wrote:
(I wrote: )
I believe a 670 machine check can also result from a read of a
non-existent I/O space. I'm not a programmer, but could that be the
problem here?
No, that's a 660. (system machine check
Alexander Leidinger writes:
On Wed, 15 Jan 2003 19:32:39 +0100
Poul-Henning Kamp [EMAIL PROTECTED] wrote:
They have a beta test access to vtune:
http://www.intel.com/software/products/vtune/vlin/
Bye,
Alexander.
I signed up and waded through tons of slow https js
Alexander Leidinger writes:
On Fri, 17 Jan 2003 10:01:35 -0500 (EST)
Andrew Gallatin [EMAIL PROTECTED] wrote:
I signed up and waded through tons of slow https js laden forms, only
to finally be presented with nothing but a self-extracting .exe file
which looks like a patch
I'm re-examining some of our driver code where we need to wire down a
portion of a user's address space for DMA for os-bypass networking.
We currently do vm_map_wire(). This is nice, as it presents a simple
interface, but I think its pretty high overhead the way our driver
calls it (a page at a
Hajimu UMEMOTO writes:
dwmalone2) The patches wouldn't compile on the alpha 'cos of some
dwmaloneifdefs in the linux emulation code. I just hadn't got around
dwmaloneto resolving this.
Oops, thank you for pointing this out. It seems that other than
linux_connect() are
Dag-Erling Smorgrav writes:
Memory modified after free 0xfc7b6000(8184)
panic: Most recently used by none
It might be interesting to add some printfs to see what the value
is currently, vs what was expected, and where in the zone the
modification happened..
This is a UP box right?
Dag-Erling Smorgrav writes:
Andrew Gallatin [EMAIL PROTECTED] writes:
This is a UP box right?
Yes, a PWS 600au.
OK, good.. ;) There had been SMP problems with the UMA debug code a
long time ago.
Drew
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current
Eric Anholt writes:
As a side note, do any of these AGP chipsets apply to sparc64, alpha,
ia64?
The alpha UP1000 has an amd-751 chipset and should be able to use
the amd agp module. I tried it shortly after it was brought into
the tree a few years ago, and it locked the box solid. I was
Chris BeHanna writes:
At 4:36 PM -0800 2003/02/13, Scott Long wrote:
- Fstress - http://www.cs.duke.edu/ari/fstress
SpecFS (NFS ops/sec benchmark)
Have you ever actually used SPECsfs97? In addition to being
encumbered, SPECsfs97 is pain to keep running (dies at the
Dag-Erling Smorgrav writes:
Dag-Erling Smorgrav [EMAIL PROTECTED] writes:
Any suggestions as to how I can figure out who used that block of
memory before it was allocated to the ess driver?
I threw in a call to Debugger(), but...
mtrash_dtor(0xfc7b6000, 8192, 0)
here's
Dag-Erling Smorgrav writes:
Andrew Gallatin [EMAIL PROTECTED] writes:
Can you look at the registers and match $ra with a line number using
addr2line or gdb? (sorry, forgot if ddb can even look at registers)
Uh, I know *where* it stops since I added the call to Debugger
Dag-Erling Smorgrav writes:
Andrew Gallatin [EMAIL PROTECTED] writes:
You might be able to get some idea of what's happening by enabling KTR
and tracing everything, then dumping the trace buffer at your
breakpoint.
Of course, the KTR-enabled kernel fails to crash.
*sigh
Can you post your kernel config please, along with with, if any,
CPUTYPE you have set in make.conf and a description of your machine
(mem size in particular)?
I'm unable to reproduce the boot panic with a GENERIC as of
today. (rest of machine is from Nov 1st). Rebuilding without
CPUTYPE
Dag-Erling Smorgrav writes:
#options CD9660 #kld
Do you preload any/all of the things you've marked as klds?
I've tried to duplicate your crash on my xp1000. I've used your
kernel config file. I've reduced my memory size to 254MB.
I've got
Dag-Erling Smorgrav writes:
Andrew Gallatin [EMAIL PROTECTED] writes:
Do you preload any/all of the things you've marked as klds?
No...
Damn. I'm sorry then, I think I've done all I can to try to duplicate
it. Would you mind doing a binary search to find out when your problem
Hi Soren,
After recent ATA commits, my Promise UDMA66 controller is now running
its drives in PIO4 mode. Previously, UDMA66 was working fine.
Here's a dmesg snippet:
atapci0: Promise UDMA66 controller port 0xdf00-0xdf3f,0xdfe0-0xdfe3,0xdfa8-0xd
faf,0xdfe4-0xdfe7,0xdff0-0xdff7 mem
Soeren Schmidt writes:
ATAPI_DEVICE is used on those controllers that cannot do ATAPI DMA,
the test here is bogusly reversed, I'll fix asap...
Aha! Thanks!
Drew
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
I just tried to buildworld, and xlint crapped out like this:
=== usr.bin/xlint/llib
lint -cghapbx -Cposix /usr/src/usr.bin/xlint/llib/llib-lposix
lint -cghapbx -Cstdc /usr/src/usr.bin/xlint/llib/llib-lstdc
llib-lposix:
llib-lstdc:
stdio.h(79): warning: struct __sFILEX never defined [233]
Lint
I'm developing a character driver which tracks a lot of state on a
per-open basis. I've got several mutexes in there which are
initialzed at open, and destroyed at close. After a few
dozen opens, witness seems to croak with:
witness_get: witness exhausted
Am I leaking something? Or is
John Baldwin writes:
Unfortunately dead witnesses may still be stuck in the lock order
hierarchy and I haven't figured out yet how to properly handle the
case of free'ing a witness structure from the tree while preserving
the correct lock orders. You can try
Ah, I think I see.
David Leimbach writes:
True... I guess I didn't state my case clearly enough that I think IP
over firewire
is in itself a good thing for clusters.
From my experience with the Apple IP over Firewire, it seems slow, and
very high overhead. A dual 800MHz G4 host which can transmit at well
1 - 100 of 302 matches
Mail list logo