Matthew Jacob wrote:
I found all of the knobs that turn off harvesting, and with those off, my
alpha 4100 boots again w/o hanging.
So- there are knobs (but nothing in UPDATING warned me to turn them off in
order to boot again).
The commit message described that the option would
Hi,
In addition to reported earlier general machine slowdown with interrupt
harvesting is turning on, ethernet entropy harvesting seriously hammers network
performance as well. Ftping big file over my 10M network now about 15% slower
with ethernet harvesting turning on.
Mark, please get slow
In addition to reported earlier general machine slowdown with
interrupt harvesting is turning on, ethernet entropy harvesting
seriously hammers network performance as well. Ftping big file over my
10M network now about 15% slower with ethernet harvesting turning on.
Even with the Rijndael
Mark Murray wrote:
In addition to reported earlier general machine slowdown with
interrupt harvesting is turning on, ethernet entropy harvesting
seriously hammers network performance as well. Ftping big file over my
10M network now about 15% slower with ethernet harvesting turning on.
Can everyone please test this patch against OpenSSL, which should fix
the problems observed with -j builds as well as cleaning out the
temporary ASM files.
This fix will need to be committed to -stable prior to the release, as
it shares the same code.
Kris
Index: Makefile
David Wolfskill wrote:
Date: Sun, 11 Mar 2001 15:11:09 -0800
From: Kris Kennaway [EMAIL PROTECTED]
To: Hajimu UMEMOTO [EMAIL PROTECTED]
We really need to provide a better rc.conf hook for doing this --
expecting people to write their own script just to create a /tmp is
lame.
I
Hi,
dmesg and the output of "ident /sys/dev/aic7xxx/*" and pciconf is
attached (BTW: the -v options to pciconf isn't documented in the
synopsis section of the man page).
Do you need more information, e.g. the output of a verbose boot?
Bye,
Alexander.
--
Where do you think
Hello everybody,
I had been away for two weeks and after upgrading to the latest -CURRENT I
noticed that leaving
device midi
(and maybe device seq, I did not test separately)
in my kernel config file causes a Trap 12 with interrupts disabled on
_mtx_lock_sleep+0x29a: movl 0x1a0(%edx),%eax
Hi,
In /usr/src/etc/Makefile:
"make distribution" is still trying to copy MAKEDEV to /dev
on a system with devfs mounted to /dev.
Since devfs is default, is this behaviour correct or my
/etc/make.conf is missing something ?
regards,
--
Jean Louis Ntakpe Texas Instruments -
Maxim Sobolev wrote:
Mark Murray wrote:
In addition to reported earlier general machine slowdown with
interrupt harvesting is turning on, ethernet entropy harvesting
seriously hammers network performance as well. Ftping big file over my
10M network now about 15% slower with
In message [EMAIL PROTECTED], Jean Louis Ntakpe writes:
Hi,
In /usr/src/etc/Makefile:
"make distribution" is still trying to copy MAKEDEV to /dev
on a system with devfs mounted to /dev.
Since devfs is default, is this behaviour correct or my
/etc/make.conf is missing something ?
I think that
I changed nothing from whatever the default is. It seems like a bit of POLA to
freeze now.
But I'll check this - if I can get that machine up again :-)...
OK - if this is the entropy driver, then typing about 2 lines of shit
will unlock it.
That did not fix the problem. Only
On Fri, 9 Mar 2001, Kris Kennaway wrote:
On Fri, Mar 09, 2001 at 09:32:29AM -0800, Matthew Jacob wrote:
I changed nothing from whatever the default is. It seems like a bit of POLA to
freeze now.
But I'll check this - if I can get that machine up again :-)...
Press ^T when it
On Fri, 9 Mar 2001, John Baldwin wrote:
On 09-Mar-01 Matthew Jacob wrote:
On Fri, 9 Mar 2001, Mark Murray wrote:
I changed nothing from whatever the default is. It seems like a bit of
POLA to
freeze now.
But I'll check this - if I can get that machine up again :-)...
Wait a minute. isn't this all old mail being resent? What's going on?
I did a buildworld/installworld on an alpha yesterday, and now I'm left with:
start_init: trying /sbin/init
Entropy harvesting: interrupts ethernet.
hangs
And this is even with booting from an older
On Mon, 12 Mar 2001, Mark Murray wrote:
In addition to reported earlier general machine slowdown with
interrupt harvesting is turning on, ethernet entropy harvesting
seriously hammers network performance as well. Ftping big file over my
10M network now about 15% slower with ethernet
On Mon, 12 Mar 2001, Maxim Sobolev wrote:
Maxim Sobolev wrote:
On Mon, Mar 12, 2001 at 12:29:58AM -0300, Mario Sergio Fujikawa Ferreira wr=
ote:
Hi,
=20
Is there anything against adding support for
k6-3 to the just added CPUTYPE mechanism? :)
My little machine
Do you have WITNESS_SKIPSPIN option in your kernel config?
Here is what supposedly causing the trouble:
a) the process p_spinlocks variable is initialized to one in fork1 during
the process creation
b) the sched_lock is released later in fork_exit, but the process'
p_spinlocks field is
On 12-Mar-01 Alexander N. Kabaev wrote:
Do you have WITNESS_SKIPSPIN option in your kernel config?
Here is what supposedly causing the trouble:
a) the process p_spinlocks variable is initialized to one in fork1 during
the process creation
b) the sched_lock is released later in
Even with the Rijndael code and kern.random.sys.burst=$SMALLNUMBER ??
Rijndael stops it showing up much in top -S. I'm wondering where it is
hiding :-). kern.random.sys.burst=$SMALLNUMBER had very little effect.
The Rijndael code makes a 2-orders-of-magnitude difference to the speed
of
Please try this patch. This should solve all the random harvesting
performance issues no matter how efficient or inefficient the hash
function (untested as I do not have a -current box at the moment).
-Matt
Index: yarrow.c
On Tue, Mar 13, 2001 at 12:12:35AM +0900, Daniel C. Sobral wrote:
A while ago someone suggested a /etc/md.conf and an mdon(1) similar to
swapon(1).
Putting it in terms of this analogy make this approach sound quite
reasonable.
This solution is much more flexible than simple /tmp fs on md
In message [EMAIL PROTECTED], "David O'Brien" writes:
On Tue, Mar 13, 2001 at 12:12:35AM +0900, Daniel C. Sobral wrote:
A while ago someone suggested a /etc/md.conf and an mdon(1) similar to
swapon(1).
Putting it in terms of this analogy make this approach sound quite
reasonable.
I can fully
On Sat, Mar 10, 2001 at 10:32:23PM -0800, Dima Dorfman wrote:
Peter Wemm [EMAIL PROTECTED] writes:
The version of the patch for -current uses the softdep mount option only.
If you remove the mount option, you dont get softupdates.
In this case, it might be better to just turn it on by
Speaking of md, and such, since MFS got nuked, and I died horribly every time
I tried to use md as a tmpfs, have the panics been fixed so I can use it now
as a replacement for MFS?
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
Sorry, the last patch won't patch cleanly, I forget to update my
-current source before diffing. A new patch is attached, and it
also includes reducing the sdize of the entropy ring from 1024 to a
more reasonable 64.
Mark, what I said last month still holds... you need to
In message [EMAIL PROTECTED], Matthew
Jacob writes:
Speaking of md, and such, since MFS got nuked, and I died horribly every time
I tried to use md as a tmpfs, have the panics been fixed so I can use it now
as a replacement for MFS?
Uhm, I'm drawing a blank here. When did you have panics ?
In message [EMAIL PROTECTED], "David O'Brien" writes:
On Sat, Mar 10, 2001 at 10:32:23PM -0800, Dima Dorfman wrote:
Peter Wemm [EMAIL PROTECTED] writes:
The version of the patch for -current uses the softdep mount option only.
If you remove the mount option, you dont get softupdates.
In
Please try this patch. This should solve all the random harvesting
performance issues no matter how efficient or inefficient the hash
function (untested as I do not have a -current box at the moment).
Erm, you are behind :-)
I have already committed something that does this in a
:
: Please try this patch. This should solve all the random harvesting
: performance issues no matter how efficient or inefficient the hash
: function (untested as I do not have a -current box at the moment).
:
:Erm, you are behind :-)
:
:I have already committed something that does
On Mon, Mar 12, 2001 at 08:27:54PM +0100, Poul-Henning Kamp wrote:
If it looks like a duck, swims like a duck, quacks like a duck,
Sorry David, but it there is nothing duck-like about at all...
From a user's stand point, it acts just like the old MFS when used to
create a swap backed
Date: Mon, 12 Mar 2001 11:29:50 -0800 (PST)
From: Matthew Jacob [EMAIL PROTECTED]
Speaking of md, and such, since MFS got nuked, and I died horribly every time
I tried to use md as a tmpfs, have the panics been fixed so I can use it now
as a replacement for MFS?
Well, it appears to work OK for
On Mon, 12 Mar 2001 20:33:59 +0100, Poul-Henning Kamp wrote:
Problem is many still feel it should not be used on / .
Why not ?
Because a small root partition fills up artificially during "make
installworld" and/or "make installkernel". Everybody understands _why_
it happens, but that
In message [EMAIL PROTECTED], "David O'Brien" writes:
On Mon, Mar 12, 2001 at 08:27:54PM +0100, Poul-Henning Kamp wrote:
If it looks like a duck, swims like a duck, quacks like a duck,
Sorry David, but it there is nothing duck-like about at all...
From a user's stand point, it acts just
In message [EMAIL PROTECTED], Matthew
Jacob writes:
Speaking of md, and such, since MFS got nuked, and I died horribly every time
I tried to use md as a tmpfs, have the panics been fixed so I can use it now
as a replacement for MFS?
Uhm, I'm drawing a blank here. When did you have
Mark, something like this doesn't REQUIRE any configuration!!! Don't
add confusion to the system. Just make the default something
reasonable. There is absolutely no reason to have to be able to adjust
the interrupt seeding code if the default is made something reasonable.
Date: Mon, 12 Mar 2001 20:33:59 +0100
From: Poul-Henning Kamp [EMAIL PROTECTED]
In message [EMAIL PROTECTED], "David O'Brien" writes:
On Sat, Mar 10, 2001 at 10:32:23PM -0800, Dima Dorfman wrote:
Peter Wemm [EMAIL PROTECTED] writes:
The version of the patch for -current uses the softdep mount
In message [EMAIL PROTECTED], Matthew
Jacob writes:
In message [EMAIL PROTECTED], Matthew
Jacob writes:
Speaking of md, and such, since MFS got nuked, and I died horribly every time
I tried to use md as a tmpfs, have the panics been fixed so I can use it now
as a replacement for MFS?
The last month has been pretty rough on me, so bear with me, please.
Okay. That's a very reasonable response.
Can you resend the email please ?
I'll just retry it now. Thank you for the above- I totally understand.
My previous mail was just a "check-in" on it- not a demand.
-matt
To
Just don't use the skipspin stuff, it shouldn't hurt at all. The new witness
code will hopefully be in by the end of the week. *crosses fingers*
Cool. WITNESS_SKIPSPIN was quite useful for NETGRAPH users because of some
unregistered spin mutexes there. Julian fixed the problem already, so
subscribe
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
It doesn't panic, yet. Good! Much improved!
All is good, ahem, except that when I ran this tmp filesystem out of space, it
takes a while before the space comes back if you remove files :-)
farrago.feral.com lmdd of=/tmp/file
/tmp: write failed, file system is full
117.19 MB in 9.07
That looks like regular softupdates behaviour to me ?
In message [EMAIL PROTECTED], Matthew
Jacob writes:
It doesn't panic, yet. Good! Much improved!
All is good, ahem, except that when I ran this tmp filesystem out of space, it
takes a while before the space comes back if you remove files
That looks like regular softupdates behaviour to me ?
Huh. I didn't know that there was this big of a delay.
At any rate, thanks! I'll do some more testing and see if I can break it but I
sure am happier to have something back!
To Unsubscribe: send mail to [EMAIL PROTECTED]
with
Hi,
dmesg and the output of "ident /sys/dev/aic7xxx/*" and pciconf is
attached (BTW: the -v options to pciconf isn't documented in the
synopsis section of the man page).
Do you need more information, e.g. the output of a verbose boot?
Bye,
Alexander.
I wish I had a system that exhibited this
:Matt,
:
:At it is very obvious to me that you have not even looked at the new
:code, let alone run it, I suggest that you do both before further
:engaging in this conversation.
:
:M
:--
:Mark Murray
:Warning: this .sig is umop ap!sdn
I looked at it. I'm sorry, I don't see how your
Hi all,
Does anybody knows to say when the computer changes
to promiscous mode?
/kernel: promiscuous mode enable
[ ]´s
Ronan Lucio
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
I don't understand what is so difficult about simply rate-limiting
the code at the proper point -- at the very beginning of the
call that the interrupt harvester makes, removing most of the fixed
overhead for the case where a system is getting a large number of
ticso@cicely5:/tmp# pkg_add -f -v png-1.0.9_1.tgz
Requested space: 831096 bytes, free space: 13705216 bytes in /var/tmp/instmp.ijZaPY
extract: Package name is png-1.0.9_1
extract: CWD to /usr/local
pkg_add: extract_plist: unable to cwd to '/usr/local'
Exit 2
The reason is that /usr/local is a
Mikhail Teterin [EMAIL PROTECTED] writes:
If you can, please reproduce the panic on a kernel compiled with the
INVARIANTS, INVARIANT_SUPPORT and WITNESS options.
Well, with this options on, the machine does not crash, but the
program segfaults on startup:
The trace you're showing looks
: down and will work, SNAP, just like that?
:
:Because I need to make folks other than you happy.
:
:Lots of security minded people what _all_ the interrupt entropy
:they can get, and this method gives them that while allowing others
:to throttle the harvester back.
:
:M
:--
:Mark Murray
"Alexander N. Kabaev" [EMAIL PROTECTED] writes:
Do you have WITNESS_SKIPSPIN option in your kernel config?
Yes.
Here is what supposedly causing the trouble:
You're telling the guy who did the analysis.
DES
--
Dag-Erling Smorgrav - [EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL
Your ethernet card went into promiscuous mode presumably. Did you run
tcpdump?
Tom Veldhouse
[EMAIL PROTECTED]
- Original Message -
From: "Ronan Lucio" [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Monday, March 12, 2001 3:25 PM
Subject: promiscuous mode
Hi all,
Does anybody knows
On 12 Mar, Dag-Erling Smorgrav wrote:
= Mikhail Teterin [EMAIL PROTECTED] writes:
= If you can, please reproduce the panic on a kernel compiled with the
= INVARIANTS, INVARIANT_SUPPORT and WITNESS options.
= Well, with this options on, the machine does not crash, but the
= program segfaults
The domain austin.rr.com is running some software that seems to come from
north of San Francisco that appears to be repackaging up and re-forwarding old
mail. Twice now I've responded to what are apparently legitimate (until
I look at full headers) resends of old mail.
I've asked our
:This effectively happens.
:
:The harvest ring is a limited length, and any overflows are discarded.
:
:M
:--
:Mark Murray
:Warning: this .sig is umop ap!sdn
Are you resending this mail from 5 dats ago or is there a bounce
occuring somewhere on the list?
Currently the queue size
On Tue, Mar 13, 2001 at 05:32:36AM +1100, Bruce Evans wrote:
k6-2 is already over-engineered. The only difference between it and k6
is 3dnow, but neither gcc nor any source files support 3dnow (now :-).
3dnow support exists in several ports, though.
OTOH, k6-3 doesn't add any new features,
On Mon, 12 Mar 2001, Kris Kennaway wrote:
On Tue, Mar 13, 2001 at 05:32:36AM +1100, Bruce Evans wrote:
OTOH, bsd.cpu.mk is too under-engineered to support any compiler except
gcc. It unconditionally translates FreeBSD-specific names like k6-2 to
gcc-specific flags like -march=k6.
I'm
On 12-Mar-01 Dag-Erling Smorgrav wrote:
Mikhail Teterin [EMAIL PROTECTED] writes:
If you can, please reproduce the panic on a kernel compiled with the
INVARIANTS, INVARIANT_SUPPORT and WITNESS options.
Well, with this options on, the machine does not crash, but the
program segfaults on
Both of the patches below fix the problem mentioned in PR bin/25110. The
first one fixes it inside of kern_fork.c and would appear to apply the
corrective behaviour regardless of whether the process uses libc_r or not.
The second patch fixes the problem inside of uthread_fork.c. Whether the
I committed a miibus'ified fxp driver to the tree today, and made
it the default. If you compile fxp into your kernel statically,
you will also need "device miibus" as well, if it isn't there already.
If you notice any problems with the driver (things that were working
and are not working now),
complete fresh build, etc
da0: invalid primary partition table: no magic
start_init: trying /sbin/init
fatal kernel trap:
trap entry = 0x4 (unaligned access fault)
a0 = 0xc3615fe1a88f382
a1 = 0x29
a2 = 0x1b
pc = 0xfc467578
ra
Just today I started using DEVFS again after a long time, and it works
perfectly. From the scarce info you provide, our only apparent difference
is I don't have SCSI.
If it weren't you, I'd ask if you are sure you have the very latest
sources, but of course I wonder you have more than enough
On Tue, 13 Mar 2001, Andrea Campi wrote:
Just today I started using DEVFS again after a long time, and it works
perfectly. From the scarce info you provide, our only apparent difference
is I don't have SCSI.
If it weren't you, I'd ask if you are sure you have the very latest
sources, but
This server runs only ssh, bind and natd, but I´m certain in my
case the problem is with bind and it have cause me a problem,
the machine delay a lot to resolve a DNS when it change to
promiscous mode, but I don´t why it is change to.
Ronan Lucio
Many programs put the ethernet card into
"David O'Brien" [EMAIL PROTECTED] writes:
On Sat, Mar 10, 2001 at 10:32:23PM -0800, Dima Dorfman wrote:
Peter Wemm [EMAIL PROTECTED] writes:
The version of the patch for -current uses the softdep mount option only.
If you remove the mount option, you dont get softupdates.
In this
"David O'Brien" [EMAIL PROTECTED] writes:
As long as someone that is familiar with all the "cool" and more esoteric
uses of `md' was consulted to ensure the framework is sufficiently
capable.
If everyone (well, I guess mostly everyone) can agree on a suitable
format for this md.conf, I'll
On Mon, Mar 12, 2001 at 05:12:13PM -0800, Dima Dorfman wrote:
There's always the 'nosoftdep' mount option. It's also possible to
enable it by default on everything except the root filesystem, but
that's a [minor] POLA violation.
I fail to see what is wrong with defaulting to `off'.
--
--
"David O'Brien" [EMAIL PROTECTED] writes:
On Mon, Mar 12, 2001 at 05:12:13PM -0800, Dima Dorfman wrote:
There's always the 'nosoftdep' mount option. It's also possible to
enable it by default on everything except the root filesystem, but
that's a [minor] POLA violation.
I fail to see
On Mon, 12 Mar 2001, Mark Murray wrote:
Lots of security minded people what _all_ the interrupt entropy
they can get, and this method gives them that while allowing others
to throttle the harvester back.
Lots of -CURRENT users want to be able to use their systems to write code
without
In message [EMAIL PROTECTED], Jean Louis Ntakpe writes:
Hi,
In /usr/src/etc/Makefile:
"make distribution" is still trying to copy MAKEDEV to /dev
on a system with devfs mounted to /dev.
Since devfs is default, is this behaviour correct or my
/etc/make.conf is missing something ?
I
I mostly agree with this, but let's also remember that this is -CURRENT too
and that Kris Mark others have been pretty good about feeling sorry for
you when you get hung up... ahem- responding to and fixing issues :-)
On Mon, 12 Mar 2001, Matthew N. Dodd wrote:
On Mon, 12 Mar 2001,
Kris Kennaway [EMAIL PROTECTED] writes:
Can everyone please test this patch against OpenSSL, which should fix
the problems observed with -j builds as well as cleaning out the
temporary ASM files.
I can confirm that this fixes -j-enabled builds on a -current host a
few days old. Thanks!
On Mon, Mar 12, 2001 at 07:55:01PM -0800, Dima Dorfman wrote:
Kris Kennaway [EMAIL PROTECTED] writes:
Can everyone please test this patch against OpenSSL, which should fix
the problems observed with -j builds as well as cleaning out the
temporary ASM files.
I can confirm that this fixes
I have 100% reproducable trap 12 panic in kernel. I thouhgt it appeared
somewhere after 5th of february, but i was wrong. The problem is that when
i try to compile something with "make -j 2" - it panics and i can't
backtrace the first fault point. :( If i simply use make, i have a chanse
to
Let me be clear about what I mean by interrupt rate limiting:
interrupt()
{
harvester(...)
}
It does that already.
harvester(...)
{
if (queue is not full) {
... add data to queue (reasonably sized queue, like 32 entries)
}
}
It
On my system (dual PII/400 running -current), I've noticed for some time that
if I build a kernel with too many device drivers in it (where "too many" seems
to correspond to text size 3M for the resulting kernel), the system reboots
itself immediately upon booting with the new kernel. Other
kris We really need to provide a better rc.conf hook for doing this --
kris expecting people to write their own script just to create a /tmp
kris is lame.
Here is just an example script (CAUTION: not well-tested) to create a
filesystem/swap with mdconfig(8). This script support not only '/tmp
78 matches
Mail list logo