Linux 2.4.19-rc3

Linux 2.4.19-rc3

[Marcelo Tosatti]

So here goes rc3. Another -rc is going to come only in the case of really
critical problem(s).

I'm attaching the rc2->rc3 changelog only because the full changelog got
too big (I guess thats why my -rc2 announce mail didnt go to lk).


Summary of changes from v2.4.19-rc2 to v2.4.19-rc3
============================================

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.622)
	[PATCH] PATCH: fix aha152x

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.623)
	[PATCH] PATCH: update atp870u driver

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.624)
	[PATCH] PATCH: new PCI idents for gart

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.625)
	[PATCH] PATCH: AMD766/768 $PIR PCI table entries

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.626)
	[PATCH] PATCH: fix unload crash in AF_ROSE

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.627)
	[PATCH] PATCH: wrong flag type

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.628)
	[PATCH] PATCH: more scsi blacklist

<bunk@fs.tum.de> (02/07/17 1.629)
	[PATCH] fix .text.exit error in orinoco_plx.c

<rusty@rustcorp.com.au> (02/07/17 1.630)
	[PATCH] The real netfilter conntrack SMP overrun fix

<geert@linux-m68k.org> (02/07/17 1.631)
	[PATCH] 2.4.19-rc2 and !CONFIG_BLK_DEV_IDEPCI

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.632)
	[PATCH] PATCH: APM freeze on resume with SMP kernel cure

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.633)
	[PATCH] PATCH: fix oops printing

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.634)
	[PATCH] PATCH: fpu fault

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.635)
	[PATCH] PATCH: HP 01.09 megaraid

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.636)
	[PATCH] PATCH: maestro hang

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.637)
	[PATCH] PATCH: (low prio) stereo reporting

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.638)
	[PATCH] PATCH: hotplug

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.639)
	[PATCH] PATCH: Dunord PCI decode workaround

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.640)
	[PATCH] PATCH: (low priority) bose speakers

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.641)
	[PATCH] PATCH: fix make kernel-doc

<alan@lxorguk.ukuu.org.uk> (02/07/17 1.642)
	[PATCH] PATCH: personality clashes

<hugh@veritas.com> (02/07/17 1.643)
	[PATCH] tmpfs double kunmap

<alan@lxorguk.ukuu.org.uk> (02/07/19 1.644)
	[PATCH] Add missing PCI ID

<marcelo@plucky.distro.conectiva> (02/07/19 1.645)
	Add missing ALI chipset enum in AGP

<marcelo@plucky.distro.conectiva> (02/07/19 1.646)
	Fix wrong #ifdef in ide-pci.c: Was causing problems with FastTrak

<marcelo@plucky.distro.conectiva> (02/07/19 1.647)
	Changed EXTRAVERSION to -rc3

still having smp/snat problems (Re: Linux 2.4.19-rc3)

[glynis]

[-- Attachment #1: Type: text/plain, Size: 2144 bytes --]

On Fri, Jul 19, 2002 at 09:20:49PM -0300, Marcelo Tosatti wrote:
> <rusty@rustcorp.com.au> (02/07/17 1.630)
> 	[PATCH] The real netfilter conntrack SMP overrun fix

assuming this should have fixed this:
LIST_DELETE: ip_conntrack_core.c:165
`&ct->tuplehash[IP_CT_DIR_REPLY]'(dd8f2e90) not in &ip_conntrack_hash
[hash_conntrack(&ct->tuplehash[IP_CT_DIR_REPLY].tuple)].

it doesn't seem to have taken.   i see this error and the following
crash when i use these rules:
iptables -t nat -A POSTROUTING -j SNAT --to 64.192.31.41
iptables -t nat -A POSTROUTING -j SNAT --to 192.168.1.1
iptables -t nat -A POSTROUTING -j SNAT --to 127.0.0.1

when i use the masquerading instead of snat, it doesn't exhibit the
error and crash.

here's the info for my _stable_ config on my dual athlon 1800+:
---
If some fields are empty or look unusual you may have an old version.
Compare to the current minimal requirements in Documentation/Changes.
 
Linux butterfly 2.4.19-rc3 #1 SMP Fri Jul 19 21:30:52 EDT 2002 i686 unknown unknown GNU/Linux
 
Gnu C                  gcc (GCC) 3.1.1 20020703 (Debian prerelease) Copyright (C) 2002 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Gnu make               3.79.1
util-linux             2.11n
mount                  2.11n
modutils               2.4.15
e2fsprogs              1.27
Linux C Library        2.2.5
Dynamic linker (ldd)   2.2.5
Procps                 2.0.7
Net-tools              1.60
Console-tools          0.2.3
Sh-utils               2.0.12
Modules Loaded         snd-mixer-oss snd-pcm-oss snd-ens1371 snd-rawmidi snd-seq-device snd-ac97-codec snd-pcm snd-timer snd soundcore ipt_REJECT iptable_filter ipt_MASQUERADE iptable_nat ip_conntrack ip_tables autofs4 cpia_usb cpia videodev serial af_packet usb-ohci usbcore 3c59x ne2k-pci 8390
---
-- 
____________________}John Flinchbaugh{______________________
| glynis@hjsoft.com         http://www.hjsoft.com/~glynis/ |
~~Powered by Linux: Reboots are for hardware upgrades only~~

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: still having smp/snat problems (Re: Linux 2.4.19-rc3)

[Martin Josefsson]

[-- Attachment #1: Type: text/plain, Size: 1469 bytes --]

On Sat, 2002-07-20 at 06:17, glynis@butterfly.hjsoft.com wrote:
> On Fri, Jul 19, 2002 at 09:20:49PM -0300, Marcelo Tosatti wrote:
> > <rusty@rustcorp.com.au> (02/07/17 1.630)
> > 	[PATCH] The real netfilter conntrack SMP overrun fix
> 
> assuming this should have fixed this:
> LIST_DELETE: ip_conntrack_core.c:165
> `&ct->tuplehash[IP_CT_DIR_REPLY]'(dd8f2e90) not in &ip_conntrack_hash
> [hash_conntrack(&ct->tuplehash[IP_CT_DIR_REPLY].tuple)].
> 
> it doesn't seem to have taken.   i see this error and the following
> crash when i use these rules:
> iptables -t nat -A POSTROUTING -j SNAT --to 64.192.31.41
> iptables -t nat -A POSTROUTING -j SNAT --to 192.168.1.1
> iptables -t nat -A POSTROUTING -j SNAT --to 127.0.0.1
> 
> when i use the masquerading instead of snat, it doesn't exhibit the
> error and crash.
> 
> here's the info for my _stable_ config on my dual athlon 1800+:

Hmm, this is something people have reported a long time but we've never
been able to reproduce it. People have even reported it on UP machines
(maybe they were running with preempt?).
Rusty has been over the locking in conntrack a few times without finding
anything that could explain it.

Could you please apply the attached patch and give it a go and send the
results to netfilter-devel@lists.netfilter.org ?
(I hope it applies, was diffed against my quite patched kernel)

-- 
/Martin

Never argue with an idiot. They drag you down to their level, then beat
you with experience.

[-- Attachment #2: listhelp.h-patch --]
[-- Type: text/plain, Size: 590 bytes --]

--- linux/include/linux/netfilter_ipv4/listhelp.h.orig	Sat Jul 20 11:46:47 2002
+++ linux/include/linux/netfilter_ipv4/listhelp.h	Sat Jul 20 11:48:17 2002
@@ -54,9 +54,11 @@
 #define LIST_DELETE(head, oldentry)					\
 do {									\
 	ASSERT_WRITE_LOCK(head);					\
-	if (!list_inlist(head, oldentry))				\
+	if (!list_inlist(head, oldentry)) {				\
 		printk("LIST_DELETE: %s:%u `%s'(%p) not in %s.\n",	\
 		       __FILE__, __LINE__, #oldentry, oldentry, #head);	\
+		BUG();							\
+	}								\
         else list_del((struct list_head *)oldentry);			\
 } while(0)
 #else

Re: still having smp/snat problems (Re: Linux 2.4.19-rc3)

[Rusty Russell]

On Sat, 20 Jul 2002 00:17:01 -0400
glynis@butterfly.hjsoft.com wrote:

> On Fri, Jul 19, 2002 at 09:20:49PM -0300, Marcelo Tosatti wrote:
> > <rusty@rustcorp.com.au> (02/07/17 1.630)
> > 	[PATCH] The real netfilter conntrack SMP overrun fix
> 
> assuming this should have fixed this:
> LIST_DELETE: ip_conntrack_core.c:165
> `&ct->tuplehash[IP_CT_DIR_REPLY]'(dd8f2e90) not in &ip_conntrack_hash
> [hash_conntrack(&ct->tuplehash[IP_CT_DIR_REPLY].tuple)].

Nope, that's the mysterious half-deleted bug.

> Gnu C                  gcc (GCC) 3.1.1 20020703 (Debian prerelease) Copyright (C) 2002 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

And it's always associated with a wierd compiler version like this.

1) Does this happen reliably enough for you to decide whether something
   fixes it?
2) If so, please try dropping compiler versions.

Rusty.
-- 
   there are those who do and those who hang on and you don't see too
   many doers quoting their contemporaries.  -- Larry McVoy

Re: still having smp/snat problems (Re: Linux 2.4.19-rc3)

[John M Flinchbaugh]

[-- Attachment #1: Type: text/plain, Size: 1410 bytes --]

On Mon, Mar 10, 2003 at 11:39:21AM +0200, Dan Broscoi wrote:
> i'm writing you after reading your post on 
http://hypermail.idiosynkrasia.net/linux-kernel/archived/2002/week28/142
9.html
> I have the same problem, compiling a 2.4.20 kernel on Debian 3.0
> Do you know the fix for this ?

no one could really offer any help in the end.  i think having the
multiple snat rules matching on all interfaces may have been
triggering a race in the conntrack code.  i should probably resubmit
my findings.

anyway, limitting the rules by matching them only to their appropriate
interfaces seemed to alleviate the problem.

so now i use something similar to this:
iptables -t nat -A POSTROUTING -o eth0 -s 192.168.1.0/24 \
-j SNAT --to 10.1.1.15
iptables -t nat -A POSTROUTING -o eth1 -s 192.168.1.0/24 \
-j SNAT --to 192.168.1.1
iptables -t nat -A POSTROUTING -o eth2 -s 192.168.2.0/24 \
-j SNAT --to 192.168.2.1
---

note the -o interface rules.  this keeps it from applying more than
one snat rule to a packet.  it only applies them to the packet going
out the proper interface.

ps. i'm actually going to bounce this to the kernel list again to see
if i can stir up any more noise on this one.
-- 
____________________}John Flinchbaugh{______________________
| glynis@hjsoft.com         http://www.hjsoft.com/~glynis/ |
~~Powered by Linux: Reboots are for hardware upgrades only~~

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]