You said:
"Perhaps we've finally finished ;)"
First, I would like to thank you for your excellent, _EXCELLENT WORK_, Alan!!!
Under your tender-loving care, Linux 2.0.36 has matured, and become one of the
_MOST_ stable OS ever!
There maybe, some "wrinkles" left however, as I have counted 4 so far.
IMHO, if them "wrinkles" got ironed out as well, Linux 2.0.36 would be better
than ever.
Summary of the 4 "wrinkles":
Wrinkle #1: Swap Bug.
See attached message A below for more info.
Wrinkle #2: "Free Page" Problem.
See attached message B below for more info.
Wrinkle #3: The PS/2 Mouse Bug.
See attached message C below for more info.
Wrinkle #4: "Kernel Instability" Problem.
See attached message D below for more info.
*Please note*
The "wrinkles" may have, or haven't been resolved.
They may or may not be hardware related.
If they have been resolved, please accept my apology.
Thank you again, Mr. Cox, for the wonderful job you've done!!!!
Most respectfully,
Pete
teamwork@freemail.c3.hu
------------------------------
Attached: Message A
------------------------------
From: Hubert Tonneau <tonneau.heliosam@hol.fr>
Date: Fri, 30 Oct 1998 12:08:09 +0000
Subject: Swap bug in 2.0.35 / 2.0.36pre15
The following program, witch consumes 128 Mb of memory,
will exhaust the 192 Mb (64 Mb physical + 128 Mb swap)
of a box (SMP or not) running a 2.0.35 or 2.0.36-pre15 kernel,
after less than an hour.
As a result, typing 'ls' in another terminal will core dump !
Please notice that everything works fine with the 2.1 kernels.
#include <stdio.h>
#include <stdlib.h>
#define size 128*1024*1024
int r(int range) {
static int seed=8353898;
seed=seed*79353547^38753853;
return (seed>=0 ? seed : -seed)%range; }
main() {
char *mem; char c; int i;
mem=malloc(size);
printf("init\n");
for(i=0; i<size; i++)
mem[i]=r(255);
printf("shake\n");
while(1)
c=mem[r(size)]; }
------------------------------
Attached: Message B
------------------------------
From: =3D?iso-8859-1?Q?Lars_Marowsky-Br=3DE9e?=3D <lmb@pointer.teuto.de=
>
Date: Mon, 9 Nov 1998 16:48:53 +0100
Subject: 2.0.36pre15 dies with "Couldn't get a free page"
Good Morning,
the latest 2.0.36pre kernels have been dieing on me on one machine with
the above error message.
It is only this machine, works fine on all others. I am trying to track=
down why. We already swapped the hardware once.
It is running mysql (threads?), spong (a network monitoring tool heavily
modified by us - opens lots of outgoing connections to check services etc),
NeMaC (gathers network accounting data via SNMP from lots of hosts). Otherwise
it is Red Hat 5.1.
Standard kernel - syncookies etc, just what you need on a server.
Hardwarewise it is a PII 128MB, ethernet card is a
tulip.c:v0.89H 5/23/98 becker@cesdis.gsfc.nasa.gov
eth0: Digital DS21140 Tulip at 0x6800, 00 40 05 36 0f a7, IRQ 9.
SCSI storage controller: Adaptec AIC-7880U (rev 0).
Any idea what might be going on?
Mit freundlichen Gr=FC=DFen,
Lars Marowsky-Br=E9e
Lars Marowsky-Br=E9e
Netzwerk Management
teuto.net Netzdienste GmbH - DPN Verbund-Partner
------------------------------
Attached: Message C
------------------------------
From: Jonathan H N Chin <jc254@newton.cam.ac.uk>
Date: Thu, 12 Nov 1998 11:06:44 GMT
Subject: Re: various problems with 2.0.36pre (SCSI, floppy)
Alan = > ; Tomasz = >>
>> and floppy. Another strange thing about this machine (P200 MMX, Award 4.51PG
>> BIOS, 430HX chipset) is that it hangs every time PS/2 mouse is moved before
>> its driver is loaded. Is it hardware problem, or should I upgrade the BIOS?
>> Win95 do not have any problems though.
>Very strange indeed. I really dont know.
Peter Benie discovered this problem this time last year.
Not sure how/if he fixed it. He wrote:
|> What the also have in common is a PS/2 mouse, but during the
|> installation, I'm using a kernel without a PS/2 mouse driver.
|>
|> The keyboard driver is roughly:
|> On IRQ1:
|> {
|> while (data is waiting on IO port and data is not mouse data)
|> {
|> read data from kbd controller
|> process data
|> }
|> }
|>
|> The mouse driver is roughly:
|> On IRQ12:
|> {
|> while (data is waiting on IO port)
|> {
|> read data from kbd controller
|> process data
|> }
|> }
|>
|> If you move the mouse, the keyboard controller gets some mouse data
|> which the keyboard driver refuses to read. Without the mouse driver
|> installed, no more data can be read.
|>
|> The result is a keyboard lockup, and since the machine is being
|> installed, it appears that it has completely crashed.
- -jonathan
- --
Jonathan H N Chin, 1 kyu | deputy computer | Newton Institute, Cambridge, UK
<jc254@newton.cam.ac.uk> | systems mangler | tel/fax: +44 1223 335986/330508
"respondeo etsi mutabor" --Rosenstock-Huessy
------------------------------
Attached: Message D
------------------------------
From: Hugo Embrechts <Hugo.Embrechts@KULNET.kuleuven.ac.be>
Date: 12 Nov 1998 16:16:10 -0000
Subject: kernel instability on a large volume NAT
(Oops, missing subject in previous post.)
Dear fellows,
Since a few years we use a linux-NAT to provide internet access for a
university campus network (average load of 1.5 Mbit/s). This runs
reasonably well except for regular kernel crashes, that leave the
machine dead.
We have had the problem with several kernels (2.0.24, 2.0.30, 2.0.33,
2.0.34 and now 2.0.36pre19). An indication of the problem may be the
warnings which are listed below. It problem is related to load. Now it
crashes every few days, whereas during holiday periods when the load
is about half as high the machine can stay up for more than a month.
The machine is a few years old pentium. We tried network cards from
3com (905) and from dlink (DFE-500TX).
We use a rather strict IP-filter and log denied trafic.
Thanks for any help,
Hugo
Hugo Embrechts
- ------- KULeuvenNet -------------------------------------------------------
Ludit, W. de Croylaan 52A, B-3001 Heverlee
Hugo.Embrechts@kulnet.kuleuven.ac.be
016 / 32 27 85 (Tel.), 32 29 99 (Fax.)
Kernel messages
- ---------------
(Linux ds9 2.0.36 #2 Tue Nov 10 11:16:50 CET 1998 i586 unknown)
Patch --> 2.0.36pre19
I.
Nov 10 13:42:15 ds9 kernel: Warning: kfree_skb passed an skb still on a list
(from 00ffcf3c).
Nov 10 13:42:15 ds9 kernel: invalid operand: 0000
Nov 10 13:42:15 ds9 kernel: CPU: 0
Nov 10 13:42:15 ds9 kernel: EIP: 0010:[<0008df85>]
Nov 10 13:42:15 ds9 kernel: EFLAGS: 00010293
Nov 10 13:42:15 ds9 kernel: eax: 0008dcf5 ebx: 00ffcce4 ecx: 00ffcf3c edx:
03c62414
Nov 10 13:42:15 ds9 kernel: esi: 00ffcefc edi: 00000000 ebp: 001ff0d4 esp:
001ff06c
Nov 10 13:42:15 ds9 kernel: ds: 0018 es: 0018 fs: 002b gs: 0018 ss: 0018
Nov 10 13:42:15 ds9 kernel: Process swapper (pid: 0, process nr: 0,
stackpage=001fd1bc)
Nov 10 13:42:15 ds9 kernel: Stack: 00139a83 00ffcce4 00ffcce4 00ffcf3c 00139d83
00ffcce4 00ffcf3c 00ffcf3c
Nov 10 13:42:15 ds9 kernel: 00ffcefc 0000003d 001422de 00ffcf3c 00000000
00000000 00ffcf3c 002053b0
Nov 10 13:42:15 ds9 kernel: 00000000 0022be70 00000000 00000000 00000000
00ffcd00 ffffff00 42a9d8cc
Nov 10 13:42:15 ds9 kernel: Call Trace: [sock_wfree+35/44] [kfree_skb+183/244]
[ip_rcv+1474/1492] [vremap+72/128] [net_bh+252/284] [do_bottom_half+59/96]
[handle_bottom_half+11/24]
Nov 10 13:42:15 ds9 kernel: [sys_idle+92/112] [system_call+85/124]
[init+0/612] [start_kernel+429/440] [it_real_fn+0/72] [schedule+564/652]
Nov 10 13:42:15 ds9 kernel: Code: fe ff 83 c4 08 c3 90 53 8b 44 24 08 8b 98 3c
09 00 00 66 8b
Nov 10 13:42:15 ds9 kernel: Aiee, killing interrupt handler
Nov 10 13:42:15 ds9 kernel: kfree of non-kmalloced memory: 001ff204, next=
00000018, order=1951796
Nov 10 13:42:15 ds9 kernel: kfree of non-kmalloced memory: 001ff1f4, next=
00000018, order=1951796
Nov 10 13:42:15 ds9 kernel: kfree of non-kmalloced memory: 001ff708, next=
00000018, order=1951796
Nov 10 13:42:15 ds9 kernel: idle task may not sleep
Nov 10 13:42:15 ds9 last message repeated 4 times
II.
Nov 11 16:05:08 ds9 kernel: Warning: kfree_skb passed an skb still on a list
(from 00ffcf3c).
Nov 11 16:05:08 ds9 kernel: Oops: 0002
Nov 11 16:05:08 ds9 kernel: CPU: 0
Nov 11 16:05:08 ds9 kernel: EIP: 0010:[<0008dc40>]
Nov 11 16:05:08 ds9 kernel: EFLAGS: 00010a87
Nov 11 16:05:08 ds9 kernel: eax: 0008dc01 ebx: 00ffcce4 ecx: 00ffcf3c edx:
03c62414
Nov 11 16:05:08 ds9 kernel: esi: 00ffcefc edi: 00000000 ebp: 001fe40c esp:
001fe3a4
Nov 11 16:05:08 ds9 kernel: ds: 0018 es: 0018 fs: 002b gs: 0000 ss: 0018
Nov 11 16:05:08 ds9 kernel: Process swapper (pid: 0, process nr: 0,
stackpage=001fd1bc)
Nov 11 16:05:08 ds9 kernel: Stack: 00139a83 00ffcce4 00ffcce4 00ffcf3c 00139d83
00ffcce4 00ffcf3c 00ffcf3c
Nov 11 16:05:08 ds9 kernel: 00ffcefc 0000003e 001422de 00ffcf3c 00000000
00000000 00ffcf3c 002053b0
Nov 11 16:05:08 ds9 kernel: 0000fc80 00ffcce4 00000000 00000000 00000000
00ffcd00 ffffff00 42a9d8cc
Nov 11 16:05:08 ds9 kernel: Call Trace: [sock_wfree+35/44] [kfree_skb+183/244]
[ip_rcv+1474/1492] [vremap+72/128] [net_bh+252/284] [do_bottom_half+59/96]
[handle_bottom_half+11/24]
Nov 11 16:05:08 ds9 kernel: [exit_notify+62/472] [do_exit+456/508]
[ip_masq_raudio+75497472/1571520512] [die_if_kernel+695/704]
[ip_masq_raudio+83886080/1571520512] [ip_masq_raudio+75497472/1571520512]
[BusLogic_ComputeResultCode+84/336] [do_page_fault+718/736]
Nov 11 16:05:08 ds9 kernel: [do_page_fault+0/736] [error_code+64/72]
[sock_wfree+35/44] [kfree_skb+183/244] [ip_rcv+1474/1492]
[ip_masq_raudio+76020864/1571520512] [vremap+72/128] [net_bh+252/284]
Nov 11 16:05:08 ds9 kernel: [do_bottom_half+59/96]
[handle_bottom_half+11/24] [exit_notify+62/472] [do_exit+456/508]
[ip_masq_raudio+75497472/1571520512] [die_if_kernel+695/704]
[ip_masq_raudio+83886080/1571520512] [ip_masq_raudio+75497472/1571520512]
Nov 11 16:05:08 ds9 kernel: [tty_set_ldisc+200/448] [do_invalid_op+61/68]
[do_invalid_op+0/68] [error_code+64/72] [sock_wfree+35/44] [kfree_skb+183/244]
[ip_rcv+1474/1492] [net_bh+252/284]
Nov 11 16:05:08 ds9 kernel: [vremap+72/128] [net_bh+252/284]
[do_bottom_half+59/96] [handle_bottom_half+11/24] [exit_notify+59/472]
[do_exit+456/508] [ip_masq_raudio+75497472/1571520512] [die_if_kernel+695/704]
Nov 11 16:05:08 ds9 kernel: [ip_masq_raudio+83886080/1571520512]
[ip_masq_raudio+75497472/1571520512] [tty_set_ldisc+200/448]
[do_invalid_op+61/68] [do_invalid_op+0/68] [error_code+64/72] [sock_wfree+35/44]
[kfree_skb+183/244]
Nov 11 16:05:08 ds9 kernel: [ip_rcv+1474/1492] [net_bh+252/284]
[vremap+72/128] [net_bh+252/284] [do_bottom_half+59/96]
[handle_bottom_half+11/24] [exit_notify+59/472] [do_exit+456/508]
Nov 11 16:05:08 ds9 kernel: [ip_masq_raudio+75497472/1571520512]
[die_if_kernel+695/704] [ip_masq_raudio+83886080/1571520512]
[ip_masq_raudio+75497472/1571520512] [do_general_protection+124/172]
[do_general_protection+0/172] [error_code+64/72] [sock_wfree+35/44]
Nov 11 16:05:08 ds9 kernel: [ip_rcv+1474/1492] [net_bh+252/284]
[do_bottom_half+59/96] [handle_bottom_half+11/24] [exit_notify+62/472]
[do_exit+456/508] [ip_masq_raudio+75497472/1571520512] [die_if_kernel+695/704]
Nov 11 16:05:08 ds9 kernel: [ip_masq_raudio+83886080/1571520512]
[ip_masq_raudio+75497472/1571520512] [do_invalid_op+61/68] [do_invalid_op+0/68]
[error_code+64/72] [sock_wfree+35/44] [kfree_skb+183/244] [ip_rcv+1474/1492]
Nov 11 16:05:08 ds9 kernel: [arp_bind_cache+540/692] [net_bh+252/284]
[do_bottom_half+59/96] [handle_bottom_half+11/24] [acct_process+434/440]
[schedule+21/652] [exit_notify+219/472] [do_exit+46/508]
Nov 11 16:05:08 ds9 kernel: [ip_masq_raudio+75497472/1571520512]
[die_if_kernel+695/704] [ip_masq_raudio+83886080/1571520512]
[ip_masq_raudio+75497472/1571520512] [tty_set_ldisc+200/448]
[do_invalid_op+61/68] [do_invalid_op+0/68] [error_code+64/72]
Nov 11 16:05:08 ds9 kernel: [sock_wfree+35/44] [kfree_skb+183/244]
[ip_rcv+1474/1492] [net_bh+252/284] [do_bottom_half+59/96]
[handle_bottom_half+11/24] [notify_parent+30/48] [exit_notify+219/472]
Nov 11 16:05:08 ds9 kernel: [do_exit+456/508]
[ip_masq_raudio+75497472/1571520512] [die_if_kernel+695/704]
[ip_masq_raudio+83886080/1571520512] [ip_masq_raudio+75497472/1571520512]
[tty_set_ldisc+200/448] [do_invalid_op+61/68] [do_invalid_op+0/68]
Nov 11 16:05:08 ds9 kernel: [error_code+64/72] [sock_wfree+35/44]
[kfree_skb+183/244] [ip_rcv+1474/1492] [net_bh+252/284] [do_bottom_half+59/96]
[handle_bottom_half+11/24] [exit_notify+62/472]
Nov 11 16:05:08 ds9 kernel: [do_exit+456/508] [die_if_kernel+695/704]
[ip_masq_raudio+83886080/1571520512] [ip_masq_raudio+75497472/1571520512]
Nov 11 16:05:08 ds9 kernel: Code: 00 2c 9c 02 00 04 00 00 00 38 9c 02 00 04 00
00 00 44 9c 02
Nov 11 16:05:08 ds9 kernel: Aiee, killing interrupt handler
Kernel configuration
- --------------------
#
# Loadable module support
#
CONFIG_MODULES=y
CONFIG_KERNELD=y
#
# General setup
#
CONFIG_NET=y
CONFIG_PCI=y
CONFIG_SYSVIPC=y
CONFIG_BINFMT_AOUT=y
CONFIG_BINFMT_ELF=y
CONFIG_KERNEL_ELF=y
CONFIG_M586=y
#
# Floppy, IDE, and other block devices
#
CONFIG_BLK_DEV_FD=y
CONFIG_BLK_DEV_IDE=y
#
# Please see Documentation/ide.txt for help/info on IDE drives
#
CONFIG_BLK_DEV_IDECD=y
CONFIG_BLK_DEV_IDEFLOPPY=y
CONFIG_BLK_DEV_CMD640=y
CONFIG_BLK_DEV_RZ1000=y
CONFIG_BLK_DEV_TRITON=y
#
# Additional Block Devices
#
CONFIG_BLK_DEV_LOOP=y
CONFIG_BLK_DEV_MD=y
CONFIG_MD_LINEAR=m
CONFIG_MD_STRIPED=y
#
# Networking options
#
CONFIG_FIREWALL=y
CONFIG_NET_ALIAS=y
CONFIG_INET=y
CONFIG_IP_FORWARD=y
CONFIG_IP_MULTICAST=y
CONFIG_SYN_COOKIES=y
CONFIG_IP_FIREWALL=y
CONFIG_IP_FIREWALL_VERBOSE=y
CONFIG_IP_MASQUERADE=y
#
# Protocol-specific masquerading support will be built as modules.
#
CONFIG_IP_MASQUERADE_ICMP=y
CONFIG_IP_ALWAYS_DEFRAG=y
CONFIG_IP_ACCT=y
CONFIG_IP_ROUTER=y
CONFIG_IP_ALIAS=y
#
# (it is safe to leave these untouched)
#
CONFIG_IP_NOSR=y
CONFIG_SKB_LARGE=y
#
# SCSI support
#
CONFIG_SCSI=y
#
# SCSI support type (disk, tape, CD-ROM)
#
CONFIG_BLK_DEV_SD=y
CONFIG_CHR_DEV_ST=y
CONFIG_BLK_DEV_SR=y
CONFIG_CHR_DEV_SG=y
#
# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
#
CONFIG_SCSI_MULTI_LUN=y
CONFIG_SCSI_CONSTANTS=y
#
# SCSI low-level drivers
#
CONFIG_SCSI_AIC7XXX=y
CONFIG_AIC7XXX_RESET_DELAY=5
CONFIG_SCSI_BUSLOGIC=y
#
# Network device support
#
CONFIG_NETDEVICES=y
CONFIG_DUMMY=m
CONFIG_PPP=y
#
# CCP compressors for PPP are only built as modules.
#
CONFIG_NET_ETHERNET=y
CONFIG_NET_VENDOR_3COM=y
CONFIG_EL3=m
CONFIG_VORTEX=y
CONFIG_NET_PCI=y
CONFIG_EEXPRESS_PRO100B=m
CONFIG_DEC_ELCP=m
CONFIG_NET_ISA=y
CONFIG_NE2000=y
#
# Filesystems
#
# CONFIG_QUOTA is not set
CONFIG_MINIX_FS=y
CONFIG_EXT2_FS=y
CONFIG_NLS=y
CONFIG_ISO9660_FS=y
CONFIG_FAT_FS=y
CONFIG_MSDOS_FS=y
CONFIG_NLS_CODEPAGE_437=y
CONFIG_NLS_CODEPAGE_850=y
CONFIG_NLS_ISO8859_1=y
CONFIG_PROC_FS=y
CONFIG_NFS_FS=y
CONFIG_SMB_FS=m
#
# Character devices
#
CONFIG_SERIAL=y
CONFIG_PRINTER=m
CONFIG_MOUSE=y
CONFIG_PSMOUSE=m
CONFIG_WATCHDOG=y
CONFIG_SOFT_WATCHDOG=y
------------------------------
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/