Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 11 Sep 1999 13:19:33 +0000
From:      greg <greg@jules.res.cmu.edu>
To:        freebsd-smp@FreeBSD.ORG
Subject:   repeated deadlocks in FS with SMP machine...
Message-ID:  <199909111319.NAA00652@jules.res.cmu.edu>

next in thread | raw e-mail | index | archive | help
Hi - I'm running a 3.3-RC machine right now an a dual P2 with an IBM 
DDRS-34560D DC1B through an adaptec 2940 u2w card.

For the past 3 weeks the machine has been freezing up in a similar manner.  
When the disk is getting between very hard, I'll see processes enter the 
namecache state, or the inode state & never come back to life.  Other 
processes run fine until they need to do read or write something.

I'm not using softupdates or doing anything over NFS (though the NFS code is 
enabled in the kernel).  I've included the config file at the end...  I wasn't 
quite sure what I should set the SMP options for, so I only set the NCPU to 2, 
The motherboard is a supermicro p6dbe.

the latest happened when I was running 2 instances of du - one of them within 
a subdirectory of the other.  Both went into the namecache state at the same 
time & just stayed there.  I went to ddb & panicked, so I have a whole bunch 
of data now... I'm just not sure about where to begin looking.  The problem is 
fairly repeatable for me - I just need to really beat the fs...

I did run2 du's in parallel (which caused deadlock in early versions of 3.0, 
3.1 because of a hold & wait (or a race condition, i can't recall)) but this 
didn't cause any problem.  Of course, I only ran it once, so I could be 
wrong...

Has anybody else experienced problems like this?  I've seen other reports of 
problems with NFS & softupdates, but none this simple...

Does anybody know where I should begin digging?  Unfortunately, I'm not to 
familiar with the kernel source area (though I am familiar with kernels...).  
Also, should I enable the INVARIANTS option?  It already takes maybe 10 
minutes before I can get a lockup anyway & I don't even know if anything would 
pop up (though I guess I can only try...).

thanks alot...
Greg


machine		"i386"

ident		JULES
maxusers	32

makeoptions	DEBUG="-g"

options		"MAXDSIZ=(256*1024*1024)"
options		"DFLDSIZ=(256*1024*1024)"
#options		PQ_LARGECACHE		# color for 512k/16k cache

options		FAILSAFE
options         INCLUDE_CONFIG_FILE     # Include this file in kernel

config		kernel	root on da0 dumps on da0

options		SMP			# Symmetric MultiProcessor Kernel
options		APIC_IO			# Symmetric (APIC) I/O
options		NCPU=2

cpu		"I686_CPU"

options		USER_LDT
options		"COMPAT_43"

options		SYSVSHM
options		SYSVSEM
options		SYSVMSG

options		"MD5"

options		"VM86"

options		DDB

options		KTRACE			#kernel tracing

#options		INVARIANTS
#options		INVARIANT_SUPPORT
options		DIAGNOSTIC

options		PERFMON

options		UCONSOLE
options		INET			#Internet communications protocols

pseudo-device	ether			#Generic Ethernet
pseudo-device	loop			#Network loopback device
pseudo-device	bpfilter 4		#Berkeley packet filter
pseudo-device	disc			#Discard device
pseudo-device	tun	1		#Tunnel driver (user process ppp(8))
pseudo-device	ppp	2		#Point-to-point protocol
options PPP_BSDCOMP			#PPP BSD-compress support
options PPP_DEFLATE			#PPP zlib/deflate/gzip support
options PPP_FILTER			#enable bpf filtering (needs bpfilter)

options         "ICMP_BANDLIM"

options		FFS			#Fast filesystem
options		MFS			#Memory File System
options		NFS			#Network File System
options		"CD9660"		#ISO 9660 filesystem
options		FDESC			#File descriptor filesystem
options		KERNFS			#Kernel filesystem
options		MSDOSFS			#MS DOS File System
options		FFS_ROOT		#FFS usable as root device
options		DEVFS			#devices filesystem

#options		SOFTUPDATES

options		NSWAPDEV=5


options		CODA			#CODA filesystem.
pseudo-device	vcoda	4		#coda minicache <-> venus comm.

options		"P1003_1B"
options		"_KPOSIX_PRIORITY_SCHEDULING"
options		"_KPOSIX_VERSION=199309L"

controller	scbus0	#base SCSI code
device		da0	#SCSI direct access devices (aka disks)
device		cd0	#SCSI CD-ROMs

pseudo-device	pty	64	#Pseudo ttys - can go as high as 256
pseudo-device	speaker		#Play IBM BASIC-style noises out your speaker
pseudo-device	gzip		#Exec gzipped a.out's
pseudo-device	snp	3	#Snoop device - to look at pty/vty/etc..
pseudo-device	vn		#Vnode driver (turns a file into a device)
pseudo-device	ccd	4	#Concatenated disk driver

controller	isa0
options		PPS_SYNC
controller	atkbdc0	at isa? port IO_KBD tty
device		atkbd0	at isa? tty irq 1
device		psm0	at isa? tty irq 12
device		vga0	at isa? port ? conflicts
pseudo-device	splash

device		sc0	at isa? tty
options		MAXCONS=16		# number of virtual consoles
options		"STD8X16FONT"		# Compile font in
makeoptions	"STD8X16FONT"="cp850"
options		SC_HISTORY_SIZE=200	# number of history buffer lines

device		npx0	at isa? port IO_NPX iosiz 0x0 flags 0x0 irq 13

controller	wdc0	at isa? port "IO_WD1" bio irq 14
disk		wd0	at wdc0 drive 0
disk		wd1	at wdc0 drive 1
options         ATAPI   #Enable ATAPI support for IDE bus
options		ATAPI_STATIC	#Don't do it as an LKM

options		IDE_DELAY=8000	# Be optimistic about Joe IDE device

device          acd0

controller	fdc0	at isa? port "IO_FD1" bio irq 6 drq 2
disk		fd0	at fdc0 drive 0
device		sio0	at isa? port "IO_COM1" tty flags 0x10 irq 4

device ep0 at isa? port 0x300 net irq 10

device pcm0 at isa? port ? tty irq 10 drq 1 flags 0x0
device pca0 at isa? port "IO_TIMER1" tty

controller	pci0
controller	ahc1

device		pn0
options		COMPAT_LINUX




To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-smp" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199909111319.NAA00652>