kernel crash on resume from S3 suspend

Binary package hint: linux-image-2.6-686-smp

Hi, I have a D820 latitude, which works fine under dapper. Suspend/resume is a bit hit or miss, a few times it works fine, then it fails, then again a few times ok. I have noticed a kernel crash at resume time. Here is the stacktrace (note this is a dual core machine, maybe that is a clue?):

Sep 13 11:01:20 localhost kernel: [34105.144000] ------------[ cut here ]------------
Sep 13 11:01:20 localhost kernel: [34105.144000] kernel BUG at drivers/scsi/libata-core.c:1099!
Sep 13 11:01:20 localhost kernel: [34105.144000] invalid operand: 0000 [#1]
Sep 13 11:01:20 localhost kernel: [34105.144000] PREEMPT SMP
Sep 13 11:01:20 localhost kernel: [34105.144000] Modules linked in: acpi_sbs i2c_acpi_ec battery ac thermal fan button michael_mic arc4 ieee80211_crypt_tkip af_packet binfmt_misc rfcomm l2cap bluetooth
ipv6 irtty_sir sir_dev ppdev speedstep_centrino cpufreq_userspace cpufreq_stats freq_table cpufreq_powersave cpufreq_ondemand cpufreq_conservative video tc1100_wmi sony_acpi pcc_acpi hotkey dev_acpi container irda crc_ccitt dm_mod md_mod sbp2 parport_pc lp parport joydev pcmcia tsdev pcspkr ieee80211_crypt yenta_socket rsrc_nonstatic psmouse serio_raw pcmcia_core nvidia i2c_core snd_hda_intel snd_hda_codec intel_agp agpgart snd_pcm_oss snd_mixer_oss snd_pcm snd_timer hw_random snd soundcore snd_page_alloc shpchp pci_hotplug evdev sr_mod cdrom sg ext3 jbd ide_generic ohci1394 ieee1394 usbcore sd_mod
generic ata_piix libata scsi_mod processor capability commoncap vesafb fbcon tileblit font bitblit softcursor
Sep 13 11:01:20 localhost kernel: [34105.144000] CPU: 0
Sep 13 11:01:20 localhost kernel: [34105.144000] EIP: 0060:[pg0+947023345/1069167616] Tainted: P VLI
Sep 13 11:01:20 localhost kernel: [34105.144000] EFLAGS: 00210046 (2.6.15-26-686)
Sep 13 11:01:20 localhost kernel: [34105.144000] EIP is at ata_exec_internal+0x1a1/0x1b0 [libata]
Sep 13 11:01:20 localhost kernel: [34105.144000] eax: 00000000 ebx: c235ff38 ecx: 00000000 edx: dfd5f2cc
Sep 13 11:01:20 localhost kernel: [34105.144000] esi: 00000000 edi: 00000200 ebp: dfd5f2cc esp: c235feb4
Sep 13 11:01:20 localhost kernel: [34105.144000] ds: 007b es: 007b ss: 0068
Sep 13 11:01:20 localhost kernel: [34105.144000] Process kacpid-work-5 (pid: 8844, threadinfo=c235e000 task=f5d17560)
Sep 13 11:01:20 localhost kernel: [34105.144000] Stack: dfd5f2cc dfd5f350 dfd5f2cc a101869f 00200202 00000000 00000001 00000000
Sep 13 11:01:20 localhost kernel: [34105.144000] c235fed4 c235fed4 00000001 c235ff38 dfd5f350 c235ff4c dfd5f364 f8b81fdb
Sep 13 11:01:20 localhost kernel: [34105.144000] dfd5f2cc dfd5f350 c235ff38 00000002 dfd5f364 00000200 c235ff18 dfd5f57c
Sep 13 11:01:20 localhost kernel: [34105.144000] Call Trace:
Sep 13 11:01:20 localhost kernel: [34105.144000] [pg0+947023835/1069167616] ata_dev_identify+0x11b/0x580 [libata]
Sep 13 11:01:20 localhost kernel: [34105.144000] [pg0+947022463/1069167616] ata_std_dev_select+0x4f/0x60 [libata]
Sep 13 11:01:20 localhost kernel: [34105.144000] [pg0+947025197/1069167616] ata_bus_probe+0x3d/0xc0 [libata]
Sep 13 11:01:20 localhost kernel: [34105.144000] [acpi_os_execute_worker+0/206] acpi_os_execute_worker+0x0/0xce
Sep 13 11:01:20 localhost kernel: [34105.144000] [pg0+947063138/1069167616] libata_acpi_notify+0x52/0x80 [libata]
Sep 13 11:01:20 localhost kernel: [34105.144000] [acpi_ev_notify_dispatch+78/91] acpi_ev_notify_dispatch+0x4e/0x5b
Sep 13 11:01:20 localhost kernel: [34105.144000] [acpi_os_execute_worker+159/206] acpi_os_execute_worker+0x9f/0xce
Sep 13 11:01:20 localhost kernel: [34105.144000] [default_wake_function+0/32] default_wake_function+0x0/0x20
Sep 13 11:01:20 localhost kernel: [34105.144000] [kthread+200/208] kthread+0xc8/0xd0
Sep 13 11:01:20 localhost kernel: [34105.144000] [kthread+0/208] kthread+0x0/0xd0
Sep 13 11:01:20 localhost kernel: [34105.144000] [kernel_thread_helper+5/16] kernel_thread_helper+0x5/0x10
Sep 13 11:01:20 localhost kernel: [34105.144000] Code: 86 8c 00 00 00 89 34 24 e8 bd 33 00 00 0f b6 44 24 0f 89 44 24 08 8b 45 0c c7 04 24 d8 be b8 f8 89 44 24 04 e8 81 30 5a c7 eb a5 <0f> 0b 4b 04 c2 cf b8 f8 e9 cb fe ff ff 89 f6 56 ba 4c 1d 00 00
Sep 13 11:01:20 localhost kernel: [34105.144000] <6>note: kacpid-work-5[8844] exited with preempt_count 1

This is a clean reinstall of dapper from yesterday. Here is the contents of /etc/default/acpi-support:

jos@duet:~$ cat /etc/default/acpi-support

# Uncomment the next line to enable ACPI suspend to RAM
ACPI_SLEEP=true

# Comment the next line to disable suspend to disk
#ACPI_HIBERNATE=true

# Change the following to "standby" to use ACPI S1 sleep, rather than S3.
# This will save less power, but may work on more machines
ACPI_SLEEP_MODE=mem

# Add modules to this list to have them removed before suspend and reloaded
# on resume. An example would be MODULES="em8300 yenta_socket"
#
# Note that network cards and USB controllers will automatically be unloaded
# unless they're listed in MODULES_WHITELIST
MODULES="uhci_hcd ehci_hcd"

# Add modules to this list to leave them in the kernel over suspend/resume
MODULES_WHITELIST=""

# Should we save and restore state using the VESA BIOS Extensions?
#SAVE_VBE_STATE=true

# The file that we use to save the vbestate
VBESTATE=/var/lib/acpi-support/vbestate

# Should we attempt to warm-boot the video hardware on resume?
POST_VIDEO=false

# Save and restore video state?
SAVE_VIDEO_PCI_STATE=true

# Should we switch the screen off with DPMS on suspend?
USE_DPMS=true

# Use Radeontool to switch the screen off? Seems to be needed on some machines
# RADEON_LIGHT=true

# Uncomment the next line to switch away from X and back again after resume.
# This is needed for some hardware, but should be unnecessary on most.
DOUBLE_CONSOLE_SWITCH=true

# Set the following to "platform" if you want to use ACPI to shut down
# your machine on hibernation
HIBERNATE_MODE=shutdown

# Comment this out to disable screen locking on resume
LOCK_SCREEN=true

# Uncomment this line to have DMA disabled before suspend and reenabled
# afterwards
# DISABLE_DMA=true

# Uncomment this line to attempt to reset the drive on resume. This seems
# to be needed for some Sonys
RESET_DRIVE=true

# Add services to this list to stop them before suspend and restart them in
# the resume process.
STOP_SERVICES=" "

# Restart Infra Red services on resume - off by default as it crashes some
# machines
RESTART_IRDA=false

# Switch to laptop-mode on battery power - off by default as it causes odd
# hangs on some machines
ENABLE_LAPTOP_MODE=false
jos@duet:~$