i386/87026: [hang] Bootup hang on atkbdc on Compaq 1850R
between 6.0-CURRENT-SNAP004 and 6.0-BETA5
Frank Durda IV
uhclem at freebsd.org
Thu Mar 30 18:00:44 UTC 2006
The following reply was made to PR i386/87026; it has been noted by GNATS.
From: uhclem at freebsd.org (Frank Durda IV)
To: bug-followup at FreeBSD.org, nevans at talkpoint.com
Cc:
Subject: Re: i386/87026: [hang] Bootup hang on atkbdc on Compaq 1850R between 6.0-CURRENT-SNAP004 and 6.0-BETA5
Date: Thu, 30 Mar 2006 11:55:00 -0600 (CST)
As some additional data points on this problem, some systems I
have come into posession of demonstrate similar if not identical
failures, sometimes BOTH on FreeBSD 5.4 RELEASE and 6.1BETA4, but on
more machines on 6.1BETA4. (I have only tested 5.4 RELEASE and 6.1BETA4.)
The first system in question is a Compaq DL380R01 P800-256 128 US
Dual 800/133 MHz with 256Kbyte cache processors, both CPUs present.
Problem was originally noted with the Compaq Remote Insight card
installed (PCI card device), but it was removed and it didn't
change/improve the situation. Machine previously only ran NT.
All PCI slots are empty.
Verbose boot messages noted in FreeBSD 6.1BETA 4 (hand transcribed)
COMPAQ System BIOS - P17 (12/18/2002)
...
...
ioapic0; routing intpin 1 (ISA IRQ 1) to vector 54
atkbd0: [GIANT-LOCKED]
psm0: current command byte:0065
kdbc: TEST_AUX_PORT status:0003
psm0: strange result for test aux port (3).
kbdc: RESET_AUX return code:0000
The system hangs at this point. Will not respond to CTRL-ALT-DEL.
Must cycle power since Compaqs don't have a reset button.
Does not fail 100% of the time, but failure rate exceeds 90%. By chance,
I was eventually able to boot from CD-ROM and then installed 6.1BETA4,
but have been unable to boot direct from the hard disk or CD-ROM since,
even with Verbose logging turned on.
Booting the same box with FreeBSD 5.4 RELEASE also fails with a few
more verbose messages: (hand transcribed)
COMPAQ System BIOS - P17 (12/18/2002)
...
...
atkbdc0: <Keyboard controller (i8042)> port 0x64,0x60 irq 1 on acpi0
atkbd0: <AT Keyboard> irq 1 on atkbdc0
atkbd: the current kbd controller command byte 0065
atkbd: keyboard ID 0x41ab (2)
kbdc: RESET_KBD return code:00fa
kbdc: RESET_KBD status:00aa
kbd0 at atkbd0
kbd0: atkbd0, AT 101/102 (2), config:0x0, flags:0x3d0000
atkbd0: [GIANT-LOCKED]
psm0: current command byte:0065
kdbc: TEST_AUX_PORT status:0003
psm0: strange result for test aux port (3).
kbdc: RESET_AUX return code:0000
and we die here.
Meanwhile, the NT image on disk that I was unable to over-write boots
and works, so the keyboard/mouse hardware appears functional.
- - - -
I also have a Compaq PL1850R 6/450 512 64 US, (1 CPU installed) which
seems to be the immediate predecessor to the DL380 above (same chassis),
and it also fails in a similar way, on 6.1BETA4, but boots and
installs fine on 5.4 RELEASE.
Here are the verbose boot messages from 6.1BETA4 (which hangs)
COMPAQ System BIOS - P07 (02/14/2002)
...
acpi: bad RSDP checksum (224)
...
PNP Identify complete
unknown: status reg test failed ff
unknown: status reg test failed ff
unknown: status reg test failed 0
unknown: status reg test failed ff
unknown: status reg test failed ff
unknown: status reg test failed ff
aha_isa_probe 2: ioport 0x2c00 alloc failed
sc: sc0 already exists; skipping it
...
aic0: not probed (disabled)
atkbdc0: <Keyboard controller (i8042)> port 0x64,0x60 irq 1 on acpi0
atkbd0: <AT Keyboard> irq 1 on atkbdc0
atkbd: the current kbd controller command byte 0065
and it hangs there.
Here are the verbose boot messages from 5.4-RELEASE (which boots and
installed and rebooted)
COMPAQ System BIOS - P07 (02/14/2002)
...
acpi: bad RSDP checksum (224)
...
ata: ata1 already exists; skipping it
Trying Read_Port at 203
Trying Read_Port at 243
Trying Read_Port at 283
Trying Read_Port at 2c3
Trying Read_Port at 303
Trying Read_Port at 343
Trying Read_Port at 383
Trying Read_Port at 3c3
ex_isa_identify()
unknown: status reg test failed ff
unknown: status reg test failed ff
unknown: status reg test failed 0
unknown: status reg test failed ff
unknown: status reg test failed ff
unknown: status reg test failed ff
sc: sc0 already exists; skipping it
...
aic0: not probed (disabled)
atkbdc0: <Keyboard controller (i8042)> port 0x64,0x60 irq 1 on acpi0
atkbd0: <AT Keyboard> irq 1 on atkbdc0
atkbd: the current kbd controller command byte 0065
atkbd: keyboard ID 0x41ab (2)
kbd0 at atkbd0
kbd0: atkbd0, AT 101/102 (2), config:0x0, flags:0x3d0000
atkbd0: [GIANT-LOCKED]
psm0: current command byte:0065
psm0: <PS/2 Mouse> irq 12 on atkbdc0
psm0: [GIANT-LOCKED]
psm0: model IntelliMouse Explorer, device ID 4-00, 5 buttons
psm0: config:00000000, flags:00000008, packet size:4
psm0: syncmask:08, syncbits:00
bt0: not probed (disabled)
cs0: not probed (disabled)
...
Testing was done with all PCI card slots empty.
This machine also previously only ran NT.
- - - -
Next, I have a Compaq DL380R01 P1000 256 US which boots on 6.1BETA4
consistently. Inside the case, the only obvious change from the 800MHz
machine mentioned above to this one is the processors are now 1GHz vs
800MHz. The BIOS release also claims to be the same. Verbose boot
messages for this successful boot look like this:
COMPAQ System BIOS - P17 (12/18/2002)
...
atkbdc0: <Keyboard controller (i8042)> port 0x64,0x60 irq 1 on acpi0
atkbd0: <AT Keyboard> irq 1 on atkbdc0
atkbd: the current kbd controller command byte 0065
atkbd: keyboard ID 0x41ab (2)
kbd0 at atkbd0
kbd0: atkbd0, AT 101/102 (2), config:0x0, flags:0x3d0000
atkbd0: [GIANT-LOCKED]
psm0: current command byte:0065
psm0: <PS/2 Mouse> irq 12 on atkbdc0
ioapic0: routing intpin (ISA IRQ 12) to vector 55
psm0: [GIANT-LOCKED]
psm0: model IntelliMouse Explorer, device ID 4-00, 5 buttons
psm0: config:00000000, flags:00000008, packet size:4
psm0: syncmask:08, syncbits:00
ppc0; using extended I/O port range
ppc0: SPP
...
and the boot succeeds.
- - - -
All three systems were tested above used the same KVM, keyboard and
mouse (Logitech optical) attached. No network or other external
devices were attached.
This all makes me somewhat suspect a race/timing issue, which tends to
cause more frequent problems on 6.1BETA4 (compared to 5.4R) on the
slower processors.
- - - -
I'll also mention that despite the fact that the 1GHz system does boot
and function predictably, it takes an extremely long time to boot, pausing
for almost 90 seconds late in the probe/start process, after the second
CPU is launched. Verbose logging shows it actually stalls out after:
...
INTR: Assigning IRQ 19 to local APIC 0
ioapic0: Assigning PCI IRQ 19 to local APIC 0
(... 88 to 90 second pause ...)
Trying to mount root from ufs:/dev/idad0s1a
Perhaps this issue that bears some thought since it does look like the
system has hung and some people might not wait that long before giving
up. ([CTRL]-T does not display anything during the pause either, so
we aren't that far along.)
Also, the DL380 models also have a variable speed processor/chassis fan
that runs at full tilt under FreeBSD but under NT it would slow down too
a slower and be much quieter (aka "office friendly") speed during the
NT boot process and would remain quiet until a subsequent reboot.
It would be nice if FreeBSD could knock this fan speed down so I don't
have to put the FreeBSD boxes in far away places just because of the
noise.
Because the above machines that won't reliably boot FreeBSD are useless
to me, I have set them aside, but will eventually have to dispose of
them. (They are part of a larger collection of Compaq/HP machines dumped
in my lap as part of a corporate split, and I only need a portion of them
to do what needs to be done.) If you would like some further debugging
attempted on the above machines, please request this ASAP before I put
these things in storage or they get sold.
Thanks.
Frank Durda IV uhclem at freebsd.org |"The Knights who say "LETNi"
or uhclemLOSE.apr06%nemesis.lonestar.org| demand... A SEGMENT REGISTER!!!"
You must remove the "LOSE" to mail me. |"A what?"
http://nemesis.lonestar.org |"LETNi! LETNi! LETNi!" - 1983
More information about the freebsd-i386
mailing list