Bug #1127

cdrom drive not detected

Added by tgr about 6 years ago. Updated about 5 years ago.

Status:NewStart date:
Priority:HighDue date:
Assignee:corecode% Done:

0%

Category:-
Target version:-

Description

Hi, I've got a bit of a problem with the install CD. I've tested this on
2.0 and 2.1.0-DEV (i.e. the daily iso downloadable on 2008.08.22). This
happens way before I can make any sort of FS, but the same machine's
booted up XP, Kubuntu and FBSD5.4.

Anyway, what I can see on the bootup screen of that machine now is as
follows:

<quote>
iso_mountroot: can't find rootvp
Root mount failed: 6
Mounting root from cd9660:acd0
no disk named 'acd0'
setrootbyname failed
iso_mountroot: can't find rootvp
Root mount failed: 6
Mounting root from cd9660:acd1
no disk named 'acd1'
setrootbyname failed
iso_mountroot: can't find rootvp
Root mount failed: 6
Mounting root from cd9660:/dev/acd0a
no disk named 'acd0a'
setrootbyname failed
iso_mountroot: can't find rootvp
Root mount failed: 6
</quote>

When I then type "panic", the following appears:

<quote>
panic: panic from console
Trace beginning at frame 0xc872ca8
panic(c05ce64a,c06bc720,c05d0be3,c0872cd8,6) at panic+0x99
panic(c05d0be3,c05ce4a6,696e6170,c0560063,8) at panic+0x99
vfs_mountroot_ask(c4360b70,c06702dc,d7886c3c,87d000,c0872d98) at
vfs_mountroot_ask+0xd7
vfs_mountroot(0,87d000,86fc00,87d000,0) at fvs_mountroot+oxcf
mi_startup(86f000,d,c06b2a38,c0872c24,c0872c0c) at mi_startup+0x99
begin() at begin+0x42
Debugger("panic")
Stopped at Debugger+0x44: movb $0,in_Debugger.0
</quote>

I've run this twice in a row just to verify all the values, in case they
changed. They didn't.

If there is any more information I can provide, please do tell. This
system is not in use in any other way, all data on the HDs etc is
discardable, so anything goes.

dmesg_with_ahci (17.2 KB) tgr, 08/22/2008 09:28 PM

dmesg_with_ide (17.1 KB) tgr, 08/22/2008 09:28 PM

dmesg_with_raid (8.77 KB) tgr, 08/22/2008 09:28 PM

mountroot.txt Magnifier (977 Bytes) tgr, 08/23/2008 10:19 PM

pciconf.txt Magnifier (1.81 KB) tgr, 08/23/2008 10:58 PM

dmesg.txt Magnifier (32.6 KB) tgr, 08/25/2008 12:24 PM

History

#1 Updated by tgr about 6 years ago

Just to clarify this, I'm getting this when I boot up with the install
CD. i.e. I'm just booting the computer up and waiting for the install
process to pop up with an interactive screen, and it pops up with a
mountroot CLI, which is where I'm typing "panic" to get at the kernel
debugger.

A bit of info on the hardware itself. It's an Asus P5B BIOS Revision
0701 with an intel core2 duo 6700 cpu (2.66ghz), which I believe uses a
jmicron sata/IDE adapter for the IDE interface (which is where my CDROM
is connected). The optical drive is a Sony DVD RW DW-U10A, although I'm
not sure if that matters much. I'm not sure if it's relevant, but I see
"ata8" right before the errors in the first quoteblock, followed by the
mountroot interface. I'm guessing the problem's mounting the drive,
although the fact that XP, Kubuntu and FBSD5.4 has no problem mounting
it makes me wonder why dfbsd should do so as well.

I wish I had some way of getting more information than this from the
actual bootup process, but unless someone knows of a way to get a
dmesg-like output from inside the mountroot, I'm stuck.

Actually, I just made a weird change here, I discovered that the chipset
was set to run in RAID mode, I set it back to IDE mode, and it made a
difference. Now it's just spewing forth a whole lot of

<quote>
intr 11 at <number between 2000 and 7000>/20000 hz, livelock removed
intr 11 at 40001/40000hz, livelocked limit engaged!
</quote>

and at the end the following two lines came a few seconds apart:

acd0: WARNING - TEST_UNIT_READY taskqueue timeout - completing request
directly
acd0: WARNING - TEST_UNIT_READY freeing taskquaue zombie request

Also, the DVD drive won't respond to any requests for ejecting the tray
anymore, not sure if that's the standard *nix mounting routine which has
partially engaged or what, but it's another symptom I thought I'd throw
out there.

Also, installed FBSD 7.0 here just to see if I could get a dmesg output
from that, and that seems to work better. It still complains, but it
does move on after a few seconds. Just to make it even better, the
jmicron card has 3 modes it can run in, RAID, IDE and AHCI. I have no
idea why they make such a difference in dmesg size, but I'm attaching
them for completeness.

Debian 3.0 just stops up during the detection phase, whereas Debian
4.0r4 continues on without a hitch.

I sincerely hope this helps giving this issue a bit more attention. But
as before, if there's any info I should get, tell me and I'll make sure
I collect it.

#2 Updated by tgr about 6 years ago

Just did a bit more testing.

Windows XP: works fine.
Debian 3.0: hangs during bootup of install cd
Debian 4.0r4: works fine.
FBSD7.0: complains with the following:
<quote>
acd0: TIMEOUT - READ_BIG retrying (1 retry left)
acd0: TIMEOUT - READ_BIG retrying (0 retries left)
acd0: FAILURE - READ_BIG timed out
GEOM_LABEL: Label for provider acd0t01 is iso9660/FreeBSD_Install.
</quote>
before apparently switching to a compatible mode of reading which works.
DFBSD1.1, 2.0, 2.1-DEVEL/DAILY (downloaded 2008.08.22): pops into
mountroot when it tries to mount the ide dvd drive. Tried with 2
different (and new) IDE cables as well, and also tried in a quadcore
asus p5q with the exact same result. An old dual P3 box which I use as a
firewall booted up just fine.

And just now I even tried to install it on the firewall (which doesn't
die on itself during bootup of the install cd). That went fine, I tested
that it supported a reboot, but when I moved the harddisk over to the
asus p5n box in question, it dies during bootup. The last thing I see
here is in the attached file "mountroot.txt".

So basically, something on the Asus p5n and p5q motherboards are too new
to be supported, or there's a bug in the NATA module after it was ported
from fbsd, because fbsd 7.0 works just fine.

#3 Updated by tgr about 6 years ago

Attached is the pciconf taken from fbsd7.0

#4 Updated by qhwt+dfly about 6 years ago

Can you try booting with ACPI disabled and see anything changes?

Cheers.

#5 Updated by tgr about 6 years ago

I was fairly certain I had tried (but forgotten to mention it), but I
just tried again. normal, without acpi and in safemode, they all end up
with mountroot, unfortunately.

#6 Updated by tgr about 6 years ago

So, I finally managed to get a verbose boot dmesg after about a week's
worth of trying random shit just to get it to boot, where even pxe was
b0rked.

Enjoy the dmesg.txt, I sure didn't. I hope that'll be helpful in getting
things one step further.

#7 Updated by corecode over 5 years ago

Could you try with a recent snapshot whether the problem persists? This might
have been fixed with my last changes to nata.

#8 Updated by swildner over 5 years ago

I just investigated with tgr, and the machine boots up fine using a kernel with
SMP and APIC_IO.

#9 Updated by tpreitzel about 5 years ago

I receive the following error when attempting to boot the CDROM iso:

iso_mountroot: can't find rootvp

This error sends the boot process into the debugger. Actually, I've had this
problem with the last several versions of DragonFly BSD. Apparently, this
problem is significant as others have experienced it also.

BTW, my system is a Compaq SR5113WM. However, I've added another hard drive to
the system so the drives aren't in OEM boot order. The OEM drive (Hitachi) is
PATA and the added drive is SATA.

#10 Updated by nobody about 5 years ago

I have a similar error on my home "debranded HP media PC" which is a
Core2Duo intel chipset based motherboard, SATA hard disk and DVDRW, with
bios option set to AHCI mode instead of IDE mode. With default boot
options (no keyboard input) the CD fails to boot with same rootvp can't
be found, apparently trying acd0, which doesn't exist under AHCI, and
ends debugger prompt. If I watch the boot process and choose the "no
AHCI" boot option, It finds the DVD rom and works fine, presumably
because acd0 exists under nata. On a slightly older version, It would
fail to the mountroot prompt and typing in cd9660:cd0 works sometimes or
not with an "Device is in process of getting ready" type of message,
showing up...

> tpreitzel <> added the comment:
>
> I receive the following error when attempting to boot the CDROM iso:
>
> iso_mountroot: can't find rootvp
>
> This error sends the boot process into the debugger. Actually, I've had this
> problem with the last several versions of DragonFly BSD. Apparently, this
> problem is significant as others have experienced it also.
>

#11 Updated by corecode about 5 years ago

I just tried installing 2.4 on a Dell Optiplex, but it didn't find my DVD drive.

I tried both default and AHCI disabled - but it didn't work either way. I'll check the BIOS settings later.

The USB keyboard also doesn't work reliably.

cheers
simon

#12 Updated by dillon about 5 years ago

:I have a similar error on my home "debranded HP media PC" which is a
:Core2Duo intel chipset based motherboard, SATA hard disk and DVDRW, with
:bios option set to AHCI mode instead of IDE mode. With default boot
:options (no keyboard input) the CD fails to boot with same rootvp can't
:be found, apparently trying acd0, which doesn't exist under AHCI, and
:ends debugger prompt. If I watch the boot process and choose the "no
:AHCI" boot option, It finds the DVD rom and works fine, presumably
:because acd0 exists under nata. On a slightly older version, It would
:fail to the mountroot prompt and typing in cd9660:cd0 works sometimes or
:not with an "Device is in process of getting ready" type of message,
:showing up...

Hmm. A CD/DVD detected via the AHCI driver will probe as 'cd0'.
The root mount should try acd0, cd0, acd1, cd1, and cd8 (for USB
attached CD/DVDs).

The AHCI driver is supposed to probe everything before mountroot
runs. If it isn't one way to test and debug the situation is
to type '?' + <return> at the mountroot> prompt several times.
If the CD/DVD is still being probed then it should show up in the
list after a few tries and you should then be able to specify it
as the root mount.

The question would then become: Why didn't it probe before the
mountroot? The AHCI driver explicitly probes devices before letting
CAM finish up and mountroot to proceed. To figure that out a verbose
boot is probably needed. Sometimes a verbose boot changes the timing
enough such that the problem might not show up at all though :-(.

-Matt

#13 Updated by corecode about 5 years ago

Simon 'corecode' Schubert wrote:
> I just tried installing 2.4 on a Dell Optiplex, but it didn't find my
> DVD drive.
>
> I tried both default and AHCI disabled - but it didn't work either way.
> I'll check the BIOS settings later.

It is a PATA drive. Booting without ACPI works.

> The USB keyboard also doesn't work reliably.

Sometimes it works on boot, sometimes it doesn't :/

Also available in: Atom PDF