Bug 64923

Summary: RH 7.3 aic7xxx_mod problem
Product: [Retired] Red Hat Linux Reporter: Robin Sun <robin.sun>
Component: kernelAssignee: Doug Ledford <dledford>
Status: CLOSED CURRENTRELEASE QA Contact: Brock Organ <borgan>
Severity: medium Docs Contact:
Priority: medium    
Version: 7.3CC: enslaver, etbonick, rb_home
Target Milestone: ---   
Target Release: ---   
Hardware: i386   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2004-09-30 15:39:34 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Robin Sun 2002-05-14 15:44:21 UTC
From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0)

Description of problem:
When installing RH 7.3 from SCSI CDROM select aic7xxx driver, it kept looping 
back to SCSI/Other option then select driver list. With RH 7.3 a new item old 
aix7xxx (aix7xxx_old) is added to the list which cause the problem. I had 
installed RH 7.2 on the same computer with identical configuration, the problem 
did not exist.

Version-Release number of selected component (if applicable):


How reproducible:
Always

Steps to Reproduce:
1.Boot up from CDROM, at install device source, select CDROM than SCSI
2.Select Adaptec 29??, ???? (aic7xxx), than install driver box appear, than 
loop back to select SCSI/Others box.
3.Select old aic7xxx, a bunch of numbers appears, than back to select CDROM 
type screen SCSI/Others.
	

Actual Results:  can't get out of the loop. Had to power off the computer.

Expected Results:  detected and installed the driver and installation continous.
I think in the RH 7.2 anarcron or something was loaded. 
In RH 7.2 installation went on without any problem.


Additional info:

Computer: HP NetServer LC 5/66 (Pentium 166) 112MB, 2 SCSI HDD, Toshiba SCSI 
CDROM, HP SCSI Tape drive. Adaptec aic7770 v2.12 BIOS.
I had also tried to install Mandrake's LM 8.1 distro onto the same machine. 
Identical problem happened. RH 7.2 allowed me to install successfully which I 
had tested for a few month. Mandake new release LM 8.2 had fixed the problem, 
which I installed and now testing.
I downloaded the new RH 7.3 release and to my surprise, it won't allow me to go 
further i.e. the problem that LM 8.1 had. I use the LM 8.2 current installed on 
my computer to look into the drvblock.img file in the images/ of the CD on both 
RH 7.2 and RH 7.3 disk1. It appeared that RH 7.2 had a newer aic7xxx_mod 
driver. RH 7.3 come with a aic7xxx_old module, but without the aic7xxx_mod.
RH 7.2 did not have an extra old aic7xxx item in the driver selection list. The 
aic7xxx_mod was working fine in RH 7.2.
So, how do I switch to aic7xxx_mod of RH 7.2 and continue installation on the 
RH 7.3

Comment 1 Doug Ledford 2002-05-16 19:31:26 UTC
aic7xxx_mod is now called just aic7xxx in 7.3  The old aic7xxx driver is now
called aic7xxx_old.  Both of those should be present.

In order to provide more help I would have to know about any error messages that
come up on vt4 during the install.  In order to see them, select the CD-ROM
install, then select SCSI driver, then select aic7xxx driver, then press Alt-F4
to see the kernel messages.  Post any error messages that scroll up on the
screen here and I might be able to figure out what's going on.

Comment 2 Robin Sun 2002-05-18 20:04:20 UTC
after selecting Adaptec AHA-2740, 28xx, 29xx, 39xx (aic7xxx)
Loading aic7xxx driver... pop up then
Initializing CDROM screen pop up then
CD Type selection box reappeared, back to above.
after selecting Old Adaptec 28xx, 29xx, 39xx, (aic7xxx_old)
Loading aic7xxx driver... pop up then these populated the screen, i was only 
able to see some as the CD Type selection screen reappeared. this is on the 
screen:
Process loader (pid:34, stackpage = c68a9000)
stack : c680f000 c680db9800 c680f110 c68a9da4 c7817e6d ... more addresses

call trace:[<c7817e6d>]
[<c020ae2a>]
[<c020ae2a>]
... and more on the left side of the screen, then at the bottom
code: 8b 58 3c 31 f6 56 53 51 e8 5d 7a 96 f8 83 c4 0c 5b 5f 5d
after i hit enter, the driver screen reappeared. i can't get out of either 
Adaptec drivers, just keep looping back to the select CD type screen and start 
all over again.





Comment 3 Robin Sun 2002-05-18 20:06:44 UTC
after selecting Adaptec AHA-2740, 28xx, 29xx, 39xx (aic7xxx) from the list
Loading aic7xxx driver... pop up then
Initializing CDROM screen pop up then
CD Type selection box reappeared, back to CD type screen.
select SCSI, then
after selecting Old Adaptec 28xx, 29xx, 39xx, (aic7xxx_old)
Loading aic7xxx driver... pop up then these populated the screen, i was only 
able to see some as the CD Type selection screen reappeared. this is on the 
screen:
Process loader (pid:34, stackpage = c68a9000)
stack : c680f000 c680db9800 c680f110 c68a9da4 c7817e6d ... more addresses

call trace:[<c7817e6d>]
[<c020ae2a>]
[<c020ae2a>]
... and more on the left side of the screen, then at the bottom
code: 8b 58 3c 31 f6 56 53 51 e8 5d 7a 96 f8 83 c4 0c 5b 5f 5d
after i hit enter, the driver screen reappeared. i can't get out of either 
Adaptec drivers, just keep looping back to the select CD type screen and start 
all over again.


Comment 4 Doug Ledford 2002-05-20 20:39:53 UTC
You can't load both the aic7xxx and aic7xxx_old drivers, it results in them both
trying to operate on the same SCSI controller and confusing things up, resulting
in all sorts of kernel errors.  What I need is for you to *just* load the
aic7xxx driver, then switch to vt4 (by pressing Alt-F4) and tell me what
messages come up on that screen from the aic7xxx driver.

Comment 5 Robin Sun 2002-05-22 02:40:56 UTC
i did not load both driver at the same time. loading the first aic7xxx driver 
the installation just keep looping back to the select SCSI/Other CDROM option. 
power off the system, select the old adaptec driver(the second one on the list) 
the screen just populated with addresses (see comment posted 2002-5-18).
Today, do as you instructed, switched to vt4, screen message:
<4> ide-floppy driver 0.99 newide
<6> usb.c: registered new driver usbdevfs
<6> usb.c registered new driver hub
<6> usb-uhci.c: $Revision: 1.275 $ time 06:55:51 Apr 18 2002
<6> usb-uhci.c: High bandwidth mode enable
... continued loading and detecting other devices, the last two line
<4> VFS: Mounted root (ext2 filesystem)
<6> SCSI subsystem driver Revision: 1.00
and hang here
power down computer and start from cold boot.
select old adaptec driver, addresses populated the screen, switch to vt4 same 
messages as above.
however, when i switched to vt3, this is what i found
* no pcic controller found
* probing buses
* found nothing
* 127 keymaps are available
* load 9 keymap tables
* starting to STEP_URL
*/tmp/aic7xxx.o init_module:
Hint: insmod errors can be caused by incorrect module parameters, including 
invaild IO or IRQ parameters
* probing for floppy devices
* first non-detached floppy is fd0
* system floppy device is fd0
* fail to insert /tmp/aic7xxx.o
* load module set done
switched back to vt1, SCSI/Other CDROM and reselect the same driver, it just 
start logging from 
* starting to STEP_URL to
...
* load module set done.
power down and start computer
selected the old adaptec produces the same message in vt3 with the exception 
that the screen get populated with addresses.
Now my investigation went further.
I boot up the computer with the installed LM8.2 on the hard drive, and open up 
the Disk1 /images/drvblock.img file and compared the two RH versions.
In RH 7.2:
the line 
aic7xxx_mod
     " New (experimental) Adaptec 28xx, 29xx, 39xx"
than some where further down, the corresponding line to above
aic7xxx_mod: scsi_mod

In RH 7.3
there is no aic7xxx_mod line at the top. there is just the
aic7xxx_old
     "Old Adaptec 28xx, 29xx, 39xx"
and the corresponding line further below
aic7xxx_old: scsi_mod
regardless for what, the RH 7.2 with the description "New experimental) ..." 
works. i switch to vt3 while running RH 7.2 installation and it continued pass 
where RH 7.3 had hanged. after starting to STEP_URL, the next log loads the 
aic7xxx.o (NULL path), than other drivers etc. switched back to vt1 screen it 
continued to load anaconda, and than running anaconda etc to complete the rest 
of the installation.

Comment 6 Robin Sun 2002-06-06 13:32:07 UTC
is there a solution to this?

Comment 7 furick1 2002-06-08 00:39:52 UTC
Try "linux apic" .  It worked for me when linux noprobe and loading the 
aic7xxx and aic7xxx_old wouldn't work.  I recalled doing this to get 7.2 to 
install as well but had forgot until I went back to the old bugs and found the 
work around.

Comment 8 Need Real Name 2002-06-21 03:38:49 UTC
I have a HP Netserver 5/166 with 7770 onboard, identical symptoms to the 
Netserver with the same chipset already contained in this bug report.

I can boot to the RH 7.3 CD1.  All goes well until it asks for installation 
source.  After selecting 7xxx (tried new, and tried old), I end up back at the 
menu to choose a source again.  <power switch>

Have not tried the APIC thing yet, though I did make it through a 440 mainboard 
installation not more than two months ago(RH 7.2).  The fix wasn't APIC, though 
that did work;  the fix at the time was a BIOS update from Intel to address PCI 
routing issues.

Is this the same thing all over again?

Please keep me posted.

Thanks in advance,

Ryan Beisner
rb_home


Comment 9 Robin Sun 2002-06-21 14:06:39 UTC
Hi Ryan,
Glad to hear that you encountered the same problem. Tried Linux APIC suggested 
by furick1, didn't work. I'm pretty sure it was the file location problem.
Just out of curiousity, I wanted to see the installation log of RH7.2, at 
installing CDROM, SCSI selection, there were multiple lines of insmod and with 
a message <Null path> after every successful insert. With RH7.3 installation, 
insmod was searching for drivers at /tmp/aic7xxx.o; a specific location or path 
which led to an error
* fail to insert /tmp/aic7xxx.o
then it just keep looping back to the CDROM selection screen. Which indicated 
that the installation is looking for files at a specific location, can't find 
it, go back to look for it again, the symthom of infinite loop.
As I mentioned to RH in this issue earlier. I was able to install RH7.2 on the 
same machine without any problem. I did encountered same problems with Linux 
Mandrake 8.1. However, LM8.2 has fixed the problem. RH seemed to have gone 
backward.

Comment 10 Need Real Name 2002-06-21 16:38:55 UTC
*LINUX APIC does not fix this.
*LINUX NOPROBE does not fix this.

This is always reproducible on 7.2 & 7.3.  I'm trying 7.1 right now...


Comment 11 Need Real Name 2002-06-21 17:28:56 UTC
Same story in 7.1.

I just installed Mandrake 8.0 (Kernel 2.4.3) with no problem.  Booted to CD, 
selected packages, and it's booted to KDE desktop right now.

I don't want Mandrake, I want Redhat!   = )




Comment 12 Robin Sun 2002-06-21 22:14:24 UTC
I never install LM8.0 onto my HP, but, LM8.1 gave me the similar problem as 
RH7.3. LM8.2 has fixed the problem.

Yes, I like to stay with RH too.

Comment 13 Need Real Name 2002-06-21 22:48:13 UTC
FYI  Debian Potato (2.2 rev 6) also had no problem initializing the cdrom.  I 
just installed it, manually partitioned, formatted; and all is well.

FYI  I am only testing these different distros so that RH may pull some more 
clues out of these trials.

To me, it's simple:  look at how everybody else is handling this chipset, and 
make the necessary adjustments.  I hope they can do that quickly.

My goal is to end up with RH on the box.  But in the mean time, I am going to 
proceed with a different distro until I see that this bug has been properly 
addressed.  I can't have this Netserver sitting on my workbench staring at me 
until someone unbugs it's current woes with RH 7.x.
  = )

Patiently waiting, eager to help,
    -RB

Comment 14 Need Real Name 2002-06-22 02:58:00 UTC
Is this really a kernel problem?  ...look at the *path* error?  I know just 
enough to be dangerous so watch out.
-RB


Comment 15 Robin Sun 2002-06-24 14:26:21 UTC
I don't think its a kernel problem. I'm quite certain it's a path problem in 
the installation script. Fron the installation log (Alt F3) you will see each 
stages of installation. The CDROM installation is caught in an infinite loop 
during an insmod task looking to insert the aic7xxx.o module from a specific 
location. RH7.3 path is set to install drivers from /tmp which is not on the 
CD. Referred back to RH7.2, path is set to <NULL PATH> which allow flexible 
search of drivers.

I've only use RedHat and Mandrake distros. I've no desire testing other distros 
at this time, particularly Caldera (I had a very bad experience with SCO back 
in 94. Too bad Caldera now owned by SCO).

My goal is pretty similar to yours. First, I'd like to see a smooth and trouble 
free installation. Afterall, installation is the first key exposure to users. A 
frustrating installation turns new users away, old users to look for other 
distros. I had switched from LM to RH and now back to LM simply becasue I'm 
able to install LM successfully. I would like to switch back to RH as they 
actually run faster and smoother on the older systems once they are up and 
running.

Comment 16 Doug Ledford 2002-06-26 20:36:31 UTC
The kernel messages you posted (specifically these):

*/tmp/aic7xxx.o init_module:
Hint: insmod errors can be caused by incorrect module parameters, including 
invaild IO or IRQ parameters

indicate that the installer is indeed finding and loading the aic7xxx module. 
It is during the card detection phase that the aic7xxx driver is deciding you
don't have any aic7xxx compatible controllers in your system.  In order to get
more information on this out of the aic7xxx driver, you need to boot the system
with the option "linux noprobe" then go in an manually select the old aic7xxx
driver, then before you actually load the module, check the box that allows you
to specify module options and pass the option aic7xxx=verbose:0xffff to the
module.  That should produce a *lot* more output on vt4 so that we can tell why
the driver is refusing to recognize your card.  Once you have that output, post
it here and I'll see if I can't give you a workaround to make things run
smoothly for you.


Comment 17 Robin Sun 2002-06-28 00:48:04 UTC
Tried as you suggested, same error as posted on 2002-5-18 16:04:44

code: 8b 58 3c 31 f6 56 53 51 e8 5d 7a 96 f8 83 c4 0c 5b 5e 5f 5d

and loop back to the DialogBox CDROM type with 2 options SCSI, Other CDROM

goto vt3, last line of the log

failed to insert /tmp/aic7xxx_old.o

About your hint: 
incorrect module parameters. possibly
incorrect IO or IRQ parameters. I disagree. As I can pop in the RH7.2, LM8.1 or 
even MS NT4 or W2K, and it will have no problem installing them.


Comment 18 Doug Ledford 2002-06-28 01:03:34 UTC
That error (the code: line) is only one small part of the problem.  The entire
output, starting with the kernel oops: line is the entire kernel oops report. 
The problem is that both aic7xxx drivers are oopsing on your computer and
without a detail listing of the entire oops report, I have 0 chance of figuring
it out.  I need you to try booting up with the option no_probe, then select the
aic7xxx_old driver (and *only* the _old driver, don't try the new one at all),
then let me know if it just fails to find your controller or if it oopses the
computer.  You only need to try and load the driver once.  Go to vt4 after the
attempt and see if there is an oops there on the screen.  If so, I need that, in
its entirety.

Comment 19 Robin Sun 2002-06-28 13:52:50 UTC
That was exactly what I did last night. Booting up and typed Linux noprobe at 
the prompt. At the CDROM drivers selection screen, selected *only* the _old 
aic7xxx driver, did not try the new one. I don't recall any oops lines. I'll do 
as you suggested again and look for that oops report.

Comment 20 Robin Sun 2002-06-28 23:30:26 UTC
Tried it again as suggested, this is whats on the entire vt4 screen

---------------------- Top of screen -----------------------------------------
<>Process loader (pid:30, stackpage=c689000)
<4>Stack: c680f000 c6db9800 c680f110 c6897da4 c7817ebd c680f000 c6897da4 0000000
0
<4>       00000000 33323130 c680f000 42413938 46454443 4a494847 4e4d4c4b 5251504
f
<4>       56555453 5a595857 00080000 00519005 00000008 33323130 37363534 6261393
8
<4>Call Trace: [<c7817e6d>]
<4>[<c020ae2a>]
<4>[<c014c4f0>]
<4>[<c01176f1>]
<4>[<c01176f1>]
<4>[<c7844100>]
<4>[<c7810547>]
<4>[<c783db46>]
<4>[<c7844100>]
<4>[<c7844100>]
<4>[<c0118425>]
<4>[<c78414a0>]
<4>[<c7829060>]
<4>[<c01085e7>]
<4>
<4>
<4>Code: 8b 58 3c 31 f6 56 53 51 e8 5d 7a 96 f8 83 c4 0c 5b 5e 5f 5d
<4>
------------- bottom of screen -----------------------------------------------


Comment 21 Ethan Bonick 2002-08-23 05:27:47 UTC
MSI model ms-6347 has same adaptec chip onboard. In 7.3 it freezes teh system 
trying to load the module and the same for 7.1. I haven't tried APIC yet but 
will sometime soon.

Comment 22 Joshua Trimm 2003-11-06 20:50:20 UTC
Is there any fix for this or is this issue still open? I am having the
exact same trouble and have tried everything here.

Comment 23 Bugzilla owner 2004-09-30 15:39:34 UTC
Thanks for the bug report. However, Red Hat no longer maintains this version of
the product. Please upgrade to the latest version and open a new bug if the problem
persists.

The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases, 
and if you believe this bug is interesting to them, please report the problem in
the bug tracker at: http://bugzilla.fedora.us/