Bug 54386

Summary: aic7xxx driver crashes installer
Product: [Retired] Red Hat Linux Reporter: Need Real Name <steve>
Component: kernelAssignee: Arjan van de Ven <arjanv>
Status: CLOSED DUPLICATE QA Contact: Brock Organ <borgan>
Severity: high Docs Contact:
Priority: medium    
Version: 7.1CC: ckjohnson, dledford, msf
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: 2001-10-16 20:25:03 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 Need Real Name 2001-10-05 13:33:22 UTC
From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 5.01; Windows 98)

Description of problem:
I have the boxed set of RedHat Linux Deluxe Workstation.  I have already 
reported this problem through tech support (ticket #191896), but they 
suggested I report the problem here.

My system is as follows:

ASUS CUV4X-E Motherboard
1 GHz P3
Adaptec 29160 SCSI Controller
Two IBM 36G 10,000 RPM U160 Hard Drives

I am raiding the two hard drives with separate RAID1 partitions 
for /, /boot, and /home.

The "use linear mode" box is checked, but I am not specifying any kernel 
options.

I am selecting the "everything" option on the package selection page.

The installer always crashes during formatting the drives or near the 
beginning of package installation.  When it crashes, the machine locks up 
completely.
  
I tried using the aic7xxx_mod driver, but I get the same results.

I tried using the boot disk that supposedly fixes the aic7xxx problems 
with 440GX motherboards, but that didn't help either.

Here's a dump of virtual console 4 when the system crashes:

<6>RAID level 1 does not need chunksize! Continuing anyway.
<6>md2: max total readahead window set to 508k
<6>md2: 1 data-disks, max readahead per data-disk: 508k
<6>raid1: device sda7 operational as mirror 1
<6>raid1: device sdb7 operational as mirror 0
<4>raid1: raid set md2 not clean; reconstructing mirrors
<6>raid1: raid set md2 active with 2 out of 2 mirrors
<6>md: updating md2 RAID superblock on device
<4>sda7 [events: 00000001](write) sda7's sb offset: 31575616
<6>md: serializing resync, md2 shares one or more physical units with md1!
<4>sdb7 [events: 00000001](write) sdb7's sb offset: 31575616
<4>.
invalid operand: 0000
CPU:    0
EIP:    0010:[<f905b64f>]
EFLAGS: 00010202
eax: ee1f23d0   ebx: 00000002   ecx: 0000002e   edx: 0000002e
esi: f7120cb0   edi: ee1f2250   ebp: f7120c00   esp: f30e9bb8
ds: 0018   es: 0018   ss: 0018
Process mke2fs (pid: 85, stackpage=f30e9000)
Stack: f9006fae ee1f2250 00000001 f7120c00 00000040 00000000 00000000 
f90072a8
       f7120c00 00000001 0000002c 00000001 00000001 0000006c 00000009 
f729d618
       00000040 00000000 00000001 00000040 f7120c00 00000040 00000001 
00000000
Call Trace: [<f900fae>] [<f90072a8>] [<f9010cc1>] [<f9006663>] 
[<f9035c4e>] [<f90df19>] [<c010a019>]
       [<c010a168>] [<c0108f58>] [<c0120018>] [<c012fb91>] [<c01341ab>] 
[<f90128e0>] [<f9006e50>] [<f9007066>]
       [<f90014bb>] [<f90076c5>] [<f90128e0>] [<f9006e50>] [<f9007066>] 
[<f905a643>] [<f9006e50>] [<f9007066>]
       [<f90072a8>] [<f9010cc1>] [<f9006663>] [<f903df1f>] [<c012e335>] 
[<c010a168>] [<c010a188>] [<c0108eaf>]
       [<c010002b>]

Code: c7 6a 00 66 8b 47 08 66 c1 e8 09 0f b7 c0 50 53 e8 c4 52 14
Kernel panic: Aiee, killing interrupt handler!
In interrupt handler - not syncing



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


How reproducible:
Always

Steps to Reproduce:
1. Run installer.
2.
3.
	

Actual Results:  Installer crashes while formatting drives or while 
installing packages.

Expected Results:  Installer shouldn't crash.


Additional info:

Comment 1 Arjan van de Ven 2001-10-05 13:38:46 UTC
Asus CUV4X: could you please check the "MPS" setting in the bios? I have several
reports that setting that to "none" fixes quite a few problems with this board.

Also, if you type "linux noprobe" at the syslinux prompt of the installation CD,
you get a choice of aic7xxx drivers to install (we ship 2 different ones). Could
you try "aic7xxx_mod" instead of "aic7xxx" ?

Comment 2 Need Real Name 2001-10-05 14:56:54 UTC
I assume by "MPS" you mean Multiple Processor Support?  I couldn't find any 
options for this in my BIOS.  As far as I know, this motherboard is not multi-
processor capable (there is only one socket on the motherboard, and it is 
filled).

With regards to booting with "linux noprobe" and using the aic7xxx_mod driver, 
I tried that and I get the same results as when using the aic7xxx driver.  
Here's the dump from virtual console 4 when using the aic7xxx_mod driver.

<6>md2: 1 data-disks, max readahead per data-disk: 508k
<6>raid1: device sdb8 operational as mirror 1
<6>raid1: device sda8 operational as mirror 0
<4>raid1: raid set md2 not clean; reconstructing mirrors
<6>raid1: raid set md2 active with 2 out of 2 mirrors
<6>md: updating md2 RAID superblock on device
<4>sdb8 [events: 00000001](write) sdb8's sb offset: 26097472
<6>md: serializing resync, md2 shares one or more physical units with md1!
<4>sda8 [events: 00000001](write) sda8's sb offset: 26097472
<4>.
<4>(scsi0:A:1:0): Locking max tag count at 128
<4>(scsi0:A:0:0): Locking max tag count at 128
Oops: 0000
CPU:    0
EIP:    0010:[<c0110ada>]
EFLAGS: 00010006
eax: e2b86ee8   ebx: e2b87eec   ecx: 00000000   edx: e2b86eec
esi: e2b86ea0   edi: e2b87ee4   ebp: f2fedb90   esp: f2fedb74
ds: 0018   es: 0018   ss: 0018
Process mke2fs (pid: 86, stackpage=f2fed000)
Stack: e2b86eec 00000001 00000286 00000003 01b81002 e2b86ea0 e9796830 00000246
       f90585c0 e2b86ea0 00000001 f90128e0 f71d5160 f729de00 00000000 e9796830
       e9796730 00000001 f6952c00 f90586e0 e9796830 00000001 00000002 f6952cb0
Call Trace: [<f90585c0>] [<f90128e0>] [<f90586e0>] [<f9006fae>] [<f90072a8>] 
[<f9010cc1>] [<f9001be3>]
       [<f9001a56>] [<c0116b44>] [<c0116a79>] [<c0116993>] [<c010a19b>] 
[<c0108f58>] [<c0120018>] [<c012fb91>]
       [<c01341ab>] [<c011966a>] [<f9004a3e>] [<f903506e>] [<f9004a3e>] 
[<f900196b>] [<f9035a8d>] [<f9001484>]
       [<c010a019>] [<f903506e>] [<f9001484>] [<f90077e4>] [<f90128e0>] 
[<f9006e50>] [<f9007066>] [<f90072a8>]
       [<f9010cc1>] [<c0116b44>] [<c012e335>] [<c0108eaf>]

Code: 8b 01 85 45 f0 8b 1b 74 50 31 c0 9c 5e fa c7 01 00 00 00 00
Kernel panic: Aiee, killing interrupt handler!
In interrupt handler - not syncing



Comment 3 Doug Ledford 2001-10-16 17:03:21 UTC
I attempted to see what was oopsing by decoding the output of the oops. 
Unfortunately, without a module load map, I have no clue which modules are
loaded where and the oops report you have given spans at least 3 or 4 different
modules.  So, the only way I know to proceed from here is to try and get a
module load map of the system before it oopses, then cause it to oops and attach
the matching module map and oops to this bug report.  I'm probably not the best
person to try and detail how to get that done.  Someone from the install team
might have better ideas about how to do this than I do (currently, the only
thing I know to do is to cat /proc/ksyms after the aic7xxx module has been
loaded and try to save it to the hard disk, which assumes the machine will
survive the mke2fs process, or maybe you could boot from CD-ROM, then when you
have a working shell on vt2 you might be able to save the output to a formatted
floppy disk or something like that).

Comment 4 Need Real Name 2001-10-16 17:56:52 UTC
I'd be glad to send you any information you need.  Please give me more detailed 
instructions on what you want me to do, or put me in touch with someone who can 
give me the necessary instructions.  If the instructions are complicated, it 
may be easier to do over the phone, which should be included in the support 
package of my RH7.1 Deluxe Workstation.  Doing some of this over the phone may 
make things go more quickly.  I reported this problem over two weeks ago, and 
we still haven't made much progress at all.  I hope we can move forward quickly 
on this as I have a brand new server that is powered off right now because we 
can't install the OS.  Let me know how to proceed from here.

-Steve


Comment 5 Need Real Name 2001-10-16 20:24:59 UTC
I just talked to my boss and he reiterated that this is urgent.  We would be 
willing to FedEx the problematic PC to you so you could solve this problem 
faster.  Is this an option you would be interested in?


Comment 6 Need Real Name 2001-10-26 02:25:59 UTC
Aha!  The SCSI stuff was just a red herring.  The real problem was the ASUS 
CUV4X-E motherboard.  I set the System/SDRAM Frequency Ratio in the BIOS from 
1/1 to 4/3.  This changes the SDRAM bus speed from 133MHz to 100MHz.  It still 
sucks that I have to run at 100MHz, but hey, at least it works now.  This is a 
duplicate of Bug #37654 and I am marking at as such.  That bug is marked as 
closed, but what bothers me is that Windows 98 installed and ran just fine on 
this system with the memory at 133MHz.  That suggests to me that the problem 
isn't entirely the motherboard's fault.


*** This bug has been marked as a duplicate of 37654 ***