Bug 166381 - Booting from CD #1 with 64 bit media with linux dd does not assign usb floppy
Summary: Booting from CD #1 with 64 bit media with linux dd does not assign usb floppy
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Red Hat Enterprise Linux 3
Classification: Red Hat
Component: kernel
Version: 3.0
Hardware: x86_64
OS: Linux
medium
medium
Target Milestone: ---
Assignee: Pete Zaitcev
QA Contact: Brian Brock
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2005-08-19 22:43 UTC by Alex Bruno
Modified: 2007-11-30 22:07 UTC (History)
3 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2005-09-17 04:45:04 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)
Screenshot capture of error (2.25 MB, application/octet-stream)
2005-08-24 14:42 UTC, Alex Bruno
no flags Details

Description Alex Bruno 2005-08-19 22:43:29 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686 (x86_64); en-US; rv:1.7.10) Gecko/20050719 Red Hat/1.0.6-1.4.1 Firefox/1.0.6

Description of problem:
Customer boots with CD #1 of RHEL 3 U5 media for 64 bit.  Chooses the 

linux dd

option becuase they have to provide a driver.  When it gets to the screen asking the user to insert the disk, it says to install it on

/dev/hda

Placing the diskette in the usb drive accomplishes nothing, not found, address is not being assigned/accepted and the error code seen on the other virtual terminals is error code -110

Version-Release number of selected component (if applicable):
RHEL 3 AS U5 for x86_64

How reproducible:
Always

Steps to Reproduce:
1.boot with RHEL 3 U5 64 bit media and have usb floppy
2.choose linux dd
3.when at screen asking for disk insertion in /dev/hda insert disk...nothing happens
  

Actual Results:  According to other console output, device not assigned address, error code -110

Expected Results:  Should have recognized the usb floppy as a different designation (NOT /dev/hda) and should have been able to have been assigned an address for the device.

Additional info:

This does NOT happen with the 32 bit versions of the media.  64 bit only.

Comment 1 Jeremy Katz 2005-08-22 17:20:13 UTC
The installer code is exactly the same.  The exact error message from tty4 will
probably be helpful in tracking down the bug (which is probably usb
driver/usb-storage)

Comment 2 Alex Bruno 2005-08-24 14:41:07 UTC
Customer reports the error output:

Last four lines of console 4 messages are as follows: 
 
<6>hub.c: new USB device 00:03.1-2, assigned address 7 
<4>usb_control/bulk_msg: timeout 
<3>usb-ohci.c: unlink URB timeout 
<3>usb.c: USB device not accepting new address=7 (error=-110) 
 
I have also attached a screen print of console 4.

Comment 3 Alex Bruno 2005-08-24 14:42:43 UTC
Created attachment 118061 [details]
Screenshot capture of error

Comment 4 Alex Bruno 2005-08-31 12:51:21 UTC
Customer comments:

I found an article from IBM that helped get this going.  At the install prompt,
I typed "linux dd acpi=noirq".  This then asked me for a driver disk and gave me
hda and sda as an option.  I selected sda and away it went.

http://www-1.ibm.com/support/docview.wss?uid=psg1MIGR-59187 
 
This article has pretty detailed installation instructions for the hardware and
platform combination.

Comment 5 Neil Horman 2005-08-31 16:01:37 UTC
Where did they get the idea that /dev/hda was going to be the device for the USB
attached floppy drive?  USB storage is always exported as a scsi device
(/dev/sd<x>).  /dev/hda is an IDE device.  From reading the instructions you
provide above from IBM's web site, it seems as though thats pretty clear.

Comment 6 Pete Zaitcev 2005-09-17 04:43:41 UTC
Ugh, Neil, they are trying to install on /dev/hda, not read drivers off it :-)

The -110 is wrong interrupt routing. 99% of cases it's broken BIOS.

The bug looks like WONTFIX material to me. I think that playing with
BIOS table parsing in -BOOT kernel of RHEL 3 is so risky as to be crazy.
So the only viable solution is to use a trick like "acpi=noirq" to get
the system installed and then perhaps SMP kernel would work automatically.
If not, some command line trick like "noacpi" or "acpi=noirq" is called for.
That's about all I can offer.


Note You need to log in before you can comment on or make changes to this bug.