amd64/159222: unusual behavior writing boot0 from single user mode

[Andriy Gapon]

on 27/07/2011 09:22 tim newsham said the following:
>> Number:         159222
>> Category:       amd64
>> Synopsis:       unusual behavior writing boot0 from single user mode
>> Confidential:   no
>> Severity:       serious
>> Priority:       low
>> Responsible:    freebsd-amd64
>> State:          open
>> Quarter:        
>> Keywords:       
>> Date-Required:
>> Class:          sw-bug
>> Submitter-Id:   current-users
>> Arrival-Date:   Wed Jul 27 06:30:08 UTC 2011
>> Closed-Date:
>> Last-Modified:
>> Originator:     tim newsham
>> Release:        8.2
>> Organization:
>> Environment:
> FreeBSD hpsux.x0d99.com 8.2-RELEASE FreeBSD 8.2-RELEASE #0: Thu Feb 17 02:41:51 UTC 2011     root at mason.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC  amd64
> 
>> Description:
> I boot single user, then I run "df" to see which disk is root.  Then I run:
> 
>    # fdisk -B -b /boot/boot0 /dev/da0
> 
> when its done it says "/boot/boot0: Device not configured" (error varies)
> and then afterwards most commands I type (like "ls") ellicit: "vnode_pager_getpages: I/O read error" from the kernel.
> 
> I get the same behavior on my real machine as well as in vmware.  The vmware has a clean 8.2 release installed off of DVD (8.2 release disk) and then updated with "freebsd-update".
> 
>> How-To-Repeat:
> Install 8.2 release for amd64.
> Update with freebsd-update (I don't know if this matters).
> Boot single user.
> Run "fdisk -B -b /boot/boot0 /dev/da0" (or whatever the root disk drive is).
> Run "ls" after it completes.
>> Fix:

Picking on you at random, nothing personal.
Have you verified that this problem is amd64 specific?
I do not see anything in your report about trying i386.
Why then did you pick amd64 as a category?
If you don't know for sure that an issue is specific to some architecture, but the
issue seems to be in kernel, then you can use just kern category.
I wish that more users actually read this article:
http://www.freebsd.org/doc/en/articles/problem-reports/article.html
Section 4.4 provides quite clear guidelines on filling out the PR template.

Among other things it says:
If you are convinced that the problem will *only* [*] occur under the processor
architecture you are using, select one of the architecture-specific categories:
commonly i386 for Intel-compatible machines in 32-bit mode; amd64 for AMD machines
running in 64-bit mode (this also includes Intel-compatible machines running in
EMT64 mode); and less commonly arm, ia64, powerpc, and sparc64.

[*] - emphasis is mine

Maybe we should prominently feature a link to this article in send-pr and in the
web interface.

-- 
Andriy Gapon