Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 20 Mar 2009 16:02:40 -0700
From:      Marcel Moolenaar <xcllnt@mac.com>
To:        Ivan Voras <ivoras@freebsd.org>
Cc:        Jeremy Lea <reg@freebsd.org>, svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org, Marcel Moolenaar <marcel@freebsd.org>
Subject:   Re: svn commit: r190182 - head
Message-ID:  <8C44C8B6-FBBD-48CC-BAE0-5EA2EAC84609@mac.com>
In-Reply-To: <9bbcef730903201554x773f46bdy890c8ab48760dce2@mail.gmail.com>
References:  <200903202151.n2KLpRMB014702@svn.freebsd.org> <20090320221211.GA5205@flint.openpave.org> <9bbcef730903201554x773f46bdy890c8ab48760dce2@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help

On Mar 20, 2009, at 3:54 PM, Ivan Voras wrote:

> 2009/3/20 Jeremy Lea <reg@freebsd.org>:
>> Hi,
>>
>> On Fri, Mar 20, 2009 at 09:51:27PM +0000, Ivan Voras wrote:
>>> +20090320:
>>> +     GEOM_PART has become the default partition slicer for  
>>> storage devices,
>>> +     replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT  
>>> slicers. It
>>> +     introduces some changes: the devices created from MSDOS  
>>> extended
>>> +     partition entries (EBR) can be named differently than with  
>>> GEOM_MBR
>>> +     and are now symlinks to devices with offset-based names, and  
>>> kernel
>>> +     dumps are now not allowed to be written to devices whose  
>>> partition
>>> +     types indicate they are meant to be used for file systems.
>>> +
>>
>> This should probably also note the issue with stale disklabels from  
>> old
>> DD mode disks, which causes the machine to be unbootable.  It should
>> also mention the "geometry does not match label (255h,63s != 16h, 
>> 63s)"
>> message (which appears to be harmless), and it should also probably
>> mention that the c partition now covers the entire disk, not just the
>> "DOS partition"...  And it should be noted that these changes took  
>> place
>> in Decemeber, so that people have a better chance of knowing if  
>> they are
>> going to be bitten.
>
> Ok. Should I re-sort the note into December (probably not since it
> mentions development from a large period) or just mention it in text?

I would mention it in text.

> @Marcel: is the warning important (I didn't encounter it yet; I guess
> it probably a quirk of UFS?)

The warning could be important. It also makes us aware that
there's a discrepancy, which indicates that we have bugs.

> and is the "c" change intended?

The only change is that the 'c' partition cannot be used directly
anymore. If you want the whole disk, you use the provider. The
meaning of 'c' didn't change.

> (isn't it
> time to abandon the special case of "c"?)

I chose to hide 'c' rather than give it a new meaning. The
reason being that disklabel/bsdlabel still assumes the old
meaning. By hiding the 'c' partition we avoid that people
abuse the 'c' partition. There's no need to abuse 'c'
anymore because the partition table can hold 20 partitions
instead of 8.

-- 
Marcel Moolenaar
xcllnt@mac.com






Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?8C44C8B6-FBBD-48CC-BAE0-5EA2EAC84609>