To a broader audience.
Was a question on [EMAIL PROTECTED] about importing kernel
crashdump CTF data while using mdb to read a disk device (image). Since
Phil suggested fsdb and I pushed back ...
Makes a lot of sense from my point of view - what do you folks think ?
---------- Forwarded message ----------
Date: Mon, 29 Jan 2007 12:06:03 +0000 (GMT Standard Time)
From: Frank Hofmann <[EMAIL PROTECTED]>
To: Phill Hughes - PTS Kernel EMEA <[EMAIL PROTECTED]>
Cc: Frank Hofmann <[EMAIL PROTECTED]>, [EMAIL PROTECTED],
[EMAIL PROTECTED], [EMAIL PROTECTED]
Subject: Re: [mdb-discuss] Using kernel CTF with raw disk
On Mon, 29 Jan 2007, Phill Hughes - PTS Kernel EMEA wrote:
[ ... me saying "rubbish" ... ]
I stand well and truly corrected.
To be a little more diplomatic: There's a reason why the UFS group once had a
project running to replace fsdb with a set of "ufs mdb dcmds". These days, it
makes little sense to learn two terse debuggers (mdb and fsdb), so why not
apply the better one to disks just as well ?
FrankB. might know whether this is public somewhere. If not, we'll reinvent it
soon enough, it's just needed. mdb is the best debugger bundled with Solaris,
and except for compatibility with ancient scripts calling it I don't see any
reason why efforts should be put into promoting (or enhancing) fsdb.
But willing to enter a broader discussion:
Let fsdb die in favour of "mdb on-disk fs data access modules" ?
Gets my vote, in any case.
FrankH.
_______________________________________________
ufs-discuss mailing list
[email protected]
