Re: Consistent kernel panic on volume mount
Re: Consistent kernel panic on volume mount
- Subject: Re: Consistent kernel panic on volume mount
- From: Carsten G Pedersen <email@hidden>
- Date: Wed, 19 Jan 2005 21:59:30 +0100
William Kucharski <email@hidden> wrote on
01/19/2005 05:23:07 PM:
> On Jan 19, 2005, at 2:20 AM, Carsten G Pedersen wrote:
>
> The bigger issue is that there are all kinds of file system errors
that
> will cause the kernel to panic, and many of them are only fixable
by
> AlSoft's DiskWarrior, if at all.
>
Well it's certainly thought provoking that vfs_journal.c, which seems to
be the culprit for me, contains nearly 60 panic calls.Surely some of these
could and should be substituted by more sophisticated ways of bailing out.
Also studying the exact lines of code that triggers my panic makes you
wonder a bit:
if (overlap > 0) {
if (overlap % jhdr_size
!= 0) {
panic("jnl:
do_overlap: overlap of %d is not multiple of %d\n", overlap, jhdr_size);
}
Getting the message "overlap of 0 is not multiple of 24" from
this code doesn't really make sense - my best guess is that the 64bit integer
overlap can't be correctly displayed by %d, does anybody know whether that
might be so?
And yes Felix - I will take your advice and file a bug - just want to gather
some more information first. I'm seriously considering building a version
of the kernel that stays away from journaling altogether - of course if
someone else know about an easier way to avoid this code being called I'd
be most interested to hear about it.
Carsten
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden