Re: [bug #48685] update-grub triggers filesystem server panics

[Justus Winter]

"Brent W. Baccala" <cosine@freesoft.org> writes:

> On Tue, Aug 2, 2016 at 9:52 PM, Justus Winter <INVALID.NOREPLY@gnu.org>
> wrote:
>
>>
>> The isofs crash is fixed in 21ee0e79.  Please update your Hurd packages.
>>
>> The fact that you are seeing these messages is a (quite nice actually)
>> side-effect of 234568d0.  I guess that if you were doing this on a Linux
>> machine, you would see these in the kernel ring buffer.
>>
>>
> OK, I updated my Hurd system using the Debian ports repository.  The
> package is hurd-0.8.git20160522, and it includes 21ee0e79.
>
> Sure enough, the crash is gone, but I still get the error messages:

Good.

> ext2fs: /dev/hd0s2: panic: main: device too small for superblock (1024
> bytes)
> /hurd/fatfs: Invalid number of bytes per sector
> /hurd/iso9660fs: Could not find valid superblock
>
> Are you suggesting that these are to be expected?

Yes.

> Is this a bug in update-grub?

No.

> These messages seem awfully severe for a standard system operation.

And as I demonstrated, the very same kind of messages are generated on a
Linux machine.  The filesystem servers open the Mach console and write
the messages there.  You just happen to work on a console where those
messages are relayed.  If you were to do update-grub from an xterm, you
wouldn't see these.

Cheers,
Justus

signature.asc
Description: PGP signature