From mboxrd@z Thu Jan  1 00:00:00 1970
From: erik quanstrom <quanstro@quanstro.net>
Date: Tue,  5 Nov 2013 14:17:09 -0500
To: 9fans@9fans.net
Message-ID: <c4247651917d82ac9bc32f772422171f@mikro>
In-Reply-To: <117b4d1db2043054bf1dd9a2abe7e410@gmx.de>
References: <117b4d1db2043054bf1dd9a2abe7e410@gmx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Subject: Re: [9fans] rb kernel issue
Topicbox-Message-UUID: 89bd73be-ead8-11e9-9d60-3106f5b1d025

On Tue Nov  5 13:06:06 EST 2013, cinap_lenrek@gmx.de wrote:
> how is this possible? execing a zero byte file should fail
> header parsing, which is portable code in sysexec(). can
> you reproduce the panic?
>
> i get this on a pc:
>
> term% ./xxx
> ./xxx: exec header invalid

this is not what happened.  what happened was
/sys/src/cmd/vl/v.out was a valid mips executable.
unfortuntely, vl creates/truncates (i haven't looked yet)
v.out  but when the name of vl is actually v.out,
the process nukes its own text.  so the next page fault in text/data
should be unfixable.  this should terminate the running
process, not panic the kernel.

- erik