[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Xen-devel] [PATCH] tools: hvmloader: attempt to SHUTDOWN_crash on BUG
- To: Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>, Ian Campbell <Ian.Campbell@xxxxxxxxxxxxx>
- From: Keir Fraser <keir.xen@xxxxxxxxx>
- Date: Wed, 20 Apr 2011 20:18:30 +0100
- Cc: "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>
- Delivery-date: Wed, 20 Apr 2011 12:19:12 -0700
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:cc:message-id :thread-topic:thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; bh=0rXAgSDItWU0obAXTkCb2/HIRMsi/dXs8WdZ+JcN06U=; b=WDBo4bAoSqN1QpmjuwelXTn0+F5neXH/fmWBS6h3a5luO2+BByslD6s66Ctu3bWQI9 THreWtalE3pxdiwl9hLgsWHqm4SDI15C3QAILnVObUtYCSMJs5/0oXap3VMxwHmFxnYr 3zb7kDO/cd+aMfq9j6+U2kgl2skHu64AE0/5U=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:cc:message-id:thread-topic :thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; b=pmEALCBfxnk/5EDJL3kzB8Ewj09VMgPqJMx8gBfnjv0GQgFHGdrWyTuvpvFZC5jbf/ SnYRtK1HhAVO7T8KxHRtFPnWHmi3REe3Hssk/cWnm2WPl4YgorUtB4JYZ9v9toNua9PB uPwk874b8cuAiPFlXbWOQTmCXiKemu/X74j7M=
- List-id: Xen developer discussion <xen-devel.lists.xensource.com>
- Thread-index: Acv/j7trrySATtncgkSuXqEmfNHzxA==
- Thread-topic: [Xen-devel] [PATCH] tools: hvmloader: attempt to SHUTDOWN_crash on BUG
On 20/04/2011 19:35, "Ian Jackson" <Ian.Jackson@xxxxxxxxxxxxx> wrote:
> Ian Campbell writes ("Re: [Xen-devel] [PATCH] tools: hvmloader: attempt to
> SHUTDOWN_crash on BUG"):
>> The toolstack can't tell the difference between hvmloader and the
>> subsequent bootloader then OS doing stuff and they may not be PV'd up.
>
> Indeed. But hvmloader can put the reason code back just before it
> launches the guest bootloader (or perhaps just before it enters the
> BIOS).
Currently the reason code cannot be changed after it has been latched by the
hypervisor. Secondly, even if it could be changed, we wouldn't want to
change it to a new value after hvmloader has run; we'd want to reset it to
no-code. That would be cleanest with a new sched_op. However it is arguable
whether this is worthwhile, since I've never heard of hvmloader crashing
except on an explicit bug/assert statement (and even those are rare, and
during development work). So it's kind of a non problem.
Also I already applied Ian's patch. :-) (I still maintain tools/firmware/,
at least until seabios goes in).
-- Keir
> Ian.
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-devel
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel