RE: Re: [PATCH 0/2] passthrough: make management of PCI D-states by guest optional

From: "Cui, Dexuan" <dexuan.cui@intel.com>
To: Yuji Shimada <shimada-yxb@necst.nec.co.jp>,
	Kouya Shimura <kouya@jp.fujitsu.com>
Cc: Ian,
	"xen-devel@lists.xensource.com" <xen-devel@lists.xensource.com>,
	Jackson <Ian.Jackson@eu.citrix.com>
Subject: RE: Re: [PATCH 0/2] passthrough: make management of PCI D-states by guest optional
Date: Fri, 6 Mar 2009 21:41:50 +0800	[thread overview]
Message-ID: <EADF0A36011179459010BDF5142A457507A4DD14@pdsmsx502.ccr.corp.intel.com> (raw)
In-Reply-To: <20090306163623.6E2E.27C06F64@necst.nec.co.jp>

Yuji Shimada wrote:
> On Fri, 6 Mar 2009 14:23:20 +0900
> Kouya Shimura <kouya@jp.fujitsu.com> wrote:
> >
>> Using D3hot state of PCI devices in xen is not mature yet.
>> I met domain destruction/creation troubles in some PCI cards since
>> the PCI config registers can't be read with D3hot state. xend.log is
>> attached. 
>>
>> This patch set makes the management of PCI D-states by guest optional.
>> 
>> The default is "pci_power_mgmt=0" which disables the guest OS from
>> managing D-states because it would be better to avoid the trouble than
>> advantage of low power consumption.
Is this the only reason to add the guest config parameter?
If yes, I don't think it is worthwhile to have two over-200-lines patches.
IMO we should fix the code in xend.

> 
> xend.log shows error occurred on resetting the device.
> I think programming the device to D0 state is needed before resetting
> the device.
I think so, too. 

> Another approach is that qemu programs the device to D0 state on
> shutdowning of guest domain, instead of booting of guest domain.
> Then, xend can reset device successfully.
I think it's not good to do this in ioemu. One reason is: ioemu doesn't always know when a guest is being shut down.e.g., we may "xm destroy" a guest.
xend should be a better place.

Thanks,
-- Dexuan