[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH 3/5] RTC: Add UIP(update in progress) check logic

>>> On 06.03.12 at 01:26, "Zhang, Yang Z" <yang.z.zhang@xxxxxxxxx> wrote:
> Actually, in the draft patch, I also think it is unnecessary. And I don't 
> find any real use mode to use RTC in this way.
> But someone points out that the RTC datasheet allows us to check the update 
> cycle by UIP. So we need to support it.

Doesn't Xen itself make use of this (and would hence be affected [non-
fatally] when running nested)? See xen/arch/x86/time.c:get_cmos_time().


>> -----Original Message-----
>> From: Ian Campbell [mailto:Ian.Campbell@xxxxxxxxxx]
>> Sent: Tuesday, March 06, 2012 2:36 AM
>> To: Zhang, Yang Z
>> Cc: xen-devel@xxxxxxxxxxxxxxxxxxx; Jan Beulich
>> Subject: Re: [Xen-devel] [PATCH 3/5] RTC: Add UIP(update in progress) check
>> logic
>> On Mon, 2012-03-05 at 03:25 -0500, Zhang, Yang Z wrote:
>> > The UIP(update in progress) is set when RTC is in updating. And the
>> > update cycle begins 244us later after UIP is set.
>> Not arguing against this patch but OOI how important is it to emulate this
>> behaviour to this level of accuracy? On real hardware the UIP tells you that 
> the
>> data may not be valid/consistent but I don't think the emulated RTC has that
>> property (does it?). Do guest OSes rely on seeing the UIP bit occasionally 
> or have
>> you seen a specific issue which caused you to make this change?
>> Ian.

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.