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

RE: [Xen-ia64-devel] [PATCH] bug fix new_tlbflush_clock_period()


  • To: "Isaku Yamahata" <yamahata@xxxxxxxxxxxxx>
  • From: "Xu, Anthony" <anthony.xu@xxxxxxxxx>
  • Date: Tue, 6 Feb 2007 13:15:22 +0800
  • Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Mon, 05 Feb 2007 21:14:43 -0800
  • List-id: Discussion of the ia64 port of Xen <xen-ia64-devel.lists.xensource.com>
  • Thread-index: AcdJqvBvU9u6a0xOS5abl50izwNppQAAIUNg
  • Thread-topic: [Xen-ia64-devel] [PATCH] bug fix new_tlbflush_clock_period()

Isaku Yamahata write on 2007年2月6日 12:55:
> It consumes only several mega bytes (We can assume at least
> several giga bytes on a box.) and the tlb flush clock overflow
> would takes very long time
> The current implementation is intended to allow per-vcpu vhpt
> domain coexist with per-pcpu vhpt domain.

I understand your intention now,

> Currently we can only turn off per-vcpu vhpt system widely with
> "pervcpu_vhpt" boot option.
> The only missing part is how to pass the user configuration to xen
> domain creation and initialize it in per-domain fashion.
> 
> So your proposal is to remove such dynamic checks and
> make the per-vcpu vhpt code compile-time-only configuration.
> Or to allocate per-pcpu vhpt when per-pcph vhpt domain is created
> and to add check if per-pcph vhpt is allocated.
> Do you think it's worth while?

Because tlbflush_clock_local_flush is called unfrequently,
This will not impact overall performance.

I would like to support both per-vcpu and per-pcpu vhpt.


- Anthony

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel


 


Rackspace

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