[PATCH 0/3] GDT virtualization performance

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

 



(Note... reply list substantially reduced)

What is the status for getting virtualization preparatory
changes into the kernel?  Approx. how many have/are going
into which releases? At what release is it expected that
Linux for Xen (or the Vmware PVI) will be buildable in
a released kernel?

I have a virtualization prep patch for Linux/ia64 (entirely
contained in archdep for ia64) and have been getting
push-back from Tony (the Linux/ia64 maintainer) that he
isn't in any hurry to apply it in part because the x86 stuff
isn't going in yet.

So I'd appreciate any info.

Thanks,
Dan

> -----Original Message-----
> From: virtualization-bounces@xxxxxxxxxxxxxx 
> [mailto:virtualization-bounces@xxxxxxxxxxxxxx] On Behalf Of 
> Zachary Amsden
> Sent: Thursday, September 22, 2005 1:35 AM
> To: Linus Torvalds; Jeffrey Sheldon; Ole Agesen; Shai 
> Fultheim; Andrew Morton; Jack Lo; Ingo Molnar; Linux Kernel 
> Mailing List; Virtualization Mailing List; Chris Wright; 
> Martin Bligh; Pratap Subrahmanyam; Christopher Li; H. Peter 
> Anvin; Zwane Mwaikambo; Andi Kleen; Zachary Amsden
> Subject: [PATCH 0/3] GDT virtualization performance
> 
> Three patches to clean up GDT access in Linux to make it friendly to
> virtualization environments.  The basic problem is that the GDT must
> be write protected, which causes spurious overhead when the GDT lies
> on the same page as other data.  This problem exists both for VMware
> and Xen; Xen actually requires page isolation, so we have implemented
> the most general and compatible solution.
> 
> Patch 1 deprecates a broken GDT reference;
> Patch 2 adds a per-cpu GDT accessor;
> Patch 3 moves the GDT out of the per-cpu area and makes it page
>  padded and page aligned.
> 
> The GDTs for secondary processors are allocated dynamically to avoid
> bloating kernel static data with GDTs for not-present processors.
> 
> This could be adapted to drop and reallocate GDTs for CPU hotplug
> if desired, although the space savings (1 page) are dubious, so
> I have not implemented that at this time.
> 
> Testing: 4 way SMP boot-halts, kernel compiles, stress.
> 
> Zachary Amsden <zach@xxxxxxxxxx>
> 


[Index of Archives]     [KVM Development]     [Libvirt Development]     [Libvirt Users]     [CentOS Virtualization]     [Netdev]     [Ethernet Bridging]     [Linux Wireless]     [Kernel Newbies]     [Security]     [Linux for Hams]     [Netfilter]     [Bugtraq]     [Yosemite Forum]     [MIPS Linux]     [ARM Linux]     [Linux RAID]     [Linux Admin]     [Samba]

  Powered by Linux