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

[Xen-devel] why each process needs a separate page for kernel's pgd table in PV x86_64?


  • To: xen-devel <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: weiming <zephyr.zhao@xxxxxxxxx>
  • Date: Mon, 24 Mar 2008 00:07:39 -0400
  • Delivery-date: Sun, 23 Mar 2008 21:08:04 -0700
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=message-id:date:from:to:subject:mime-version:content-type; b=M6zYNmYC/q7+DZd4u+tFHeMZAtErxRNoP+winVaFn++uvCCHCmoqBUQUtPFrfSJH3jukWjlop46Sg9OalERlyPFt1UxZZX4+YL1oUlgivvs9jA+w4VIxpxGnGkQjyOo5rw2xM5GtE6pe7ga0MBoAKhbGqEwllUFjTk8z1MgpBZY=
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>

Hi,

I notice that in the paravirtualized linux for x86_64, each process allocates 2 pages for mm->pgd: one for kernel's page table, the other for user's page table.

Why not let all processes share the kernel's page table? i.e., add a field in mm struct (like mm->kernel_pgd, ) and let it point to a page, which will be referenced by all process. This can save some 4K for each process forked.

Please correct me if I'm wrong.

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

 


Rackspace

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