Re: [Qemu-devel] [RFC PATCH] pc: align gpa-hpa on 1GB boundary by splitting RAM on several regions

2013-11-07 Thread Igor Mammedov
On Wed, 30 Oct 2013 21:44:12 -0200 Marcelo Tosatti mtosa...@redhat.com wrote: On Wed, Oct 30, 2013 at 08:56:17PM +0100, Igor Mammedov wrote: On Wed, 30 Oct 2013 16:51:29 -0200 Marcelo Tosatti mtosa...@redhat.com wrote: [...] btw: now if QEMU can't allocate hugepages for whole guest

Re: [Qemu-devel] [RFC PATCH] pc: align gpa-hpa on 1GB boundary by splitting RAM on several regions

2013-10-30 Thread Igor Mammedov
On Tue, 29 Oct 2013 19:38:44 -0200 Marcelo Tosatti mtosa...@redhat.com wrote: On Tue, Oct 29, 2013 at 07:18:49PM +0100, Igor Mammedov wrote: Otherwise 1GB TLBs cannot be cached for the range. This fails to back non-1GB-aligned gpas, but 2MB aligned, with 2MB large pages. With current

Re: [Qemu-devel] [RFC PATCH] pc: align gpa-hpa on 1GB boundary by splitting RAM on several regions

2013-10-30 Thread Marcelo Tosatti
On Wed, Oct 30, 2013 at 05:49:49PM +0100, Igor Mammedov wrote: On Tue, 29 Oct 2013 19:38:44 -0200 Marcelo Tosatti mtosa...@redhat.com wrote: On Tue, Oct 29, 2013 at 07:18:49PM +0100, Igor Mammedov wrote: Otherwise 1GB TLBs cannot be cached for the range. This fails to back

Re: [Qemu-devel] [RFC PATCH] pc: align gpa-hpa on 1GB boundary by splitting RAM on several regions

2013-10-30 Thread Marcelo Tosatti
On Wed, Oct 30, 2013 at 04:51:29PM -0200, Marcelo Tosatti wrote: On Wed, Oct 30, 2013 at 05:49:49PM +0100, Igor Mammedov wrote: On Tue, 29 Oct 2013 19:38:44 -0200 Marcelo Tosatti mtosa...@redhat.com wrote: On Tue, Oct 29, 2013 at 07:18:49PM +0100, Igor Mammedov wrote: Otherwise 1GB

Re: [Qemu-devel] [RFC PATCH] pc: align gpa-hpa on 1GB boundary by splitting RAM on several regions

2013-10-30 Thread Marcelo Tosatti
On Wed, Oct 30, 2013 at 05:49:49PM +0100, Igor Mammedov wrote: On Tue, 29 Oct 2013 19:38:44 -0200 Marcelo Tosatti mtosa...@redhat.com wrote: On Tue, Oct 29, 2013 at 07:18:49PM +0100, Igor Mammedov wrote: Otherwise 1GB TLBs cannot be cached for the range. This fails to back

Re: [Qemu-devel] [RFC PATCH] pc: align gpa-hpa on 1GB boundary by splitting RAM on several regions

2013-10-30 Thread Igor Mammedov
On Wed, 30 Oct 2013 16:51:29 -0200 Marcelo Tosatti mtosa...@redhat.com wrote: On Wed, Oct 30, 2013 at 05:49:49PM +0100, Igor Mammedov wrote: On Tue, 29 Oct 2013 19:38:44 -0200 Marcelo Tosatti mtosa...@redhat.com wrote: On Tue, Oct 29, 2013 at 07:18:49PM +0100, Igor Mammedov wrote:

Re: [Qemu-devel] [RFC PATCH] pc: align gpa-hpa on 1GB boundary by splitting RAM on several regions

2013-10-30 Thread Paolo Bonzini
Il 30/10/2013 20:31, Marcelo Tosatti ha scritto: I do not believe that unability to deal with memory aliases on the future ability to create machine from configuration is strong enough reason to justify pushing mixed 1GB and 2MB pages size to the user (1GB pages are allocated on the host

Re: [Qemu-devel] [RFC PATCH] pc: align gpa-hpa on 1GB boundary by splitting RAM on several regions

2013-10-30 Thread Marcelo Tosatti
On Wed, Oct 30, 2013 at 08:56:17PM +0100, Igor Mammedov wrote: On Wed, 30 Oct 2013 16:51:29 -0200 Marcelo Tosatti mtosa...@redhat.com wrote: On Wed, Oct 30, 2013 at 05:49:49PM +0100, Igor Mammedov wrote: On Tue, 29 Oct 2013 19:38:44 -0200 Marcelo Tosatti mtosa...@redhat.com wrote:

[Qemu-devel] [RFC PATCH] pc: align gpa-hpa on 1GB boundary by splitting RAM on several regions

2013-10-29 Thread Igor Mammedov
Otherwise 1GB TLBs cannot be cached for the range. PS: as side effect we are not wasting ~1Gb of memory if 1Gb hugepages are used and -m hpagesize(in Mb)*n + 1 Signed-off-by: Igor Mammedov imamm...@redhat.com --- PS2: As RFC it's yet without compatibility changes noted by Paolo --- exec.c

Re: [Qemu-devel] [RFC PATCH] pc: align gpa-hpa on 1GB boundary by splitting RAM on several regions

2013-10-29 Thread Marcelo Tosatti
On Tue, Oct 29, 2013 at 07:18:49PM +0100, Igor Mammedov wrote: Otherwise 1GB TLBs cannot be cached for the range. This fails to back non-1GB-aligned gpas, but 2MB aligned, with 2MB large pages. Since hugetlbfs allocation is static, it requires the user to inform different 1GB and 2MB sized