Re: [Qemu-devel] Performance regression using KVM/ARM

2016-06-13 Thread Auger Eric
Hi, Le 22/04/2016 à 12:15, Christoffer Dall a écrit : > On Fri, Apr 22, 2016 at 12:06:52PM +0200, Alexander Graf wrote: >> On 04/22/2016 12:01 PM, Christoffer Dall wrote: >>> On Thu, Apr 21, 2016 at 09:50:05PM +0200, Alexander Graf wrote: On 21.04.16 18:23, Christoffer Dall wrote: >

Re: [Qemu-devel] Performance regression using KVM/ARM

2016-04-22 Thread Alexander Graf
On 04/22/2016 01:16 PM, Andrew Jones wrote: On Fri, Apr 22, 2016 at 12:26:52PM +0200, Christoffer Dall wrote: On Fri, Apr 22, 2016 at 11:17:47AM +0100, Peter Maydell wrote: On 22 April 2016 at 11:15, Christoffer Dall wrote: Peter just pointed me to a change I remember doing for ARM, so perhap

Re: [Qemu-devel] Performance regression using KVM/ARM

2016-04-22 Thread Andrew Jones
On Fri, Apr 22, 2016 at 12:26:52PM +0200, Christoffer Dall wrote: > On Fri, Apr 22, 2016 at 11:17:47AM +0100, Peter Maydell wrote: > > On 22 April 2016 at 11:15, Christoffer Dall > > wrote: > > > Peter just pointed me to a change I remember doing for ARM, so perhaps > > > this fix is the right on

Re: [Qemu-devel] Performance regression using KVM/ARM

2016-04-22 Thread Christoffer Dall
On Fri, Apr 22, 2016 at 11:17:47AM +0100, Peter Maydell wrote: > On 22 April 2016 at 11:15, Christoffer Dall > wrote: > > Peter just pointed me to a change I remember doing for ARM, so perhaps > > this fix is the right one? > > > > > > diff --git a/util/oslib-posix.c b/util/oslib-posix.c > > inde

Re: [Qemu-devel] Performance regression using KVM/ARM

2016-04-22 Thread Peter Maydell
On 22 April 2016 at 11:15, Christoffer Dall wrote: > Peter just pointed me to a change I remember doing for ARM, so perhaps > this fix is the right one? > > > diff --git a/util/oslib-posix.c b/util/oslib-posix.c > index d25f671..a36e734 100644 > --- a/util/oslib-posix.c > +++ b/util/oslib-posix.c

Re: [Qemu-devel] Performance regression using KVM/ARM

2016-04-22 Thread Christoffer Dall
On Fri, Apr 22, 2016 at 12:06:52PM +0200, Alexander Graf wrote: > On 04/22/2016 12:01 PM, Christoffer Dall wrote: > >On Thu, Apr 21, 2016 at 09:50:05PM +0200, Alexander Graf wrote: > >> > >>On 21.04.16 18:23, Christoffer Dall wrote: > >>>Hi, > >>> > >>>Commit 9fac18f (oslib: allocate PROT_NONE page

Re: [Qemu-devel] Performance regression using KVM/ARM

2016-04-22 Thread Alexander Graf
On 04/22/2016 12:01 PM, Christoffer Dall wrote: On Thu, Apr 21, 2016 at 09:50:05PM +0200, Alexander Graf wrote: On 21.04.16 18:23, Christoffer Dall wrote: Hi, Commit 9fac18f (oslib: allocate PROT_NONE pages on top of RAM, 2015-09-10) had the unfortunate side effect that memory slots registere

Re: [Qemu-devel] Performance regression using KVM/ARM

2016-04-22 Thread Christoffer Dall
Hi Laszlo, On Thu, Apr 21, 2016 at 11:58:07PM +0200, Laszlo Ersek wrote: > On 04/21/16 18:23, Christoffer Dall wrote: > > Hi, > > > > Commit 9fac18f (oslib: allocate PROT_NONE pages on top of RAM, > > 2015-09-10) had the unfortunate side effect that memory slots registered > > with KVM no longer

Re: [Qemu-devel] Performance regression using KVM/ARM

2016-04-22 Thread Christoffer Dall
On Thu, Apr 21, 2016 at 09:50:05PM +0200, Alexander Graf wrote: > > > On 21.04.16 18:23, Christoffer Dall wrote: > > Hi, > > > > Commit 9fac18f (oslib: allocate PROT_NONE pages on top of RAM, > > 2015-09-10) had the unfortunate side effect that memory slots registered > > with KVM no longer cont

Re: [Qemu-devel] Performance regression using KVM/ARM

2016-04-21 Thread Laszlo Ersek
On 04/21/16 18:23, Christoffer Dall wrote: > Hi, > > Commit 9fac18f (oslib: allocate PROT_NONE pages on top of RAM, > 2015-09-10) had the unfortunate side effect that memory slots registered > with KVM no longer contain a userspace address that is aligned to a 2M > boundary, causing the use of THP

Re: [Qemu-devel] Performance regression using KVM/ARM

2016-04-21 Thread Alexander Graf
On 21.04.16 18:23, Christoffer Dall wrote: > Hi, > > Commit 9fac18f (oslib: allocate PROT_NONE pages on top of RAM, > 2015-09-10) had the unfortunate side effect that memory slots registered > with KVM no longer contain a userspace address that is aligned to a 2M > boundary, causing the use of T

[Qemu-devel] Performance regression using KVM/ARM

2016-04-21 Thread Christoffer Dall
Hi, Commit 9fac18f (oslib: allocate PROT_NONE pages on top of RAM, 2015-09-10) had the unfortunate side effect that memory slots registered with KVM no longer contain a userspace address that is aligned to a 2M boundary, causing the use of THP to fail in the kernel. I fail to see where in the QEM