All of lore.kernel.org
 help / color / mirror / Atom feed
* [Qemu-devel] [RFC] Is it useful to create a seperate address space for cpu-memory on x86 platform?
@ 2017-05-26  9:59 Gonglei (Arei)
  2017-05-26 11:06 ` Paolo Bonzini
  0 siblings, 1 reply; 2+ messages in thread
From: Gonglei (Arei) @ 2017-05-26  9:59 UTC (permalink / raw)
  To: Paolo Bonzini, peter.crosthwaite, peter.maydell, edgar.iglesias
  Cc: qemu-devel, Huangweidong (C)

Hi all,

I found that the memory consumption increase about 5M since below commit on Qemu process.

commit 6731d864f80938e404dc3e5eb7f6b76b891e3e43
Author: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Date:   Thu Jan 21 14:15:06 2016 +0000

    qom/cpu: Add MemoryRegion property
    
    Add a MemoryRegion property, which if set is used to construct
    the CPU's initial (default) AddressSpace.
    
    Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
    [PMM: code is moved from qom/cpu.c to exec.c to avoid having to
     make qom/cpu.o be a non-common object file; code to use the
     MemoryRegion and to default it to system_memory added.]
    Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
    Acked-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

The result before this commit:

linux-arei:/mnt/sdb/gonglei/opensource/qemu # ps aux|grep qemu
root     12658 17.2  0.0 375996 21412 pts/1    Sl+  12:53   0:00 ./x86_64-softmmu/qemu-system-x86_64 -machine pc-i440fx-2.5,accel=kvm,usb=off -nodefaults

after the commit:
 
linux-arei:/mnt/sdb/gonglei/opensource/qemu # ps aux|grep qemu
root     23201 17.2  0.0 384196 25736 pts/1    Sl+  12:54   0:00 ./x86_64-softmmu/qemu-system-x86_64 -machine pc-i440fx-2.5,accel=kvm,usb=off -nodefaults

I couldn't find any benefits on x86 platform (I'm not familiar with ARM), am I missing something?

And It works well if I revert this commit. Can I?

Thanks,
-Gonglei

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [Qemu-devel] [RFC] Is it useful to create a seperate address space for cpu-memory on x86 platform?
  2017-05-26  9:59 [Qemu-devel] [RFC] Is it useful to create a seperate address space for cpu-memory on x86 platform? Gonglei (Arei)
@ 2017-05-26 11:06 ` Paolo Bonzini
  0 siblings, 0 replies; 2+ messages in thread
From: Paolo Bonzini @ 2017-05-26 11:06 UTC (permalink / raw)
  To: Gonglei (Arei), peter.crosthwaite, peter.maydell, edgar.iglesias
  Cc: qemu-devel, Huangweidong (C)



On 26/05/2017 11:59, Gonglei (Arei) wrote:
> The result before this commit:
> 
> linux-arei:/mnt/sdb/gonglei/opensource/qemu # ps aux|grep qemu
> root     12658 17.2  0.0 375996 21412 pts/1    Sl+  12:53   0:00 ./x86_64-softmmu/qemu-system-x86_64 -machine pc-i440fx-2.5,accel=kvm,usb=off -nodefaults
> 
> after the commit:
>  
> linux-arei:/mnt/sdb/gonglei/opensource/qemu # ps aux|grep qemu
> root     23201 17.2  0.0 384196 25736 pts/1    Sl+  12:54   0:00 ./x86_64-softmmu/qemu-system-x86_64 -machine pc-i440fx-2.5,accel=kvm,usb=off -nodefaults
> 
> I couldn't find any benefits on x86 platform (I'm not familiar with ARM), am I missing something?

Hi,

this was also reported by Anthony Xu.  I am working on reducing the time
and space cost of multiple address spaces by sharing the FlatView and
AddressSpaceDispatch.

Paolo

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2017-05-26 11:06 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-05-26  9:59 [Qemu-devel] [RFC] Is it useful to create a seperate address space for cpu-memory on x86 platform? Gonglei (Arei)
2017-05-26 11:06 ` Paolo Bonzini

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.