qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Andreas Färber" <afaerber@suse.de>
To: Paolo Bonzini <pbonzini@redhat.com>,
	qemu-devel@nongnu.org, Chen Fan <chen.fan.fnst@cn.fujitsu.com>
Cc: ehabkost@redhat.com, hutao@cn.fujitsu.com, mtosatti@redhat.com,
	imammedo@redhat.com, a.motakis@virtualopensystems.com,
	gaowanlong@cn.fujitsu.com
Subject: Re: [Qemu-devel] [PATCH 2.1 00/28] Current state of NUMA series, and hostmem improvements
Date: Wed, 05 Mar 2014 12:05:06 +0100	[thread overview]
Message-ID: <531704E2.3010406@suse.de> (raw)
In-Reply-To: <1393941656-29068-1-git-send-email-pbonzini@redhat.com>

Am 04.03.2014 15:00, schrieb Paolo Bonzini:
> This series includes all the pending work on QOMifying the memory
> backends.
[snip]

There's also a recent RFC from Chen Fan about how to model the
association between NUMA nodes and CPU socket/core/thread that
would/should influence this series if we're aiming for 2.1 now.

I didn't review it in-depth yet, but minor technical issues apart, I
think we need to keep NUMA and CPU separate, which then brings up the
question Chen Fan asked about whether we need to support splitting CPU
threads of one core or CPU cores of one socket onto different NUMA
nodes. If we can stop supporting this, 2.0 would be a good point in time
to catch this with an error message at least, even if the remodeling
depending on it happens post-2.0.

For example, assuming associativity at CPU socket level, I'd imagine:

/machine
  /numa-node[0]
    cpu[0] -> /machine/unassigned/device[0]
  /unassigned
    /device[0]
      /core[0]
        /thread[0]

I.e. the CPU socket is a self-contained object in /machine/unassigned
when machine-created or in /machine/peripheral when device_add'ed.

Hotplug will need a link<> property somewhere, and for a standard PC
this can be on /machine (Qseven modules, SoCs etc. would require it a
level down on their container but we don't support those yet). Having
the link<cpu>s on the NUMA nodes corresponds to having multiple buses in
the qdev world.

Compare that to:

/machine
  /node[0] # This is not really telling!
    /socket[0]
      /core[0]
        /thread[0] # So CPUState != thread?
          cpu -> /machine/unassigned/device[0]
  /unassigned
    /device[0]

Note that according to my interpretation of QOM ABI stability rules we
can't just turn a link<cpu> into a child<cpu> without renaming, thus
trying to be forward-looking for where we want to go design-wise.

Regards,
Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg

  parent reply	other threads:[~2014-03-05 11:05 UTC|newest]

Thread overview: 70+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-03-04 14:00 [Qemu-devel] [PATCH 2.1 00/28] Current state of NUMA series, and hostmem improvements Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 01/28] NUMA: move numa related code to new file numa.c Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 02/28] NUMA: check if the total numa memory size is equal to ram_size Paolo Bonzini
2014-03-04 17:00   ` Eric Blake
2014-03-04 17:19     ` Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 03/28] NUMA: Add numa_info structure to contain numa nodes info Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 04/28] NUMA: convert -numa option to use OptsVisitor Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 05/28] NUMA: expand MAX_NODES from 64 to 128 Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 06/28] man: improve -numa doc Paolo Bonzini
2014-03-11 18:53   ` Eduardo Habkost
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 07/28] qemu-option: introduce qemu_find_opts_singleton Paolo Bonzini
2014-03-05 10:08   ` Andreas Färber
2014-03-07  2:27   ` Hu Tao
2014-03-11 18:55   ` Eduardo Habkost
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 08/28] vl: convert -m to QemuOpts Paolo Bonzini
2014-03-05 10:06   ` Andreas Färber
2014-03-05 10:31     ` Paolo Bonzini
2014-03-05 15:09     ` Igor Mammedov
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 09/28] vl: redo -object parsing Paolo Bonzini
2014-03-07  2:56   ` Hu Tao
2014-03-07  7:39     ` Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 10/28] qmp: allow object-add completion handler to get canonical path Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 11/28] qmp: improve error reporting for -object and object-add Paolo Bonzini
2014-03-07  3:07   ` Hu Tao
2014-03-07  7:57     ` Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 12/28] pc: pass QEMUMachineInitArgs to pc_memory_init Paolo Bonzini
2014-03-07  3:09   ` Hu Tao
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 13/28] numa: introduce memory_region_allocate_system_memory Paolo Bonzini
2014-03-07  3:18   ` Hu Tao
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 14/28] add memdev backend infrastructure Paolo Bonzini
2014-03-07  3:31   ` Hu Tao
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 15/28] numa: add -numa node, memdev= option Paolo Bonzini
2014-03-04 17:52   ` Eric Blake
2014-03-07  5:33   ` Hu Tao
2014-03-07  7:41     ` Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 16/28] memory: reorganize file-based allocation Paolo Bonzini
2014-03-07  6:09   ` Hu Tao
2014-03-07  6:34     ` Hu Tao
2014-03-07  7:47     ` Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 17/28] memory: move mem_path handling to memory_region_allocate_system_memory Paolo Bonzini
2014-03-11  3:50   ` Hu Tao
2014-03-11  8:03     ` Paolo Bonzini
2014-03-12  2:08       ` Marcelo Tosatti
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 18/28] memory: add error propagation to file-based RAM allocation Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 19/28] memory: move preallocation code out of exec.c Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 20/28] memory: move RAM_PREALLOC_MASK to exec.c, rename Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 21/28] hostmem: add file-based HostMemoryBackend Paolo Bonzini
2014-03-04 17:38   ` Eric Blake
2014-03-04 18:12     ` Paolo Bonzini
2014-03-07  6:57   ` Hu Tao
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 22/28] hostmem: separate allocation from UserCreatable complete method Paolo Bonzini
2014-03-07  7:08   ` Hu Tao
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 23/28] hostmem: add merge and dump properties Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 24/28] hostmem: allow preallocation of any memory region Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 25/28] hostmem: add property to map memory with MAP_SHARED Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 26/28] configure: add Linux libnuma detection Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 27/28] hostmem: add properties for NUMA memory policy Paolo Bonzini
2014-03-04 14:00 ` [Qemu-devel] [PATCH 2.1 28/28] qmp: add query-memdev Paolo Bonzini
2014-03-04 17:37   ` Eric Blake
2014-03-04 18:11     ` Paolo Bonzini
2014-03-05  3:50       ` Hu Tao
2014-03-05  8:17         ` Paolo Bonzini
2014-03-05  3:48   ` Hu Tao
2014-03-05 11:05 ` Andreas Färber [this message]
2014-03-05 11:30   ` [Qemu-devel] [PATCH 2.1 00/28] Current state of NUMA series, and hostmem improvements Paolo Bonzini
2014-03-07 11:59     ` Andreas Färber
2014-03-07 12:20       ` Paolo Bonzini
2014-03-07 12:56         ` Igor Mammedov
2014-03-07 13:35           ` Paolo Bonzini
2014-03-07 14:54             ` Igor Mammedov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=531704E2.3010406@suse.de \
    --to=afaerber@suse.de \
    --cc=a.motakis@virtualopensystems.com \
    --cc=chen.fan.fnst@cn.fujitsu.com \
    --cc=ehabkost@redhat.com \
    --cc=gaowanlong@cn.fujitsu.com \
    --cc=hutao@cn.fujitsu.com \
    --cc=imammedo@redhat.com \
    --cc=mtosatti@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).