qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: David Hildenbrand <david@redhat.com>
To: qemu-devel@nongnu.org
Cc: "Igor Mammedov" <imammedo@redhat.com>,
	"Michael S . Tsirkin" <mst@redhat.com>,
	"Paolo Bonzini" <pbonzini@redhat.com>,
	"David Hildenbrand" <david@redhat.com>,
	"Chenyi Qiang" <chenyi.qiang@intel.com>,
	"Michal Privoznik" <mprivozn@redhat.com>,
	"Philippe Mathieu-Daudé" <philmd@linaro.org>
Subject: [GIT PULL 4/4] hostmem: Honor multiple preferred nodes if possible
Date: Mon,  2 Jan 2023 12:29:21 +0100	[thread overview]
Message-ID: <20230102112921.68077-5-david@redhat.com> (raw)
In-Reply-To: <20230102112921.68077-1-david@redhat.com>

From: Michal Privoznik <mprivozn@redhat.com>

If a memory-backend is configured with mode
HOST_MEM_POLICY_PREFERRED then
host_memory_backend_memory_complete() calls mbind() as:

  mbind(..., MPOL_PREFERRED, nodemask, ...);

Here, 'nodemask' is a bitmap of host NUMA nodes and corresponds
to the .host-nodes attribute. Therefore, there can be multiple
nodes specified. However, the documentation to MPOL_PREFERRED
says:

  MPOL_PREFERRED
    This mode sets the preferred node for allocation. ...
    If nodemask specifies more than one node ID, the first node
    in the mask will be selected as the preferred node.

Therefore, only the first node is honored and the rest is
silently ignored. Well, with recent changes to the kernel and
numactl we can do better.

The Linux kernel added in v5.15 via commit cfcaa66f8032
("mm/hugetlb: add support for mempolicy MPOL_PREFERRED_MANY")
support for MPOL_PREFERRED_MANY, which accepts multiple preferred
NUMA nodes instead.

Then, numa_has_preferred_many() API was introduced to numactl
(v2.0.15~26) allowing applications to query kernel support.

Wiring this all together, we can pass MPOL_PREFERRED_MANY to the
mbind() call instead and stop ignoring multiple nodes, silently.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Message-Id: <a0b4adce1af5bd2344c2218eb4a04b3ff7bcfdb4.1671097918.git.mprivozn@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
---
 backends/hostmem.c | 19 +++++++++++++++++--
 meson.build        |  5 +++++
 2 files changed, 22 insertions(+), 2 deletions(-)

diff --git a/backends/hostmem.c b/backends/hostmem.c
index 8640294c10..747e7838c0 100644
--- a/backends/hostmem.c
+++ b/backends/hostmem.c
@@ -23,7 +23,12 @@
 
 #ifdef CONFIG_NUMA
 #include <numaif.h>
+#include <numa.h>
 QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_DEFAULT != MPOL_DEFAULT);
+/*
+ * HOST_MEM_POLICY_PREFERRED may either translate to MPOL_PREFERRED or
+ * MPOL_PREFERRED_MANY, see comments further below.
+ */
 QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_PREFERRED != MPOL_PREFERRED);
 QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_BIND != MPOL_BIND);
 QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_INTERLEAVE != MPOL_INTERLEAVE);
@@ -346,6 +351,7 @@ host_memory_backend_memory_complete(UserCreatable *uc, Error **errp)
          * before mbind(). note: MPOL_MF_STRICT is ignored on hugepages so
          * this doesn't catch hugepage case. */
         unsigned flags = MPOL_MF_STRICT | MPOL_MF_MOVE;
+        int mode = backend->policy;
 
         /* check for invalid host-nodes and policies and give more verbose
          * error messages than mbind(). */
@@ -369,9 +375,18 @@ host_memory_backend_memory_complete(UserCreatable *uc, Error **errp)
                BITS_TO_LONGS(MAX_NODES + 1) * sizeof(unsigned long));
         assert(maxnode <= MAX_NODES);
 
+#ifdef HAVE_NUMA_HAS_PREFERRED_MANY
+        if (mode == MPOL_PREFERRED && numa_has_preferred_many() > 0) {
+            /*
+             * Replace with MPOL_PREFERRED_MANY otherwise the mbind() below
+             * silently picks the first node.
+             */
+            mode = MPOL_PREFERRED_MANY;
+        }
+#endif
+
         if (maxnode &&
-            mbind(ptr, sz, backend->policy, backend->host_nodes, maxnode + 1,
-                  flags)) {
+            mbind(ptr, sz, mode, backend->host_nodes, maxnode + 1, flags)) {
             if (backend->policy != MPOL_DEFAULT || errno != ENOSYS) {
                 error_setg_errno(errp, errno,
                                  "cannot bind memory to host NUMA nodes");
diff --git a/meson.build b/meson.build
index 4c6f8a674a..3f31db5963 100644
--- a/meson.build
+++ b/meson.build
@@ -1858,6 +1858,11 @@ config_host_data.set('CONFIG_LINUX_AIO', libaio.found())
 config_host_data.set('CONFIG_LINUX_IO_URING', linux_io_uring.found())
 config_host_data.set('CONFIG_LIBPMEM', libpmem.found())
 config_host_data.set('CONFIG_NUMA', numa.found())
+if numa.found()
+  config_host_data.set('HAVE_NUMA_HAS_PREFERRED_MANY',
+                       cc.has_function('numa_has_preferred_many',
+                                       dependencies: numa))
+endif
 config_host_data.set('CONFIG_OPENGL', opengl.found())
 config_host_data.set('CONFIG_PROFILER', get_option('profiler'))
 config_host_data.set('CONFIG_RBD', rbd.found())
-- 
2.39.0



  parent reply	other threads:[~2023-01-02 11:31 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-02 11:29 [GIT PULL 0/4] Host Memory Backends and Memory devices queue 2023-01-02 David Hildenbrand
2023-01-02 11:29 ` [GIT PULL 1/4] virtio-mem: Fix the bitmap index of the section offset David Hildenbrand
2023-01-02 11:29 ` [GIT PULL 2/4] virtio-mem: Fix the iterator variable in a vmem->rdl_list loop David Hildenbrand
2023-01-02 11:29 ` [GIT PULL 3/4] virtio-mem: Fix typo in function name David Hildenbrand
2023-01-02 11:29 ` David Hildenbrand [this message]
2023-01-05 16:58 ` [GIT PULL 0/4] Host Memory Backends and Memory devices queue 2023-01-02 Peter Maydell

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=20230102112921.68077-5-david@redhat.com \
    --to=david@redhat.com \
    --cc=chenyi.qiang@intel.com \
    --cc=imammedo@redhat.com \
    --cc=mprivozn@redhat.com \
    --cc=mst@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=philmd@linaro.org \
    --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).