qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: qemu-devel@nongnu.org
Cc: v.maffione@gmail.com
Subject: [Qemu-devel] [PATCH 4/8] memory: avoid unnecessary object_ref/unref
Date: Wed, 16 Dec 2015 11:59:56 +0100	[thread overview]
Message-ID: <1450263601-2828-5-git-send-email-pbonzini@redhat.com> (raw)
In-Reply-To: <1450263601-2828-1-git-send-email-pbonzini@redhat.com>

For the common case of DMA into non-hotplugged RAM, it is unnecessary
but expensive to do object_ref/unref.  Add back an owner field to
MemoryRegion, so that these memory regions can skip the reference
counting.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 include/exec/memory.h |  1 +
 memory.c              | 28 ++++++++++++----------------
 2 files changed, 13 insertions(+), 16 deletions(-)

diff --git a/include/exec/memory.h b/include/exec/memory.h
index 5b1fd12..24b7cba 100644
--- a/include/exec/memory.h
+++ b/include/exec/memory.h
@@ -172,6 +172,7 @@ struct MemoryRegion {
     bool global_locking;
     uint8_t dirty_log_mask;
     ram_addr_t ram_addr;
+    Object *owner;
     const MemoryRegionIOMMUOps *iommu_ops;
 
     const MemoryRegionOps *ops;
diff --git a/memory.c b/memory.c
index c0770a6..1783300 100644
--- a/memory.c
+++ b/memory.c
@@ -908,20 +908,22 @@ void memory_region_init(MemoryRegion *mr,
                         const char *name,
                         uint64_t size)
 {
-    if (!owner) {
-        owner = container_get(qdev_get_machine(), "/unattached");
-    }
-
     object_initialize(mr, sizeof(*mr), TYPE_MEMORY_REGION);
     mr->size = int128_make64(size);
     if (size == UINT64_MAX) {
         mr->size = int128_2_64();
     }
     mr->name = g_strdup(name);
+    mr->owner = owner;
 
     if (name) {
         char *escaped_name = memory_region_escape_name(name);
         char *name_array = g_strdup_printf("%s[*]", escaped_name);
+
+        if (!owner) {
+            owner = container_get(qdev_get_machine(), "/unattached");
+        }
+
         object_property_add_child(owner, name_array, OBJECT(mr), &error_abort);
         object_unref(OBJECT(mr));
         g_free(name_array);
@@ -1341,24 +1343,18 @@ void memory_region_ref(MemoryRegion *mr)
      * The memory region is a child of its owner.  As long as the
      * owner doesn't call unparent itself on the memory region,
      * ref-ing the owner will also keep the memory region alive.
-     * Memory regions without an owner are supposed to never go away,
-     * but we still ref/unref them for debugging purposes.
+     * Memory regions without an owner are supposed to never go away;
+     * we do not ref/unref them because it slows down DMA sensibly.
      */
-    Object *obj = OBJECT(mr);
-    if (obj && obj->parent) {
-        object_ref(obj->parent);
-    } else {
-        object_ref(obj);
+    if (mr && mr->owner) {
+        object_ref(mr->owner);
     }
 }
 
 void memory_region_unref(MemoryRegion *mr)
 {
-    Object *obj = OBJECT(mr);
-    if (obj && obj->parent) {
-        object_unref(obj->parent);
-    } else {
-        object_unref(obj);
+    if (mr && mr->owner) {
+        object_unref(mr->owner);
     }
 }
 
-- 
2.5.0

  parent reply	other threads:[~2015-12-16 11:00 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-16 10:59 [Qemu-devel] [PATCH 0/8] Optimize address_space_read/write/map Paolo Bonzini
2015-12-16 10:59 ` [Qemu-devel] [PATCH 1/8] exec: always call qemu_get_ram_ptr within rcu_read_lock Paolo Bonzini
2015-12-16 10:59 ` [Qemu-devel] [PATCH 2/8] exec: make qemu_ram_ptr_length more similar to qemu_get_ram_ptr Paolo Bonzini
2015-12-16 10:59 ` [Qemu-devel] [PATCH 3/8] memory: reorder MemoryRegion fields Paolo Bonzini
2015-12-16 10:59 ` Paolo Bonzini [this message]
2015-12-16 10:59 ` [Qemu-devel] [PATCH 5/8] memory: split address_space_read and address_space_write Paolo Bonzini
2015-12-16 10:59 ` [Qemu-devel] [PATCH 6/8] memory: extract first iteration of " Paolo Bonzini
2015-12-16 10:59 ` [Qemu-devel] [PATCH 7/8] memory: inline a few small accessors Paolo Bonzini
2015-12-16 11:00 ` [Qemu-devel] [PATCH 8/8] memory: try to inline constant-length reads Paolo Bonzini
2015-12-16 11:00 ` [Qemu-devel] [PATCH] " Paolo Bonzini

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=1450263601-2828-5-git-send-email-pbonzini@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=v.maffione@gmail.com \
    /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).