qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Michael Roth <mdroth@linux.vnet.ibm.com>
To: qemu-devel@nongnu.org
Cc: qemu-stable@nongnu.org, Paolo Bonzini <pbonzini@redhat.com>
Subject: [Qemu-devel] [PATCH 23/55] memory: seek FlatView sharing candidates among children subregions
Date: Wed,  6 Dec 2017 13:16:16 -0600	[thread overview]
Message-ID: <20171206191648.18208-24-mdroth@linux.vnet.ibm.com> (raw)
In-Reply-To: <20171206191648.18208-1-mdroth@linux.vnet.ibm.com>

From: Paolo Bonzini <pbonzini@redhat.com>

A container can be used instead of an alias to allow switching between
multiple subregions.  In this case we cannot directly share the
subregions (since they only belong to a single parent), but if the
subregions are aliases we can in turn walk those.

This is not enough to remove all source of quadratic FlatView creation,
but it enables sharing of the PCI bus master FlatViews (and their
AddressSpaceDispatch structures) across all PCI devices.  For 112
virtio-net-pci devices, boot time is reduced from 25 to 10 seconds and
memory consumption from 1.4 to 1 G.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
(cherry picked from commit e673ba9af9bf8fd8e0f44025ac738b8285b3ed27)
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
---
 memory.c | 40 ++++++++++++++++++++++++++++++++++------
 1 file changed, 34 insertions(+), 6 deletions(-)

diff --git a/memory.c b/memory.c
index 774a17f60a..231bb78fa7 100644
--- a/memory.c
+++ b/memory.c
@@ -733,12 +733,40 @@ static void render_memory_region(FlatView *view,
 
 static MemoryRegion *memory_region_get_flatview_root(MemoryRegion *mr)
 {
-    while (mr->alias && !mr->alias_offset &&
-           int128_ge(mr->size, mr->alias->size)) {
-        /* The alias is included in its entirety.  Use it as
-         * the "real" root, so that we can share more FlatViews.
-         */
-        mr = mr->alias;
+    while (mr->enabled) {
+        if (mr->alias) {
+            if (!mr->alias_offset && int128_ge(mr->size, mr->alias->size)) {
+                /* The alias is included in its entirety.  Use it as
+                 * the "real" root, so that we can share more FlatViews.
+                 */
+                mr = mr->alias;
+                continue;
+            }
+        } else if (!mr->terminates) {
+            unsigned int found = 0;
+            MemoryRegion *child, *next = NULL;
+            QTAILQ_FOREACH(child, &mr->subregions, subregions_link) {
+                if (child->enabled) {
+                    if (++found > 1) {
+                        next = NULL;
+                        break;
+                    }
+                    if (!child->addr && int128_ge(mr->size, child->size)) {
+                        /* A child is included in its entirety.  If it's the only
+                         * enabled one, use it in the hope of finding an alias down the
+                         * way. This will also let us share FlatViews.
+                         */
+                        next = child;
+                    }
+                }
+            }
+            if (next) {
+                mr = next;
+                continue;
+            }
+        }
+
+        break;
     }
 
     return mr;
-- 
2.11.0

  parent reply	other threads:[~2017-12-06 19:17 UTC|newest]

Thread overview: 56+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-12-06 19:15 [Qemu-devel] [PATCH 00/55] Patch Round-up for stable 2.10.2, freeze on 2017-12-13 Michael Roth
2017-12-06 19:15 ` [Qemu-devel] [PATCH 01/55] hw/ppc: CAS reset on early device hotplug Michael Roth
2017-12-06 19:15 ` [Qemu-devel] [PATCH 02/55] hw/usb/bus: Remove bad object_unparent() from usb_try_create_simple() Michael Roth
2017-12-06 19:15 ` [Qemu-devel] [PATCH 03/55] block/mirror: check backing in bdrv_mirror_top_flush Michael Roth
2017-12-06 19:15 ` [Qemu-devel] [PATCH 04/55] kvmclock: use the updated system_timer_msr Michael Roth
2017-12-06 19:15 ` [Qemu-devel] [PATCH 05/55] block: Perform copy-on-read in loop Michael Roth
2017-12-06 19:15 ` [Qemu-devel] [PATCH 06/55] exec: Explicitly export target AS from address_space_translate_internal Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 07/55] memory: Open code FlatView rendering Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 08/55] memory: Move FlatView allocation to a helper Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 09/55] memory: Move AddressSpaceDispatch from AddressSpace to FlatView Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 10/55] memory: Remove AddressSpace pointer from AddressSpaceDispatch Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 11/55] memory: avoid "resurrection" of dead FlatViews Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 12/55] memory: Switch memory from using AddressSpace to FlatView Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 13/55] memory: Cleanup after switching " Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 14/55] memory: Rename mem_begin/mem_commit/mem_add helpers Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 15/55] memory: Store physical root MR in FlatView Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 16/55] memory: Alloc dispatch tree where topology is generared Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 17/55] memory: Move address_space_update_ioeventfds Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 18/55] memory: Share FlatView's and dispatch trees between address spaces Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 19/55] memory: Do not allocate FlatView in address_space_init Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 20/55] memory: Get rid of address_space_init_shareable Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 21/55] memory: Create FlatView directly Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 22/55] memory: trace FlatView creation and destruction Michael Roth
2017-12-06 19:16 ` Michael Roth [this message]
2017-12-06 19:16 ` [Qemu-devel] [PATCH 24/55] memory: Share special empty FlatView Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 25/55] exec: add page_mask for flatview_do_translate Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 26/55] exec: simplify address_space_get_iotlb_entry Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 27/55] memory: fix off-by-one error in memory_region_notify_one() Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 28/55] hw/sd: fix out-of-bounds check for multi block reads Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 29/55] qcow2: Fix unaligned preallocated truncation Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 30/55] qcow2: Always execute preallocate() in a coroutine Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 31/55] iotests: Add cluster_size=64k to 125 Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 32/55] nios2: define tcg_env Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 33/55] io: monitor encoutput buffer size from websocket GSource Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 34/55] ppc: fix setting of compat mode Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 35/55] translate.c: Fix usermode big-endian AArch32 LDREXD and STREXD Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 36/55] hw/intc/arm_gicv3_its: Don't abort on table save failure Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 37/55] net/socket: fix coverity issue Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 38/55] net: fix check for number of parameters to -netdev socket Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 39/55] nbd/client: Use error_prepend() correctly Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 40/55] util/stats64: Fix min/max comparisons Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 41/55] virtio: Add queue interface to restore avail index from vring used index Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 42/55] vhost: restore avail index from vring used index on disconnection Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 43/55] hw/ppc: clear pending_events on machine reset Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 44/55] spapr: reset DRCs after devices Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 45/55] scripts/make-release: ship u-boot source as a tarball Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 46/55] block/nfs: fix nfs_client_open for filesize greater than 1TB Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 47/55] virtio-net: don't touch virtqueue if vm is stopped Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 48/55] nbd/server: CVE-2017-15119 Reject options larger than 32M Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 49/55] nbd/server: CVE-2017-15118 Stack smash on large export name Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 50/55] vhost: fix error check in vhost_verify_ring_mappings() Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 51/55] nbd/server: fix nbd_negotiate_handle_info Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 52/55] nbd-client: Refuse read-only client with BDRV_O_RDWR Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 53/55] nbd/client: Don't hard-disconnect on ESHUTDOWN from server Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 54/55] vga: drop line_offset variable Michael Roth
2017-12-06 19:16 ` [Qemu-devel] [PATCH 55/55] vga: handle cirrus vbe mode wraparounds Michael Roth

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=20171206191648.18208-24-mdroth@linux.vnet.ibm.com \
    --to=mdroth@linux.vnet.ibm.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-stable@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).