From: Peter Xu <peterx@redhat.com>
To: Peter Maydell <peter.maydell@linaro.org>, qemu-devel@nongnu.org
Cc: Fabiano Rosas <farosas@suse.de>,
peterx@redhat.com, David Hildenbrand <david@redhat.com>,
Paolo Bonzini <pbonzini@redhat.com>,
qemu-stable@nongnu.org
Subject: [PULL 35/45] physmem: Destroy all CPU AddressSpaces on unrealize
Date: Fri, 3 Oct 2025 11:39:38 -0400 [thread overview]
Message-ID: <20251003153948.1304776-36-peterx@redhat.com> (raw)
In-Reply-To: <20251003153948.1304776-1-peterx@redhat.com>
From: Peter Maydell <peter.maydell@linaro.org>
When we unrealize a CPU object (which happens on vCPU hot-unplug), we
should destroy all the AddressSpace objects we created via calls to
cpu_address_space_init() when the CPU was realized.
Commit 24bec42f3d6eae added a function to do this for a specific
AddressSpace, but did not add any places where the function was
called.
Since we always want to destroy all the AddressSpaces on unrealize,
regardless of the target architecture, we don't need to try to keep
track of how many are still undestroyed, or make the target
architecture code manually call a destroy function for each AS it
created. Instead we can adjust the function to always completely
destroy the whole cpu->ases array, and arrange for it to be called
during CPU unrealize as part of the common code.
Without this fix, AddressSanitizer will report a leak like this
from a run where we hot-plugged and then hot-unplugged an x86 KVM
vCPU:
Direct leak of 416 byte(s) in 1 object(s) allocated from:
#0 0x5b638565053d in calloc (/data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/qemu-system-x86_64+0x1ee153d) (BuildId: c1cd6022b195142106e1bffeca23498c2b752bca)
#1 0x7c28083f77b1 in g_malloc0 (/lib/x86_64-linux-gnu/libglib-2.0.so.0+0x637b1) (BuildId: 1eb6131419edb83b2178b682829a6913cf682d75)
#2 0x5b6386999c7c in cpu_address_space_init /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../system/physmem.c:797:25
#3 0x5b638727f049 in kvm_cpu_realizefn /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../target/i386/kvm/kvm-cpu.c:102:5
#4 0x5b6385745f40 in accel_cpu_common_realize /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../accel/accel-common.c:101:13
#5 0x5b638568fe3c in cpu_exec_realizefn /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../hw/core/cpu-common.c:232:10
#6 0x5b63874a2cd5 in x86_cpu_realizefn /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../target/i386/cpu.c:9321:5
#7 0x5b6387a0469a in device_set_realized /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../hw/core/qdev.c:494:13
#8 0x5b6387a27d9e in property_set_bool /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../qom/object.c:2375:5
#9 0x5b6387a2090b in object_property_set /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../qom/object.c:1450:5
#10 0x5b6387a35b05 in object_property_set_qobject /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../qom/qom-qobject.c:28:10
#11 0x5b6387a21739 in object_property_set_bool /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../qom/object.c:1520:15
#12 0x5b63879fe510 in qdev_realize /data_nvme1n1/linaro/qemu-from-laptop/qemu/build/x86-tgts-asan/../../hw/core/qdev.c:276:12
Cc: qemu-stable@nongnu.org
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2517
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20250929144228.1994037-4-peter.maydell@linaro.org
Signed-off-by: Peter Xu <peterx@redhat.com>
---
include/exec/cpu-common.h | 10 +++++-----
include/hw/core/cpu.h | 1 -
hw/core/cpu-common.c | 1 +
stubs/cpu-destroy-address-spaces.c | 15 ++++++++++++++
system/physmem.c | 32 ++++++++++++++----------------
stubs/meson.build | 1 +
6 files changed, 37 insertions(+), 23 deletions(-)
create mode 100644 stubs/cpu-destroy-address-spaces.c
diff --git a/include/exec/cpu-common.h b/include/exec/cpu-common.h
index f373781ae0..b96ac49844 100644
--- a/include/exec/cpu-common.h
+++ b/include/exec/cpu-common.h
@@ -123,13 +123,13 @@ size_t qemu_ram_pagesize_largest(void);
void cpu_address_space_init(CPUState *cpu, int asidx,
const char *prefix, MemoryRegion *mr);
/**
- * cpu_address_space_destroy:
- * @cpu: CPU for which address space needs to be destroyed
- * @asidx: integer index of this address space
+ * cpu_destroy_address_spaces:
+ * @cpu: CPU for which address spaces need to be destroyed
*
- * Note that with KVM only one address space is supported.
+ * Destroy all address spaces associated with this CPU; this
+ * is called as part of unrealizing the CPU.
*/
-void cpu_address_space_destroy(CPUState *cpu, int asidx);
+void cpu_destroy_address_spaces(CPUState *cpu);
void cpu_physical_memory_rw(hwaddr addr, void *buf,
hwaddr len, bool is_write);
diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
index c9f40c2539..0fcbc923f3 100644
--- a/include/hw/core/cpu.h
+++ b/include/hw/core/cpu.h
@@ -515,7 +515,6 @@ struct CPUState {
QSIMPLEQ_HEAD(, qemu_work_item) work_list;
struct CPUAddressSpace *cpu_ases;
- int cpu_ases_count;
int num_ases;
AddressSpace *as;
MemoryRegion *memory;
diff --git a/hw/core/cpu-common.c b/hw/core/cpu-common.c
index 41a339903c..8c306c89e4 100644
--- a/hw/core/cpu-common.c
+++ b/hw/core/cpu-common.c
@@ -294,6 +294,7 @@ void cpu_exec_unrealizefn(CPUState *cpu)
* accel_cpu_common_unrealize, which may free fields using call_rcu.
*/
accel_cpu_common_unrealize(cpu);
+ cpu_destroy_address_spaces(cpu);
}
static void cpu_common_initfn(Object *obj)
diff --git a/stubs/cpu-destroy-address-spaces.c b/stubs/cpu-destroy-address-spaces.c
new file mode 100644
index 0000000000..dc6813f5bd
--- /dev/null
+++ b/stubs/cpu-destroy-address-spaces.c
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+
+#include "qemu/osdep.h"
+#include "exec/cpu-common.h"
+
+/*
+ * user-mode CPUs never create address spaces with
+ * cpu_address_space_init(), so the cleanup function doesn't
+ * need to do anything. We need this stub because cpu-common.c
+ * is built-once so it can't #ifndef CONFIG_USER around the
+ * call; the real function is in physmem.c which is system-only.
+ */
+void cpu_destroy_address_spaces(CPUState *cpu)
+{
+}
diff --git a/system/physmem.c b/system/physmem.c
index ae8ecd50ea..dbb2a4e017 100644
--- a/system/physmem.c
+++ b/system/physmem.c
@@ -795,7 +795,6 @@ void cpu_address_space_init(CPUState *cpu, int asidx,
if (!cpu->cpu_ases) {
cpu->cpu_ases = g_new0(CPUAddressSpace, cpu->num_ases);
- cpu->cpu_ases_count = cpu->num_ases;
}
newas = &cpu->cpu_ases[asidx];
@@ -809,30 +808,29 @@ void cpu_address_space_init(CPUState *cpu, int asidx,
}
}
-void cpu_address_space_destroy(CPUState *cpu, int asidx)
+void cpu_destroy_address_spaces(CPUState *cpu)
{
CPUAddressSpace *cpuas;
+ int asidx;
assert(cpu->cpu_ases);
- assert(asidx >= 0 && asidx < cpu->num_ases);
- cpuas = &cpu->cpu_ases[asidx];
- if (tcg_enabled()) {
- memory_listener_unregister(&cpuas->tcg_as_listener);
- }
+ /* convenience alias just points to some cpu_ases[n] */
+ cpu->as = NULL;
- address_space_destroy(cpuas->as);
- g_free_rcu(cpuas->as, rcu);
-
- if (asidx == 0) {
- /* reset the convenience alias for address space 0 */
- cpu->as = NULL;
+ for (asidx = 0; asidx < cpu->num_ases; asidx++) {
+ cpuas = &cpu->cpu_ases[asidx];
+ if (!cpuas->as) {
+ /* This index was never initialized; no deinit needed */
+ continue;
+ }
+ if (tcg_enabled()) {
+ memory_listener_unregister(&cpuas->tcg_as_listener);
+ }
+ g_clear_pointer(&cpuas->as, address_space_destroy_free);
}
- if (--cpu->cpu_ases_count == 0) {
- g_free(cpu->cpu_ases);
- cpu->cpu_ases = NULL;
- }
+ g_clear_pointer(&cpu->cpu_ases, g_free);
}
AddressSpace *cpu_get_address_space(CPUState *cpu, int asidx)
diff --git a/stubs/meson.build b/stubs/meson.build
index cef046e685..5d577467bf 100644
--- a/stubs/meson.build
+++ b/stubs/meson.build
@@ -55,6 +55,7 @@ endif
if have_user
# Symbols that are used by hw/core.
stub_ss.add(files('cpu-synchronize-state.c'))
+ stub_ss.add(files('cpu-destroy-address-spaces.c'))
# Stubs for QAPI events. Those can always be included in the build, but
# they are not built at all for --disable-system builds.
--
2.50.1
next prev parent reply other threads:[~2025-10-03 15:49 UTC|newest]
Thread overview: 67+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-10-03 15:39 [PULL 00/45] Staging patches Peter Xu
2025-10-03 15:39 ` [PULL 01/45] migration: push Error **errp into vmstate_subsection_load() Peter Xu
2025-10-03 15:39 ` [PULL 02/45] migration: push Error **errp into vmstate_load_state() Peter Xu
2025-10-21 15:43 ` Peter Maydell
2025-10-21 16:16 ` Peter Xu
2025-10-21 16:21 ` Peter Maydell
2025-10-21 16:46 ` Peter Xu
2025-10-21 16:49 ` Peter Maydell
2025-10-21 17:05 ` Peter Xu
2025-10-21 17:22 ` Peter Maydell
2025-10-21 18:09 ` Peter Xu
2025-10-23 21:50 ` Arun Menon
2025-10-24 15:48 ` Peter Xu
2025-10-03 15:39 ` [PULL 03/45] migration: push Error **errp into qemu_loadvm_state_header() Peter Xu
2025-10-03 15:39 ` [PULL 04/45] migration: push Error **errp into vmstate_load() Peter Xu
2025-10-03 15:39 ` [PULL 05/45] migration: push Error **errp into loadvm_process_command() Peter Xu
2025-10-03 15:39 ` [PULL 06/45] migration: push Error **errp into loadvm_handle_cmd_packaged() Peter Xu
2025-10-03 15:39 ` [PULL 07/45] migration: push Error **errp into qemu_loadvm_state() Peter Xu
2025-10-03 15:39 ` [PULL 08/45] migration: push Error **errp into qemu_load_device_state() Peter Xu
2025-10-03 15:39 ` [PULL 09/45] migration: push Error **errp into qemu_loadvm_state_main() Peter Xu
2025-10-03 15:39 ` [PULL 10/45] migration: push Error **errp into qemu_loadvm_section_start_full() Peter Xu
2025-10-03 15:39 ` [PULL 11/45] migration: push Error **errp into qemu_loadvm_section_part_end() Peter Xu
2025-10-03 15:39 ` [PULL 12/45] migration: Update qemu_file_get_return_path() docs and remove dead checks Peter Xu
2025-10-03 15:39 ` [PULL 13/45] migration: make loadvm_postcopy_handle_resume() void Peter Xu
2025-10-03 15:39 ` [PULL 14/45] migration: push Error **errp into ram_postcopy_incoming_init() Peter Xu
2025-10-03 15:39 ` [PULL 15/45] migration: push Error **errp into loadvm_postcopy_handle_advise() Peter Xu
2025-10-03 15:39 ` [PULL 16/45] migration: push Error **errp into loadvm_postcopy_handle_listen() Peter Xu
2025-10-03 15:39 ` [PULL 17/45] migration: push Error **errp into loadvm_postcopy_handle_run() Peter Xu
2025-10-03 15:39 ` [PULL 18/45] migration: push Error **errp into loadvm_postcopy_ram_handle_discard() Peter Xu
2025-10-03 15:39 ` [PULL 19/45] migration: push Error **errp into loadvm_handle_recv_bitmap() Peter Xu
2025-10-03 15:39 ` [PULL 20/45] migration: Return -1 on memory allocation failure in ram.c Peter Xu
2025-10-03 15:39 ` [PULL 21/45] migration: push Error **errp into loadvm_process_enable_colo() Peter Xu
2025-10-03 15:39 ` [PULL 22/45] migration: push Error **errp into loadvm_postcopy_handle_switchover_start() Peter Xu
2025-10-03 15:39 ` [PULL 23/45] migration: Capture error in postcopy_ram_listen_thread() Peter Xu
2025-10-21 14:53 ` Peter Maydell
2025-10-21 15:37 ` Peter Xu
2025-10-28 5:46 ` Arun Menon
2025-10-03 15:39 ` [PULL 24/45] migration: Remove error variant of vmstate_save_state() function Peter Xu
2025-10-03 15:39 ` [PULL 25/45] migration: Add error-parameterized function variants in VMSD struct Peter Xu
2025-10-03 15:39 ` [PULL 26/45] backends/tpm: Propagate vTPM error on migration failure Peter Xu
2025-10-03 15:39 ` [PULL 27/45] io/crypto: Move tls premature termination handling into QIO layer Peter Xu
2025-10-10 8:00 ` iotest 233 is failing (was: [PULL 27/45] io/crypto: Move tls premature termination handling into QIO layer) Thomas Huth
2025-10-10 8:35 ` iotest 233 is failing Thomas Huth
2025-10-03 15:39 ` [PULL 28/45] migration: Make migration_has_failed() work even for CANCELLING Peter Xu
2025-10-03 15:39 ` [PULL 29/45] migration: HMP: Adjust the order of output fields Peter Xu
2025-10-03 15:39 ` [PULL 30/45] migration/multifd/tls: Cleanup BYE message processing on sender side Peter Xu
2025-10-03 15:39 ` [PULL 31/45] migration: Fix state transition in postcopy_start() error handling Peter Xu
2025-10-03 15:39 ` [PULL 32/45] migration: ensure APIC is loaded prior to VFIO PCI devices Peter Xu
2025-10-03 15:39 ` [PULL 33/45] include/system/memory.h: Clarify address_space_destroy() behaviour Peter Xu
2025-10-03 15:39 ` [PULL 34/45] memory: New AS helper to serialize destroy+free Peter Xu
2025-10-03 15:39 ` Peter Xu [this message]
2025-10-03 15:39 ` [PULL 36/45] migration: simplify error reporting after channel read Peter Xu
2025-10-03 15:39 ` [PULL 37/45] migration: multi-mode notifier Peter Xu
2025-10-03 15:39 ` [PULL 38/45] migration: add cpr_walk_fd Peter Xu
2025-10-03 15:39 ` [PULL 39/45] oslib: qemu_clear_cloexec Peter Xu
2025-10-03 15:39 ` [PULL 40/45] migration: cpr-exec-command parameter Peter Xu
2025-10-23 15:41 ` Peter Maydell
2025-10-23 16:08 ` Peter Xu
2025-10-03 15:39 ` [PULL 41/45] migration: cpr-exec save and load Peter Xu
2025-10-21 14:59 ` Peter Maydell
2025-10-03 15:39 ` [PULL 42/45] migration: cpr-exec mode Peter Xu
2025-10-21 15:34 ` Peter Maydell
2025-10-21 15:59 ` Peter Xu
2025-10-03 15:39 ` [PULL 43/45] migration: cpr-exec docs Peter Xu
2025-10-03 15:39 ` [PULL 44/45] vfio: cpr-exec mode Peter Xu
2025-10-03 15:39 ` [PULL 45/45] migration-test: test cpr-exec Peter Xu
2025-10-04 17:53 ` [PULL 00/45] Staging patches Richard Henderson
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=20251003153948.1304776-36-peterx@redhat.com \
--to=peterx@redhat.com \
--cc=david@redhat.com \
--cc=farosas@suse.de \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--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).