rpms/qemu-kvm
7
0
Fork 0
Code Issues Pull Requests Releases Activity

import OL qemu-kvm-9.1.0-15.el9_6.7

Browse Source
This commit is contained in:
Andrew Lukoshko 2025-08-07 11:00:30 +00:00
parent ec7054e054
commit 14e6eca34d
35 changed files with 2164 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

48
SOURCES/kvm-block-skip-automatic-zero-init-of-large-array-in-ioq.patch Normal file
View File

@ -0,0 +1,48 @@
From 9f8ff1d0ef010b9c0339869f655ee9af6b10dcd5 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:41 +0100
Subject: [PATCH 04/31] block: skip automatic zero-init of large array in
ioq_submit
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [3/30] 0a24695ab7f3a11ab61b12ae2b95bd45a7babc05
The 'ioq_submit' method has a struct array that is 8k in size.
Skip the automatic zero-init of this array to eliminate the
performance overhead in the I/O hot path.
The 'iocbs' array will selectively initialized when processing
the I/O data.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-4-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 83750c1da807c973b0b11d977d61df7e41122d03)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
block/linux-aio.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/block/linux-aio.c b/block/linux-aio.c
index e3b5ec9aba..26d9f086d2 100644
--- a/block/linux-aio.c
+++ b/block/linux-aio.c
@@ -291,7 +291,7 @@ static void ioq_submit(LinuxAioState *s)
{
int ret, len;
struct qemu_laiocb *aiocb;
- struct iocb *iocbs[MAX_EVENTS];
+ QEMU_UNINITIALIZED struct iocb *iocbs[MAX_EVENTS];
QSIMPLEQ_HEAD(, qemu_laiocb) completed;
do {
--
2.39.3

49
SOURCES/kvm-chardev-char-fd-skip-automatic-zero-init-of-large-ar.patch Normal file
View File

@ -0,0 +1,49 @@
From c336909ad95540147d9cfed843874ddd986ad917 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:42 +0100
Subject: [PATCH 05/31] chardev/char-fd: skip automatic zero-init of large
array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [4/30] e9adf42d47ddd90cc15862e711de0a90d1d25e0a
The 'fd_chr_read' method has a 4k byte array used for copying
data between the socket and device. Skip the automatic zero-init
of this array to eliminate the performance overhead in the I/O
hot path.
The 'buf' array will be fully initialized when reading data off
the network socket.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-5-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit a503bdc22b91869e3bf45522e36b122889465306)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
chardev/char-fd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/chardev/char-fd.c b/chardev/char-fd.c
index d2c4923359..8dd662c066 100644
--- a/chardev/char-fd.c
+++ b/chardev/char-fd.c
@@ -50,7 +50,7 @@ static gboolean fd_chr_read(QIOChannel *chan, GIOCondition cond, void *opaque)
Chardev *chr = CHARDEV(opaque);
FDChardev *s = FD_CHARDEV(opaque);
int len;
- uint8_t buf[CHR_READ_BUF_LEN];
+ QEMU_UNINITIALIZED uint8_t buf[CHR_READ_BUF_LEN];
ssize_t ret;
len = sizeof(buf);
--
2.39.3

49
SOURCES/kvm-chardev-char-pty-skip-automatic-zero-init-of-large-a.patch Normal file
View File

@ -0,0 +1,49 @@
From 489ead1e7d721c0f689626e6d5d22241ffdd7bc8 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:43 +0100
Subject: [PATCH 06/31] chardev/char-pty: skip automatic zero-init of large
array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [5/30] 09be65fa8c25dadbd472321cacc96badf0a2d963
The 'pty_chr_read' method has a 4k byte array used for copying
data between the PTY and device. Skip the automatic zero-init
of this array to eliminate the performance overhead in the I/O
hot path.
The 'buf' array will be fully initialized when reading data off
the PTY.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-6-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 45bb7fb21c8d18294a9f92da99d01ab3c67c7df2)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
chardev/char-pty.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/chardev/char-pty.c b/chardev/char-pty.c
index cc2f7617fe..3319ad215d 100644
--- a/chardev/char-pty.c
+++ b/chardev/char-pty.c
@@ -152,7 +152,7 @@ static gboolean pty_chr_read(QIOChannel *chan, GIOCondition cond, void *opaque)
Chardev *chr = CHARDEV(opaque);
PtyChardev *s = PTY_CHARDEV(opaque);
gsize len;
- uint8_t buf[CHR_READ_BUF_LEN];
+ QEMU_UNINITIALIZED uint8_t buf[CHR_READ_BUF_LEN];
ssize_t ret;
len = sizeof(buf);
--
2.39.3

49
SOURCES/kvm-chardev-char-socket-skip-automatic-zero-init-of-larg.patch Normal file
View File

@ -0,0 +1,49 @@
From 67cf0b18b68071b7b0a036b715f9d406f0bc1ec7 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:44 +0100
Subject: [PATCH 07/31] chardev/char-socket: skip automatic zero-init of large
array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [6/30] fa1d406d2f9be389090614d666466fb94f480f84
The 'tcp_chr_read' method has a 4k byte array used for copying
data between the socket and device. Skip the automatic zero-init
of this array to eliminate the performance overhead in the I/O
hot path.
The 'buf' array will be fully initialized when reading data off
the network socket.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-7-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 9a23075cef1ac6e73a95a489ac72f41c573ceb9b)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
chardev/char-socket.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/chardev/char-socket.c b/chardev/char-socket.c
index 1ca9441b1b..99d644e89f 100644
--- a/chardev/char-socket.c
+++ b/chardev/char-socket.c
@@ -497,7 +497,7 @@ static gboolean tcp_chr_read(QIOChannel *chan, GIOCondition cond, void *opaque)
{
Chardev *chr = CHARDEV(opaque);
SocketChardev *s = SOCKET_CHARDEV(opaque);
- uint8_t buf[CHR_READ_BUF_LEN];
+ QEMU_UNINITIALIZED uint8_t buf[CHR_READ_BUF_LEN];
int len, size;
if ((s->state != TCP_CHARDEV_STATE_CONNECTED) ||
--
2.39.3

42
SOURCES/kvm-file-posix-Define-DM_MPATH_PROBE_PATHS.patch Normal file
View File

@ -0,0 +1,42 @@
From 762f24e92b93c0e8cbb5b0abe135d29fb444737d Mon Sep 17 00:00:00 2001
From: Kevin Wolf <kwolf@redhat.com>
Date: Tue, 29 Apr 2025 17:05:41 +0200
Subject: [PATCH 1/2] file-posix: Define DM_MPATH_PROBE_PATHS
RH-Author: Kevin Wolf <kwolf@redhat.com>
RH-MergeRequest: 456: file-posix: Fix multipath failover with SCSI passthrough [9.6.z]
RH-Jira: RHEL-95407
RH-Acked-by: Hanna Czenczek <hreitz@redhat.com>
RH-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
RH-Commit: [1/2] 0d9ec74bf3bb999c8baa929e0d25682680fa1731 (kmwolf/rhel-qemu-kvm)
While the kernel side isn't merged yet and we're still using old kernel
headers, just define DM_MPATH_PROBE_PATHS manually.
This is a downstream-only patch that can be removed after the next minor
release.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
block/file-posix.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/block/file-posix.c b/block/file-posix.c
index 0cb4e922c0..6a5c506549 100644
--- a/block/file-posix.c
+++ b/block/file-posix.c
@@ -134,6 +134,11 @@
#define RAW_LOCK_PERM_BASE 100
#define RAW_LOCK_SHARED_BASE 200
+/* TODO Remove this when the kernel side is merged */
+#if !defined(DM_MPATH_PROBE_PATHS) && defined(DM_GET_TARGET_VERSION)
+#define DM_MPATH_PROBE_PATHS _IO(DM_IOCTL, DM_GET_TARGET_VERSION_CMD + 1)
+#endif
+
typedef struct BDRVRawState {
int fd;
bool use_lock;
--
2.48.1

215
SOURCES/kvm-file-posix-Probe-paths-and-retry-SG_IO-on-potential-.patch Normal file
View File

@ -0,0 +1,215 @@
From 50b1a3ec7cfea5a92069e043e8f77a2595480d20 Mon Sep 17 00:00:00 2001
From: Kevin Wolf <kwolf@redhat.com>
Date: Thu, 22 May 2025 15:08:03 +0200
Subject: [PATCH 2/2] file-posix: Probe paths and retry SG_IO on potential path
errors
RH-Author: Kevin Wolf <kwolf@redhat.com>
RH-MergeRequest: 456: file-posix: Fix multipath failover with SCSI passthrough [9.6.z]
RH-Jira: RHEL-95407
RH-Acked-by: Hanna Czenczek <hreitz@redhat.com>
RH-Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
RH-Commit: [2/2] 2f4aed9004889e9df25e30c2273f2524a67c666c (kmwolf/rhel-qemu-kvm)
When scsi-block is used on a host multipath device, it runs into the
problem that the kernel dm-mpath doesn't know anything about SCSI or
SG_IO and therefore can't decide if a SG_IO request returned an error
and needs to be retried on a different path. Instead of getting working
failover, an error is returned to scsi-block and handled according to
the configured error policy. Obviously, this is not what users want,
they want working failover.
QEMU can parse the SG_IO result and determine whether this could have
been a path error, but just retrying the same request could just send it
to the same failing path again and result in the same error.
With a kernel that supports the DM_MPATH_PROBE_PATHS ioctl on dm-mpath
block devices (queued in the device mapper tree for Linux 6.16), we can
tell the kernel to probe all paths and tell us if any usable paths
remained. If so, we can now retry the SG_IO ioctl and expect it to be
sent to a working path.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20250522130803.34738-1-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
(cherry picked from commit bf627788ef17721955bfcfba84209a07ae5f54ea)
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
block/file-posix.c | 115 ++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 114 insertions(+), 1 deletion(-)
diff --git a/block/file-posix.c b/block/file-posix.c
index 6a5c506549..f17a3f4d10 100644
--- a/block/file-posix.c
+++ b/block/file-posix.c
@@ -41,6 +41,7 @@
#include "scsi/pr-manager.h"
#include "scsi/constants.h"
+#include "scsi/utils.h"
#if defined(__APPLE__) && (__MACH__)
#include <sys/ioctl.h>
@@ -72,6 +73,7 @@
#include <linux/blkzoned.h>
#endif
#include <linux/cdrom.h>
+#include <linux/dm-ioctl.h>
#include <linux/fd.h>
#include <linux/fs.h>
#include <linux/hdreg.h>
@@ -139,6 +141,22 @@
#define DM_MPATH_PROBE_PATHS _IO(DM_IOCTL, DM_GET_TARGET_VERSION_CMD + 1)
#endif
+/*
+ * Multiple retries are mostly meant for two separate scenarios:
+ *
+ * - DM_MPATH_PROBE_PATHS returns success, but before SG_IO completes, another
+ * path goes down.
+ *
+ * - DM_MPATH_PROBE_PATHS failed all paths in the current path group, so we have
+ * to send another SG_IO to switch to another path group to probe the paths in
+ * it.
+ *
+ * Even if each path is in a separate path group (path_grouping_policy set to
+ * failover), it's rare to have more than eight path groups - and even then
+ * pretty unlikely that only bad path groups would be chosen in eight retries.
+ */
+#define SG_IO_MAX_RETRIES 8
+
typedef struct BDRVRawState {
int fd;
bool use_lock;
@@ -166,6 +184,7 @@ typedef struct BDRVRawState {
bool use_linux_aio:1;
bool has_laio_fdsync:1;
bool use_linux_io_uring:1;
+ bool use_mpath:1;
int page_cache_inconsistent; /* errno from fdatasync failure */
bool has_fallocate;
bool needs_alignment;
@@ -4248,15 +4267,105 @@ hdev_open_Mac_error:
/* Since this does ioctl the device must be already opened */
bs->sg = hdev_is_sg(bs);
+ /* sg devices aren't even block devices and can't use dm-mpath */
+ s->use_mpath = !bs->sg;
+
return ret;
}
#if defined(__linux__)
+#if defined(DM_MPATH_PROBE_PATHS)
+static bool coroutine_fn sgio_path_error(int ret, sg_io_hdr_t *io_hdr)
+{
+ if (ret < 0) {
+ switch (ret) {
+ case -ENODEV:
+ return true;
+ case -EAGAIN:
+ /*
+ * The device is probably suspended. This happens while the dm table
+ * is reloaded, e.g. because a path is added or removed. This is an
+ * operation that should complete within 1ms, so just wait a bit and
+ * retry.
+ *
+ * If the device was suspended for another reason, we'll wait and
+ * retry SG_IO_MAX_RETRIES times. This is a tolerable delay before
+ * we return an error and potentially stop the VM.
+ */
+ qemu_co_sleep_ns(QEMU_CLOCK_REALTIME, 1000000);
+ return true;
+ default:
+ return false;
+ }
+ }
+
+ if (io_hdr->host_status != SCSI_HOST_OK) {
+ return true;
+ }
+
+ switch (io_hdr->status) {
+ case GOOD:
+ case CONDITION_GOOD:
+ case INTERMEDIATE_GOOD:
+ case INTERMEDIATE_C_GOOD:
+ case RESERVATION_CONFLICT:
+ case COMMAND_TERMINATED:
+ return false;
+ case CHECK_CONDITION:
+ return !scsi_sense_buf_is_guest_recoverable(io_hdr->sbp,
+ io_hdr->mx_sb_len);
+ default:
+ return true;
+ }
+}
+
+static bool coroutine_fn hdev_co_ioctl_sgio_retry(RawPosixAIOData *acb, int ret)
+{
+ BDRVRawState *s = acb->bs->opaque;
+ RawPosixAIOData probe_acb;
+
+ if (!s->use_mpath) {
+ return false;
+ }
+
+ if (!sgio_path_error(ret, acb->ioctl.buf)) {
+ return false;
+ }
+
+ probe_acb = (RawPosixAIOData) {
+ .bs = acb->bs,
+ .aio_type = QEMU_AIO_IOCTL,
+ .aio_fildes = s->fd,
+ .aio_offset = 0,
+ .ioctl = {
+ .buf = NULL,
+ .cmd = DM_MPATH_PROBE_PATHS,
+ },
+ };
+
+ ret = raw_thread_pool_submit(handle_aiocb_ioctl, &probe_acb);
+ if (ret == -ENOTTY) {
+ s->use_mpath = false;
+ } else if (ret == -EAGAIN) {
+ /* The device might be suspended for a table reload, worth retrying */
+ return true;
+ }
+
+ return ret == 0;
+}
+#else
+static bool coroutine_fn hdev_co_ioctl_sgio_retry(RawPosixAIOData *acb, int ret)
+{
+ return false;
+}
+#endif /* DM_MPATH_PROBE_PATHS */
+
static int coroutine_fn
hdev_co_ioctl(BlockDriverState *bs, unsigned long int req, void *buf)
{
BDRVRawState *s = bs->opaque;
RawPosixAIOData acb;
+ int retries = SG_IO_MAX_RETRIES;
int ret;
ret = fd_open(bs);
@@ -4284,7 +4393,11 @@ hdev_co_ioctl(BlockDriverState *bs, unsigned long int req, void *buf)
},
};
- return raw_thread_pool_submit(handle_aiocb_ioctl, &acb);
+ do {
+ ret = raw_thread_pool_submit(handle_aiocb_ioctl, &acb);
+ } while (req == SG_IO && retries-- && hdev_co_ioctl_sgio_retry(&acb, ret));
+
+ return ret;
}
#endif /* linux */
--
2.48.1

57
SOURCES/kvm-hw-audio-ac97-skip-automatic-zero-init-of-large-arra.patch Normal file
View File

@ -0,0 +1,57 @@
From d589e33ff7dea4aeb69fe205bea02fb6bd7da618 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:45 +0100
Subject: [PATCH 08/31] hw/audio/ac97: skip automatic zero-init of large arrays
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [7/30] a2898256b990c1916082a9938740d0fe53da5325
The 'read_audio' & 'write_audio' methods have a 4k byte array used
for copying data between the audio backend and device. Skip the
automatic zero-init of these arrays to eliminate the performance
overhead in the I/O hot path.
The 'tmpbuf' array will be fully initialized when reading data from
the audio backend and/or device memory.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-8-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 2553d2d26a9d0f46386bf8c37d184567e5cede6c)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/audio/ac97.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/hw/audio/ac97.c b/hw/audio/ac97.c
index 3f0053f94d..681b5752a1 100644
--- a/hw/audio/ac97.c
+++ b/hw/audio/ac97.c
@@ -886,7 +886,7 @@ static void nabm_writel(void *opaque, uint32_t addr, uint32_t val)
static int write_audio(AC97LinkState *s, AC97BusMasterRegs *r,
int max, int *stop)
{
- uint8_t tmpbuf[4096];
+ QEMU_UNINITIALIZED uint8_t tmpbuf[4096];
uint32_t addr = r->bd.addr;
uint32_t temp = r->picb << 1;
uint32_t written = 0;
@@ -959,7 +959,7 @@ static void write_bup(AC97LinkState *s, int elapsed)
static int read_audio(AC97LinkState *s, AC97BusMasterRegs *r,
int max, int *stop)
{
- uint8_t tmpbuf[4096];
+ QEMU_UNINITIALIZED uint8_t tmpbuf[4096];
uint32_t addr = r->bd.addr;
uint32_t temp = r->picb << 1;
uint32_t nread = 0;
--
2.39.3

59
SOURCES/kvm-hw-audio-cs4231a-skip-automatic-zero-init-of-large-a.patch Normal file
View File

@ -0,0 +1,59 @@
From 63094ee4645705be09f68c547e3f1775ca528951 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:46 +0100
Subject: [PATCH 09/31] hw/audio/cs4231a: skip automatic zero-init of large
arrays
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [8/30] c6117831ac2ec8e0a700207d5f101eddb67a24a4
The 'cs_write_audio' method has a pair of byte arrays, one 4k in size
and one 8k, which are used in converting audio samples. Skip the
automatic zero-init of these arrays to eliminate the performance
overhead in the I/O hot path.
The 'tmpbuf' array will be fully initialized when reading a block of
data from the guest. The 'linbuf' array will be fully initialized
when converting the audio samples.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-9-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit ca2cc0385d97cea66cd54ee42553f385c403d4a6)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/audio/cs4231a.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/hw/audio/cs4231a.c b/hw/audio/cs4231a.c
index 9ef57f042d..5c312642cc 100644
--- a/hw/audio/cs4231a.c
+++ b/hw/audio/cs4231a.c
@@ -528,7 +528,7 @@ static int cs_write_audio (CSState *s, int nchan, int dma_pos,
int dma_len, int len)
{
int temp, net;
- uint8_t tmpbuf[4096];
+ QEMU_UNINITIALIZED uint8_t tmpbuf[4096];
IsaDmaClass *k = ISADMA_GET_CLASS(s->isa_dma);
temp = len;
@@ -547,7 +547,7 @@ static int cs_write_audio (CSState *s, int nchan, int dma_pos,
copied = k->read_memory(s->isa_dma, nchan, tmpbuf, dma_pos, to_copy);
if (s->tab) {
int i;
- int16_t linbuf[4096];
+ QEMU_UNINITIALIZED int16_t linbuf[4096];
for (i = 0; i < copied; ++i)
linbuf[i] = s->tab[tmpbuf[i]];
--
2.39.3

49
SOURCES/kvm-hw-audio-es1370-skip-automatic-zero-init-of-large-ar.patch Normal file
View File

@ -0,0 +1,49 @@
From 8b487658db35f371e4a527ed18a2ae63b4048f83 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:47 +0100
Subject: [PATCH 10/31] hw/audio/es1370: skip automatic zero-init of large
array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [9/30] d6be11e78b1af782f12d7a25fc6c297370aafd4a
The 'es1370_transfer_audio' method has a 4k byte array used for
copying data between the audio backend and device. Skip the automatic
zero-init of this array to eliminate the performance overhead in
the I/O hot path.
The 'tmpbuf' array will be fully initialized when reading data from
the audio backend and/or device memory.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-10-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 8236e206084b832d1d7ec947a4798b818f4cdf1f)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/audio/es1370.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/audio/es1370.c b/hw/audio/es1370.c
index 4ab61d3b9d..6aea934f54 100644
--- a/hw/audio/es1370.c
+++ b/hw/audio/es1370.c
@@ -604,7 +604,7 @@ static uint64_t es1370_read(void *opaque, hwaddr addr, unsigned size)
static void es1370_transfer_audio (ES1370State *s, struct chan *d, int loop_sel,
int max, bool *irq)
{
- uint8_t tmpbuf[4096];
+ QEMU_UNINITIALIZED uint8_t tmpbuf[4096];
size_t to_transfer;
uint32_t addr = d->frame_addr;
int sc = d->scount & 0xffff;
--
2.39.3

48
SOURCES/kvm-hw-audio-gus-skip-automatic-zero-init-of-large-array.patch Normal file
View File

@ -0,0 +1,48 @@
From 6aac6e3888bd249856dc5bc91a40d9b4eb60f732 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:48 +0100
Subject: [PATCH 11/31] hw/audio/gus: skip automatic zero-init of large array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [10/30] 1af0f37dbbd3f3703988dfd6548e2ba018f05ee7
The 'GUS_read_DMA' method has a 4k byte array used for copying
data between the audio backend and device. Skip the automatic
zero-init of this array to eliminate the performance overhead in
the I/O hot path.
The 'tmpbuf' array will be fully initialized when reading data
from device memory.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-11-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 2e438da4929018c62609381e1156aac0b2fe3de3)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/audio/gus.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/audio/gus.c b/hw/audio/gus.c
index 4beb3fd74e..e8b0b85d44 100644
--- a/hw/audio/gus.c
+++ b/hw/audio/gus.c
@@ -183,7 +183,7 @@ static int GUS_read_DMA (void *opaque, int nchan, int dma_pos, int dma_len)
{
GUSState *s = opaque;
IsaDmaClass *k = ISADMA_GET_CLASS(s->isa_dma);
- char tmpbuf[4096];
+ QEMU_UNINITIALIZED char tmpbuf[4096];
int pos = dma_pos, mode, left = dma_len - dma_pos;
ldebug ("read DMA %#x %d\n", dma_pos, dma_len);
--
2.39.3

50
SOURCES/kvm-hw-audio-marvell_88w8618-skip-automatic-zero-init-of.patch Normal file
View File

@ -0,0 +1,50 @@
From ad0ae4edded2db9e7fea4c82cb22d47798b34528 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:49 +0100
Subject: [PATCH 12/31] hw/audio/marvell_88w8618: skip automatic zero-init of
large array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [11/30] 22754626ac42807e6306becab38fbaf564a84660
The 'mv88w8618_audio_callback' method has a 4k byte array used for
copying data between the audio backend and device. Skip the automatic
zero-init of this array to eliminate the performance overhead in
the I/O hot path.
The 'buf' array will be fully initialized when reading data from
device memory.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-12-berrange@redhat.com
[Fixed hw/audio/gus in commit message --Stefan]
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 5b6cd5c5df4229972d8a0fd9dd9a089a1644d6ba)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/audio/marvell_88w8618.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/audio/marvell_88w8618.c b/hw/audio/marvell_88w8618.c
index cc285444bc..b7b4b27272 100644
--- a/hw/audio/marvell_88w8618.c
+++ b/hw/audio/marvell_88w8618.c
@@ -66,7 +66,7 @@ static void mv88w8618_audio_callback(void *opaque, int free_out, int free_in)
{
mv88w8618_audio_state *s = opaque;
int16_t *codec_buffer;
- int8_t buf[4096];
+ QEMU_UNINITIALIZED int8_t buf[4096];
int8_t *mem_buffer;
int pos, block_size;
--
2.39.3

48
SOURCES/kvm-hw-audio-sb16-skip-automatic-zero-init-of-large-arra.patch Normal file
View File

@ -0,0 +1,48 @@
From 06915675e69c2ff0d1a686dcea200429f54fee74 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:50 +0100
Subject: [PATCH 13/31] hw/audio/sb16: skip automatic zero-init of large array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [12/30] 93e9b488f1ec99b425cc149c3aba8f824d000c73
The 'write_audio' method has a 4k byte array used for copying data
between the audio backend and device. Skip the automatic zero-init
of this array to eliminate the performance overhead in the I/O hot
path.
The 'tmpbuf' array will be fully initialized when reading data from
device memory.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-13-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 30c82f6657c1ee9fbb5473924b4d3273f214bd6f)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/audio/sb16.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/audio/sb16.c b/hw/audio/sb16.c
index fd76e78d18..04c818ed3d 100644
--- a/hw/audio/sb16.c
+++ b/hw/audio/sb16.c
@@ -1181,7 +1181,7 @@ static int write_audio (SB16State *s, int nchan, int dma_pos,
IsaDma *isa_dma = nchan == s->dma ? s->isa_dma : s->isa_hdma;
IsaDmaClass *k = ISADMA_GET_CLASS(isa_dma);
int temp, net;
- uint8_t tmpbuf[4096];
+ QEMU_UNINITIALIZED uint8_t tmpbuf[4096];
temp = len;
net = 0;
--
2.39.3

49
SOURCES/kvm-hw-audio-via-ac97-skip-automatic-zero-init-of-large-.patch Normal file
View File

@ -0,0 +1,49 @@
From 180cb8f07e5a7a1c4bfe01709b27cbf8d080f1a8 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:51 +0100
Subject: [PATCH 14/31] hw/audio/via-ac97: skip automatic zero-init of large
array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [13/30] d27c1e2fbd89df34bde1248d1b053eca40625838
The 'out_cb' method has a 4k byte array used for copying data
between the audio backend and device. Skip the automatic zero-init
of this array to eliminate the performance overhead in the I/O hot
path.
The 'tmpbuf' array will be fully initialized when reading data from
device memory.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-14-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit bb71d9fe1419f44529c91d1b09464718d157e647)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/audio/via-ac97.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/audio/via-ac97.c b/hw/audio/via-ac97.c
index 4c127a1def..e8fcf44e5d 100644
--- a/hw/audio/via-ac97.c
+++ b/hw/audio/via-ac97.c
@@ -175,7 +175,7 @@ static void out_cb(void *opaque, int avail)
ViaAC97SGDChannel *c = &s->aur;
int temp, to_copy, copied;
bool stop = false;
- uint8_t tmpbuf[4096];
+ QEMU_UNINITIALIZED uint8_t tmpbuf[4096];
if (c->stat & STAT_PAUSED) {
return;
--
2.39.3

49
SOURCES/kvm-hw-char-sclpconsole-lm-skip-automatic-zero-init-of-l.patch Normal file
View File

@ -0,0 +1,49 @@
From 03b3f8f230e30399e952c2c5eafde3efa7c015b3 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:52 +0100
Subject: [PATCH 15/31] hw/char/sclpconsole-lm: skip automatic zero-init of
large array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [14/30] fca3dbfc00277d1c35ecdb65a56d26c95c8cb6bb
The 'process_mdb' method has a 4k byte array used for copying data
between the guest and the chardev backend. Skip the automatic zero-init
of this array to eliminate the performance overhead in the I/O hot
path.
The 'buffer' array will be selectively initialized when data is converted
between EBCDIC and ASCII.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-15-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 8b1dac1ad57082611419b0e2f347acd96115d25f)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/char/sclpconsole-lm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/char/sclpconsole-lm.c b/hw/char/sclpconsole-lm.c
index 7719f438f6..19e64b92f6 100644
--- a/hw/char/sclpconsole-lm.c
+++ b/hw/char/sclpconsole-lm.c
@@ -214,7 +214,7 @@ static int process_mdb(SCLPEvent *event, MDBO *mdbo)
{
int rc;
int len;
- uint8_t buffer[SIZE_BUFFER];
+ QEMU_UNINITIALIZED uint8_t buffer[SIZE_BUFFER];
len = be16_to_cpu(mdbo->length);
len -= sizeof(mdbo->length) + sizeof(mdbo->type)
--
2.39.3

49
SOURCES/kvm-hw-display-vmware_vga-skip-automatic-zero-init-of-la.patch Normal file
View File

@ -0,0 +1,49 @@
From 4c88a2da13473491d30328f24239fe305ff037ef Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:54 +0100
Subject: [PATCH 17/31] hw/display/vmware_vga: skip automatic zero-init of
large struct
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [16/30] 7b1073ec54071782a9ad32e17293b141bee639c4
The 'vmsvga_fifo_run' method has a struct which is a little over 20k
in size, used for holding image data for cursor changes. Skip the
automatic zero-init of this struct to eliminate the performance
overhead in the I/O hot path.
The cursor variable will be fully initialized only when processing
a cursor definition message from the guest.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-17-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 7048e70f391df76d009eecca25f8027858f9f304)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/display/vmware_vga.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/display/vmware_vga.c b/hw/display/vmware_vga.c
index 3db3ff98f7..69afe98a2f 100644
--- a/hw/display/vmware_vga.c
+++ b/hw/display/vmware_vga.c
@@ -618,7 +618,7 @@ static void vmsvga_fifo_run(struct vmsvga_state_s *s)
uint32_t cmd, colour;
int args, len, maxloop = 1024;
int x, y, dx, dy, width, height;
- struct vmsvga_cursor_definition_s cursor;
+ QEMU_UNINITIALIZED struct vmsvga_cursor_definition_s cursor;
uint32_t cmd_start;
len = vmsvga_fifo_length(s);
--
2.39.3

47
SOURCES/kvm-hw-dma-xlnx_csu_dma-skip-automatic-zero-init-of-larg.patch Normal file
View File

@ -0,0 +1,47 @@
From 7208c85b822d45410d21ec42acdc872550a57262 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:53 +0100
Subject: [PATCH 16/31] hw/dma/xlnx_csu_dma: skip automatic zero-init of large
array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [15/30] 8ffc738b76ad72de3d0925c5e44408b0f712fcef
The 'xlnx_csu_dma_src_notify' method has a 4k byte array used for
copying DMA data. Skip the automatic zero-init of this array to
eliminate the performance overhead in the I/O hot path.
The 'buf' array will be fully initialized when data is copied.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-16-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit ce14f24611aa0469b464a9512e192b4fd51dca2b)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/dma/xlnx_csu_dma.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/dma/xlnx_csu_dma.c b/hw/dma/xlnx_csu_dma.c
index ae307482f2..9d1cccc5ca 100644
--- a/hw/dma/xlnx_csu_dma.c
+++ b/hw/dma/xlnx_csu_dma.c
@@ -287,7 +287,7 @@ static uint32_t xlnx_csu_dma_advance(XlnxCSUDMA *s, uint32_t len)
static void xlnx_csu_dma_src_notify(void *opaque)
{
XlnxCSUDMA *s = XLNX_CSU_DMA(opaque);
- unsigned char buf[4 * 1024];
+ QEMU_UNINITIALIZED unsigned char buf[4 * 1024];
size_t rlen = 0;
ptimer_transaction_begin(s->src_timer);
--
2.39.3

56
SOURCES/kvm-hw-hyperv-syndbg-skip-automatic-zero-init-of-large-a.patch Normal file
View File

@ -0,0 +1,56 @@
From c7b6fe3f924396dd49bdf13485696a536aa34fb0 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:55 +0100
Subject: [PATCH 18/31] hw/hyperv/syndbg: skip automatic zero-init of large
array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [17/30] 4202f998aff3fd784508ab6e52658cecd197924b
The 'handle_recv_msg' method has a 4k byte array used for copying
data between the network socket and guest memory. Skip the automatic
zero-init of this array to eliminate the performance overhead in the
I/O hot path.
The 'data_buf' array will be fully initialized when data is read
off the network socket.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-18-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 5a1f614d0cd0bcc8e84e0b7ab6af63d56bd348a2)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Conflicts:
hw/hyperv/syndbg.c
Context conflict due to missing commit 3efb9d226221
("hw/hyperv/syndbg: common compilation unit") downstream. There is no
need to backport the commit because it's not a bug fix.
---
hw/hyperv/syndbg.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/hyperv/syndbg.c b/hw/hyperv/syndbg.c
index 065e12fb1e..c7c43c8009 100644
--- a/hw/hyperv/syndbg.c
+++ b/hw/hyperv/syndbg.c
@@ -188,7 +188,7 @@ static uint16_t handle_recv_msg(HvSynDbg *syndbg, uint64_t outgpa,
uint64_t timeout, uint32_t *retrieved_count)
{
uint16_t ret;
- uint8_t data_buf[TARGET_PAGE_SIZE - UDP_PKT_HEADER_SIZE];
+ QEMU_UNINITIALIZED uint8_t data_buf[TARGET_PAGE_SIZE - UDP_PKT_HEADER_SIZE];
hwaddr out_len;
void *out_data;
ssize_t recv_byte_count;
--
2.39.3

87
SOURCES/kvm-hw-i386-Fix-machine-type-compatibility.patch Normal file
View File

@ -0,0 +1,87 @@
From 9bd4a89d3e1410b3a5994ab2b33f4332a4246955 Mon Sep 17 00:00:00 2001
From: Sebastian Ott <sebott@redhat.com>
Date: Thu, 15 May 2025 18:45:51 +0200
Subject: [PATCH] hw/i386: Fix machine type compatibility
RH-Author: Sebastian Ott <sebott@redhat.com>
RH-MergeRequest: 452: hw/i386: Fix machine type compatibility
RH-Jira: RHEL-92077
RH-Acked-by: Cornelia Huck <cohuck@redhat.com>
RH-Acked-by: Jon Maloy <jmaloy@redhat.com>
RH-Commit: [1/1] d594f142e8ce616b6fd1accd6950ab5cebc34984
Upstream Status: RHEL only
Ensure compatibility of rhel specific i440fx and q35 machine types.
Pick up missing bits from pc_compat_9_0 upstream.
Signed-off-by: Sebastian Ott <sebott@redhat.com>
---
hw/i386/pc.c | 8 ++++++++
hw/i386/pc_piix.c | 2 ++
hw/i386/pc_q35.c | 2 ++
include/hw/i386/pc.h | 3 +++
4 files changed, 15 insertions(+)
diff --git a/hw/i386/pc.c b/hw/i386/pc.c
index fa0e42d072..d8f1b2d899 100644
--- a/hw/i386/pc.c
+++ b/hw/i386/pc.c
@@ -298,6 +298,14 @@ GlobalProperty pc_rhel_compat[] = {
};
const size_t pc_rhel_compat_len = G_N_ELEMENTS(pc_rhel_compat);
+GlobalProperty pc_rhel_9_6_compat[] = {
+ /* pc_rhel_9_6_compat from pc_compat_9_0 */
+ { TYPE_X86_CPU, "x-amd-topoext-features-only", "false" },
+ { TYPE_X86_CPU, "x-l1-cache-per-thread", "false" },
+ { TYPE_X86_CPU, "legacy-multi-node", "on" },
+};
+const size_t pc_rhel_9_6_compat_len = G_N_ELEMENTS(pc_rhel_9_6_compat);
+
GlobalProperty pc_rhel_9_5_compat[] = {
/* pc_rhel_9_5_compat from pc_compat_pc_9_0 (backported from 9.1) */
{ TYPE_X86_CPU, "guest-phys-bits", "0" },
diff --git a/hw/i386/pc_piix.c b/hw/i386/pc_piix.c
index 656abb5d39..80d366bf17 100644
--- a/hw/i386/pc_piix.c
+++ b/hw/i386/pc_piix.c
@@ -885,6 +885,8 @@ static void pc_i440fx_rhel_machine_7_6_0_options(MachineClass *m)
compat_props_add(m->compat_props, hw_compat_rhel_9_6,
hw_compat_rhel_9_6_len);
+ compat_props_add(m->compat_props, pc_rhel_9_6_compat,
+ pc_rhel_9_6_compat_len);
compat_props_add(m->compat_props, pc_rhel_9_5_compat,
pc_rhel_9_5_compat_len);
compat_props_add(m->compat_props, hw_compat_rhel_9_5,
diff --git a/hw/i386/pc_q35.c b/hw/i386/pc_q35.c
index 578f63524f..e3653b44cd 100644
--- a/hw/i386/pc_q35.c
+++ b/hw/i386/pc_q35.c
@@ -701,6 +701,8 @@ static void pc_q35_rhel_machine_9_4_0_options(MachineClass *m)
compat_props_add(m->compat_props, hw_compat_rhel_9_6,
hw_compat_rhel_9_6_len);
+ compat_props_add(m->compat_props, pc_rhel_9_6_compat,
+ pc_rhel_9_6_compat_len);
compat_props_add(m->compat_props, pc_rhel_9_5_compat,
pc_rhel_9_5_compat_len);
compat_props_add(m->compat_props, hw_compat_rhel_9_5,
diff --git a/include/hw/i386/pc.h b/include/hw/i386/pc.h
index 75c9271cdd..2b7c18f2b0 100644
--- a/include/hw/i386/pc.h
+++ b/include/hw/i386/pc.h
@@ -305,6 +305,9 @@ extern const size_t pc_compat_2_3_len;
extern GlobalProperty pc_rhel_compat[];
extern const size_t pc_rhel_compat_len;
+extern GlobalProperty pc_rhel_9_6_compat[];
+extern const size_t pc_rhel_9_6_compat_len;
+
extern GlobalProperty pc_rhel_9_5_compat[];
extern const size_t pc_rhel_9_5_compat_len;
--
2.48.1

57
SOURCES/kvm-hw-misc-aspeed_hace-skip-automatic-zero-init-of-larg.patch Normal file
View File

@ -0,0 +1,57 @@
From ee75cefe77904ffc659bdb2df32feef7e01a914e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:56 +0100
Subject: [PATCH 19/31] hw/misc/aspeed_hace: skip automatic zero-init of large
array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [18/30] fb1e9ede27fca69cf9074c2590c221dce3633a68
The 'do_hash_operation' method has a 256 element iovec array used for
holding pointers to data that is to be hashed. Skip the automatic
zero-init of this array to eliminate the performance overhead in the
I/O hot path.
The 'iovec' array will be selectively initialized based on data that
needs to be hashed.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-19-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 6992c886838282f36b20deee44b666bbfc573a8f)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Conflicts:
hw/misc/aspeed_hace.c
Context conflict due to missing commit b9ccbe212e24
("hw/misc/aspeed_hace: Extract accumulation-mode hash execution into
helper function") downstream. The commit is not a bug fix, so there is
no need to backport it.
---
hw/misc/aspeed_hace.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/misc/aspeed_hace.c b/hw/misc/aspeed_hace.c
index c06c04ddc6..d2118f1864 100644
--- a/hw/misc/aspeed_hace.c
+++ b/hw/misc/aspeed_hace.c
@@ -188,7 +188,7 @@ static int gen_acc_mode_iov(AspeedHACEState *s, struct iovec *iov, int id,
static void do_hash_operation(AspeedHACEState *s, int algo, bool sg_mode,
bool acc_mode)
{
- struct iovec iov[ASPEED_HACE_MAX_SG];
+ QEMU_UNINITIALIZED struct iovec iov[ASPEED_HACE_MAX_SG];
g_autofree uint8_t *digest_buf = NULL;
size_t digest_len = 0;
int niov = 0;
--
2.39.3

48
SOURCES/kvm-hw-net-rtl8139-skip-automatic-zero-init-of-large-arr.patch Normal file
View File

@ -0,0 +1,48 @@
From 09fe29d40b3d1e30e0e921e186a797c4da5ac583 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:57 +0100
Subject: [PATCH 20/31] hw/net/rtl8139: skip automatic zero-init of large array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [19/30] 3c072f59b9a283b40327117585d9f01d32ecc081
The 'rtl8139_transmit_one' method has a 8k byte array used for
copying data between guest and host. Skip the automatic zero-init
of this array to eliminate the performance overhead in the I/O
hot path.
The 'txbuffer' will be fully initialized when reading PCI DMA
buffers.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-20-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 3ccc6489dd4925ddd1f3066bd3751389169cd7aa)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/net/rtl8139.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/net/rtl8139.c b/hw/net/rtl8139.c
index f2fe057535..a2732bf1c1 100644
--- a/hw/net/rtl8139.c
+++ b/hw/net/rtl8139.c
@@ -1818,7 +1818,7 @@ static int rtl8139_transmit_one(RTL8139State *s, int descriptor)
PCIDevice *d = PCI_DEVICE(s);
int txsize = s->TxStatus[descriptor] & 0x1fff;
- uint8_t txbuffer[0x2000];
+ QEMU_UNINITIALIZED uint8_t txbuffer[0x2000];
DPRINTF("+++ transmit reading %d bytes from host memory at 0x%08x\n",
txsize, s->TxAddr[descriptor]);
--
2.39.3

47
SOURCES/kvm-hw-net-tulip-skip-automatic-zero-init-of-large-array.patch Normal file
View File

@ -0,0 +1,47 @@
From f9a1a355dbd2d59bbd80d33e713579d93fe3932f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:58 +0100
Subject: [PATCH 21/31] hw/net/tulip: skip automatic zero-init of large array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [20/30] ff60b673f4e06a25fd5efbc51f5536da5d9c99f5
The 'tulip_setup_frame' method has a 4k byte array used for copynig
DMA data from the device. Skip the automatic zero-init of this array
to eliminate the performance overhead in the I/O hot path.
The 'buf' array will be fully initialized when reading data from the
device.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-21-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit e1afd5ee6eb2954f4baf3c97820e4aaf7de97d2a)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/net/tulip.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/net/tulip.c b/hw/net/tulip.c
index 1f2ef20977..5cf2b96fbd 100644
--- a/hw/net/tulip.c
+++ b/hw/net/tulip.c
@@ -629,7 +629,7 @@ static void tulip_setup_filter_addr(TULIPState *s, uint8_t *buf, int n)
static void tulip_setup_frame(TULIPState *s,
struct tulip_descriptor *desc)
{
- uint8_t buf[4096];
+ QEMU_UNINITIALIZED uint8_t buf[4096];
int len = (desc->control >> TDES1_BUF1_SIZE_SHIFT) & TDES1_BUF1_SIZE_MASK;
int i;
--
2.39.3

54
SOURCES/kvm-hw-net-virtio-net-skip-automatic-zero-init-of-large-.patch Normal file
View File

@ -0,0 +1,54 @@
From 9ecc539204dd6ab7a1124089f9e557248e321282 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:36:59 +0100
Subject: [PATCH 22/31] hw/net/virtio-net: skip automatic zero-init of large
arrays
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [21/30] f093a50cc162bd376fd74ea47c6274d7e718ba69
The 'virtio_net_receive_rcu' method has three arrays with
VIRTQUEUE_MAX_SIZE elements, which are apprixmately 32k in
size used for copying data between guest and host. Skip the
automatic zero-init of these arrays to eliminate the
performance overhead in the I/O hot path.
The three arrays will be selectively initialized as required
when processing network buffers.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-22-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 21cf31c51a7aeff4270c9b30b37e019c536d54b2)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/net/virtio-net.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/hw/net/virtio-net.c b/hw/net/virtio-net.c
index 3d2b2460ad..086ea20ea0 100644
--- a/hw/net/virtio-net.c
+++ b/hw/net/virtio-net.c
@@ -1895,9 +1895,9 @@ static ssize_t virtio_net_receive_rcu(NetClientState *nc, const uint8_t *buf,
VirtIONet *n = qemu_get_nic_opaque(nc);
VirtIONetQueue *q = virtio_net_get_subqueue(nc);
VirtIODevice *vdev = VIRTIO_DEVICE(n);
- VirtQueueElement *elems[VIRTQUEUE_MAX_SIZE];
- size_t lens[VIRTQUEUE_MAX_SIZE];
- struct iovec mhdr_sg[VIRTQUEUE_MAX_SIZE];
+ QEMU_UNINITIALIZED VirtQueueElement *elems[VIRTQUEUE_MAX_SIZE];
+ QEMU_UNINITIALIZED size_t lens[VIRTQUEUE_MAX_SIZE];
+ QEMU_UNINITIALIZED struct iovec mhdr_sg[VIRTQUEUE_MAX_SIZE];
struct virtio_net_hdr_v1_hash extra_hdr;
unsigned mhdr_cnt = 0;
size_t offset, i, guest_offset, j;
--
2.39.3

47
SOURCES/kvm-hw-net-xgamc-skip-automatic-zero-init-of-large-array.patch Normal file
View File

@ -0,0 +1,47 @@
From 3b39fa3e031d5b8a89c05302f3d73e7d4748bf58 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:37:00 +0100
Subject: [PATCH 23/31] hw/net/xgamc: skip automatic zero-init of large array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [22/30] d83a91284970ced6f60964ded15d04405845e8bb
The 'xgmac_enet_send' method has a 8k byte array used for copying
data between guest and host. Skip the automatic zero-init of this
array to eliminate the performance overhead in the I/O hot path.
The 'frame' buffer will be fully initialized when reading guest
memory to fetch the data to send.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-23-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 8b723287b84a62bb5d1a7799ef0959ca8e6c293a)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/net/xgmac.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/net/xgmac.c b/hw/net/xgmac.c
index ffe3fc8dbe..eff8022aca 100644
--- a/hw/net/xgmac.c
+++ b/hw/net/xgmac.c
@@ -207,7 +207,7 @@ static void xgmac_enet_send(XgmacState *s)
struct desc bd;
int frame_size;
int len;
- uint8_t frame[8192];
+ QEMU_UNINITIALIZED uint8_t frame[8192];
uint8_t *ptr;
ptr = frame;
--
2.39.3

72
SOURCES/kvm-hw-nvme-ctrl-skip-automatic-zero-init-of-large-array.patch Normal file
View File

@ -0,0 +1,72 @@
From 794d838efddc7e96f6e40c1c4bb2b1baf3c95cfb Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:37:01 +0100
Subject: [PATCH 24/31] hw/nvme/ctrl: skip automatic zero-init of large arrays
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [23/30] 97877d2e280daf654f5893461c0bc9e6f6caa77d
The 'nvme_map_sgl' method has a 256 element array used for copying
data from the device. Skip the automatic zero-init of this array
to eliminate the performance overhead in the I/O hot path.
The 'segment' array will be fully initialized when reading data from
the device.
The 'nme_changed_nslist' method has a 4k byte array that is manually
initialized with memset(). The compiler ought to be intelligent
enough to turn the memset() into a static initialization operation,
and thus not duplicate the automatic zero-init. Replacing memset()
with '{}' makes it unambiguous that the array is statically initialized.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Message-id: 20250610123709.835102-24-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 7eeb1d3acc175813ad3d5e824f26123e0992093a)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/nvme/ctrl.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
index 9f277b81d8..f000e2246f 100644
--- a/hw/nvme/ctrl.c
+++ b/hw/nvme/ctrl.c
@@ -1047,7 +1047,8 @@ static uint16_t nvme_map_sgl(NvmeCtrl *n, NvmeSg *sg, NvmeSglDescriptor sgl,
*/
#define SEG_CHUNK_SIZE 256
- NvmeSglDescriptor segment[SEG_CHUNK_SIZE], *sgld, *last_sgld;
+ QEMU_UNINITIALIZED NvmeSglDescriptor segment[SEG_CHUNK_SIZE];
+ NvmeSglDescriptor *sgld, *last_sgld;
uint64_t nsgld;
uint32_t seg_len;
uint16_t status;
@@ -5029,7 +5030,7 @@ static uint16_t nvme_error_info(NvmeCtrl *n, uint8_t rae, uint32_t buf_len,
static uint16_t nvme_changed_nslist(NvmeCtrl *n, uint8_t rae, uint32_t buf_len,
uint64_t off, NvmeRequest *req)
{
- uint32_t nslist[1024];
+ uint32_t nslist[1024] = {};
uint32_t trans_len;
int i = 0;
uint32_t nsid;
@@ -5039,7 +5040,6 @@ static uint16_t nvme_changed_nslist(NvmeCtrl *n, uint8_t rae, uint32_t buf_len,
return NVME_INVALID_FIELD | NVME_DNR;
}
- memset(nslist, 0x0, sizeof(nslist));
trans_len = MIN(sizeof(nslist) - off, buf_len);
while ((nsid = find_first_bit(n->changed_nsids, NVME_CHANGED_NSID_SIZE)) !=
--
2.39.3

52
SOURCES/kvm-hw-ppc-spapr_tpm_proxy-skip-automatic-zero-init-of-l.patch Normal file
View File

@ -0,0 +1,52 @@
From 087151816f810052c013d496e32be1011e5c01ef Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:37:03 +0100
Subject: [PATCH 25/31] hw/ppc/spapr_tpm_proxy: skip automatic zero-init of
large arrays
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [24/30] a2360ae956c03481af7aceb34d26c7d8ba33a1d7
The 'tpm_execute' method has a pair of 4k arrays used for copying
data between guest and host. Skip the automatic zero-init of these
arrays to eliminate the performance overhead in the I/O hot path.
The two arrays will be fully initialized when reading data from
guest memory or reading data from the proxy FD.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-26-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 5dd9087fff74b5672526cad254e76f790fb35c7a)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/ppc/spapr_tpm_proxy.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/hw/ppc/spapr_tpm_proxy.c b/hw/ppc/spapr_tpm_proxy.c
index e10af35a18..88833d9e2e 100644
--- a/hw/ppc/spapr_tpm_proxy.c
+++ b/hw/ppc/spapr_tpm_proxy.c
@@ -41,8 +41,8 @@ static ssize_t tpm_execute(SpaprTpmProxy *tpm_proxy, target_ulong *args)
target_ulong data_in_size = args[2];
uint64_t data_out = ppc64_phys_to_real(args[3]);
target_ulong data_out_size = args[4];
- uint8_t buf_in[TPM_SPAPR_BUFSIZE];
- uint8_t buf_out[TPM_SPAPR_BUFSIZE];
+ QEMU_UNINITIALIZED uint8_t buf_in[TPM_SPAPR_BUFSIZE];
+ QEMU_UNINITIALIZED uint8_t buf_out[TPM_SPAPR_BUFSIZE];
ssize_t ret;
trace_spapr_tpm_execute(data_in, data_in_size, data_out, data_out_size);
--
2.39.3

49
SOURCES/kvm-hw-scsi-lsi53c895a-skip-automatic-zero-init-of-large.patch Normal file
View File

@ -0,0 +1,49 @@
From 3179e7e183295b91e314c32db8697d8cf0947367 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:37:05 +0100
Subject: [PATCH 27/31] hw/scsi/lsi53c895a: skip automatic zero-init of large
array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [26/30] 92ec0b18f26956767d2f3d0284d712696d60852e
The 'lsi_memcpy' method has a 4k byte array used for copying data
to/from the device. Skip the automatic zero-init of this array to
eliminate the performance overhead in the I/O hot path.
The 'buf' array will be fully initialized when data is copied.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-28-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 55243edf42ee87bce9f36ca251f3ab9cda1563e4)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/scsi/lsi53c895a.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/scsi/lsi53c895a.c b/hw/scsi/lsi53c895a.c
index f1935e5328..f165705f8a 100644
--- a/hw/scsi/lsi53c895a.c
+++ b/hw/scsi/lsi53c895a.c
@@ -1112,7 +1112,7 @@ bad:
static void lsi_memcpy(LSIState *s, uint32_t dest, uint32_t src, int count)
{
int n;
- uint8_t buf[LSI_BUF_SIZE];
+ QEMU_UNINITIALIZED uint8_t buf[LSI_BUF_SIZE];
trace_lsi_memcpy(dest, src, count);
while (count) {
--
2.39.3

73
SOURCES/kvm-hw-scsi-megasas-skip-automatic-zero-init-of-large-ar.patch Normal file
View File

@ -0,0 +1,73 @@
From 30a938409f664179061b039796354628c714229e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:37:06 +0100
Subject: [PATCH 28/31] hw/scsi/megasas: skip automatic zero-init of large
arrays
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [27/30] 70c3c002601bc54ff81d000092e7be2bdb1eb82a
The 'megasas_dcmd_pd_get_list' and 'megasas_dcmd_get_properties'
methods have 4k structs used for copying data from the device.
Skip the automatic zero-init of this array to eliminate the
performance overhead in the I/O hot path.
The 'info' structs are manually initialized with memset(). The
compiler ought to be intelligent enough to turn the memset()
into a static initialization operation, and thus not duplicate
the automatic zero-init. Replacing memset() with '{}' makes it
unambiguous that the arrays are statically initialized.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-29-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit ca0559e2350c618048f7caf80cb79c1259e7cfd2)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/scsi/megasas.c | 7 ++-----
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/hw/scsi/megasas.c b/hw/scsi/megasas.c
index 2d0c607177..91b65accbc 100644
--- a/hw/scsi/megasas.c
+++ b/hw/scsi/megasas.c
@@ -981,13 +981,11 @@ static int megasas_event_wait(MegasasState *s, MegasasCmd *cmd)
static int megasas_dcmd_pd_get_list(MegasasState *s, MegasasCmd *cmd)
{
- struct mfi_pd_list info;
- size_t dcmd_size = sizeof(info);
+ struct mfi_pd_list info = {};
BusChild *kid;
uint32_t offset, dcmd_limit, num_pd_disks = 0, max_pd_disks;
dma_addr_t residual;
- memset(&info, 0, dcmd_size);
offset = 8;
dcmd_limit = offset + sizeof(struct mfi_pd_address);
if (cmd->iov_size < dcmd_limit) {
@@ -1429,11 +1427,10 @@ static int megasas_dcmd_cfg_read(MegasasState *s, MegasasCmd *cmd)
static int megasas_dcmd_get_properties(MegasasState *s, MegasasCmd *cmd)
{
- struct mfi_ctrl_props info;
+ struct mfi_ctrl_props info = {};
size_t dcmd_size = sizeof(info);
dma_addr_t residual;
- memset(&info, 0x0, dcmd_size);
if (cmd->iov_size < dcmd_size) {
trace_megasas_dcmd_invalid_xfer_len(cmd->index, cmd->iov_size,
dcmd_size);
--
2.39.3

50
SOURCES/kvm-hw-ufs-lu-skip-automatic-zero-init-of-large-array.patch Normal file
View File

@ -0,0 +1,50 @@
From 36046b3119eb2338c811f43b2956c3aa787a2e3c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:37:07 +0100
Subject: [PATCH 29/31] hw/ufs/lu: skip automatic zero-init of large array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [28/30] 4d68ab4596b8fb97106d07e5af11dc3fdcb25e96
The 'ufs_emulate_scsi_cmd' method has a 4k byte array used for
copying data from the device. Skip the automatic zero-init of
this array to eliminate the performance overhead in the I/O hot
path.
The 'outbuf' array will be fully initialized when data is copied
from the guest.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-30-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 7708e298180550eac262c1fd742e6e80c711a5d8)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/ufs/lu.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/ufs/lu.c b/hw/ufs/lu.c
index 81bfff9b4e..caad82dcc4 100644
--- a/hw/ufs/lu.c
+++ b/hw/ufs/lu.c
@@ -194,7 +194,7 @@ static int ufs_emulate_wlun_inquiry(UfsRequest *req, uint8_t *outbuf,
static UfsReqResult ufs_emulate_scsi_cmd(UfsLu *lu, UfsRequest *req)
{
uint8_t lun = lu->lun;
- uint8_t outbuf[4096];
+ QEMU_UNINITIALIZED uint8_t outbuf[4096];
uint8_t sense_buf[UFS_SENSE_SIZE];
uint8_t scsi_status;
int len = 0;
--
2.39.3

50
SOURCES/kvm-hw-usb-hcd-ohci-skip-automatic-zero-init-of-large-ar.patch Normal file
View File

@ -0,0 +1,50 @@
From 31886b02875b9d2d61710c14d6cdc0ab20d6dfc5 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:37:04 +0100
Subject: [PATCH 26/31] hw/usb/hcd-ohci: skip automatic zero-init of large
array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [25/30] a69a9e0467a1b695b7fa07cdf30733b709867bd2
The 'ohci_service_iso_td' method has a 8k byte array used for copying
data between guest and host. Skip the automatic zero-init of this
array to eliminate the performance overhead in the I/O hot path.
The 'buf' array will be fully initialized when reading data from guest
memory.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-27-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 14997d521d1cd0bb36c902ef1032f0d3f2a3c912)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/usb/hcd-ohci.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/usb/hcd-ohci.c b/hw/usb/hcd-ohci.c
index 71b54914d3..72a9f9f474 100644
--- a/hw/usb/hcd-ohci.c
+++ b/hw/usb/hcd-ohci.c
@@ -577,7 +577,7 @@ static int ohci_service_iso_td(OHCIState *ohci, struct ohci_ed *ed)
USBDevice *dev;
USBEndpoint *ep;
USBPacket *pkt;
- uint8_t buf[8192];
+ QEMU_UNINITIALIZED uint8_t buf[8192];
bool int_req;
struct ohci_iso_td iso_td;
uint32_t addr;
--
2.39.3

73
SOURCES/kvm-hw-virtio-virtio-avoid-cost-of-ftrivial-auto-var-ini.patch Normal file
View File

@ -0,0 +1,73 @@
From ea242d728ed1716602b4cdd01d3fabb5ab260781 Mon Sep 17 00:00:00 2001
From: Stefan Hajnoczi <stefanha@redhat.com>
Date: Tue, 10 Jun 2025 13:36:40 +0100
Subject: [PATCH 03/31] hw/virtio/virtio: avoid cost of -ftrivial-auto-var-init
in hot path
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [2/30] a90fb4e14c182ace7a28c8335858895b4257f37b
Since commit 7ff9ff039380 ("meson: mitigate against use of uninitialize
stack for exploits") the -ftrivial-auto-var-init=zero compiler option is
used to zero local variables. While this reduces security risks
associated with uninitialized stack data, it introduced a measurable
bottleneck in the virtqueue_split_pop() and virtqueue_packed_pop()
functions.
These virtqueue functions are in the hot path. They are called for each
element (request) that is popped from a VIRTIO device's virtqueue. Using
__attribute__((uninitialized)) on large stack variables in these
functions improves fio randread bs=4k iodepth=64 performance from 304k
to 332k IOPS (+9%).
This issue was found using perf-top(1). virtqueue_split_pop() was one of
the top CPU consumers and the "annotate" feature showed that the memory
zeroing instructions at the beginning of the functions were hot.
Fixes: 7ff9ff039380 ("meson: mitigate against use of uninitialize stack for exploits")
Cc: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20250610123709.835102-3-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit ba2868ce091cd4abe4be6de4b7e44b3be303b352)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
hw/virtio/virtio.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index 10f24a58dd..7f7b178a50 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -1680,8 +1680,8 @@ static void *virtqueue_split_pop(VirtQueue *vq, size_t sz)
VirtIODevice *vdev = vq->vdev;
VirtQueueElement *elem = NULL;
unsigned out_num, in_num, elem_entries;
- hwaddr addr[VIRTQUEUE_MAX_SIZE];
- struct iovec iov[VIRTQUEUE_MAX_SIZE];
+ hwaddr QEMU_UNINITIALIZED addr[VIRTQUEUE_MAX_SIZE];
+ struct iovec QEMU_UNINITIALIZED iov[VIRTQUEUE_MAX_SIZE];
VRingDesc desc;
int rc;
@@ -1826,8 +1826,8 @@ static void *virtqueue_packed_pop(VirtQueue *vq, size_t sz)
VirtIODevice *vdev = vq->vdev;
VirtQueueElement *elem = NULL;
unsigned out_num, in_num, elem_entries;
- hwaddr addr[VIRTQUEUE_MAX_SIZE];
- struct iovec iov[VIRTQUEUE_MAX_SIZE];
+ hwaddr QEMU_UNINITIALIZED addr[VIRTQUEUE_MAX_SIZE];
+ struct iovec QEMU_UNINITIALIZED iov[VIRTQUEUE_MAX_SIZE];
VRingPackedDesc desc;
uint16_t id;
int rc;
--
2.39.3

80
SOURCES/kvm-include-qemu-compiler-add-QEMU_UNINITIALIZED-attribu.patch Normal file
View File

@ -0,0 +1,80 @@
From 73f85b945f09ae118f2c1479110f2e34906e084b Mon Sep 17 00:00:00 2001
From: Stefan Hajnoczi <stefanha@redhat.com>
Date: Tue, 10 Jun 2025 13:36:39 +0100
Subject: [PATCH 02/31] include/qemu/compiler: add QEMU_UNINITIALIZED attribute
macro
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [1/30] 6b6151625fdf6636cbf352731906f643e2fbfd35
The QEMU_UNINITIALIZED macro is to be used to skip the default compiler
variable initialization done by -ftrivial-auto-var-init=zero.
Use this in cases where there a method in the device I/O path (or other
important hot paths), that has large variables on the stack. A rule of
thumb is that "large" means a method with 4kb data in the local stack
frame. Any variables which are KB in size, should be annotated with this
attribute, to pre-emptively eliminate any potential overhead from the
compiler zero'ing memory.
Given that this turns off a security hardening feature, when using this
to flag variables, it is important that the code is double-checked to
ensure there is no possible use of uninitialized data in the method.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20250610123709.835102-2-berrange@redhat.com
[DB: split off patch & rewrite guidance on when to use the annotation]
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit c653b67d1863b7ebfa67f7c9f4aec209d7b5ced5)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Conflicts:
include/qemu/compiler.h
Context conflict due to clang Thread Safety Analysis macros.
---
include/qemu/compiler.h | 20 ++++++++++++++++++++
1 file changed, 20 insertions(+)
diff --git a/include/qemu/compiler.h b/include/qemu/compiler.h
index c06954ccb4..cc193d5b82 100644
--- a/include/qemu/compiler.h
+++ b/include/qemu/compiler.h
@@ -212,6 +212,26 @@
# define QEMU_USED
#endif
+/*
+ * Disable -ftrivial-auto-var-init on a local variable.
+ *
+ * Use this in cases where there a method in the device I/O path (or other
+ * important hot paths), that has large variables on the stack. A rule of
+ * thumb is that "large" means a method with 4kb data in the local stack
+ * frame. Any variables which are KB in size, should be annotated with this
+ * attribute, to pre-emptively eliminate any potential overhead from the
+ * compiler's implicit zero'ing of memory.
+ *
+ * Given that this turns off a security hardening feature, when using this
+ * to flag variables, it is important that the code is double-checked to
+ * ensure there is no possible use of uninitialized data in the method.
+ */
+#if __has_attribute(uninitialized)
+# define QEMU_UNINITIALIZED __attribute__((uninitialized))
+#else
+# define QEMU_UNINITIALIZED
+#endif
+
/*
* Ugly CPP trick that is like "defined FOO", but also works in C
* code. Useful to replace #ifdef with "if" statements; assumes
--
2.39.3

49
SOURCES/kvm-net-socket-skip-automatic-zero-init-of-large-array.patch Normal file
View File

@ -0,0 +1,49 @@
From 9a941183c365f4c6698c46aa13594e9bb360e8e9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:37:08 +0100
Subject: [PATCH 30/31] net/socket: skip automatic zero-init of large array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [29/30] e7f35ae85a0ca50367487b1e7ed92b395b247cda
The 'net_socket_send' method has a 68k byte array used for copying
data between guest and host. Skip the automatic zero-init of this
array to eliminate the performance overhead in the I/O hot path.
The 'buf1' array will be fully initialized when reading data off
the network socket.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-31-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 751b0e79f1e0e7f88fad2fe2f22595ad03d78859)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
net/socket.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/socket.c b/net/socket.c
index 8e3702e1f3..784dda686f 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -157,7 +157,7 @@ static void net_socket_send(void *opaque)
NetSocketState *s = opaque;
int size;
int ret;
- uint8_t buf1[NET_BUFSIZE];
+ QEMU_UNINITIALIZED uint8_t buf1[NET_BUFSIZE];
const uint8_t *buf;
size = recv(s->fd, buf1, sizeof(buf1), 0);
--
2.39.3

49
SOURCES/kvm-net-stream-skip-automatic-zero-init-of-large-array.patch Normal file
View File

@ -0,0 +1,49 @@
From 790b862712841e4b363874e00f4da993ae045a53 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Daniel=20P=2E=20Berrang=C3=A9?= <berrange@redhat.com>
Date: Tue, 10 Jun 2025 13:37:09 +0100
Subject: [PATCH 31/31] net/stream: skip automatic zero-init of large array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Stefan Hajnoczi <stefanha@redhat.com>
RH-MergeRequest: 461: Solve -ftrivial-auto-var-init performance regression with QEMU_UNINITIALIZED
RH-Jira: RHEL-99887
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Commit: [30/30] 005a9dfc74c8fed3d1aa2e0cdc0473f67f888249
The 'net_stream_send' method has a 68k byte array used for copying
data between guest and host. Skip the automatic zero-init of this
array to eliminate the performance overhead in the I/O hot path.
The 'buf1' array will be fully initialized when reading data off
the network socket.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Message-id: 20250610123709.835102-32-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
(cherry picked from commit 837b87c4c5ba9ac7a255133c6642b8d578272a70)
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
net/stream.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/stream.c b/net/stream.c
index 97e6ec6679..12384ffee5 100644
--- a/net/stream.c
+++ b/net/stream.c
@@ -148,7 +148,7 @@ static gboolean net_stream_send(QIOChannel *ioc,
NetStreamState *s = data;
int size;
int ret;
- char buf1[NET_BUFSIZE];
+ QEMU_UNINITIALIZED char buf1[NET_BUFSIZE];
const char *buf;
size = qio_channel_read(s->ioc, buf1, sizeof(buf1), NULL);
--
2.39.3

97
SOURCES/kvm-ui-vnc-Update-display-update-interval-when-VM-state-.patch Normal file
View File

@ -0,0 +1,97 @@
From e2931430d3f10dd521e6b4cc7505842fbc8296ec Mon Sep 17 00:00:00 2001
From: Juraj Marcin <jmarcin@redhat.com>
Date: Wed, 21 May 2025 17:16:13 +0200
Subject: [PATCH 01/31] ui/vnc: Update display update interval when VM state
changes to RUNNING
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Juraj Marcin <None>
RH-MergeRequest: 463: ui/vnc: Update display update interval when VM state changes to RUNNING
RH-Jira: RHEL-100767
RH-Acked-by: Peter Xu <peterx@redhat.com>
RH-Acked-by: Marc-André Lureau <marcandre.lureau@redhat.com>
RH-Commit: [1/1] 60b1a7921296e82b616d055691fe8ac0f2e283b1
If a virtual machine is paused for an extended period time, for example,
due to an incoming migration, there are also no changes on the screen.
VNC in such case increases the display update interval by
VNC_REFRESH_INTERVAL_INC (50 ms). The update interval can then grow up
to VNC_REFRESH_INTERVAL_MAX (3000 ms).
When the machine resumes, it can then take up to 3 seconds for the first
display update. Furthermore, the update interval is then halved with
each display update with changes on the screen. If there are moving
elements on the screen, such as a video, this can be perceived as
freezing and stuttering for few seconds before the movement is smooth
again.
This patch resolves this issue, by adding a listener to VM state changes
and changing the update interval when the VM state changes to RUNNING.
The update_displaychangelistener() function updates the internal timer,
and the display is refreshed immediately if the timer is expired.
Signed-off-by: Juraj Marcin <jmarcin@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Link: https://lore.kernel.org/r/20250521151616.3951178-1-jmarcin@redhat.com
Signed-off-by: Peter Xu <peterx@redhat.com>
(cherry picked from commit 0310d594d98b39f9dde79b87fd8b0ad16e7c5459)
JIRA: https://issues.redhat.com/browse/RHEL-100767
Signed-off-by: Juraj Marcin <jmarcin@redhat.com>
---
ui/vnc.c | 12 ++++++++++++
ui/vnc.h | 2 ++
2 files changed, 14 insertions(+)
diff --git a/ui/vnc.c b/ui/vnc.c
index 5057ec8680..4afc925a18 100644
--- a/ui/vnc.c
+++ b/ui/vnc.c
@@ -3386,6 +3386,16 @@ static const DisplayChangeListenerOps dcl_ops = {
.dpy_cursor_define = vnc_dpy_cursor_define,
};
+static void vmstate_change_handler(void *opaque, bool running, RunState state)
+{
+ VncDisplay *vd = opaque;
+
+ if (state != RUN_STATE_RUNNING) {
+ return;
+ }
+ update_displaychangelistener(&vd->dcl, VNC_REFRESH_INTERVAL_BASE);
+}
+
void vnc_display_init(const char *id, Error **errp)
{
VncDisplay *vd;
@@ -3422,6 +3432,8 @@ void vnc_display_init(const char *id, Error **errp)
vd->dcl.ops = &dcl_ops;
register_displaychangelistener(&vd->dcl);
vd->kbd = qkbd_state_init(vd->dcl.con);
+ vd->vmstate_handler_entry = qemu_add_vm_change_state_handler(
+ &vmstate_change_handler, vd);
}
diff --git a/ui/vnc.h b/ui/vnc.h
index e5fa2efa3e..e9da707dbc 100644
--- a/ui/vnc.h
+++ b/ui/vnc.h
@@ -186,6 +186,8 @@ struct VncDisplay
#endif
AudioState *audio_state;
+
+ VMChangeStateEntry *vmstate_handler_entry;
};
typedef struct VncTight {
--
2.39.3

118
SPECS/qemu-kvm.spec
View File

@ -149,7 +149,7 @@ Obsoletes: %{name}-block-ssh <= %{epoch}:%{version} \
Summary: QEMU is a machine emulator and virtualizer
Name: qemu-kvm
Version: 9.1.0
Release: 15%{?rcrel}%{?dist}%{?cc_suffix}.4
Release: 15%{?rcrel}%{?dist}%{?cc_suffix}.7
# Epoch because we pushed a qemu-1.0 package. AIUI this can't ever be dropped
# Epoch 15 used for RHEL 8
# Epoch 17 used for RHEL 9 (due to release versioning offset in RHEL 8.5)
@ -475,6 +475,74 @@ Patch151: kvm-file-posix-probe-discard-alignment-on-Linux-block-de.patch
Patch152: kvm-block-io-skip-head-tail-requests-on-EINVAL.patch
# For RHEL-87734 - QEMU sends unaligned discards on 4K devices [rhel-9.6.z]
Patch153: kvm-file-posix-Fix-crash-on-discard_granularity-0.patch
# For RHEL-92077 - Fix x86 M-type compats [rhel-9.6.z]
Patch154: kvm-hw-i386-Fix-machine-type-compatibility.patch
# For RHEL-95407 - Support multipath failover with scsi-block [rhel-9.6.z]
Patch155: kvm-file-posix-Define-DM_MPATH_PROBE_PATHS.patch
# For RHEL-95407 - Support multipath failover with scsi-block [rhel-9.6.z]
Patch156: kvm-file-posix-Probe-paths-and-retry-SG_IO-on-potential-.patch
# For RHEL-100767 - Video stuck after switchover phase when play one video during migration [rhel-9.6.z]
Patch157: kvm-ui-vnc-Update-display-update-interval-when-VM-state-.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch158: kvm-include-qemu-compiler-add-QEMU_UNINITIALIZED-attribu.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch159: kvm-hw-virtio-virtio-avoid-cost-of-ftrivial-auto-var-ini.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch160: kvm-block-skip-automatic-zero-init-of-large-array-in-ioq.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch161: kvm-chardev-char-fd-skip-automatic-zero-init-of-large-ar.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch162: kvm-chardev-char-pty-skip-automatic-zero-init-of-large-a.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch163: kvm-chardev-char-socket-skip-automatic-zero-init-of-larg.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch164: kvm-hw-audio-ac97-skip-automatic-zero-init-of-large-arra.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch165: kvm-hw-audio-cs4231a-skip-automatic-zero-init-of-large-a.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch166: kvm-hw-audio-es1370-skip-automatic-zero-init-of-large-ar.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch167: kvm-hw-audio-gus-skip-automatic-zero-init-of-large-array.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch168: kvm-hw-audio-marvell_88w8618-skip-automatic-zero-init-of.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch169: kvm-hw-audio-sb16-skip-automatic-zero-init-of-large-arra.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch170: kvm-hw-audio-via-ac97-skip-automatic-zero-init-of-large-.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch171: kvm-hw-char-sclpconsole-lm-skip-automatic-zero-init-of-l.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch172: kvm-hw-dma-xlnx_csu_dma-skip-automatic-zero-init-of-larg.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch173: kvm-hw-display-vmware_vga-skip-automatic-zero-init-of-la.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch174: kvm-hw-hyperv-syndbg-skip-automatic-zero-init-of-large-a.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch175: kvm-hw-misc-aspeed_hace-skip-automatic-zero-init-of-larg.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch176: kvm-hw-net-rtl8139-skip-automatic-zero-init-of-large-arr.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch177: kvm-hw-net-tulip-skip-automatic-zero-init-of-large-array.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch178: kvm-hw-net-virtio-net-skip-automatic-zero-init-of-large-.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch179: kvm-hw-net-xgamc-skip-automatic-zero-init-of-large-array.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch180: kvm-hw-nvme-ctrl-skip-automatic-zero-init-of-large-array.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch181: kvm-hw-ppc-spapr_tpm_proxy-skip-automatic-zero-init-of-l.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch182: kvm-hw-usb-hcd-ohci-skip-automatic-zero-init-of-large-ar.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch183: kvm-hw-scsi-lsi53c895a-skip-automatic-zero-init-of-large.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch184: kvm-hw-scsi-megasas-skip-automatic-zero-init-of-large-ar.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch185: kvm-hw-ufs-lu-skip-automatic-zero-init-of-large-array.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch186: kvm-net-socket-skip-automatic-zero-init-of-large-array.patch
# For RHEL-99887 - -ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z]
Patch187: kvm-net-stream-skip-automatic-zero-init-of-large-array.patch
%if %{have_clang}
BuildRequires: clang
@ -1541,6 +1609,54 @@ useradd -r -u 107 -g qemu -G kvm -d / -s /sbin/nologin \
%endif
%changelog
* Fri Jul 04 2025 Miroslav Rezanina <mrezanin@redhat.com> - 9.1.0-15.el9_6.7
- kvm-ui-vnc-Update-display-update-interval-when-VM-state-.patch [RHEL-100767]
- kvm-include-qemu-compiler-add-QEMU_UNINITIALIZED-attribu.patch [RHEL-99887]
- kvm-hw-virtio-virtio-avoid-cost-of-ftrivial-auto-var-ini.patch [RHEL-99887]
- kvm-block-skip-automatic-zero-init-of-large-array-in-ioq.patch [RHEL-99887]
- kvm-chardev-char-fd-skip-automatic-zero-init-of-large-ar.patch [RHEL-99887]
- kvm-chardev-char-pty-skip-automatic-zero-init-of-large-a.patch [RHEL-99887]
- kvm-chardev-char-socket-skip-automatic-zero-init-of-larg.patch [RHEL-99887]
- kvm-hw-audio-ac97-skip-automatic-zero-init-of-large-arra.patch [RHEL-99887]
- kvm-hw-audio-cs4231a-skip-automatic-zero-init-of-large-a.patch [RHEL-99887]
- kvm-hw-audio-es1370-skip-automatic-zero-init-of-large-ar.patch [RHEL-99887]
- kvm-hw-audio-gus-skip-automatic-zero-init-of-large-array.patch [RHEL-99887]
- kvm-hw-audio-marvell_88w8618-skip-automatic-zero-init-of.patch [RHEL-99887]
- kvm-hw-audio-sb16-skip-automatic-zero-init-of-large-arra.patch [RHEL-99887]
- kvm-hw-audio-via-ac97-skip-automatic-zero-init-of-large-.patch [RHEL-99887]
- kvm-hw-char-sclpconsole-lm-skip-automatic-zero-init-of-l.patch [RHEL-99887]
- kvm-hw-dma-xlnx_csu_dma-skip-automatic-zero-init-of-larg.patch [RHEL-99887]
- kvm-hw-display-vmware_vga-skip-automatic-zero-init-of-la.patch [RHEL-99887]
- kvm-hw-hyperv-syndbg-skip-automatic-zero-init-of-large-a.patch [RHEL-99887]
- kvm-hw-misc-aspeed_hace-skip-automatic-zero-init-of-larg.patch [RHEL-99887]
- kvm-hw-net-rtl8139-skip-automatic-zero-init-of-large-arr.patch [RHEL-99887]
- kvm-hw-net-tulip-skip-automatic-zero-init-of-large-array.patch [RHEL-99887]
- kvm-hw-net-virtio-net-skip-automatic-zero-init-of-large-.patch [RHEL-99887]
- kvm-hw-net-xgamc-skip-automatic-zero-init-of-large-array.patch [RHEL-99887]
- kvm-hw-nvme-ctrl-skip-automatic-zero-init-of-large-array.patch [RHEL-99887]
- kvm-hw-ppc-spapr_tpm_proxy-skip-automatic-zero-init-of-l.patch [RHEL-99887]
- kvm-hw-usb-hcd-ohci-skip-automatic-zero-init-of-large-ar.patch [RHEL-99887]
- kvm-hw-scsi-lsi53c895a-skip-automatic-zero-init-of-large.patch [RHEL-99887]
- kvm-hw-scsi-megasas-skip-automatic-zero-init-of-large-ar.patch [RHEL-99887]
- kvm-hw-ufs-lu-skip-automatic-zero-init-of-large-array.patch [RHEL-99887]
- kvm-net-socket-skip-automatic-zero-init-of-large-array.patch [RHEL-99887]
- kvm-net-stream-skip-automatic-zero-init-of-large-array.patch [RHEL-99887]
- Resolves: RHEL-100767
(Video stuck after switchover phase when play one video during migration [rhel-9.6.z])
- Resolves: RHEL-99887
(-ftrivial-auto-var-init=zero reduced performance [rhel-9.6.z])
* Mon Jun 09 2025 Jon Maloy <jmaloy@redhat.com> - 9.1.0-15.el9_6.6
- kvm-file-posix-Define-DM_MPATH_PROBE_PATHS.patch [RHEL-95407]
- kvm-file-posix-Probe-paths-and-retry-SG_IO-on-potential-.patch [RHEL-95407]
- Resolves: RHEL-95407
(Support multipath failover with scsi-block [rhel-9.6.z])
* Mon May 26 2025 Jon Maloy <jmaloy@redhat.com> - 9.1.0-15.el9_6.5
- kvm-hw-i386-Fix-machine-type-compatibility.patch [RHEL-92077]
- Resolves: RHEL-92077
(Fix x86 M-type compats [rhel-9.6.z])
* Mon May 05 2025 Jon Maloy <jmaloy@redhat.com> - 9.1.0-15.el9_6.4
- kvm-file-posix-probe-discard-alignment-on-Linux-block-de.patch [RHEL-87734]
- kvm-block-io-skip-head-tail-requests-on-EINVAL.patch [RHEL-87734]