From: "Daniel P. Berrangé" <berrange@redhat.com>
To: John Snow <jsnow@redhat.com>
Cc: qemu-devel@nongnu.org, "Alex Bennée" <alex.bennee@linaro.org>,
kvm@vger.kernel.org,
"Dmitry Osipenko" <dmitry.osipenko@collabora.com>,
"Fabiano Rosas" <farosas@suse.de>,
"Alistair Francis" <alistair.francis@wdc.com>,
"Pierrick Bouvier" <pierrick.bouvier@oss.qualcomm.com>,
"Stefan Weil" <sw@weilnetz.de>, "Peter Xu" <peterx@redhat.com>,
"Richard Henderson" <richard.henderson@linaro.org>,
"Michael S. Tsirkin" <mst@redhat.com>,
"Philippe Mathieu-Daudé" <philmd@oss.qualcomm.com>,
"David Gibson" <david@gibson.dropbear.id.au>,
"Paolo Bonzini" <pbonzini@redhat.com>,
"Akihiko Odaki" <odaki@rsg.ci.i.u-tokyo.ac.jp>,
"Marcelo Tosatti" <mtosatti@redhat.com>,
"Marc-André Lureau" <marcandre.lureau@redhat.com>
Subject: Re: [PATCH v2 1/7] debug/mem: add utility to name anonymous memory mappings
Date: Tue, 17 Nov 2026 13:39:33 +0000 [thread overview]
Message-ID: <avxZFbMNjEmnY2yK@redhat.com> (raw)
In-Reply-To: <20261116165002.30939-2-jsnow@redhat.com>
On Mon, Nov 16, 2026 at 11:49:56AM -0500, John Snow wrote:
> Add support to name mmap'd memory regions using
> PR_SET_VMA_ANON_NAME. On unsupported platforms, these QEMU utility
> functions are no-ops.
>
> This is useful primarily for debugging and profiling QEMU; being able
> to see and correlate each VMA segment with a specific use within QEMU.
>
> Support for PR_SET_VMA_ANON_NAME was added in Linux 5.17, and requires
> CONFIG_ANON_VMA_NAME=y to be set in the kernel configuration. Fedora
> (and RHEL) do not enable this feature by default. See also
> https://man7.org/linux/man-pages/man2/pr_set_vma.2const.html
>
> Signed-off-by: John Snow <jsnow@redhat.com>
> ---
> include/qemu/mmap-vma.h | 18 +++++++++++
> util/mmap-vma.c | 71 +++++++++++++++++++++++++++++++++++++++++
> util/meson.build | 1 +
> 3 files changed, 90 insertions(+)
> create mode 100644 include/qemu/mmap-vma.h
> create mode 100644 util/mmap-vma.c
>
> diff --git a/include/qemu/mmap-vma.h b/include/qemu/mmap-vma.h
> new file mode 100644
> index 00000000000..763ea5a5cab
> --- /dev/null
> +++ b/include/qemu/mmap-vma.h
> @@ -0,0 +1,18 @@
> +/*
> + * Utility for naming VMA allocations in linux
> + *
> + * Copyright (c) Red Hat, Inc 2026
> + *
> + * SPDX-License-Identifier: LGPL-2.1-or-later
> + */
> +
> +#ifndef __QEMU_MMAP_VMA_H
> +#define __QEMU_MMAP_VMA_H
> +
> +#include <stddef.h>
> +#include "osdep.h"
> +
> +int qemu_name_vma(void *ptr, size_t sz, const char *fmt, ...)
> + G_GNUC_PRINTF(3, 4);
> +
> +#endif
> diff --git a/util/mmap-vma.c b/util/mmap-vma.c
> new file mode 100644
> index 00000000000..7a50965d496
> --- /dev/null
> +++ b/util/mmap-vma.c
> @@ -0,0 +1,71 @@
> +/*
> + * Utility for naming VMA allocations in linux
> + *
> + * Copyright (c) Red Hat, Inc 2026
> + *
> + * SPDX-License-Identifier: LGPL-2.1-or-later
> + */
> +#include "qemu/mmap-vma.h"
> +#include "qemu/osdep.h"
> +
> +#include <errno.h>
> +#include <stddef.h>
> +#include <stdio.h>
> +
> +#if defined(__linux__)
> +#include <sys/prctl.h>
> +#endif
> +
> +#define VMA_PREFIX " qemu: "
> +#define VMA_MAXLEN 80
> +
> +/*
> + * Give a human-readable name to a previously mmap'd memory segment.
> + * This feature is only available on Linux >= 5.17, and only when the
> + * kernel was configured with CONFIG_ANON_VMA_NAME=y.
> + *
> + * For systems without this option, prctl() will return -1 and set
> + * errno to EINVAL, and can be safely ignored.
> + *
> + * For systems with this option, these human-readable names will be
> + * visible in /proc/$pid/[s]maps, pmap output, etc.
> + */
> +G_GNUC_PRINTF(3, 0)
> +static int qemu_vname_vma(void *ptr, size_t sz, const char *fmt, va_list args)
As the comment says above, we can't assume the kernel will have
the feature enabled.
It is harmless if we try to set the VMA name and it fails
though. We merely loose a bit of debug info. Thus I'd
suggest we make the function return type 'void', so that
it is clear that we do NOT expect callers to check for
errors.
> +{
> + int rc = 0;
> +
> +#if defined(PR_SET_VMA_ANON_NAME)
> + char strbuf[VMA_MAXLEN];
> +
> + assert(sizeof(VMA_PREFIX) < sizeof(strbuf));
> + strncpy(strbuf, VMA_PREFIX, sizeof(strbuf));
> + rc = vsnprintf(strbuf + sizeof(VMA_PREFIX),
> + sizeof(strbuf) - sizeof(VMA_PREFIX), fmt, args);
> + if (rc < 0) {
> + return -errno;
> + } else if (rc >= (sizeof(strbuf) - sizeof(VMA_PREFIX))) {
> + sprintf(strbuf + sizeof(strbuf) - 4, "...");
> + }
> +
> + rc = prctl(PR_SET_VMA, PR_SET_VMA_ANON_NAME, ptr, sz, strbuf);
> + if (rc) {
> + return -errno;
> + }
> +#endif
> +
> + return rc;
> +}
> +
> +G_GNUC_PRINTF(3, 4)
> +int qemu_name_vma(void *ptr, size_t sz, const char *fmt, ...)
> +{
> + int rc;
> + va_list args;
> +
> + va_start(args, fmt);
> + rc = qemu_vname_vma(ptr, sz, fmt, args);
> + va_end(args);
> +
> + return rc;
> +}
> diff --git a/util/meson.build b/util/meson.build
> index fa174c07a57..861ee228d5b 100644
> --- a/util/meson.build
> +++ b/util/meson.build
> @@ -1,5 +1,6 @@
> util_ss.add(files('osdep.c', 'cutils.c', 'unicode.c', 'qemu-timer-common.c'))
> util_ss.add(files('thread-context.c'), numa)
> +util_ss.add(files('mmap-vma.c'))
> if host_os != 'windows'
> util_ss.add(files('aio-posix.c'))
> util_ss.add(files('fdmon-poll.c'))
> --
> 2.56.1
>
With regards,
Daniel
--
|: https://berrange.com ~~ https://hachyderm.io/@berrange :|
|: https://libvirt.org ~~ https://entangle-photo.org :|
|: https://pixelfed.art/berrange ~~ https://fstop138.berrange.com :|
next prev parent reply other threads:[~2026-11-17 13:39 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-16 16:49 [PATCH v2 0/7] debug: add PR_SET_VMA_ANON_NAME support John Snow
2026-11-16 16:49 ` [PATCH v2 1/7] debug/mem: add utility to name anonymous memory mappings John Snow
2026-11-17 13:39 ` Daniel P. Berrangé [this message]
2026-11-16 16:49 ` [PATCH v2 2/7] debug/mem: add tracing to qemu_vname_vma() John Snow
2026-11-16 16:49 ` [PATCH v2 3/7] debug/mem: name <reserved> regions in mmap-alloc John Snow
2026-11-16 16:49 ` [PATCH v2 4/7] debug/mem: name pthreads John Snow
2026-11-17 13:46 ` Daniel P. Berrangé
2026-11-17 13:47 ` Daniel P. Berrangé
2026-11-16 16:50 ` [PATCH v2 5/7] debug/mem: name coroutine stacks John Snow
2026-11-17 13:47 ` Daniel P. Berrangé
2026-11-16 16:50 ` [PATCH v2 6/7] debug/mem: name qemu_anon_ram_alloc/qemu_ram_mmap pages John Snow
2026-11-16 16:50 ` [PATCH v2 7/7] debug/mem: HACK add qemu_name_buffer John Snow
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=avxZFbMNjEmnY2yK@redhat.com \
--to=berrange@redhat.com \
--cc=alex.bennee@linaro.org \
--cc=alistair.francis@wdc.com \
--cc=david@gibson.dropbear.id.au \
--cc=dmitry.osipenko@collabora.com \
--cc=farosas@suse.de \
--cc=jsnow@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=marcandre.lureau@redhat.com \
--cc=mst@redhat.com \
--cc=mtosatti@redhat.com \
--cc=odaki@rsg.ci.i.u-tokyo.ac.jp \
--cc=pbonzini@redhat.com \
--cc=peterx@redhat.com \
--cc=philmd@oss.qualcomm.com \
--cc=pierrick.bouvier@oss.qualcomm.com \
--cc=qemu-devel@nongnu.org \
--cc=richard.henderson@linaro.org \
--cc=sw@weilnetz.de \
/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