From: "Daniel P. Berrangé" <berrange@redhat.com>
To: qemu-devel@nongnu.org
Cc: "Manos Pitsidianakis" <manos.pitsidianakis@linaro.org>,
"Stefan Weil" <sw@weilnetz.de>,
"Dr. David Alan Gilbert" <dave@treblig.org>,
"Pierrick Bouvier" <pierrick.bouvier@linaro.org>,
devel@lists.libvirt.org,
"Philippe Mathieu-Daudé" <philmd@linaro.org>,
"Marc-André Lureau" <marcandre.lureau@redhat.com>,
"Hanna Reitz" <hreitz@redhat.com>,
"Kevin Wolf" <kwolf@redhat.com>,
qemu-block@nongnu.org, qemu-rust@nongnu.org,
"Paolo Bonzini" <pbonzini@redhat.com>,
"Markus Armbruster" <armbru@redhat.com>,
"Gerd Hoffmann" <kraxel@redhat.com>,
"Daniel P. Berrangé" <berrange@redhat.com>,
"Christian Schoenebeck" <qemu_oss@crudebyte.com>,
"Richard Henderson" <richard.henderson@linaro.org>
Subject: [PATCH v6 11/27] util: introduce some API docs for logging APIs
Date: Wed, 11 Feb 2026 15:24:52 +0000 [thread overview]
Message-ID: <20260211152508.732487-12-berrange@redhat.com> (raw)
In-Reply-To: <20260211152508.732487-1-berrange@redhat.com>
There is a gotcha with qemu_log() usage in a threaded process.
If fragments of a log message are output via qemu_log() it is
possible for messages from two threads to get mixed up. To
prevent this qemu_log_trylock() should be used, along with
fprintf(f) calls.
This is a subtle problem that needs to be explained in the
API docs to ensure correct usage.
In the Rust code, the log_mask_ln method which is conceptually
equivalent to the C qemu_log() call will unconditionally append
a newline so must only ever be used for complete log messages.
Reported-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
include/qemu/log-for-trace.h | 17 ++++++++++++++++-
include/qemu/log.h | 32 ++++++++++++++++++++++++++++++++
rust/util/src/log.rs | 6 ++++++
3 files changed, 54 insertions(+), 1 deletion(-)
diff --git a/include/qemu/log-for-trace.h b/include/qemu/log-for-trace.h
index f3a8791f1d..6861a1a4b7 100644
--- a/include/qemu/log-for-trace.h
+++ b/include/qemu/log-for-trace.h
@@ -29,7 +29,22 @@ static inline bool qemu_loglevel_mask(int mask)
return (qemu_loglevel & mask) != 0;
}
-/* main logging function */
+/**
+ * qemu_log: report a log message
+ * @fmt: the format string for the message
+ * @...: the format string arguments
+ *
+ * This will emit a log message to the current output stream.
+ *
+ * The @fmt string should normally represent a complete line
+ * of text, and thus end with a newline character.
+ *
+ * While it is possible to incrementally output fragments of
+ * a complete line using qemu_log, this is inefficient and
+ * races with other threads. For outputting fragments it is
+ * strongly preferred to use the qemu_log_trylock() method
+ * combined with fprintf().
+ */
void G_GNUC_PRINTF(1, 2) qemu_log(const char *fmt, ...);
#endif
diff --git a/include/qemu/log.h b/include/qemu/log.h
index 7effba4da4..b49a1a5fe3 100644
--- a/include/qemu/log.h
+++ b/include/qemu/log.h
@@ -41,7 +41,39 @@ bool qemu_log_separate(void);
/* Lock/unlock output. */
+/**
+ * Acquires a lock on the current log output stream.
+ * The returned FILE object should be used with the
+ * fprintf() function to output the log message, and
+ * then qemu_log_unlock() called to release the lock.
+ *
+ * The primary use case is to be able to incrementally
+ * output fragments of a complete log message in an
+ * efficient and race free manner.
+ *
+ * The simpler qemu_log() method should normally only
+ * be used to output complete log messages, and not
+ * within scope of a qemu_log_trylock() call.
+ *
+ * A typical usage pattern would be
+ *
+ * FILE *f = qemu_log_trylock()
+ *
+ * fprintf(f, "Something ");
+ * fprintf(f, "Something ");
+ * fprintf(f, "Something ");
+ * fprintf(f, "The end\n");
+ *
+ * qemu_log_unlock(f);
+ *
+ * Returns: the current FILE if available, NULL on error
+ */
FILE *qemu_log_trylock(void) G_GNUC_WARN_UNUSED_RESULT;
+
+/**
+ * Releases the lock on the log output, previously
+ * acquired by qemu_log_trylock().
+ */
void qemu_log_unlock(FILE *fd);
/* Logging functions: */
diff --git a/rust/util/src/log.rs b/rust/util/src/log.rs
index 0a4bc4249a..a85db804a7 100644
--- a/rust/util/src/log.rs
+++ b/rust/util/src/log.rs
@@ -134,6 +134,12 @@ fn drop(&mut self) {
/// "Address 0x{:x} out of range",
/// error_address,
/// );
+///
+/// The `log_mask_ln` macro must only be used for emitting complete
+/// log messages. Where it is required to incrementally output string
+/// fragments to construct a complete message, `LogGuard::new()` must
+/// be directly used in combination with `writeln()` to avoid output
+/// races with other QEMU threads.
/// ```
#[macro_export]
macro_rules! log_mask_ln {
--
2.53.0
next prev parent reply other threads:[~2026-02-11 15:26 UTC|newest]
Thread overview: 59+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-02-11 15:24 [PATCH v6 00/27] util: sync error_report & qemu_log output more closely Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 01/27] meson: don't access 'cxx' object without checking cpp lang Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 02/27] qemu-options: remove extraneous [] around arg values Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 03/27] include: define constant for early constructor priority Daniel P. Berrangé
2026-02-18 9:22 ` Markus Armbruster
2026-02-18 10:46 ` Daniel P. Berrangé
2026-02-18 13:23 ` Markus Armbruster
2026-02-11 15:24 ` [PATCH v6 04/27] monitor: initialize global data from a constructor Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 05/27] system: unconditionally enable thread naming Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 06/27] util: fix race setting thread name on Win32 Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 07/27] util: expose qemu_thread_set_name Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 08/27] audio: make jackaudio use qemu_thread_set_name Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 09/27] util: set the name for the 'main' thread on Windows Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 10/27] util: add API to fetch the current thread name Daniel P. Berrangé
2026-02-11 15:24 ` Daniel P. Berrangé [this message]
2026-02-18 9:38 ` [PATCH v6 11/27] util: introduce some API docs for logging APIs Markus Armbruster
2026-02-11 15:24 ` [PATCH v6 12/27] util: avoid repeated prefix on incremental qemu_log calls Daniel P. Berrangé
2026-02-18 9:52 ` Markus Armbruster
2026-02-18 10:45 ` Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 13/27] util/log: add missing error reporting in qemu_log_trylock_with_err Daniel P. Berrangé
2026-02-18 10:45 ` Markus Armbruster
2026-02-11 15:24 ` [PATCH v6 14/27] ui: add proper error reporting for password changes Daniel P. Berrangé
2026-02-18 12:10 ` Markus Armbruster
2026-02-25 16:08 ` Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 15/27] ui: remove redundant use of error_printf_unless_qmp() Daniel P. Berrangé
2026-02-18 12:12 ` Markus Armbruster
2026-02-11 15:24 ` [PATCH v6 16/27] monitor: remove redundant error_[v]printf_unless_qmp Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 17/27] monitor: refactor error_vprintf() Daniel P. Berrangé
2026-02-11 15:24 ` [PATCH v6 18/27] monitor: move error_vprintf back to error-report.c Daniel P. Berrangé
2026-02-11 15:25 ` [PATCH v6 19/27] util: fix interleaving of error & trace output Daniel P. Berrangé
2026-02-18 12:41 ` Markus Armbruster
2026-02-18 12:45 ` Markus Armbruster
2026-02-11 15:25 ` [PATCH v6 20/27] util: don't skip error prefixes when QMP is active Daniel P. Berrangé
2026-02-18 12:47 ` Markus Armbruster
2026-02-11 15:25 ` [PATCH v6 21/27] util: fix interleaving of error prefixes Daniel P. Berrangé
2026-02-11 15:25 ` [PATCH v6 22/27] util: introduce common helper for error-report & log code Daniel P. Berrangé
2026-02-18 14:04 ` Markus Armbruster
2026-02-25 16:18 ` Daniel P. Berrangé
2026-02-25 17:51 ` Markus Armbruster
2026-02-11 15:25 ` [PATCH v6 23/27] util: convert error-report & log to message API for timestamp Daniel P. Berrangé
2026-02-11 15:25 ` [PATCH v6 24/27] util: add support for formatting a workload name in messages Daniel P. Berrangé
2026-02-11 15:25 ` [PATCH v6 25/27] util: add support for formatting a program " Daniel P. Berrangé
2026-02-19 10:08 ` Markus Armbruster
2026-02-25 16:24 ` Daniel P. Berrangé
2026-02-26 7:11 ` Markus Armbruster
2026-02-19 10:23 ` Peter Maydell
2026-02-25 16:38 ` Daniel P. Berrangé
2026-02-25 17:43 ` Peter Maydell
2026-02-25 17:47 ` Daniel P. Berrangé
2026-02-11 15:25 ` [PATCH v6 26/27] util: add support for formatting thread info " Daniel P. Berrangé
2026-02-19 10:14 ` Markus Armbruster
2026-02-25 16:33 ` Daniel P. Berrangé
2026-02-19 10:29 ` Peter Maydell
2026-02-25 16:30 ` Daniel P. Berrangé
2026-02-25 17:39 ` Peter Maydell
2026-02-11 15:25 ` [PATCH v6 27/27] util: add brackets around guest name in message context Daniel P. Berrangé
2026-02-19 10:16 ` Markus Armbruster
2026-02-26 9:51 ` [PATCH v6 00/27] util: sync error_report & qemu_log output more closely Markus Armbruster
2026-02-26 9:58 ` Daniel P. Berrangé
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=20260211152508.732487-12-berrange@redhat.com \
--to=berrange@redhat.com \
--cc=armbru@redhat.com \
--cc=dave@treblig.org \
--cc=devel@lists.libvirt.org \
--cc=hreitz@redhat.com \
--cc=kraxel@redhat.com \
--cc=kwolf@redhat.com \
--cc=manos.pitsidianakis@linaro.org \
--cc=marcandre.lureau@redhat.com \
--cc=pbonzini@redhat.com \
--cc=philmd@linaro.org \
--cc=pierrick.bouvier@linaro.org \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=qemu-rust@nongnu.org \
--cc=qemu_oss@crudebyte.com \
--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