From: "Alex Bennée" <alex.bennee@linaro.org>
To: qemu-devel@nongnu.org
Cc: "Paolo Bonzini" <pbonzini@redhat.com>,
"Richard Henderson" <rth@twiddle.net>,
"Doug Gale" <doug16k@gmail.com>,
"Alex Bennée" <alex.bennee@linaro.org>,
"Peter Maydell" <peter.maydell@linaro.org>
Subject: [PATCH] cpus: kick all vCPUs when running thread=single
Date: Tue, 1 Oct 2019 17:04:26 +0100 [thread overview]
Message-ID: <20191001160426.26644-1-alex.bennee@linaro.org> (raw)
qemu_cpu_kick is used for a number of reasons including to indicate
there is work to be done. However when thread=single the old
qemu_cpu_kick_rr_cpu only advanced the vCPU to the next executing one
which can lead to a hang in the case that:
a) the kick is from outside the vCPUs (e.g. iothread)
b) the timers are paused (i.e. iothread calling run_on_cpu)
To avoid this lets split qemu_cpu_kick_rr into two functions. One for
the timer which continues to advance to the next timeslice and another
for all other kicks.
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Cc: Doug Gale <doug16k@gmail.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Peter Maydell <peter.maydell@linaro.org>
---
cpus.c | 24 ++++++++++++++++++------
1 file changed, 18 insertions(+), 6 deletions(-)
diff --git a/cpus.c b/cpus.c
index d2c61ff155..bee7209134 100644
--- a/cpus.c
+++ b/cpus.c
@@ -949,8 +949,8 @@ static inline int64_t qemu_tcg_next_kick(void)
return qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + TCG_KICK_PERIOD;
}
-/* Kick the currently round-robin scheduled vCPU */
-static void qemu_cpu_kick_rr_cpu(void)
+/* Kick the currently round-robin scheduled vCPU to next */
+static void qemu_cpu_kick_rr_next_cpu(void)
{
CPUState *cpu;
do {
@@ -961,6 +961,16 @@ static void qemu_cpu_kick_rr_cpu(void)
} while (cpu != atomic_mb_read(&tcg_current_rr_cpu));
}
+/* Kick all RR vCPUs */
+static void qemu_cpu_kick_rr_cpus(void)
+{
+ CPUState *cpu;
+
+ CPU_FOREACH(cpu) {
+ cpu_exit(cpu);
+ };
+}
+
static void do_nothing(CPUState *cpu, run_on_cpu_data unused)
{
}
@@ -993,7 +1003,7 @@ void qemu_timer_notify_cb(void *opaque, QEMUClockType type)
static void kick_tcg_thread(void *opaque)
{
timer_mod(tcg_kick_vcpu_timer, qemu_tcg_next_kick());
- qemu_cpu_kick_rr_cpu();
+ qemu_cpu_kick_rr_next_cpu();
}
static void start_tcg_kick_timer(void)
@@ -1828,9 +1838,11 @@ void qemu_cpu_kick(CPUState *cpu)
{
qemu_cond_broadcast(cpu->halt_cond);
if (tcg_enabled()) {
- cpu_exit(cpu);
- /* NOP unless doing single-thread RR */
- qemu_cpu_kick_rr_cpu();
+ if (qemu_tcg_mttcg_enabled()) {
+ cpu_exit(cpu);
+ } else {
+ qemu_cpu_kick_rr_cpus();
+ }
} else {
if (hax_enabled()) {
/*
--
2.20.1
next reply other threads:[~2019-10-01 16:48 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-01 16:04 Alex Bennée [this message]
2019-10-01 16:38 ` [PATCH] cpus: kick all vCPUs when running thread=single Paolo Bonzini
2019-10-01 17:40 ` Alex Bennée
2019-10-06 16:05 ` Paolo Bonzini
2019-10-07 14:00 ` Richard Henderson
2019-10-07 14:56 ` Paolo Bonzini
2019-10-07 15:30 ` Alex Bennée
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=20191001160426.26644-1-alex.bennee@linaro.org \
--to=alex.bennee@linaro.org \
--cc=doug16k@gmail.com \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=qemu-devel@nongnu.org \
--cc=rth@twiddle.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).