From: Paolo Bonzini <pbonzini@redhat.com>
To: qemu-devel@nongnu.org
Cc: qemu-stable@nongnu.org, qemu-block@nongnu.org
Subject: [Qemu-devel] [PATCH] blockjob: prevent double enter (with dangling access on the second)
Date: Wed, 23 Dec 2015 11:48:26 +0100 [thread overview]
Message-ID: <1450867706-19860-4-git-send-email-pbonzini@redhat.com> (raw)
Because block_job_sleep_ns marks the job as non-busy, it is
possible to enter it again from block_job_enter. However, the
same coroutine may then be re-entered from co_sleep_cb, and this
time the CoSleepCB is not on the stack anymore.
Fix this by open coding the sleep in block_job_sleep_ns, and
deleting the timer immediately after the coroutine is re-entered.
Cc: qemu-stable@nongnu.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
blockjob.c | 27 +++++++++++++++++++++++----
include/block/blockjob.h | 3 +--
2 files changed, 24 insertions(+), 6 deletions(-)
diff --git a/blockjob.c b/blockjob.c
index 80adb9d..5e59a8f 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -329,8 +329,21 @@ int block_job_complete_sync(BlockJob *job, Error **errp)
return block_job_finish_sync(job, &block_job_complete, errp);
}
+typedef struct CoSleepCB {
+ QEMUTimer *ts;
+ Coroutine *co;
+} CoSleepCB;
+
+static void co_sleep_cb(void *opaque)
+{
+ CoSleepCB *sleep_cb = opaque;
+
+ qemu_coroutine_enter(sleep_cb->co, NULL);
+}
+
void block_job_sleep_ns(BlockJob *job, QEMUClockType type, int64_t ns)
{
+ CoSleepCB sleep_cb = { .ts = NULL };
assert(job->busy);
/* Check cancellation *before* setting busy = false, too! */
@@ -339,10 +352,16 @@ void block_job_sleep_ns(BlockJob *job, QEMUClockType type, int64_t ns)
}
job->busy = false;
- if (block_job_is_paused(job)) {
- qemu_coroutine_yield();
- } else {
- co_aio_sleep_ns(bdrv_get_aio_context(job->bs), type, ns);
+ if (!block_job_is_paused(job)) {
+ sleep_cb.co = qemu_coroutine_self(),
+ sleep_cb.ts = aio_timer_new(bdrv_get_aio_context(job->bs), type,
+ SCALE_NS, co_sleep_cb, &sleep_cb);
+ timer_mod(sleep_cb.ts, qemu_clock_get_ns(type) + ns);
+ }
+ qemu_coroutine_yield();
+ if (sleep_cb.ts) {
+ timer_del(sleep_cb.ts);
+ timer_free(sleep_cb.ts);
}
job->busy = true;
}
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index d84ccd8..82063f2 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -120,8 +120,7 @@ struct BlockJob {
/**
* Set to false by the job while it is in a quiescent state, where
- * no I/O is pending and the job has yielded on any condition
- * that is not detected by #aio_poll, such as a timer.
+ * no I/O is pending. The job can be reentered with qemu_coroutine_enter.
*/
bool busy;
--
2.5.0
next reply other threads:[~2015-12-23 10:48 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-12-23 10:48 Paolo Bonzini [this message]
2015-12-23 10:51 ` [Qemu-devel] [PATCH] blockjob: prevent double enter (with dangling access on the second) Paolo Bonzini
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=1450867706-19860-4-git-send-email-pbonzini@redhat.com \
--to=pbonzini@redhat.com \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=qemu-stable@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).