qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Ioannis Angelakopoulos <iangelak@redhat.com>
To: qemu-devel@nongnu.org, virtio-fs@redhat.com
Cc: iangelak@redhat.com, stefanha@redhat.com, dgilbert@redhat.com,
	vgoyal@redhat.com
Subject: [PATCH 0/6] virtiofsd: Support for remote blocking posix locks
Date: Wed, 16 Jun 2021 15:39:15 -0400	[thread overview]
Message-ID: <20210616193921.608720-1-iangelak@redhat.com> (raw)

Hi,

Here is the implementation for the remote blocking posix locks in
virtiofsd.

This patch series also addresses the issues that were not fixed in:
[Virtio-fs] [PATCH v2 0/5] [RFC] virtiofsd, vhost-user-fs: Add support
for notification queue
https://patchwork.kernel.org/project/qemu-devel/cover/20191204190836.31324-1-vgoyal@redhat.com/

The main issue was that the guest could not reboot when a thread of the
virtiofsd process was blocked, waiting to acquire a lock. The
introduction of a custom threadpool dedicated to servicing only locking
requests addresses the aforementioned issue.

Description:

Through the addition of a notification queue it is possible for a
guest now to block waiting for a remote lock held by another guest.

When the lock is available virtiofsd sends a notification to the guest
that unblocks the process/thread waiting for the lock in order to
acquire it.

This implementation also handles cases where the guest that is
waiting for a lock, hard-reboots through SYSRQ
(echo "b" > /proc/sysrq-trigger). Specifically, a custom threadpool
(posix theads) in virtiofsd is used to service the locking requests.
If one of the threads is blocked waiting for a lock and the guest
hard-reboots then virtiofsd sends a signal to the blocked thread waking 
it up so that it exits and then virtiofsd cleans up its state and
re-initializes.

Finally, the custom threadpool enables virtiofsd to concurrently service
other requests that are non-blocking or block for a limited time while
the custom threadpool services locking requests that might block. This is
especially useful in the case where the GThreadPool is not utilized.

Thanks,
Ioannis

Ioannis Angelakopoulos (2):
  virtiofsd: Thread state cleanup when blocking posix locks are used
  virtiofsd: Custom threadpool for remote blocking posix locks requests

Vivek Goyal (4):
  virtiofsd: Release file locks using F_UNLCK
  virtiofsd: Create a notification queue
  virtiofsd: Specify size of notification buffer using config space
  virtiofsd: Implement blocking posix locks

 hw/virtio/vhost-user-fs.c                  |  57 +-
 include/hw/virtio/vhost-user-fs.h          |   4 +-
 include/standard-headers/linux/fuse.h      |   8 +
 include/standard-headers/linux/virtio_fs.h |   5 +
 tools/virtiofsd/fuse_i.h                   |   2 +
 tools/virtiofsd/fuse_lowlevel.c            |  40 +-
 tools/virtiofsd/fuse_lowlevel.h            |  26 +
 tools/virtiofsd/fuse_virtio.c              | 626 +++++++++++++++++++--
 tools/virtiofsd/passthrough_ll.c           |  90 ++-
 9 files changed, 804 insertions(+), 54 deletions(-)

-- 
2.27.0



             reply	other threads:[~2021-06-16 20:52 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-16 19:39 Ioannis Angelakopoulos [this message]
2021-06-16 19:39 ` [PATCH 1/6] virtiofsd: Release file locks using F_UNLCK Ioannis Angelakopoulos
2021-06-16 19:39 ` [PATCH 2/6] virtiofsd: Create a notification queue Ioannis Angelakopoulos
2021-06-16 19:39 ` [PATCH 3/6] virtiofsd: Specify size of notification buffer using config space Ioannis Angelakopoulos
2021-06-16 19:39 ` [PATCH 4/6] virtiofsd: Implement blocking posix locks Ioannis Angelakopoulos
2021-06-16 19:39 ` [PATCH 5/6] virtiofsd: Thread state cleanup when blocking posix locks are used Ioannis Angelakopoulos
2021-06-16 19:39 ` [PATCH 6/6] virtiofsd: Custom threadpool for remote blocking posix locks requests Ioannis Angelakopoulos

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=20210616193921.608720-1-iangelak@redhat.com \
    --to=iangelak@redhat.com \
    --cc=dgilbert@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@redhat.com \
    --cc=vgoyal@redhat.com \
    --cc=virtio-fs@redhat.com \
    /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).