public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Gregory Haskins <ghaskins@novell.com>
To: kvm@vger.kernel.org
Cc: viro@ZenIV.linux.org.uk, linux-kernel@vger.kernel.org,
	avi@redhat.com, davidel@xmailserver.org
Subject: [KVM PATCH v7 0/3] kvm: eventfd interfaces (formerly irqfd)
Date: Tue, 12 May 2009 14:26:45 -0400	[thread overview]
Message-ID: <20090512181134.26131.10023.stgit@dev.haskins.net> (raw)

(Applies to kvm.git:b5e725fa)

This is v7 of the series.  We have generalized the name of the series (as well
as some of the hunks in the series) to reflect the fact that we have multiple
eventfd based components.

This series has been tested and appears to be working as intended.  You can
download the unit-test used to verify this here:

ftp://ftp.novell.com/dev/ghaskins/kvm-eventfd.tar.bz

This will also serve as an example on how to use the new interfaces.

[ Changelog:

   v7:
        *) Added "iofd" to allow PIO/MMIO writes to generate an eventfd
           signal.  This was previously discussed as "hypercallfd", but
           since explicit hypercalls are not looking to be very popular,
           and based on the fact that they were not going to carry payload
           anyway, I named them "iofd".
        *) Generalized some of the code so that irqfd and iofd could be
           logically grouped together.  For instance
           s/KVM_CAP_IRQFD/KVM_CAP_EVENTFD and
           virt/kvm/irqfd.c becomes virt/kvm/eventfd.c
        *) Added support for "deassign" operations to ensure we can properly
           support hot-unplug.
	*) Reinstated the eventfd EXPORT_SYMBOL patch since we need it again
           for supporting iofd.
        *) Rebased to kvm.git:b5e725fa 

   v6:
        *) Moved eventfd creation back to userspace, per Avi's request
        *) Dropped no longer necessary supporting patches from series
        *) Rebased to kvm.git:833367b57

   v5:
        *) Added padding to the ioctl structure
        *) Added proper ref-count increment to the file before returning
           success. (Needs review by Al Viro, Davide Libenzi)
        *) Cleaned up error-handling path to make sure we remove ourself
           from the waitq if necessary.
        *) Make sure we only add ourselves to kvm->irqfds if successful
           creating the irqfd in the first place.
        *) Rebased to kvm.git:66b0aed4 

   v4:
        *) Changed allocation model to create the new fd last, after
           we get past the last potential error point by using Davide's
           new eventfd_file_create interface (Al Viro, Davide Libenzi)
	*) We no longer export sys_eventfd2() since it is replaced
           functionally with eventfd_file_create();
        *) Rebased to kvm.git:7da2e3ba

   v3:
        *) The kernel now allocates the eventfd (need to export sys_eventfd2)
        *) Added a flags field for future expansion to kvm_irqfd()
        *) We properly toggle the irq level 1+0.
        *) We re-use the USERSPACE_SRC_ID instead of creating our own
        *) Properly check for failures establishing a poll-table with eventfd
	*) Fixed fd/file leaks on failure
	*) Rebased to lateste kvm.git::41b76d8d04

   v2:
	*) Dropped notifier_chain based callbacks in favor of
	   wait_queue_t::func and file::poll based callbacks (Thanks to
	   Davide for the suggestion)

   v1:
        *) Initial release

]

[ Todo:
        *) Implement the bus_io_unregister() function so the iofd hot-unplug
           path may be completed
        *) Test the hot-unplug path
]


---

Gregory Haskins (3):
      kvm: add iofd support
      kvm: add support for irqfd via eventfd-notification interface
      eventfd: export eventfd interfaces for module use


 arch/x86/kvm/Makefile    |    2 
 arch/x86/kvm/x86.c       |    1 
 fs/eventfd.c             |    3 
 include/linux/kvm.h      |   22 +++
 include/linux/kvm_host.h |    7 +
 virt/kvm/eventfd.c       |  294 ++++++++++++++++++++++++++++++++++++++++++++++
 virt/kvm/kvm_main.c      |   33 +++++
 7 files changed, 361 insertions(+), 1 deletions(-)
 create mode 100644 virt/kvm/eventfd.c

-- 
Signature

             reply	other threads:[~2009-05-12 18:27 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-05-12 18:26 Gregory Haskins [this message]
2009-05-12 18:26 ` [KVM PATCH v7 1/3] eventfd: export eventfd interfaces for module use Gregory Haskins
2009-05-12 19:02   ` Davide Libenzi
2009-05-12 18:26 ` [KVM PATCH v7 2/3] kvm: add support for irqfd via eventfd-notification interface Gregory Haskins
2009-05-14  9:47   ` Avi Kivity
2009-05-14 11:52     ` Gregory Haskins
2009-05-14 12:20       ` Avi Kivity
2009-05-14 13:12         ` Gregory Haskins
2009-05-14 11:22   ` Avi Kivity
2009-05-14 15:52     ` Gregory Haskins
2009-05-15  3:22       ` Davide Libenzi
2009-05-15  3:35         ` Gregory Haskins
2009-05-12 18:27 ` [KVM PATCH v7 3/3] kvm: add iofd support Gregory Haskins
2009-05-12 19:05   ` Gregory Haskins
2009-05-12 19:29   ` [KVM PATCH v7.1] " Gregory Haskins
2009-05-12 22:17   ` [KVM PATCH v7.2] " Gregory Haskins
2009-05-13  2:46     ` Gregory Haskins
2009-05-14 11:11   ` [KVM PATCH v7 3/3] " Avi Kivity
2009-05-14 12:02     ` Gregory Haskins
2009-05-14 12:22       ` Avi Kivity

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=20090512181134.26131.10023.stgit@dev.haskins.net \
    --to=ghaskins@novell.com \
    --cc=avi@redhat.com \
    --cc=davidel@xmailserver.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=viro@ZenIV.linux.org.uk \
    /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