renesas/virt/virtio-spec.git
17 months agocontent: support SR-IOV master
Tiwei Bie [Thu, 24 May 2018 16:49:51 +0000 (00:49 +0800)]
content: support SR-IOV

Allocate a feature bit for virtio devices which support SR-IOV.
Note: at proposer's request, driver requirements have been
omitted - a reworked version is now being discussed.

Suggested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/11
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
18 months agoVIRTIO_F_IO_BARRIER: use I/O barriers in driver
Tiwei Bie [Thu, 10 May 2018 15:41:26 +0000 (23:41 +0800)]
VIRTIO_F_IO_BARRIER: use I/O barriers in driver

There will be hardware virtio devices in the future, which
require drivers to use the barriers suitable for I/O devices,
compared with software virtio devices which just require
drivers to use the barriers suitable for CPU cores.

To fix the ordering issue for hardware virtio devices, add
a new feature: VIRTIO_F_IO_BARRIER. When negotiated, driver
will use the barriers suitable for I/O devices.

Fixes: https://github.com/oasis-tcs/virtio-spec/issues/10
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
18 months agocontent: document hugetlb page allocation counts
Jonathan Helman [Wed, 18 Apr 2018 17:10:19 +0000 (10:10 -0700)]
content: document hugetlb page allocation counts

The Linux kernel exports the number of successful and failed
hugetlb page allocations via the virtio balloon driver. These
two counts need to be documented in this specification.

Fixes: https://github.com/oasis-tcs/virtio-spec/issues/8
Signed-off-by: Jonathan Helman <jonathan.helman@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
18 months agoREADME: supply link to the authoritative source
Michael S. Tsirkin [Wed, 23 May 2018 17:22:00 +0000 (20:22 +0300)]
README: supply link to the authoritative source

authoritative source is the tagged version.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
18 months agosplit-ring: Fix typo in used ring description
Christoffer Dall [Tue, 22 May 2018 19:11:05 +0000 (21:11 +0200)]
split-ring: Fix typo in used ring description

Only the device updates the used ring.

This looks like a typo as a result of copy-paste from the description of
the Available Ring.

Fixes: https://github.com/oasis-tcs/virtio-spec/issues/12
Signed-off-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
18 months agoRevert "VIRTIO_F_NOTIFICATION_DATA: extra data to devices"
Michael S. Tsirkin [Wed, 25 Apr 2018 21:43:20 +0000 (00:43 +0300)]
Revert "VIRTIO_F_NOTIFICATION_DATA: extra data to devices"

This reverts commit 396b1957817e0bd14697e28d9fe61bfc80f1350f.

Was pushed by mistake.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agotex: enable build on systems without ellipsis.sty
Michael S. Tsirkin [Mon, 16 Apr 2018 20:14:23 +0000 (23:14 +0300)]
tex: enable build on systems without ellipsis.sty

Unfortunately some systems (e.g. RHEL7.5 and derivatives) lack
the ellipsis package. As it's function is merely cosmetic,
support these systems by loading the package conditionally.

Tested-by: Mihai Carabas <mihai.carabas@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agotex: drop unused todonotes
Michael S. Tsirkin [Mon, 16 Apr 2018 20:12:35 +0000 (23:12 +0300)]
tex: drop unused todonotes

We don't really use todonotes, this dependency can be dropped.  We do
need to import xcolor which was previously pulled in by todonotes.

Tested-by: Mihai Carabas <mihai.carabas@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agoVIRTIO_F_NOTIFICATION_DATA: extra data to devices
Michael S. Tsirkin [Tue, 27 Mar 2018 02:56:04 +0000 (05:56 +0300)]
VIRTIO_F_NOTIFICATION_DATA: extra data to devices

Some devices benefit from ability to find out the number of available
descriptors in the ring: for efficiency or as a debugging aid.

To help with these optimizations, add a new feature:
VIRTIO_F_NOTIFICATION_DATA. When negotiated, driver notifications to the
device include this extra information.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
19 months agopacked-ring: rephrase "an example driver code"
Stefan Hajnoczi [Wed, 14 Mar 2018 15:46:18 +0000 (15:46 +0000)]
packed-ring: rephrase "an example driver code"

Either "an example of driver code" or "a driver code example" would
work.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/7
Approved-by: https://www.oasis-open.org/committees/ballot.php?id=3184
19 months agopacked-ring: fix "the descriptor flags +field+ is updated"
Stefan Hajnoczi [Wed, 14 Mar 2018 15:46:17 +0000 (15:46 +0000)]
packed-ring: fix "the descriptor flags +field+ is updated"

Either it needs to be "the descriptors flags are update" (since "flags"
is plural) or "the descriptor flags field is updated".  Since "flags" is
a name of a specific field, I chose the latter solution.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/7
Approved-by: https://www.oasis-open.org/committees/ballot.php?id=3184
19 months agopacked-ring: fix plurals
Stefan Hajnoczi [Wed, 14 Mar 2018 15:46:16 +0000 (15:46 +0000)]
packed-ring: fix plurals

There are two instances where plural needs to be used instead of
singular:

1. "descriptor(s)" is used later in the text, so use it consistently
   instead of starting with "descriptor" and switching later.

2. "What follows is the requirements" ->
   "What follows are the requirements"

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/7
Approved-by: https://www.oasis-open.org/committees/ballot.php?id=3184
19 months agopacked-ring: fix capitalization of names
Stefan Hajnoczi [Wed, 14 Mar 2018 15:46:15 +0000 (15:46 +0000)]
packed-ring: fix capitalization of names

Names are capitalized, like the Element Length field.  General terms
like descriptor are not capitalized.

This patch fixes two instances where this is inconsistent.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/7
Approved-by: https://www.oasis-open.org/committees/ballot.php?id=3184
19 months agopacked-ring: fix 'contigious' typo
Stefan Hajnoczi [Wed, 14 Mar 2018 15:46:14 +0000 (15:46 +0000)]
packed-ring: fix 'contigious' typo

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/7
Approved-by: https://www.oasis-open.org/committees/ballot.php?id=3184
19 months agopacked-ring: add missing comma
Stefan Hajnoczi [Wed, 14 Mar 2018 15:46:13 +0000 (15:46 +0000)]
packed-ring: add missing comma

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/7
Approved-by: https://www.oasis-open.org/committees/ballot.php?id=3184
19 months agopacked-ring: add missing articles to sentences
Stefan Hajnoczi [Wed, 14 Mar 2018 15:46:12 +0000 (15:46 +0000)]
packed-ring: add missing articles to sentences

There are several sentence structures used throughout the text where an
article is missing.  For example:

  ... allocated by driver.

or

  Driver then notifies the device.

Standards documents normally use complete sentences so this patch adds
the missing articles required before the noun ("driver"):

  ... allocated by the driver.

or

  The driver then notifies the device.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/7
Approved-by: https://www.oasis-open.org/committees/ballot.php?id=3184
20 months agovirtio-net: Add missing flag define to struct
sameehj [Thu, 21 Dec 2017 14:45:04 +0000 (16:45 +0200)]
virtio-net: Add missing flag define to struct

The flag VIRTIO_NET_HDR_F_DATA_VALID is aleady  mentioned in the virtio-net
spec however it is not defined in the struct. This flag is already defined
and used in virtio-net device in Qemu as well as some of it's drivers.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/6
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3182
20 months agocontent: document balloon memory statistics
Tomáš Golembiovský [Thu, 22 Feb 2018 12:50:52 +0000 (13:50 +0100)]
content: document balloon memory statistics

Linux kernel provides some balloon memory statistics that were not
included in the specs. Include them to avoid any ID clashes in the
future.

Signed-off-by: Tomáš Golembiovský <tgolembi@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/4
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3180
20 months agovirtio-blk: add discard and write zeroes features to specification
Changpeng Liu [Fri, 9 Mar 2018 02:44:23 +0000 (10:44 +0800)]
virtio-blk: add discard and write zeroes features to specification

Existing virtio-blk protocol doesn't have DISCARD/WRITE ZEROES support,
this will impact the performance when using SSD backend over file systems.

Here is the proposal to extend existing virtio-blk protocol to support
DISCARD/WRITE ZEROES commands.

Basic idea here is using 16 Bytes payload to support 1 descriptor, users
can put several segments together with 1 DISCARD/WRITE ZEROES command.

struct virtio_blk_discard_write_zeroes {
       le64 sector;
       le32 num_sectors;
       struct {
               le32 unmap:1;
               le32 reserved:31;
       } flags;
};

For the purpose to support such feature, we need to introduce 2 new feature
flags: VIRTIO_BLK_F_DISCARD/VIRTIO_BLK_F_WRITE_ZEROES, and 2 new command
types: VIRTIO_BLK_T_DISCARD/VIRTIO_BLK_T_WRITE_ZEROES. Also we introduce
several new parameters in the configuration space of virtio-blk:
max_discard_sectors/max_discard_seg/max_write_zeroes_sectors.
These parameters will tell the OS what's the granularity when
issuing such commands.

If both DISCARD and WRITE ZEROES are supported, unmap flag bit maybe used
for WRITE ZEROES command with DISCARD bit enabled.

Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3181
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/5

20 months agosplit-ring: in order feature
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:40 +0000 (23:23 +0200)]
split-ring: in order feature

For a split ring, require that drivers use descriptors in order too.
This allows devices to skip reading the available ring.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agopacked-ring: add in order support
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:38 +0000 (23:23 +0200)]
packed-ring: add in order support

Support in-order requests for packed rings.
This allows selective write-out of used descriptors.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agocontent: in-order buffer use
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:36 +0000 (23:23 +0200)]
content: in-order buffer use

Using descriptors in-order is sometimes beneficial.  Add an option for
that - per-format detail allowing more optimizations will be added by
follow-up patches.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agopacked virtqueues: more efficient virtqueue layout
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:35 +0000 (23:23 +0200)]
packed virtqueues: more efficient virtqueue layout

Performance analysis of this is in my kvm forum 2016 presentation.  The
idea is to have a r/w descriptor in a ring structure, replacing the used
and available ring, index and descriptor buffer.

This is also easier for devices to implement than the 1.0 layout.
Several more enhancements will be necessary to actually make this
efficient for devices to use.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agosplit-ring: typo: aligment
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:35 +0000 (23:23 +0200)]
split-ring: typo: aligment
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agosplit-ring: generalize text
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:34 +0000 (23:23 +0200)]
split-ring: generalize text

Update generic text to talk about available/used buffers, not rings.
Move some split-ring specific text to the correct section.

Update conformance section with link to the new conformance clause.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agocontent: generalize rest of text
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:33 +0000 (23:23 +0200)]
content: generalize rest of text

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agocontent: generalize transport ring part naming
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:32 +0000 (23:23 +0200)]
content: generalize transport ring part naming

Replace descriptor table/available ring/used ring
with descriptor area/driver area/device area
in all transports.

Document what's in which area.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agocontent: len -> used length, used ring -> vq
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:32 +0000 (23:23 +0200)]
content: len -> used length, used ring -> vq

Document buffer used len and use that terminology everywhere in the
generic section.

Further, drop the 'used ring' terminology and just say virtqueue.

Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agocontent: move virtqueue operation description
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:31 +0000 (23:23 +0200)]
content: move virtqueue operation description

virtqueue operation description is specific to the virtqueue
format. Move it out to split-ring.tex and update all
references.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agocontent: move ring text out to a separate file
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:30 +0000 (23:23 +0200)]
content: move ring text out to a separate file

Will be easier to manage this way.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agocontent: move 1.0 queue format out to a separate section
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:30 +0000 (23:23 +0200)]
content: move 1.0 queue format out to a separate section

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agointroduction: document bitfield notation
Michael S. Tsirkin [Fri, 9 Mar 2018 21:23:29 +0000 (23:23 +0200)]
introduction: document bitfield notation

Bitfields are a useful and familiar way to specify sub-byte structure
layout. The only issue is that bitfield order isn't portable across
architectures.  Document that we list bitfields from least to
most significant one, and warn about portability issues.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3177
Fixes: https://github.com/oasis-tcs/virtio-spec/issues/3

20 months agogithub: PULL_REQUEST_TEMPLATE.md
Michael S. Tsirkin [Wed, 14 Mar 2018 03:28:47 +0000 (05:28 +0200)]
github: PULL_REQUEST_TEMPLATE.md

Document that we don't accept pull requests at this point.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
20 months agovirtio-scsi: fix typos
Paolo Bonzini [Wed, 30 Nov 2016 21:22:13 +0000 (22:22 +0100)]
virtio-scsi: fix typos

Fixes: VIRTIO-155
Approved-by: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3118
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
20 months agoREADME.md: add hints for contributors
Michael S. Tsirkin [Fri, 2 Mar 2018 20:08:35 +0000 (22:08 +0200)]
README.md: add hints for contributors

Add links to admin repo and mailing lists.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
23 months agocontent: reserve virtio device ID for IOMMU
Jean-Philippe Brucker [Tue, 31 Oct 2017 15:13:21 +0000 (15:13 +0000)]
content: reserve virtio device ID for IOMMU

Reserve device ID 23 for virtio-iommu devices.

Resolved VIRTIO-162
Approved https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3155

Signed-off-by: Jean-Philippe Brucker <jean-philippe.brucker@arm.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
2 years agonet: relax buf size requirements for mrg buffers
Michael S. Tsirkin [Mon, 27 Mar 2017 23:09:08 +0000 (02:09 +0300)]
net: relax buf size requirements for mrg buffers

VIRTIO-84 added a requirement that each buffer is > header size but
there's no reason to require that for merge-able buffers since data can
be placed in follow-up buffers.

The current wording seems confusing enough that it seems likely
no one relies on this.

VIRTIO-159

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <1490654778-11857-1-git-send-email-mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
2 years agonet: clarify device rules for mergeable buffers
Michael S. Tsirkin [Mon, 27 Mar 2017 23:09:11 +0000 (02:09 +0300)]
net: clarify device rules for mergeable buffers

The idea behind mergeable buffers was to simply use them in a way
similar to a chain of descriptors.  Unfortunately the current text does
not say so - apparently nothing says device can't spread a packet over
as many buffers as it likes - but this didn't prevent drivers from
relying on buffers being used as a chain of descriptors, completely -
and blindly accessing it without checking the length at least for the
packet header.

Let's just make the spec match this reality - if devices ever want more
flexibility, we can add a feature bit.

Further, correct all misuses of a "descriptor" to "buffer" as that
is the entity that is being merged.

VIRTIO-160

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <1490655576-12173-1-git-send-email-mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
2 years agovirtio-net: fix virtqueue numbering error
Cornelia Huck [Tue, 11 Jul 2017 11:34:08 +0000 (13:34 +0200)]
virtio-net: fix virtqueue numbering error

receiveq(n) is at position 2(n-1), not 2n. Same error for transmitq(n)
and controlq.

VIRTIO-161

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Halil Pasic <pasic@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
2 years agotitle, acknowledgements: update my affiliation
Cornelia Huck [Fri, 8 Sep 2017 08:50:29 +0000 (10:50 +0200)]
title, acknowledgements: update my affiliation

Signed-off-by: Cornelia Huck <cohuck@redhat.com>
2 years agoMerge https://github.com/oasis-tcs/virtio-spec into trunk
Michael S. Tsirkin [Fri, 8 Sep 2017 03:44:22 +0000 (06:44 +0300)]
Merge https://github.com/oasis-tcs/virtio-spec into trunk

This merges in the OASIS boilerplate files in preparation to
moving virtio spec development to git, as approved at:

https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3120

Files added:

CONTRIBUTING.md
LICENSE.md
README.md

TODO: tweak boilerplate files as appropriate/needed.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2 years agotypo, case
Robin Cover [Thu, 7 Sep 2017 14:18:26 +0000 (09:18 -0500)]
typo, case

2 years agocreate boilerplate LICENSE text
robincover [Thu, 7 Sep 2017 14:04:03 +0000 (09:04 -0500)]
create boilerplate LICENSE text

2 years agoCreate CONTRIBUTING.md
robincover [Thu, 7 Sep 2017 14:03:19 +0000 (09:03 -0500)]
Create CONTRIBUTING.md

2 years agoupdate README with noilerplate
Robin Cover [Thu, 7 Sep 2017 14:02:02 +0000 (09:02 -0500)]
update README with noilerplate

2 years agoInitial commit
robincover [Thu, 7 Sep 2017 13:57:57 +0000 (08:57 -0500)]
Initial commit

2 years agocontext: Document VIRTIO_F_IOMMU_PLATFORM reserved feature bit (33)
mstsirkin [Tue, 10 Jan 2017 05:28:07 +0000 (05:28 +0000)]
context: Document VIRTIO_F_IOMMU_PLATFORM reserved feature bit (33)

On most virtio systems, physical addresses are provided from the driver
to the device and subsequently used to access memory directly. However,
some systems feature an IOMMU that complicates things slightly:

  - The IOMMU might require configuration for DMA to work at all. Once
    configured, the driver must pass bus addresses to the device, which
    are then translated by the IOMMU into physical addresses in memory.

  - The IOMMU might be bypassed by virtio DMA. In this case, physical
    addresses must still be provided to the device by the driver, even
    though an IOMMU appears to be present.

This has been solved by the addition of a new reserved feature bit,
VIRTIO_F_IOMMU_PLATFORM, to describe the two cases above. Legacy systems
will not negotiate the feature, and therefore continue to operate
directly on physical addresses.

VIRTIO-154

Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/trunk@587 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agogit-svn: instructions for faster init
mstsirkin [Tue, 11 Oct 2016 01:49:42 +0000 (01:49 +0000)]
git-svn: instructions for faster init

initial svn clone takes ages. A faster way is to copy history
from a git mirror. Document this.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/trunk@586 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agovirtio-net: add Max MTU configuration field
mstsirkin [Tue, 11 Oct 2016 00:58:19 +0000 (00:58 +0000)]
virtio-net: add Max MTU configuration field

It is helpful for a host to indicate it's MTU to be set on guest NICs
other than the assumed 1500 byte value.  This helps in situations where
the host network is using Jumbo Frames, or aiding in PMTU discovery by
configuring a homogenous network.  It is also helpful for sizing receive
buffers correctly.

The change adds a new field to configuration area of network
devices.  It will be used to pass a maximum MTU from the device to
the driver.  This will be used by the driver as a maximum value for
packet sizes during transmission, without segmentation offloading.

In addition, in order to support backward and forward compatibility,
we introduce a new feature bit called VIRTIO_NET_F_MTU.

VIRTIO-152

Signed-off-by: Aaron Conole <aconole@redhat.com>
Cc: Victor Kaplansky <victork@redhat.com>
Reviewed-by: Hannes Reiencke <hare@suse.de>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/trunk@585 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agocontent: Reserve virtio device ID for pstore
mstsirkin [Mon, 10 Oct 2016 19:37:47 +0000 (19:37 +0000)]
content: Reserve virtio device ID for pstore

This patch just reserve next available device ID for pstore device
type.  The device specification for pstore will come later.

VIRTIO-150

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Message-Id: <1468941217-12904-1-git-send-email-namhyung@gmail.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/trunk@583 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agocontent: reserve virtio device ID for sdm
mstsirkin [Mon, 10 Oct 2016 19:37:42 +0000 (19:37 +0000)]
content: reserve virtio device ID for sdm

VIRTIO-149

Signed-off-by: Christian Pinto <c.pinto@virtualopensystems.com>
Signed-off-by: Baptiste Reynal <b.reynal@virtualopensystems.com>
Message-Id: <1467111824-11548-2-git-send-email-c.pinto@virtualopensystems.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/trunk@582 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agovirtio: reserve device id for virtio crypto
mstsirkin [Mon, 10 Oct 2016 19:37:37 +0000 (19:37 +0000)]
virtio: reserve device id for virtio crypto

Reserve a device ID for virtio crypto device to
avoid collisions with other new virtio devices.

Because 19 had been reserved for virtio socket
device, I'd like to reserve 20 for crypto device.

VIRTIO-148

Suggested-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Message-Id: <1460685060-5904-1-git-send-email-arei.gonglei@huawei.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/trunk@581 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agovsock: reserve device ID
mstsirkin [Mon, 10 Oct 2016 19:37:31 +0000 (19:37 +0000)]
vsock: reserve device ID

Add a device ID for the Socket device.  The draft specification is
currently under review but it helps to reserve the ID to avoid
collisions with other new virtio devices.

VIRTIO-147

Suggested-by: Ian Campbell <ian.campbell@docker.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <1458729661-7514-1-git-send-email-stefanha@redhat.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/trunk@580 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agoccw: add CCW_CMD_READ_STATUS
mstsirkin [Mon, 10 Oct 2016 19:37:25 +0000 (19:37 +0000)]
ccw: add CCW_CMD_READ_STATUS

ccw currently allows the driver to update the status via the
CCW_CMD_WRITE_STATUS command; however, it does not allow the driver to
retrieve the current status at the device, which is needed to properly
support DEVICE_NEEDS_RESET.

Therefore, provide a new command CCW_CMD_READ_STATUS allowing the driver
to retrieve the device status. Provide this command when revision 2 has
been negotiated.

VIRTIO-117

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Message-Id: <1445243513-2263-2-git-send-email-cornelia.huck@de.ibm.com>

git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/trunk@579 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agoMerge remote-tracking branch 'origin/v1.0'
mstsirkin [Fri, 22 Jul 2016 21:40:05 +0000 (21:40 +0000)]
Merge remote-tracking branch 'origin/v1.0'

git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/trunk@572 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agoREVISION: update for cs04
mstsirkin [Fri, 22 Jul 2016 21:36:09 +0000 (21:36 +0000)]
REVISION: update for cs04

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@570 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agotitle: drop Rusty's email from chair/editor list
mstsirkin [Tue, 16 Feb 2016 19:33:33 +0000 (19:33 +0000)]
title: drop Rusty's email from chair/editor list

I kept Rusty's name in the acknowledgements section,
but listing him as Chair/Editor does not help since
the email listed bounces.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@569 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agoacknowledgements: acknowledge review by Kevin Lo
mstsirkin [Tue, 16 Feb 2016 19:33:08 +0000 (19:33 +0000)]
acknowledgements: acknowledge review by Kevin Lo

Cc: Kevin Lo <kevlo@kevlo.org>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@568 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agoacknowledgements: update affiliation for Thomas Huth
mstsirkin [Tue, 16 Feb 2016 19:32:32 +0000 (19:32 +0000)]
acknowledgements: update affiliation for Thomas Huth

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@567 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agomakediff: suppress diff for chater names
mstsirkin [Tue, 16 Feb 2016 18:09:33 +0000 (18:09 +0000)]
makediff: suppress diff for chater names

Temporary hack since we have nameref pointing
to a chapter name, and that conflicts with latexdiff.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@566 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agocl: list changes since CSPRD05
mstsirkin [Tue, 16 Feb 2016 17:31:33 +0000 (17:31 +0000)]
cl: list changes since CSPRD05

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@565 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agomakediff: cherry pick file name change
mstsirkin [Tue, 16 Feb 2016 17:12:41 +0000 (17:12 +0000)]
makediff: cherry pick file name change

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@563 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agofixupdiff: actually apply the DIFadd/DIFdel hack
mstsirkin [Tue, 16 Feb 2016 17:12:18 +0000 (17:12 +0000)]
fixupdiff: actually apply the DIFadd/DIFdel hack

looks like recent changes made it inactive: must
change $line, not $_.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@562 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agovirtio header: fix diff generation
mstsirkin [Tue, 16 Feb 2016 17:11:57 +0000 (17:11 +0000)]
virtio header: fix diff generation

add old cross reference as well as an chapter parameter
to avoid latexdiff errors when chapter is renamed.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@561 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agofixupdiff: update for latest perl
mstsirkin [Tue, 16 Feb 2016 17:11:35 +0000 (17:11 +0000)]
fixupdiff: update for latest perl

Escape {} in regexes to avoid warnings in latest perl.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@560 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agoinit: sort status bits
mstsirkin [Tue, 16 Feb 2016 17:11:14 +0000 (17:11 +0000)]
init: sort status bits

Status bit order is inconsistent: they are neither in increasing
order nor in the order they are likely to be used.

The second approach seems more useful since there aren't
that many bits, so the numerical order does not help much.

A typical order of use would be:

ACKNOWLEDGE
DRIVER
then either FAILED or FEATURES_OK
then either FAILED or DRIVER_OK
then DEVICE_NEEDS_RESET (if device detects an error)

Sort the bits accordingly.

Reported-by: Kevin Lo <kevlo@kevlo.org>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@559 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agovirtio-ring.h -> virtio-queue.h
mstsirkin [Tue, 16 Feb 2016 17:10:51 +0000 (17:10 +0000)]
virtio-ring.h -> virtio-queue.h

Since vring* and VRING* have been replaced with virtq* and VIRTQ*
respectively, rename the header virtio_ring.h to virtio_queue.h.

Suggested-by: Kevin Lo <kevlo@kevlo.org>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@558 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agoRevert "headerfile rename virtio_ring to virtio queue"
mstsirkin [Tue, 16 Feb 2016 17:10:13 +0000 (17:10 +0000)]
Revert "headerfile rename virtio_ring to virtio queue"

This immediately reverts the previous commit, solely for the
benefit of latexdiff.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@557 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agoheaderfile: rename virtio_ring to virtio queue
mstsirkin [Tue, 16 Feb 2016 17:09:52 +0000 (17:09 +0000)]
headerfile: rename virtio_ring to virtio queue

Done as a separate commit for benefit of latexdiff.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@556 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agovirtio header: tweak change motivation
mstsirkin [Tue, 16 Feb 2016 17:09:29 +0000 (17:09 +0000)]
virtio header: tweak change motivation

The changes are not just to remove Linux assumptions,
we have also renamed ring->queue.
Tweak the header description accordingly.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@555 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agovirtio-net: fix inconsistent legacy header size
mstsirkin [Tue, 16 Feb 2016 17:09:07 +0000 (17:09 +0000)]
virtio-net: fix inconsistent legacy header size

Current text says:
The legacy driver only presented num_buffers in the struct
virtio_net_hdr when VIRTIO_NET_F_MRG_RXBUF was not negotiated;

Should be:
"... was negotiated ..." instead of "... was not negotiated ..."

To be consistent with the following:
without that feature the structure was 2 bytes shorter.

Reported-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@554 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agoREVISION: update for csprd05
mstsirkin [Tue, 16 Feb 2016 17:08:45 +0000 (17:08 +0000)]
REVISION: update for csprd05

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@553 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

3 years agotitle: fix previous link to cs03
mstsirkin [Tue, 16 Feb 2016 17:08:18 +0000 (17:08 +0000)]
title: fix previous link to cs03

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@552 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agocl: document resolution for VIRTIO-146.
mstsirkin [Wed, 21 Oct 2015 14:22:09 +0000 (14:22 +0000)]
cl: document resolution for VIRTIO-146.

add advice on transition from earlier drafts.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@551 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agocl: document resolution for VIRTIO-145.
mstsirkin [Wed, 21 Oct 2015 14:21:41 +0000 (14:21 +0000)]
cl: document resolution for VIRTIO-145.

pci: clarify configuration access capability rules

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@550 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agocl: document resolution for VIRTIO-144
mstsirkin [Wed, 21 Oct 2015 14:21:16 +0000 (14:21 +0000)]
cl: document resolution for VIRTIO-144

virtio-blk: restore VIRTIO_BLK_F_FLUSH and VIRTIO_BLK_F_CONFIG_WCE

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@549 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agointro: add advice on transition from legacy interfaces
mstsirkin [Wed, 21 Oct 2015 14:20:51 +0000 (14:20 +0000)]
intro: add advice on transition from legacy interfaces

Reading legacy chapters gives a hint about what changed,
let's help readers discover this useful shortcut.

This resolves VIRTIO-146.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Accepted by electronic ballot:
https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=2866

git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@548 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agopci: clarify configuration access capability rules
mstsirkin [Wed, 21 Oct 2015 14:20:27 +0000 (14:20 +0000)]
pci: clarify configuration access capability rules

The point of the configuration access capability is to enable
access to other capabilities.  The intent never was to allow
writes to a random place within device BARs.
Limiting drivers simplifies devices - and devices can always
add another capability if drivers ever want to access
some other range.

This resolves VIRTIO-145.

Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Accepted by electronic ballot:
https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=2865

git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@547 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agovirtio-blk: restore VIRTIO_BLK_F_FLUSH and VIRTIO_BLK_F_CONFIG_WCE
mstsirkin [Wed, 21 Oct 2015 14:20:04 +0000 (14:20 +0000)]
virtio-blk: restore VIRTIO_BLK_F_FLUSH and VIRTIO_BLK_F_CONFIG_WCE

VIRTIO_BLK_F_CONFIG_WCE is important in order to achieve good performance
(up to 2x, though more realistically +30-40%) in latency-bound workloads.
However, it was removed by mistake together with VIRTIO_BLK_F_FLUSH.

In addition, even removing VIRTIO_BLK_F_FLUSH was probably not a great
idea, because it simplifies simple drivers (e.g. firmware) that are okay
with a writethrough cache but still need data to persist after power loss.
What really should have been removed is just the possibility that devices
not propose VIRTIO_BLK_F_FLUSH, but even that only deserves a "SHOULD" in
the new world of conformance statements.

Restore these, with the following changes:

* clarify and use conformance statements in order to define writeback
and writethrough caching according to what is commonly done by high-end
storage.

* clarify (with conformance statements) the influence of the
VIRTIO_BLK_F_FLUSH feature on caching and how to proceed if only one of
VIRTIO_BLK_F_FLUSH and VIRTIO_BLK_F_CONFIG_WCE is negotiated.

* strengthen the requirement for persisting writes to MUST after
a VIRTIO_BLK_T_FLUSH request (and in other cases too involving the
new features).

The suggested behavior upon feature negotiation is okay for the Linux
implementation of virtio1, even after the implementation is modified to
support the two new features.

This fixes VIRTIO-144.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Accepted by electronic ballot:
https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=2864

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@546 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agoRevert: makediff: cleanup using begingroup/endgroup
mstsirkin [Wed, 21 Oct 2015 14:19:35 +0000 (14:19 +0000)]
Revert: makediff: cleanup using begingroup/endgroup

This reverts commit
commit ef519a86f046d1be22f82f32d845653a850c21dd
Author: mstsirkin <mstsirkin@0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652>
Date:   Thu Jun 26 16:50:13 2014 +0000

makediff: cleanup using begingroup/endgroup

This seemed like a good idea originally: replace perl hacks with
tex hacks. However adding \begingroup and \endgroup within macros
used by latex-diff breaks its assumption that these are low level
tex and can be used in arbitrary situations, e.g.  cross the
boundaries of environments, use mis-matching begin and end
instructions (DIFaddbegin with DIFaddendFL), etc.

Let's go back to the original work-around:

commit b665b3165b454b98c782617e37a128b53b56c89c
Author: mstsirkin <mstsirkin@0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652>
Date:   Thu Jun 26 12:34:38 2014 +0000

    work around xetex bug

    Too many \color directives produce corrupted output
    and this warning:
    WARNING ** Color stack overflow. Just ignore.

    Use script to reduce # of these directives.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@545 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agocl: list changes 540-542
mstsirkin [Sun, 11 Oct 2015 15:13:49 +0000 (15:13 +0000)]
cl: list changes 540-542

all of them non-material

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@544 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agocl: clear it out for cs03
mstsirkin [Sun, 11 Oct 2015 15:13:22 +0000 (15:13 +0000)]
cl: clear it out for cs03

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@543 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agovirtio-blk: flush: Fix driver vs. device typo
mstsirkin [Sun, 11 Oct 2015 14:56:13 +0000 (14:56 +0000)]
virtio-blk: flush: Fix driver vs. device typo

This is a simple non-material change.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@542 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agovirtio-blk: fix typo in legacy framing requirements section
mstsirkin [Sun, 11 Oct 2015 14:55:44 +0000 (14:55 +0000)]
virtio-blk: fix typo in legacy framing requirements section

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@541 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agovirtqueues: fix trivial typo
mstsirkin [Sun, 11 Oct 2015 14:55:13 +0000 (14:55 +0000)]
virtqueues: fix trivial typo

Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@540 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agoMerge remote-tracking branch 'origin/v1.0'
mstsirkin [Sun, 2 Aug 2015 11:41:54 +0000 (11:41 +0000)]
Merge remote-tracking branch 'origin/v1.0'

Update trunk to virtio-v1.0-cs03

git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/trunk@539 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agoadmin: set date and revision for cs03
mstsirkin [Sun, 2 Aug 2015 11:31:12 +0000 (11:31 +0000)]
admin: set date and revision for cs03

Approved Aug 2:
https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=2843

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@537 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agotitle: update previous version link
mstsirkin [Thu, 21 May 2015 11:39:27 +0000 (11:39 +0000)]
title: update previous version link

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@536 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agofixupdiff: better diff for lstlisting
mstsirkin [Thu, 21 May 2015 10:13:40 +0000 (10:13 +0000)]
fixupdiff: better diff for lstlisting

Actually strike out deleted text.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@535 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agofixupdiff.pl: more work-arounds for latexdiff
mstsirkin [Thu, 21 May 2015 10:13:22 +0000 (10:13 +0000)]
fixupdiff.pl: more work-arounds for latexdiff

work around more latexdiff issues:
- force FL style within headers
- fix beginFL ending with end and vice versa
  across begin/end of enum or itemize

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@534 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agomakediff: mark enumerate/itemize as floats
mstsirkin [Thu, 21 May 2015 10:13:04 +0000 (10:13 +0000)]
makediff: mark enumerate/itemize as floats

Helps detect begin/end crossing these environments
and fix up automatically.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@533 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agocontent: replace \\note with note environment
mstsirkin [Thu, 21 May 2015 10:12:45 +0000 (10:12 +0000)]
content: replace \\note with note environment

\\note just before start of subsection seems to
confuse latex, resulting in incorrect text alignment.
Replace with \begin{note} \end{note}.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@532 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agocl-os: VIRTIO-123, VIRTIO-124, VIRTIO-126, VIRTIO-143
mstsirkin [Thu, 21 May 2015 10:12:25 +0000 (10:12 +0000)]
cl-os: VIRTIO-123, VIRTIO-124, VIRTIO-126, VIRTIO-143

Document resolutions to VIRTIO-123, VIRTIO-124, VIRTIO-126 and
VIRTIO-143.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@531 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agoconformance: link in new conformance statements
mstsirkin [Thu, 21 May 2015 10:12:02 +0000 (10:12 +0000)]
conformance: link in new conformance statements

Link new conformance statements added by commit
    network device: document VIRTIO_NET_F_CTRL_RX_EXTRA

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@530 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agonetwork device: document VIRTIO_NET_F_CTRL_RX_EXTRA
mstsirkin [Mon, 18 May 2015 20:50:15 +0000 (20:50 +0000)]
network device: document VIRTIO_NET_F_CTRL_RX_EXTRA

VIRTIO-124

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@529 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agonetwork device: xmit/receive cleanup
mstsirkin [Mon, 18 May 2015 20:49:51 +0000 (20:49 +0000)]
network device: xmit/receive cleanup

Fix up multiple issues in xmit/receive sections:
- drop MAY/MUST/SHOULD outside normative statements
- spell out conformance requirements for both drivers and
  devices, for xmit and receive paths
- document the missing VIRTIO_NET_HDR_F_DATA_VALID
- document handling of unrecognized flag bits so we can extend
  flags in the future, similar to VIRTIO_NET_HDR_F_DATA_VALID

VIRTIO-123

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@528 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agoballoon: document deflate on oom
mstsirkin [Mon, 18 May 2015 20:49:25 +0000 (20:49 +0000)]
balloon: document deflate on oom

Document the new option, and also clarify behaviour
without it.

In particular, I noticed that actual field is not the
actual number of pages in the balloon as
driver might do inflate followed by deflate.

Also, device isn't always driven by interrupts,
driver can inflate/deflate in response to e.g.
memory compaction.

VIRTIO-126

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@527 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agoballoon: transitional device support
mstsirkin [Mon, 18 May 2015 20:48:58 +0000 (20:48 +0000)]
balloon: transitional device support

Virtio 1.0 cs02 doesn't include a modern balloon device.  At some
point we'll likely define an incompatible interface with a
different ID and different semantics.  But for now, it's not a
big effort to support a transitional balloon device: this has the
advantage of supporting existing drivers, transparently, as well
as transports that don't allow mixing virtio 0 and virtio 1
devices. And balloon is an easy device to test, so it's also
useful for people to test virtio core handling of transitional
devices.

Three issues with legacy hypervisors have been identified:
1. Actual value is actually used, and is necessary for management
   to work. Luckily 4 byte config space writes are now atomic.
   When using old guests, hypervisors can detect access to the last byte.
   When using old hypervisors, drivers can use atomic 4-byte accesses.
2. Hypervisors actually didn't ignore the stats from the first
   buffer supplied. This means the values there would be
   incorrect until hypervisor resends the request.
   Add a note suggesting hypervisors ignore the 1st buffer.
3. QEMU simply over-writes stats from each buffer it gets.
   Thus if driver supplies a different subset of stats
   on each request, stale values will be there.
   Require drivers to supply the same subset on each
   request. This also gives us a simple way to figure out
   which stats are supported.

VIRTIO-143

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@526 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agocl: document VIRTIO-134, VIRTIO-140, VIRTIO-141, VIRTIO-142
mstsirkin [Wed, 29 Apr 2015 11:10:14 +0000 (11:10 +0000)]
cl: document VIRTIO-134, VIRTIO-140, VIRTIO-141, VIRTIO-142

document changes made by commits r520 to r523 inclusive

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@525 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652

4 years agoconformance: add missing links
mstsirkin [Wed, 29 Apr 2015 11:09:56 +0000 (11:09 +0000)]
conformance: add missing links

previous commit "used ring: specify legacy behaviour for len field"
to resolve VIRTIO-141 added new sections with normative
statements for legacy balloon and scsi devices. Link them from
the conformance section.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio/branches/v1.0@524 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652