diff mbox

[v2] virtio-spec: Add virtio-ccw spec.

Message ID 1346771641-12136-2-git-send-email-cornelia.huck@de.ibm.com (mailing list archive)
State New, archived
Headers show

Commit Message

Cornelia Huck Sept. 4, 2012, 3:14 p.m. UTC 
Add specifications for the new s390 specific virtio-ccw transport.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
---

Changes v1->v2:
- support more than 32 feature bits
- allow to allocate less than requested queue size
- allow to transfer alignment

---
 virtio-spec.lyx |  534 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 534 insertions(+), 0 deletions(-)

Comments

Rusty Russell Sept. 6, 2012, 2:26 a.m. UTC | #1 
Cornelia Huck <cornelia.huck@de.ibm.com> writes:

> Add specifications for the new s390 specific virtio-ccw transport.
>
> Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>

Excellent.  I have applied and pushed this revision.

I was tempted to ask for an explicit endian marker, as switching to
explicit little endian was high on the TODO list for virtio2.  On the
other hand, we could say virtio2-pci config space is always
little-endian, and virtio-ccw config space is always big-endian.

That covers the config endian, and and if we introduce a new ring
format, that endian debate can be separate.

Cheers,
Rusty.
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Benjamin Herrenschmidt Sept. 6, 2012, 3:09 a.m. UTC | #2 
On Thu, 2012-09-06 at 11:56 +0930, Rusty Russell wrote:
> Excellent.  I have applied and pushed this revision.
> 
> I was tempted to ask for an explicit endian marker, as switching to
> explicit little endian was high on the TODO list for virtio2.  On the
> other hand, we could say virtio2-pci config space is always
> little-endian, and virtio-ccw config space is always big-endian.
> 
> That covers the config endian, and and if we introduce a new ring
> format, that endian debate can be separate. 

I'm not too happy with having different endians for the transport. That
somewhat relies on making sure "client" code always use the proper
size-based accessors to any element in the config space, whether it's
the common one or the device specific part.

I've seen that assumption broken multiple times in the past (iirc, maybe
I'm just mis-remembering here :-)

I would think that it would be much better to define the endianness of
the virtio config space once and for all regardless of the underlying
transport.

It's not like a few byteswaps on config accesses are going to hurt s390
performance...

Cheers,
Ben.


--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Cornelia Huck Sept. 6, 2012, 1:51 p.m. UTC | #3 
On Thu, 06 Sep 2012 11:56:31 +0930
Rusty Russell <rusty@rustcorp.com.au> wrote:

> Cornelia Huck <cornelia.huck@de.ibm.com> writes:
> 
> > Add specifications for the new s390 specific virtio-ccw transport.
> >
> > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> 
> Excellent.  I have applied and pushed this revision.

Cool.

Unfortunately, we just discovered that we probably need to expand the
protocol to allow for a configuration-changed marker in addition to the
indicator bits. Can we easily amend the spec later on?

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Rusty Russell Sept. 6, 2012, 11:57 p.m. UTC | #4 
Cornelia Huck <cornelia.huck@de.ibm.com> writes:

> On Thu, 06 Sep 2012 11:56:31 +0930
> Rusty Russell <rusty@rustcorp.com.au> wrote:
>
>> Cornelia Huck <cornelia.huck@de.ibm.com> writes:
>> 
>> > Add specifications for the new s390 specific virtio-ccw transport.
>> >
>> > Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
>> 
>> Excellent.  I have applied and pushed this revision.
>
> Cool.
>
> Unfortunately, we just discovered that we probably need to expand the
> protocol to allow for a configuration-changed marker in addition to the
> indicator bits. Can we easily amend the spec later on?

Hmm, I'll delay the point-release of the spec until I have that patch.

You can amend the spec because you're the only implementation, but it's
poor form :)

Cheers,
Rusty.
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/virtio-spec.lyx b/virtio-spec.lyx
index 7a073f4..8247d2e 100644
--- a/virtio-spec.lyx
+++ b/virtio-spec.lyx
@@ -57,6 +57,7 @@ 
 \html_css_as_file 0
 \html_be_strict false
 \author -608949062 "Rusty Russell,,," 
+\author -385801441 "Cornelia Huck" cornelia.huck@de.ibm.com
 \author 1531152142 "Paolo Bonzini,,," 
 \end_header
 
@@ -9350,8 +9351,541 @@  tatus register description is asserted.
  After the interrupt is handled, the driver must acknowledge it by writing
  a bit mask corresponding to the serviced interrupt to the InterruptACK
  register.
+\change_inserted -385801441 1343732742
+
 \end_layout
 
 \end_deeper
+\begin_layout Chapter*
+
+\change_inserted -385801441 1343732726
+Appendix Y: virtio-ccw
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+S/390 based virtual machines support neither PCI nor MMIO, so a different
+ transport is needed there.
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+The old s390-virtio mechanism used a special page mapped above the guest's
+ memory and several diagnose calls (hypercalls); it does have some drawbacks,
+ however, like a rather limited number of devices and very restricted hotplug
+ support.
+ Moreover, device discovery and operation differ from other environments
+ on the S/390 platform.
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+virtio-ccw uses the standard channel I/O based mechanism used for the majority
+ of devices on S/390.
+ A virtual channel device with a special control unit type acts as proxy
+ to the virtio device (similar to the way virtio-pci uses a PCI device)
+ and configuration and operation of the virtio device is accomplished (mostly)
+ via channel commands.
+ This means virtio devices are discoverable via standard operating system
+ algorithms, and adding virtio support is mainly a question of supporting
+ a new control unit type.
+\end_layout
+
+\begin_layout Subsection*
+
+\change_inserted -385801441 1343732726
+Basic Concepts
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732817
+As a proxy device, virtio-ccw uses a channel-attached I/O control unit with
+ a special control unit type (0x3832) and a control unit model corresponding
+ to the attached virtio device's subsystem device ID, accessed via a virtual
+ I/O subchannel and a virtual channel path of type 0x32.
+ This proxy device is discoverable via normal channel subsystem device discovery
+ (usually a STORE SUBCHANNEL loop) and answers to the basic channel commands,
+ most importantly SENSE ID.
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+In addition to the basic channel commands, virtio-ccw defines a set of channel
+ commands related to configuration and operation of virtio:
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+\begin_inset listings
+inline false
+status open
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+#define CCW_CMD_SET_VQ 0x13
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+#define CCW_CMD_VDEV_RESET 0x33
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+#define CCW_CMD_SET_IND 0x43
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+#define CCW_CMD_READ_FEAT 0x12
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+#define CCW_CMD_WRITE_FEAT 0x11
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+#define CCW_CMD_READ_CONF 0x22 
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+#define CCW_CMD_WRITE_CONF 0x21
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+#define CCW_CMD_WRITE_STATUS 0x31
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+#define CCW_CMD_READ_VQ_CONF 0x32
+\end_layout
+
+\end_inset
+
+
+\end_layout
+
+\begin_layout Subsection*
+
+\change_inserted -385801441 1343732726
+Device Initialization
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+virtio-ccw uses several channel commands to set up a device.
+\end_layout
+
+\begin_layout Subsubsection*
+
+\change_inserted -385801441 1343732726
+Configuring a Virtqueue
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+CCW_CMD_READ_VQ_CONF is issued by the guest to obtain information about
+ a queue.
+ It uses the following structure for communicating:
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+\begin_inset listings
+inline false
+status open
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+struct vq_config_block { 
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027276
+
+	__u16 index; 
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027280
+
+	__u16 max_num; 
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+} __attribute__ ((packed));
+\end_layout
+
+\end_inset
+
+
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1345020625
+The requested number of buffers for queue 
+\family typewriter
+index
+\family default
+ is returned in 
+\family typewriter
+max_num
+\family default
+.
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+Afterwards, CCW_CMD_SET_VQ is issued by the guest to inform the host about
+ the location used for its queue.
+ The transmitted structure is
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+\begin_inset listings
+inline false
+status open
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+struct vq_info_block { 
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027266
+
+	__u64 queue; 
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027268
+
+	__u32 align;
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027270
+
+	__u16 index;
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027271
+
+	__u16 num; 
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1343732726
+
+} __attribute__ ((packed)); 
+\end_layout
+
+\end_inset
+
+
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1345020957
+
+\family typewriter
+queue
+\family default
+ contains the guest address for queue 
+\family typewriter
+index
+\family default
+.
+ The actual number of allocated buffers is transmitted in 
+\family typewriter
+num
+\family default
+ and their alignment in 
+\family typewriter
+align
+\family default
+.
+\end_layout
+
+\begin_layout Subsubsection*
+
+\change_inserted -385801441 1343732726
+Communicating Status Information
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+The guest can change the status of a device via the CCW_CMD_WRITE_STATUS
+ command, which transmits an 8 bit status value.
+\end_layout
+
+\begin_layout Subsubsection*
+
+\change_inserted -385801441 1343732726
+Handling Device Features
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1345027079
+Feature bits are arranged in an array of 32 bit values, making for a total
+ of 8192 feature bits.
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1345027108
+The CCW commands dealing with features use the following communication block:
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027350
+\begin_inset listings
+inline false
+status open
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027350
+
+struct virtio_feature_desc { 
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027350
+
+	__u32 features; 
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027350
+
+	__u8 index;
+\end_layout
+
+\begin_layout LyX-Code
+
+\change_inserted -385801441 1345027350
+
+} __attribute__ ((packed)); 
+\change_unchanged
+
+\end_layout
+
+\end_inset
+
+
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1345027491
+
+\family typewriter
+features
+\family default
+ are the 32 bits of features currently accessed, while 
+\family typewriter
+index
+\family default
+ describes which of the feature bit values is to be accessed.
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1345027568
+The guest may obtain the host's device feature set via the CCW_CMD_READ_FEAT
+ command.
+ The host stores the features at 
+\family typewriter
+index
+\family default
+ to 
+\family typewriter
+features
+\family default
+.
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1345027622
+For communicating its device features to the host, the guest may use the
+ CCW_CMD_WRITE_FEAT command, denoting a 
+\family typewriter
+features
+\family default
+/
+\family typewriter
+index
+\family default
+ combination.
+\end_layout
+
+\begin_layout Subsubsection*
+
+\change_inserted -385801441 1343732726
+Device Configuration
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+The device's configuration space is located in host memory.
+ It is the same size as the standard PCI configuration space.
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+To obtain information from the configuration space, the guest may use CCW_CMD_RE
+AD_CONF, specifying the guest memory for the host to write to.
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+For changing configuration information, the guest may use CCW_CMD_WRITE_CONF,
+ specifying the guest memory for the host to read from.
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+In both cases, the complete configuration space is transmitted.
+\end_layout
+
+\begin_layout Subsubsection*
+
+\change_inserted -385801441 1343732726
+Setting Up Indicators
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+To communicate the location of the indicator bits for host->guest notification,
+ the guest uses the CCW_CMD_SET_IND command which sends the guest address
+ of the indicators in a 64 bit value.
+\end_layout
+
+\begin_layout Subsection*
+
+\change_inserted -385801441 1343732726
+Device Operation
+\end_layout
+
+\begin_layout Subsubsection*
+
+\change_inserted -385801441 1343732726
+Host->Guest Notification
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+For notifying the guest of virtqueue buffers, the host sets the corresponding
+ bit in the guest-provided indicators.
+ If an interrupt is not already pending for the subchannel, the host generates
+ an unsolicited I/O interrupt.
+\end_layout
+
+\begin_layout Subsubsection*
+
+\change_inserted -385801441 1343732726
+Guest->Host Notification
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+For notifying the host of virtqueue buffers, the guest unfortunately can't
+ use a channel command (the asynchronous characteristics of channel I/O
+ interact badly with the host block I/O backend).
+ Instead, it uses a diagnose 0x500 call with subcode 3 specifying the queue.
+\end_layout
+
+\begin_layout Subsubsection*
+
+\change_inserted -385801441 1343732726
+Early printk for Virtio Consoles
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+For the early printk mechanism, diagnose 0x500 with subcode 0 is used.
+\end_layout
+
+\begin_layout Subsubsection*
+
+\change_inserted -385801441 1343732726
+Resetting Devices
+\end_layout
+
+\begin_layout Standard
+
+\change_inserted -385801441 1343732726
+In order to reset a device, a guest may send the CCW_CMD_VDEV_RESET command.
+\change_unchanged
+
+\end_layout
+
 \end_body
 \end_document