OASIS Virtual I/O Device (VIRTIO) TC

OK, this builds on Michael's endianness patches and my previous effort, does some sweeps through the document for consistency and adds Legacy Interface notes. This assumes we go for "LE everywhere" rather than "LE for everything except CCW". Cornelia is still away, so I'm making that call for the moment. Cheers, Rusty. Michael S. Tsirkin (2): configuration space endian-ness virtqueue endian-ness Rusty Russell (4): Use le types in all non-legacy places. Thorough endianness notes for legacy. Remove __ annotations on pci. CCW: Make it very clear that CCW channel commands are big-endian. virtio-v1.0-wd01-part1-specification.txt 398 +++++++++++++++++++------------ 1 file changed, 248 insertions(+), 150 deletions(-) -- 1.8.1.2

We use le32 etc elsewhere, make it consistent. Signed-off-by: Rusty Russell <rusty@au1.ibm.com> --- virtio-v1.0-wd01-part1-specification.txt 50 ++++++++++++++++---------------- 1 file changed, 25 insertions(+), 25 deletions(-) diff --git a/virtio-v1.0-wd01-part1-specification.txt b/virtio-v1.0-wd01-part1-specification.txt index 54ada26..096993e 100644 --- a/virtio-v1.0-wd01-part1-specification.txt +++ b/virtio-v1.0-wd01-part1-specification.txt @@ -774,24 +774,24 @@ Common configuration structure layout is documented below: struct virtio_pci_common_cfg { /* About the whole device. */ - __le32 device_feature_select; /* read-write */ - __le32 device_feature; /* read-only */ - __le32 guest_feature_select; /* read-write */ - __le32 guest_feature; /* read-write */ - __le16 msix_config; /* read-write */ - __le16 num_queues; /* read-only */ - __u8 device_status; /* read-write */ - __u8 unused1; + le32 device_feature_select; /* read-write */ + le32 device_feature; /* read-only */ + le32 guest_feature_select; /* read-write */ + le32 guest_feature; /* read-write */ + le16 msix_config; /* read-write */ + le16 num_queues; /* read-only */ + u8 device_status; /* read-write */ + u8 unused1; /* About a specific virtqueue. */ - __le16 queue_select; /* read-write */ - __le16 queue_size; /* read-write, power of 2, or 0. */ - __le16 queue_msix_vector; /* read-write */ - __le16 queue_enable; /* read-write */ - __le16 queue_notify_off; /* read-only */ - __le64 queue_desc; /* read-write */ - __le64 queue_avail; /* read-write */ - __le64 queue_used; /* read-write */ + le16 queue_select; /* read-write */ + le16 queue_size; /* read-write, power of 2, or 0. */ + le16 queue_msix_vector; /* read-write */ + le16 queue_enable; /* read-write */ + le16 queue_notify_off; /* read-only */ + le64 queue_desc; /* read-write */ + le64 queue_avail; /* read-write */ + le64 queue_used; /* read-write */ }; device_feature_select @@ -985,14 +985,14 @@ This virtio structure capability uses little-endian format; all bits are read-only: struct virtio_pci_cap { - __u8 cap_vndr; /* Generic PCI field: PCI_CAP_ID_VNDR */ - __u8 cap_next; /* Generic PCI field: next ptr. */ - __u8 cap_len; /* Generic PCI field: capability length */ - __u8 cfg_type; /* Identifies the structure. */ - __u8 bar; /* Where to find it. */ - __u8 padding[3];/* Pad to full dword. */ - __le32 offset; /* Offset within bar. */ - __le32 length; /* Length of the structure, in bytes. */ + u8 cap_vndr; /* Generic PCI field: PCI_CAP_ID_VNDR */ + u8 cap_next; /* Generic PCI field: next ptr. */ + u8 cap_len; /* Generic PCI field: capability length */ + u8 cfg_type; /* Identifies the structure. */ + u8 bar; /* Where to find it. */ + u8 padding[3]; /* Pad to full dword. */ + le32 offset; /* Offset within bar. */ + le32 length; /* Length of the structure, in bytes. */ }; This structure can optionally followed by extra data, depending on @@ -1071,7 +1071,7 @@ by additional fields: struct virtio_pci_notify_cap { struct virtio_pci_cap cap; - __le32 notify_off_multiplier; /* Multiplier for queue_notify_off. */ + le32 notify_off_multiplier; /* Multiplier for queue_notify_off. */ }; notify_off_multiplier -- 1.8.1.2

From: "Michael S. Tsirkin" <mst@redhat.com> make all ring structures LE resolves VIRTIO-9 Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Rusty Russell <rusty@au1.ibm.com> --- virtio-v1.0-wd01-part1-specification.txt 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/virtio-v1.0-wd01-part1-specification.txt b/virtio-v1.0-wd01-part1-specification.txt index 8340c42..7a49d6d 100644 --- a/virtio-v1.0-wd01-part1-specification.txt +++ b/virtio-v1.0-wd01-part1-specification.txt @@ -308,14 +308,12 @@ The legacy virtqueue layout structure therefore looks like this: struct vring_used used; }; -2.1.4.1. A Note on Virtqueue Endianness +2.1.4.1. Legacy Interfaces: A Note on Virtqueue Endianness -------------------------------------- Note that the endian of fields and in the virtqueue is the native -endian of the guest, not little-endian as PCI normally is. This makes -for simpler guest code, and it is assumed that the host already has to -be deeply aware of the guest endian so such an “endian-aware” device -is not a significant issue. +endian of the guest, not little-endian as PCI normally is. +It is assumed that the host is already aware of the guest endian. 2.1.4.2. Message Framing ----------------------- -- 1.8.1.2

From: "Michael S. Tsirkin" <mst@redhat.com> make it all LE and require natural width accesses resolves VIRTIO-9 Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Rusty Russell <rusty@au1.ibm.com> (rebase) --- virtio-v1.0-wd01-part1-specification.txt 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/virtio-v1.0-wd01-part1-specification.txt b/virtio-v1.0-wd01-part1-specification.txt index d1b15c6..8340c42 100644 --- a/virtio-v1.0-wd01-part1-specification.txt +++ b/virtio-v1.0-wd01-part1-specification.txt @@ -88,6 +88,8 @@ o Feature bits o Configuration space o One or more virtqueues +Unless explicitly specified otherwise, all multi-byte fields are little-endian. + 2.1.1. Device Status Field ------------------------- @@ -203,7 +205,13 @@ Interface' in the section title. Configuration space is generally used for rarely-changing or initialization-time parameters. -Note that this space is generally the guest's native endian, +Note that configuration space generally uses the little-endian format +for multi-byte fields. + +100.100.4.1. Legacy Interface: A Note on Configuration Space endian-ness +-------------------------------------- + +Note that for legacy interfaces, configuration space is generally the guest's native endian, rather than PCI's little-endian. 2.1.4. Virtqueues @@ -888,8 +896,8 @@ device-specific region can be accessed using any width accesses, and should obtain the same results. Note that this is possible because while the virtio header is PCI -(i.e. little) endian, the device-specific region is encoded in -the native endian of the guest (where such distinction is +(i.e. little) endian, when using the legacy interface the device-specific +region is encoded in the native endian of the guest (where such distinction is applicable). When used through the legacy interface, the virtio header looks as follows: -- 1.8.1.2

Emphasize that legacy endian was different, but change all the basic types to use le32 etc as a constant reminder. Signed-off-by: Rusty Russell <rusty@au1.ibm.com> --- virtio-v1.0-wd01-part1-specification.txt 227 ++++++++++++++++--------------- 1 file changed, 119 insertions(+), 108 deletions(-) diff --git a/virtio-v1.0-wd01-part1-specification.txt b/virtio-v1.0-wd01-part1-specification.txt index 7a49d6d..db79dc7 100644 --- a/virtio-v1.0-wd01-part1-specification.txt +++ b/virtio-v1.0-wd01-part1-specification.txt @@ -89,6 +89,7 @@ o Configuration space o One or more virtqueues Unless explicitly specified otherwise, all multi-byte fields are little-endian. +To reinforce this the examples use typenames like "le16" instead of "uint16_t". 2.1.1. Device Status Field ------------------------- @@ -211,8 +212,8 @@ for multi-byte fields. 100.100.4.1. Legacy Interface: A Note on Configuration Space endian-ness -------------------------------------- -Note that for legacy interfaces, configuration space is generally the guest's native endian, -rather than PCI's little-endian. +Note that for legacy interfaces, configuration space is generally the +guest's native endian, rather than PCI's little-endian. 2.1.4. Virtqueues ---------------- @@ -312,7 +313,7 @@ The legacy virtqueue layout structure therefore looks like this: -------------------------------------- Note that the endian of fields and in the virtqueue is the native -endian of the guest, not little-endian as PCI normally is. +endian of the guest, not little-endian as specified by this standard. It is assumed that the host is already aware of the guest endian. 2.1.4.2. Message Framing @@ -355,13 +356,18 @@ can be chained via the next field. Each descriptor describes a buffer which is read-only or write-only, but a chain of descriptors can contain both read-only and write-only buffers. +The actual contents of the memory offered to the device depends on the +device type. Most common is to begin the data with a header +(containing little-endian fields) for the device to read, and postfix +it with a status tailer for the device to write. + No descriptor chain may be more than 2^32 bytes long in total. struct vring_desc { /* Address (guest-physical). */ - u64 addr; + le64 addr; /* Length. */ - u32 len; + le32 len; /* This marks a buffer as continuing via the next field. */ #define VRING_DESC_F_NEXT 1 @@ -370,9 +376,9 @@ No descriptor chain may be more than 2^32 bytes long in total. /* This means the buffer contains a list of buffer descriptors. */ #define VRING_DESC_F_INDIRECT 4 /* The flags as indicated above. */ - u16 flags; + le16 flags; /* Next field if flags & NEXT */ - u16 next; + le16 next; }; The number of descriptors in the table is defined by the queue size @@ -429,10 +435,10 @@ entry (modulo the queue size). This starts at 0, and increases. struct vring_avail { #define VRING_AVAIL_F_NO_INTERRUPT 1 - u16 flags; - u16 idx; - u16 ring[ /* Queue Size */ ]; - u16 used_event; /* Only if VIRTIO_RING_F_EVENT_IDX */ + le16 flags; + le16 idx; + le16 ring[ /* Queue Size */ ]; + le16 used_event; /* Only if VIRTIO_RING_F_EVENT_IDX */ }; 2.1.4.5. The Virtqueue Used Ring @@ -458,20 +464,20 @@ for guests using untrusted buffers: if you do not know exactly how much has been written by the device, you usually have to zero the buffer to ensure no data leakage occurs. - /* u32 is used here for ids for padding reasons. */ + /* le32 is used here for ids for padding reasons. */ struct vring_used_elem { /* Index of start of used descriptor chain. */ - u32 id; + le32 id; /* Total length of the descriptor chain which was used (written to) */ - u32 len; + le32 len; }; struct vring_used { #define VRING_USED_F_NO_NOTIFY 1 - u16 flags; - u16 idx; + le16 flags; + le16 idx; struct vring_used_elem ring[ /* Queue Size */]; - u16 avail_event; /* Only if VIRTIO_RING_F_EVENT_IDX */ + le16 avail_event; /* Only if VIRTIO_RING_F_EVENT_IDX */ }; 2.1.4.6. Helpers for Operating Virtqueues @@ -606,7 +612,8 @@ free descriptors before beginning the mappings. ------------------------------------- The head of the buffer we mapped is the first d in the algorithm -above. A naive implementation would do the following: +above. A naive implementation would do the following (with the +appropriate conversion to-and-from little-endian assumed): avail->ring[avail->idx % qsz] = head; @@ -694,11 +701,11 @@ suppressed by the device: vring_disable_interrupts(vq); for (;;) { - if (vq->last_seen_used != vring->used.idx) { + if (vq->last_seen_used != le16_to_cpu(vring->used.idx)) { vring_enable_interrupts(vq); mb(); - if (vq->last_seen_used != vring->used.idx) + if (vq->last_seen_used != le16_to_cpu(vring->used.idx)) break; } @@ -1843,7 +1850,7 @@ VIRTIO_NET_S_LINK_UP and VIRTIO_NET_S_ANNOUNCE. struct virtio_net_config { u8 mac[6]; - u16 status; + le16 status; }; 2.4.1.4. Device Initialization @@ -1906,12 +1913,12 @@ case, the packet itself is preceeded by a header: #define VIRTIO_NET_HDR_GSO_TCPV6 4 #define VIRTIO_NET_HDR_GSO_ECN 0x80 u8 gso_type; - u16 hdr_len; - u16 gso_size; - u16 csum_start; - u16 csum_offset; + le16 hdr_len; + le16 gso_size; + le16 csum_start; + le16 csum_offset; /* Only if VIRTIO_NET_F_MRG_RXBUF: */ - u16 num_buffers; + le16 num_buffers; }; The controlq is used to control device features such as @@ -2078,7 +2085,7 @@ off. The command-specific-data is one byte containing 0 (off) or ----------------------------------------- struct virtio_net_ctrl_mac { - u32 entries; + le32 entries; u8 macs[entries][ETH_ALEN]; }; @@ -2144,7 +2151,7 @@ send control commands for dynamic offloads state configuration. 2.4.1.5.4.3.1. Setting Offloads State ------------------------------------- - u64 offloads; + le64 offloads; #define VIRTIO_NET_F_GUEST_CSUM 1 #define VIRTIO_NET_F_GUEST_TSO4 7 @@ -2158,7 +2165,7 @@ send control commands for dynamic offloads state configuration. The class VIRTIO_NET_CTRL_GUEST_OFFLOADS has one command: VIRTIO_NET_CTRL_GUEST_OFFLOADS_SET applies the new offloads configuration. -u64 value passed as command data is a bitmask, bits set define +le64 value passed as command data is a bitmask, bits set define offloads to be enabled, bits cleared - offloads to be disabled. There is a corresponding device feature for each offload. Upon feature @@ -2217,15 +2224,15 @@ present. The availability of the others all depend on various feature bits as indicated above. struct virtio_blk_config { - u64 capacity; - u32 size_max; - u32 seg_max; + le64 capacity; + le32 size_max; + le32 seg_max; struct virtio_blk_geometry { - u16 cylinders; + le16 cylinders; u8 heads; u8 sectors; } geometry; - u32 blk_size; + le32 blk_size; }; @@ -2252,9 +2259,9 @@ The driver queues requests to the virtqueue, and they are used by the device (not necessarily in order). Each request is of form: struct virtio_blk_req { - u32 type; - u32 reserved; - u64 sector; + le32 type; + le32 reserved; + le64 sector; char data[][512]; u8 status; }; @@ -2408,9 +2415,9 @@ data and outgoing characters are placed in the transmit queue. be fetched. struct virtio_console_config { - u16 cols; - u16 rows; - u32 max_nr_ports; + le16 cols; + le16 rows; + le32 max_nr_ports; }; 2.4.3.5. Device Initialization @@ -2465,9 +2472,9 @@ data and outgoing characters are placed in the transmit queue. buffer and the events associated are: struct virtio_console_control { - uint32_t id; /* Port number */ - uint16_t event; /* The kind of control event */ - uint16_t value; /* Extra information for the event */ + le32 id; /* Port number */ + le16 event; /* The kind of control event */ + le16 value; /* Extra information for the event */ }; /* Some events for the internal messages (control packets) */ @@ -2546,14 +2553,18 @@ guest memory statistics to the host. 2.4.5.4. Device configuration layout ----------------------------------- Both fields of this configuration - are always available. Note that they are little endian, despite - convention that device fields are guest endian: + are always available. struct virtio_balloon_config { - u32 num_pages; - u32 actual; + le32 num_pages; + le32 actual; }; +2.4.5.4.1. Legacy Interface: Device configuration layout +----------------------------------- +Note that these fields are always little endian, despite convention +that legacy device fields are guest endian. + 2.4.5.5. Device Initialization ----------------------------- @@ -2640,8 +2651,8 @@ as follows: #define VIRTIO_BALLOON_S_MINFLT 3 #define VIRTIO_BALLOON_S_MEMFREE 4 #define VIRTIO_BALLOON_S_MEMTOT 5 - u16 tag; - u64 val; + le16 tag; + le64 val; } __attribute__((packed)); 2.4.5.6.2. Memory Statistics Tags @@ -2715,16 +2726,16 @@ targets that receive and process the requests. and cdb_size are writable by the guest. struct virtio_scsi_config { - u32 num_queues; - u32 seg_max; - u32 max_sectors; - u32 cmd_per_lun; - u32 event_info_size; - u32 sense_size; - u32 cdb_size; - u16 max_channel; - u16 max_target; - u32 max_lun; + le32 num_queues; + le32 seg_max; + le32 max_sectors; + le32 cmd_per_lun; + le32 event_info_size; + le32 sense_size; + le32 cdb_size; + le16 max_channel; + le16 max_target; + le32 max_lun; }; num_queues is the total number of request virtqueues exposed by @@ -2796,16 +2807,16 @@ Requests have the following format: struct virtio_scsi_req_cmd { // Read-only u8 lun[8]; - u64 id; + le64 id; u8 task_attr; u8 prio; u8 crn; char cdb[cdb_size]; char dataout[]; // Write-only part - u32 sense_len; - u32 residual; - u16 status_qualifier; + le32 sense_len; + le32 residual; + le16 status_qualifier; u8 status; u8 response; u8 sense[sense_size]; @@ -2912,7 +2923,7 @@ The controlq is used for other SCSI transport operations. Requests have the following format: struct virtio_scsi_ctrl { - u32 type; + le32 type; ... u8 response; }; @@ -2946,12 +2957,12 @@ The following commands are defined: struct virtio_scsi_ctrl_tmf { // Read-only part - u32 type; - u32 subtype; - u8 lun[8]; - u64 id; + le32 type; + le32 subtype; + u8 lun[8]; + le64 id; // Write-only part - u8 response; + u8 response; } /* command-specific response values */ @@ -2981,12 +2992,12 @@ The following commands are defined: struct virtio_scsi_ctrl_an { // Read-only part - u32 type; - u8 lun[8]; - u32 event_requested; + le32 type; + u8 lun[8]; + le32 event_requested; // Write-only part - u32 event_actual; - u8 response; + le32 event_actual; + u8 response; } #define VIRTIO_SCSI_EVT_ASYNC_OPERATIONAL_CHANGE 2 @@ -3014,12 +3025,12 @@ The following commands are defined: struct virtio_scsi_ctrl_an { // Read-only part - u32 type; - u8 lun[8]; - u32 event_requested; + le32 type; + u8 lun[8]; + le32 event_requested; // Write-only part - u32 event_actual; - u8 response; + le32 event_actual; + u8 response; } By sending this command, the driver asks the specified LUN to @@ -3064,9 +3075,9 @@ following format: struct virtio_scsi_event { // Write-only part - u32 event; + le32 event; u8 lun[8]; - i32 reason; + le32 reason; } If bit 31 is set in the event field, the device failed to report @@ -3303,35 +3314,35 @@ VIRTIO_F_ANY_LAYOUT (27) This feature indicates that the device * These can chain together via "next". */ struct vring_desc { /* Address (guest-physical). */ - uint64_t addr; + le64 addr; /* Length. */ - uint32_t len; + le32 len; /* The flags as indicated above. */ - uint16_t flags; + le16 flags; /* We chain unused descriptors via this, too */ - uint16_t next; + le16 next; }; struct vring_avail { - uint16_t flags; - uint16_t idx; - uint16_t ring[]; - /* Only if VIRTIO_RING_F_EVENT_IDX: uint16_t used_event; */ + le16 flags; + le16 idx; + le16 ring[]; + /* Only if VIRTIO_RING_F_EVENT_IDX: le16 used_event; */ }; -/* u32 is used here for ids for padding reasons. */ +/* le32 is used here for ids for padding reasons. */ struct vring_used_elem { /* Index of start of used descriptor chain. */ - uint32_t id; + le32 id; /* Total length of the descriptor chain which was written to. */ - uint32_t len; + le32 len; }; struct vring_used { - uint16_t flags; - uint16_t idx; + le16 flags; + le16 idx; struct vring_used_elem ring[]; - /* Only if VIRTIO_RING_F_EVENT_IDX: uint16_t avail_event; */ + /* Only if VIRTIO_RING_F_EVENT_IDX: le16 avail_event; */ }; struct vring { @@ -3350,19 +3361,19 @@ struct vring { * struct vring_desc desc[num]; * * // A ring of available descriptor heads with free-running index. - * __u16 avail_flags; - * __u16 avail_idx; - * __u16 available[num]; - * __u16 used_event_idx; // Only if VIRTIO_RING_F_EVENT_IDX + * le16 avail_flags; + * le16 avail_idx; + * le16 available[num]; + * le16 used_event_idx; // Only if VIRTIO_RING_F_EVENT_IDX * * // Padding to the next align boundary. * char pad[]; * * // A ring of used descriptor heads with free-running index. - * __u16 used_flags; - * __u16 used_idx; + * le16 used_flags; + * le16 used_idx; * struct vring_used_elem used[num]; - * __u16 avail_event_idx; // Only if VIRTIO_RING_F_EVENT_IDX + * le16 avail_event_idx; // Only if VIRTIO_RING_F_EVENT_IDX * }; * Note: for virtio PCI, align is 4096. */ @@ -3372,16 +3383,16 @@ static inline void vring_init(struct vring *vr, unsigned int num, void *p, vr->num = num; vr->desc = p; vr->avail = p + num*sizeof(struct vring_desc); - vr->used = (void *)(((unsigned long)&vr->avail->ring[num] + sizeof(uint16_t) + vr->used = (void *)(((unsigned long)&vr->avail->ring[num] + sizeof(le16) + align-1) & ~(align - 1)); } static inline unsigned vring_size(unsigned int num, unsigned long align) { - return ((sizeof(struct vring_desc)*num + sizeof(uint16_t)*(3+num) + return ((sizeof(struct vring_desc)*num + sizeof(le16)*(3+num) + align - 1) & ~(align - 1)) - + sizeof(uint16_t)*3 + sizeof(struct vring_used_elem)*num; + + sizeof(le16)*3 + sizeof(struct vring_used_elem)*num; } static inline int vring_need_event(uint16_t event_idx, uint16_t new_idx, uint16_t old_idx) @@ -3390,16 +3401,16 @@ static inline int vring_need_event(uint16_t event_idx, uint16_t new_idx, uint16_ } /* Get location of event indices (only with VIRTIO_RING_F_EVENT_IDX) */ -static inline uint16_t *vring_used_event(struct vring *vr) +static inline le16 *vring_used_event(struct vring *vr) { /* For backwards compat, used event index is at *end* of avail ring. */ return &vr->avail->ring[vr->num]; } -static inline uint16_t *vring_avail_event(struct vring *vr) +static inline le16 *vring_avail_event(struct vring *vr) { /* For backwards compat, avail event index is at *end* of used ring. */ - return (uint16_t *)&vr->used->ring[vr->num]; + return (le16 *)&vr->used->ring[vr->num]; } #endif /* VIRTIO_RING_H */ -- 1.8.1.2

Noone is likely to get confused, but let's be explicit since the rest of the spec is little-endian. Signed-off-by: Rusty Russell <rusty@au1.ibm.com> --- virtio-v1.0-wd01-part1-specification.txt 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) diff --git a/virtio-v1.0-wd01-part1-specification.txt b/virtio-v1.0-wd01-part1-specification.txt index 096993e..6024f85 100644 --- a/virtio-v1.0-wd01-part1-specification.txt +++ b/virtio-v1.0-wd01-part1-specification.txt @@ -1516,6 +1516,10 @@ discoverable via standard operating system algorithms, and adding virtio support is mainly a question of supporting a new control unit type. +As the S/390 is a big endian machine, these channel commands are +big-endian: this is made clear by use of the types be16, be32 and +be64. + 2.3.3.1. Basic Concepts ----------------------- @@ -1555,8 +1559,8 @@ CCW_CMD_READ_VQ_CONF is issued by the guest to obtain information about a queue. It uses the following structure for communicating: struct vq_config_block { - __u16 index; - __u16 max_num; + be16 index; + be16 max_num; } __attribute__ ((packed)); The requested number of buffers for queue index is returned in @@ -1567,10 +1571,10 @@ host about the location used for its queue. The transmitted structure is struct vq_info_block { - __u64 queue; - __u32 align; - __u16 index; - __u16 num; + be64 queue; + be32 align; + be16 index; + be16 num; } __attribute__ ((packed)); queue contains the guest address for queue index. The actual @@ -1615,8 +1619,8 @@ The CCW commands dealing with features use the following communication block: struct virtio_feature_desc { - __u32 features; - __u8 index; + be32 features; + u8 index; } __attribute__ ((packed)); features are the 32 bits of features currently accessed, while -- 1.8.1.2

This is noted at the beginning of the spec, but let's make it explicit everywhere le32 etc is used. Signed-off-by: Rusty Russell <rusty@au1.ibm.com> --- virtio-v1.0-wd01-part1-specification.txt 83 ++++++++++++++++++++++++++++++-- 1 file changed, 80 insertions(+), 3 deletions(-) diff --git a/virtio-v1.0-wd01-part1-specification.txt b/virtio-v1.0-wd01-part1-specification.txt index db79dc7..54ada26 100644 --- a/virtio-v1.0-wd01-part1-specification.txt +++ b/virtio-v1.0-wd01-part1-specification.txt @@ -1853,6 +1853,12 @@ VIRTIO_NET_S_LINK_UP and VIRTIO_NET_S_ANNOUNCE. le16 status; }; +100.4.1.4.1. Legacy Interface: Device configuration layout +-------------------- +For legacy devices, the status field in struct virtio_net_config is the +native endian of the guest rather than (necessarily) little-endian. + + 2.4.1.4. Device Initialization ----------------------------- @@ -1924,6 +1930,11 @@ case, the packet itself is preceeded by a header: The controlq is used to control device features such as filtering. +100.4.1.5.1 Legacy Interface: Device Operation +------------------------ +For legacy devices, the fields in struct virtio_net_hdr are the +native endian of the guest rather than (necessarily) little-endian. + 2.4.1.5.1. Packet Transmission ----------------------------- @@ -2099,6 +2110,11 @@ command-specific-data is two variable length tables of 6-byte MAC addresses. The first table contains unicast addresses, and the second contains multicast addresses. +2.4.1.5.2.2.1. Legacy Interface: Setting MAC Address Filtering +----------------------------------------- +For legacy devices, the entries field in struct virtio_net_ctrl_mac is the +native endian of the guest rather than (necessarily) little-endian. + 2.4.1.5.3.3. VLAN Filtering -------------------------- @@ -2110,7 +2126,12 @@ can control a VLAN filter table in the device. #define VIRTIO_NET_CTRL_VLAN_DEL 1 Both the VIRTIO_NET_CTRL_VLAN_ADD and VIRTIO_NET_CTRL_VLAN_DEL -command take a 16-bit VLAN id as the command-specific-data. +command take a little-endian 16-bit VLAN id as the command-specific-data. + +2.4.1.5.3.3.1. Legacy Interface: VLAN Filtering +----------------------------------------- +For legacy devices, the VLAN id is in the +native endian of the guest rather than (necessarily) little-endian. 2.4.1.5.3.4. Gratuitous Packet Sending ------------------------------------- @@ -2176,6 +2197,12 @@ Corresponding feature must be negotiated at startup in order to allow dynamic change of specific offload state. +2.4.1.5.4.3.1.1. Legacy Interface: Setting Offloads State +------------------------------------- +For legacy devices, the offloads field is the +native endian of the guest rather than (necessarily) little-endian. + + 2.4.2. Block Device ================== @@ -2236,6 +2263,12 @@ bits as indicated above. }; +100.2.4.2.5.1. Legacy Interface: Device configuration layout +-------------------- +For legacy devices, the fields in struct virtio_blk_config are the +native endian of the guest rather than (necessarily) little-endian. + + 2.4.2.4. Device Initialization ----------------------------- @@ -2289,6 +2322,9 @@ error or VIRTIO_BLK_S_UNSUPP for a request unsupported by host: 2.4.2.5.1 Legacy Interface: Device Operation ------------------------ +For legacy devices, the fields in struct virtio_blk_req are the +native endian of the guest rather than (necessarily) little-endian. + The 'reserved' field was previously called ioprio. The ioprio field is a hint about the relative priorities of requests to the device: higher numbers indicate more important requests. @@ -2296,6 +2332,7 @@ higher numbers indicate more important requests. If the device has VIRTIO_BLK_F_SCSI feature, it can also support scsi packet command requests, each of these requests is of form: + /* All fields are in guest's native endian. */ struct virtio_scsi_pc_req { u32 type; u32 ioprio; @@ -2420,6 +2457,11 @@ data and outgoing characters are placed in the transmit queue. le32 max_nr_ports; }; +2.4.3.4.1. Legacy Interface: Device configuration layout +----------------------------------- +For legacy devices, the fields in struct virtio_console_config are the +native endian of the guest rather than (necessarily) little-endian. + 2.4.3.5. Device Initialization ----------------------------- @@ -2487,6 +2529,12 @@ data and outgoing characters are placed in the transmit queue. #define VIRTIO_CONSOLE_PORT_OPEN 6 #define VIRTIO_CONSOLE_PORT_NAME 7 +2.4.3.6.1. Legacy Interface: Device Operation +------------------------ +For legacy devices, the fields in struct virtio_console_control are the +native endian of the guest rather than (necessarily) little-endian. + + 2.4.4. Entropy Device ==================== @@ -2639,8 +2687,7 @@ as follows: subsequent request) and consumes the statistics. Memory Statistics Format Each statistic consists of a 16 bit - tag and a 64 bit value. Both quantities are represented in the - native endian of the guest. All statistics are optional and the + tag and a 64 bit value. All statistics are optional and the driver may choose which ones to supply. To guarantee backwards compatibility, unsupported statistics should be omitted. @@ -2655,6 +2702,12 @@ as follows: le64 val; } __attribute__((packed)); +2.4.5.6.1. Legacy Interface: Memory Statistics +--------------------------- +For legacy devices, the fields in struct virtio_balloon_stat are the +native endian of the guest rather than (necessarily) little-endian. + + 2.4.5.6.2. Memory Statistics Tags -------------------------------- @@ -2774,6 +2827,11 @@ targets that receive and process the requests. as hints to constrain scanning the logical units on the host.h +2.4.6.4.1. Legacy Interface: Device configuration layout +----------------------------------- +For legacy devices, the fields in struct virtio_scsi_config are the +native endian of the guest rather than (necessarily) little-endian. + 2.4.6.5. Device Initialization ----------------------------- @@ -2916,6 +2974,11 @@ following: request will be immediately returned with a response equal to VIRTIO_SCSI_S_FAILURE. +2.4.6.6.1. Legacy Interface: Device Operation: Request Queues +------------------------------------------ +For legacy devices, the fields in struct virtio_scsi_req_cmd are the +native endian of the guest rather than (necessarily) little-endian. + 2.4.6.6.2. Device Operation: controlq ------------------------------------ @@ -3048,6 +3111,15 @@ The following commands are defined: No command-specific values are defined for the response byte. +2.4.6.6.2.1. Legacy Interface: Device Operation: controlq +------------------------------------------ + +For legacy devices, the fields in struct virtio_scsi_ctrl, struct +virtio_scsi_ctrl_tmf, struct virtio_scsi_ctrl_an and struct +virtio_scsi_ctrl_an are the native endian of the guest rather than +(necessarily) little-endian. + + 2.4.6.6.3. Device Operation: eventq ---------------------------------- @@ -3198,6 +3270,11 @@ contents of the event field. The following events are defined: For simplicity, as of this version of the specification the host must never report this event for MMC devices. +2.4.6.6.3.1. Legacy Interface: Device Operation: eventq +---------------------------------- +For legacy devices, the fields in struct virtio_scsi_event are the +native endian of the guest rather than (necessarily) little-endian. + 2.5. Reserved Feature Bits ========================= -- 1.8.1.2