* [Qemu-devel] [PATCH 0/3] Add support for vhost-scsi-pci
@ 2013-03-27 21:59 Nicholas A. Bellinger
2013-03-27 21:59 ` [Qemu-devel] [PATCH 1/3] virtio-scsi: create VirtIOSCSICommon Nicholas A. Bellinger
` (3 more replies)
0 siblings, 4 replies; 16+ messages in thread
From: Nicholas A. Bellinger @ 2013-03-27 21:59 UTC (permalink / raw)
To: target-devel
Cc: kvm-devel, Michael S. Tsirkin, qemu-devel, Nicholas Bellinger,
lf-virt, Anthony Liguori, Stefan Hajnoczi, Paolo Bonzini,
Asias He
From: Nicholas Bellinger <nab@linux-iscsi.org>
Hi QEMU folks,
The following is the patch series to support vhost-scsi-pci within the
upstream QEMU tree.
This includes the refactoring of existing virtio-scsi code from Paolo
to allow a VirtIOSCSICommon structure that is shared amoungst existing
virtio-scsi-pci device and new vhost-scsi-pci device code.
Currently this code requires Asias's seabios bugfixes (commit 5a7730db5
and b44a7be17), and another tcm_vhost patch to reset vq->last_used_idx
after seabios handoff that's currently in flight for v3.9.0 kernel code.
A big thanks to Paolo, Asias, MST, and Stefan for all of their efforts
on this series.
Thank you,
--nab
Nicholas Bellinger (1):
vhost: Only call vhost_verify_ring_mappings when adding vhost memory
Paolo Bonzini (2):
virtio-scsi: create VirtIOSCSICommon
vhost-scsi: new device supporting the tcm_vhost Linux kernel module
configure | 15 +++-
hw/Makefile.objs | 5 +-
hw/s390x/s390-virtio-bus.c | 35 +++++++
hw/vhost-scsi.c | 241 ++++++++++++++++++++++++++++++++++++++++++++
hw/vhost-scsi.h | 66 ++++++++++++
hw/vhost.c | 2 +-
hw/virtio-pci.c | 62 +++++++++++
hw/virtio-scsi.c | 190 +++++++++--------------------------
hw/virtio-scsi.h | 132 +++++++++++++++++++++++-
include/qemu/osdep.h | 4 +
10 files changed, 602 insertions(+), 150 deletions(-)
create mode 100644 hw/vhost-scsi.c
create mode 100644 hw/vhost-scsi.h
--
1.7.2.5
^ permalink raw reply [flat|nested] 16+ messages in thread
* [Qemu-devel] [PATCH 1/3] virtio-scsi: create VirtIOSCSICommon
2013-03-27 21:59 [Qemu-devel] [PATCH 0/3] Add support for vhost-scsi-pci Nicholas A. Bellinger
@ 2013-03-27 21:59 ` Nicholas A. Bellinger
2013-03-27 21:59 ` [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module Nicholas A. Bellinger
` (2 subsequent siblings)
3 siblings, 0 replies; 16+ messages in thread
From: Nicholas A. Bellinger @ 2013-03-27 21:59 UTC (permalink / raw)
To: target-devel
Cc: kvm-devel, Michael S. Tsirkin, qemu-devel, lf-virt,
Anthony Liguori, Stefan Hajnoczi, Paolo Bonzini, Asias He
From: Paolo Bonzini <pbonzini@redhat.com>
This patch refactors existing virtio-scsi code into VirtIOSCSICommon
in order to allow virtio_scsi_init_common() to be used by both internal
virtio_scsi_init() and external vhost-scsi-pci code.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Asias He <asias@redhat.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
---
hw/virtio-scsi.c | 190 +++++++++++++-------------------------------------
hw/virtio-scsi.h | 130 +++++++++++++++++++++++++++++++++--
include/qemu/osdep.h | 4 +
3 files changed, 177 insertions(+), 147 deletions(-)
diff --git a/hw/virtio-scsi.c b/hw/virtio-scsi.c
index 8620712..9dc7150 100644
--- a/hw/virtio-scsi.c
+++ b/hw/virtio-scsi.c
@@ -18,118 +18,6 @@
#include <hw/scsi.h>
#include <hw/scsi-defs.h>
-#define VIRTIO_SCSI_VQ_SIZE 128
-#define VIRTIO_SCSI_CDB_SIZE 32
-#define VIRTIO_SCSI_SENSE_SIZE 96
-#define VIRTIO_SCSI_MAX_CHANNEL 0
-#define VIRTIO_SCSI_MAX_TARGET 255
-#define VIRTIO_SCSI_MAX_LUN 16383
-
-/* Response codes */
-#define VIRTIO_SCSI_S_OK 0
-#define VIRTIO_SCSI_S_OVERRUN 1
-#define VIRTIO_SCSI_S_ABORTED 2
-#define VIRTIO_SCSI_S_BAD_TARGET 3
-#define VIRTIO_SCSI_S_RESET 4
-#define VIRTIO_SCSI_S_BUSY 5
-#define VIRTIO_SCSI_S_TRANSPORT_FAILURE 6
-#define VIRTIO_SCSI_S_TARGET_FAILURE 7
-#define VIRTIO_SCSI_S_NEXUS_FAILURE 8
-#define VIRTIO_SCSI_S_FAILURE 9
-#define VIRTIO_SCSI_S_FUNCTION_SUCCEEDED 10
-#define VIRTIO_SCSI_S_FUNCTION_REJECTED 11
-#define VIRTIO_SCSI_S_INCORRECT_LUN 12
-
-/* Controlq type codes. */
-#define VIRTIO_SCSI_T_TMF 0
-#define VIRTIO_SCSI_T_AN_QUERY 1
-#define VIRTIO_SCSI_T_AN_SUBSCRIBE 2
-
-/* Valid TMF subtypes. */
-#define VIRTIO_SCSI_T_TMF_ABORT_TASK 0
-#define VIRTIO_SCSI_T_TMF_ABORT_TASK_SET 1
-#define VIRTIO_SCSI_T_TMF_CLEAR_ACA 2
-#define VIRTIO_SCSI_T_TMF_CLEAR_TASK_SET 3
-#define VIRTIO_SCSI_T_TMF_I_T_NEXUS_RESET 4
-#define VIRTIO_SCSI_T_TMF_LOGICAL_UNIT_RESET 5
-#define VIRTIO_SCSI_T_TMF_QUERY_TASK 6
-#define VIRTIO_SCSI_T_TMF_QUERY_TASK_SET 7
-
-/* Events. */
-#define VIRTIO_SCSI_T_EVENTS_MISSED 0x80000000
-#define VIRTIO_SCSI_T_NO_EVENT 0
-#define VIRTIO_SCSI_T_TRANSPORT_RESET 1
-#define VIRTIO_SCSI_T_ASYNC_NOTIFY 2
-#define VIRTIO_SCSI_T_PARAM_CHANGE 3
-
-/* Reasons for transport reset event */
-#define VIRTIO_SCSI_EVT_RESET_HARD 0
-#define VIRTIO_SCSI_EVT_RESET_RESCAN 1
-#define VIRTIO_SCSI_EVT_RESET_REMOVED 2
-
-/* SCSI command request, followed by data-out */
-typedef struct {
- uint8_t lun[8]; /* Logical Unit Number */
- uint64_t tag; /* Command identifier */
- uint8_t task_attr; /* Task attribute */
- uint8_t prio;
- uint8_t crn;
- uint8_t cdb[];
-} QEMU_PACKED VirtIOSCSICmdReq;
-
-/* Response, followed by sense data and data-in */
-typedef struct {
- uint32_t sense_len; /* Sense data length */
- uint32_t resid; /* Residual bytes in data buffer */
- uint16_t status_qualifier; /* Status qualifier */
- uint8_t status; /* Command completion status */
- uint8_t response; /* Response values */
- uint8_t sense[];
-} QEMU_PACKED VirtIOSCSICmdResp;
-
-/* Task Management Request */
-typedef struct {
- uint32_t type;
- uint32_t subtype;
- uint8_t lun[8];
- uint64_t tag;
-} QEMU_PACKED VirtIOSCSICtrlTMFReq;
-
-typedef struct {
- uint8_t response;
-} QEMU_PACKED VirtIOSCSICtrlTMFResp;
-
-/* Asynchronous notification query/subscription */
-typedef struct {
- uint32_t type;
- uint8_t lun[8];
- uint32_t event_requested;
-} QEMU_PACKED VirtIOSCSICtrlANReq;
-
-typedef struct {
- uint32_t event_actual;
- uint8_t response;
-} QEMU_PACKED VirtIOSCSICtrlANResp;
-
-typedef struct {
- uint32_t event;
- uint8_t lun[8];
- uint32_t reason;
-} QEMU_PACKED VirtIOSCSIEvent;
-
-typedef struct {
- uint32_t num_queues;
- uint32_t seg_max;
- uint32_t max_sectors;
- uint32_t cmd_per_lun;
- uint32_t event_info_size;
- uint32_t sense_size;
- uint32_t cdb_size;
- uint16_t max_channel;
- uint16_t max_target;
- uint32_t max_lun;
-} QEMU_PACKED VirtIOSCSIConfig;
-
typedef struct VirtIOSCSIReq {
VirtIOSCSI *dev;
VirtQueue *vq;
@@ -178,7 +66,7 @@ static void virtio_scsi_complete_req(VirtIOSCSIReq *req)
scsi_req_unref(req->sreq);
}
g_free(req);
- virtio_notify(&s->vdev, vq);
+ virtio_notify(&s->vs.vdev, vq);
}
static void virtio_scsi_bad_req(void)
@@ -237,7 +125,7 @@ static void virtio_scsi_save_request(QEMUFile *f, SCSIRequest *sreq)
VirtIOSCSIReq *req = sreq->hba_private;
uint32_t n = virtio_queue_get_id(req->vq) - 2;
- assert(n < req->dev->conf->num_queues);
+ assert(n < req->dev->vs.conf->num_queues);
qemu_put_be32s(f, &n);
qemu_put_buffer(f, (unsigned char *)&req->elem, sizeof(req->elem));
}
@@ -251,9 +139,9 @@ static void *virtio_scsi_load_request(QEMUFile *f, SCSIRequest *sreq)
req = g_malloc(sizeof(*req));
qemu_get_be32s(f, &n);
- assert(n < s->conf->num_queues);
+ assert(n < s->vs.conf->num_queues);
qemu_get_buffer(f, (unsigned char *)&req->elem, sizeof(req->elem));
- virtio_scsi_parse_req(s, s->cmd_vqs[n], req);
+ virtio_scsi_parse_req(s, s->vs.cmd_vqs[n], req);
scsi_req_ref(sreq);
req->sreq = sreq;
@@ -468,8 +356,8 @@ static void virtio_scsi_handle_cmd(VirtIODevice *vdev, VirtQueue *vq)
out_size = req->elem.out_sg[0].iov_len;
in_size = req->elem.in_sg[0].iov_len;
- if (out_size < sizeof(VirtIOSCSICmdReq) + s->cdb_size ||
- in_size < sizeof(VirtIOSCSICmdResp) + s->sense_size) {
+ if (out_size < sizeof(VirtIOSCSICmdReq) + s->vs.cdb_size ||
+ in_size < sizeof(VirtIOSCSICmdResp) + s->vs.sense_size) {
virtio_scsi_bad_req();
}
@@ -511,7 +399,7 @@ static void virtio_scsi_get_config(VirtIODevice *vdev,
uint8_t *config)
{
VirtIOSCSIConfig *scsiconf = (VirtIOSCSIConfig *)config;
- VirtIOSCSI *s = (VirtIOSCSI *)vdev;
+ VirtIOSCSICommon *s = (VirtIOSCSICommon *)vdev;
stl_raw(&scsiconf->num_queues, s->conf->num_queues);
stl_raw(&scsiconf->seg_max, 128 - 2);
@@ -537,8 +425,8 @@ static void virtio_scsi_set_config(VirtIODevice *vdev,
exit(1);
}
- s->sense_size = ldl_raw(&scsiconf->sense_size);
- s->cdb_size = ldl_raw(&scsiconf->cdb_size);
+ s->vs.sense_size = ldl_raw(&scsiconf->sense_size);
+ s->vs.cdb_size = ldl_raw(&scsiconf->cdb_size);
}
static uint32_t virtio_scsi_get_features(VirtIODevice *vdev,
@@ -555,8 +443,8 @@ static void virtio_scsi_reset(VirtIODevice *vdev)
qbus_reset_all(&s->bus.qbus);
s->resetting--;
- s->sense_size = VIRTIO_SCSI_SENSE_SIZE;
- s->cdb_size = VIRTIO_SCSI_CDB_SIZE;
+ s->vs.sense_size = VIRTIO_SCSI_SENSE_SIZE;
+ s->vs.cdb_size = VIRTIO_SCSI_CDB_SIZE;
s->events_dropped = false;
}
@@ -566,7 +454,7 @@ static void virtio_scsi_reset(VirtIODevice *vdev)
static void virtio_scsi_save(QEMUFile *f, void *opaque)
{
VirtIOSCSI *s = opaque;
- virtio_save(&s->vdev, f);
+ virtio_save(&s->vs.vdev, f);
}
static int virtio_scsi_load(QEMUFile *f, void *opaque, int version_id)
@@ -574,7 +462,7 @@ static int virtio_scsi_load(QEMUFile *f, void *opaque, int version_id)
VirtIOSCSI *s = opaque;
int ret;
- ret = virtio_load(&s->vdev, f);
+ ret = virtio_load(&s->vs.vdev, f);
if (ret) {
return ret;
}
@@ -584,11 +472,11 @@ static int virtio_scsi_load(QEMUFile *f, void *opaque, int version_id)
static void virtio_scsi_push_event(VirtIOSCSI *s, SCSIDevice *dev,
uint32_t event, uint32_t reason)
{
- VirtIOSCSIReq *req = virtio_scsi_pop_req(s, s->event_vq);
+ VirtIOSCSIReq *req = virtio_scsi_pop_req(s, s->vs.event_vq);
VirtIOSCSIEvent *evt;
int in_size;
- if (!(s->vdev.status & VIRTIO_CONFIG_S_DRIVER_OK)) {
+ if (!(s->vs.vdev.status & VIRTIO_CONFIG_S_DRIVER_OK)) {
return;
}
@@ -643,7 +531,7 @@ static void virtio_scsi_change(SCSIBus *bus, SCSIDevice *dev, SCSISense sense)
{
VirtIOSCSI *s = container_of(bus, VirtIOSCSI, bus);
- if (((s->vdev.guest_features >> VIRTIO_SCSI_F_CHANGE) & 1) &&
+ if (((s->vs.vdev.guest_features >> VIRTIO_SCSI_F_CHANGE) & 1) &&
dev->type != TYPE_ROM) {
virtio_scsi_push_event(s, dev, VIRTIO_SCSI_T_PARAM_CHANGE,
sense.asc | (sense.ascq << 8));
@@ -654,7 +542,7 @@ static void virtio_scsi_hotplug(SCSIBus *bus, SCSIDevice *dev)
{
VirtIOSCSI *s = container_of(bus, VirtIOSCSI, bus);
- if ((s->vdev.guest_features >> VIRTIO_SCSI_F_HOTPLUG) & 1) {
+ if ((s->vs.vdev.guest_features >> VIRTIO_SCSI_F_HOTPLUG) & 1) {
virtio_scsi_push_event(s, dev, VIRTIO_SCSI_T_TRANSPORT_RESET,
VIRTIO_SCSI_EVT_RESET_RESCAN);
}
@@ -664,7 +552,7 @@ static void virtio_scsi_hot_unplug(SCSIBus *bus, SCSIDevice *dev)
{
VirtIOSCSI *s = container_of(bus, VirtIOSCSI, bus);
- if ((s->vdev.guest_features >> VIRTIO_SCSI_F_HOTPLUG) & 1) {
+ if ((s->vs.vdev.guest_features >> VIRTIO_SCSI_F_HOTPLUG) & 1) {
virtio_scsi_push_event(s, dev, VIRTIO_SCSI_T_TRANSPORT_RESET,
VIRTIO_SCSI_EVT_RESET_REMOVED);
}
@@ -686,35 +574,53 @@ static struct SCSIBusInfo virtio_scsi_scsi_info = {
.load_request = virtio_scsi_load_request,
};
-VirtIODevice *virtio_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf)
+VirtIOSCSICommon *virtio_scsi_init_common(DeviceState *dev, VirtIOSCSIConf *proxyconf,
+ size_t sz)
{
- VirtIOSCSI *s;
- static int virtio_scsi_id;
- size_t sz;
+ VirtIOSCSICommon *s;
int i;
- sz = sizeof(VirtIOSCSI) + proxyconf->num_queues * sizeof(VirtQueue *);
- s = (VirtIOSCSI *)virtio_common_init("virtio-scsi", VIRTIO_ID_SCSI,
- sizeof(VirtIOSCSIConfig), sz);
+ /* Place the cmd_vqs array at the end of the block that the caller
+ * requested. Align the array properly, just in case.
+ */
+ sz = ROUND_UP(sz, sizeof(void *));
+ s = (VirtIOSCSICommon *)
+ virtio_common_init("virtio-scsi", VIRTIO_ID_SCSI,
+ sizeof(VirtIOSCSIConfig),
+ sz + sizeof(VirtQueue *) * proxyconf->num_queues);
s->qdev = dev;
s->conf = proxyconf;
+ s->sense_size = VIRTIO_SCSI_SENSE_SIZE;
+ s->cdb_size = VIRTIO_SCSI_CDB_SIZE;
- /* TODO set up vdev function pointers */
s->vdev.get_config = virtio_scsi_get_config;
- s->vdev.set_config = virtio_scsi_set_config;
s->vdev.get_features = virtio_scsi_get_features;
- s->vdev.reset = virtio_scsi_reset;
s->ctrl_vq = virtio_add_queue(&s->vdev, VIRTIO_SCSI_VQ_SIZE,
virtio_scsi_handle_ctrl);
s->event_vq = virtio_add_queue(&s->vdev, VIRTIO_SCSI_VQ_SIZE,
virtio_scsi_handle_event);
+
+ s->cmd_vqs = (VirtQueue **) ((char *)s + sz);
for (i = 0; i < s->conf->num_queues; i++) {
s->cmd_vqs[i] = virtio_add_queue(&s->vdev, VIRTIO_SCSI_VQ_SIZE,
virtio_scsi_handle_cmd);
}
+ return s;
+}
+
+VirtIODevice *virtio_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf)
+{
+ VirtIOSCSI *s;
+ static int virtio_scsi_id;
+
+ s = (VirtIOSCSI *)virtio_scsi_init_common(dev, proxyconf, sizeof(VirtIOSCSI));
+
+ s->vs.vdev.set_config = virtio_scsi_set_config;
+ s->vs.vdev.reset = virtio_scsi_reset;
+
scsi_bus_new(&s->bus, dev, &virtio_scsi_scsi_info);
if (!dev->hotplugged) {
scsi_bus_legacy_handle_cmdline(&s->bus);
@@ -723,12 +629,12 @@ VirtIODevice *virtio_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf)
register_savevm(dev, "virtio-scsi", virtio_scsi_id++, 1,
virtio_scsi_save, virtio_scsi_load, s);
- return &s->vdev;
+ return &s->vs.vdev;
}
void virtio_scsi_exit(VirtIODevice *vdev)
{
VirtIOSCSI *s = (VirtIOSCSI *)vdev;
- unregister_savevm(s->qdev, "virtio-scsi", s);
+ unregister_savevm(s->vs.qdev, "virtio-scsi", s);
virtio_cleanup(vdev);
}
diff --git a/hw/virtio-scsi.h b/hw/virtio-scsi.h
index ccf1e42..2bcf606 100644
--- a/hw/virtio-scsi.h
+++ b/hw/virtio-scsi.h
@@ -26,25 +26,142 @@
#define VIRTIO_SCSI_F_HOTPLUG 1
#define VIRTIO_SCSI_F_CHANGE 2
+#define VIRTIO_SCSI_VQ_SIZE 128
+#define VIRTIO_SCSI_CDB_SIZE 32
+#define VIRTIO_SCSI_SENSE_SIZE 96
+#define VIRTIO_SCSI_MAX_CHANNEL 0
+#define VIRTIO_SCSI_MAX_TARGET 255
+#define VIRTIO_SCSI_MAX_LUN 16383
+
+/* Response codes */
+#define VIRTIO_SCSI_S_OK 0
+#define VIRTIO_SCSI_S_OVERRUN 1
+#define VIRTIO_SCSI_S_ABORTED 2
+#define VIRTIO_SCSI_S_BAD_TARGET 3
+#define VIRTIO_SCSI_S_RESET 4
+#define VIRTIO_SCSI_S_BUSY 5
+#define VIRTIO_SCSI_S_TRANSPORT_FAILURE 6
+#define VIRTIO_SCSI_S_TARGET_FAILURE 7
+#define VIRTIO_SCSI_S_NEXUS_FAILURE 8
+#define VIRTIO_SCSI_S_FAILURE 9
+#define VIRTIO_SCSI_S_FUNCTION_SUCCEEDED 10
+#define VIRTIO_SCSI_S_FUNCTION_REJECTED 11
+#define VIRTIO_SCSI_S_INCORRECT_LUN 12
+
+/* Controlq type codes. */
+#define VIRTIO_SCSI_T_TMF 0
+#define VIRTIO_SCSI_T_AN_QUERY 1
+#define VIRTIO_SCSI_T_AN_SUBSCRIBE 2
+
+/* Valid TMF subtypes. */
+#define VIRTIO_SCSI_T_TMF_ABORT_TASK 0
+#define VIRTIO_SCSI_T_TMF_ABORT_TASK_SET 1
+#define VIRTIO_SCSI_T_TMF_CLEAR_ACA 2
+#define VIRTIO_SCSI_T_TMF_CLEAR_TASK_SET 3
+#define VIRTIO_SCSI_T_TMF_I_T_NEXUS_RESET 4
+#define VIRTIO_SCSI_T_TMF_LOGICAL_UNIT_RESET 5
+#define VIRTIO_SCSI_T_TMF_QUERY_TASK 6
+#define VIRTIO_SCSI_T_TMF_QUERY_TASK_SET 7
+
+/* Events. */
+#define VIRTIO_SCSI_T_EVENTS_MISSED 0x80000000
+#define VIRTIO_SCSI_T_NO_EVENT 0
+#define VIRTIO_SCSI_T_TRANSPORT_RESET 1
+#define VIRTIO_SCSI_T_ASYNC_NOTIFY 2
+#define VIRTIO_SCSI_T_PARAM_CHANGE 3
+
+/* Reasons for transport reset event */
+#define VIRTIO_SCSI_EVT_RESET_HARD 0
+#define VIRTIO_SCSI_EVT_RESET_RESCAN 1
+#define VIRTIO_SCSI_EVT_RESET_REMOVED 2
+
+/* SCSI command request, followed by data-out */
+typedef struct {
+ uint8_t lun[8]; /* Logical Unit Number */
+ uint64_t tag; /* Command identifier */
+ uint8_t task_attr; /* Task attribute */
+ uint8_t prio;
+ uint8_t crn;
+ uint8_t cdb[];
+} QEMU_PACKED VirtIOSCSICmdReq;
+
+/* Response, followed by sense data and data-in */
+typedef struct {
+ uint32_t sense_len; /* Sense data length */
+ uint32_t resid; /* Residual bytes in data buffer */
+ uint16_t status_qualifier; /* Status qualifier */
+ uint8_t status; /* Command completion status */
+ uint8_t response; /* Response values */
+ uint8_t sense[];
+} QEMU_PACKED VirtIOSCSICmdResp;
+
+/* Task Management Request */
+typedef struct {
+ uint32_t type;
+ uint32_t subtype;
+ uint8_t lun[8];
+ uint64_t tag;
+} QEMU_PACKED VirtIOSCSICtrlTMFReq;
+
+typedef struct {
+ uint8_t response;
+} QEMU_PACKED VirtIOSCSICtrlTMFResp;
+
+/* Asynchronous notification query/subscription */
+typedef struct {
+ uint32_t type;
+ uint8_t lun[8];
+ uint32_t event_requested;
+} QEMU_PACKED VirtIOSCSICtrlANReq;
+
+typedef struct {
+ uint32_t event_actual;
+ uint8_t response;
+} QEMU_PACKED VirtIOSCSICtrlANResp;
+
+typedef struct {
+ uint32_t event;
+ uint8_t lun[8];
+ uint32_t reason;
+} QEMU_PACKED VirtIOSCSIEvent;
+
+typedef struct {
+ uint32_t num_queues;
+ uint32_t seg_max;
+ uint32_t max_sectors;
+ uint32_t cmd_per_lun;
+ uint32_t event_info_size;
+ uint32_t sense_size;
+ uint32_t cdb_size;
+ uint16_t max_channel;
+ uint16_t max_target;
+ uint32_t max_lun;
+} QEMU_PACKED VirtIOSCSIConfig;
+
struct VirtIOSCSIConf {
uint32_t num_queues;
uint32_t max_sectors;
uint32_t cmd_per_lun;
};
-typedef struct VirtIOSCSI {
+typedef struct VirtIOSCSICommon {
VirtIODevice vdev;
DeviceState *qdev;
VirtIOSCSIConf *conf;
- SCSIBus bus;
uint32_t sense_size;
uint32_t cdb_size;
- int resetting;
- bool events_dropped;
VirtQueue *ctrl_vq;
VirtQueue *event_vq;
- VirtQueue *cmd_vqs[0];
+ VirtQueue **cmd_vqs;
+} VirtIOSCSICommon;
+
+typedef struct {
+ VirtIOSCSICommon vs;
+
+ SCSIBus bus;
+ int resetting;
+ bool events_dropped;
} VirtIOSCSI;
#define DEFINE_VIRTIO_SCSI_PROPERTIES(_state, _features_field, _conf_field) \
@@ -55,4 +172,7 @@ typedef struct VirtIOSCSI {
DEFINE_PROP_BIT("hotplug", _state, _features_field, VIRTIO_SCSI_F_HOTPLUG, true), \
DEFINE_PROP_BIT("param_change", _state, _features_field, VIRTIO_SCSI_F_CHANGE, true)
+VirtIOSCSICommon *virtio_scsi_init_common(DeviceState *dev, VirtIOSCSIConf *proxyconf,
+ size_t sz);
+
#endif /* _QEMU_VIRTIO_SCSI_H */
diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h
index df24400..8b465fd 100644
--- a/include/qemu/osdep.h
+++ b/include/qemu/osdep.h
@@ -67,6 +67,10 @@ typedef signed int int_fast16_t;
#define MAX(a, b) (((a) > (b)) ? (a) : (b))
#endif
+#ifndef ROUND_UP
+#define ROUND_UP(n,d) (((n) + (d) - 1) & -(d))
+#endif
+
#ifndef DIV_ROUND_UP
#define DIV_ROUND_UP(n,d) (((n) + (d) - 1) / (d))
#endif
--
1.7.2.5
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-27 21:59 [Qemu-devel] [PATCH 0/3] Add support for vhost-scsi-pci Nicholas A. Bellinger
2013-03-27 21:59 ` [Qemu-devel] [PATCH 1/3] virtio-scsi: create VirtIOSCSICommon Nicholas A. Bellinger
@ 2013-03-27 21:59 ` Nicholas A. Bellinger
2013-03-27 22:28 ` Michael S. Tsirkin
2013-03-27 21:59 ` [Qemu-devel] [PATCH 3/3] vhost: Only call vhost_verify_ring_mappings when adding vhost memory Nicholas A. Bellinger
2013-03-27 22:16 ` [Qemu-devel] [PATCH 0/3] Add support for vhost-scsi-pci Michael S. Tsirkin
3 siblings, 1 reply; 16+ messages in thread
From: Nicholas A. Bellinger @ 2013-03-27 21:59 UTC (permalink / raw)
To: target-devel
Cc: kvm-devel, Michael S. Tsirkin, qemu-devel, lf-virt,
Anthony Liguori, Stefan Hajnoczi, Paolo Bonzini, Asias He
From: Paolo Bonzini <pbonzini@redhat.com>
The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
available from the QEMU command-line. Instead, I hardcode it to zero.
Changes in V4:
- Set event_idx=off by default (nab, thanks asias)
- Disable hotplug feature bit for v3.9 tcm_vhost kernel code, need to
re-enable in v3.10 (nab)
- Update to latest qemu.git/master HEAD
Changes in V3:
- Drop ioeventfd vhost_scsi_properties (asias, thanks stefanha)
- Add CONFIG_VHOST_SCSI (asias, thanks stefanha)
- Add hotplug feature bit
Changes in V2:
- Add backend guest masking support (nab)
- Bump ABI_VERSION to 1 (nab)
- Set up set_guest_notifiers (asias)
- Set up vs->dev.vq_index (asias)
- Drop vs->vs.vdev.{set,clear}_vhost_endpoint (asias)
- Drop VIRTIO_CONFIG_S_DRIVER check in vhost_scsi_set_status (asias)
Howto:
Use the latest seabios, at least commit b44a7be17b
git clone git://git.seabios.org/seabios.git
make
cp out/bios.bin /usr/share/qemu/bios.bin
qemu -device vhost-scsi-pci,wwpn=naa.6001405bd4e8476d,event_idx=off
...
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Asias He <asias@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
---
configure | 15 +++-
hw/Makefile.objs | 5 +-
hw/s390x/s390-virtio-bus.c | 35 +++++++
hw/vhost-scsi.c | 241 ++++++++++++++++++++++++++++++++++++++++++++
hw/vhost-scsi.h | 66 ++++++++++++
hw/virtio-pci.c | 62 +++++++++++
hw/virtio-scsi.h | 2 +
7 files changed, 424 insertions(+), 2 deletions(-)
create mode 100644 hw/vhost-scsi.c
create mode 100644 hw/vhost-scsi.h
diff --git a/configure b/configure
index f2af714..6b67e35 100755
--- a/configure
+++ b/configure
@@ -169,6 +169,7 @@ libattr=""
xfs=""
vhost_net="no"
+vhost_scsi="no"
kvm="no"
gprof="no"
debug_tcg="no"
@@ -532,6 +533,7 @@ Haiku)
usb="linux"
kvm="yes"
vhost_net="yes"
+ vhost_scsi="yes"
if [ "$cpu" = "i386" -o "$cpu" = "x86_64" ] ; then
audio_possible_drivers="$audio_possible_drivers fmod"
fi
@@ -858,6 +860,10 @@ for opt do
;;
--enable-vhost-net) vhost_net="yes"
;;
+ --disable-vhost-scsi) vhost_scsi="no"
+ ;;
+ --enable-vhost-scsi) vhost_scsi="yes"
+ ;;
--disable-glx) glx="no"
;;
--enable-glx) glx="yes"
@@ -3078,7 +3084,7 @@ fi
# __sync_fetch_and_and requires at least -march=i486. Many toolchains
# use i686 as default anyway, but for those that don't, an explicit
# specification is necessary
-if test "$vhost_net" = "yes" && test "$cpu" = "i386"; then
+if (test "$vhost_net" = "yes" -o "$vhost_scsi" = "yes") && test "$cpu" = "i386"; then
cat > $TMPC << EOF
static int sfaa(int *ptr)
{
@@ -3424,6 +3430,7 @@ echo "sigev_thread_id $sigev_thread_id"
echo "uuid support $uuid"
echo "libcap-ng support $cap_ng"
echo "vhost-net support $vhost_net"
+echo "vhost-scsi support $vhost_scsi"
echo "Trace backend $trace_backend"
echo "Trace output file $trace_file-<pid>"
echo "spice support $spice ($spice_protocol_version/$spice_server_version)"
@@ -3697,6 +3704,9 @@ fi
if test "$virtfs" = "yes" ; then
echo "CONFIG_VIRTFS=y" >> $config_host_mak
fi
+if test "$vhost_scsi" = "yes" ; then
+ echo "CONFIG_VHOST_SCSI=y" >> $config_host_mak
+fi
if test "$blobs" = "yes" ; then
echo "INSTALL_BLOBS=yes" >> $config_host_mak
fi
@@ -4175,6 +4185,9 @@ case "$target_arch2" in
if test "$vhost_net" = "yes" ; then
echo "CONFIG_VHOST_NET=y" >> $config_target_mak
fi
+ if test "$vhost_scsi" = "yes" ; then
+ echo "CONFIG_VHOST_SCSI=y" >> $config_target_mak
+ fi
fi
esac
case "$target_arch2" in
diff --git a/hw/Makefile.objs b/hw/Makefile.objs
index d0b2ecb..0a38eb5 100644
--- a/hw/Makefile.objs
+++ b/hw/Makefile.objs
@@ -205,8 +205,11 @@ common-obj-$(CONFIG_XEN_BACKEND) += xen_console.o xenfb.o xen_disk.o xen_nic.o
obj-$(CONFIG_VIRTIO) += dataplane/
obj-$(CONFIG_VIRTIO) += virtio.o virtio-blk.o virtio-balloon.o virtio-net.o
obj-$(CONFIG_VIRTIO) += virtio-serial-bus.o virtio-scsi.o
+ifeq ($(CONFIG_VIRTIO), y)
+obj-$(CONFIG_LINUX) += vhost-scsi.o
+endif
obj-$(CONFIG_SOFTMMU) += vhost_net.o
-obj-$(CONFIG_VHOST_NET) += vhost.o
+obj-$(CONFIG_LINUX) += vhost.o
obj-$(CONFIG_REALLY_VIRTFS) += 9pfs/
obj-$(CONFIG_VGA) += vga.o
diff --git a/hw/s390x/s390-virtio-bus.c b/hw/s390x/s390-virtio-bus.c
index c5d5456..73d93fd 100644
--- a/hw/s390x/s390-virtio-bus.c
+++ b/hw/s390x/s390-virtio-bus.c
@@ -28,6 +28,8 @@
#include "hw/virtio-rng.h"
#include "hw/virtio-serial.h"
#include "hw/virtio-net.h"
+#include "hw/virtio-scsi.h"
+#include "hw/vhost-scsi.h"
#include "hw/sysbus.h"
#include "sysemu/kvm.h"
@@ -214,6 +216,18 @@ static int s390_virtio_scsi_init(VirtIOS390Device *dev)
return s390_virtio_device_init(dev, vdev);
}
+static int s390_vhost_scsi_init(VirtIOS390Device *dev)
+{
+ VirtIODevice *vdev;
+
+ vdev = vhost_scsi_init((DeviceState *)dev, &dev->scsi);
+ if (!vdev) {
+ return -1;
+ }
+
+ return s390_virtio_device_init(dev, vdev);
+}
+
static int s390_virtio_rng_init(VirtIOS390Device *dev)
{
VirtIODevice *vdev;
@@ -537,6 +551,27 @@ static const TypeInfo virtio_s390_device_info = {
.abstract = true,
};
+static Property s390_vhost_scsi_properties[] = {
+ DEFINE_VIRTIO_SCSI_PROPERTIES(VirtIOS390Device, host_features, scsi),
+ DEFINE_PROP_END_OF_LIST(),
+};
+
+static void s390_vhost_scsi_class_init(ObjectClass *klass, void *data)
+{
+ DeviceClass *dc = DEVICE_CLASS(klass);
+ VirtIOS390DeviceClass *k = VIRTIO_S390_DEVICE_CLASS(klass);
+
+ k->init = s390_vhost_scsi_init;
+ dc->props = s390_vhost_scsi_properties;
+}
+
+static const TypeInfo s390_vhost_scsi = {
+ .name = "vhost-scsi-s390",
+ .parent = TYPE_VIRTIO_S390_DEVICE,
+ .instance_size = sizeof(VirtIOS390Device),
+ .class_init = s390_vhost_scsi_class_init,
+};
+
static Property s390_virtio_scsi_properties[] = {
DEFINE_VIRTIO_SCSI_PROPERTIES(VirtIOS390Device, host_features, scsi),
DEFINE_PROP_END_OF_LIST(),
diff --git a/hw/vhost-scsi.c b/hw/vhost-scsi.c
new file mode 100644
index 0000000..70e42fc
--- /dev/null
+++ b/hw/vhost-scsi.c
@@ -0,0 +1,241 @@
+/*
+ * vhost_scsi host device
+ *
+ * Copyright IBM, Corp. 2011
+ *
+ * Authors:
+ * Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
+ *
+ * Changes for QEMU mainline + tcm_vhost kernel upstream:
+ * Nicholas Bellinger <nab@risingtidesystems.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+
+#include <sys/ioctl.h>
+#include "config.h"
+#include "qemu/queue.h"
+#include "monitor/monitor.h"
+#include "migration/migration.h"
+#include "vhost-scsi.h"
+#include "vhost.h"
+#include "virtio-scsi.h"
+
+typedef struct VHostSCSI {
+ VirtIOSCSICommon vs;
+
+ Error *migration_blocker;
+
+ struct vhost_dev dev;
+} VHostSCSI;
+
+static int vhost_scsi_set_endpoint(VirtIODevice *vdev)
+{
+ VHostSCSI *vs = (VHostSCSI *)vdev;
+ struct vhost_scsi_target backend;
+ int ret;
+
+ memset(&backend, 0, sizeof(backend));
+ pstrcpy(backend.vhost_wwpn, sizeof(backend.vhost_wwpn), vs->vs.conf->wwpn);
+ ret = ioctl(vs->dev.control, VHOST_SCSI_SET_ENDPOINT, &backend);
+ if (ret < 0) {
+ return -errno;
+ }
+ return 0;
+}
+
+static void vhost_scsi_clear_endpoint(VirtIODevice *vdev)
+{
+ VHostSCSI *vs = (VHostSCSI *)vdev;
+ struct vhost_scsi_target backend;
+
+ memset(&backend, 0, sizeof(backend));
+ pstrcpy(backend.vhost_wwpn, sizeof(backend.vhost_wwpn), vs->vs.conf->wwpn);
+ ioctl(vs->dev.control, VHOST_SCSI_CLEAR_ENDPOINT, &backend);
+}
+
+static int vhost_scsi_start(VHostSCSI *vs, VirtIODevice *vdev)
+{
+ int ret, abi_version;
+
+ if (!vdev->binding->set_guest_notifiers) {
+ error_report("binding does not support guest notifiers");
+ return -ENOSYS;
+ }
+
+ ret = ioctl(vs->dev.control, VHOST_SCSI_GET_ABI_VERSION, &abi_version);
+ if (ret < 0) {
+ return -errno;
+ }
+ if (abi_version > VHOST_SCSI_ABI_VERSION) {
+ error_report("vhost-scsi: The running tcm_vhost kernel abi_version:"
+ " %d is greater than vhost_scsi userspace supports: %d, please"
+ " upgrade your version of QEMU\n", abi_version,
+ VHOST_SCSI_ABI_VERSION);
+ return -ENOSYS;
+ }
+
+ ret = vhost_dev_enable_notifiers(&vs->dev, vdev);
+ if (ret < 0) {
+ return ret;
+ }
+
+ ret = vhost_dev_start(&vs->dev, vdev);
+ if (ret < 0) {
+ error_report("Error start vhost dev");
+ goto err_notifiers;
+ }
+
+ ret = vhost_scsi_set_endpoint(vdev);
+ if (ret < 0) {
+ error_report("Error set vhost-scsi endpoint");
+ goto err_vhost_stop;
+ }
+
+ ret = vdev->binding->set_guest_notifiers(vdev->binding_opaque, vs->dev.nvqs, true);
+ if (ret < 0) {
+ error_report("Error binding guest notifier");
+ goto err_endpoint;
+ }
+ return ret;
+
+err_endpoint:
+ vhost_scsi_clear_endpoint(vdev);
+err_vhost_stop:
+ vhost_dev_stop(&vs->dev, vdev);
+err_notifiers:
+ vhost_dev_disable_notifiers(&vs->dev, vdev);
+ return ret;
+}
+
+static void vhost_scsi_stop(VHostSCSI *vs, VirtIODevice *vdev)
+{
+ int ret = 0;
+
+ if (!vdev->binding->set_guest_notifiers) {
+ ret = vdev->binding->set_guest_notifiers(vdev->binding_opaque,
+ vs->dev.nvqs, false);
+ if (ret < 0) {
+ error_report("vhost guest notifier cleanup failed: %d\n", ret);
+ }
+ }
+ assert(ret >= 0);
+
+ vhost_scsi_clear_endpoint(vdev);
+ vhost_dev_stop(&vs->dev, vdev);
+ vhost_dev_disable_notifiers(&vs->dev, vdev);
+}
+
+static void vhost_scsi_set_config(VirtIODevice *vdev,
+ const uint8_t *config)
+{
+ VirtIOSCSIConfig *scsiconf = (VirtIOSCSIConfig *)config;
+ VHostSCSI *vs = (VHostSCSI *)vdev;
+
+ if ((uint32_t) ldl_raw(&scsiconf->sense_size) != vs->vs.sense_size ||
+ (uint32_t) ldl_raw(&scsiconf->cdb_size) != vs->vs.cdb_size) {
+ error_report("vhost-scsi does not support changing the sense data and CDB sizes");
+ exit(1);
+ }
+}
+
+static void vhost_scsi_set_status(VirtIODevice *vdev, uint8_t val)
+{
+ VHostSCSI *vs = (VHostSCSI *)vdev;
+ bool start = (val & VIRTIO_CONFIG_S_DRIVER_OK);
+
+ if (vs->dev.started == start) {
+ return;
+ }
+
+ if (start) {
+ int ret;
+
+ ret = vhost_scsi_start(vs, vdev);
+ if (ret < 0) {
+ error_report("virtio-scsi: unable to start vhost: %s\n",
+ strerror(-ret));
+
+ /* There is no userspace virtio-scsi fallback so exit */
+ exit(1);
+ }
+ } else {
+ vhost_scsi_stop(vs, vdev);
+ }
+}
+
+static void vhost_scsi_guest_notifier_mask(VirtIODevice *vdev, int idx,
+ bool mask)
+{
+ VHostSCSI *vs = (VHostSCSI *)vdev;
+
+ vhost_virtqueue_mask(&vs->dev, vdev, idx, mask);
+}
+
+static bool vhost_scsi_guest_notifier_pending(VirtIODevice *vdev, int idx)
+{
+ VHostSCSI *vs = (VHostSCSI *)vdev;
+
+ return vhost_virtqueue_pending(&vs->dev, idx);
+}
+
+VirtIODevice *vhost_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf)
+{
+ VHostSCSI *vs;
+ int vhostfd = -1;
+ int ret;
+
+ if (!proxyconf->wwpn) {
+ error_report("vhost-scsi: missing wwpn\n");
+ return NULL;
+ }
+
+ if (proxyconf->vhostfd) {
+ vhostfd = monitor_handle_fd_param(cur_mon, proxyconf->vhostfd);
+ if (vhostfd == -1) {
+ error_report("vhost-scsi: unable to parse vhostfd\n");
+ return NULL;
+ }
+ }
+
+ vs = (VHostSCSI *)virtio_scsi_init_common(dev, proxyconf,
+ sizeof(VHostSCSI));
+
+ vs->vs.vdev.set_config = vhost_scsi_set_config;
+ vs->vs.vdev.set_status = vhost_scsi_set_status;
+ vs->vs.vdev.guest_notifier_mask = vhost_scsi_guest_notifier_mask;
+ vs->vs.vdev.guest_notifier_pending = vhost_scsi_guest_notifier_pending;
+
+ vs->dev.nvqs = VHOST_SCSI_VQ_NUM_FIXED + vs->vs.conf->num_queues;
+ vs->dev.vqs = g_new(struct vhost_virtqueue, vs->dev.nvqs);
+ vs->dev.vq_index = 0;
+
+ ret = vhost_dev_init(&vs->dev, vhostfd, "/dev/vhost-scsi", true);
+ if (ret < 0) {
+ error_report("vhost-scsi: vhost initialization failed: %s\n",
+ strerror(-ret));
+ return NULL;
+ }
+ vs->dev.backend_features = 0;
+ vs->dev.acked_features = 0;
+
+ error_setg(&vs->migration_blocker,
+ "vhost-scsi does not support migration");
+ migrate_add_blocker(vs->migration_blocker);
+
+ return &vs->vs.vdev;
+}
+
+void vhost_scsi_exit(VirtIODevice *vdev)
+{
+ VHostSCSI *vs = (VHostSCSI *)vdev;
+ migrate_del_blocker(vs->migration_blocker);
+ error_free(vs->migration_blocker);
+
+ /* This will stop vhost backend. */
+ vhost_scsi_set_status(vdev, 0);
+ g_free(vs->dev.vqs);
+ virtio_cleanup(vdev);
+}
diff --git a/hw/vhost-scsi.h b/hw/vhost-scsi.h
new file mode 100644
index 0000000..b01f012
--- /dev/null
+++ b/hw/vhost-scsi.h
@@ -0,0 +1,66 @@
+/*
+ * vhost_scsi host device
+ *
+ * Copyright IBM, Corp. 2011
+ *
+ * Authors:
+ * Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+
+#ifndef VHOST_SCSI_H
+#define VHOST_SCSI_H
+
+#include "qemu-common.h"
+#include "qdev.h"
+#include "virtio-scsi.h"
+
+/*
+ * Used by QEMU userspace to ensure a consistent vhost-scsi ABI.
+ *
+ * ABI Rev 0: July 2012 version starting point for v3.6-rc merge candidate +
+ * RFC-v2 vhost-scsi userspace. Add GET_ABI_VERSION ioctl usage
+ * ABI Rev 1: January 2013. Ignore vhost_tpgt filed in struct vhost_scsi_target.
+ * All the targets under vhost_wwpn can be seen and used by guest.
+ */
+
+#define VHOST_SCSI_ABI_VERSION 1
+
+/* TODO #include <linux/vhost.h> properly */
+/* For VHOST_SCSI_SET_ENDPOINT/VHOST_SCSI_CLEAR_ENDPOINT ioctl */
+struct vhost_scsi_target {
+ int abi_version;
+ char vhost_wwpn[224];
+ unsigned short vhost_tpgt;
+ unsigned short reserved;
+};
+
+enum vhost_scsi_vq_list {
+ VHOST_SCSI_VQ_CONTROL = 0,
+ VHOST_SCSI_VQ_EVENT = 1,
+ VHOST_SCSI_VQ_NUM_FIXED = 2,
+};
+
+#define VHOST_VIRTIO 0xAF
+#define VHOST_SCSI_SET_ENDPOINT _IOW(VHOST_VIRTIO, 0x40, struct vhost_scsi_target)
+#define VHOST_SCSI_CLEAR_ENDPOINT _IOW(VHOST_VIRTIO, 0x41, struct vhost_scsi_target)
+#define VHOST_SCSI_GET_ABI_VERSION _IOW(VHOST_VIRTIO, 0x42, int)
+
+#define DEFINE_VHOST_SCSI_PROPERTIES(_state, _features_field, _conf_field) \
+ DEFINE_PROP_BIT("indirect_desc", _state, _features_field, VIRTIO_RING_F_INDIRECT_DESC, true), \
+ DEFINE_PROP_BIT("event_idx", _state, _features_field, VIRTIO_RING_F_EVENT_IDX, false), \
+ DEFINE_PROP_STRING("vhostfd", _state, _conf_field.vhostfd), \
+ DEFINE_PROP_STRING("wwpn", _state, _conf_field.wwpn), \
+ DEFINE_PROP_UINT32("num_queues", _state, _conf_field.num_queues, 1), \
+ DEFINE_PROP_UINT32("max_sectors", _state, _conf_field.max_sectors, 0xFFFF), \
+ DEFINE_PROP_UINT32("cmd_per_lun", _state, _conf_field.cmd_per_lun, 128), \
+ DEFINE_PROP_BIT("hotplug", _state, _features_field, VIRTIO_SCSI_F_HOTPLUG, true)
+
+VirtIODevice *vhost_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf);
+void vhost_scsi_exit(VirtIODevice *vdev);
+
+
+#endif
diff --git a/hw/virtio-pci.c b/hw/virtio-pci.c
index f3ece78..0d67b84 100644
--- a/hw/virtio-pci.c
+++ b/hw/virtio-pci.c
@@ -22,6 +22,7 @@
#include "hw/virtio-net.h"
#include "hw/virtio-serial.h"
#include "hw/virtio-scsi.h"
+#include "hw/vhost-scsi.h"
#include "hw/pci/pci.h"
#include "qemu/error-report.h"
#include "hw/pci/msi.h"
@@ -1247,6 +1248,64 @@ static const TypeInfo virtio_scsi_info = {
.class_init = virtio_scsi_class_init,
};
+#ifdef CONFIG_VHOST_SCSI
+static int vhost_scsi_init_pci(PCIDevice *pci_dev)
+{
+ VirtIOPCIProxy *proxy = DO_UPCAST(VirtIOPCIProxy, pci_dev, pci_dev);
+ VirtIODevice *vdev;
+
+ vdev = vhost_scsi_init(&pci_dev->qdev, &proxy->scsi);
+ if (!vdev) {
+ return -EINVAL;
+ }
+
+ vdev->nvectors = proxy->nvectors == DEV_NVECTORS_UNSPECIFIED
+ ? proxy->scsi.num_queues + 3
+ : proxy->nvectors;
+ virtio_init_pci(proxy, vdev);
+
+ /* make the actual value visible */
+ proxy->nvectors = vdev->nvectors;
+ return 0;
+}
+
+static void vhost_scsi_exit_pci(PCIDevice *pci_dev)
+{
+ VirtIOPCIProxy *proxy = DO_UPCAST(VirtIOPCIProxy, pci_dev, pci_dev);
+
+ vhost_scsi_exit(proxy->vdev);
+ virtio_exit_pci(pci_dev);
+}
+
+static Property vhost_scsi_properties[] = {
+ DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors, DEV_NVECTORS_UNSPECIFIED),
+ DEFINE_VHOST_SCSI_PROPERTIES(VirtIOPCIProxy, host_features, scsi),
+ DEFINE_PROP_END_OF_LIST(),
+};
+
+static void vhost_scsi_class_init(ObjectClass *klass, void *data)
+{
+ DeviceClass *dc = DEVICE_CLASS(klass);
+ PCIDeviceClass *k = PCI_DEVICE_CLASS(klass);
+
+ k->init = vhost_scsi_init_pci;
+ k->exit = vhost_scsi_exit_pci;
+ k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
+ k->device_id = PCI_DEVICE_ID_VIRTIO_SCSI;
+ k->revision = 0x00;
+ k->class_id = PCI_CLASS_STORAGE_SCSI;
+ dc->reset = virtio_pci_reset;
+ dc->props = vhost_scsi_properties;
+}
+
+static const TypeInfo vhost_scsi_info = {
+ .name = "vhost-scsi-pci",
+ .parent = TYPE_PCI_DEVICE,
+ .instance_size = sizeof(VirtIOPCIProxy),
+ .class_init = vhost_scsi_class_init,
+};
+#endif
+
#ifdef CONFIG_VIRTFS
static int virtio_9p_init_pci(PCIDevice *pci_dev)
{
@@ -1495,6 +1554,9 @@ static void virtio_pci_register_types(void)
type_register_static(&virtio_serial_info);
type_register_static(&virtio_balloon_info);
type_register_static(&virtio_scsi_info);
+#ifdef CONFIG_VHOST_SCSI
+ type_register_static(&vhost_scsi_info);
+#endif
type_register_static(&virtio_rng_info);
type_register_static(&virtio_pci_bus_info);
type_register_static(&virtio_pci_info);
diff --git a/hw/virtio-scsi.h b/hw/virtio-scsi.h
index 2bcf606..356bf38 100644
--- a/hw/virtio-scsi.h
+++ b/hw/virtio-scsi.h
@@ -142,6 +142,8 @@ struct VirtIOSCSIConf {
uint32_t num_queues;
uint32_t max_sectors;
uint32_t cmd_per_lun;
+ char *vhostfd;
+ char *wwpn;
};
typedef struct VirtIOSCSICommon {
--
1.7.2.5
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [Qemu-devel] [PATCH 3/3] vhost: Only call vhost_verify_ring_mappings when adding vhost memory
2013-03-27 21:59 [Qemu-devel] [PATCH 0/3] Add support for vhost-scsi-pci Nicholas A. Bellinger
2013-03-27 21:59 ` [Qemu-devel] [PATCH 1/3] virtio-scsi: create VirtIOSCSICommon Nicholas A. Bellinger
2013-03-27 21:59 ` [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module Nicholas A. Bellinger
@ 2013-03-27 21:59 ` Nicholas A. Bellinger
2013-03-27 22:15 ` Michael S. Tsirkin
2013-03-27 22:16 ` [Qemu-devel] [PATCH 0/3] Add support for vhost-scsi-pci Michael S. Tsirkin
3 siblings, 1 reply; 16+ messages in thread
From: Nicholas A. Bellinger @ 2013-03-27 21:59 UTC (permalink / raw)
To: target-devel
Cc: kvm-devel, Michael S. Tsirkin, qemu-devel, Nicholas Bellinger,
lf-virt, Anthony Liguori, Stefan Hajnoczi, Paolo Bonzini,
Asias He
From: Nicholas Bellinger <nab@linux-iscsi.org>
This patch fixes a bug where vhost_verify_ring_mappings() could be
called after vhost_dev_unassign_memory() has been called during
vhost_set_memory(..., add=false).
This appears to be fallout from when dev->started = true assignment
occurs before vhost_set_memory(..., add=false) is called in:
commit 24f4fe345c1b80bab1ee18573914123d8028a9e6
Author: Michael S. Tsirkin <mst@redhat.com>
Date: Tue Dec 25 17:41:07 2012 +0200
vhost: set started flag while start is in progress
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Asias He <asias@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
---
hw/vhost.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/hw/vhost.c b/hw/vhost.c
index 4d6aee3..687a689 100644
--- a/hw/vhost.c
+++ b/hw/vhost.c
@@ -421,7 +421,7 @@ static void vhost_set_memory(MemoryListener *listener,
return;
}
- if (dev->started) {
+ if (dev->started && add) {
r = vhost_verify_ring_mappings(dev, start_addr, size);
assert(r >= 0);
}
--
1.7.2.5
^ permalink raw reply related [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 3/3] vhost: Only call vhost_verify_ring_mappings when adding vhost memory
2013-03-27 21:59 ` [Qemu-devel] [PATCH 3/3] vhost: Only call vhost_verify_ring_mappings when adding vhost memory Nicholas A. Bellinger
@ 2013-03-27 22:15 ` Michael S. Tsirkin
0 siblings, 0 replies; 16+ messages in thread
From: Michael S. Tsirkin @ 2013-03-27 22:15 UTC (permalink / raw)
To: Nicholas A. Bellinger
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Wed, Mar 27, 2013 at 09:59:46PM +0000, Nicholas A. Bellinger wrote:
> From: Nicholas Bellinger <nab@linux-iscsi.org>
>
> This patch fixes a bug where vhost_verify_ring_mappings() could be
> called after vhost_dev_unassign_memory() has been called during
> vhost_set_memory(..., add=false).
>
Sorry, I NAK'd this already.
Is a memory including the rings unassigned then?
How can this happen?
We need to debug wht's going on here.
> This appears to be fallout from when dev->started = true assignment
> occurs before vhost_set_memory(..., add=false) is called in:
>
> commit 24f4fe345c1b80bab1ee18573914123d8028a9e6
> Author: Michael S. Tsirkin <mst@redhat.com>
> Date: Tue Dec 25 17:41:07 2012 +0200
>
> vhost: set started flag while start is in progress
>
> Cc: Michael S. Tsirkin <mst@redhat.com>
> Cc: Asias He <asias@redhat.com>
> Cc: Paolo Bonzini <pbonzini@redhat.com>
> Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
-net seems to work fine with that patch.
> ---
> hw/vhost.c | 2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/hw/vhost.c b/hw/vhost.c
> index 4d6aee3..687a689 100644
> --- a/hw/vhost.c
> +++ b/hw/vhost.c
> @@ -421,7 +421,7 @@ static void vhost_set_memory(MemoryListener *listener,
> return;
> }
>
> - if (dev->started) {
> + if (dev->started && add) {
> r = vhost_verify_ring_mappings(dev, start_addr, size);
> assert(r >= 0);
> }
> --
> 1.7.2.5
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 0/3] Add support for vhost-scsi-pci
2013-03-27 21:59 [Qemu-devel] [PATCH 0/3] Add support for vhost-scsi-pci Nicholas A. Bellinger
` (2 preceding siblings ...)
2013-03-27 21:59 ` [Qemu-devel] [PATCH 3/3] vhost: Only call vhost_verify_ring_mappings when adding vhost memory Nicholas A. Bellinger
@ 2013-03-27 22:16 ` Michael S. Tsirkin
3 siblings, 0 replies; 16+ messages in thread
From: Michael S. Tsirkin @ 2013-03-27 22:16 UTC (permalink / raw)
To: Nicholas A. Bellinger
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Wed, Mar 27, 2013 at 09:59:43PM +0000, Nicholas A. Bellinger wrote:
> From: Nicholas Bellinger <nab@linux-iscsi.org>
>
> Hi QEMU folks,
>
> The following is the patch series to support vhost-scsi-pci within the
> upstream QEMU tree.
>
> This includes the refactoring of existing virtio-scsi code from Paolo
> to allow a VirtIOSCSICommon structure that is shared amoungst existing
> virtio-scsi-pci device and new vhost-scsi-pci device code.
>
> Currently this code requires Asias's seabios bugfixes (commit 5a7730db5
> and b44a7be17), and another tcm_vhost patch to reset vq->last_used_idx
> after seabios handoff that's currently in flight for v3.9.0 kernel code.
>
> A big thanks to Paolo, Asias, MST, and Stefan for all of their efforts
> on this series.
>
> Thank you,
>
> --nab
Looks sane but 3/3 is a hack, needs to be debugged and fixed properly.
> Nicholas Bellinger (1):
> vhost: Only call vhost_verify_ring_mappings when adding vhost memory
>
> Paolo Bonzini (2):
> virtio-scsi: create VirtIOSCSICommon
> vhost-scsi: new device supporting the tcm_vhost Linux kernel module
>
> configure | 15 +++-
> hw/Makefile.objs | 5 +-
> hw/s390x/s390-virtio-bus.c | 35 +++++++
> hw/vhost-scsi.c | 241 ++++++++++++++++++++++++++++++++++++++++++++
> hw/vhost-scsi.h | 66 ++++++++++++
> hw/vhost.c | 2 +-
> hw/virtio-pci.c | 62 +++++++++++
> hw/virtio-scsi.c | 190 +++++++++--------------------------
> hw/virtio-scsi.h | 132 +++++++++++++++++++++++-
> include/qemu/osdep.h | 4 +
> 10 files changed, 602 insertions(+), 150 deletions(-)
> create mode 100644 hw/vhost-scsi.c
> create mode 100644 hw/vhost-scsi.h
>
> --
> 1.7.2.5
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-27 21:59 ` [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module Nicholas A. Bellinger
@ 2013-03-27 22:28 ` Michael S. Tsirkin
2013-03-27 22:45 ` Nicholas A. Bellinger
0 siblings, 1 reply; 16+ messages in thread
From: Michael S. Tsirkin @ 2013-03-27 22:28 UTC (permalink / raw)
To: Nicholas A. Bellinger
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Wed, Mar 27, 2013 at 09:59:45PM +0000, Nicholas A. Bellinger wrote:
> From: Paolo Bonzini <pbonzini@redhat.com>
>
> The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
> The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
> available from the QEMU command-line. Instead, I hardcode it to zero.
>
> Changes in V4:
> - Set event_idx=off by default (nab, thanks asias)
Why? What's going on here?
> - Disable hotplug feature bit for v3.9 tcm_vhost kernel code, need to
> re-enable in v3.10 (nab)
Userspace needs to support detecting host features
at runtime, based on dev->features.
In particular this applies to event_idx, but also hotplug.
See vhost_net_get_features.
> - Update to latest qemu.git/master HEAD
>
> Changes in V3:
> - Drop ioeventfd vhost_scsi_properties (asias, thanks stefanha)
> - Add CONFIG_VHOST_SCSI (asias, thanks stefanha)
> - Add hotplug feature bit
>
> Changes in V2:
> - Add backend guest masking support (nab)
> - Bump ABI_VERSION to 1 (nab)
> - Set up set_guest_notifiers (asias)
> - Set up vs->dev.vq_index (asias)
> - Drop vs->vs.vdev.{set,clear}_vhost_endpoint (asias)
> - Drop VIRTIO_CONFIG_S_DRIVER check in vhost_scsi_set_status (asias)
>
> Howto:
> Use the latest seabios, at least commit b44a7be17b
> git clone git://git.seabios.org/seabios.git
> make
> cp out/bios.bin /usr/share/qemu/bios.bin
> qemu -device vhost-scsi-pci,wwpn=naa.6001405bd4e8476d,event_idx=off
> ...
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> Signed-off-by: Asias He <asias@redhat.com>
> Cc: Michael S. Tsirkin <mst@redhat.com>
> Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
> ---
> configure | 15 +++-
> hw/Makefile.objs | 5 +-
> hw/s390x/s390-virtio-bus.c | 35 +++++++
> hw/vhost-scsi.c | 241 ++++++++++++++++++++++++++++++++++++++++++++
> hw/vhost-scsi.h | 66 ++++++++++++
> hw/virtio-pci.c | 62 +++++++++++
> hw/virtio-scsi.h | 2 +
> 7 files changed, 424 insertions(+), 2 deletions(-)
> create mode 100644 hw/vhost-scsi.c
> create mode 100644 hw/vhost-scsi.h
>
> diff --git a/configure b/configure
> index f2af714..6b67e35 100755
> --- a/configure
> +++ b/configure
> @@ -169,6 +169,7 @@ libattr=""
> xfs=""
>
> vhost_net="no"
> +vhost_scsi="no"
> kvm="no"
> gprof="no"
> debug_tcg="no"
> @@ -532,6 +533,7 @@ Haiku)
> usb="linux"
> kvm="yes"
> vhost_net="yes"
> + vhost_scsi="yes"
> if [ "$cpu" = "i386" -o "$cpu" = "x86_64" ] ; then
> audio_possible_drivers="$audio_possible_drivers fmod"
> fi
> @@ -858,6 +860,10 @@ for opt do
> ;;
> --enable-vhost-net) vhost_net="yes"
> ;;
> + --disable-vhost-scsi) vhost_scsi="no"
> + ;;
> + --enable-vhost-scsi) vhost_scsi="yes"
> + ;;
> --disable-glx) glx="no"
> ;;
> --enable-glx) glx="yes"
> @@ -3078,7 +3084,7 @@ fi
> # __sync_fetch_and_and requires at least -march=i486. Many toolchains
> # use i686 as default anyway, but for those that don't, an explicit
> # specification is necessary
> -if test "$vhost_net" = "yes" && test "$cpu" = "i386"; then
> +if (test "$vhost_net" = "yes" -o "$vhost_scsi" = "yes") && test "$cpu" = "i386"; then
> cat > $TMPC << EOF
> static int sfaa(int *ptr)
> {
> @@ -3424,6 +3430,7 @@ echo "sigev_thread_id $sigev_thread_id"
> echo "uuid support $uuid"
> echo "libcap-ng support $cap_ng"
> echo "vhost-net support $vhost_net"
> +echo "vhost-scsi support $vhost_scsi"
> echo "Trace backend $trace_backend"
> echo "Trace output file $trace_file-<pid>"
> echo "spice support $spice ($spice_protocol_version/$spice_server_version)"
> @@ -3697,6 +3704,9 @@ fi
> if test "$virtfs" = "yes" ; then
> echo "CONFIG_VIRTFS=y" >> $config_host_mak
> fi
> +if test "$vhost_scsi" = "yes" ; then
> + echo "CONFIG_VHOST_SCSI=y" >> $config_host_mak
> +fi
> if test "$blobs" = "yes" ; then
> echo "INSTALL_BLOBS=yes" >> $config_host_mak
> fi
> @@ -4175,6 +4185,9 @@ case "$target_arch2" in
> if test "$vhost_net" = "yes" ; then
> echo "CONFIG_VHOST_NET=y" >> $config_target_mak
> fi
> + if test "$vhost_scsi" = "yes" ; then
> + echo "CONFIG_VHOST_SCSI=y" >> $config_target_mak
> + fi
> fi
> esac
> case "$target_arch2" in
> diff --git a/hw/Makefile.objs b/hw/Makefile.objs
> index d0b2ecb..0a38eb5 100644
> --- a/hw/Makefile.objs
> +++ b/hw/Makefile.objs
> @@ -205,8 +205,11 @@ common-obj-$(CONFIG_XEN_BACKEND) += xen_console.o xenfb.o xen_disk.o xen_nic.o
> obj-$(CONFIG_VIRTIO) += dataplane/
> obj-$(CONFIG_VIRTIO) += virtio.o virtio-blk.o virtio-balloon.o virtio-net.o
> obj-$(CONFIG_VIRTIO) += virtio-serial-bus.o virtio-scsi.o
> +ifeq ($(CONFIG_VIRTIO), y)
> +obj-$(CONFIG_LINUX) += vhost-scsi.o
> +endif
> obj-$(CONFIG_SOFTMMU) += vhost_net.o
> -obj-$(CONFIG_VHOST_NET) += vhost.o
> +obj-$(CONFIG_LINUX) += vhost.o
> obj-$(CONFIG_REALLY_VIRTFS) += 9pfs/
> obj-$(CONFIG_VGA) += vga.o
>
> diff --git a/hw/s390x/s390-virtio-bus.c b/hw/s390x/s390-virtio-bus.c
> index c5d5456..73d93fd 100644
> --- a/hw/s390x/s390-virtio-bus.c
> +++ b/hw/s390x/s390-virtio-bus.c
> @@ -28,6 +28,8 @@
> #include "hw/virtio-rng.h"
> #include "hw/virtio-serial.h"
> #include "hw/virtio-net.h"
> +#include "hw/virtio-scsi.h"
> +#include "hw/vhost-scsi.h"
> #include "hw/sysbus.h"
> #include "sysemu/kvm.h"
>
> @@ -214,6 +216,18 @@ static int s390_virtio_scsi_init(VirtIOS390Device *dev)
> return s390_virtio_device_init(dev, vdev);
> }
>
> +static int s390_vhost_scsi_init(VirtIOS390Device *dev)
> +{
> + VirtIODevice *vdev;
> +
> + vdev = vhost_scsi_init((DeviceState *)dev, &dev->scsi);
> + if (!vdev) {
> + return -1;
> + }
> +
> + return s390_virtio_device_init(dev, vdev);
> +}
> +
> static int s390_virtio_rng_init(VirtIOS390Device *dev)
> {
> VirtIODevice *vdev;
> @@ -537,6 +551,27 @@ static const TypeInfo virtio_s390_device_info = {
> .abstract = true,
> };
>
> +static Property s390_vhost_scsi_properties[] = {
> + DEFINE_VIRTIO_SCSI_PROPERTIES(VirtIOS390Device, host_features, scsi),
> + DEFINE_PROP_END_OF_LIST(),
> +};
> +
> +static void s390_vhost_scsi_class_init(ObjectClass *klass, void *data)
> +{
> + DeviceClass *dc = DEVICE_CLASS(klass);
> + VirtIOS390DeviceClass *k = VIRTIO_S390_DEVICE_CLASS(klass);
> +
> + k->init = s390_vhost_scsi_init;
> + dc->props = s390_vhost_scsi_properties;
> +}
> +
> +static const TypeInfo s390_vhost_scsi = {
> + .name = "vhost-scsi-s390",
> + .parent = TYPE_VIRTIO_S390_DEVICE,
> + .instance_size = sizeof(VirtIOS390Device),
> + .class_init = s390_vhost_scsi_class_init,
> +};
> +
> static Property s390_virtio_scsi_properties[] = {
> DEFINE_VIRTIO_SCSI_PROPERTIES(VirtIOS390Device, host_features, scsi),
> DEFINE_PROP_END_OF_LIST(),
> diff --git a/hw/vhost-scsi.c b/hw/vhost-scsi.c
> new file mode 100644
> index 0000000..70e42fc
> --- /dev/null
> +++ b/hw/vhost-scsi.c
> @@ -0,0 +1,241 @@
> +/*
> + * vhost_scsi host device
> + *
> + * Copyright IBM, Corp. 2011
> + *
> + * Authors:
> + * Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
> + *
> + * Changes for QEMU mainline + tcm_vhost kernel upstream:
> + * Nicholas Bellinger <nab@risingtidesystems.com>
> + *
> + * This work is licensed under the terms of the GNU LGPL, version 2 or later.
> + * See the COPYING.LIB file in the top-level directory.
> + *
> + */
> +
> +#include <sys/ioctl.h>
> +#include "config.h"
> +#include "qemu/queue.h"
> +#include "monitor/monitor.h"
> +#include "migration/migration.h"
> +#include "vhost-scsi.h"
> +#include "vhost.h"
> +#include "virtio-scsi.h"
> +
> +typedef struct VHostSCSI {
> + VirtIOSCSICommon vs;
> +
> + Error *migration_blocker;
> +
> + struct vhost_dev dev;
> +} VHostSCSI;
> +
> +static int vhost_scsi_set_endpoint(VirtIODevice *vdev)
> +{
> + VHostSCSI *vs = (VHostSCSI *)vdev;
> + struct vhost_scsi_target backend;
> + int ret;
> +
> + memset(&backend, 0, sizeof(backend));
> + pstrcpy(backend.vhost_wwpn, sizeof(backend.vhost_wwpn), vs->vs.conf->wwpn);
> + ret = ioctl(vs->dev.control, VHOST_SCSI_SET_ENDPOINT, &backend);
> + if (ret < 0) {
> + return -errno;
> + }
> + return 0;
> +}
> +
> +static void vhost_scsi_clear_endpoint(VirtIODevice *vdev)
> +{
> + VHostSCSI *vs = (VHostSCSI *)vdev;
> + struct vhost_scsi_target backend;
> +
> + memset(&backend, 0, sizeof(backend));
> + pstrcpy(backend.vhost_wwpn, sizeof(backend.vhost_wwpn), vs->vs.conf->wwpn);
> + ioctl(vs->dev.control, VHOST_SCSI_CLEAR_ENDPOINT, &backend);
> +}
> +
> +static int vhost_scsi_start(VHostSCSI *vs, VirtIODevice *vdev)
> +{
> + int ret, abi_version;
> +
> + if (!vdev->binding->set_guest_notifiers) {
> + error_report("binding does not support guest notifiers");
> + return -ENOSYS;
> + }
> +
> + ret = ioctl(vs->dev.control, VHOST_SCSI_GET_ABI_VERSION, &abi_version);
> + if (ret < 0) {
> + return -errno;
> + }
> + if (abi_version > VHOST_SCSI_ABI_VERSION) {
> + error_report("vhost-scsi: The running tcm_vhost kernel abi_version:"
> + " %d is greater than vhost_scsi userspace supports: %d, please"
> + " upgrade your version of QEMU\n", abi_version,
> + VHOST_SCSI_ABI_VERSION);
> + return -ENOSYS;
> + }
> +
> + ret = vhost_dev_enable_notifiers(&vs->dev, vdev);
> + if (ret < 0) {
> + return ret;
> + }
> +
> + ret = vhost_dev_start(&vs->dev, vdev);
> + if (ret < 0) {
> + error_report("Error start vhost dev");
> + goto err_notifiers;
> + }
> +
> + ret = vhost_scsi_set_endpoint(vdev);
> + if (ret < 0) {
> + error_report("Error set vhost-scsi endpoint");
> + goto err_vhost_stop;
> + }
> +
> + ret = vdev->binding->set_guest_notifiers(vdev->binding_opaque, vs->dev.nvqs, true);
> + if (ret < 0) {
> + error_report("Error binding guest notifier");
> + goto err_endpoint;
> + }
> + return ret;
> +
> +err_endpoint:
> + vhost_scsi_clear_endpoint(vdev);
> +err_vhost_stop:
> + vhost_dev_stop(&vs->dev, vdev);
> +err_notifiers:
> + vhost_dev_disable_notifiers(&vs->dev, vdev);
> + return ret;
> +}
> +
> +static void vhost_scsi_stop(VHostSCSI *vs, VirtIODevice *vdev)
> +{
> + int ret = 0;
> +
> + if (!vdev->binding->set_guest_notifiers) {
> + ret = vdev->binding->set_guest_notifiers(vdev->binding_opaque,
> + vs->dev.nvqs, false);
> + if (ret < 0) {
> + error_report("vhost guest notifier cleanup failed: %d\n", ret);
> + }
> + }
> + assert(ret >= 0);
> +
> + vhost_scsi_clear_endpoint(vdev);
> + vhost_dev_stop(&vs->dev, vdev);
> + vhost_dev_disable_notifiers(&vs->dev, vdev);
> +}
> +
> +static void vhost_scsi_set_config(VirtIODevice *vdev,
> + const uint8_t *config)
> +{
> + VirtIOSCSIConfig *scsiconf = (VirtIOSCSIConfig *)config;
> + VHostSCSI *vs = (VHostSCSI *)vdev;
> +
> + if ((uint32_t) ldl_raw(&scsiconf->sense_size) != vs->vs.sense_size ||
> + (uint32_t) ldl_raw(&scsiconf->cdb_size) != vs->vs.cdb_size) {
> + error_report("vhost-scsi does not support changing the sense data and CDB sizes");
> + exit(1);
> + }
> +}
> +
> +static void vhost_scsi_set_status(VirtIODevice *vdev, uint8_t val)
> +{
> + VHostSCSI *vs = (VHostSCSI *)vdev;
> + bool start = (val & VIRTIO_CONFIG_S_DRIVER_OK);
> +
> + if (vs->dev.started == start) {
> + return;
> + }
> +
> + if (start) {
> + int ret;
> +
> + ret = vhost_scsi_start(vs, vdev);
> + if (ret < 0) {
> + error_report("virtio-scsi: unable to start vhost: %s\n",
> + strerror(-ret));
> +
> + /* There is no userspace virtio-scsi fallback so exit */
> + exit(1);
> + }
> + } else {
> + vhost_scsi_stop(vs, vdev);
> + }
> +}
> +
> +static void vhost_scsi_guest_notifier_mask(VirtIODevice *vdev, int idx,
> + bool mask)
> +{
> + VHostSCSI *vs = (VHostSCSI *)vdev;
> +
> + vhost_virtqueue_mask(&vs->dev, vdev, idx, mask);
> +}
> +
> +static bool vhost_scsi_guest_notifier_pending(VirtIODevice *vdev, int idx)
> +{
> + VHostSCSI *vs = (VHostSCSI *)vdev;
> +
> + return vhost_virtqueue_pending(&vs->dev, idx);
> +}
> +
> +VirtIODevice *vhost_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf)
> +{
> + VHostSCSI *vs;
> + int vhostfd = -1;
> + int ret;
> +
> + if (!proxyconf->wwpn) {
> + error_report("vhost-scsi: missing wwpn\n");
> + return NULL;
> + }
> +
> + if (proxyconf->vhostfd) {
> + vhostfd = monitor_handle_fd_param(cur_mon, proxyconf->vhostfd);
> + if (vhostfd == -1) {
> + error_report("vhost-scsi: unable to parse vhostfd\n");
> + return NULL;
> + }
> + }
> +
> + vs = (VHostSCSI *)virtio_scsi_init_common(dev, proxyconf,
> + sizeof(VHostSCSI));
> +
> + vs->vs.vdev.set_config = vhost_scsi_set_config;
> + vs->vs.vdev.set_status = vhost_scsi_set_status;
> + vs->vs.vdev.guest_notifier_mask = vhost_scsi_guest_notifier_mask;
> + vs->vs.vdev.guest_notifier_pending = vhost_scsi_guest_notifier_pending;
> +
> + vs->dev.nvqs = VHOST_SCSI_VQ_NUM_FIXED + vs->vs.conf->num_queues;
> + vs->dev.vqs = g_new(struct vhost_virtqueue, vs->dev.nvqs);
> + vs->dev.vq_index = 0;
> +
> + ret = vhost_dev_init(&vs->dev, vhostfd, "/dev/vhost-scsi", true);
> + if (ret < 0) {
> + error_report("vhost-scsi: vhost initialization failed: %s\n",
> + strerror(-ret));
> + return NULL;
> + }
> + vs->dev.backend_features = 0;
> + vs->dev.acked_features = 0;
> +
> + error_setg(&vs->migration_blocker,
> + "vhost-scsi does not support migration");
> + migrate_add_blocker(vs->migration_blocker);
> +
> + return &vs->vs.vdev;
> +}
> +
> +void vhost_scsi_exit(VirtIODevice *vdev)
> +{
> + VHostSCSI *vs = (VHostSCSI *)vdev;
> + migrate_del_blocker(vs->migration_blocker);
> + error_free(vs->migration_blocker);
> +
> + /* This will stop vhost backend. */
> + vhost_scsi_set_status(vdev, 0);
> + g_free(vs->dev.vqs);
> + virtio_cleanup(vdev);
> +}
> diff --git a/hw/vhost-scsi.h b/hw/vhost-scsi.h
> new file mode 100644
> index 0000000..b01f012
> --- /dev/null
> +++ b/hw/vhost-scsi.h
> @@ -0,0 +1,66 @@
> +/*
> + * vhost_scsi host device
> + *
> + * Copyright IBM, Corp. 2011
> + *
> + * Authors:
> + * Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
> + *
> + * This work is licensed under the terms of the GNU LGPL, version 2 or later.
> + * See the COPYING.LIB file in the top-level directory.
> + *
> + */
> +
> +#ifndef VHOST_SCSI_H
> +#define VHOST_SCSI_H
> +
> +#include "qemu-common.h"
> +#include "qdev.h"
> +#include "virtio-scsi.h"
> +
> +/*
> + * Used by QEMU userspace to ensure a consistent vhost-scsi ABI.
> + *
> + * ABI Rev 0: July 2012 version starting point for v3.6-rc merge candidate +
> + * RFC-v2 vhost-scsi userspace. Add GET_ABI_VERSION ioctl usage
> + * ABI Rev 1: January 2013. Ignore vhost_tpgt filed in struct vhost_scsi_target.
> + * All the targets under vhost_wwpn can be seen and used by guest.
> + */
> +
> +#define VHOST_SCSI_ABI_VERSION 1
> +
> +/* TODO #include <linux/vhost.h> properly */
> +/* For VHOST_SCSI_SET_ENDPOINT/VHOST_SCSI_CLEAR_ENDPOINT ioctl */
> +struct vhost_scsi_target {
> + int abi_version;
> + char vhost_wwpn[224];
> + unsigned short vhost_tpgt;
> + unsigned short reserved;
> +};
> +
> +enum vhost_scsi_vq_list {
> + VHOST_SCSI_VQ_CONTROL = 0,
> + VHOST_SCSI_VQ_EVENT = 1,
> + VHOST_SCSI_VQ_NUM_FIXED = 2,
> +};
> +
> +#define VHOST_VIRTIO 0xAF
> +#define VHOST_SCSI_SET_ENDPOINT _IOW(VHOST_VIRTIO, 0x40, struct vhost_scsi_target)
> +#define VHOST_SCSI_CLEAR_ENDPOINT _IOW(VHOST_VIRTIO, 0x41, struct vhost_scsi_target)
> +#define VHOST_SCSI_GET_ABI_VERSION _IOW(VHOST_VIRTIO, 0x42, int)
> +
> +#define DEFINE_VHOST_SCSI_PROPERTIES(_state, _features_field, _conf_field) \
> + DEFINE_PROP_BIT("indirect_desc", _state, _features_field, VIRTIO_RING_F_INDIRECT_DESC, true), \
> + DEFINE_PROP_BIT("event_idx", _state, _features_field, VIRTIO_RING_F_EVENT_IDX, false), \
Maybe this works around some bug?
Needs a comment with motivation.
Ideally we won't do this, use DEFINE_VIRTIO_COMMON_FEATURES
so that adding properties to all devices will be easier
in the future.
> + DEFINE_PROP_STRING("vhostfd", _state, _conf_field.vhostfd), \
> + DEFINE_PROP_STRING("wwpn", _state, _conf_field.wwpn), \
> + DEFINE_PROP_UINT32("num_queues", _state, _conf_field.num_queues, 1), \
> + DEFINE_PROP_UINT32("max_sectors", _state, _conf_field.max_sectors, 0xFFFF), \
> + DEFINE_PROP_UINT32("cmd_per_lun", _state, _conf_field.cmd_per_lun, 128), \
> + DEFINE_PROP_BIT("hotplug", _state, _features_field, VIRTIO_SCSI_F_HOTPLUG, true)
Seems unused?
> +
> +VirtIODevice *vhost_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf);
> +void vhost_scsi_exit(VirtIODevice *vdev);
> +
> +
> +#endif
> diff --git a/hw/virtio-pci.c b/hw/virtio-pci.c
> index f3ece78..0d67b84 100644
> --- a/hw/virtio-pci.c
> +++ b/hw/virtio-pci.c
> @@ -22,6 +22,7 @@
> #include "hw/virtio-net.h"
> #include "hw/virtio-serial.h"
> #include "hw/virtio-scsi.h"
> +#include "hw/vhost-scsi.h"
> #include "hw/pci/pci.h"
> #include "qemu/error-report.h"
> #include "hw/pci/msi.h"
> @@ -1247,6 +1248,64 @@ static const TypeInfo virtio_scsi_info = {
> .class_init = virtio_scsi_class_init,
> };
>
> +#ifdef CONFIG_VHOST_SCSI
> +static int vhost_scsi_init_pci(PCIDevice *pci_dev)
> +{
> + VirtIOPCIProxy *proxy = DO_UPCAST(VirtIOPCIProxy, pci_dev, pci_dev);
> + VirtIODevice *vdev;
> +
> + vdev = vhost_scsi_init(&pci_dev->qdev, &proxy->scsi);
> + if (!vdev) {
> + return -EINVAL;
> + }
> +
> + vdev->nvectors = proxy->nvectors == DEV_NVECTORS_UNSPECIFIED
> + ? proxy->scsi.num_queues + 3
> + : proxy->nvectors;
> + virtio_init_pci(proxy, vdev);
> +
> + /* make the actual value visible */
> + proxy->nvectors = vdev->nvectors;
> + return 0;
> +}
> +
> +static void vhost_scsi_exit_pci(PCIDevice *pci_dev)
> +{
> + VirtIOPCIProxy *proxy = DO_UPCAST(VirtIOPCIProxy, pci_dev, pci_dev);
> +
> + vhost_scsi_exit(proxy->vdev);
> + virtio_exit_pci(pci_dev);
> +}
> +
> +static Property vhost_scsi_properties[] = {
> + DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors, DEV_NVECTORS_UNSPECIFIED),
> + DEFINE_VHOST_SCSI_PROPERTIES(VirtIOPCIProxy, host_features, scsi),
> + DEFINE_PROP_END_OF_LIST(),
> +};
> +
> +static void vhost_scsi_class_init(ObjectClass *klass, void *data)
> +{
> + DeviceClass *dc = DEVICE_CLASS(klass);
> + PCIDeviceClass *k = PCI_DEVICE_CLASS(klass);
> +
> + k->init = vhost_scsi_init_pci;
> + k->exit = vhost_scsi_exit_pci;
> + k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
> + k->device_id = PCI_DEVICE_ID_VIRTIO_SCSI;
> + k->revision = 0x00;
> + k->class_id = PCI_CLASS_STORAGE_SCSI;
> + dc->reset = virtio_pci_reset;
> + dc->props = vhost_scsi_properties;
> +}
> +
> +static const TypeInfo vhost_scsi_info = {
> + .name = "vhost-scsi-pci",
> + .parent = TYPE_PCI_DEVICE,
> + .instance_size = sizeof(VirtIOPCIProxy),
> + .class_init = vhost_scsi_class_init,
> +};
> +#endif
> +
> #ifdef CONFIG_VIRTFS
> static int virtio_9p_init_pci(PCIDevice *pci_dev)
> {
> @@ -1495,6 +1554,9 @@ static void virtio_pci_register_types(void)
> type_register_static(&virtio_serial_info);
> type_register_static(&virtio_balloon_info);
> type_register_static(&virtio_scsi_info);
> +#ifdef CONFIG_VHOST_SCSI
> + type_register_static(&vhost_scsi_info);
> +#endif
> type_register_static(&virtio_rng_info);
> type_register_static(&virtio_pci_bus_info);
> type_register_static(&virtio_pci_info);
> diff --git a/hw/virtio-scsi.h b/hw/virtio-scsi.h
> index 2bcf606..356bf38 100644
> --- a/hw/virtio-scsi.h
> +++ b/hw/virtio-scsi.h
> @@ -142,6 +142,8 @@ struct VirtIOSCSIConf {
> uint32_t num_queues;
> uint32_t max_sectors;
> uint32_t cmd_per_lun;
> + char *vhostfd;
> + char *wwpn;
> };
>
> typedef struct VirtIOSCSICommon {
> --
> 1.7.2.5
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-27 22:28 ` Michael S. Tsirkin
@ 2013-03-27 22:45 ` Nicholas A. Bellinger
2013-03-27 22:50 ` Michael S. Tsirkin
0 siblings, 1 reply; 16+ messages in thread
From: Nicholas A. Bellinger @ 2013-03-27 22:45 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Thu, 2013-03-28 at 00:28 +0200, Michael S. Tsirkin wrote:
> On Wed, Mar 27, 2013 at 09:59:45PM +0000, Nicholas A. Bellinger wrote:
> > From: Paolo Bonzini <pbonzini@redhat.com>
> >
> > The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
> > The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
> > available from the QEMU command-line. Instead, I hardcode it to zero.
> >
> > Changes in V4:
> > - Set event_idx=off by default (nab, thanks asias)
>
> Why? What's going on here?
>
Not disabling event_idx by default, or disabling from the command line
ends up resulting in ->handle_kick() not getting called for subsequent
commands..
I spent some time trying to track this down recently with no luck, and
AFAICT it's always been required in order for vhost-scsi to function.
> > - Disable hotplug feature bit for v3.9 tcm_vhost kernel code, need to
> > re-enable in v3.10 (nab)
>
> Userspace needs to support detecting host features
> at runtime, based on dev->features.
> In particular this applies to event_idx, but also hotplug.
> See vhost_net_get_features.
<nod>, will get this fixed up.
>
> > - Update to latest qemu.git/master HEAD
> >
> > Changes in V3:
> > - Drop ioeventfd vhost_scsi_properties (asias, thanks stefanha)
> > - Add CONFIG_VHOST_SCSI (asias, thanks stefanha)
> > - Add hotplug feature bit
> >
> > Changes in V2:
> > - Add backend guest masking support (nab)
> > - Bump ABI_VERSION to 1 (nab)
> > - Set up set_guest_notifiers (asias)
> > - Set up vs->dev.vq_index (asias)
> > - Drop vs->vs.vdev.{set,clear}_vhost_endpoint (asias)
> > - Drop VIRTIO_CONFIG_S_DRIVER check in vhost_scsi_set_status (asias)
> >
> > Howto:
> > Use the latest seabios, at least commit b44a7be17b
> > git clone git://git.seabios.org/seabios.git
> > make
> > cp out/bios.bin /usr/share/qemu/bios.bin
> > qemu -device vhost-scsi-pci,wwpn=naa.6001405bd4e8476d,event_idx=off
> > ...
> >
> > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> > Signed-off-by: Asias He <asias@redhat.com>
> > Cc: Michael S. Tsirkin <mst@redhat.com>
> > Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
> > ---
> > configure | 15 +++-
> > hw/Makefile.objs | 5 +-
> > hw/s390x/s390-virtio-bus.c | 35 +++++++
> > hw/vhost-scsi.c | 241 ++++++++++++++++++++++++++++++++++++++++++++
> > hw/vhost-scsi.h | 66 ++++++++++++
> > hw/virtio-pci.c | 62 +++++++++++
> > hw/virtio-scsi.h | 2 +
> > 7 files changed, 424 insertions(+), 2 deletions(-)
> > create mode 100644 hw/vhost-scsi.c
> > create mode 100644 hw/vhost-scsi.h
> >
<SNIP>
> > diff --git a/hw/vhost-scsi.h b/hw/vhost-scsi.h
> > new file mode 100644
> > index 0000000..b01f012
> > --- /dev/null
> > +++ b/hw/vhost-scsi.h
> > @@ -0,0 +1,66 @@
> > +/*
> > + * vhost_scsi host device
> > + *
> > + * Copyright IBM, Corp. 2011
> > + *
> > + * Authors:
> > + * Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
> > + *
> > + * This work is licensed under the terms of the GNU LGPL, version 2 or later.
> > + * See the COPYING.LIB file in the top-level directory.
> > + *
> > + */
> > +
> > +#ifndef VHOST_SCSI_H
> > +#define VHOST_SCSI_H
> > +
> > +#include "qemu-common.h"
> > +#include "qdev.h"
> > +#include "virtio-scsi.h"
> > +
> > +/*
> > + * Used by QEMU userspace to ensure a consistent vhost-scsi ABI.
> > + *
> > + * ABI Rev 0: July 2012 version starting point for v3.6-rc merge candidate +
> > + * RFC-v2 vhost-scsi userspace. Add GET_ABI_VERSION ioctl usage
> > + * ABI Rev 1: January 2013. Ignore vhost_tpgt filed in struct vhost_scsi_target.
> > + * All the targets under vhost_wwpn can be seen and used by guest.
> > + */
> > +
> > +#define VHOST_SCSI_ABI_VERSION 1
> > +
> > +/* TODO #include <linux/vhost.h> properly */
> > +/* For VHOST_SCSI_SET_ENDPOINT/VHOST_SCSI_CLEAR_ENDPOINT ioctl */
> > +struct vhost_scsi_target {
> > + int abi_version;
> > + char vhost_wwpn[224];
> > + unsigned short vhost_tpgt;
> > + unsigned short reserved;
> > +};
> > +
> > +enum vhost_scsi_vq_list {
> > + VHOST_SCSI_VQ_CONTROL = 0,
> > + VHOST_SCSI_VQ_EVENT = 1,
> > + VHOST_SCSI_VQ_NUM_FIXED = 2,
> > +};
> > +
> > +#define VHOST_VIRTIO 0xAF
> > +#define VHOST_SCSI_SET_ENDPOINT _IOW(VHOST_VIRTIO, 0x40, struct vhost_scsi_target)
> > +#define VHOST_SCSI_CLEAR_ENDPOINT _IOW(VHOST_VIRTIO, 0x41, struct vhost_scsi_target)
> > +#define VHOST_SCSI_GET_ABI_VERSION _IOW(VHOST_VIRTIO, 0x42, int)
> > +
> > +#define DEFINE_VHOST_SCSI_PROPERTIES(_state, _features_field, _conf_field) \
> > + DEFINE_PROP_BIT("indirect_desc", _state, _features_field, VIRTIO_RING_F_INDIRECT_DESC, true), \
> > + DEFINE_PROP_BIT("event_idx", _state, _features_field, VIRTIO_RING_F_EVENT_IDX, false), \
>
> Maybe this works around some bug?
> Needs a comment with motivation.
>
> Ideally we won't do this, use DEFINE_VIRTIO_COMMON_FEATURES
> so that adding properties to all devices will be easier
> in the future.
AFAICT vhost-scsi has never worked without event_idx=off, so disabling
it by default here made sense. ;)
>
> > + DEFINE_PROP_STRING("vhostfd", _state, _conf_field.vhostfd), \
> > + DEFINE_PROP_STRING("wwpn", _state, _conf_field.wwpn), \
> > + DEFINE_PROP_UINT32("num_queues", _state, _conf_field.num_queues, 1), \
> > + DEFINE_PROP_UINT32("max_sectors", _state, _conf_field.max_sectors, 0xFFFF), \
> > + DEFINE_PROP_UINT32("cmd_per_lun", _state, _conf_field.cmd_per_lun, 128), \
> > + DEFINE_PROP_BIT("hotplug", _state, _features_field, VIRTIO_SCSI_F_HOTPLUG, true)
>
> Seems unused?
Looks like left-over debug code from Asias..
Dropping this now..
>
> > +
> > +VirtIODevice *vhost_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf);
> > +void vhost_scsi_exit(VirtIODevice *vdev);
> > +
> > +
> > +#endif
> > diff --git a/hw/virtio-pci.c b/hw/virtio-pci.c
> > index f3ece78..0d67b84 100644
> > --- a/hw/virtio-pci.c
> > +++ b/hw/virtio-pci.c
> > @@ -22,6 +22,7 @@
> > #include "hw/virtio-net.h"
> > #include "hw/virtio-serial.h"
> > #include "hw/virtio-scsi.h"
> > +#include "hw/vhost-scsi.h"
> > #include "hw/pci/pci.h"
> > #include "qemu/error-report.h"
> > #include "hw/pci/msi.h"
> > @@ -1247,6 +1248,64 @@ static const TypeInfo virtio_scsi_info = {
> > .class_init = virtio_scsi_class_init,
> > };
> >
> > +#ifdef CONFIG_VHOST_SCSI
> > +static int vhost_scsi_init_pci(PCIDevice *pci_dev)
> > +{
> > + VirtIOPCIProxy *proxy = DO_UPCAST(VirtIOPCIProxy, pci_dev, pci_dev);
> > + VirtIODevice *vdev;
> > +
> > + vdev = vhost_scsi_init(&pci_dev->qdev, &proxy->scsi);
> > + if (!vdev) {
> > + return -EINVAL;
> > + }
> > +
> > + vdev->nvectors = proxy->nvectors == DEV_NVECTORS_UNSPECIFIED
> > + ? proxy->scsi.num_queues + 3
> > + : proxy->nvectors;
> > + virtio_init_pci(proxy, vdev);
> > +
> > + /* make the actual value visible */
> > + proxy->nvectors = vdev->nvectors;
> > + return 0;
> > +}
> > +
> > +static void vhost_scsi_exit_pci(PCIDevice *pci_dev)
> > +{
> > + VirtIOPCIProxy *proxy = DO_UPCAST(VirtIOPCIProxy, pci_dev, pci_dev);
> > +
> > + vhost_scsi_exit(proxy->vdev);
> > + virtio_exit_pci(pci_dev);
> > +}
> > +
> > +static Property vhost_scsi_properties[] = {
> > + DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors, DEV_NVECTORS_UNSPECIFIED),
> > + DEFINE_VHOST_SCSI_PROPERTIES(VirtIOPCIProxy, host_features, scsi),
> > + DEFINE_PROP_END_OF_LIST(),
> > +};
> > +
> > +static void vhost_scsi_class_init(ObjectClass *klass, void *data)
> > +{
> > + DeviceClass *dc = DEVICE_CLASS(klass);
> > + PCIDeviceClass *k = PCI_DEVICE_CLASS(klass);
> > +
> > + k->init = vhost_scsi_init_pci;
> > + k->exit = vhost_scsi_exit_pci;
> > + k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
> > + k->device_id = PCI_DEVICE_ID_VIRTIO_SCSI;
> > + k->revision = 0x00;
> > + k->class_id = PCI_CLASS_STORAGE_SCSI;
> > + dc->reset = virtio_pci_reset;
> > + dc->props = vhost_scsi_properties;
> > +}
> > +
> > +static const TypeInfo vhost_scsi_info = {
> > + .name = "vhost-scsi-pci",
> > + .parent = TYPE_PCI_DEVICE,
> > + .instance_size = sizeof(VirtIOPCIProxy),
> > + .class_init = vhost_scsi_class_init,
> > +};
> > +#endif
> > +
> > #ifdef CONFIG_VIRTFS
> > static int virtio_9p_init_pci(PCIDevice *pci_dev)
> > {
> > @@ -1495,6 +1554,9 @@ static void virtio_pci_register_types(void)
> > type_register_static(&virtio_serial_info);
> > type_register_static(&virtio_balloon_info);
> > type_register_static(&virtio_scsi_info);
> > +#ifdef CONFIG_VHOST_SCSI
> > + type_register_static(&vhost_scsi_info);
> > +#endif
> > type_register_static(&virtio_rng_info);
> > type_register_static(&virtio_pci_bus_info);
> > type_register_static(&virtio_pci_info);
> > diff --git a/hw/virtio-scsi.h b/hw/virtio-scsi.h
> > index 2bcf606..356bf38 100644
> > --- a/hw/virtio-scsi.h
> > +++ b/hw/virtio-scsi.h
> > @@ -142,6 +142,8 @@ struct VirtIOSCSIConf {
> > uint32_t num_queues;
> > uint32_t max_sectors;
> > uint32_t cmd_per_lun;
> > + char *vhostfd;
> > + char *wwpn;
> > };
> >
> > typedef struct VirtIOSCSICommon {
> > --
> > 1.7.2.5
> --
> To unsubscribe from this list: send the line "unsubscribe target-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-27 22:45 ` Nicholas A. Bellinger
@ 2013-03-27 22:50 ` Michael S. Tsirkin
2013-03-27 22:53 ` Michael S. Tsirkin
2013-03-27 23:05 ` Nicholas A. Bellinger
0 siblings, 2 replies; 16+ messages in thread
From: Michael S. Tsirkin @ 2013-03-27 22:50 UTC (permalink / raw)
To: Nicholas A. Bellinger
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Wed, Mar 27, 2013 at 03:45:42PM -0700, Nicholas A. Bellinger wrote:
> On Thu, 2013-03-28 at 00:28 +0200, Michael S. Tsirkin wrote:
> > On Wed, Mar 27, 2013 at 09:59:45PM +0000, Nicholas A. Bellinger wrote:
> > > From: Paolo Bonzini <pbonzini@redhat.com>
> > >
> > > The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
> > > The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
> > > available from the QEMU command-line. Instead, I hardcode it to zero.
> > >
> > > Changes in V4:
> > > - Set event_idx=off by default (nab, thanks asias)
> >
> > Why? What's going on here?
> >
>
> Not disabling event_idx by default, or disabling from the command line
> ends up resulting in ->handle_kick() not getting called for subsequent
> commands..
>
> I spent some time trying to track this down recently with no luck, and
> AFAICT it's always been required in order for vhost-scsi to function.
Hmm this is a bug in kernel then. A better work-around is
to disable EVENT_IDX in kernel. Let's do it for 3.9?
Then when we fix it, things just work for userspace.
> > > - Disable hotplug feature bit for v3.9 tcm_vhost kernel code, need to
> > > re-enable in v3.10 (nab)
> >
> > Userspace needs to support detecting host features
> > at runtime, based on dev->features.
> > In particular this applies to event_idx, but also hotplug.
> > See vhost_net_get_features.
>
> <nod>, will get this fixed up.
>
> >
> > > - Update to latest qemu.git/master HEAD
> > >
> > > Changes in V3:
> > > - Drop ioeventfd vhost_scsi_properties (asias, thanks stefanha)
> > > - Add CONFIG_VHOST_SCSI (asias, thanks stefanha)
> > > - Add hotplug feature bit
> > >
> > > Changes in V2:
> > > - Add backend guest masking support (nab)
> > > - Bump ABI_VERSION to 1 (nab)
> > > - Set up set_guest_notifiers (asias)
> > > - Set up vs->dev.vq_index (asias)
> > > - Drop vs->vs.vdev.{set,clear}_vhost_endpoint (asias)
> > > - Drop VIRTIO_CONFIG_S_DRIVER check in vhost_scsi_set_status (asias)
> > >
> > > Howto:
> > > Use the latest seabios, at least commit b44a7be17b
> > > git clone git://git.seabios.org/seabios.git
> > > make
> > > cp out/bios.bin /usr/share/qemu/bios.bin
> > > qemu -device vhost-scsi-pci,wwpn=naa.6001405bd4e8476d,event_idx=off
> > > ...
> > >
> > > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> > > Signed-off-by: Asias He <asias@redhat.com>
> > > Cc: Michael S. Tsirkin <mst@redhat.com>
> > > Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
> > > ---
> > > configure | 15 +++-
> > > hw/Makefile.objs | 5 +-
> > > hw/s390x/s390-virtio-bus.c | 35 +++++++
> > > hw/vhost-scsi.c | 241 ++++++++++++++++++++++++++++++++++++++++++++
> > > hw/vhost-scsi.h | 66 ++++++++++++
> > > hw/virtio-pci.c | 62 +++++++++++
> > > hw/virtio-scsi.h | 2 +
> > > 7 files changed, 424 insertions(+), 2 deletions(-)
> > > create mode 100644 hw/vhost-scsi.c
> > > create mode 100644 hw/vhost-scsi.h
> > >
>
> <SNIP>
>
> > > diff --git a/hw/vhost-scsi.h b/hw/vhost-scsi.h
> > > new file mode 100644
> > > index 0000000..b01f012
> > > --- /dev/null
> > > +++ b/hw/vhost-scsi.h
> > > @@ -0,0 +1,66 @@
> > > +/*
> > > + * vhost_scsi host device
> > > + *
> > > + * Copyright IBM, Corp. 2011
> > > + *
> > > + * Authors:
> > > + * Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
> > > + *
> > > + * This work is licensed under the terms of the GNU LGPL, version 2 or later.
> > > + * See the COPYING.LIB file in the top-level directory.
> > > + *
> > > + */
> > > +
> > > +#ifndef VHOST_SCSI_H
> > > +#define VHOST_SCSI_H
> > > +
> > > +#include "qemu-common.h"
> > > +#include "qdev.h"
> > > +#include "virtio-scsi.h"
> > > +
> > > +/*
> > > + * Used by QEMU userspace to ensure a consistent vhost-scsi ABI.
> > > + *
> > > + * ABI Rev 0: July 2012 version starting point for v3.6-rc merge candidate +
> > > + * RFC-v2 vhost-scsi userspace. Add GET_ABI_VERSION ioctl usage
> > > + * ABI Rev 1: January 2013. Ignore vhost_tpgt filed in struct vhost_scsi_target.
> > > + * All the targets under vhost_wwpn can be seen and used by guest.
> > > + */
> > > +
> > > +#define VHOST_SCSI_ABI_VERSION 1
> > > +
> > > +/* TODO #include <linux/vhost.h> properly */
> > > +/* For VHOST_SCSI_SET_ENDPOINT/VHOST_SCSI_CLEAR_ENDPOINT ioctl */
> > > +struct vhost_scsi_target {
> > > + int abi_version;
> > > + char vhost_wwpn[224];
> > > + unsigned short vhost_tpgt;
> > > + unsigned short reserved;
> > > +};
> > > +
> > > +enum vhost_scsi_vq_list {
> > > + VHOST_SCSI_VQ_CONTROL = 0,
> > > + VHOST_SCSI_VQ_EVENT = 1,
> > > + VHOST_SCSI_VQ_NUM_FIXED = 2,
> > > +};
> > > +
> > > +#define VHOST_VIRTIO 0xAF
> > > +#define VHOST_SCSI_SET_ENDPOINT _IOW(VHOST_VIRTIO, 0x40, struct vhost_scsi_target)
> > > +#define VHOST_SCSI_CLEAR_ENDPOINT _IOW(VHOST_VIRTIO, 0x41, struct vhost_scsi_target)
> > > +#define VHOST_SCSI_GET_ABI_VERSION _IOW(VHOST_VIRTIO, 0x42, int)
> > > +
> > > +#define DEFINE_VHOST_SCSI_PROPERTIES(_state, _features_field, _conf_field) \
> > > + DEFINE_PROP_BIT("indirect_desc", _state, _features_field, VIRTIO_RING_F_INDIRECT_DESC, true), \
> > > + DEFINE_PROP_BIT("event_idx", _state, _features_field, VIRTIO_RING_F_EVENT_IDX, false), \
> >
> > Maybe this works around some bug?
> > Needs a comment with motivation.
> >
> > Ideally we won't do this, use DEFINE_VIRTIO_COMMON_FEATURES
> > so that adding properties to all devices will be easier
> > in the future.
>
> AFAICT vhost-scsi has never worked without event_idx=off, so disabling
> it by default here made sense. ;)
>
> >
> > > + DEFINE_PROP_STRING("vhostfd", _state, _conf_field.vhostfd), \
> > > + DEFINE_PROP_STRING("wwpn", _state, _conf_field.wwpn), \
> > > + DEFINE_PROP_UINT32("num_queues", _state, _conf_field.num_queues, 1), \
> > > + DEFINE_PROP_UINT32("max_sectors", _state, _conf_field.max_sectors, 0xFFFF), \
> > > + DEFINE_PROP_UINT32("cmd_per_lun", _state, _conf_field.cmd_per_lun, 128), \
> > > + DEFINE_PROP_BIT("hotplug", _state, _features_field, VIRTIO_SCSI_F_HOTPLUG, true)
> >
> > Seems unused?
>
> Looks like left-over debug code from Asias..
>
> Dropping this now..
>
> >
> > > +
> > > +VirtIODevice *vhost_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf);
> > > +void vhost_scsi_exit(VirtIODevice *vdev);
> > > +
> > > +
> > > +#endif
> > > diff --git a/hw/virtio-pci.c b/hw/virtio-pci.c
> > > index f3ece78..0d67b84 100644
> > > --- a/hw/virtio-pci.c
> > > +++ b/hw/virtio-pci.c
> > > @@ -22,6 +22,7 @@
> > > #include "hw/virtio-net.h"
> > > #include "hw/virtio-serial.h"
> > > #include "hw/virtio-scsi.h"
> > > +#include "hw/vhost-scsi.h"
> > > #include "hw/pci/pci.h"
> > > #include "qemu/error-report.h"
> > > #include "hw/pci/msi.h"
> > > @@ -1247,6 +1248,64 @@ static const TypeInfo virtio_scsi_info = {
> > > .class_init = virtio_scsi_class_init,
> > > };
> > >
> > > +#ifdef CONFIG_VHOST_SCSI
> > > +static int vhost_scsi_init_pci(PCIDevice *pci_dev)
> > > +{
> > > + VirtIOPCIProxy *proxy = DO_UPCAST(VirtIOPCIProxy, pci_dev, pci_dev);
> > > + VirtIODevice *vdev;
> > > +
> > > + vdev = vhost_scsi_init(&pci_dev->qdev, &proxy->scsi);
> > > + if (!vdev) {
> > > + return -EINVAL;
> > > + }
> > > +
> > > + vdev->nvectors = proxy->nvectors == DEV_NVECTORS_UNSPECIFIED
> > > + ? proxy->scsi.num_queues + 3
> > > + : proxy->nvectors;
> > > + virtio_init_pci(proxy, vdev);
> > > +
> > > + /* make the actual value visible */
> > > + proxy->nvectors = vdev->nvectors;
> > > + return 0;
> > > +}
> > > +
> > > +static void vhost_scsi_exit_pci(PCIDevice *pci_dev)
> > > +{
> > > + VirtIOPCIProxy *proxy = DO_UPCAST(VirtIOPCIProxy, pci_dev, pci_dev);
> > > +
> > > + vhost_scsi_exit(proxy->vdev);
> > > + virtio_exit_pci(pci_dev);
> > > +}
> > > +
> > > +static Property vhost_scsi_properties[] = {
> > > + DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors, DEV_NVECTORS_UNSPECIFIED),
> > > + DEFINE_VHOST_SCSI_PROPERTIES(VirtIOPCIProxy, host_features, scsi),
> > > + DEFINE_PROP_END_OF_LIST(),
> > > +};
> > > +
> > > +static void vhost_scsi_class_init(ObjectClass *klass, void *data)
> > > +{
> > > + DeviceClass *dc = DEVICE_CLASS(klass);
> > > + PCIDeviceClass *k = PCI_DEVICE_CLASS(klass);
> > > +
> > > + k->init = vhost_scsi_init_pci;
> > > + k->exit = vhost_scsi_exit_pci;
> > > + k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
> > > + k->device_id = PCI_DEVICE_ID_VIRTIO_SCSI;
> > > + k->revision = 0x00;
> > > + k->class_id = PCI_CLASS_STORAGE_SCSI;
> > > + dc->reset = virtio_pci_reset;
> > > + dc->props = vhost_scsi_properties;
> > > +}
> > > +
> > > +static const TypeInfo vhost_scsi_info = {
> > > + .name = "vhost-scsi-pci",
> > > + .parent = TYPE_PCI_DEVICE,
> > > + .instance_size = sizeof(VirtIOPCIProxy),
> > > + .class_init = vhost_scsi_class_init,
> > > +};
> > > +#endif
> > > +
> > > #ifdef CONFIG_VIRTFS
> > > static int virtio_9p_init_pci(PCIDevice *pci_dev)
> > > {
> > > @@ -1495,6 +1554,9 @@ static void virtio_pci_register_types(void)
> > > type_register_static(&virtio_serial_info);
> > > type_register_static(&virtio_balloon_info);
> > > type_register_static(&virtio_scsi_info);
> > > +#ifdef CONFIG_VHOST_SCSI
> > > + type_register_static(&vhost_scsi_info);
> > > +#endif
> > > type_register_static(&virtio_rng_info);
> > > type_register_static(&virtio_pci_bus_info);
> > > type_register_static(&virtio_pci_info);
> > > diff --git a/hw/virtio-scsi.h b/hw/virtio-scsi.h
> > > index 2bcf606..356bf38 100644
> > > --- a/hw/virtio-scsi.h
> > > +++ b/hw/virtio-scsi.h
> > > @@ -142,6 +142,8 @@ struct VirtIOSCSIConf {
> > > uint32_t num_queues;
> > > uint32_t max_sectors;
> > > uint32_t cmd_per_lun;
> > > + char *vhostfd;
> > > + char *wwpn;
> > > };
> > >
> > > typedef struct VirtIOSCSICommon {
> > > --
> > > 1.7.2.5
> > --
> > To unsubscribe from this list: send the line "unsubscribe target-devel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-27 22:50 ` Michael S. Tsirkin
@ 2013-03-27 22:53 ` Michael S. Tsirkin
2013-03-27 23:24 ` Nicholas A. Bellinger
2013-03-27 23:05 ` Nicholas A. Bellinger
1 sibling, 1 reply; 16+ messages in thread
From: Michael S. Tsirkin @ 2013-03-27 22:53 UTC (permalink / raw)
To: Nicholas A. Bellinger
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Thu, Mar 28, 2013 at 12:50:21AM +0200, Michael S. Tsirkin wrote:
> On Wed, Mar 27, 2013 at 03:45:42PM -0700, Nicholas A. Bellinger wrote:
> > On Thu, 2013-03-28 at 00:28 +0200, Michael S. Tsirkin wrote:
> > > On Wed, Mar 27, 2013 at 09:59:45PM +0000, Nicholas A. Bellinger wrote:
> > > > From: Paolo Bonzini <pbonzini@redhat.com>
> > > >
> > > > The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
> > > > The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
> > > > available from the QEMU command-line. Instead, I hardcode it to zero.
> > > >
> > > > Changes in V4:
> > > > - Set event_idx=off by default (nab, thanks asias)
> > >
> > > Why? What's going on here?
> > >
> >
> > Not disabling event_idx by default, or disabling from the command line
> > ends up resulting in ->handle_kick() not getting called for subsequent
> > commands..
> >
> > I spent some time trying to track this down recently with no luck, and
> > AFAICT it's always been required in order for vhost-scsi to function.
>
> Hmm this is a bug in kernel then. A better work-around is
> to disable EVENT_IDX in kernel. Let's do it for 3.9?
But before we do, can you check that SET_FEATURES is
called with this bit set if you enable event_idx?
If not that's your bug then ...
--
MST
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-27 22:50 ` Michael S. Tsirkin
2013-03-27 22:53 ` Michael S. Tsirkin
@ 2013-03-27 23:05 ` Nicholas A. Bellinger
1 sibling, 0 replies; 16+ messages in thread
From: Nicholas A. Bellinger @ 2013-03-27 23:05 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Thu, 2013-03-28 at 00:50 +0200, Michael S. Tsirkin wrote:
> On Wed, Mar 27, 2013 at 03:45:42PM -0700, Nicholas A. Bellinger wrote:
> > On Thu, 2013-03-28 at 00:28 +0200, Michael S. Tsirkin wrote:
> > > On Wed, Mar 27, 2013 at 09:59:45PM +0000, Nicholas A. Bellinger wrote:
> > > > From: Paolo Bonzini <pbonzini@redhat.com>
> > > >
> > > > The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
> > > > The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
> > > > available from the QEMU command-line. Instead, I hardcode it to zero.
> > > >
> > > > Changes in V4:
> > > > - Set event_idx=off by default (nab, thanks asias)
> > >
> > > Why? What's going on here?
> > >
> >
> > Not disabling event_idx by default, or disabling from the command line
> > ends up resulting in ->handle_kick() not getting called for subsequent
> > commands..
> >
> > I spent some time trying to track this down recently with no luck, and
> > AFAICT it's always been required in order for vhost-scsi to function.
>
> Hmm this is a bug in kernel then. A better work-around is
> to disable EVENT_IDX in kernel. Let's do it for 3.9?
> Then when we fix it, things just work for userspace.
>
Mmmm, the following patch yields the same results with the event_idx=on
default for DEFINE_VHOST_SCSI_PROPERTIES.
Is there another method you had in mind to disable EVENT_IDX..?
diff --git a/drivers/vhost/tcm_vhost.c b/drivers/vhost/tcm_vhost.c
index 0524267..757cbda 100644
--- a/drivers/vhost/tcm_vhost.c
+++ b/drivers/vhost/tcm_vhost.c
@@ -981,7 +981,7 @@ static void vhost_scsi_flush(struct vhost_scsi *vs)
static int vhost_scsi_set_features(struct vhost_scsi *vs, u64 features)
{
- if (features & ~VHOST_FEATURES)
+ if (features & ~VHOST_TCM_FEATURES)
return -EOPNOTSUPP;
mutex_lock(&vs->dev.mutex);
@@ -1027,7 +1027,7 @@ static long vhost_scsi_ioctl(struct file *f, unsigned int ioctl,
return -EFAULT;
return 0;
case VHOST_GET_FEATURES:
- features = VHOST_FEATURES;
+ features = VHOST_TCM_FEATURES;
if (copy_to_user(featurep, &features, sizeof features))
return -EFAULT;
return 0;
diff --git a/drivers/vhost/vhost.h b/drivers/vhost/vhost.h
index 17261e2..588536e 100644
--- a/drivers/vhost/vhost.h
+++ b/drivers/vhost/vhost.h
@@ -204,6 +204,9 @@ enum {
VHOST_NET_FEATURES = VHOST_FEATURES |
(1ULL << VHOST_NET_F_VIRTIO_NET_HDR) |
(1ULL << VIRTIO_NET_F_MRG_RXBUF),
+ VHOST_TCM_FEATURES = (1ULL << VIRTIO_F_NOTIFY_ON_EMPTY) |
+ (1ULL << VIRTIO_RING_F_INDIRECT_DESC) |
+ (1ULL << VHOST_F_LOG_ALL)
};
static inline int vhost_has_feature(struct vhost_dev *dev, int bit)
^ permalink raw reply related [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-27 22:53 ` Michael S. Tsirkin
@ 2013-03-27 23:24 ` Nicholas A. Bellinger
2013-03-28 0:21 ` Nicholas A. Bellinger
2013-03-28 6:19 ` Michael S. Tsirkin
0 siblings, 2 replies; 16+ messages in thread
From: Nicholas A. Bellinger @ 2013-03-27 23:24 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Thu, 2013-03-28 at 00:53 +0200, Michael S. Tsirkin wrote:
> On Thu, Mar 28, 2013 at 12:50:21AM +0200, Michael S. Tsirkin wrote:
> > On Wed, Mar 27, 2013 at 03:45:42PM -0700, Nicholas A. Bellinger wrote:
> > > On Thu, 2013-03-28 at 00:28 +0200, Michael S. Tsirkin wrote:
> > > > On Wed, Mar 27, 2013 at 09:59:45PM +0000, Nicholas A. Bellinger wrote:
> > > > > From: Paolo Bonzini <pbonzini@redhat.com>
> > > > >
> > > > > The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
> > > > > The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
> > > > > available from the QEMU command-line. Instead, I hardcode it to zero.
> > > > >
> > > > > Changes in V4:
> > > > > - Set event_idx=off by default (nab, thanks asias)
> > > >
> > > > Why? What's going on here?
> > > >
> > >
> > > Not disabling event_idx by default, or disabling from the command line
> > > ends up resulting in ->handle_kick() not getting called for subsequent
> > > commands..
> > >
> > > I spent some time trying to track this down recently with no luck, and
> > > AFAICT it's always been required in order for vhost-scsi to function.
> >
> > Hmm this is a bug in kernel then. A better work-around is
> > to disable EVENT_IDX in kernel. Let's do it for 3.9?
>
> But before we do, can you check that SET_FEATURES is
> called with this bit set if you enable event_idx?
> If not that's your bug then ...
>
Ok, SET_FEATURES is currently not setting any bits at all based upon
vhost_dev->features, so it looks like a vhost-scsi-pci bug..
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-27 23:24 ` Nicholas A. Bellinger
@ 2013-03-28 0:21 ` Nicholas A. Bellinger
2013-03-28 6:19 ` Michael S. Tsirkin
1 sibling, 0 replies; 16+ messages in thread
From: Nicholas A. Bellinger @ 2013-03-28 0:21 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Wed, 2013-03-27 at 16:24 -0700, Nicholas A. Bellinger wrote:
> On Thu, 2013-03-28 at 00:53 +0200, Michael S. Tsirkin wrote:
> > On Thu, Mar 28, 2013 at 12:50:21AM +0200, Michael S. Tsirkin wrote:
> > > On Wed, Mar 27, 2013 at 03:45:42PM -0700, Nicholas A. Bellinger wrote:
> > > > On Thu, 2013-03-28 at 00:28 +0200, Michael S. Tsirkin wrote:
> > > > > On Wed, Mar 27, 2013 at 09:59:45PM +0000, Nicholas A. Bellinger wrote:
> > > > > > From: Paolo Bonzini <pbonzini@redhat.com>
> > > > > >
> > > > > > The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
> > > > > > The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
> > > > > > available from the QEMU command-line. Instead, I hardcode it to zero.
> > > > > >
> > > > > > Changes in V4:
> > > > > > - Set event_idx=off by default (nab, thanks asias)
> > > > >
> > > > > Why? What's going on here?
> > > > >
> > > >
> > > > Not disabling event_idx by default, or disabling from the command line
> > > > ends up resulting in ->handle_kick() not getting called for subsequent
> > > > commands..
> > > >
> > > > I spent some time trying to track this down recently with no luck, and
> > > > AFAICT it's always been required in order for vhost-scsi to function.
> > >
> > > Hmm this is a bug in kernel then. A better work-around is
> > > to disable EVENT_IDX in kernel. Let's do it for 3.9?
> >
> > But before we do, can you check that SET_FEATURES is
> > called with this bit set if you enable event_idx?
> > If not that's your bug then ...
> >
>
> Ok, SET_FEATURES is currently not setting any bits at all based upon
> vhost_dev->features, so it looks like a vhost-scsi-pci bug..
>
>
Adding the following patch for a vhost_scsi specific ->get_features()
call in order to strip off the unsupported host feature bits. Please
review.
Also, providing the proper EVENT_IDX bit to SET_FEATURES results in the
same lost ->handle_kick() for vhost-scsi-pci devices, so the previous
VHOST_TCM_FEATURES patch to strip off EVENT_IDX is still required to
work.. Sending this out separately, please review and ACK.
Thanks,
--nab
diff --git a/hw/vhost-scsi.c b/hw/vhost-scsi.c
index 70e42fc..adc929b 100644
--- a/hw/vhost-scsi.c
+++ b/hw/vhost-scsi.c
@@ -128,6 +128,28 @@ static void vhost_scsi_stop(VHostSCSI *vs, VirtIODevice *vdev)
vhost_dev_disable_notifiers(&vs->dev, vdev);
}
+static uint32_t vhost_scsi_get_features(VirtIODevice *vdev,
+ uint32_t features)
+{
+ VHostSCSI *vs = (VHostSCSI *)vdev;
+
+ /* Clear features not supported by host kernel. */
+ if (!(vs->dev.features & (1 << VIRTIO_F_NOTIFY_ON_EMPTY))) {
+ features &= ~(1 << VIRTIO_F_NOTIFY_ON_EMPTY);
+ }
+ if (!(vs->dev.features & (1 << VIRTIO_RING_F_INDIRECT_DESC))) {
+ features &= ~(1 << VIRTIO_RING_F_INDIRECT_DESC);
+ }
+ if (!(vs->dev.features & (1 << VIRTIO_RING_F_EVENT_IDX))) {
+ features &= ~(1 << VIRTIO_RING_F_EVENT_IDX);
+ }
+ if (!(vs->dev.features & (1 << VIRTIO_SCSI_F_HOTPLUG))) {
+ features &= ~(1 << VIRTIO_SCSI_F_HOTPLUG);
+ }
+
+ return features;
+}
+
static void vhost_scsi_set_config(VirtIODevice *vdev,
const uint8_t *config)
{
@@ -203,6 +225,7 @@ VirtIODevice *vhost_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf)
vs = (VHostSCSI *)virtio_scsi_init_common(dev, proxyconf,
sizeof(VHostSCSI));
+ vs->vs.vdev.get_features = vhost_scsi_get_features;
vs->vs.vdev.set_config = vhost_scsi_set_config;
vs->vs.vdev.set_status = vhost_scsi_set_status;
vs->vs.vdev.guest_notifier_mask = vhost_scsi_guest_notifier_mask;
@@ -219,7 +242,7 @@ VirtIODevice *vhost_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf)
return NULL;
}
vs->dev.backend_features = 0;
- vs->dev.acked_features = 0;
+ vs->dev.acked_features = vs->dev.features;
error_setg(&vs->migration_blocker,
"vhost-scsi does not support migration");
diff --git a/hw/vhost-scsi.h b/hw/vhost-scsi.h
index b01f012..70f4a58 100644
--- a/hw/vhost-scsi.h
+++ b/hw/vhost-scsi.h
@@ -50,14 +50,12 @@ enum vhost_scsi_vq_list {
#define VHOST_SCSI_GET_ABI_VERSION _IOW(VHOST_VIRTIO, 0x42, int)
#define DEFINE_VHOST_SCSI_PROPERTIES(_state, _features_field, _conf_field) \
- DEFINE_PROP_BIT("indirect_desc", _state, _features_field, VIRTIO_RING_F_INDIRECT_DESC, true), \
- DEFINE_PROP_BIT("event_idx", _state, _features_field, VIRTIO_RING_F_EVENT_IDX, false), \
+ DEFINE_VIRTIO_COMMON_FEATURES(_state, _features_field), \
DEFINE_PROP_STRING("vhostfd", _state, _conf_field.vhostfd), \
DEFINE_PROP_STRING("wwpn", _state, _conf_field.wwpn), \
DEFINE_PROP_UINT32("num_queues", _state, _conf_field.num_queues, 1), \
DEFINE_PROP_UINT32("max_sectors", _state, _conf_field.max_sectors, 0xFFFF), \
- DEFINE_PROP_UINT32("cmd_per_lun", _state, _conf_field.cmd_per_lun, 128), \
- DEFINE_PROP_BIT("hotplug", _state, _features_field, VIRTIO_SCSI_F_HOTPLUG, true)
+ DEFINE_PROP_UINT32("cmd_per_lun", _state, _conf_field.cmd_per_lun, 128)
VirtIODevice *vhost_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf);
void vhost_scsi_exit(VirtIODevice *vdev);
diff --git a/hw/virtio-scsi.c b/hw/virtio-scsi.c
index 9dc7150..c59e9c6 100644
--- a/hw/virtio-scsi.c
+++ b/hw/virtio-scsi.c
@@ -595,7 +595,6 @@ VirtIOSCSICommon *virtio_scsi_init_common(DeviceState *dev, VirtIOSCSIConf *prox
s->cdb_size = VIRTIO_SCSI_CDB_SIZE;
s->vdev.get_config = virtio_scsi_get_config;
- s->vdev.get_features = virtio_scsi_get_features;
s->ctrl_vq = virtio_add_queue(&s->vdev, VIRTIO_SCSI_VQ_SIZE,
virtio_scsi_handle_ctrl);
@@ -618,6 +617,7 @@ VirtIODevice *virtio_scsi_init(DeviceState *dev, VirtIOSCSIConf *proxyconf)
s = (VirtIOSCSI *)virtio_scsi_init_common(dev, proxyconf, sizeof(VirtIOSCSI));
+ s->vs.vdev.get_features = virtio_scsi_get_features;
s->vs.vdev.set_config = virtio_scsi_set_config;
s->vs.vdev.reset = virtio_scsi_reset;
^ permalink raw reply related [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-27 23:24 ` Nicholas A. Bellinger
2013-03-28 0:21 ` Nicholas A. Bellinger
@ 2013-03-28 6:19 ` Michael S. Tsirkin
2013-03-28 6:28 ` Nicholas A. Bellinger
1 sibling, 1 reply; 16+ messages in thread
From: Michael S. Tsirkin @ 2013-03-28 6:19 UTC (permalink / raw)
To: Nicholas A. Bellinger
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Wed, Mar 27, 2013 at 04:24:22PM -0700, Nicholas A. Bellinger wrote:
> On Thu, 2013-03-28 at 00:53 +0200, Michael S. Tsirkin wrote:
> > On Thu, Mar 28, 2013 at 12:50:21AM +0200, Michael S. Tsirkin wrote:
> > > On Wed, Mar 27, 2013 at 03:45:42PM -0700, Nicholas A. Bellinger wrote:
> > > > On Thu, 2013-03-28 at 00:28 +0200, Michael S. Tsirkin wrote:
> > > > > On Wed, Mar 27, 2013 at 09:59:45PM +0000, Nicholas A. Bellinger wrote:
> > > > > > From: Paolo Bonzini <pbonzini@redhat.com>
> > > > > >
> > > > > > The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
> > > > > > The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
> > > > > > available from the QEMU command-line. Instead, I hardcode it to zero.
> > > > > >
> > > > > > Changes in V4:
> > > > > > - Set event_idx=off by default (nab, thanks asias)
> > > > >
> > > > > Why? What's going on here?
> > > > >
> > > >
> > > > Not disabling event_idx by default, or disabling from the command line
> > > > ends up resulting in ->handle_kick() not getting called for subsequent
> > > > commands..
> > > >
> > > > I spent some time trying to track this down recently with no luck, and
> > > > AFAICT it's always been required in order for vhost-scsi to function.
> > >
> > > Hmm this is a bug in kernel then. A better work-around is
> > > to disable EVENT_IDX in kernel. Let's do it for 3.9?
> >
> > But before we do, can you check that SET_FEATURES is
> > called with this bit set if you enable event_idx?
> > If not that's your bug then ...
> >
>
> Ok, SET_FEATURES is currently not setting any bits at all based upon
> vhost_dev->features, so it looks like a vhost-scsi-pci bug..
>
>
If you call SET_FEATURES, does it start working even without
disabling EVENT_IDX in kernel?
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-28 6:19 ` Michael S. Tsirkin
@ 2013-03-28 6:28 ` Nicholas A. Bellinger
2013-03-28 16:54 ` Michael S. Tsirkin
0 siblings, 1 reply; 16+ messages in thread
From: Nicholas A. Bellinger @ 2013-03-28 6:28 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Thu, 2013-03-28 at 08:19 +0200, Michael S. Tsirkin wrote:
> On Wed, Mar 27, 2013 at 04:24:22PM -0700, Nicholas A. Bellinger wrote:
> > On Thu, 2013-03-28 at 00:53 +0200, Michael S. Tsirkin wrote:
> > > On Thu, Mar 28, 2013 at 12:50:21AM +0200, Michael S. Tsirkin wrote:
> > > > On Wed, Mar 27, 2013 at 03:45:42PM -0700, Nicholas A. Bellinger wrote:
> > > > > On Thu, 2013-03-28 at 00:28 +0200, Michael S. Tsirkin wrote:
> > > > > > On Wed, Mar 27, 2013 at 09:59:45PM +0000, Nicholas A. Bellinger wrote:
> > > > > > > From: Paolo Bonzini <pbonzini@redhat.com>
> > > > > > >
> > > > > > > The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
> > > > > > > The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
> > > > > > > available from the QEMU command-line. Instead, I hardcode it to zero.
> > > > > > >
> > > > > > > Changes in V4:
> > > > > > > - Set event_idx=off by default (nab, thanks asias)
> > > > > >
> > > > > > Why? What's going on here?
> > > > > >
> > > > >
> > > > > Not disabling event_idx by default, or disabling from the command line
> > > > > ends up resulting in ->handle_kick() not getting called for subsequent
> > > > > commands..
> > > > >
> > > > > I spent some time trying to track this down recently with no luck, and
> > > > > AFAICT it's always been required in order for vhost-scsi to function.
> > > >
> > > > Hmm this is a bug in kernel then. A better work-around is
> > > > to disable EVENT_IDX in kernel. Let's do it for 3.9?
> > >
> > > But before we do, can you check that SET_FEATURES is
> > > called with this bit set if you enable event_idx?
> > > If not that's your bug then ...
> > >
> >
> > Ok, SET_FEATURES is currently not setting any bits at all based upon
> > vhost_dev->features, so it looks like a vhost-scsi-pci bug..
> >
> >
>
> If you call SET_FEATURES, does it start working even without
> disabling EVENT_IDX in kernel?
>
Unfortunately, no. The same strangeness persists unless event_idx=off
is passed, or EVENT_IDX is disabled in the kernel and stripped off with
the vhost_scsi_get_features() patch.
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module
2013-03-28 6:28 ` Nicholas A. Bellinger
@ 2013-03-28 16:54 ` Michael S. Tsirkin
0 siblings, 0 replies; 16+ messages in thread
From: Michael S. Tsirkin @ 2013-03-28 16:54 UTC (permalink / raw)
To: Nicholas A. Bellinger
Cc: kvm-devel, qemu-devel, lf-virt, Anthony Liguori, target-devel,
Stefan Hajnoczi, Paolo Bonzini, Asias He
On Wed, Mar 27, 2013 at 11:28:56PM -0700, Nicholas A. Bellinger wrote:
> On Thu, 2013-03-28 at 08:19 +0200, Michael S. Tsirkin wrote:
> > On Wed, Mar 27, 2013 at 04:24:22PM -0700, Nicholas A. Bellinger wrote:
> > > On Thu, 2013-03-28 at 00:53 +0200, Michael S. Tsirkin wrote:
> > > > On Thu, Mar 28, 2013 at 12:50:21AM +0200, Michael S. Tsirkin wrote:
> > > > > On Wed, Mar 27, 2013 at 03:45:42PM -0700, Nicholas A. Bellinger wrote:
> > > > > > On Thu, 2013-03-28 at 00:28 +0200, Michael S. Tsirkin wrote:
> > > > > > > On Wed, Mar 27, 2013 at 09:59:45PM +0000, Nicholas A. Bellinger wrote:
> > > > > > > > From: Paolo Bonzini <pbonzini@redhat.com>
> > > > > > > >
> > > > > > > > The WWPN specified in configfs is passed to "-device vhost-scsi-pci".
> > > > > > > > The tgpt field of the SET_ENDPOINT ioctl is obsolete now, so it is not
> > > > > > > > available from the QEMU command-line. Instead, I hardcode it to zero.
> > > > > > > >
> > > > > > > > Changes in V4:
> > > > > > > > - Set event_idx=off by default (nab, thanks asias)
> > > > > > >
> > > > > > > Why? What's going on here?
> > > > > > >
> > > > > >
> > > > > > Not disabling event_idx by default, or disabling from the command line
> > > > > > ends up resulting in ->handle_kick() not getting called for subsequent
> > > > > > commands..
> > > > > >
> > > > > > I spent some time trying to track this down recently with no luck, and
> > > > > > AFAICT it's always been required in order for vhost-scsi to function.
> > > > >
> > > > > Hmm this is a bug in kernel then. A better work-around is
> > > > > to disable EVENT_IDX in kernel. Let's do it for 3.9?
> > > >
> > > > But before we do, can you check that SET_FEATURES is
> > > > called with this bit set if you enable event_idx?
> > > > If not that's your bug then ...
> > > >
> > >
> > > Ok, SET_FEATURES is currently not setting any bits at all based upon
> > > vhost_dev->features, so it looks like a vhost-scsi-pci bug..
> > >
> > >
> >
> > If you call SET_FEATURES, does it start working even without
> > disabling EVENT_IDX in kernel?
> >
>
> Unfortunately, no. The same strangeness persists unless event_idx=off
> is passed, or EVENT_IDX is disabled in the kernel and stripped off with
> the vhost_scsi_get_features() patch.
>
I think it's important to fix it (usually exposes other bugs)
but I agree, lets get it working without first.
^ permalink raw reply [flat|nested] 16+ messages in thread
end of thread, other threads:[~2013-03-28 16:53 UTC | newest]
Thread overview: 16+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-03-27 21:59 [Qemu-devel] [PATCH 0/3] Add support for vhost-scsi-pci Nicholas A. Bellinger
2013-03-27 21:59 ` [Qemu-devel] [PATCH 1/3] virtio-scsi: create VirtIOSCSICommon Nicholas A. Bellinger
2013-03-27 21:59 ` [Qemu-devel] [PATCH 2/3] vhost-scsi: new device supporting the tcm_vhost Linux kernel module Nicholas A. Bellinger
2013-03-27 22:28 ` Michael S. Tsirkin
2013-03-27 22:45 ` Nicholas A. Bellinger
2013-03-27 22:50 ` Michael S. Tsirkin
2013-03-27 22:53 ` Michael S. Tsirkin
2013-03-27 23:24 ` Nicholas A. Bellinger
2013-03-28 0:21 ` Nicholas A. Bellinger
2013-03-28 6:19 ` Michael S. Tsirkin
2013-03-28 6:28 ` Nicholas A. Bellinger
2013-03-28 16:54 ` Michael S. Tsirkin
2013-03-27 23:05 ` Nicholas A. Bellinger
2013-03-27 21:59 ` [Qemu-devel] [PATCH 3/3] vhost: Only call vhost_verify_ring_mappings when adding vhost memory Nicholas A. Bellinger
2013-03-27 22:15 ` Michael S. Tsirkin
2013-03-27 22:16 ` [Qemu-devel] [PATCH 0/3] Add support for vhost-scsi-pci Michael S. Tsirkin
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).