]> xenbits.xensource.com Git - people/dariof/qemu-xen.git/commitdiff
Remove VXHS block device
authorMarc-André Lureau <marcandre.lureau@redhat.com>
Sat, 11 Jul 2020 06:59:26 +0000 (10:59 +0400)
committerKevin Wolf <kwolf@redhat.com>
Fri, 17 Jul 2020 12:20:57 +0000 (14:20 +0200)
The vxhs code doesn't compile since v2.12.0. There's no point in fixing
and then adding CI for a config that our users have demonstrated that
they do not use; better to just remove it.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20200711065926.2204721-1-marcandre.lureau@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
18 files changed:
block/Makefile.objs
block/trace-events
block/vxhs.c [deleted file]
configure
docs/system/deprecated.rst
qapi/block-core.json
tests/qemu-iotests/017
tests/qemu-iotests/029
tests/qemu-iotests/073
tests/qemu-iotests/114
tests/qemu-iotests/130
tests/qemu-iotests/134
tests/qemu-iotests/156
tests/qemu-iotests/158
tests/qemu-iotests/282
tests/qemu-iotests/check
tests/qemu-iotests/common.filter
tests/qemu-iotests/common.rc

index 577e578bc220cace83475ff2c886777e4c677a15..19c6f371c9dfa2cceab994291aa4ef7bded19265 100644 (file)
@@ -31,7 +31,6 @@ block-obj-$(CONFIG_LIBNFS) += nfs.o
 block-obj-$(CONFIG_CURL) += curl.o
 block-obj-$(CONFIG_RBD) += rbd.o
 block-obj-$(CONFIG_GLUSTERFS) += gluster.o
-block-obj-$(CONFIG_VXHS) += vxhs.o
 block-obj-$(CONFIG_LIBSSH) += ssh.o
 block-obj-y += accounting.o dirty-bitmap.o
 block-obj-y += write-threshold.o
@@ -61,7 +60,6 @@ rbd.o-cflags       := $(RBD_CFLAGS)
 rbd.o-libs         := $(RBD_LIBS)
 gluster.o-cflags   := $(GLUSTERFS_CFLAGS)
 gluster.o-libs     := $(GLUSTERFS_LIBS)
-vxhs.o-libs        := $(VXHS_LIBS)
 ssh.o-cflags       := $(LIBSSH_CFLAGS)
 ssh.o-libs         := $(LIBSSH_LIBS)
 block-obj-dmg-bz2-$(CONFIG_BZIP2) += dmg-bz2.o
index dbe76a7613688858f1342de8e4c6c125558cbd2b..d3533ca8963cb644c80dbdbc2733595df89fe095 100644 (file)
@@ -136,23 +136,6 @@ qed_aio_write_prefill(void *s, void *acb, uint64_t start, size_t len, uint64_t o
 qed_aio_write_postfill(void *s, void *acb, uint64_t start, size_t len, uint64_t offset) "s %p acb %p start %"PRIu64" len %zu offset %"PRIu64
 qed_aio_write_main(void *s, void *acb, int ret, uint64_t offset, size_t len) "s %p acb %p ret %d offset %"PRIu64" len %zu"
 
-# vxhs.c
-vxhs_iio_callback(int error) "ctx is NULL: error %d"
-vxhs_iio_callback_chnfail(int err, int error) "QNIO channel failed, no i/o %d, %d"
-vxhs_iio_callback_unknwn(int opcode, int err) "unexpected opcode %d, errno %d"
-vxhs_aio_rw_invalid(int req) "Invalid I/O request iodir %d"
-vxhs_aio_rw_ioerr(char *guid, int iodir, uint64_t size, uint64_t off, void *acb, int ret, int err) "IO ERROR (vDisk %s) FOR : Read/Write = %d size = %"PRIu64" offset = %"PRIu64" ACB = %p. Error = %d, errno = %d"
-vxhs_get_vdisk_stat_err(char *guid, int ret, int err) "vDisk (%s) stat ioctl failed, ret = %d, errno = %d"
-vxhs_get_vdisk_stat(char *vdisk_guid, uint64_t vdisk_size) "vDisk %s stat ioctl returned size %"PRIu64
-vxhs_complete_aio(void *acb, uint64_t ret) "aio failed acb %p ret %"PRIu64
-vxhs_parse_uri_filename(const char *filename) "URI passed via bdrv_parse_filename %s"
-vxhs_open_vdiskid(const char *vdisk_id) "Opening vdisk-id %s"
-vxhs_open_hostinfo(char *of_vsa_addr, int port) "Adding host %s:%d to BDRVVXHSState"
-vxhs_open_iio_open(const char *host) "Failed to connect to storage agent on host %s"
-vxhs_parse_uri_hostinfo(char *host, int port) "Host: IP %s, Port %d"
-vxhs_close(char *vdisk_guid) "Closing vdisk %s"
-vxhs_get_creds(const char *cacert, const char *client_key, const char *client_cert) "cacert %s, client_key %s, client_cert %s"
-
 # nvme.c
 nvme_kick(void *s, int queue) "s %p queue %d"
 nvme_dma_flush_queue_wait(void *s) "s %p"
diff --git a/block/vxhs.c b/block/vxhs.c
deleted file mode 100644 (file)
index dc0e254..0000000
+++ /dev/null
@@ -1,587 +0,0 @@
-/*
- * QEMU Block driver for Veritas HyperScale (VxHS)
- *
- * Copyright (c) 2017 Veritas Technologies LLC.
- *
- * This work is licensed under the terms of the GNU GPL, version 2 or later.
- * See the COPYING file in the top-level directory.
- *
- */
-
-#include "qemu/osdep.h"
-#include <qnio/qnio_api.h>
-#include <sys/param.h>
-#include "block/block_int.h"
-#include "block/qdict.h"
-#include "qapi/qmp/qerror.h"
-#include "qapi/qmp/qdict.h"
-#include "qapi/qmp/qstring.h"
-#include "trace.h"
-#include "qemu/module.h"
-#include "qemu/uri.h"
-#include "qapi/error.h"
-#include "qemu/uuid.h"
-#include "crypto/tlscredsx509.h"
-#include "sysemu/replay.h"
-
-#define VXHS_OPT_FILENAME           "filename"
-#define VXHS_OPT_VDISK_ID           "vdisk-id"
-#define VXHS_OPT_SERVER             "server"
-#define VXHS_OPT_HOST               "host"
-#define VXHS_OPT_PORT               "port"
-
-/* Only accessed under QEMU global mutex */
-static uint32_t vxhs_ref;
-
-typedef enum {
-    VDISK_AIO_READ,
-    VDISK_AIO_WRITE,
-} VDISKAIOCmd;
-
-/*
- * HyperScale AIO callbacks structure
- */
-typedef struct VXHSAIOCB {
-    BlockAIOCB common;
-    int err;
-} VXHSAIOCB;
-
-typedef struct VXHSvDiskHostsInfo {
-    void *dev_handle; /* Device handle */
-    char *host; /* Host name or IP */
-    int port; /* Host's port number */
-} VXHSvDiskHostsInfo;
-
-/*
- * Structure per vDisk maintained for state
- */
-typedef struct BDRVVXHSState {
-    VXHSvDiskHostsInfo vdisk_hostinfo; /* Per host info */
-    char *vdisk_guid;
-    char *tlscredsid; /* tlscredsid */
-} BDRVVXHSState;
-
-static void vxhs_complete_aio_bh(void *opaque)
-{
-    VXHSAIOCB *acb = opaque;
-    BlockCompletionFunc *cb = acb->common.cb;
-    void *cb_opaque = acb->common.opaque;
-    int ret = 0;
-
-    if (acb->err != 0) {
-        trace_vxhs_complete_aio(acb, acb->err);
-        ret = (-EIO);
-    }
-
-    qemu_aio_unref(acb);
-    cb(cb_opaque, ret);
-}
-
-/*
- * Called from a libqnio thread
- */
-static void vxhs_iio_callback(void *ctx, uint32_t opcode, uint32_t error)
-{
-    VXHSAIOCB *acb = NULL;
-
-    switch (opcode) {
-    case IRP_READ_REQUEST:
-    case IRP_WRITE_REQUEST:
-
-        /*
-         * ctx is VXHSAIOCB*
-         * ctx is NULL if error is QNIOERROR_CHANNEL_HUP
-         */
-        if (ctx) {
-            acb = ctx;
-        } else {
-            trace_vxhs_iio_callback(error);
-            goto out;
-        }
-
-        if (error) {
-            if (!acb->err) {
-                acb->err = error;
-            }
-            trace_vxhs_iio_callback(error);
-        }
-
-        replay_bh_schedule_oneshot_event(bdrv_get_aio_context(acb->common.bs),
-                                         vxhs_complete_aio_bh, acb);
-        break;
-
-    default:
-        if (error == QNIOERROR_HUP) {
-            /*
-             * Channel failed, spontaneous notification,
-             * not in response to I/O
-             */
-            trace_vxhs_iio_callback_chnfail(error, errno);
-        } else {
-            trace_vxhs_iio_callback_unknwn(opcode, error);
-        }
-        break;
-    }
-out:
-    return;
-}
-
-static QemuOptsList runtime_opts = {
-    .name = "vxhs",
-    .head = QTAILQ_HEAD_INITIALIZER(runtime_opts.head),
-    .desc = {
-        {
-            .name = VXHS_OPT_FILENAME,
-            .type = QEMU_OPT_STRING,
-            .help = "URI to the Veritas HyperScale image",
-        },
-        {
-            .name = VXHS_OPT_VDISK_ID,
-            .type = QEMU_OPT_STRING,
-            .help = "UUID of the VxHS vdisk",
-        },
-        {
-            .name = "tls-creds",
-            .type = QEMU_OPT_STRING,
-            .help = "ID of the TLS/SSL credentials to use",
-        },
-        { /* end of list */ }
-    },
-};
-
-static QemuOptsList runtime_tcp_opts = {
-    .name = "vxhs_tcp",
-    .head = QTAILQ_HEAD_INITIALIZER(runtime_tcp_opts.head),
-    .desc = {
-        {
-            .name = VXHS_OPT_HOST,
-            .type = QEMU_OPT_STRING,
-            .help = "host address (ipv4 addresses)",
-        },
-        {
-            .name = VXHS_OPT_PORT,
-            .type = QEMU_OPT_NUMBER,
-            .help = "port number on which VxHSD is listening (default 9999)",
-            .def_value_str = "9999"
-        },
-        { /* end of list */ }
-    },
-};
-
-/*
- * Parse incoming URI and populate *options with the host
- * and device information
- */
-static int vxhs_parse_uri(const char *filename, QDict *options)
-{
-    URI *uri = NULL;
-    char *port;
-    int ret = 0;
-
-    trace_vxhs_parse_uri_filename(filename);
-    uri = uri_parse(filename);
-    if (!uri || !uri->server || !uri->path) {
-        uri_free(uri);
-        return -EINVAL;
-    }
-
-    qdict_put_str(options, VXHS_OPT_SERVER ".host", uri->server);
-
-    if (uri->port) {
-        port = g_strdup_printf("%d", uri->port);
-        qdict_put_str(options, VXHS_OPT_SERVER ".port", port);
-        g_free(port);
-    }
-
-    qdict_put_str(options, "vdisk-id", uri->path);
-
-    trace_vxhs_parse_uri_hostinfo(uri->server, uri->port);
-    uri_free(uri);
-
-    return ret;
-}
-
-static void vxhs_parse_filename(const char *filename, QDict *options,
-                                Error **errp)
-{
-    if (qdict_haskey(options, "vdisk-id") || qdict_haskey(options, "server")) {
-        error_setg(errp, "vdisk-id/server and a file name may not be specified "
-                         "at the same time");
-        return;
-    }
-
-    if (strstr(filename, "://")) {
-        int ret = vxhs_parse_uri(filename, options);
-        if (ret < 0) {
-            error_setg(errp, "Invalid URI. URI should be of the form "
-                       "  vxhs://<host_ip>:<port>/<vdisk-id>");
-        }
-    }
-}
-
-static void vxhs_refresh_limits(BlockDriverState *bs, Error **errp)
-{
-    /* XXX Does VXHS support AIO on less than 512-byte alignment? */
-    bs->bl.request_alignment = 512;
-}
-
-static int vxhs_init_and_ref(void)
-{
-    if (vxhs_ref++ == 0) {
-        if (iio_init(QNIO_VERSION, vxhs_iio_callback)) {
-            return -ENODEV;
-        }
-    }
-    return 0;
-}
-
-static void vxhs_unref(void)
-{
-    if (--vxhs_ref == 0) {
-        iio_fini();
-    }
-}
-
-static void vxhs_get_tls_creds(const char *id, char **cacert,
-                               char **key, char **cert, Error **errp)
-{
-    Object *obj;
-    QCryptoTLSCreds *creds;
-    QCryptoTLSCredsX509 *creds_x509;
-
-    obj = object_resolve_path_component(
-        object_get_objects_root(), id);
-
-    if (!obj) {
-        error_setg(errp, "No TLS credentials with id '%s'",
-                   id);
-        return;
-    }
-
-    creds_x509 = (QCryptoTLSCredsX509 *)
-        object_dynamic_cast(obj, TYPE_QCRYPTO_TLS_CREDS_X509);
-
-    if (!creds_x509) {
-        error_setg(errp, "Object with id '%s' is not TLS credentials",
-                   id);
-        return;
-    }
-
-    creds = &creds_x509->parent_obj;
-
-    if (creds->endpoint != QCRYPTO_TLS_CREDS_ENDPOINT_CLIENT) {
-        error_setg(errp,
-                   "Expecting TLS credentials with a client endpoint");
-        return;
-    }
-
-    /*
-     * Get the cacert, client_cert and client_key file names.
-     */
-    if (!creds->dir) {
-        error_setg(errp, "TLS object missing 'dir' property value");
-        return;
-    }
-
-    *cacert = g_strdup_printf("%s/%s", creds->dir,
-                              QCRYPTO_TLS_CREDS_X509_CA_CERT);
-    *cert = g_strdup_printf("%s/%s", creds->dir,
-                            QCRYPTO_TLS_CREDS_X509_CLIENT_CERT);
-    *key = g_strdup_printf("%s/%s", creds->dir,
-                           QCRYPTO_TLS_CREDS_X509_CLIENT_KEY);
-}
-
-static int vxhs_open(BlockDriverState *bs, QDict *options,
-                     int bdrv_flags, Error **errp)
-{
-    BDRVVXHSState *s = bs->opaque;
-    void *dev_handlep;
-    QDict *backing_options = NULL;
-    QemuOpts *opts = NULL;
-    QemuOpts *tcp_opts = NULL;
-    char *of_vsa_addr = NULL;
-    Error *local_err = NULL;
-    const char *vdisk_id_opt;
-    const char *server_host_opt;
-    int ret = 0;
-    char *cacert = NULL;
-    char *client_key = NULL;
-    char *client_cert = NULL;
-
-    ret = vxhs_init_and_ref();
-    if (ret < 0) {
-        ret = -EINVAL;
-        goto out;
-    }
-
-    /* Create opts info from runtime_opts and runtime_tcp_opts list */
-    opts = qemu_opts_create(&runtime_opts, NULL, 0, &error_abort);
-    tcp_opts = qemu_opts_create(&runtime_tcp_opts, NULL, 0, &error_abort);
-
-    if (!qemu_opts_absorb_qdict(opts, options, errp)) {
-        ret = -EINVAL;
-        goto out;
-    }
-
-    /* vdisk-id is the disk UUID */
-    vdisk_id_opt = qemu_opt_get(opts, VXHS_OPT_VDISK_ID);
-    if (!vdisk_id_opt) {
-        error_setg(errp, QERR_MISSING_PARAMETER, VXHS_OPT_VDISK_ID);
-        ret = -EINVAL;
-        goto out;
-    }
-
-    /* vdisk-id may contain a leading '/' */
-    if (strlen(vdisk_id_opt) > UUID_FMT_LEN + 1) {
-        error_setg(errp, "vdisk-id cannot be more than %d characters",
-                   UUID_FMT_LEN);
-        ret = -EINVAL;
-        goto out;
-    }
-
-    s->vdisk_guid = g_strdup(vdisk_id_opt);
-    trace_vxhs_open_vdiskid(vdisk_id_opt);
-
-    /* get the 'server.' arguments */
-    qdict_extract_subqdict(options, &backing_options, VXHS_OPT_SERVER".");
-
-    if (!qemu_opts_absorb_qdict(tcp_opts, backing_options, errp)) {
-        ret = -EINVAL;
-        goto out;
-    }
-
-    server_host_opt = qemu_opt_get(tcp_opts, VXHS_OPT_HOST);
-    if (!server_host_opt) {
-        error_setg(errp, QERR_MISSING_PARAMETER,
-                   VXHS_OPT_SERVER"."VXHS_OPT_HOST);
-        ret = -EINVAL;
-        goto out;
-    }
-
-    if (strlen(server_host_opt) > MAXHOSTNAMELEN) {
-        error_setg(errp, "server.host cannot be more than %d characters",
-                   MAXHOSTNAMELEN);
-        ret = -EINVAL;
-        goto out;
-    }
-
-    /* check if we got tls-creds via the --object argument */
-    s->tlscredsid = g_strdup(qemu_opt_get(opts, "tls-creds"));
-    if (s->tlscredsid) {
-        vxhs_get_tls_creds(s->tlscredsid, &cacert, &client_key,
-                           &client_cert, &local_err);
-        if (local_err != NULL) {
-            ret = -EINVAL;
-            goto out;
-        }
-        trace_vxhs_get_creds(cacert, client_key, client_cert);
-    }
-
-    s->vdisk_hostinfo.host = g_strdup(server_host_opt);
-    s->vdisk_hostinfo.port = g_ascii_strtoll(qemu_opt_get(tcp_opts,
-                                                          VXHS_OPT_PORT),
-                                                          NULL, 0);
-
-    trace_vxhs_open_hostinfo(s->vdisk_hostinfo.host,
-                             s->vdisk_hostinfo.port);
-
-    of_vsa_addr = g_strdup_printf("of://%s:%d",
-                                  s->vdisk_hostinfo.host,
-                                  s->vdisk_hostinfo.port);
-
-    /*
-     * Open qnio channel to storage agent if not opened before
-     */
-    dev_handlep = iio_open(of_vsa_addr, s->vdisk_guid, 0,
-                           cacert, client_key, client_cert);
-    if (dev_handlep == NULL) {
-        trace_vxhs_open_iio_open(of_vsa_addr);
-        ret = -ENODEV;
-        goto out;
-    }
-    s->vdisk_hostinfo.dev_handle = dev_handlep;
-
-out:
-    g_free(of_vsa_addr);
-    qobject_unref(backing_options);
-    qemu_opts_del(tcp_opts);
-    qemu_opts_del(opts);
-    g_free(cacert);
-    g_free(client_key);
-    g_free(client_cert);
-
-    if (ret < 0) {
-        vxhs_unref();
-        g_free(s->vdisk_hostinfo.host);
-        g_free(s->vdisk_guid);
-        g_free(s->tlscredsid);
-        s->vdisk_guid = NULL;
-    }
-
-    return ret;
-}
-
-static const AIOCBInfo vxhs_aiocb_info = {
-    .aiocb_size = sizeof(VXHSAIOCB)
-};
-
-/*
- * This allocates QEMU-VXHS callback for each IO
- * and is passed to QNIO. When QNIO completes the work,
- * it will be passed back through the callback.
- */
-static BlockAIOCB *vxhs_aio_rw(BlockDriverState *bs, uint64_t offset,
-                               QEMUIOVector *qiov, uint64_t size,
-                               BlockCompletionFunc *cb, void *opaque,
-                               VDISKAIOCmd iodir)
-{
-    VXHSAIOCB *acb = NULL;
-    BDRVVXHSState *s = bs->opaque;
-    int iio_flags = 0;
-    int ret = 0;
-    void *dev_handle = s->vdisk_hostinfo.dev_handle;
-
-    acb = qemu_aio_get(&vxhs_aiocb_info, bs, cb, opaque);
-
-    /*
-     * Initialize VXHSAIOCB.
-     */
-    acb->err = 0;
-
-    iio_flags = IIO_FLAG_ASYNC;
-
-    switch (iodir) {
-    case VDISK_AIO_WRITE:
-            ret = iio_writev(dev_handle, acb, qiov->iov, qiov->niov,
-                             offset, size, iio_flags);
-            break;
-    case VDISK_AIO_READ:
-            ret = iio_readv(dev_handle, acb, qiov->iov, qiov->niov,
-                            offset, size, iio_flags);
-            break;
-    default:
-            trace_vxhs_aio_rw_invalid(iodir);
-            goto errout;
-    }
-
-    if (ret != 0) {
-        trace_vxhs_aio_rw_ioerr(s->vdisk_guid, iodir, size, offset,
-                                acb, ret, errno);
-        goto errout;
-    }
-    return &acb->common;
-
-errout:
-    qemu_aio_unref(acb);
-    return NULL;
-}
-
-static BlockAIOCB *vxhs_aio_preadv(BlockDriverState *bs,
-                                   uint64_t offset, uint64_t bytes,
-                                   QEMUIOVector *qiov, int flags,
-                                   BlockCompletionFunc *cb, void *opaque)
-{
-    return vxhs_aio_rw(bs, offset, qiov, bytes, cb, opaque, VDISK_AIO_READ);
-}
-
-static BlockAIOCB *vxhs_aio_pwritev(BlockDriverState *bs,
-                                    uint64_t offset, uint64_t bytes,
-                                    QEMUIOVector *qiov, int flags,
-                                    BlockCompletionFunc *cb, void *opaque)
-{
-    return vxhs_aio_rw(bs, offset, qiov, bytes, cb, opaque, VDISK_AIO_WRITE);
-}
-
-static void vxhs_close(BlockDriverState *bs)
-{
-    BDRVVXHSState *s = bs->opaque;
-
-    trace_vxhs_close(s->vdisk_guid);
-
-    g_free(s->vdisk_guid);
-    s->vdisk_guid = NULL;
-
-    /*
-     * Close vDisk device
-     */
-    if (s->vdisk_hostinfo.dev_handle) {
-        iio_close(s->vdisk_hostinfo.dev_handle);
-        s->vdisk_hostinfo.dev_handle = NULL;
-    }
-
-    vxhs_unref();
-
-    /*
-     * Free the dynamically allocated host string etc
-     */
-    g_free(s->vdisk_hostinfo.host);
-    g_free(s->tlscredsid);
-    s->tlscredsid = NULL;
-    s->vdisk_hostinfo.host = NULL;
-    s->vdisk_hostinfo.port = 0;
-}
-
-static int64_t vxhs_get_vdisk_stat(BDRVVXHSState *s)
-{
-    int64_t vdisk_size = -1;
-    int ret = 0;
-    void *dev_handle = s->vdisk_hostinfo.dev_handle;
-
-    ret = iio_ioctl(dev_handle, IOR_VDISK_STAT, &vdisk_size, 0);
-    if (ret < 0) {
-        trace_vxhs_get_vdisk_stat_err(s->vdisk_guid, ret, errno);
-        return -EIO;
-    }
-
-    trace_vxhs_get_vdisk_stat(s->vdisk_guid, vdisk_size);
-    return vdisk_size;
-}
-
-/*
- * Returns the size of vDisk in bytes. This is required
- * by QEMU block upper block layer so that it is visible
- * to guest.
- */
-static int64_t vxhs_getlength(BlockDriverState *bs)
-{
-    BDRVVXHSState *s = bs->opaque;
-    int64_t vdisk_size;
-
-    vdisk_size = vxhs_get_vdisk_stat(s);
-    if (vdisk_size < 0) {
-        return -EIO;
-    }
-
-    return vdisk_size;
-}
-
-static const char *const vxhs_strong_runtime_opts[] = {
-    VXHS_OPT_VDISK_ID,
-    "tls-creds",
-    VXHS_OPT_HOST,
-    VXHS_OPT_PORT,
-    VXHS_OPT_SERVER".",
-
-    NULL
-};
-
-static BlockDriver bdrv_vxhs = {
-    .format_name                  = "vxhs",
-    .protocol_name                = "vxhs",
-    .instance_size                = sizeof(BDRVVXHSState),
-    .bdrv_file_open               = vxhs_open,
-    .bdrv_parse_filename          = vxhs_parse_filename,
-    .bdrv_refresh_limits          = vxhs_refresh_limits,
-    .bdrv_close                   = vxhs_close,
-    .bdrv_getlength               = vxhs_getlength,
-    .bdrv_aio_preadv              = vxhs_aio_preadv,
-    .bdrv_aio_pwritev             = vxhs_aio_pwritev,
-    .strong_runtime_opts          = vxhs_strong_runtime_opts,
-};
-
-static void bdrv_vxhs_init(void)
-{
-    bdrv_register(&bdrv_vxhs);
-}
-
-block_init(bdrv_vxhs_init);
index b751c853f54c394d2e1c2adee409da1f47c47ac2..8227962b457bb9cc083dec27b38266618045b051 100755 (executable)
--- a/configure
+++ b/configure
@@ -501,7 +501,6 @@ numa=""
 tcmalloc="no"
 jemalloc="no"
 replication="yes"
-vxhs=""
 bochs="yes"
 cloop="yes"
 dmg="yes"
@@ -1541,10 +1540,6 @@ for opt do
   ;;
   --enable-replication) replication="yes"
   ;;
-  --disable-vxhs) vxhs="no"
-  ;;
-  --enable-vxhs) vxhs="yes"
-  ;;
   --disable-bochs) bochs="no"
   ;;
   --enable-bochs) bochs="yes"
@@ -1932,7 +1927,6 @@ disabled with --disable-FEATURE, default is enabled if available:
   xfsctl          xfsctl support
   qom-cast-debug  cast debugging support
   tools           build qemu-io, qemu-nbd and qemu-img tools
-  vxhs            Veritas HyperScale vDisk backend support
   bochs           bochs image format support
   cloop           cloop image format support
   dmg             dmg image format support
@@ -6249,33 +6243,6 @@ if compile_prog "" "" ; then
     have_sysmacros=yes
 fi
 
-##########################################
-# Veritas HyperScale block driver VxHS
-# Check if libvxhs is installed
-
-if test "$vxhs" != "no" ; then
-  cat > $TMPC <<EOF
-#include <stdint.h>
-#include <qnio/qnio_api.h>
-
-void *vxhs_callback;
-
-int main(void) {
-    iio_init(QNIO_VERSION, vxhs_callback);
-    return 0;
-}
-EOF
-  vxhs_libs="-lvxhs -lssl"
-  if compile_prog "" "$vxhs_libs" ; then
-    vxhs=yes
-  else
-    if test "$vxhs" = "yes" ; then
-      feature_not_found "vxhs block device" "Install libvxhs See github"
-    fi
-    vxhs=no
-  fi
-fi
-
 ##########################################
 # check for _Static_assert()
 
@@ -7033,7 +7000,6 @@ echo "jemalloc support  $jemalloc"
 echo "avx2 optimization $avx2_opt"
 echo "avx512f optimization $avx512f_opt"
 echo "replication support $replication"
-echo "VxHS block device $vxhs"
 echo "bochs support     $bochs"
 echo "cloop support     $cloop"
 echo "dmg support       $dmg"
@@ -7884,11 +7850,6 @@ elif test "$pthread_setname_np_wo_tid" = "yes" ; then
   echo "CONFIG_PTHREAD_SETNAME_NP_WO_TID=y" >> $config_host_mak
 fi
 
-if test "$vxhs" = "yes" ; then
-  echo "CONFIG_VXHS=y" >> $config_host_mak
-  echo "VXHS_LIBS=$vxhs_libs" >> $config_host_mak
-fi
-
 if test "$libpmem" = "yes" ; then
   echo "CONFIG_LIBPMEM=y" >> $config_host_mak
 fi
index 971b65be75838fc63f264e03b766d851e88d63c8..851dbdeb8abac77c7d2f6aaa22f47e7c4b1fc4e5 100644 (file)
@@ -618,3 +618,11 @@ to achieve the same fake NUMA effect or a properly configured
 New machine versions (since 5.1) will not accept the option but it will still
 work with old machine types. User can check the QAPI schema to see if the legacy
 option is supported by looking at MachineInfo::numa-mem-supported property.
+
+Block devices
+-------------
+
+VXHS backend (removed in 5.1)
+'''''''''''''''''''''''''''''
+
+The VXHS code does not compile since v2.12.0. It was removed in 5.1.
index 463ffd83daf73a145e59a527d890106ee89ac41c..ab7bf3c612b75a5bf6b5c09ad8ee0b86e7de12b5 100644 (file)
 #
 # Drivers that are supported in block device operations.
 #
-# @vxhs: Since 2.10
 # @throttle: Since 2.11
 # @nvme: Since 2.12
 # @copy-on-read: Since 3.0
             'qcow', 'qcow2', 'qed', 'quorum', 'raw', 'rbd',
             { 'name': 'replication', 'if': 'defined(CONFIG_REPLICATION)' },
             'sheepdog',
-            'ssh', 'throttle', 'vdi', 'vhdx', 'vmdk', 'vpc', 'vvfat', 'vxhs' ] }
+            'ssh', 'throttle', 'vdi', 'vhdx', 'vmdk', 'vpc', 'vvfat' ] }
 
 ##
 # @BlockdevOptionsFile:
   'base': 'BlockdevOptionsGenericFormat',
   'data': { '*offset': 'int', '*size': 'int' } }
 
-##
-# @BlockdevOptionsVxHS:
-#
-# Driver specific block device options for VxHS
-#
-# @vdisk-id: UUID of VxHS volume
-# @server: vxhs server IP, port
-# @tls-creds: TLS credentials ID
-#
-# Since: 2.10
-##
-{ 'struct': 'BlockdevOptionsVxHS',
-  'data': { 'vdisk-id': 'str',
-            'server': 'InetSocketAddressBase',
-            '*tls-creds': 'str' } }
-
 ##
 # @BlockdevOptionsThrottle:
 #
       'vhdx':       'BlockdevOptionsGenericFormat',
       'vmdk':       'BlockdevOptionsGenericCOWFormat',
       'vpc':        'BlockdevOptionsGenericFormat',
-      'vvfat':      'BlockdevOptionsVVFAT',
-      'vxhs':       'BlockdevOptionsVxHS'
+      'vvfat':      'BlockdevOptionsVVFAT'
   } }
 
 ##
index 585512bb296b73739169f6e32943fb6f4608f57a..3413e34f27d5b0c7d35aa3a9f45449a9b71c36bf 100755 (executable)
@@ -40,7 +40,6 @@ trap "_cleanup; exit \$status" 0 1 2 3 15
 # Any format supporting backing files
 _supported_fmt qcow qcow2 vmdk qed
 _supported_proto generic
-_unsupported_proto vxhs
 _unsupported_imgopts "subformat=monolithicFlat" "subformat=twoGbMaxExtentFlat" \
                      "subformat=streamOptimized"
 
index 2161a4b87a52333f9711690ec9755d9ea8a7b4c0..61d78c00a4c5975ea5f7c7a8ad385a3f5ad08a96 100755 (executable)
@@ -41,7 +41,6 @@ trap "_cleanup; exit \$status" 0 1 2 3 15
 # Any format supporting intenal snapshots
 _supported_fmt qcow2
 _supported_proto generic
-_unsupported_proto vxhs
 # Internal snapshots are (currently) impossible with refcount_bits=1,
 # and generally impossible with external data files
 _unsupported_imgopts 'refcount_bits=1[^0-9]' data_file
index 1dce478709142c5f600f261cc569dd87c182cf5f..68517821e875aed4384de94ee3b1cc3c2c66d8a4 100755 (executable)
@@ -38,7 +38,6 @@ trap "_cleanup; exit \$status" 0 1 2 3 15
 
 _supported_fmt qcow2
 _supported_proto generic
-_unsupported_proto vxhs
 # External data files do not support compressed clusters
 # (TODO: Consider writing a version for external data files that does
 #        not test compressed clusters)
index 5a7b0a49982e925cec51ba1d42b5bbbf75702251..80e5e5e5910eb8298aced88d8fa7994223d43d32 100755 (executable)
@@ -38,7 +38,6 @@ trap "_cleanup; exit \$status" 0 1 2 3 15
 
 _supported_fmt qcow2
 _supported_proto generic
-_unsupported_proto vxhs
 # At least OpenBSD doesn't seem to have truncate
 _supported_os Linux
 # qcow2.py does not work too well with external data files
index 77ad2aa13a06094f26d2c8991e4862bba62eaa9c..a7b365701c490b4b4894eb0c26a8d997620e7454 100755 (executable)
@@ -42,7 +42,6 @@ trap "_cleanup; exit \$status" 0 1 2 3 15
 
 _supported_fmt qcow2
 _supported_proto generic
-_unsupported_proto vxhs
 _supported_os Linux
 # We are going to use lazy-refcounts
 _unsupported_imgopts 'compat=0.10'
index 5162d21662488b230a27e29816b763a548f4bd4a..17fe1d6ed48d7fe42c0bc21054dc42ba259e94ac 100755 (executable)
@@ -38,7 +38,6 @@ trap "_cleanup; exit \$status" 0 1 2 3 15
 
 _supported_fmt qcow qcow2
 _supported_proto generic
-_unsupported_proto vxhs
 
 
 size=128M
index 7c69a6c3fa32ba460371eff760164a93d3d7450d..9c7878dd2dd30e8226c78125c72cefc43bce3f51 100755 (executable)
@@ -50,7 +50,6 @@ trap "_cleanup; exit \$status" 0 1 2 3 15
 
 _supported_fmt qcow2 qed
 _supported_proto generic
-_unsupported_proto vxhs
 # Copying files around with cp does not work with external data files
 _unsupported_imgopts data_file
 
index 3175968e2b6a1f463551550adbf0f11b583b9587..cf23742c5946fadf0f704bd0ad3143acf7393ea1 100755 (executable)
@@ -38,7 +38,6 @@ trap "_cleanup; exit \$status" 0 1 2 3 15
 
 _supported_fmt qcow qcow2
 _supported_proto generic
-_unsupported_proto vxhs
 
 
 size=128M
index 081eb120802154f90fcb603cd499ee555f7f1202..27da2a002397a61a8c86a78c1ec2310279a947b4 100755 (executable)
@@ -38,7 +38,6 @@ trap "_cleanup; exit \$status" 0 1 2 3 15
 
 _supported_fmt luks
 _supported_proto generic
-_unsupported_proto vxhs
 
 echo "== Create non-UTF8 secret =="
 echo -n -e '\x3a\x3c\x3b\xff' > non_utf8_secret
index 9c461cf76de8f9bebd54273349f804df7241d774..e0d8049012e7a1dd9e4c76b34929d2af41004a0b 100755 (executable)
@@ -272,7 +272,6 @@ image protocol options
     -nbd                test nbd
     -ssh                test ssh
     -nfs                test nfs
-    -vxhs               test vxhs
 
 other options
     -xdiff              graphical mode diff
@@ -383,11 +382,6 @@ testlist options
             xpand=false
             ;;
 
-        -vxhs)
-            IMGPROTO=vxhs
-            xpand=false
-            ;;
-
         -ssh)
             IMGPROTO=ssh
             xpand=false
@@ -646,10 +640,6 @@ if [ -z $QEMU_NBD_PROG ]; then
 fi
 export QEMU_NBD_PROG="$(type -p "$QEMU_NBD_PROG")"
 
-if [ -z "$QEMU_VXHS_PROG" ]; then
-    export QEMU_VXHS_PROG="$(set_prog_path qnio_server)"
-fi
-
 if [ -x "$build_iotests/socket_scm_helper" ]
 then
     export SOCKET_SCM_HELPER="$build_iotests/socket_scm_helper"
index 9b772245cdce4760ce20d42028619910a4b86b6d..c6912be009d9452d5c7da1bd8f6f5e92ceb34d19 100644 (file)
@@ -227,7 +227,6 @@ _filter_img_info()
         -e "s#$IMGFMT#IMGFMT#g" \
         -e 's#nbd://127.0.0.1:[0-9]\\+$#TEST_DIR/t.IMGFMT#g' \
         -e 's#nbd+unix:///\??socket=SOCK_DIR/nbd#TEST_DIR/t.IMGFMT#g' \
-        -e 's#json.*vdisk-id.*vxhs"}}#TEST_DIR/t.IMGFMT#' \
         -e "/encrypted: yes/d" \
         -e "/cluster_size: [0-9]\\+/d" \
         -e "/table_size: [0-9]\\+/d" \
index 7ac46edc1f9e8f5a64dcfcf451227db00d82b838..494490a2725ffcb456a93d44142eb83b70039646 100644 (file)
@@ -124,7 +124,6 @@ fi
 : ${VALGRIND_QEMU_IMG=$VALGRIND_QEMU}
 : ${VALGRIND_QEMU_IO=$VALGRIND_QEMU}
 : ${VALGRIND_QEMU_NBD=$VALGRIND_QEMU}
-: ${VALGRIND_QEMU_VXHS=$VALGRIND_QEMU}
 
 # The Valgrind own parameters may be set with
 # its environment variable VALGRIND_OPTS, e.g.
@@ -212,19 +211,6 @@ _qemu_nbd_wrapper()
     return $RETVAL
 }
 
-_qemu_vxhs_wrapper()
-{
-    local VALGRIND_LOGFILE="${TEST_DIR}"/$$.valgrind
-    (
-        echo $BASHPID > "${TEST_DIR}/qemu-vxhs.pid"
-        VALGRIND_QEMU="${VALGRIND_QEMU_VXHS}" _qemu_proc_exec "${VALGRIND_LOGFILE}" \
-            "$QEMU_VXHS_PROG" $QEMU_VXHS_OPTIONS "$@"
-    )
-    RETVAL=$?
-    _qemu_proc_valgrind_log "${VALGRIND_LOGFILE}" $RETVAL
-    return $RETVAL
-}
-
 # Valgrind bug #409141 https://bugs.kde.org/show_bug.cgi?id=409141
 # Until valgrind 3.16+ is ubiquitous, we must work around a hang in
 # valgrind when issuing sigkill. Disable valgrind for this invocation.
@@ -237,7 +223,6 @@ export QEMU=_qemu_wrapper
 export QEMU_IMG=_qemu_img_wrapper
 export QEMU_IO=_qemu_io_wrapper
 export QEMU_NBD=_qemu_nbd_wrapper
-export QEMU_VXHS=_qemu_vxhs_wrapper
 
 if [ "$IMGOPTSSYNTAX" = "true" ]; then
     DRIVER="driver=$IMGFMT"
@@ -279,9 +264,6 @@ else
         TEST_IMG_FILE=$TEST_DIR/t.$IMGFMT
         REMOTE_TEST_DIR="nfs://127.0.0.1$TEST_DIR"
         TEST_IMG="nfs://127.0.0.1$TEST_IMG_FILE"
-    elif [ "$IMGPROTO" = "vxhs" ]; then
-        TEST_IMG_FILE=$TEST_DIR/t.$IMGFMT
-        TEST_IMG="vxhs://127.0.0.1:9999/t.$IMGFMT"
     else
         TEST_IMG=$IMGPROTO:$TEST_DIR/t.$IMGFMT
     fi
@@ -436,12 +418,6 @@ _make_test_img()
         eval "$QEMU_NBD -v -t -k '$SOCK_DIR/nbd' -f $IMGFMT -e 42 -x '' $TEST_IMG_FILE >/dev/null &"
         sleep 1 # FIXME: qemu-nbd needs to be listening before we continue
     fi
-
-    # Start QNIO server on image directory for vxhs protocol
-    if [ $IMGPROTO = "vxhs" ]; then
-        eval "$QEMU_VXHS -d  $TEST_DIR > /dev/null &"
-        sleep 1 # Wait for server to come up.
-    fi
 }
 
 _rm_test_img()
@@ -468,15 +444,6 @@ _cleanup_test_img()
             _stop_nbd_server
             rm -f "$TEST_IMG_FILE"
             ;;
-        vxhs)
-            if [ -f "${TEST_DIR}/qemu-vxhs.pid" ]; then
-                local QEMU_VXHS_PID
-                read QEMU_VXHS_PID < "${TEST_DIR}/qemu-vxhs.pid"
-                kill ${QEMU_VXHS_PID} >/dev/null 2>&1
-                rm -f "${TEST_DIR}/qemu-vxhs.pid"
-            fi
-            rm -f "$TEST_IMG_FILE"
-            ;;
 
         file)
             _rm_test_img "$TEST_DIR/t.$IMGFMT"