Re: [Qemu-devel] [PATCHv4] block: add native support for NFS

[Fam Zheng <famz@redhat.com>]

On 2013年12月21日 00:04, Peter Lieven wrote:
> This patch adds native support for accessing images on NFS shares without
> the requirement to actually mount the entire NFS share on the host.
>
> NFS Images can simply be specified by an url of the form:
> nfs://<host>/<export>/<filename>
>
> For example:
> qemu-img create -f qcow2 nfs://10.0.0.1/qemu-images/test.qcow2
>
> You need LibNFS from Ronnie Sahlberg available at:
>     git://github.com/sahlberg/libnfs.git
> for this to work.
>
> During configure it is automatically probed for libnfs and support
> is enabled on-the-fly. You can forbid or enforce libnfs support
> with --disable-libnfs or --enable-libnfs respectively.
>
> Due to NFS restrictions you might need to execute your binaries
> as root, allow them to open priviledged ports (<1024) or specify
> insecure option on the NFS server.
>

What are the error messages like, if no privilege. Is root always 
required for this to work?

> LibNFS currently support NFS version 3 only.
>
> Signed-off-by: Peter Lieven <pl@kamp.de>
> ---
> v3->v4:
>   - finally added full implementation of bdrv_get_allocated_file_size [Stefan]
>   - removed trailing \n from error statements [Stefan]
>
> v2->v3:
>   - rebased the stefanha/block
>   - use pkg_config to check for libnfs (ignoring cflags which are broken in 1.8.0) [Stefan]
>   - fixed NFSClient declaration [Stefan]
>   - renamed Task variables to task [Stefan]
>   - renamed NFSTask to NFSRPC [Ronnie]
>   - do not update bs->total_sectors in nfs_co_writev [Stefan]
>   - return -ENOMEM on all async call failures [Stefan,Ronnie]
>   - fully implement ftruncate
>   - use util/uri.c for URL parsing [Stefan]
>   - reworked nfs_file_open_common to nfs_client_open which works on NFSClient [Stefan]
>   - added a comment ot the connect message that libnfs support NFSv3 only at the moment.
>   - DID NOT add full implementation of bdrv_get_allocated_file_size because
>     we are not in a coroutine context and I cannot do an async call here.
>     I could do a sync call if there would be a guarantee that no requests
>     are in flight. [Stefan]
>
> v1->v2:
>   - fixed block/Makefile.objs [Ronnie]
>   - do not always register a read handler [Ronnie]
>   - add support for reading beyond EOF [Fam]
>   - fixed struct and paramter naming [Fam]
>   - fixed overlong lines and whitespace errors [Fam]
>   - return return status from libnfs whereever possible [Fam]
>   - added comment why we set allocated_file_size to -ENOTSUP after write [Fam]
>   - avoid segfault when parsing filname [Fam]
>   - remove unused close_bh from NFSClient [Fam]
>   - avoid dividing and mutliplying total_size by BDRV_SECTOR_SIZE in nfs_file_create [Fam]
>
>   MAINTAINERS         |    5 +
>   block/Makefile.objs |    1 +
>   block/nfs.c         |  414 +++++++++++++++++++++++++++++++++++++++++++++++++++
>   configure           |   26 ++++
>   4 files changed, 446 insertions(+)
>   create mode 100644 block/nfs.c
>
> diff --git a/MAINTAINERS b/MAINTAINERS
> index a5ab8f8..09996ab 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -935,6 +935,11 @@ M: Peter Lieven <pl@kamp.de>
>   S: Supported
>   F: block/iscsi.c
>
> +NFS
> +M: Peter Lieven <pl@kamp.de>
> +S: Maintained
> +F: block/nfs.c
> +
>   SSH
>   M: Richard W.M. Jones <rjones@redhat.com>
>   S: Supported
> diff --git a/block/Makefile.objs b/block/Makefile.objs
> index 4e8c91e..e254a21 100644
> --- a/block/Makefile.objs
> +++ b/block/Makefile.objs
> @@ -12,6 +12,7 @@ block-obj-$(CONFIG_LINUX_AIO) += linux-aio.o
>   ifeq ($(CONFIG_POSIX),y)
>   block-obj-y += nbd.o nbd-client.o sheepdog.o
>   block-obj-$(CONFIG_LIBISCSI) += iscsi.o
> +block-obj-$(CONFIG_LIBNFS) += nfs.o
>   block-obj-$(CONFIG_CURL) += curl.o
>   block-obj-$(CONFIG_RBD) += rbd.o
>   block-obj-$(CONFIG_GLUSTERFS) += gluster.o
> diff --git a/block/nfs.c b/block/nfs.c
> new file mode 100644
> index 0000000..78fd8a1
> --- /dev/null
> +++ b/block/nfs.c
> @@ -0,0 +1,414 @@
> +/*
> + * QEMU Block driver for native access to files on NFS shares
> + *
> + * Copyright (c) 2013 Peter Lieven <pl@kamp.de>
> + *
> + * Permission is hereby granted, free of charge, to any person obtaining a copy
> + * of this software and associated documentation files (the "Software"), to deal
> + * in the Software without restriction, including without limitation the rights
> + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
> + * copies of the Software, and to permit persons to whom the Software is
> + * furnished to do so, subject to the following conditions:
> + *
> + * The above copyright notice and this permission notice shall be included in
> + * all copies or substantial portions of the Software.
> + *
> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
> + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
> + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
> + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
> + * THE SOFTWARE.
> + */
> +
> +#include "config-host.h"
> +
> +#include <poll.h>
> +#include "qemu-common.h"
> +#include "qemu/config-file.h"
> +#include "qemu/error-report.h"
> +#include "block/block_int.h"
> +#include "trace.h"
> +#include "qemu/iov.h"
> +#include "qemu/uri.h"
> +#include "sysemu/sysemu.h"
> +
> +#include <nfsc/libnfs-zdr.h>
> +#include <nfsc/libnfs.h>
> +#include <nfsc/libnfs-raw.h>
> +#include <nfsc/libnfs-raw-mount.h>
> +
> +typedef struct NFSClient {
> +    struct nfs_context *context;
> +    struct nfsfh *fh;
> +    int events;
> +    bool has_zero_init;
> +} NFSClient;
> +
> +typedef struct NFSRPC {
> +    int status;
> +    int complete;
> +    QEMUIOVector *iov;
> +    struct stat *st;
> +    Coroutine *co;
> +    QEMUBH *bh;
> +} NFSRPC;
> +
> +static void nfs_process_read(void *arg);
> +static void nfs_process_write(void *arg);
> +
> +static void nfs_set_events(NFSClient *client)
> +{
> +    int ev = nfs_which_events(client->context);
> +    if (ev != client->events) {
> +        qemu_aio_set_fd_handler(nfs_get_fd(client->context),
> +                      (ev & POLLIN) ? nfs_process_read : NULL,
> +                      (ev & POLLOUT) ? nfs_process_write : NULL,
> +                      client);
> +
> +    }
> +    client->events = ev;
> +}
> +
> +static void nfs_process_read(void *arg)
> +{
> +    NFSClient *client = arg;
> +    nfs_service(client->context, POLLIN);
> +    nfs_set_events(client);
> +}
> +
> +static void nfs_process_write(void *arg)
> +{
> +    NFSClient *client = arg;
> +    nfs_service(client->context, POLLOUT);
> +    nfs_set_events(client);
> +}
> +
> +static void nfs_co_init_task(NFSClient *client, NFSRPC *task)
> +{
> +    *task = (NFSRPC) {
> +        .co         = qemu_coroutine_self(),
> +    };
> +}
> +
> +static void nfs_co_generic_bh_cb(void *opaque)
> +{
> +    NFSRPC *task = opaque;
> +    qemu_bh_delete(task->bh);
> +    qemu_coroutine_enter(task->co, NULL);
> +}
> +
> +static void
> +nfs_co_generic_cb(int status, struct nfs_context *nfs, void *data,
> +                  void *private_data)
> +{
> +    NFSRPC *task = private_data;
> +    task->complete = 1;
> +    task->status = status;
> +    if (task->status > 0 && task->iov) {
> +        if (task->status <= task->iov->size) {
> +            qemu_iovec_from_buf(task->iov, 0, data, task->status);
> +        } else {
> +            task->status = -EIO;
> +        }
> +    }
> +    if (task->status == 0 && task->st) {
> +        memcpy(task->st, data, sizeof(struct stat));
> +    }
> +    if (task->co) {
> +        task->bh = qemu_bh_new(nfs_co_generic_bh_cb, task);
> +        qemu_bh_schedule(task->bh);
> +    }
> +}
> +
> +static int coroutine_fn nfs_co_readv(BlockDriverState *bs,
> +                                     int64_t sector_num, int nb_sectors,
> +                                     QEMUIOVector *iov)
> +{
> +    NFSClient *client = bs->opaque;
> +    NFSRPC task;
> +
> +    nfs_co_init_task(client, &task);
> +    task.iov = iov;
> +
> +    if (nfs_pread_async(client->context, client->fh,
> +                        sector_num * BDRV_SECTOR_SIZE,
> +                        nb_sectors * BDRV_SECTOR_SIZE,
> +                        nfs_co_generic_cb, &task) != 0) {
> +        return -ENOMEM;
> +    }
> +
> +    while (!task.complete) {
> +        nfs_set_events(client);
> +        qemu_coroutine_yield();
> +    }
> +
> +    if (task.status < 0) {
> +        return task.status;
> +    }
> +
> +    return 0;
> +}
> +
> +static int coroutine_fn nfs_co_writev(BlockDriverState *bs,
> +                                        int64_t sector_num, int nb_sectors,
> +                                        QEMUIOVector *iov)
> +{
> +    NFSClient *client = bs->opaque;
> +    NFSRPC task;
> +    char *buf = NULL;
> +
> +    nfs_co_init_task(client, &task);
> +
> +    buf = g_malloc(nb_sectors * BDRV_SECTOR_SIZE);
> +    qemu_iovec_to_buf(iov, 0, buf, nb_sectors * BDRV_SECTOR_SIZE);
> +
> +    if (nfs_pwrite_async(client->context, client->fh,
> +                         sector_num * BDRV_SECTOR_SIZE,
> +                         nb_sectors * BDRV_SECTOR_SIZE,
> +                         buf, nfs_co_generic_cb, &task) != 0) {
> +        g_free(buf);
> +        return -ENOMEM;
> +    }
> +
> +    while (!task.complete) {
> +        nfs_set_events(client);
> +        qemu_coroutine_yield();
> +    }
> +
> +    g_free(buf);
> +
> +    if (task.status != nb_sectors * BDRV_SECTOR_SIZE) {
> +        return task.status < 0 ? task.status : -EIO;
> +    }
> +
> +    return 0;
> +}
> +
> +static int coroutine_fn nfs_co_flush(BlockDriverState *bs)
> +{
> +    NFSClient *client = bs->opaque;
> +    NFSRPC task;
> +
> +    nfs_co_init_task(client, &task);
> +
> +    if (nfs_fsync_async(client->context, client->fh, nfs_co_generic_cb,
> +                        &task) != 0) {
> +        return -ENOMEM;
> +    }
> +
> +    while (!task.complete) {
> +        nfs_set_events(client);
> +        qemu_coroutine_yield();
> +    }
> +
> +    return task.status;
> +}
> +
> +static QemuOptsList runtime_opts = {
> +    .name = "nfs",
> +    .head = QTAILQ_HEAD_INITIALIZER(runtime_opts.head),
> +    .desc = {
> +        {
> +            .name = "filename",
> +            .type = QEMU_OPT_STRING,
> +            .help = "URL to the NFS file",
> +        },
> +        { /* end of list */ }
> +    },
> +};
> +
> +static void nfs_client_close(NFSClient *client)
> +{
> +    if (client->context) {
> +        if (client->fh) {
> +            nfs_close(client->context, client->fh);
> +        }
> +        qemu_aio_set_fd_handler(nfs_get_fd(client->context), NULL, NULL, NULL);
> +        nfs_destroy_context(client->context);
> +    }
> +    memset(client, 0, sizeof(NFSClient));
> +}
> +
> +static void nfs_file_close(BlockDriverState *bs)
> +{
> +    NFSClient *client = bs->opaque;
> +    nfs_client_close(client);
> +}
> +
> +static int64_t nfs_client_open(NFSClient *client, const char *filename,
> +                               int flags, Error **errp)
> +{
> +    int ret = -EINVAL;
> +    URI *uri;
> +    char *file = NULL, *strp = NULL;
> +    struct stat st;
> +
> +    uri = uri_parse(filename);
> +    if (!uri) {
> +        error_setg(errp, "Invalid URL specified.");

Could drop the period in the end of error message, but I don't mind 
merging it as is:

Reviewed-by: Fam Zheng <famz@redhat.com>

Thanks

> +        goto fail;
> +    }
> +    strp = strrchr(uri->path, '/');
> +    if (strp == NULL) {
> +        error_setg(errp, "Invalid URL specified.");
> +        goto fail;
> +    }
> +    file = g_strdup(strp);
> +    *strp = 0;
> +
> +    client->context = nfs_init_context();
> +    if (client->context == NULL) {
> +        error_setg(errp, "Failed to init NFS context");
> +        goto fail;
> +    }
> +
> +    ret = nfs_mount(client->context, uri->server, uri->path);
> +    if (ret < 0) {
> +        error_setg(errp, "Failed to mount nfs share: %s",
> +                    nfs_get_error(client->context));
> +        goto fail;
> +    }
> +
> +    if (flags & O_CREAT) {
> +        ret = nfs_creat(client->context, file, 0600, &client->fh);
> +        if (ret < 0) {
> +            error_setg(errp, "Failed to create file: %s",
> +                             nfs_get_error(client->context));
> +            goto fail;
> +        }
> +    } else {
> +        ret = nfs_open(client->context, file, flags, &client->fh);
> +        if (ret < 0) {
> +            error_setg(errp, "Failed to open file : %s",
> +                       nfs_get_error(client->context));
> +            goto fail;
> +        }
> +    }
> +
> +    ret = nfs_fstat(client->context, client->fh, &st);
> +    if (ret < 0) {
> +        error_setg(errp, "Failed to fstat file: %s",
> +                   nfs_get_error(client->context));
> +        goto fail;
> +    }
> +
> +    ret = DIV_ROUND_UP(st.st_size, BDRV_SECTOR_SIZE);
> +    client->has_zero_init = S_ISREG(st.st_mode);
> +    goto out;
> +fail:
> +    nfs_client_close(client);
> +out:
> +    uri_free(uri);
> +    g_free(file);
> +    return ret;
> +}
> +
> +static int nfs_file_open(BlockDriverState *bs, QDict *options, int flags,
> +                         Error **errp) {
> +    NFSClient *client = bs->opaque;
> +    int64_t ret;
> +    QemuOpts *opts;
> +    Error *local_err = NULL;
> +
> +    opts = qemu_opts_create_nofail(&runtime_opts);
> +    qemu_opts_absorb_qdict(opts, options, &local_err);
> +    if (error_is_set(&local_err)) {
> +        qerror_report_err(local_err);
> +        error_free(local_err);
> +        return -EINVAL;
> +    }
> +    ret = nfs_client_open(client, qemu_opt_get(opts, "filename"),
> +                          (flags & BDRV_O_RDWR) ? O_RDWR : O_RDONLY,
> +                          errp);
> +    if (ret < 0) {
> +        return ret;
> +    }
> +    bs->total_sectors = ret;
> +    return 0;
> +}
> +
> +static int nfs_file_create(const char *filename, QEMUOptionParameter *options,
> +                           Error **errp)
> +{
> +    int ret = 0;
> +    int64_t total_size = 0;
> +    NFSClient *client = g_malloc0(sizeof(NFSClient));
> +
> +    /* Read out options */
> +    while (options && options->name) {
> +        if (!strcmp(options->name, "size")) {
> +            total_size = options->value.n;
> +        }
> +        options++;
> +    }
> +
> +    ret = nfs_client_open(client, filename, O_CREAT, errp);
> +    if (ret < 0) {
> +        goto out;
> +    }
> +    ret = nfs_ftruncate(client->context, client->fh, total_size);
> +out:
> +    nfs_client_close(client);
> +    g_free(client);
> +    return ret;
> +}
> +
> +static int nfs_has_zero_init(BlockDriverState *bs)
> +{
> +    NFSClient *client = bs->opaque;
> +    return client->has_zero_init;
> +}
> +
> +static int64_t nfs_get_allocated_file_size(BlockDriverState *bs)
> +{
> +    NFSClient *client = bs->opaque;
> +    NFSRPC task = {0};
> +    struct stat st;
> +
> +    task.st = &st;
> +    if (nfs_fstat_async(client->context, client->fh, nfs_co_generic_cb,
> +                        &task) != 0) {
> +        return -ENOMEM;
> +    }
> +
> +    while (!task.complete) {
> +        nfs_set_events(client);
> +        qemu_aio_wait();
> +    }
> +
> +    return (task.status < 0 ? task.status : st.st_blocks * st.st_blksize);
> +}
> +
> +static int nfs_file_truncate(BlockDriverState *bs, int64_t offset)
> +{
> +    NFSClient *client = bs->opaque;
> +    return nfs_ftruncate(client->context, client->fh, offset);
> +}
> +
> +static BlockDriver bdrv_nfs = {
> +    .format_name     = "nfs",
> +    .protocol_name   = "nfs",
> +
> +    .instance_size   = sizeof(NFSClient),
> +    .bdrv_needs_filename = true,
> +    .bdrv_has_zero_init = nfs_has_zero_init,
> +    .bdrv_get_allocated_file_size = nfs_get_allocated_file_size,
> +    .bdrv_truncate = nfs_file_truncate,
> +
> +    .bdrv_file_open  = nfs_file_open,
> +    .bdrv_close      = nfs_file_close,
> +    .bdrv_create     = nfs_file_create,
> +
> +    .bdrv_co_readv         = nfs_co_readv,
> +    .bdrv_co_writev        = nfs_co_writev,
> +    .bdrv_co_flush_to_disk = nfs_co_flush,
> +};
> +
> +static void nfs_block_init(void)
> +{
> +    bdrv_register(&bdrv_nfs);
> +}
> +
> +block_init(nfs_block_init);
> diff --git a/configure b/configure
> index d97556c..bae1d2b 100755
> --- a/configure
> +++ b/configure
> @@ -251,6 +251,7 @@ vss_win32_sdk=""
>   win_sdk="no"
>   want_tools="yes"
>   libiscsi=""
> +libnfs=""
>   coroutine=""
>   coroutine_pool=""
>   seccomp=""
> @@ -839,6 +840,10 @@ for opt do
>     ;;
>     --enable-libiscsi) libiscsi="yes"
>     ;;
> +  --disable-libnfs) libnfs="no"
> +  ;;
> +  --enable-libnfs) libnfs="yes"
> +  ;;
>     --enable-profiler) profiler="yes"
>     ;;
>     --disable-cocoa) cocoa="no"
> @@ -1210,6 +1215,8 @@ echo "  --enable-spice           enable spice"
>   echo "  --enable-rbd             enable building the rados block device (rbd)"
>   echo "  --disable-libiscsi       disable iscsi support"
>   echo "  --enable-libiscsi        enable iscsi support"
> +echo "  --disable-libnfs         disable nfs support"
> +echo "  --enable-libnfs          enable nfs support"
>   echo "  --disable-smartcard-nss  disable smartcard nss support"
>   echo "  --enable-smartcard-nss   enable smartcard nss support"
>   echo "  --disable-libusb         disable libusb (for usb passthrough)"
> @@ -3596,6 +3603,20 @@ elif test "$debug" = "no" ; then
>     CFLAGS="-O2 -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=2 $CFLAGS"
>   fi
>
> +##########################################
> +# Do we have libnfs
> +if test "$libnfs" != "no" ; then
> +  if $pkg_config --atleast-version=1.8.0 libnfs; then
> +    libnfs="yes"
> +    libnfs_libs=$($pkg_config --libs libnfs)
> +    LIBS="$LIBS $libnfs_libs"
> +  else
> +    if test "$libnfs" = "yes" ; then
> +      feature_not_found "libnfs"
> +    fi
> +    libnfs="no"
> +  fi
> +fi
>
>   # Disable zero malloc errors for official releases unless explicitly told to
>   # enable/disable
> @@ -3825,6 +3846,7 @@ echo "libiscsi support  $libiscsi (1.4.0)"
>   else
>   echo "libiscsi support  $libiscsi"
>   fi
> +echo "libnfs support    $libnfs"
>   echo "build guest agent $guest_agent"
>   echo "QGA VSS support   $guest_agent_with_vss"
>   echo "seccomp support   $seccomp"
> @@ -4161,6 +4183,10 @@ if test "$libiscsi" = "yes" ; then
>     fi
>   fi
>
> +if test "$libnfs" = "yes" ; then
> +  echo "CONFIG_LIBNFS=y" >> $config_host_mak
> +fi
> +
>   if test "$seccomp" = "yes"; then
>     echo "CONFIG_SECCOMP=y" >> $config_host_mak
>   fi
>