| Message ID | 20210628075545.702106-2-vshankar@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | ceph: new mount device syntax | expand |
[ As I said, I didn't fully reviewed this patch. Just sending out a few comments. ] On Mon, Jun 28, 2021 at 01:25:42PM +0530, Venky Shankar wrote: > Old mount device syntax (source) has the following problems: > > - mounts to the same cluster but with different fsnames > and/or creds have identical device string which can > confuse xfstests. > > - Userspace mount helper tool resolves monitor addresses > and fill in mon addrs automatically, but that means the > device shown in /proc/mounts is different than what was > used for mounting. > > New device syntax is as follows: > > cephuser@fsid.mycephfs2=/path > > Note, there is no "monitor address" in the device string. > That gets passed in as mount option. This keeps the device > string same when monitor addresses change (on remounts). > > Also note that the userspace mount helper tool is backward > compatible. I.e., the mount helper will fallback to using > old syntax after trying to mount with the new syntax. > > Signed-off-by: Venky Shankar <vshankar@redhat.com> > --- > fs/ceph/super.c | 117 +++++++++++++++++++++++++++++++++++++++++++----- > fs/ceph/super.h | 3 ++ > 2 files changed, 110 insertions(+), 10 deletions(-) > > diff --git a/fs/ceph/super.c b/fs/ceph/super.c > index 9b1b7f4cfdd4..950a28ad9c59 100644 > --- a/fs/ceph/super.c > +++ b/fs/ceph/super.c > @@ -145,6 +145,7 @@ enum { > Opt_mds_namespace, > Opt_recover_session, > Opt_source, > + Opt_mon_addr, > /* string args above */ > Opt_dirstat, > Opt_rbytes, > @@ -196,6 +197,7 @@ static const struct fs_parameter_spec ceph_mount_parameters[] = { > fsparam_u32 ("rsize", Opt_rsize), > fsparam_string ("snapdirname", Opt_snapdirname), > fsparam_string ("source", Opt_source), > + fsparam_string ("mon_addr", Opt_mon_addr), > fsparam_u32 ("wsize", Opt_wsize), > fsparam_flag_no ("wsync", Opt_wsync), > {} > @@ -226,10 +228,68 @@ static void canonicalize_path(char *path) > path[j] = '\0'; > } > > +static int ceph_parse_old_source(const char *dev_name, const char *dev_name_end, > + struct fs_context *fc) > +{ > + int r; > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > + struct ceph_mount_options *fsopt = pctx->opts; > + > + if (*dev_name_end != ':') > + return invalfc(fc, "separator ':' missing in source"); > + > + r = ceph_parse_mon_ips(dev_name, dev_name_end - dev_name, > + pctx->copts, fc->log.log); > + if (r) > + return r; > + > + fsopt->new_dev_syntax = false; > + return 0; > +} > + > +static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end, > + struct fs_context *fc) > +{ > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > + struct ceph_mount_options *fsopt = pctx->opts; > + char *fsid_start, *fs_name_start; > + > + if (*dev_name_end != '=') > + return invalfc(fc, "separator '=' missing in source"); An annoying thing is that we'll always see this error message when falling back to the old_source method. (Also, is there a good reason for using '=' instead of ':'? I probably missed this discussion somewhere else already...) > + > + fsid_start = strchr(dev_name, '@'); > + if (!fsid_start) > + return invalfc(fc, "missing cluster fsid"); > + ++fsid_start; /* start of cluster fsid */ > + > + fs_name_start = strchr(fsid_start, '.'); > + if (!fs_name_start) > + return invalfc(fc, "missing file system name"); > + > + ++fs_name_start; /* start of file system name */ > + fsopt->mds_namespace = kstrndup(fs_name_start, > + dev_name_end - fs_name_start, GFP_KERNEL); > + if (!fsopt->mds_namespace) > + return -ENOMEM; > + dout("file system (mds namespace) '%s'\n", fsopt->mds_namespace); > + > + fsopt->new_dev_syntax = true; > + return 0; > +} > + > /* > - * Parse the source parameter. Distinguish the server list from the path. > + * Parse the source parameter for new device format. Distinguish the device > + * spec from the path. Try parsing new device format and fallback to old > + * format if needed. > * > - * The source will look like: > + * New device syntax will looks like: > + * <device_spec>=/<path> > + * where > + * <device_spec> is name@fsid.fsname > + * <path> is optional, but if present must begin with '/' > + * (monitor addresses are passed via mount option) > + * > + * Old device syntax is: > * <server_spec>[,<server_spec>...]:[<path>] > * where > * <server_spec> is <ip>[:<port>] > @@ -262,22 +322,48 @@ static int ceph_parse_source(struct fs_parameter *param, struct fs_context *fc) > dev_name_end = dev_name + strlen(dev_name); > } > > - dev_name_end--; /* back up to ':' separator */ > - if (dev_name_end < dev_name || *dev_name_end != ':') > - return invalfc(fc, "No path or : separator in source"); > + dev_name_end--; /* back up to separator */ > + if (dev_name_end < dev_name) > + return invalfc(fc, "path missing in source"); > > dout("device name '%.*s'\n", (int)(dev_name_end - dev_name), dev_name); > if (fsopt->server_path) > dout("server path '%s'\n", fsopt->server_path); > > - ret = ceph_parse_mon_ips(param->string, dev_name_end - dev_name, > - pctx->copts, fc->log.log); > + dout("trying new device syntax"); > + ret = ceph_parse_new_source(dev_name, dev_name_end, fc); > + if (ret == 0) > + goto done; > + > + dout("trying old device syntax"); > + ret = ceph_parse_old_source(dev_name, dev_name_end, fc); > if (ret) > - return ret; > + goto out; > > + done: > fc->source = param->string; > param->string = NULL; > - return 0; > + out: > + return ret; > +} > + > +static int ceph_parse_mon_addr(struct fs_parameter *param, > + struct fs_context *fc) > +{ > + int r; > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > + struct ceph_mount_options *fsopt = pctx->opts; > + > + kfree(fsopt->mon_addr); > + fsopt->mon_addr = kstrdup(param->string, GFP_KERNEL); > + if (!fsopt->mon_addr) > + return -ENOMEM; > + > + strreplace(param->string, '/', ','); > + r = ceph_parse_mon_ips(param->string, strlen(param->string), > + pctx->copts, fc->log.log); > + param->string = NULL; I think this will result in a memory leak. Why don't we simply set fsopt->mon_addr to param->string instead of kstrdup'ing it? Cheers, -- Luís > + return r; > } > > static int ceph_parse_mount_param(struct fs_context *fc, > @@ -322,6 +408,8 @@ static int ceph_parse_mount_param(struct fs_context *fc, > if (fc->source) > return invalfc(fc, "Multiple sources specified"); > return ceph_parse_source(param, fc); > + case Opt_mon_addr: > + return ceph_parse_mon_addr(param, fc); > case Opt_wsize: > if (result.uint_32 < PAGE_SIZE || > result.uint_32 > CEPH_MAX_WRITE_SIZE) > @@ -473,6 +561,7 @@ static void destroy_mount_options(struct ceph_mount_options *args) > kfree(args->mds_namespace); > kfree(args->server_path); > kfree(args->fscache_uniq); > + kfree(args->mon_addr); > kfree(args); > } > > @@ -516,6 +605,10 @@ static int compare_mount_options(struct ceph_mount_options *new_fsopt, > if (ret) > return ret; > > + ret = strcmp_null(fsopt1->mon_addr, fsopt2->mon_addr); > + if (ret) > + return ret; > + > return ceph_compare_options(new_opt, fsc->client); > } > > @@ -571,9 +664,13 @@ static int ceph_show_options(struct seq_file *m, struct dentry *root) > if ((fsopt->flags & CEPH_MOUNT_OPT_NOCOPYFROM) == 0) > seq_puts(m, ",copyfrom"); > > - if (fsopt->mds_namespace) > + /* dump mds_namespace when old device syntax is in use */ > + if (fsopt->mds_namespace && !fsopt->new_dev_syntax) > seq_show_option(m, "mds_namespace", fsopt->mds_namespace); > > + if (fsopt->mon_addr) > + seq_printf(m, ",mon_addr=%s", fsopt->mon_addr); > + > if (fsopt->flags & CEPH_MOUNT_OPT_CLEANRECOVER) > seq_show_option(m, "recover_session", "clean"); > > diff --git a/fs/ceph/super.h b/fs/ceph/super.h > index 839e6b0239ee..557348ff3203 100644 > --- a/fs/ceph/super.h > +++ b/fs/ceph/super.h > @@ -88,6 +88,8 @@ struct ceph_mount_options { > unsigned int max_readdir; /* max readdir result (entries) */ > unsigned int max_readdir_bytes; /* max readdir result (bytes) */ > > + bool new_dev_syntax; > + > /* > * everything above this point can be memcmp'd; everything below > * is handled in compare_mount_options() > @@ -97,6 +99,7 @@ struct ceph_mount_options { > char *mds_namespace; /* default NULL */ > char *server_path; /* default NULL (means "/") */ > char *fscache_uniq; /* default NULL */ > + char *mon_addr; > }; > > struct ceph_fs_client { > -- > 2.27.0 >
On Tue, Jun 29, 2021 at 5:02 PM Luis Henriques <lhenriques@suse.de> wrote: > > [ As I said, I didn't fully reviewed this patch. Just sending out a few > comments. ] > > On Mon, Jun 28, 2021 at 01:25:42PM +0530, Venky Shankar wrote: > > Old mount device syntax (source) has the following problems: > > > > - mounts to the same cluster but with different fsnames > > and/or creds have identical device string which can > > confuse xfstests. > > > > - Userspace mount helper tool resolves monitor addresses > > and fill in mon addrs automatically, but that means the > > device shown in /proc/mounts is different than what was > > used for mounting. > > > > New device syntax is as follows: > > > > cephuser@fsid.mycephfs2=/path > > > > Note, there is no "monitor address" in the device string. > > That gets passed in as mount option. This keeps the device > > string same when monitor addresses change (on remounts). > > > > Also note that the userspace mount helper tool is backward > > compatible. I.e., the mount helper will fallback to using > > old syntax after trying to mount with the new syntax. > > > > Signed-off-by: Venky Shankar <vshankar@redhat.com> > > --- > > fs/ceph/super.c | 117 +++++++++++++++++++++++++++++++++++++++++++----- > > fs/ceph/super.h | 3 ++ > > 2 files changed, 110 insertions(+), 10 deletions(-) > > > > diff --git a/fs/ceph/super.c b/fs/ceph/super.c > > index 9b1b7f4cfdd4..950a28ad9c59 100644 > > --- a/fs/ceph/super.c > > +++ b/fs/ceph/super.c > > @@ -145,6 +145,7 @@ enum { > > Opt_mds_namespace, > > Opt_recover_session, > > Opt_source, > > + Opt_mon_addr, > > /* string args above */ > > Opt_dirstat, > > Opt_rbytes, > > @@ -196,6 +197,7 @@ static const struct fs_parameter_spec ceph_mount_parameters[] = { > > fsparam_u32 ("rsize", Opt_rsize), > > fsparam_string ("snapdirname", Opt_snapdirname), > > fsparam_string ("source", Opt_source), > > + fsparam_string ("mon_addr", Opt_mon_addr), > > fsparam_u32 ("wsize", Opt_wsize), > > fsparam_flag_no ("wsync", Opt_wsync), > > {} > > @@ -226,10 +228,68 @@ static void canonicalize_path(char *path) > > path[j] = '\0'; > > } > > > > +static int ceph_parse_old_source(const char *dev_name, const char *dev_name_end, > > + struct fs_context *fc) > > +{ > > + int r; > > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > > + struct ceph_mount_options *fsopt = pctx->opts; > > + > > + if (*dev_name_end != ':') > > + return invalfc(fc, "separator ':' missing in source"); > > + > > + r = ceph_parse_mon_ips(dev_name, dev_name_end - dev_name, > > + pctx->copts, fc->log.log); > > + if (r) > > + return r; > > + > > + fsopt->new_dev_syntax = false; > > + return 0; > > +} > > + > > +static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end, > > + struct fs_context *fc) > > +{ > > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > > + struct ceph_mount_options *fsopt = pctx->opts; > > + char *fsid_start, *fs_name_start; > > + > > + if (*dev_name_end != '=') > > + return invalfc(fc, "separator '=' missing in source"); > > An annoying thing is that we'll always see this error message when falling > back to the old_source method. > > (Also, is there a good reason for using '=' instead of ':'? I probably > missed this discussion somewhere else already...) > > > + > > + fsid_start = strchr(dev_name, '@'); > > + if (!fsid_start) > > + return invalfc(fc, "missing cluster fsid"); > > + ++fsid_start; /* start of cluster fsid */ > > + > > + fs_name_start = strchr(fsid_start, '.'); > > + if (!fs_name_start) > > + return invalfc(fc, "missing file system name"); > > + > > + ++fs_name_start; /* start of file system name */ > > + fsopt->mds_namespace = kstrndup(fs_name_start, > > + dev_name_end - fs_name_start, GFP_KERNEL); > > + if (!fsopt->mds_namespace) > > + return -ENOMEM; > > + dout("file system (mds namespace) '%s'\n", fsopt->mds_namespace); > > + > > + fsopt->new_dev_syntax = true; > > + return 0; > > +} > > + > > /* > > - * Parse the source parameter. Distinguish the server list from the path. > > + * Parse the source parameter for new device format. Distinguish the device > > + * spec from the path. Try parsing new device format and fallback to old > > + * format if needed. > > * > > - * The source will look like: > > + * New device syntax will looks like: > > + * <device_spec>=/<path> > > + * where > > + * <device_spec> is name@fsid.fsname > > + * <path> is optional, but if present must begin with '/' > > + * (monitor addresses are passed via mount option) > > + * > > + * Old device syntax is: > > * <server_spec>[,<server_spec>...]:[<path>] > > * where > > * <server_spec> is <ip>[:<port>] > > @@ -262,22 +322,48 @@ static int ceph_parse_source(struct fs_parameter *param, struct fs_context *fc) > > dev_name_end = dev_name + strlen(dev_name); > > } > > > > - dev_name_end--; /* back up to ':' separator */ > > - if (dev_name_end < dev_name || *dev_name_end != ':') > > - return invalfc(fc, "No path or : separator in source"); > > + dev_name_end--; /* back up to separator */ > > + if (dev_name_end < dev_name) > > + return invalfc(fc, "path missing in source"); > > > > dout("device name '%.*s'\n", (int)(dev_name_end - dev_name), dev_name); > > if (fsopt->server_path) > > dout("server path '%s'\n", fsopt->server_path); > > > > - ret = ceph_parse_mon_ips(param->string, dev_name_end - dev_name, > > - pctx->copts, fc->log.log); > > + dout("trying new device syntax"); > > + ret = ceph_parse_new_source(dev_name, dev_name_end, fc); > > + if (ret == 0) > > + goto done; > > + > > + dout("trying old device syntax"); > > + ret = ceph_parse_old_source(dev_name, dev_name_end, fc); > > if (ret) > > - return ret; > > + goto out; > > > > + done: > > fc->source = param->string; > > param->string = NULL; > > - return 0; > > + out: > > + return ret; > > +} > > + > > +static int ceph_parse_mon_addr(struct fs_parameter *param, > > + struct fs_context *fc) > > +{ > > + int r; > > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > > + struct ceph_mount_options *fsopt = pctx->opts; > > + > > + kfree(fsopt->mon_addr); > > + fsopt->mon_addr = kstrdup(param->string, GFP_KERNEL); > > + if (!fsopt->mon_addr) > > + return -ENOMEM; > > + > > + strreplace(param->string, '/', ','); > > + r = ceph_parse_mon_ips(param->string, strlen(param->string), > > + pctx->copts, fc->log.log); > > + param->string = NULL; > > I think this will result in a memory leak. Why don't we simply set > fsopt->mon_addr to param->string instead of kstrdup'ing it? Sure. As discussed over irc, will do away with the alloc. > > Cheers, > -- > Luís > > > + return r; > > } > > > > static int ceph_parse_mount_param(struct fs_context *fc, > > @@ -322,6 +408,8 @@ static int ceph_parse_mount_param(struct fs_context *fc, > > if (fc->source) > > return invalfc(fc, "Multiple sources specified"); > > return ceph_parse_source(param, fc); > > + case Opt_mon_addr: > > + return ceph_parse_mon_addr(param, fc); > > case Opt_wsize: > > if (result.uint_32 < PAGE_SIZE || > > result.uint_32 > CEPH_MAX_WRITE_SIZE) > > @@ -473,6 +561,7 @@ static void destroy_mount_options(struct ceph_mount_options *args) > > kfree(args->mds_namespace); > > kfree(args->server_path); > > kfree(args->fscache_uniq); > > + kfree(args->mon_addr); > > kfree(args); > > } > > > > @@ -516,6 +605,10 @@ static int compare_mount_options(struct ceph_mount_options *new_fsopt, > > if (ret) > > return ret; > > > > + ret = strcmp_null(fsopt1->mon_addr, fsopt2->mon_addr); > > + if (ret) > > + return ret; > > + > > return ceph_compare_options(new_opt, fsc->client); > > } > > > > @@ -571,9 +664,13 @@ static int ceph_show_options(struct seq_file *m, struct dentry *root) > > if ((fsopt->flags & CEPH_MOUNT_OPT_NOCOPYFROM) == 0) > > seq_puts(m, ",copyfrom"); > > > > - if (fsopt->mds_namespace) > > + /* dump mds_namespace when old device syntax is in use */ > > + if (fsopt->mds_namespace && !fsopt->new_dev_syntax) > > seq_show_option(m, "mds_namespace", fsopt->mds_namespace); > > > > + if (fsopt->mon_addr) > > + seq_printf(m, ",mon_addr=%s", fsopt->mon_addr); > > + > > if (fsopt->flags & CEPH_MOUNT_OPT_CLEANRECOVER) > > seq_show_option(m, "recover_session", "clean"); > > > > diff --git a/fs/ceph/super.h b/fs/ceph/super.h > > index 839e6b0239ee..557348ff3203 100644 > > --- a/fs/ceph/super.h > > +++ b/fs/ceph/super.h > > @@ -88,6 +88,8 @@ struct ceph_mount_options { > > unsigned int max_readdir; /* max readdir result (entries) */ > > unsigned int max_readdir_bytes; /* max readdir result (bytes) */ > > > > + bool new_dev_syntax; > > + > > /* > > * everything above this point can be memcmp'd; everything below > > * is handled in compare_mount_options() > > @@ -97,6 +99,7 @@ struct ceph_mount_options { > > char *mds_namespace; /* default NULL */ > > char *server_path; /* default NULL (means "/") */ > > char *fscache_uniq; /* default NULL */ > > + char *mon_addr; > > }; > > > > struct ceph_fs_client { > > -- > > 2.27.0 > > >
On Tue, 2021-06-29 at 19:24 +0530, Venky Shankar wrote: > On Tue, Jun 29, 2021 at 5:02 PM Luis Henriques <lhenriques@suse.de> wrote: > > > > [ As I said, I didn't fully reviewed this patch. Just sending out a few > > comments. ] > > > > On Mon, Jun 28, 2021 at 01:25:42PM +0530, Venky Shankar wrote: > > > Old mount device syntax (source) has the following problems: > > > > > > - mounts to the same cluster but with different fsnames > > > and/or creds have identical device string which can > > > confuse xfstests. > > > > > > - Userspace mount helper tool resolves monitor addresses > > > and fill in mon addrs automatically, but that means the > > > device shown in /proc/mounts is different than what was > > > used for mounting. > > > > > > New device syntax is as follows: > > > > > > cephuser@fsid.mycephfs2=/path > > > > > > Note, there is no "monitor address" in the device string. > > > That gets passed in as mount option. This keeps the device > > > string same when monitor addresses change (on remounts). > > > > > > Also note that the userspace mount helper tool is backward > > > compatible. I.e., the mount helper will fallback to using > > > old syntax after trying to mount with the new syntax. > > > > > > Signed-off-by: Venky Shankar <vshankar@redhat.com> > > > --- > > > fs/ceph/super.c | 117 +++++++++++++++++++++++++++++++++++++++++++----- > > > fs/ceph/super.h | 3 ++ > > > 2 files changed, 110 insertions(+), 10 deletions(-) > > > > > > diff --git a/fs/ceph/super.c b/fs/ceph/super.c > > > index 9b1b7f4cfdd4..950a28ad9c59 100644 > > > --- a/fs/ceph/super.c > > > +++ b/fs/ceph/super.c > > > @@ -145,6 +145,7 @@ enum { > > > Opt_mds_namespace, > > > Opt_recover_session, > > > Opt_source, > > > + Opt_mon_addr, > > > /* string args above */ > > > Opt_dirstat, > > > Opt_rbytes, > > > @@ -196,6 +197,7 @@ static const struct fs_parameter_spec ceph_mount_parameters[] = { > > > fsparam_u32 ("rsize", Opt_rsize), > > > fsparam_string ("snapdirname", Opt_snapdirname), > > > fsparam_string ("source", Opt_source), > > > + fsparam_string ("mon_addr", Opt_mon_addr), > > > fsparam_u32 ("wsize", Opt_wsize), > > > fsparam_flag_no ("wsync", Opt_wsync), > > > {} > > > @@ -226,10 +228,68 @@ static void canonicalize_path(char *path) > > > path[j] = '\0'; > > > } > > > > > > +static int ceph_parse_old_source(const char *dev_name, const char *dev_name_end, > > > + struct fs_context *fc) > > > +{ > > > + int r; > > > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > > > + struct ceph_mount_options *fsopt = pctx->opts; > > > + > > > + if (*dev_name_end != ':') > > > + return invalfc(fc, "separator ':' missing in source"); > > > + > > > + r = ceph_parse_mon_ips(dev_name, dev_name_end - dev_name, > > > + pctx->copts, fc->log.log); > > > + if (r) > > > + return r; > > > + > > > + fsopt->new_dev_syntax = false; > > > + return 0; > > > +} > > > + > > > +static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end, > > > + struct fs_context *fc) > > > +{ > > > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > > > + struct ceph_mount_options *fsopt = pctx->opts; > > > + char *fsid_start, *fs_name_start; > > > + > > > + if (*dev_name_end != '=') > > > + return invalfc(fc, "separator '=' missing in source"); > > > > An annoying thing is that we'll always see this error message when falling > > back to the old_source method. > > True. I'd rather this fallback be silent. > > (Also, is there a good reason for using '=' instead of ':'? I probably > > missed this discussion somewhere else already...) > > Yes, we needed a way for the kernel to distinguish between the old and new syntax. Old kernels already reject any mount string without ":/" in it, so we needed the new format to _not_ have that to ensure a clean fallback procedure. It's not as pretty as I would have liked, but it's starting to grow on me. :) > > > + > > > + fsid_start = strchr(dev_name, '@'); > > > + if (!fsid_start) > > > + return invalfc(fc, "missing cluster fsid"); > > > + ++fsid_start; /* start of cluster fsid */ > > > + > > > + fs_name_start = strchr(fsid_start, '.'); > > > + if (!fs_name_start) > > > + return invalfc(fc, "missing file system name"); > > > + > > > + ++fs_name_start; /* start of file system name */ > > > + fsopt->mds_namespace = kstrndup(fs_name_start, > > > + dev_name_end - fs_name_start, GFP_KERNEL); > > > + if (!fsopt->mds_namespace) > > > + return -ENOMEM; > > > + dout("file system (mds namespace) '%s'\n", fsopt->mds_namespace); > > > + > > > + fsopt->new_dev_syntax = true; > > > + return 0; > > > +} > > > + > > > /* > > > - * Parse the source parameter. Distinguish the server list from the path. > > > + * Parse the source parameter for new device format. Distinguish the device > > > + * spec from the path. Try parsing new device format and fallback to old > > > + * format if needed. > > > * > > > - * The source will look like: > > > + * New device syntax will looks like: > > > + * <device_spec>=/<path> > > > + * where > > > + * <device_spec> is name@fsid.fsname > > > + * <path> is optional, but if present must begin with '/' > > > + * (monitor addresses are passed via mount option) > > > + * > > > + * Old device syntax is: > > > * <server_spec>[,<server_spec>...]:[<path>] > > > * where > > > * <server_spec> is <ip>[:<port>] > > > @@ -262,22 +322,48 @@ static int ceph_parse_source(struct fs_parameter *param, struct fs_context *fc) > > > dev_name_end = dev_name + strlen(dev_name); > > > } > > > > > > - dev_name_end--; /* back up to ':' separator */ > > > - if (dev_name_end < dev_name || *dev_name_end != ':') > > > - return invalfc(fc, "No path or : separator in source"); > > > + dev_name_end--; /* back up to separator */ > > > + if (dev_name_end < dev_name) > > > + return invalfc(fc, "path missing in source"); > > > > > > dout("device name '%.*s'\n", (int)(dev_name_end - dev_name), dev_name); > > > if (fsopt->server_path) > > > dout("server path '%s'\n", fsopt->server_path); > > > > > > - ret = ceph_parse_mon_ips(param->string, dev_name_end - dev_name, > > > - pctx->copts, fc->log.log); > > > + dout("trying new device syntax"); > > > + ret = ceph_parse_new_source(dev_name, dev_name_end, fc); > > > + if (ret == 0) > > > + goto done; > > > + > > > + dout("trying old device syntax"); > > > + ret = ceph_parse_old_source(dev_name, dev_name_end, fc); > > > if (ret) > > > - return ret; > > > + goto out; > > > > > > + done: > > > fc->source = param->string; > > > param->string = NULL; > > > - return 0; > > > + out: > > > + return ret; > > > +} > > > + > > > +static int ceph_parse_mon_addr(struct fs_parameter *param, > > > + struct fs_context *fc) > > > +{ > > > + int r; > > > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > > > + struct ceph_mount_options *fsopt = pctx->opts; > > > + > > > + kfree(fsopt->mon_addr); > > > + fsopt->mon_addr = kstrdup(param->string, GFP_KERNEL); > > > + if (!fsopt->mon_addr) > > > + return -ENOMEM; > > > + > > > + strreplace(param->string, '/', ','); > > > + r = ceph_parse_mon_ips(param->string, strlen(param->string), > > > + pctx->copts, fc->log.log); > > > + param->string = NULL; > > > > I think this will result in a memory leak. Why don't we simply set > > fsopt->mon_addr to param->string instead of kstrdup'ing it? > > Sure. As discussed over irc, will do away with the alloc. > > > > > Cheers, > > -- > > Luís > > > > > + return r; > > > } > > > > > > static int ceph_parse_mount_param(struct fs_context *fc, > > > @@ -322,6 +408,8 @@ static int ceph_parse_mount_param(struct fs_context *fc, > > > if (fc->source) > > > return invalfc(fc, "Multiple sources specified"); > > > return ceph_parse_source(param, fc); > > > + case Opt_mon_addr: > > > + return ceph_parse_mon_addr(param, fc); > > > case Opt_wsize: > > > if (result.uint_32 < PAGE_SIZE || > > > result.uint_32 > CEPH_MAX_WRITE_SIZE) > > > @@ -473,6 +561,7 @@ static void destroy_mount_options(struct ceph_mount_options *args) > > > kfree(args->mds_namespace); > > > kfree(args->server_path); > > > kfree(args->fscache_uniq); > > > + kfree(args->mon_addr); > > > kfree(args); > > > } > > > > > > @@ -516,6 +605,10 @@ static int compare_mount_options(struct ceph_mount_options *new_fsopt, > > > if (ret) > > > return ret; > > > > > > + ret = strcmp_null(fsopt1->mon_addr, fsopt2->mon_addr); > > > + if (ret) > > > + return ret; > > > + > > > return ceph_compare_options(new_opt, fsc->client); > > > } > > > > > > @@ -571,9 +664,13 @@ static int ceph_show_options(struct seq_file *m, struct dentry *root) > > > if ((fsopt->flags & CEPH_MOUNT_OPT_NOCOPYFROM) == 0) > > > seq_puts(m, ",copyfrom"); > > > > > > - if (fsopt->mds_namespace) > > > + /* dump mds_namespace when old device syntax is in use */ > > > + if (fsopt->mds_namespace && !fsopt->new_dev_syntax) > > > seq_show_option(m, "mds_namespace", fsopt->mds_namespace); > > > > > > + if (fsopt->mon_addr) > > > + seq_printf(m, ",mon_addr=%s", fsopt->mon_addr); > > > + > > > if (fsopt->flags & CEPH_MOUNT_OPT_CLEANRECOVER) > > > seq_show_option(m, "recover_session", "clean"); > > > > > > diff --git a/fs/ceph/super.h b/fs/ceph/super.h > > > index 839e6b0239ee..557348ff3203 100644 > > > --- a/fs/ceph/super.h > > > +++ b/fs/ceph/super.h > > > @@ -88,6 +88,8 @@ struct ceph_mount_options { > > > unsigned int max_readdir; /* max readdir result (entries) */ > > > unsigned int max_readdir_bytes; /* max readdir result (bytes) */ > > > > > > + bool new_dev_syntax; > > > + > > > /* > > > * everything above this point can be memcmp'd; everything below > > > * is handled in compare_mount_options() > > > @@ -97,6 +99,7 @@ struct ceph_mount_options { > > > char *mds_namespace; /* default NULL */ > > > char *server_path; /* default NULL (means "/") */ > > > char *fscache_uniq; /* default NULL */ > > > + char *mon_addr; > > > }; > > > > > > struct ceph_fs_client { > > > -- > > > 2.27.0 > > > > > > >
On Tue, Jun 29, 2021 at 11:10:02AM -0400, Jeff Layton wrote: > On Tue, 2021-06-29 at 19:24 +0530, Venky Shankar wrote: > > On Tue, Jun 29, 2021 at 5:02 PM Luis Henriques <lhenriques@suse.de> wrote: > > > > > > [ As I said, I didn't fully reviewed this patch. Just sending out a few > > > comments. ] > > > > > > On Mon, Jun 28, 2021 at 01:25:42PM +0530, Venky Shankar wrote: > > > > Old mount device syntax (source) has the following problems: > > > > > > > > - mounts to the same cluster but with different fsnames > > > > and/or creds have identical device string which can > > > > confuse xfstests. > > > > > > > > - Userspace mount helper tool resolves monitor addresses > > > > and fill in mon addrs automatically, but that means the > > > > device shown in /proc/mounts is different than what was > > > > used for mounting. > > > > > > > > New device syntax is as follows: > > > > > > > > cephuser@fsid.mycephfs2=/path > > > > > > > > Note, there is no "monitor address" in the device string. > > > > That gets passed in as mount option. This keeps the device > > > > string same when monitor addresses change (on remounts). > > > > > > > > Also note that the userspace mount helper tool is backward > > > > compatible. I.e., the mount helper will fallback to using > > > > old syntax after trying to mount with the new syntax. > > > > > > > > Signed-off-by: Venky Shankar <vshankar@redhat.com> > > > > --- > > > > fs/ceph/super.c | 117 +++++++++++++++++++++++++++++++++++++++++++----- > > > > fs/ceph/super.h | 3 ++ > > > > 2 files changed, 110 insertions(+), 10 deletions(-) > > > > > > > > diff --git a/fs/ceph/super.c b/fs/ceph/super.c > > > > index 9b1b7f4cfdd4..950a28ad9c59 100644 > > > > --- a/fs/ceph/super.c > > > > +++ b/fs/ceph/super.c > > > > @@ -145,6 +145,7 @@ enum { > > > > Opt_mds_namespace, > > > > Opt_recover_session, > > > > Opt_source, > > > > + Opt_mon_addr, > > > > /* string args above */ > > > > Opt_dirstat, > > > > Opt_rbytes, > > > > @@ -196,6 +197,7 @@ static const struct fs_parameter_spec ceph_mount_parameters[] = { > > > > fsparam_u32 ("rsize", Opt_rsize), > > > > fsparam_string ("snapdirname", Opt_snapdirname), > > > > fsparam_string ("source", Opt_source), > > > > + fsparam_string ("mon_addr", Opt_mon_addr), > > > > fsparam_u32 ("wsize", Opt_wsize), > > > > fsparam_flag_no ("wsync", Opt_wsync), > > > > {} > > > > @@ -226,10 +228,68 @@ static void canonicalize_path(char *path) > > > > path[j] = '\0'; > > > > } > > > > > > > > +static int ceph_parse_old_source(const char *dev_name, const char *dev_name_end, > > > > + struct fs_context *fc) > > > > +{ > > > > + int r; > > > > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > > > > + struct ceph_mount_options *fsopt = pctx->opts; > > > > + > > > > + if (*dev_name_end != ':') > > > > + return invalfc(fc, "separator ':' missing in source"); > > > > + > > > > + r = ceph_parse_mon_ips(dev_name, dev_name_end - dev_name, > > > > + pctx->copts, fc->log.log); > > > > + if (r) > > > > + return r; > > > > + > > > > + fsopt->new_dev_syntax = false; > > > > + return 0; > > > > +} > > > > + > > > > +static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end, > > > > + struct fs_context *fc) > > > > +{ > > > > + struct ceph_parse_opts_ctx *pctx = fc->fs_private; > > > > + struct ceph_mount_options *fsopt = pctx->opts; > > > > + char *fsid_start, *fs_name_start; > > > > + > > > > + if (*dev_name_end != '=') > > > > + return invalfc(fc, "separator '=' missing in source"); > > > > > > An annoying thing is that we'll always see this error message when falling > > > back to the old_source method. > > > > > True. I'd rather this fallback be silent. > > > > (Also, is there a good reason for using '=' instead of ':'? I probably > > > missed this discussion somewhere else already...) > > > > > Yes, we needed a way for the kernel to distinguish between the old and > new syntax. Old kernels already reject any mount string without ":/" in > it, so we needed the new format to _not_ have that to ensure a clean > fallback procedure. > > It's not as pretty as I would have liked, but it's starting to grow on > me. :) Heh. And gets even worst with using '/' for separating the mons IPs. (I understand why it can't be ',' and there's probably a reason for not using ';' either.) But yeah, that ship has sailed, I'm sure you guys discussed all this already ;-) Cheers, -- Luís
diff --git a/fs/ceph/super.c b/fs/ceph/super.c index 9b1b7f4cfdd4..950a28ad9c59 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -145,6 +145,7 @@ enum { Opt_mds_namespace, Opt_recover_session, Opt_source, + Opt_mon_addr, /* string args above */ Opt_dirstat, Opt_rbytes, @@ -196,6 +197,7 @@ static const struct fs_parameter_spec ceph_mount_parameters[] = { fsparam_u32 ("rsize", Opt_rsize), fsparam_string ("snapdirname", Opt_snapdirname), fsparam_string ("source", Opt_source), + fsparam_string ("mon_addr", Opt_mon_addr), fsparam_u32 ("wsize", Opt_wsize), fsparam_flag_no ("wsync", Opt_wsync), {} @@ -226,10 +228,68 @@ static void canonicalize_path(char *path) path[j] = '\0'; } +static int ceph_parse_old_source(const char *dev_name, const char *dev_name_end, + struct fs_context *fc) +{ + int r; + struct ceph_parse_opts_ctx *pctx = fc->fs_private; + struct ceph_mount_options *fsopt = pctx->opts; + + if (*dev_name_end != ':') + return invalfc(fc, "separator ':' missing in source"); + + r = ceph_parse_mon_ips(dev_name, dev_name_end - dev_name, + pctx->copts, fc->log.log); + if (r) + return r; + + fsopt->new_dev_syntax = false; + return 0; +} + +static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end, + struct fs_context *fc) +{ + struct ceph_parse_opts_ctx *pctx = fc->fs_private; + struct ceph_mount_options *fsopt = pctx->opts; + char *fsid_start, *fs_name_start; + + if (*dev_name_end != '=') + return invalfc(fc, "separator '=' missing in source"); + + fsid_start = strchr(dev_name, '@'); + if (!fsid_start) + return invalfc(fc, "missing cluster fsid"); + ++fsid_start; /* start of cluster fsid */ + + fs_name_start = strchr(fsid_start, '.'); + if (!fs_name_start) + return invalfc(fc, "missing file system name"); + + ++fs_name_start; /* start of file system name */ + fsopt->mds_namespace = kstrndup(fs_name_start, + dev_name_end - fs_name_start, GFP_KERNEL); + if (!fsopt->mds_namespace) + return -ENOMEM; + dout("file system (mds namespace) '%s'\n", fsopt->mds_namespace); + + fsopt->new_dev_syntax = true; + return 0; +} + /* - * Parse the source parameter. Distinguish the server list from the path. + * Parse the source parameter for new device format. Distinguish the device + * spec from the path. Try parsing new device format and fallback to old + * format if needed. * - * The source will look like: + * New device syntax will looks like: + * <device_spec>=/<path> + * where + * <device_spec> is name@fsid.fsname + * <path> is optional, but if present must begin with '/' + * (monitor addresses are passed via mount option) + * + * Old device syntax is: * <server_spec>[,<server_spec>...]:[<path>] * where * <server_spec> is <ip>[:<port>] @@ -262,22 +322,48 @@ static int ceph_parse_source(struct fs_parameter *param, struct fs_context *fc) dev_name_end = dev_name + strlen(dev_name); } - dev_name_end--; /* back up to ':' separator */ - if (dev_name_end < dev_name || *dev_name_end != ':') - return invalfc(fc, "No path or : separator in source"); + dev_name_end--; /* back up to separator */ + if (dev_name_end < dev_name) + return invalfc(fc, "path missing in source"); dout("device name '%.*s'\n", (int)(dev_name_end - dev_name), dev_name); if (fsopt->server_path) dout("server path '%s'\n", fsopt->server_path); - ret = ceph_parse_mon_ips(param->string, dev_name_end - dev_name, - pctx->copts, fc->log.log); + dout("trying new device syntax"); + ret = ceph_parse_new_source(dev_name, dev_name_end, fc); + if (ret == 0) + goto done; + + dout("trying old device syntax"); + ret = ceph_parse_old_source(dev_name, dev_name_end, fc); if (ret) - return ret; + goto out; + done: fc->source = param->string; param->string = NULL; - return 0; + out: + return ret; +} + +static int ceph_parse_mon_addr(struct fs_parameter *param, + struct fs_context *fc) +{ + int r; + struct ceph_parse_opts_ctx *pctx = fc->fs_private; + struct ceph_mount_options *fsopt = pctx->opts; + + kfree(fsopt->mon_addr); + fsopt->mon_addr = kstrdup(param->string, GFP_KERNEL); + if (!fsopt->mon_addr) + return -ENOMEM; + + strreplace(param->string, '/', ','); + r = ceph_parse_mon_ips(param->string, strlen(param->string), + pctx->copts, fc->log.log); + param->string = NULL; + return r; } static int ceph_parse_mount_param(struct fs_context *fc, @@ -322,6 +408,8 @@ static int ceph_parse_mount_param(struct fs_context *fc, if (fc->source) return invalfc(fc, "Multiple sources specified"); return ceph_parse_source(param, fc); + case Opt_mon_addr: + return ceph_parse_mon_addr(param, fc); case Opt_wsize: if (result.uint_32 < PAGE_SIZE || result.uint_32 > CEPH_MAX_WRITE_SIZE) @@ -473,6 +561,7 @@ static void destroy_mount_options(struct ceph_mount_options *args) kfree(args->mds_namespace); kfree(args->server_path); kfree(args->fscache_uniq); + kfree(args->mon_addr); kfree(args); } @@ -516,6 +605,10 @@ static int compare_mount_options(struct ceph_mount_options *new_fsopt, if (ret) return ret; + ret = strcmp_null(fsopt1->mon_addr, fsopt2->mon_addr); + if (ret) + return ret; + return ceph_compare_options(new_opt, fsc->client); } @@ -571,9 +664,13 @@ static int ceph_show_options(struct seq_file *m, struct dentry *root) if ((fsopt->flags & CEPH_MOUNT_OPT_NOCOPYFROM) == 0) seq_puts(m, ",copyfrom"); - if (fsopt->mds_namespace) + /* dump mds_namespace when old device syntax is in use */ + if (fsopt->mds_namespace && !fsopt->new_dev_syntax) seq_show_option(m, "mds_namespace", fsopt->mds_namespace); + if (fsopt->mon_addr) + seq_printf(m, ",mon_addr=%s", fsopt->mon_addr); + if (fsopt->flags & CEPH_MOUNT_OPT_CLEANRECOVER) seq_show_option(m, "recover_session", "clean"); diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 839e6b0239ee..557348ff3203 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -88,6 +88,8 @@ struct ceph_mount_options { unsigned int max_readdir; /* max readdir result (entries) */ unsigned int max_readdir_bytes; /* max readdir result (bytes) */ + bool new_dev_syntax; + /* * everything above this point can be memcmp'd; everything below * is handled in compare_mount_options() @@ -97,6 +99,7 @@ struct ceph_mount_options { char *mds_namespace; /* default NULL */ char *server_path; /* default NULL (means "/") */ char *fscache_uniq; /* default NULL */ + char *mon_addr; }; struct ceph_fs_client {
Old mount device syntax (source) has the following problems: - mounts to the same cluster but with different fsnames and/or creds have identical device string which can confuse xfstests. - Userspace mount helper tool resolves monitor addresses and fill in mon addrs automatically, but that means the device shown in /proc/mounts is different than what was used for mounting. New device syntax is as follows: cephuser@fsid.mycephfs2=/path Note, there is no "monitor address" in the device string. That gets passed in as mount option. This keeps the device string same when monitor addresses change (on remounts). Also note that the userspace mount helper tool is backward compatible. I.e., the mount helper will fallback to using old syntax after trying to mount with the new syntax. Signed-off-by: Venky Shankar <vshankar@redhat.com> --- fs/ceph/super.c | 117 +++++++++++++++++++++++++++++++++++++++++++----- fs/ceph/super.h | 3 ++ 2 files changed, 110 insertions(+), 10 deletions(-)