rpms/kernel/devel kernel-2.6.30-pnfs-rc7.patch, NONE, 1.1.2.1 config-generic, 1.284, 1.284.4.1 kernel.spec, 1.1553, 1.1553.4.1
Steve Dickson
steved at fedoraproject.org
Wed May 27 17:45:50 UTC 2009
Author: steved
Update of /cvs/pkgs/rpms/kernel/devel
In directory cvs1.fedora.phx.redhat.com:/tmp/cvs-serv13909
Modified Files:
Tag: kernel-2_6_30-pnfs_rc7
config-generic kernel.spec
Added Files:
Tag: kernel-2_6_30-pnfs_rc7
kernel-2.6.30-pnfs-rc7.patch
Log Message:
Updated to latest pNFS code (tag: v2.6.30-rc7)
kernel-2.6.30-pnfs-rc7.patch:
--- NEW FILE kernel-2.6.30-pnfs-rc7.patch ---
diff --git a/Documentation/filesystems/00-INDEX b/Documentation/filesystems/00-INDEX
index 8dd6db7..f15621e 100644
--- a/Documentation/filesystems/00-INDEX
+++ b/Documentation/filesystems/00-INDEX
@@ -66,6 +66,10 @@ mandatory-locking.txt
- info on the Linux implementation of Sys V mandatory file locking.
ncpfs.txt
- info on Novell Netware(tm) filesystem using NCP protocol.
+nfs41-server.txt
+ - info on the Linux server implementation of NFSv4 minor version 1.
+nfs-rdma.txt
+ - how to install and setup the Linux NFS/RDMA client and server software.
nfsroot.txt
- short guide on setting up a diskless box with NFS root filesystem.
nilfs2.txt
diff --git a/Documentation/spnfs.txt b/Documentation/spnfs.txt
new file mode 100644
index 0000000..be5815d
--- /dev/null
+++ b/Documentation/spnfs.txt
@@ -0,0 +1,210 @@
+(c) 2007 Network Appliance Inc.
+
+spNFS
+-----
+
+An spNFS system consists of a Meta Data Server (MDS), a number of Client machines (C) and a number of Data Servers (DS).
+
+A file system is mounted by the clients from the MDS, and all file data
+is striped across the DSs.
+
+Identify the machines that will be filling each of these roles.
+
+The spnfs kernel will be installed on all machines: clients, the MDS and DSs.
+
+
+Building and installing the spNFS kernel
+----------------------------------------
+
+Get the spNFS kernel from:
+
+ git://linux-nfs.org/~dmuntz/spnfs.git
+
+add these options to your .config file
+
+ CONFIG_NETWORK_FILESYSTEMS=y
+ CONFIG_NFS_FS=y
+ CONFIG_NFSD=y
+ CONFIG_NFS_V4_1=y
+ CONFIG_NFSD_V4_1=y
+ CONFIG_PNFS=y
+ CONFIG_PNFSD=y
+ CONFIG_SPNFS=y
+
+By default, spNFS uses whole-file layouts. Layout segments can be enabled
+by adding:
+
+ CONFIG_SPNFS_LAYOUTSEGMENTS=y
+
+to your .config file.
+
+Building and installation of kernel+modules is as usual.
+This kernel should be installed and booted on the client, MDS and DSs.
+
+
+Building nfs-utils
+------------------
+
+Get the nfs-utils package containing spnfsd from:
+
+ git://linux-nfs.org/~dmuntz/nfs-utils.git
+
+Follow the standard instructions for building nfs-utils. We HIGHLY recommend
+NOT doing an install of the binaries generated by this build. You will only
+need the spnfsd binary generated by this build, and the spnfsd.conf template.
+
+After building, the spnfsd daemon will be located in utils/spnfsd. The spnfsd
+daemon will only be needed on the MDS.
+
+
+Installation
+------------
+
+The nfs-utils package contains a default spnfsd.conf file in
+utils/spnfsd/spnfsd.conf. Copy this file to /etc/spnfsd.conf.
+
+By default, the DS-Mount-Directory is set to /spnfs (see spnfsd.conf). Under
+this directory, mount points must be created for each DS to
+be used for pNFS data stripes. These mount points are named by the ip address
+of the corresponding DS. In the sample spnfsd.conf, there are two
+DSs defined (172.16.28.134 and 172.16.28.141).
+
+Following the sample spnfsd.conf,
+
+ mkdir /spnfs
+
+on the MDS (corresponding to DS-Mount-Directory). Then
+
+ mkdir /spnfs/172.16.28.134
+ mkdir /spnfs/172.16.28.141
+
+to create the mount points for the DSs.
+
+On the DSs, chose a directory where data stripes will be created by the MDS.
+For the sample file, this directory is /pnfs, so on each DS execute:
+
+ mkdir /pnfs
+
+This directory is specified in the spnfsd.conf file by the DS*_ROOT option
+(where * is replaced by the DS number). DS_ROOT is specified relative to
+the directory being exported by the DSs. In our example, our DSs are exporting
+the root directory (/) and therefore our DS_ROOT is /pnfs. On the DSs, we have
+the following entry in /etc/exports:
+
+ / *(rw,fsid=0,insecure,no_root_squash,sync,no_subtree_check)
+
+N.B. If we had created a /exports directory and a /pnfs directory under
+/exports, and if we were exporting /exports, then DS_ROOT would still be /pnfs
+(not /exports/pnfs).
+
+It may be useful to add entries to /etc/fstab on the MDS to automatically
+mount the DS_ROOT file systems. For this example, our MDS fstab would
+contain:
+
+ 172.17.84.128:/pnfs /spnfs/172.17.84.128 nfs defaults 1 2
+ 172.17.84.122:/pnfs /spnfs/172.17.84.122 nfs defaults 1 2
+
+The DS mounts must be performed manually or via fstab at this time (automatic
+mounting, directory creation, etc. are on the todo list). To perform I/O
+through the MDS, the DS mounts MUST use NFSv3 at this time (this restriction
+will eventually be removed).
+
+
+On the MDS, choose a file system to use with spNFS and export it, e.g.:
+
+ / *(rw,fsid=0,insecure,no_root_squash,sync,no_subtree_check)
+
+Make sure nfsd and all supporting processes are running on the MDS and DSs.
+
+
+Running
+-------
+
+If rpc_pipefs is not already mounted (if you're running idmapd it probably is),
+you may want to add the following line to /etc/fstab:
+
+ rpc_pipefs /var/lib/nfs/rpc_pipefs rpc_pipefs defaults 0 0
+
+to automatically mount rpc_pipefs.
+
+With spnfsd.conf configured for your environment and the mounts mounted as
+described above, spnfsd can now be started.
+
+On the MDS, execute spnfsd:
+
+ spnfsd
+
+The executable is located in the directory where it was built, and
+may also have been installed elsewhere depending on how you built nfs-utils.
+It will run in the foreground by default, and in fact will do so despite
+any options suggesting the contrary (it's still a debugging build).
+
+On the client, make sure the nfslayoutdriver module is loaded:
+
+ modprobe nfslayoutdriver
+
+Then mount the file system from the MDS:
+
+ mount -t nfs4 mds:/ /mnt
+
+I/O through the MDS is now supported. To use it, do not load the
+nfslayoutdriver on the client, and mount the MDS using NFSv4 or 4.1
+(NFSv2 and v3 are not yet supported).
+
+You may now use spNFS by performing file system activities in /mnt.
+If you create files in /mnt, you should see stripe files corresponding to
+new files being created on the DSs. The current implementation names the
+stripe files based on the inode number of the file on the MDS. For example,
+if you create a file foo in /mnt and do an 'ls -li /mnt/foo':
+
+ # ls -li foo
+ 1233 -rw-r--r-- 1 root root 0 Nov 29 15:54 foo
+
+You should see stripe files on each under /pnfs (per the sample) named
+1233. The file /pnfs/1233 on DS1 will contain the first <stripe size> bytes
+of data written to foo, DS2 will contain the next <stripe size> bytes, etc.
+Removing /mnt/foo will remove the corresponding stripe files on the DSs.
+Other file system operations should behave (mostly :-) as expected.
+
+
+Layout Segments
+---------------
+
+If the kernel is compiled to support layout segments, there will
+be two files created under /proc/fs/spnfs for controlling layout
+segment functionality.
+
+To enable layout segments, write a '1' to /proc/fs/spnfs/layoutseg, e.g.:
+
[...32800 lines suppressed...]
+ len = sock->ops->sendpage(sock, virt_to_page(xbufp->head[0].iov_base),
+ (unsigned long)xbufp->head[0].iov_base & ~PAGE_MASK,
+ xbufp->head[0].iov_len, flags);
+
+ if (len != xbufp->head[0].iov_len)
+ goto out;
+
+ /*
+ * send page data
+ *
+ * Check the amount of data to be sent. If it is less than the
+ * remaining page, then send it else send the current page
+ */
+
+ size = PAGE_SIZE - base < pglen ? PAGE_SIZE - base : pglen;
+ while (pglen > 0) {
+ if (total_len == size)
+ flags = 0;
+ result = sock->ops->sendpage(sock, *pages, base, size, flags);
+ if (result > 0)
+ len += result;
+ if (result != size)
+ goto out;
+ total_len -= size;
+ pglen -= size;
+ size = PAGE_SIZE < pglen ? PAGE_SIZE : pglen;
+ base = 0;
+ pages++;
+ }
+ /*
+ * send tail
+ */
+ if (xbufp->tail[0].iov_len) {
+ result = sock->ops->sendpage(sock,
+ xbufp->tail[0].iov_base,
+ (unsigned long)xbufp->tail[0].iov_base & ~PAGE_MASK,
+ xbufp->tail[0].iov_len,
+ 0);
+
+ if (result > 0)
+ len += result;
+ }
+out:
+ if (len != xbufp->len)
+ printk(KERN_NOTICE "Error sending entire callback!\n");
+
+ return len;
+}
+
+/*
+ * The send routine. Borrows from svc_send
+ */
+static int bc_send_request(struct rpc_task *task)
+{
+ struct rpc_rqst *req = task->tk_rqstp;
+ struct rpc_xprt *bc_xprt = req->rq_xprt;
+ struct svc_xprt *xprt;
+ struct svc_sock *svsk;
+ u32 len;
+
+ dprintk("sending request with xid: %08x\n", ntohl(req->rq_xid));
+ /*
+ * Get the server socket associated with this callback xprt
+ */
+ svsk = bc_xprt->bc_sock;
+ xprt = &svsk->sk_xprt;
+
+ mutex_lock(&xprt->xpt_mutex);
+ if (test_bit(XPT_DEAD, &xprt->xpt_flags))
+ len = -ENOTCONN;
+ else
+ len = bc_sendto(req);
+ mutex_unlock(&xprt->xpt_mutex);
+
+ return 0;
+
+}
+
+/*
+ * The close routine. Since this is client initiated, we do nothing
+ */
+
+static void bc_close(struct rpc_xprt *xprt)
+{
+ return;
+}
+
+/*
+ * The xprt destroy routine. Again, because this connection is client
+ * initiated, we do nothing
+ */
+
+static void bc_destroy(struct rpc_xprt *xprt)
+{
+ return;
+}
+
static struct rpc_xprt_ops xs_udp_ops = {
.set_buffer_size = xs_udp_set_buffer_size,
.reserve_xprt = xprt_reserve_xprt_cong,
@@ -2010,11 +2375,32 @@ static struct rpc_xprt_ops xs_tcp_ops = {
.buf_free = rpc_free,
.send_request = xs_tcp_send_request,
.set_retrans_timeout = xprt_set_retrans_timeout_def,
+#if defined(CONFIG_NFS_V4_1)
+ .release_request = bc_release_request,
+#endif /* CONFIG_NFS_V4_1 */
.close = xs_tcp_close,
.destroy = xs_destroy,
.print_stats = xs_tcp_print_stats,
};
+/*
+ * The rpc_xprt_ops for the server backchannel
+ */
+
+static struct rpc_xprt_ops bc_tcp_ops = {
+ .reserve_xprt = xprt_reserve_xprt,
+ .release_xprt = xprt_release_xprt,
+ .set_port = bc_set_port,
+ .connect = bc_connect,
+ .buf_alloc = bc_malloc,
+ .buf_free = bc_free,
+ .send_request = bc_send_request,
+ .set_retrans_timeout = xprt_set_retrans_timeout_def,
+ .close = bc_close,
+ .destroy = bc_destroy,
+ .print_stats = xs_tcp_print_stats,
+};
+
static struct rpc_xprt *xs_setup_xprt(struct xprt_create *args,
unsigned int slot_table_size)
{
@@ -2147,13 +2533,29 @@ static struct rpc_xprt *xs_setup_tcp(struct xprt_create *args)
xprt->tsh_size = sizeof(rpc_fraghdr) / sizeof(u32);
xprt->max_payload = RPC_MAX_FRAGMENT_SIZE;
- xprt->bind_timeout = XS_BIND_TO;
- xprt->connect_timeout = XS_TCP_CONN_TO;
- xprt->reestablish_timeout = XS_TCP_INIT_REEST_TO;
- xprt->idle_timeout = XS_IDLE_DISC_TO;
+ if (args->bc_sock) {
+ /* backchannel */
+ xprt_set_bound(xprt);
+ INIT_DELAYED_WORK(&transport->connect_worker,
+ bc_connect_worker);
+ xprt->bind_timeout = 0;
+ xprt->connect_timeout = 0;
+ xprt->reestablish_timeout = 0;
+ xprt->idle_timeout = (~0);
- xprt->ops = &xs_tcp_ops;
- xprt->timeout = &xs_tcp_default_timeout;
+ /*
+ * The backchannel uses the same socket connection as the
+ * forechannel
+ */
+ xprt->bc_sock = args->bc_sock;
+ xprt->bc_sock->sk_bc_xprt = xprt;
+ transport->sock = xprt->bc_sock->sk_sock;
+ transport->inet = xprt->bc_sock->sk_sk;
+
+ xprt->ops = &bc_tcp_ops;
+
+ goto next;
+ }
switch (addr->sa_family) {
case AF_INET:
@@ -2161,13 +2563,29 @@ static struct rpc_xprt *xs_setup_tcp(struct xprt_create *args)
xprt_set_bound(xprt);
INIT_DELAYED_WORK(&transport->connect_worker, xs_tcp_connect_worker4);
- xs_format_ipv4_peer_addresses(xprt, "tcp", RPCBIND_NETID_TCP);
break;
case AF_INET6:
if (((struct sockaddr_in6 *)addr)->sin6_port != htons(0))
xprt_set_bound(xprt);
INIT_DELAYED_WORK(&transport->connect_worker, xs_tcp_connect_worker6);
+ break;
+ }
+ xprt->bind_timeout = XS_BIND_TO;
+ xprt->connect_timeout = XS_TCP_CONN_TO;
+ xprt->reestablish_timeout = XS_TCP_INIT_REEST_TO;
+ xprt->idle_timeout = XS_IDLE_DISC_TO;
+
+ xprt->ops = &xs_tcp_ops;
+
+next:
+ xprt->timeout = &xs_tcp_default_timeout;
+
+ switch (addr->sa_family) {
+ case AF_INET:
+ xs_format_ipv4_peer_addresses(xprt, "tcp", RPCBIND_NETID_TCP);
+ break;
+ case AF_INET6:
xs_format_ipv6_peer_addresses(xprt, "tcp", RPCBIND_NETID_TCP6);
break;
default:
Index: config-generic
===================================================================
RCS file: /cvs/pkgs/rpms/kernel/devel/config-generic,v
retrieving revision 1.284
retrieving revision 1.284.4.1
diff -u -p -r1.284 -r1.284.4.1
--- config-generic 24 May 2009 01:40:12 -0000 1.284
+++ config-generic 27 May 2009 17:45:19 -0000 1.284.4.1
@@ -3141,6 +3141,7 @@ CONFIG_NFS_FS=m
CONFIG_NFS_V3=y
CONFIG_NFS_V3_ACL=y
CONFIG_NFS_V4=y
+CONFIG_NFS_V4_1=y
CONFIG_NFS_DIRECTIO=y
CONFIG_NFSD=m
CONFIG_NFSD_V3=y
@@ -3148,6 +3149,13 @@ CONFIG_NFSD_V3_ACL=y
CONFIG_NFSD_V4=y
CONFIG_NFSD_TCP=y
CONFIG_NFS_FSCACHE=y
+CONFIG_PNFS=y
+CONFIG_PNFSD=y
+CONFIG_PNFSD_LOCAL_EXPORT=y
+CONFIG_PNFS_PANLAYOUT=y
+CONFIG_PNFS_BLOCK=y
+CONFIG_SPNFS=y
+CONFIG_SPNFS_LAYOUTSEGMENTS=y
CONFIG_LOCKD=m
CONFIG_LOCKD_V4=y
CONFIG_EXPORTFS=m
Index: kernel.spec
===================================================================
RCS file: /cvs/pkgs/rpms/kernel/devel/kernel.spec,v
retrieving revision 1.1553
retrieving revision 1.1553.4.1
diff -u -p -r1.1553 -r1.1553.4.1
--- kernel.spec 26 May 2009 16:25:44 -0000 1.1553
+++ kernel.spec 27 May 2009 17:45:19 -0000 1.1553.4.1
@@ -12,7 +12,7 @@ Summary: The Linux kernel
# that the kernel isn't the stock distribution kernel, for example,
# by setting the define to ".local" or ".bz123456"
#
-# % define buildid .local
+%define buildid .pnfs
# fedora_build defines which build revision of this kernel version we're
# building. Rather than incrementing forever, as with the prior versioning
@@ -86,7 +86,8 @@ Summary: The Linux kernel
# kernel-headers
%define with_headers %{?_without_headers: 0} %{?!_without_headers: 1}
# kernel-firmware
-%define with_firmware %{?_with_firmware: 1} %{?!_with_firmware: 0}
+#%define with_firmware %{?_with_firmware: 1} %{?!_with_firmware: 0}
+%define with_firmware 1
# kernel-debuginfo
%define with_debuginfo %{?_without_debuginfo: 0} %{?!_without_debuginfo: 1}
# kernel-bootwrapper (for creating zImages from kernel + initrd)
@@ -255,7 +256,7 @@ Summary: The Linux kernel
%define with_up 0
%define with_headers 0
%define all_arch_configs kernel-%{version}-*.config
-%define with_firmware %{?_without_firmware: 0} %{?!_without_firmware: 1}
+%define with_firmware 0
%endif
# bootwrapper is only on ppc
@@ -670,6 +671,7 @@ Patch9002: cpufreq-add-atom-to-p4-clockm
#snmp fixes
Patch10000: linux-2.6-missing-rfc2465-stats.patch
+Patch20000: kernel-2.6.30-pnfs-rc7.patch
%endif
BuildRoot: %{_tmppath}/kernel-%{KVERREL}-root
@@ -1222,6 +1224,7 @@ ApplyPatch linux-2.6-silence-acpi-blackl
#ApplyPatch cpufreq-add-atom-to-p4-clockmod.patch
+ApplyPatch kernel-2.6.30-pnfs-rc7.patch
# END OF PATCH APPLICATIONS
%endif
@@ -1815,6 +1818,9 @@ fi
# ||----w |
# || ||
%changelog
+* Wed May 27 2009 Steve Dickson <steved at redhat.com>
+- Updated to latest pNFS code (tag: v2.6.30-rc7).
+
* Tue May 26 2009 Dave Jones <davej at redhat.com>
- Various cpufreq patches from git.
More information about the scm-commits
mailing list