The spec allows backchannels for multiple clients to share the same tcp
connection. When that happens, we need to use the same xprt for all of
them. Similarly, we need the same xps.
This fixes list corruption introduced by the multipath code.
Cc: stable@vger.kernel.org
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: Trond Myklebust <trondmy@primarydata.com>
Pull string hash improvements from George Spelvin:
"This series does several related things:
- Makes the dcache hash (fs/namei.c) useful for general kernel use.
(Thanks to Bruce for noticing the zero-length corner case)
- Converts the string hashes in <linux/sunrpc/svcauth.h> to use the
above.
- Avoids 64-bit multiplies in hash_64() on 32-bit platforms. Two
32-bit multiplies will do well enough.
- Rids the world of the bad hash multipliers in hash_32.
This finishes the job started in commit 689de1d6ca ("Minimal
fix-up of bad hashing behavior of hash_64()")
The vast majority of Linux architectures have hardware support for
32x32-bit multiply and so derive no benefit from "simplified"
multipliers.
The few processors that do not (68000, h8/300 and some models of
Microblaze) have arch-specific implementations added. Those
patches are last in the series.
- Overhauls the dcache hash mixing.
The patch in commit 0fed3ac866 ("namei: Improve hash mixing if
CONFIG_DCACHE_WORD_ACCESS") was an off-the-cuff suggestion.
Replaced with a much more careful design that's simultaneously
faster and better. (My own invention, as there was noting suitable
in the literature I could find. Comments welcome!)
- Modify the hash_name() loop to skip the initial HASH_MIX(). This
would let us salt the hash if we ever wanted to.
- Sort out partial_name_hash().
The hash function is declared as using a long state, even though
it's truncated to 32 bits at the end and the extra internal state
contributes nothing to the result. And some callers do odd things:
- fs/hfs/string.c only allocates 32 bits of state
- fs/hfsplus/unicode.c uses it to hash 16-bit unicode symbols not bytes
- Modify bytemask_from_count to handle inputs of 1..sizeof(long)
rather than 0..sizeof(long)-1. This would simplify users other
than full_name_hash"
Special thanks to Bruce Fields for testing and finding bugs in v1. (I
learned some humbling lessons about "obviously correct" code.)
On the arch-specific front, the m68k assembly has been tested in a
standalone test harness, I've been in contact with the Microblaze
maintainers who mostly don't care, as the hardware multiplier is never
omitted in real-world applications, and I haven't heard anything from
the H8/300 world"
* 'hash' of git://ftp.sciencehorizons.net/linux:
h8300: Add <asm/hash.h>
microblaze: Add <asm/hash.h>
m68k: Add <asm/hash.h>
<linux/hash.h>: Add support for architecture-specific functions
fs/namei.c: Improve dcache hash function
Eliminate bad hash multipliers from hash_32() and hash_64()
Change hash_64() return value to 32 bits
<linux/sunrpc/svcauth.h>: Define hash_str() in terms of hashlen_string()
fs/namei.c: Add hashlen_string() function
Pull out string hash to <linux/stringhash.h>
Finally, the first use of previous two patches: eliminate the
separate ad-hoc string hash functions in the sunrpc code.
Now hash_str() is a wrapper around hash_string(), and hash_mem() is
likewise a wrapper around full_name_hash().
Note that sunrpc code *does* call hash_mem() with a zero length, which
is why the previous patch needed to handle that in full_name_hash().
(Thanks, Bruce, for finding that!)
This also eliminates the only caller of hash_long which asks for
more than 32 bits of output.
The comment about the quality of hashlen_string() and full_name_hash()
is jumping the gun by a few patches; they aren't very impressive now,
but will be improved greatly later in the series.
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Tested-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: J. Bruce Fields <bfields@redhat.com>
Cc: Jeff Layton <jlayton@poochiereds.net>
Cc: linux-nfs@vger.kernel.org
Pull NFS client updates from Anna Schumaker:
"Highlights include:
Features:
- Add support for the NFS v4.2 COPY operation
- Add support for NFS/RDMA over IPv6
Bugfixes and cleanups:
- Avoid race that crashes nfs_init_commit()
- Fix oops in callback path
- Fix LOCK/OPEN race when unlinking an open file
- Choose correct stateids when using delegations in setattr, read and
write
- Don't send empty SETATTR after OPEN_CREATE
- xprtrdma: Prevent server from writing a reply into memory client
has released
- xprtrdma: Support using Read list and Reply chunk in one RPC call"
* tag 'nfs-for-4.7-1' of git://git.linux-nfs.org/projects/anna/linux-nfs: (61 commits)
pnfs: pnfs_update_layout needs to consider if strict iomode checking is on
nfs/flexfiles: Use the layout segment for reading unless it a IOMODE_RW and reading is disabled
nfs/flexfiles: Helper function to detect FF_FLAGS_NO_READ_IO
nfs: avoid race that crashes nfs_init_commit
NFS: checking for NULL instead of IS_ERR() in nfs_commit_file()
pnfs: make pnfs_layout_process more robust
pnfs: rework LAYOUTGET retry handling
pnfs: lift retry logic from send_layoutget to pnfs_update_layout
pnfs: fix bad error handling in send_layoutget
flexfiles: add kerneldoc header to nfs4_ff_layout_prepare_ds
flexfiles: remove pointless setting of NFS_LAYOUT_RETURN_REQUESTED
pnfs: only tear down lsegs that precede seqid in LAYOUTRETURN args
pnfs: keep track of the return sequence number in pnfs_layout_hdr
pnfs: record sequence in pnfs_layout_segment when it's created
pnfs: don't merge new ff lsegs with ones that have LAYOUTRETURN bit set
pNFS/flexfiles: When initing reads or writes, we might have to retry connecting to DSes
pNFS/flexfiles: When checking for available DSes, conditionally check for MDS io
pNFS/flexfile: Fix erroneous fall back to read/write through the MDS
NFS: Reclaim writes via writepage are opportunistic
NFSv4: Use the right stateid for delegations in setattr, read and write
...
Currently the sysctls that allow setting the inline threshold allow
any value to be set.
Small values only make the transport run slower. The default 1KB
setting is as low as is reasonable. And the logic that decides how
to divide a Send buffer between RPC-over-RDMA header and RPC message
assumes (but does not check) that the lower bound is not crazy (say,
57 bytes).
Send and receive buffers share a page with some control information.
Values larger than about 3KB can't be supported, currently.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Tested-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
RPC-over-RDMA transports have a limit on how large a backward
direction (backchannel) RPC message can be. Ensure that the NFSv4.x
CREATE_SESSION operation advertises this limit to servers.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Tested-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Commit 176e21ee2e ("SUNRPC: Support for RPC over AF_LOCAL
transports") added a 5-character netid, but did not bump
RPCBIND_MAXNETIDLEN from 4 to 5.
Fixes: 176e21ee2e ("SUNRPC: Support for RPC over AF_LOCAL ...")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: Pass in just the piece of the svc_rqst that is needed
here.
While we're in the area, add an informative documenting comment.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Sometimes we might have a RCU managed credential pointer and don't want
to use locking to handle it. Add a function that will take a reference
to the cred iff the refcount is not already zero. Callers can dereference
the pointer under the rcu_read_lock and use that function to take a
reference only if the cred is not on its way to destruction.
Signed-off-by: Jeff Layton <jeff.layton@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Add function rpc_lookup_generic_cred, which allows lookups of a generic
credential that's not current_cred().
[jlayton: add gfp_t parm]
Signed-off-by: Jeff Layton <jeff.layton@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
We need to be able to call the generic_cred creator from different
contexts. Add a gfp_t parm to the crcreate operation and to
rpcauth_lookup_credcache. For now, we just push the gfp_t parms up
one level to the *_lookup_cred functions.
Signed-off-by: Jeff Layton <jeff.layton@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Pull nfsd updates from Bruce Fields:
"Various bugfixes, a RDMA update from Chuck Lever, and support for a
new pnfs layout type from Christoph Hellwig. The new layout type is a
variant of the block layout which uses SCSI features to offer improved
fencing and device identification.
(Also: note this pull request also includes the client side of SCSI
layout, with Trond's permission.)"
* tag 'nfsd-4.6' of git://linux-nfs.org/~bfields/linux:
sunrpc/cache: drop reference when sunrpc_cache_pipe_upcall() detects a race
nfsd: recover: fix memory leak
nfsd: fix deadlock secinfo+readdir compound
nfsd4: resfh unused in nfsd4_secinfo
svcrdma: Use new CQ API for RPC-over-RDMA server send CQs
svcrdma: Use new CQ API for RPC-over-RDMA server receive CQs
svcrdma: Remove close_out exit path
svcrdma: Hook up the logic to return ERR_CHUNK
svcrdma: Use correct XID in error replies
svcrdma: Make RDMA_ERROR messages work
rpcrdma: Add RPCRDMA_HDRLEN_ERR
svcrdma: svc_rdma_post_recv() should close connection on error
svcrdma: Close connection when a send error occurs
nfsd: Lower NFSv4.1 callback message size limit
svcrdma: Do not send Write chunk XDR pad with inline content
svcrdma: Do not write xdr_buf::tail in a Write chunk
svcrdma: Find client-provided write and reply chunks once per reply
nfsd: Update NFS server comments related to RDMA support
nfsd: Fix a memory leak when meeting unsupported state_protect_how4
nfsd4: fix bad bounds checking
Pull NFS client updates from Trond Myklebust:
"Highlights include:
Features:
- Add support for multiple NFSv4.1 callbacks in flight
- Initial patchset for RPC multipath support
- Adapt RPC/RDMA to use the new completion queue API
Bugfixes and cleanups:
- nfs4: nfs4_ff_layout_prepare_ds should return NULL if connection failed
- Cleanups to remove nfs_inode_dio_wait and nfs4_file_fsync
- Fix RPC/RDMA credit accounting
- Properly handle RDMA_ERROR replies
- xprtrdma: Do not wait if ib_post_send() fails
- xprtrdma: Segment head and tail XDR buffers on page boundaries
- xprtrdma cleanups for dprintk, physical_op_map and unused macros"
* tag 'nfs-for-4.6-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (35 commits)
nfs/blocklayout: make sure making a aligned read request
nfs4: nfs4_ff_layout_prepare_ds should return NULL if connection failed
nfs: remove nfs_inode_dio_wait
nfs: remove nfs4_file_fsync
xprtrdma: Use new CQ API for RPC-over-RDMA client send CQs
xprtrdma: Use an anonymous union in struct rpcrdma_mw
xprtrdma: Use new CQ API for RPC-over-RDMA client receive CQs
xprtrdma: Serialize credit accounting again
xprtrdma: Properly handle RDMA_ERROR replies
rpcrdma: Add RPCRDMA_HDRLEN_ERR
xprtrdma: Do not wait if ib_post_send() fails
xprtrdma: Segment head and tail XDR buffers on page boundaries
xprtrdma: Clean up dprintk format string containing a newline
xprtrdma: Clean up physical_op_map()
xprtrdma: Clean up unused RPCRDMA_INLINE_PAD_THRESH macro
NFS add callback_ops to nfs4_proc_bind_conn_to_session_callback
pnfs/NFSv4.1: Add multipath capabilities to pNFS flexfiles servers over NFSv3
SUNRPC: Allow addition of new transports to a struct rpc_clnt
NFSv4.1: nfs4_proc_bind_conn_to_session must iterate over all connections
SUNRPC: Make NFS swap work with multipath
...
NFS: NFSoRDMA Client Side Changes
These patches include several bugfixes and cleanups for the NFSoRDMA client.
This includes bugfixes for NFS v4.1, proper RDMA_ERROR handling, and fixes
from the recent workqueue swicchover. These patches also switch xprtrdma to
use the new CQ API
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
* tag 'nfs-rdma-4.6-1' of git://git.linux-nfs.org/projects/anna/nfs-rdma: (787 commits)
xprtrdma: Use new CQ API for RPC-over-RDMA client send CQs
xprtrdma: Use an anonymous union in struct rpcrdma_mw
xprtrdma: Use new CQ API for RPC-over-RDMA client receive CQs
xprtrdma: Serialize credit accounting again
xprtrdma: Properly handle RDMA_ERROR replies
rpcrdma: Add RPCRDMA_HDRLEN_ERR
xprtrdma: Do not wait if ib_post_send() fails
xprtrdma: Segment head and tail XDR buffers on page boundaries
xprtrdma: Clean up dprintk format string containing a newline
xprtrdma: Clean up physical_op_map()
xprtrdma: Clean up unused RPCRDMA_INLINE_PAD_THRESH macro
Error headers are shorter than either RDMA_MSG or RDMA_NOMSG.
Since HDRLEN_MIN is already used in several other places that would
be annoying to change, add RPCRDMA_HDRLEN_ERR for the one or two
spots where the shorter length is needed.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Devesh Sharma <devesh.sharma@broadcom.com>
Reviewed-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Calling ib_poll_cq() to sort through WCs during a completion is a
common pattern amongst RDMA consumers. Since commit 14d3a3b249
("IB: add a proper completion queue abstraction"), WC sorting can
be handled by the IB core.
By converting to this new API, svcrdma is made a better neighbor to
other RDMA consumers, as it allows the core to schedule the delivery
of completions more fairly amongst all active consumers.
This new API also aims each completion at a function that is
specific to the WR's opcode. Thus the ctxt->wr_op field and the
switch in process_context is replaced by a set of methods that
handle each completion type.
Because each ib_cqe carries a pointer to a completion method, the
core can now post operations on a consumer's QP, and handle the
completions itself.
The server's rdma_stat_sq_poll and rdma_stat_sq_prod metrics are no
longer updated.
As a clean up, the cq_event_handler, the dto_tasklet, and all
associated locking is removed, as they are no longer referenced or
used.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Tested-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Calling ib_poll_cq() to sort through WCs during a completion is a
common pattern amongst RDMA consumers. Since commit 14d3a3b249
("IB: add a proper completion queue abstraction"), WC sorting can
be handled by the IB core.
By converting to this new API, svcrdma is made a better neighbor to
other RDMA consumers, as it allows the core to schedule the delivery
of completions more fairly amongst all active consumers.
Because each ib_cqe carries a pointer to a completion method, the
core can now post operations on a consumer's QP, and handle the
completions itself.
svcrdma receive completions no longer use the dto_tasklet. Each
polled Receive WC is now handled individually in soft IRQ context.
The server transport's rdma_stat_rq_poll and rdma_stat_rq_prod
metrics are no longer updated.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
When constructing an error reply, svc_rdma_xdr_encode_error()
needs to view the client's request message so it can get the
failing request's XID.
svc_rdma_xdr_decode_req() is supposed to return a pointer to the
client's request header. But if it fails to decode the client's
message (and thus an error reply is needed) it does not return the
pointer. The server then sends a bogus XID in the error reply.
Instead, unconditionally generate the pointer to the client's header
in svc_rdma_recvfrom(), and pass that pointer to both functions.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Devesh Sharma <devesh.sharma@broadcom.com>
Tested-by: Devesh Sharma <devesh.sharma@broadcom.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Fix several issues with svc_rdma_send_error():
- Post a receive buffer to replace the one that was consumed by
the incoming request
- Posting a send should use DMA_TO_DEVICE, not DMA_FROM_DEVICE
- No need to put_page _and_ free pages in svc_rdma_put_context
- Make sure the sge is set up completely in case the error
path goes through svc_rdma_unmap_dma()
- Replace the use of ENOSYS, which has a reserved meaning
Related fixes in svc_rdma_recvfrom():
- Don't leak the ctxt associated with the incoming request
- Don't close the connection after sending an error reply
- Let svc_rdma_send_error() figure out the right header error code
As a last clean up, move svc_rdma_send_error() to svc_rdma_sendto.c
with other similar functions. There is some common logic in these
functions that could someday be combined to reduce code duplication.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Devesh Sharma <devesh.sharma@broadcom.com>
Tested-by: Devesh Sharma <devesh.sharma@broadcom.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Error headers are shorter than either RDMA_MSG or RDMA_NOMSG.
Since HDRLEN_MIN is already used in several other places that would
be annoying to change, add RPCRDMA_HDRLEN_ERR for the one or two
spots where the shorter length is needed.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Devesh Sharma <devesh.sharma@broadcom.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
