A new argument is added to __skb_recv_datagram to provide
an explicit skb destructor, invoked under the receive queue
lock.
The UDP protocol uses such argument to perform memory
reclaiming on dequeue, so that the UDP protocol does not
set anymore skb->desctructor.
Instead explicit memory reclaiming is performed at close() time and
when skbs are removed from the receive queue.
The in kernel UDP protocol users now need to call a
skb_recv_udp() variant instead of skb_recv_datagram() to
properly perform memory accounting on dequeue.
Overall, this allows acquiring only once the receive queue
lock on dequeue.
Tested using pktgen with random src port, 64 bytes packet,
wire-speed on a 10G link as sender and udp_sink as the receiver,
using an l4 tuple rxhash to stress the contention, and one or more
udp_sink instances with reuseport.
nr sinks vanilla patched
1 440 560
3 2150 2300
6 3650 3800
9 4450 4600
12 6250 6450
v1 -> v2:
- do rmem and allocated memory scheduling under the receive lock
- do bulk scheduling in first_packet_length() and in udp_destruct_sock()
- avoid the typdef for the dequeue callback
Suggested-by: Eric Dumazet <edumazet@google.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Completely avoid default sock memory accounting and replace it
with udp-specific accounting.
Since the new memory accounting model encapsulates completely
the required locking, remove the socket lock on both enqueue and
dequeue, and avoid using the backlog on enqueue.
Be sure to clean-up rx queue memory on socket destruction, using
udp its own sk_destruct.
Tested using pktgen with random src port, 64 bytes packet,
wire-speed on a 10G link as sender and udp_sink as the receiver,
using an l4 tuple rxhash to stress the contention, and one or more
udp_sink instances with reuseport.
nr readers Kpps (vanilla) Kpps (patched)
1 170 440
3 1250 2150
6 3000 3650
9 4200 4450
12 5700 6250
v4 -> v5:
- avoid unneeded test in first_packet_length
v3 -> v4:
- remove useless sk_rcvqueues_full() call
v2 -> v3:
- do not set the now unsed backlog_rcv callback
v1 -> v2:
- add memory pressure support
- fixed dropwatch accounting for ipv6
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Highlights include:
Stable bugfixes:
- sunrpc: fix writ espace race causing stalls
- NFS: Fix inode corruption in nfs_prime_dcache()
- NFSv4: Don't report revoked delegations as valid in
nfs_have_delegation()
- NFSv4: nfs4_copy_delegation_stateid() must fail if the delegation is
invalid
- NFSv4: Open state recovery must account for file permission changes
- NFSv4.2: Fix a reference leak in nfs42_proc_layoutstats_generic
Features:
- Add support for tracking multiple layout types with an ordered list
- Add support for using multiple backchannel threads on the client
- Add support for pNFS file layout session trunking
- Delay xprtrdma use of DMA API (for device driver removal)
- Add support for xprtrdma remote invalidation
- Add support for larger xprtrdma inline thresholds
- Use a scatter/gather list for sending xprtrdma RPC calls
- Add support for the CB_NOTIFY_LOCK callback
- Improve hashing sunrpc auth_creds by using both uid and gid
Bugfixes:
- Fix xprtrdma use of DMA API
- Validate filenames before adding to the dcache
- Fix corruption of xdr->nwords in xdr_copy_to_scratch
- Fix setting buffer length in xdr_set_next_buffer()
- Don't deadlock the state manager on the SEQUENCE status flags
- Various delegation and stateid related fixes
- Retry operations if an interrupted slot receives EREMOTEIO
- Make nfs boot time y2038 safe
-----BEGIN PGP SIGNATURE-----
iQIcBAABCAAGBQJX/+ZfAAoJENfLVL+wpUDr5MUP/16s2Kp9ZZZZ7ICi3yrHOzb0
9WpCOmbKUIELXl8YgkxlvPUYMzTQTIc32TwbVgdFV0g41my/0+O3z3+IiTrUGxH5
8LgouMWBZ9KKmyUB//+KQAXr3j/bvDdF6Li6wJfz8a2o+9xT4oTkK1+Js8p0kn6e
HNKfRknfCKwvE+j4tPCLfs2RX5qDyBFILXwWhj1fAbmT3rbnp+QqkXD4mWUrXb9z
DBgxciXRhOkOQQAD2KQBFd2kUqWDZ5ED23b+aYsu9D3VCW45zitBqQFAxkQWL0hp
x8Mp+MDCxlgdEaGQPUmUiDtPkG1X9ZxUJCAwaJWWsZaItwR2Il+en2sETctnTZ1X
0IAxZVFdolzSeLzIfNx3OG32JdWJdaNjUzkIZam8gO6i1f6PAmK4alR0J3CT31nJ
/OEN76o1E7acGWRMmj+MAZ2U5gPfR7EitOzyE8ZUPcHgyeGMiynjwi56WIpeSvT2
F/Sp5kRe5+D5gtnYuppGp7Srp5vYdtFaz1zgPDUKpDLcxfDweO8AHGjJf3Zmrunx
X24yia4A14CnfcUy4vKpISXRykmkG/3Z0tpWwV53uXZm4nlQfRc7gPibiW7Ay521
af8sDoItW98K3DK5NQU7IUn83ua1TStzpoqlAEafRw//g9zPMTbhHvNvOyrRfrcX
kjWn6hNblMu9M34JOjtu
=XOrF
-----END PGP SIGNATURE-----
Merge tag 'nfs-for-4.9-1' of git://git.linux-nfs.org/projects/anna/linux-nfs
Pull NFS client updates from Anna Schumaker:
"Highlights include:
Stable bugfixes:
- sunrpc: fix writ espace race causing stalls
- NFS: Fix inode corruption in nfs_prime_dcache()
- NFSv4: Don't report revoked delegations as valid in nfs_have_delegation()
- NFSv4: nfs4_copy_delegation_stateid() must fail if the delegation is invalid
- NFSv4: Open state recovery must account for file permission changes
- NFSv4.2: Fix a reference leak in nfs42_proc_layoutstats_generic
Features:
- Add support for tracking multiple layout types with an ordered list
- Add support for using multiple backchannel threads on the client
- Add support for pNFS file layout session trunking
- Delay xprtrdma use of DMA API (for device driver removal)
- Add support for xprtrdma remote invalidation
- Add support for larger xprtrdma inline thresholds
- Use a scatter/gather list for sending xprtrdma RPC calls
- Add support for the CB_NOTIFY_LOCK callback
- Improve hashing sunrpc auth_creds by using both uid and gid
Bugfixes:
- Fix xprtrdma use of DMA API
- Validate filenames before adding to the dcache
- Fix corruption of xdr->nwords in xdr_copy_to_scratch
- Fix setting buffer length in xdr_set_next_buffer()
- Don't deadlock the state manager on the SEQUENCE status flags
- Various delegation and stateid related fixes
- Retry operations if an interrupted slot receives EREMOTEIO
- Make nfs boot time y2038 safe"
* tag 'nfs-for-4.9-1' of git://git.linux-nfs.org/projects/anna/linux-nfs: (100 commits)
NFSv4.2: Fix a reference leak in nfs42_proc_layoutstats_generic
fs: nfs: Make nfs boot time y2038 safe
sunrpc: replace generic auth_cred hash with auth-specific function
sunrpc: add RPCSEC_GSS hash_cred() function
sunrpc: add auth_unix hash_cred() function
sunrpc: add generic_auth hash_cred() function
sunrpc: add hash_cred() function to rpc_authops struct
Retry operation on EREMOTEIO on an interrupted slot
pNFS: Fix atime updates on pNFS clients
sunrpc: queue work on system_power_efficient_wq
NFSv4.1: Even if the stateid is OK, we may need to recover the open modes
NFSv4: If recovery failed for a specific open stateid, then don't retry
NFSv4: Fix retry issues with nfs41_test/free_stateid
NFSv4: Open state recovery must account for file permission changes
NFSv4: Mark the lock and open stateids as invalid after freeing them
NFSv4: Don't test open_stateid unless it is set
NFSv4: nfs4_do_handle_exception() handle revoke/expiry of a single stateid
NFS: Always call nfs_inode_find_state_and_recover() when revoking a delegation
NFSv4: Fix a race when updating an open_stateid
NFSv4: Fix a race in nfs_inode_reclaim_delegation()
...
benefit from user testing:
Anna Schumacker contributed a simple NFSv4.2 COPY implementation. COPY
is already supported on the client side, so a call to copy_file_range()
on a recent client should now result in a server-side copy that doesn't
require all the data to make a round trip to the client and back.
Jeff Layton implemented callbacks to notify clients when contended locks
become available, which should reduce latency on workloads with
contended locks.
-----BEGIN PGP SIGNATURE-----
iQIcBAABAgAGBQJX/mcsAAoJECebzXlCjuG+MU0P/3SzTLGYXU5yOTAorx255/uf
fUVKQQhTzzaA2xj3gWWWztYx3y0ZJUVgwU56a+Ap5Z8/goqDQ78H+ePEc+MG7BT/
/UXS/bITvt0MP/dvPrDzhSltvqx/wpelLPBo29hGLlAQ2dsnD4Y75IbOOQccWqcC
iD2v6x7lnpWZ7j9Zhwzg/JNQHwISIb7tiLoYBjfcdNDEMU76KIyhxD0Cx9MSeBzH
9Rq/oEdwGDFS5WqVfNe2jxbngoauq1IupziQ2eQGv2D/POyXCx8fphoYjDz1XaW8
PxaJtJtM2owPGG+z2CxklJqNaS1Z4F+oppjg+nf4i/ibxmIBaTy8NluASX3vMh69
CDO1+ly+TiF0l1VqMOQJWRnqn1qGk6fLpF6P1Ac62B0oWpeLGU7nmik7XN1ORgsi
8ksxRKNAWeprZo3wl5xNrADu/wlZ7XCJTc4QoHEgYT04aHF+j8EMCHv+mtZ8+Bwn
WWiA8iItZOgXV4vitCRJlvsixjYvmF3djPIoI2Lt5KDWIg+eL89sKwzTALSfeC4m
Vjb0svzPX1MmZCNP1rCStFbl3gZYXZyqPk+uA6M7H8mjAjVeKxRPowWpMBgvYZHr
FjCPb878bAuqCeBVbIyOLLcKWBLTw8PsUWZAor3gNg454JGkMjLUyJ/S22Cz5Nbo
HdjoiTJtbPrHnCwTMXwa
=nozl
-----END PGP SIGNATURE-----
Merge tag 'nfsd-4.9' of git://linux-nfs.org/~bfields/linux
Pull nfsd updates from Bruce Fields:
"Some RDMA work and some good bugfixes, and two new features that could
benefit from user testing:
- Anna Schumacker contributed a simple NFSv4.2 COPY implementation.
COPY is already supported on the client side, so a call to
copy_file_range() on a recent client should now result in a
server-side copy that doesn't require all the data to make a round
trip to the client and back.
- Jeff Layton implemented callbacks to notify clients when contended
locks become available, which should reduce latency on workloads
with contended locks"
* tag 'nfsd-4.9' of git://linux-nfs.org/~bfields/linux:
NFSD: Implement the COPY call
nfsd: handle EUCLEAN
nfsd: only WARN once on unmapped errors
exportfs: be careful to only return expected errors.
nfsd4: setclientid_confirm with unmatched verifier should fail
nfsd: randomize SETCLIENTID reply to help distinguish servers
nfsd: set the MAY_NOTIFY_LOCK flag in OPEN replies
nfs: add a new NFS4_OPEN_RESULT_MAY_NOTIFY_LOCK constant
nfsd: add a LRU list for blocked locks
nfsd: have nfsd4_lock use blocking locks for v4.1+ locks
nfsd: plumb in a CB_NOTIFY_LOCK operation
NFSD: fix corruption in notifier registration
svcrdma: support Remote Invalidation
svcrdma: Server-side support for rpcrdma_connect_private
rpcrdma: RDMA/CM private message data structure
svcrdma: Skip put_page() when send_reply() fails
svcrdma: Tail iovec leaves an orphaned DMA mapping
nfsd: fix dprintk in nfsd4_encode_getdeviceinfo
nfsd: eliminate cb_minorversion field
nfsd: don't set a FL_LAYOUT lease for flexfiles layouts
Pull more vfs updates from Al Viro:
">rename2() work from Miklos + current_time() from Deepa"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
fs: Replace current_fs_time() with current_time()
fs: Replace CURRENT_TIME_SEC with current_time() for inode timestamps
fs: Replace CURRENT_TIME with current_time() for inode timestamps
fs: proc: Delete inode time initializations in proc_alloc_inode()
vfs: Add current_time() api
vfs: add note about i_op->rename changes to porting
fs: rename "rename2" i_op to "rename"
vfs: remove unused i_op->rename
fs: make remaining filesystems use .rename2
libfs: support RENAME_NOREPLACE in simple_rename()
fs: support RENAME_NOREPLACE for local filesystems
ncpfs: fix unused variable warning
- Updates to mlx5
- Updates to mlx4 (two conflicts, both minor and easily resolved)
- Updates to iw_cxgb4 (one conflict, not so obvious to resolve, proper
resolution is to keep the code in cxgb4_main.c as it is in Linus'
tree as attach_uld was refactored and moved into cxgb4_uld.c)
- Improvements to uAPI (moved vendor specific API elements to uAPI area)
- Add hns-roce driver and hns and hns-roce ACPI reset support
- Conversion of all rdma code away from deprecated
create_singlethread_workqueue
- Security improvement: remove unsafe ib_get_dma_mr (breaks lustre in
staging)
-----BEGIN PGP SIGNATURE-----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=LC+S
-----END PGP SIGNATURE-----
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma
Pull main rdma updates from Doug Ledford:
"This is the main pull request for the rdma stack this release. The
code has been through 0day and I had it tagged for linux-next testing
for a couple days.
Summary:
- updates to mlx5
- updates to mlx4 (two conflicts, both minor and easily resolved)
- updates to iw_cxgb4 (one conflict, not so obvious to resolve,
proper resolution is to keep the code in cxgb4_main.c as it is in
Linus' tree as attach_uld was refactored and moved into
cxgb4_uld.c)
- improvements to uAPI (moved vendor specific API elements to uAPI
area)
- add hns-roce driver and hns and hns-roce ACPI reset support
- conversion of all rdma code away from deprecated
create_singlethread_workqueue
- security improvement: remove unsafe ib_get_dma_mr (breaks lustre in
staging)"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma: (75 commits)
staging/lustre: Disable InfiniBand support
iw_cxgb4: add fast-path for small REG_MR operations
cxgb4: advertise support for FR_NSMR_TPTE_WR
IB/core: correctly handle rdma_rw_init_mrs() failure
IB/srp: Fix infinite loop when FMR sg[0].offset != 0
IB/srp: Remove an unused argument
IB/core: Improve ib_map_mr_sg() documentation
IB/mlx4: Fix possible vl/sl field mismatch in LRH header in QP1 packets
IB/mthca: Move user vendor structures
IB/nes: Move user vendor structures
IB/ocrdma: Move user vendor structures
IB/mlx4: Move user vendor structures
IB/cxgb4: Move user vendor structures
IB/cxgb3: Move user vendor structures
IB/mlx5: Move and decouple user vendor structures
IB/{core,hw}: Add constant for node_desc
ipoib: Make ipoib_warn ratelimited
IB/mlx4/alias_GUID: Remove deprecated create_singlethread_workqueue
IB/ipoib_verbs: Remove deprecated create_singlethread_workqueue
IB/ipoib: Remove deprecated create_singlethread_workqueue
...
Current supplementary groups code can massively overallocate memory and
is implemented in a way so that access to individual gid is done via 2D
array.
If number of gids is <= 32, memory allocation is more or less tolerable
(140/148 bytes). But if it is not, code allocates full page (!)
regardless and, what's even more fun, doesn't reuse small 32-entry
array.
2D array means dependent shifts, loads and LEAs without possibility to
optimize them (gid is never known at compile time).
All of the above is unnecessary. Switch to the usual
trailing-zero-len-array scheme. Memory is allocated with
kmalloc/vmalloc() and only as much as needed. Accesses become simpler
(LEA 8(gi,idx,4) or even without displacement).
Maximum number of gids is 65536 which translates to 256KB+8 bytes. I
think kernel can handle such allocation.
On my usual desktop system with whole 9 (nine) aux groups, struct
group_info shrinks from 148 bytes to 44 bytes, yay!
Nice side effects:
- "gi->gid[i]" is shorter than "GROUP_AT(gi, i)", less typing,
- fix little mess in net/ipv4/ping.c
should have been using GROUP_AT macro but this point becomes moot,
- aux group allocation is persistent and should be accounted as such.
Link: http://lkml.kernel.org/r/20160817201927.GA2096@p183.telecom.by
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Vasily Kulikov <segoon@openwall.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Replace the generic code to hash the auth_cred with the call to
the auth-specific hash function in the rpc_authops struct.
Signed-off-by: Frank Sorenson <sorenson@redhat.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Add a hash_cred() function for RPCSEC_GSS, using only the
uid from the auth_cred.
Signed-off-by: Frank Sorenson <sorenson@redhat.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Add a hash_cred() function for auth_unix, using both the
uid and gid from the auth_cred.
Signed-off-by: Frank Sorenson <sorenson@redhat.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Add a hash_cred() function for generic_auth, using both the
uid and gid from the auth_cred.
Signed-off-by: Frank Sorenson <sorenson@redhat.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
CURRENT_TIME macro is not appropriate for filesystems as it
doesn't use the right granularity for filesystem timestamps.
Use current_time() instead.
CURRENT_TIME is also not y2038 safe.
This is also in preparation for the patch that transitions
vfs timestamps to use 64 bit time and hence make them
y2038 safe. As part of the effort current_time() will be
extended to do range checks. Hence, it is necessary for all
file system timestamps to use current_time(). Also,
current_time() will be transitioned along with vfs to be
y2038 safe.
Note that whenever a single call to current_time() is used
to change timestamps in different inodes, it is because they
share the same time granularity.
Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Felipe Balbi <balbi@kernel.org>
Acked-by: Steven Whitehouse <swhiteho@redhat.com>
Acked-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Acked-by: David Sterba <dsterba@suse.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
sunrpc uses workqueue to clean cache regulary. There is no real dependency
of executing work on the cpu which queueing it.
On a idle system, especially for a heterogeneous systems like big.LITTLE,
it is observed that the big idle cpu was woke up many times just to service
this work, which against the principle of power saving. It would be better
if we can schedule it on a cpu which the scheduler believes to be the most
appropriate one.
After apply this patch, system_wq will be replaced by
system_power_efficient_wq for sunrpc. This functionality is enabled when
CONFIG_WQ_POWER_EFFICIENT is selected.
Signed-off-by: Ke Wang <ke.wang@spreadtrum.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Instead of exposing ib_get_dma_mr to ULPs and letting them use it more or
less unchecked, this moves the capability of creating a global rkey into
the RDMA core, where it can be easily audited. It also prints a warning
everytime this feature is used as well.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Reviewed-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Support Remote Invalidation. A private message is exchanged with
the client upon RDMA transport connect that indicates whether
Send With Invalidation may be used by the server to send RPC
replies. The invalidate_rkey is arbitrarily chosen from among
rkeys present in the RPC-over-RDMA header's chunk lists.
Send With Invalidate improves performance only when clients can
recognize, while processing an RPC reply, that an rkey has already
been invalidated. That has been submitted as a separate change.
In the future, the RPC-over-RDMA protocol might support Remote
Invalidation properly. The protocol needs to enable signaling
between peers to indicate when Remote Invalidation can be used
for each individual RPC.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Prepare to receive an RDMA-CM private message when handling a new
connection attempt, and send a similar message as part of connection
acceptance.
Both sides can communicate their various implementation limits.
Implementations that don't support this sideband protocol ignore it.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Message from syslogd@klimt at Aug 18 17:00:37 ...
kernel:page:ffffea0020639b00 count:0 mapcount:0 mapping: (null) index:0x0
Aug 18 17:00:37 klimt kernel: flags: 0x2fffff80000000()
Aug 18 17:00:37 klimt kernel: page dumped because: VM_BUG_ON_PAGE(page_ref_count(page) == 0)
Aug 18 17:00:37 klimt kernel: kernel BUG at /home/cel/src/linux/linux-2.6/include/linux/mm.h:445!
Aug 18 17:00:37 klimt kernel: RIP: 0010:[<ffffffffa05c21c1>] svc_rdma_sendto+0x641/0x820 [rpcrdma]
send_reply() assigns its page argument as the first page of ctxt. On
error, send_reply() already invokes svc_rdma_put_context(ctxt, 1);
which does a put_page() on that very page. No need to do that again
as svc_rdma_sendto exits.
Fixes: 3e1eeb9808 ("svcrdma: Close connection when a send error occurs")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
The ctxt's count field is overloaded to mean the number of pages in
the ctxt->page array and the number of SGEs in the ctxt->sge array.
Typically these two numbers are the same.
However, when an inline RPC reply is constructed from an xdr_buf
with a tail iovec, the head and tail often occupy the same page,
but each are DMA mapped independently. In that case, ->count equals
the number of pages, but it does not equal the number of SGEs.
There's one more SGE, for the tail iovec. Hence there is one more
DMA mapping than there are pages in the ctxt->page array.
This isn't a real problem until the server's iommu is enabled. Then
each RPC reply that has content in that iovec orphans a DMA mapping
that consists of real resources.
krb5i and krb5p always populate that tail iovec. After a couple
million sent krb5i/p RPC replies, the NFS server starts behaving
erratically. Reboot is needed to clear the problem.
Fixes: 9d11b51ce7 ("svcrdma: Fix send_reply() scatter/gather set-up")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
There is only one waiter for the completion, therefore there
is no need to use complete_all(). Let's make that clear by
using complete() instead of complete_all().
The usage pattern of the completion is:
waiter context waker context
frwr_op_unmap_sync()
reinit_completion()
ib_post_send()
wait_for_completion()
frwr_wc_localinv_wake()
complete()
Signed-off-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Cc: Anna Schumaker <Anna.Schumaker@Netapp.com>
Cc: Trond Myklebust <trond.myklebust@primarydata.com>
Cc: Chuck Lever <chuck.lever@oracle.com>
Cc: linux-nfs@vger.kernel.org
Cc: netdev@vger.kernel.org
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Use xdr->nwords to tell us how much buffer remains.
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
When we copy the first part of the data, we need to ensure that value
of xdr->nwords is updated as well. Do so by calling __xdr_inline_decode()
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Write space becoming available may race with putting the task to sleep
in xprt_wait_for_buffer_space(). The existing mechanism to avoid the
race does not work.
This (edited) partial trace illustrates the problem:
[1] rpc_task_run_action: task:43546@5 ... action=call_transmit
[2] xs_write_space <-xs_tcp_write_space
[3] xprt_write_space <-xs_write_space
[4] rpc_task_sleep: task:43546@5 ...
[5] xs_write_space <-xs_tcp_write_space
[1] Task 43546 runs but is out of write space.
[2] Space becomes available, xs_write_space() clears the
SOCKWQ_ASYNC_NOSPACE bit.
[3] xprt_write_space() attemts to wake xprt->snd_task (== 43546), but
this has not yet been queued and the wake up is lost.
[4] xs_nospace() is called which calls xprt_wait_for_buffer_space()
which queues task 43546.
[5] The call to sk->sk_write_space() at the end of xs_nospace() (which
is supposed to handle the above race) does not call
xprt_write_space() as the SOCKWQ_ASYNC_NOSPACE bit is clear and
thus the task is not woken.
Fix the race by resetting the SOCKWQ_ASYNC_NOSPACE bit in xs_nospace()
so the second call to sk->sk_write_space() calls xprt_write_space().
Suggested-by: Trond Myklebust <trondmy@primarydata.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
cc: stable@vger.kernel.org # 4.4
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: the extra layer of indirection doesn't add value.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: When converting xprtrdma to use the new CQ API, I missed a
spot. The naming convention elsewhere is:
{svc_rdma,rpcrdma}_wc_{operation}
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Tie frwr debugging messages together by always reporting the address
of the frwr.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
The Version One default inline threshold is still 1KB. But allow
testing with thresholds up to 64KB.
This maximum is somewhat arbitrary. There's no fundamental
architectural limit I'm aware of, but it's good to keep the size of
Receive buffers reasonable. Now that Send can use a s/g list, a
Send buffer is only as large as each RPC requires. Receive buffers
are always the size of the inline threshold, however.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
An RPC Call message that is sent inline but that has a data payload
(ie, one or more items in rq_snd_buf's page list) must be "pulled
up:"
- call_allocate has to reserve enough RPC Call buffer space to
accommodate the data payload
- call_transmit has to memcopy the rq_snd_buf's page list and tail
into its head iovec before it is sent
As the inline threshold is increased beyond its current 1KB default,
however, this means data payloads of more than a few KB are copied
by the host CPU. For example, if the inline threshold is increased
just to 4KB, then NFS WRITE requests up to 4KB would involve a
memcpy of the NFS WRITE's payload data into the RPC Call buffer.
This is an undesirable amount of participation by the host CPU.
The inline threshold may be much larger than 4KB in the future,
after negotiation with a peer server.
Instead of copying the components of rq_snd_buf into its head iovec,
construct a gather list of these components, and send them all in
place. The same approach is already used in the Linux server's
RPC-over-RDMA reply path.
This mechanism also eliminates the need for rpcrdma_tail_pullup,
which is used to manage the XDR pad and trailing inline content when
a Read list is present.
This requires that the pages in rq_snd_buf's page list be DMA-mapped
during marshaling, and unmapped when a data-bearing RPC is
completed. This is slightly less efficient for very small I/O
payloads, but significantly more efficient as data payload size and
inline threshold increase past a kilobyte.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Have frwr's ro_unmap_sync recognize an invalidated rkey that appears
as part of a Receive completion. Local invalidation can be skipped
for that rkey.
Use an out-of-band signaling mechanism to indicate to the server
that the client is prepared to receive RDMA Send With Invalidate.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Send an RDMA-CM private message on connect, and look for one during
a connection-established event.
Both sides can communicate their various implementation limits.
Implementations that don't support this sideband protocol ignore it.
Once the client knows the server's inline threshold maxima, it can
adjust the use of Reply chunks, and eliminate most use of Position
Zero Read chunks. Moderately-sized I/O can be done using a pure
inline RDMA Send instead of RDMA operations that require memory
registration.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: The fields in the recv_wr do not vary. There is no need to
initialize them before each ib_post_recv(). This removes a large-ish
data structure from the stack.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: Most of the fields in each send_wr do not vary. There is
no need to initialize them before each ib_post_send(). This removes
a large-ish data structure from the stack.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up.
Since commit fc66448549 ("xprtrdma: Split the completion queue"),
rpcrdma_ep_post_recv() no longer uses the "ep" argument.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up. The "ia" argument is no longer used.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Currently, each regbuf is allocated and DMA mapped at the same time.
This is done during transport creation.
When a device driver is unloaded, every DMA-mapped buffer in use by
a transport has to be unmapped, and then remapped to the new
device if the driver is loaded again. Remapping will have to be done
_after_ the connect worker has set up the new device.
But there's an ordering problem:
call_allocate, which invokes xprt_rdma_allocate which calls
rpcrdma_alloc_regbuf to allocate Send buffers, happens _before_
the connect worker can run to set up the new device.
Instead, at transport creation, allocate each buffer, but leave it
unmapped. Once the RPC carries these buffers into ->send_request, by
which time a transport connection should have been established,
check to see that the RPC's buffers have been DMA mapped. If not,
map them there.
When device driver unplug support is added, it will simply unmap all
the transport's regbufs, but it doesn't have to deallocate the
underlying memory.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
The use of DMA_BIDIRECTIONAL is discouraged by DMA-API.txt.
Fortunately, xprtrdma now knows which direction I/O is going as
soon as it allocates each regbuf.
The RPC Call and Reply buffers are no longer the same regbuf. They
can each be labeled correctly now. The RPC Reply buffer is never
part of either a Send or Receive WR, but it can be part of Reply
chunk, which is mapped and registered via ->ro_map . So it is not
DMA mapped when it is allocated (DMA_NONE), to avoid a double-
mapping.
Since Receive buffers are no longer DMA_BIDIRECTIONAL and their
contents are never modified by the host CPU, DMA-API-HOWTO.txt
suggests that a DMA sync before posting each buffer should be
unnecessary. (See my_card_interrupt_handler).
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Commit 949317464b ("xprtrdma: Limit number of RDMA segments in
RPC-over-RDMA headers") capped the number of chunks that may appear
in RPC-over-RDMA headers. The maximum header size can be estimated
and fixed to avoid allocating buffer space that is never used.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
RPC-over-RDMA needs to separate its RPC call and reply buffers.
o When an RPC Call is sent, rq_snd_buf is DMA mapped for an RDMA
Send operation using DMA_TO_DEVICE
o If the client expects a large RPC reply, it DMA maps rq_rcv_buf
as part of a Reply chunk using DMA_FROM_DEVICE
The two mappings are for data movement in opposite directions.
DMA-API.txt suggests that if these mappings share a DMA cacheline,
bad things can happen. This could occur in the final bytes of
rq_snd_buf and the first bytes of rq_rcv_buf if the two buffers
happen to share a DMA cacheline.
On x86_64 the cacheline size is typically 8 bytes, and RPC call
messages are usually much smaller than the send buffer, so this
hasn't been a noticeable problem. But the DMA cacheline size can be
larger on other platforms.
Also, often rq_rcv_buf starts most of the way into a page, thus
an additional RDMA segment is needed to map and register the end of
that buffer. Try to avoid that scenario to reduce the cost of
registering and invalidating Reply chunks.
Instead of carrying a single regbuf that covers both rq_snd_buf and
rq_rcv_buf, each struct rpcrdma_req now carries one regbuf for
rq_snd_buf and one regbuf for rq_rcv_buf.
Some incidental changes worth noting:
- To clear out some spaghetti, refactor xprt_rdma_allocate.
- The value stored in rg_size is the same as the value stored in
the iov.length field, so eliminate rg_size
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Currently there's a hidden and indirect mechanism for finding the
rpcrdma_req that goes with an rpc_rqst. It depends on getting from
the rq_buffer pointer in struct rpc_rqst to the struct
rpcrdma_regbuf that controls that buffer, and then to the struct
rpcrdma_req it goes with.
This was done back in the day to avoid the need to add a per-rqst
pointer or to alter the buf_free API when support for RPC-over-RDMA
was introduced.
I'm about to change the way regbuf's work to support larger inline
thresholds. Now is a good time to replace this indirect mechanism
with something that is more straightforward. I guess this should be
considered a clean up.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
For xprtrdma, the RPC Call and Reply buffers are involved in real
I/O operations.
To start with, the DMA direction of the I/O for a Call is opposite
that of a Reply.
In the current arrangement, the Reply buffer address is on a
four-byte alignment just past the call buffer. Would be friendlier
on some platforms if that was at a DMA cache alignment instead.
Because the current arrangement allocates a single memory region
which contains both buffers, the RPC Reply buffer often contains a
page boundary in it when the Call buffer is large enough (which is
frequent).
It would be a little nicer for setting up DMA operations (and
possible registration of the Reply buffer) if the two buffers were
separated, well-aligned, and contained as few page boundaries as
possible.
Now, I could just pad out the single memory region used for the pair
of buffers. But frequently that would mean a lot of unused space to
ensure the Reply buffer did not have a page boundary.
Add a separate pointer to rpc_rqst that points right to the RPC
Reply buffer. This makes no difference to xprtsock, but it will help
xprtrdma in subsequent patches.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
xprtrdma needs to allocate the Call and Reply buffers separately.
TBH, the reliance on using a single buffer for the pair of XDR
buffers is transport implementation-specific.
Instead of passing just the rq_buffer into the buf_free method, pass
the task structure and let buf_free take care of freeing both
XDR buffers at once.
There's a micro-optimization here. In the common case, both
xprt_release and the transport's buf_free method were checking if
rq_buffer was NULL. Now the check is done only once per RPC.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
xprtrdma needs to allocate the Call and Reply buffers separately.
TBH, the reliance on using a single buffer for the pair of XDR
buffers is transport implementation-specific.
Transports that want to allocate separate Call and Reply buffers
will ignore the "size" argument anyway. Don't bother passing it.
The buf_alloc method can't return two pointers. Instead, make the
method's return value an error code, and set the rq_buffer pointer
in the method itself.
This gives call_allocate an opportunity to terminate an RPC instead
of looping forever when a permanent problem occurs. If a request is
just bogus, or the transport is in a state where it can't allocate
resources for any request, there needs to be a way to kill the RPC
right there and not loop.
This immediately fixes a rare problem in the backchannel send path,
which loops if the server happens to send a CB request whose
call+reply size is larger than a page (which it shouldn't do yet).
One more issue: looks like xprt_inject_disconnect was incorrectly
placed in the failure path in call_allocate. It needs to be in the
success path, as it is for other call-sites.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: there is some XDR initialization logic that is common
to the forward channel and backchannel. Move it to an XDR header
so it can be shared.
rpc_rqst::rq_buffer points to a buffer containing big-endian data.
Update its annotation as part of the clean up.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up: r_xprt is already available everywhere these macros are
invoked, so just dereference that directly.
RPCRDMA_INLINE_PAD_VALUE is no longer used, so it can simply be
removed.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Use a setup function to call into the NFS layer to test an rpc_xprt
for session trunking so as to not leak the rpc_xprt_switch into
the nfs layer.
Search for the address in the rpc_xprt_switch first so as not to
put an unnecessary EXCHANGE_ID on the wire.
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Give the NFS layer access to the rpc_xprt_switch_add_xprt function
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Give the NFS layer access to the xprt_switch_put function
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
rpc_task_set_client is only called from rpc_run_task after
rpc_new_task and rpc_task_release_client is not needed as the
task is new.
When called from rpc_new_task, rpc_task_set_client also removed the
assigned rpc_xprt which is not desired.
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
