| Age | Commit message (Collapse) | Author |
|---|
|
Previously, RDS would wait until the final send WR had completed
and then handle cleanup. With silent ops, we do not know
if an atomic, rdma, or data op will be last. This patch
handles any of these cases by keeping a pointer to the last
op in the message in m_last_op.
When the TX completion event fires, rds dispatches to per-op-type
cleanup functions, and then does whole-message cleanup, if the
last op equalled m_last_op.
This patch also moves towards having op-specific functions take
the op struct, instead of the overall rm struct.
rds_ib_connection has a pointer to keep track of a a partially-
completed data send operation. This patch changes it from an
rds_message pointer to the narrower rm_data_op pointer, and
modifies places that use this pointer as needed.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
It hasn't cropped up in the field, but this code ensures it is
impossible to issue operations that pass an rdma cookie (DEST, MAP)
in the same sendmsg call that's actually initiating rdma or atomic
ops.
Disallowing this perverse-but-technically-allowed usage makes silent
RDMA heuristics slightly easier.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Add a flag to the API so users can indicate they want
silent operations. This is needed because silent ops
cannot be used with USE_ONCE MRs, so we can't just
assume silent.
Also, change send_xmit to do atomic op before rdma op if
both are present, and centralize the hairy logic to determine if
we want to attempt silent, or not.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Also, add a comment.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
When dropping ops in the send queue, we notify the client
of failed rdma ops they asked for notifications on, but not
atomic ops. It should be for both.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
rds_message_alloc_sgs() only works when nents is nonzero.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Do not allocate sgs for data for 0-length datagrams
Set data.op_active in rds_sendmsg() instead of
rds_message_copy_from_user().
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Simplify rds_send_xmit().
Send a congestion map (via xmit_cong_map) without
decrementing send_quota.
Move resetting of conn xmit variables to end of loop.
Update comments.
Implement a special case to turn off sending an rds header
when there is an atomic op and no other data.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
For consistency.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
A big changeset, but it's all pretty dumb.
struct rds_rdma_op was already embedded in struct rm_rdma_op.
Remove rds_rdma_op and put its members in rm_rdma_op. Rename
members with "op_" prefix instead of "r_", for consistency.
Of course this breaks a lot, so fixup the code accordingly.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Add atomic_free_op function, analogous to rdma_free_op,
and call it in rds_message_purge().
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
cmsg_rdma_args just calls rdma_prepare and does a little
arg checking -- not quite enough to justify its existence.
Plus, it is the only caller of rdma_prepare().
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Also, try to better-document the locking around the
rm and its m_inc in loop.c.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Maybe things worked fine with the flow control code running
even in the non-flow-control case, but making it explicitly
conditional helps the non-fc case be easier to read.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Removed unsignaled_bytes sysctl and code to signal
based on it. I believe unsignaled_wrs is more than
sufficient for our purposes.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Now that the header always goes first, it is possible to
simplify rds_ib_xmit. Instead of having a path to handle 0-byte
dgrams and another path to handle >0, these can both be handled
in one path. This lets us eliminate xmit_populate_wr().
Rename sent to bytes_sent, to differentiate better from other
variable named "send".
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
These functions were to cope with differently ordered
sg entries depending on RDS 3.0 or 3.1+. Now that
we've dropped 3.0 compatibility we no longer need them.
Also, modify usage sites for these to refer to sge[0] or [1]
directly. Reorder code to initialize header sgs first.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
RDS 3.0 connections (in OFED 1.3 and earlier) put the
header at the end. 3.1 connections put it at the head.
The code has significant added complexity in order to
handle both configurations. In OFED 1.6 we can
drop this and simplify the code by only supporting
"header-first" configuration.
This patch checks the protocol version, and if prior
to 3.1, does not complete the connection.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
both atomics and rdmas need to convert ib-specific completion codes
into RDS status codes. Rename rds_ib_rdma_send_complete to
rds_ib_send_complete, and have it take a pointer to the function to
call with the new error code.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Tidy up some whitespace issues.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
This does not appear to be necessary.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Instead of using a constant for initiator_depth and
responder_resources, read the per-QP values when the
device is enumerated, and then use these values when creating
the connection.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Implement a CMSG-based interface to do FADD and CSWP ops.
Alter send routines to handle atomic ops.
Add atomic counters to stats.
Add xmit_atomic() to struct rds_transport
Inline rds_ib_send_unmap_rdma into unmap_rm
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
The previous code was correct, but made the assumption that
if r_notifier was non-NULL then either r_recverr or r_notify
was true. Valid, but fragile. Changed to explicitly check
r_recverr (shows up in greps for recverr now, too.)
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
rds_message_alloc_sgs() now returns correctly-initialized
sg lists, so calleds need not do this themselves.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
This eliminates a separate memory alloc, although
it is now necessary to add an "r_active" flag, since
it is no longer to use the m_rdma_op pointer as an
indicator of if an rdma op is present.
rdma SGs allocated from rm sg pool.
rds_rm_size also gets bigger. It's a little inefficient to
run through CMSGs twice, but it makes later steps a lot smoother.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
RDMA is now an intrinsic part of RDS, so it's easier to just have
a single header.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
r_m_copy_from_user used to allocate the rm as well as kernel
buffers for the data, and then copy the data in. Now, sendmsg()
allocates the rm, although the data buffer alloc still happens
in r_m_copy_from_user.
SGs are still allocated with rm, but now r_m_alloc_sgs() is
used to reserve them. This allows multiple SG lists to be
allocated from the one rm -- this is important once we also
want to alloc our rdma sgl from this pool.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
First, it looks to me like the atomic_inc is wrong.
We should be decrementing refcount only once here, no? It's
already being done by the mr_put() at the end.
Second, simplify the logic a bit by bailing early (with a warning)
if !mr.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Clearly separate rdma-related variables in rm from data-related ones.
This is in anticipation of adding atomic support.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Favor "if (foo)" style over "if (foo != NULL)".
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
This fits better in connection.c, rather than threads.c.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Do not nest m_rs_lock under c_lock
Disable interrupts in {rdma,atomic}_send_complete
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
This function has been the source of numerous bugs; it's just
too complicated. Simplified to nest spinlocks cleanly within
the second loop body, and kick out early if there are no
rms to drop.
This will be a little slower because conn lock is grabbed for
each entry instead of "caching" the lock across rms, but this
should be entirely irrelevant to fastpath performance.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
On second look at this bug (OFED #2002), it seems that the
collision is not with the retransmission queue (packet acked
by the peer), but with the local send completion. A theoretical
sequence of events (from time t0 to t3) is thought to be as
follows,
Thread #1
t0:
sock_release
rds_release
rds_send_drop_to /* wait on send completion */
t2:
rds_rdma_drop_keys() /* destroy & free all mrs */
Thread #2
t1:
rds_ib_send_cq_comp_handler
rds_ib_send_unmap_rm
rds_message_unmapped /* wake up #1 @ t0 */
t3:
rds_message_put
rds_message_purge
rds_mr_put /* memory corruption detected */
The problem with the rds_rdma_drop_keys() is it could
remove a mr's refcount more than its due (i.e. repeatedly
as long as it still remains in the tree (mr->r_refcount > 0)).
Theoretically it should remove only one reference - reference
by the tree.
/* Release any MRs associated with this socket */
while ((node = rb_first(&rs->rs_rdma_keys))) {
mr = container_of(node, struct rds_mr, r_rb_node);
if (mr->r_trans == rs->rs_transport)
mr->r_invalidate = 0;
rds_mr_put(mr);
}
I think the correct way of doing it is to remove the mr from
the tree and rds_destroy_mr it first, then a rds_mr_put()
to decrement its reference count by one. Whichever thread
holds the last reference will free the mr via rds_mr_put().
Signed-off-by: Tina Yang <tina.yang@oracle.com>
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
in_interrupt() is true in softirqs. The BUG_ONs are supposed
to check for if irqs are disabled, so we should use
BUG_ON(irqs_disabled()) instead, duh.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
|
|
Blackhole routes are used when xfrm_lookup() returns -EREMOTE (error
triggered by IKE for example), hence this kind of route is always
temporary and so we should check if a better route exists for next
packets.
Bug has been introduced by commit d11a4dc18bf41719c9f0d7ed494d295dd2973b92.
Signed-off-by: Jianzhao Wang <jianzhao.wang@6wind.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Hi,
Here is one more of these warnings and a patch below:
Sep 5 23:52:33 del kernel: [46044.244833] ===================================================
Sep 5 23:52:33 del kernel: [46044.269681] [ INFO: suspicious rcu_dereference_check() usage. ]
Sep 5 23:52:33 del kernel: [46044.277000] ---------------------------------------------------
Sep 5 23:52:33 del kernel: [46044.285185] net/ipv4/fib_trie.c:1756 invoked rcu_dereference_check() without protection!
Sep 5 23:52:33 del kernel: [46044.293627]
Sep 5 23:52:33 del kernel: [46044.293632] other info that might help us debug this:
Sep 5 23:52:33 del kernel: [46044.293634]
Sep 5 23:52:33 del kernel: [46044.325333]
Sep 5 23:52:33 del kernel: [46044.325335] rcu_scheduler_active = 1, debug_locks = 0
Sep 5 23:52:33 del kernel: [46044.348013] 1 lock held by pppd/1717:
Sep 5 23:52:33 del kernel: [46044.357548] #0: (rtnl_mutex){+.+.+.}, at: [<c125dc1f>] rtnl_lock+0xf/0x20
Sep 5 23:52:33 del kernel: [46044.367647]
Sep 5 23:52:33 del kernel: [46044.367652] stack backtrace:
Sep 5 23:52:33 del kernel: [46044.387429] Pid: 1717, comm: pppd Not tainted 2.6.35.4.4a #3
Sep 5 23:52:33 del kernel: [46044.398764] Call Trace:
Sep 5 23:52:33 del kernel: [46044.409596] [<c12f9aba>] ? printk+0x18/0x1e
Sep 5 23:52:33 del kernel: [46044.420761] [<c1053969>] lockdep_rcu_dereference+0xa9/0xb0
Sep 5 23:52:33 del kernel: [46044.432229] [<c12b7235>] trie_firstleaf+0x65/0x70
Sep 5 23:52:33 del kernel: [46044.443941] [<c12b74d4>] fib_table_flush+0x14/0x170
Sep 5 23:52:33 del kernel: [46044.455823] [<c1033e92>] ? local_bh_enable_ip+0x62/0xd0
Sep 5 23:52:33 del kernel: [46044.467995] [<c12fc39f>] ? _raw_spin_unlock_bh+0x2f/0x40
Sep 5 23:52:33 del kernel: [46044.480404] [<c12b24d0>] ? fib_sync_down_dev+0x120/0x180
Sep 5 23:52:33 del kernel: [46044.493025] [<c12b069d>] fib_flush+0x2d/0x60
Sep 5 23:52:33 del kernel: [46044.505796] [<c12b06f5>] fib_disable_ip+0x25/0x50
Sep 5 23:52:33 del kernel: [46044.518772] [<c12b10d3>] fib_netdev_event+0x73/0xd0
Sep 5 23:52:33 del kernel: [46044.531918] [<c1048dfd>] notifier_call_chain+0x2d/0x70
Sep 5 23:52:33 del kernel: [46044.545358] [<c1048f0a>] raw_notifier_call_chain+0x1a/0x20
Sep 5 23:52:33 del kernel: [46044.559092] [<c124f687>] call_netdevice_notifiers+0x27/0x60
Sep 5 23:52:33 del kernel: [46044.573037] [<c124faec>] __dev_notify_flags+0x5c/0x80
Sep 5 23:52:33 del kernel: [46044.586489] [<c124fb47>] dev_change_flags+0x37/0x60
Sep 5 23:52:33 del kernel: [46044.599394] [<c12a8a8d>] devinet_ioctl+0x54d/0x630
Sep 5 23:52:33 del kernel: [46044.612277] [<c12aabb7>] inet_ioctl+0x97/0xc0
Sep 5 23:52:34 del kernel: [46044.625208] [<c123f6af>] sock_ioctl+0x6f/0x270
Sep 5 23:52:34 del kernel: [46044.638046] [<c109d2b0>] ? handle_mm_fault+0x420/0x6c0
Sep 5 23:52:34 del kernel: [46044.650968] [<c123f640>] ? sock_ioctl+0x0/0x270
Sep 5 23:52:34 del kernel: [46044.663865] [<c10c3188>] vfs_ioctl+0x28/0xa0
Sep 5 23:52:34 del kernel: [46044.676556] [<c10c38fa>] do_vfs_ioctl+0x6a/0x5c0
Sep 5 23:52:34 del kernel: [46044.688989] [<c1048676>] ? up_read+0x16/0x30
Sep 5 23:52:34 del kernel: [46044.701411] [<c1021376>] ? do_page_fault+0x1d6/0x3a0
Sep 5 23:52:34 del kernel: [46044.714223] [<c10b6588>] ? fget_light+0xf8/0x2f0
Sep 5 23:52:34 del kernel: [46044.726601] [<c1241f98>] ? sys_socketcall+0x208/0x2c0
Sep 5 23:52:34 del kernel: [46044.739140] [<c10c3eb3>] sys_ioctl+0x63/0x70
Sep 5 23:52:34 del kernel: [46044.751967] [<c12fca3d>] syscall_call+0x7/0xb
Sep 5 23:52:34 del kernel: [46044.764734] [<c12f0000>] ? cookie_v6_check+0x3d0/0x630
-------------->
This patch fixes the warning:
===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
net/ipv4/fib_trie.c:1756 invoked rcu_dereference_check() without protection!
other info that might help us debug this:
rcu_scheduler_active = 1, debug_locks = 0
1 lock held by pppd/1717:
#0: (rtnl_mutex){+.+.+.}, at: [<c125dc1f>] rtnl_lock+0xf/0x20
stack backtrace:
Pid: 1717, comm: pppd Not tainted 2.6.35.4a #3
Call Trace:
[<c12f9aba>] ? printk+0x18/0x1e
[<c1053969>] lockdep_rcu_dereference+0xa9/0xb0
[<c12b7235>] trie_firstleaf+0x65/0x70
[<c12b74d4>] fib_table_flush+0x14/0x170
...
Allow trie_firstleaf() to be called either under rcu_read_lock()
protection or with RTNL held. The same annotation is added to
node_parent_rcu() to prevent a similar warning a bit later.
Followup of commits 634a4b20 and 4eaa0e3c.
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Casts __kernel to __user pointer require __force markup, so add it. Also
sock_get/setsockopt() takes @optval and/or @optlen arguments as user pointers
but were taking kernel pointers, use new variables 'uoptval' and/or 'uoptlen'
to fix it. These remove following warnings from sparse:
net/socket.c:1922:46: warning: cast adds address space to expression (<asn:1>)
net/socket.c:3061:61: warning: incorrect type in argument 4 (different address spaces)
net/socket.c:3061:61: expected char [noderef] <asn:1>*optval
net/socket.c:3061:61: got char *optval
net/socket.c:3061:69: warning: incorrect type in argument 5 (different address spaces)
net/socket.c:3061:69: expected int [noderef] <asn:1>*optlen
net/socket.c:3061:69: got int *optlen
net/socket.c:3063:67: warning: incorrect type in argument 4 (different address spaces)
net/socket.c:3063:67: expected char [noderef] <asn:1>*optval
net/socket.c:3063:67: got char *optval
net/socket.c:3064:45: warning: incorrect type in argument 5 (different address spaces)
net/socket.c:3064:45: expected int [noderef] <asn:1>*optlen
net/socket.c:3064:45: got int *optlen
net/socket.c:3078:61: warning: incorrect type in argument 4 (different address spaces)
net/socket.c:3078:61: expected char [noderef] <asn:1>*optval
net/socket.c:3078:61: got char *optval
net/socket.c:3080:67: warning: incorrect type in argument 4 (different address spaces)
net/socket.c:3080:67: expected char [noderef] <asn:1>*optval
net/socket.c:3080:67: got char *optval
Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
When there is only one rps_cpus, skb_get_rxhash() can be eliminated.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This simple patch copies the current approach for SIOCINQ ioctl() from DCCP
into SCTP so that the userland code working with SCTP can use a similar
interface across different protocols to know how much space to allocate for
a buffer.
Signed-off-by: Diego Elio Pettenò <flameeyes@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
- Do not create expectation when forwarding the PORT
command to avoid blocking the connection. The problem is that
nf_conntrack_ftp.c:help() tries to create the same expectation later in
POST_ROUTING and drops the packet with "dropping packet" message after
failure in nf_ct_expect_related.
- Change ip_vs_update_conntrack to alter the conntrack
for related connections from real server. If we do not alter the reply in
this direction the next packet from client sent to vport 20 comes as NEW
connection. We alter it but may be some collision happens for both
conntracks and the second conntrack gets destroyed immediately. The
connection stucks too.
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The patch: "gro: fix different skb headrooms" in its part:
"2) allocate a minimal skb for head of frag_list" is buggy. The copied
skb has p->data set at the ip header at the moment, and skb_gro_offset
is the length of ip + tcp headers. So, after the change the length of
mac header is skipped. Later skb_set_mac_header() sets it into the
NET_SKB_PAD area (if it's long enough) and ip header is misaligned at
NET_SKB_PAD + NET_IP_ALIGN offset. There is no reason to assume the
original skb was wrongly allocated, so let's copy it as it was.
bugzilla : https://bugzilla.kernel.org/show_bug.cgi?id=16626
fixes commit: 3d3be4333fdf6faa080947b331a6a19bce1a4f57
Reported-by: Plamen Petrov <pvp-lsts@fs.uni-ruse.bg>
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
CC: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Tested-by: Plamen Petrov <pvp-lsts@fs.uni-ruse.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Attempt to make obvious the first-try-sleeping-then-try-deferral logic
by putting that logic into a top-level function that calls helpers.
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
|
|
Pull out some code into helper functions, fix a typo.
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
|
|
The current practice of waiting for cache updates by queueing the
whole request to be retried has (at least) two problems.
1/ With NFSv4, requests can be quite complex and re-trying a whole
request when a later part fails should only be a last-resort, not a
normal practice.
2/ Large requests, and in particular any 'write' request, will not be
queued by the current code and doing so would be undesirable.
In many cases only a very sort wait is needed before the cache gets
valid data.
So, providing the underlying transport permits it by setting
->thread_wait,
arrange to wait briefly for an upcall to be completed (as reflected in
the clearing of CACHE_PENDING).
If the short wait was not long enough and CACHE_PENDING is still set,
fall back on the old approach.
The 'thread_wait' value is set to 5 seconds when there are spare
threads, and 1 second when there are no spare threads.
These values are probably much higher than needed, but will ensure
some forward progress.
Note that as we only request an update for a non-valid item, and as
non-valid items are updated in place it is extremely unlikely that
cache_check will return -ETIMEDOUT. Normally cache_defer_req will
sleep for a short while and then find that the item is_valid.
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
|
|
This protects us from confusion when the wallclock time changes.
We convert to and from wallclock when setting or reading expiry
times.
Also use seconds since boot for last_clost time.
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
|
