Minor conflict in net/core/rtnetlink.c, David Ahern's bug fix in 'net'
overlapped the renaming of a netlink attribute in net-next.
Signed-off-by: David S. Miller <davem@davemloft.net>
The initial session number when a link is created is based on a random
value, taken from struct tipc_net->random. It is then incremented for
each link reset to avoid mixing protocol messages from different link
sessions.
However, when a bearer is reset all its links are deleted, and will
later be re-created using the same random value as the first time.
This means that if the link never went down between creation and
deletion we will still sometimes have two subsequent sessions with
the same session number. In virtual environments with potentially
long transmission times this has turned out to be a real problem.
We now fix this by randomizing the session number each time a link
is created.
With a session number size of 16 bits this gives a risk of session
collision of 1/64k. To reduce this further, we also introduce a sanity
check on the very first STATE message arriving at a link. If this has
an acknowledge value differing from 0, which is logically impossible,
we ignore the message. The final risk for session collision is hence
reduced to 1/4G, which should be sufficient.
Signed-off-by: LUU Duc Canh <canh.d.luu@dektech.com.au>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We see the following scenario:
1) Link endpoint B on node 1 discovers that its peer endpoint is gone.
Since there is a second working link, failover procedure is started.
2) Link endpoint A on node 1 sends a FAILOVER message to peer endpoint
A on node 2. The node item 1->2 goes to state FAILINGOVER.
3) Linke endpoint A/2 receives the failover, and is supposed to take
down its parallell link endpoint B/2, while producing a FAILOVER
message to send back to A/1.
4) However, B/2 has already been deleted, so no FAILOVER message can
created.
5) Node 1->2 remains in state FAILINGOVER forever, refusing to receive
any messages that can bring B/1 up again. We are left with a non-
redundant link between node 1 and 2.
We fix this with letting endpoint A/2 build a dummy FAILOVER message
to send to back to A/1, so that the situation can be resolved.
Signed-off-by: LUU Duc Canh <canh.d.luu@dektech.com.au>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Default socket receive buffer size for a listener socket is 2Mb. For
each arriving empty SYN, the linux kernel allocates a 768 bytes buffer.
This means that a listener socket can serve maximum 2700 simultaneous
empty connection setup requests before it hits a receive buffer
overflow, and much fewer if the SYN is carrying any significant
amount of data.
When this happens the setup request is rejected, and the client
receives an ECONNREFUSED error.
This commit mitigates this problem by letting the client socket try to
retransmit the SYN message multiple times when it sees it rejected with
the code TIPC_ERR_OVERLOAD. Retransmission is done at random intervals
in the range of [100 ms, setup_timeout / 4], as many times as there is
room for within the setup timeout limit.
Signed-off-by: Tung Nguyen <tung.q.nguyen@dektech.com.au>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Messages intended for intitating a connection are currently
indistinguishable from regular datagram messages. The TIPC
protocol specification defines bit 17 in word 0 as a SYN bit
to allow sanity check of such messages in the listening socket,
but this has so far never been implemented.
We do that in this commit.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We refactor the function tipc_sk_filter_connect(), both to make it
more readable and as a preparation for the next commit.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We refactor this function as a preparation for the coming commits in
the same series.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The function tipc_msg_reverse() is reversing the header of a message
while reusing the original buffer. We have seen at several occasions
that this may have unfortunate side effects when the buffer to be
reversed is a clone.
In one of the following commits we will again need to reverse cloned
buffers, so this is the right time to permanently eliminate this
problem. In this commit we let the said function always consume the
original buffer and replace it with a new one when applicable.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In tipc_link_reset() we copy the wakeup queue to input queue using
skb_queue_splice_init(link->wakeupq, link->inputq).
This is performed without holding any locks. The lists might be
simultaneously be accessed by other cpu threads in tipc_sk_rcv(),
something leading to to random missing packets.
Signed-off-by: Parthasarathy Bhuvaragan <parthasarathy.bhuvaragan@ericsson.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If we detect that under lying carrier detects errors and goes down,
we reset the bearer.
Signed-off-by: Parthasarathy Bhuvaragan <parthasarathy.bhuvaragan@ericsson.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In the case of implicit connect message with data > 1K, the flow
control accounting is incorrect. At this state, the socket does not
know the peer nodes capability and falls back to legacy flow control
by return 1, however the receiver of this message will perform the
new block accounting. This leads to a slack and eventually traffic
disturbance.
In this commit, we perform tipc_node_get_capabilities() at implicit
connect and perform accounting based on the peer's capability.
Signed-off-by: Parthasarathy Bhuvaragan <parthasarathy.bhuvaragan@ericsson.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When __tipc_dump_start() fails with running out of memory,
we have no reason to continue, especially we should avoid
calling tipc_dump_done().
Fixes: 8f5c5fcf35 ("tipc: call start and done ops directly in __tipc_nl_compat_dumpit()")
Reported-and-tested-by: syzbot+3f8324abccfbf8c74a9f@syzkaller.appspotmail.com
Cc: Jon Maloy <jon.maloy@ericsson.com>
Cc: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
An SKB is not on a list if skb->next is NULL.
Codify this convention into a helper function and use it
where we are dequeueing an SKB and need to mark it as such.
Signed-off-by: David S. Miller <davem@davemloft.net>
__tipc_nl_compat_dumpit() uses a netlink_callback on stack,
so the only way to align it with other ->dumpit() call path
is calling tipc_dump_start() and tipc_dump_done() directly
inside it. Otherwise ->dumpit() would always get NULL from
cb->args[].
But tipc_dump_start() uses sock_net(cb->skb->sk) to retrieve
net pointer, the cb->skb here doesn't set skb->sk, the net pointer
is saved in msg->net instead, so introduce a helper function
__tipc_dump_start() to pass in msg->net.
Ying pointed out cb->args[0...3] are already used by other
callbacks on this call path, so we can't use cb->args[0] any
more, use cb->args[4] instead.
Fixes: 9a07efa9ae ("tipc: switch to rhashtable iterator")
Reported-and-tested-by: syzbot+e93a2c41f91b8e2c7d9b@syzkaller.appspotmail.com
Cc: Jon Maloy <jon.maloy@ericsson.com>
Cc: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Before we unlock the sock in tipc_release(), we have to
detach sk->sk_socket from sk, otherwise a parallel
tipc_sk_fill_sock_diag() could stil read it after we
free this socket.
Fixes: c30b70deb5 ("tipc: implement socket diagnostics for AF_TIPC")
Reported-and-tested-by: syzbot+48804b87c16588ad491d@syzkaller.appspotmail.com
Cc: Jon Maloy <jon.maloy@ericsson.com>
Cc: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Trivial fix for two spelling mistakes.
Signed-off-by: Zhenbo Gao <zhenbo.gao@windriver.com>
Reviewed-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove the following obsolete parameter comments of tipc_topsrv struct:
@rcvbuf_cache
@tipc_conn_new
@tipc_conn_release
@tipc_conn_recvmsg
@imp
@type
Add the comments for the missing parameters below of tipc_topsrv struct:
@awork
@listener
Remove the unused or duplicated parameter comments of tipc_conn struct:
@outqueue_lock
@rx_action
Signed-off-by: Zhenbo Gao <zhenbo.gao@windriver.com>
Reviewed-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
syzbot reported a use-after-free in tipc_group_fill_sock_diag(),
where tipc_group_fill_sock_diag() still reads tsk->group meanwhile
tipc_group_delete() just deletes it in tipc_release().
tipc_nl_sk_walk() aims to lock this sock when walking each sock
in the hash table to close race conditions with sock changes like
this one, by acquiring tsk->sk.sk_lock.slock spinlock, unfortunately
this doesn't work at all. All non-BH call path should take
lock_sock() instead to make it work.
tipc_nl_sk_walk() brutally iterates with raw rht_for_each_entry_rcu()
where RCU read lock is required, this is the reason why lock_sock()
can't be taken on this path. This could be resolved by switching to
rhashtable iterator API's, where taking a sleepable lock is possible.
Also, the iterator API's are friendly for restartable calls like
diag dump, the last position is remembered behind the scence,
all we need to do here is saving the iterator into cb->args[].
I tested this with parallel tipc diag dump and thousands of tipc
socket creation and release, no crash or memory leak.
Reported-by: syzbot+b9c8f3ab2994b7cd1625@syzkaller.appspotmail.com
Cc: Jon Maloy <jon.maloy@ericsson.com>
Cc: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
rhashtable_walk_exit() must be paired with rhashtable_walk_enter().
Fixes: 40f9f43970 ("tipc: Fix tipc_sk_reinit race conditions")
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In function tipc_dest_push, the 32bit variables 'node' and 'port'
are stored separately in uppper and lower part of 64bit 'value'.
Then this value is assigned to dst->value which is a union like:
union
{
struct {
u32 port;
u32 node;
};
u64 value;
}
This works on little-endian machines like x86 but fails on big-endian
machines.
The fix remove the 'value' stack parameter and even the 'value'
member of the union in tipc_dest, assign the 'node' and 'port' member
directly with the input parameter to avoid the endian issue.
Fixes: a80ae5306a ("tipc: improve destination linked list")
Signed-off-by: Zhenbo Gao <zhenbo.gao@windriver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Haiqing Bai <Haiqing.Bai@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 9faa89d4ed ("tipc: make function tipc_net_finalize() thread
safe") tries to make it thread safe to set node address, so it uses
node_list_lock lock to serialize the whole process of setting node
address in tipc_net_finalize(). But it causes the following interrupt
unsafe locking scenario:
CPU0 CPU1
---- ----
rht_deferred_worker()
rhashtable_rehash_table()
lock(&(&ht->lock)->rlock)
tipc_nl_compat_doit()
tipc_net_finalize()
local_irq_disable();
lock(&(&tn->node_list_lock)->rlock);
tipc_sk_reinit()
rhashtable_walk_enter()
lock(&(&ht->lock)->rlock);
<Interrupt>
tipc_disc_rcv()
tipc_node_check_dest()
tipc_node_create()
lock(&(&tn->node_list_lock)->rlock);
*** DEADLOCK ***
When rhashtable_rehash_table() holds ht->lock on CPU0, it doesn't
disable BH. So if an interrupt happens after the lock, it can create
an inverse lock ordering between ht->lock and tn->node_list_lock. As
a consequence, deadlock might happen.
The reason causing the inverse lock ordering scenario above is because
the initial purpose of node_list_lock is not designed to do the
serialization of node address setting.
As cmpxchg() can guarantee CAS (compare-and-swap) process is atomic,
we use it to replace node_list_lock to ensure setting node address can
be atomically finished. It turns out the potential deadlock can be
avoided as well.
Fixes: 9faa89d4ed ("tipc: make function tipc_net_finalize() thread safe")
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Acked-by: Jon Maloy <maloy@donjonn.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Variables 'tn' and 'oport' are being assigned but are never used hence
they are redundant and can be removed.
Cleans up clang warnings:
warning: variable 'oport' set but not used [-Wunused-but-set-variable]
warning: variable 'tn' set but not used [-Wunused-but-set-variable]
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The wait_address argument is always directly derived from the filp
argument, so remove it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
tipc_bcast_init() is never called in atomic context.
It calls kzalloc() with GFP_ATOMIC, which is not necessary.
GFP_ATOMIC can be replaced with GFP_KERNEL.
This is found by a static analysis tool named DCNS written by myself.
Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tipc_nametbl_init() is never called in atomic context.
It calls kzalloc() with GFP_ATOMIC, which is not necessary.
GFP_ATOMIC can be replaced with GFP_KERNEL.
This is found by a static analysis tool named DCNS written by myself.
Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
when tipc_own_id failed to obtain node identity,dev_put should
be call before return -EINVAL.
Fixes: 682cd3cf94 ("tipc: confgiure and apply UDP bearer MTU on running links")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes the following sparse warnings:
net/tipc/link.c:376:5: warning: symbol 'link_bc_rcv_gap' was not declared. Should it be static?
net/tipc/link.c:823:6: warning: symbol 'link_prepare_wakeup' was not declared. Should it be static?
net/tipc/link.c:959:6: warning: symbol 'tipc_link_advance_backlog' was not declared. Should it be static?
net/tipc/link.c:1009:5: warning: symbol 'tipc_link_retrans' was not declared. Should it be static?
net/tipc/monitor.c:687:5: warning: symbol '__tipc_nl_add_monitor_peer' was not declared. Should it be static?
net/tipc/group.c:230:20: warning: symbol 'tipc_group_find_member' was not declared. Should it be static?
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The commit referred to below introduced an update of the link
capabilities field that is not safe. Given the recently added
feature to remove idle node and link items after 5 minutes, there
is a small risk that the update will happen at the very moment the
targeted link is being removed. To avoid this we have to perform
the update inside the node item's write lock protection.
Fixes: 9012de5089 ("tipc: add sequence number check for link STATE messages")
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After commit eb929a91b2 ("tipc: improve poll() for group member socket"),
it is no longer used.
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tipc_link_is_active is no longer used and can be removed.
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In some virtual environments we observe a significant higher number of
packet reordering and delays than we have been used to traditionally.
This makes it necessary with stricter checks on incoming link protocol
messages' session number, which until now only has been validated for
RESET messages.
Since the other two message types, ACTIVATE and STATE messages also
carry this number, it is easy to extend the validation check to those
messages.
We also introduce a flag indicating if a link has a valid peer session
number or not. This eliminates the mixing of 32- and 16-bit arithmethics
we are currently using to achieve this.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some switch infrastructures produce huge amounts of packet duplicates.
This becomes a problem if those messages are STATE/NACK protocol
messages, causing unnecessary retransmissions of already accepted
packets.
We now introduce a unique sequence number per STATE protocol message
so that duplicates can be identified and ignored. This will also be
useful when tracing such cases, and to avert replay attacks when TIPC
is encrypted.
For compatibility reasons we have to introduce a new capability flag
TIPC_LINK_PROTO_SEQNO to handle this new feature.
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently a link is declared stale and reset if there has been 100
repeated attempts to retransmit the same packet. However, in certain
infrastructures we see that packet (NACK) duplicates and delays may
cause such retransmit attempts to occur at a high rate, so that the
peer doesn't have a reasonable chance to acknowledge the reception
before the 100-limit is hit. This may take much less than the
stipulated link tolerance time, and despite that probe/probe replies
otherwise go through as normal.
We now extend the criteria for link reset to also being time based.
I.e., we don't reset the link until the link tolerance time is passed
AND we have made 100 retransmissions attempts.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The setting of the node address is not thread safe, meaning that
two discoverers may decide to set it simultanously, with a duplicate
entry in the name table as result. We fix that with this commit.
Fixes: 25b0b9c4e8 ("tipc: handle collisions of 32-bit node address hash values")
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The duplicate address discovery protocol is not safe against two
discoverers running in parallel. The one executing first after the
trial period is over will set the node address and change its own
message type to DSC_REQ_MSG. The one executing last may find that the
node address is already set, and never change message type, with the
result that its links may never be established.
In this commmit we ensure that the message type always is set correctly
after the trial period is over.
Fixes: 25b0b9c4e8 ("tipc: handle collisions of 32-bit node address hash values")
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With the duplicate address discovery protocol for tipc nodes addresses
we introduced a one second trial period before a node is allocated a
hash number to use as address.
Unfortunately, we miss to handle the case when a regular LINK REQUEST/
RESPONSE arrives from a cluster node during the trial period. Such
messages are not ignored as they should be, leading to links setup
attempts while the node still has no address.
Fixes: 25b0b9c4e8 ("tipc: handle collisions of 32-bit node address hash values")
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The function for checking if there is an node address conflict is
supposed to return a suggestion for a new address if it finds a
conflict, and zero otherwise. But in case the peer being checked
is previously unknown it does instead return a "suggestion" for
the checked address itself. This results in a DSC_TRIAL_FAIL_MSG
being sent unecessarily to the peer, and sometimes makes the trial
period starting over again.
Fixes: 25b0b9c4e8 ("tipc: handle collisions of 32-bit node address hash values")
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.
Warning level 2 was used: -Wimplicit-fallthrough=2
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Simple overlapping changes in stmmac driver.
Adjust skb_gro_flush_final_remcsum function signature to make GRO list
changes in net-next, as per Stephen Rothwell's example merge
resolution.
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit extends the existing TIPC socket diagnostics framework
for information related to TIPC group communication.
Acked-by: Ying Xue <ying.xue@windriver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: GhantaKrishnamurthy MohanKrishna <mohan.krishna.ghanta.krishnamurthy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A peer node is considered down if there are no
active links (or) lost contact to the node. In current implementation,
a peer node instance is deleted either if
a) TIPC module is removed (or)
b) Application can use a netlink/iproute2 interface to delete a
specific down node.
Thus, a down node instance lives in the system forever, unless the
application explicitly removes it.
We fix this by deleting the nodes which are down for
a specified amount of time (5 minutes).
Existing node supervision timer is used to achieve this.
Acked-by: Ying Xue <ying.xue@windriver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: GhantaKrishnamurthy MohanKrishna <mohan.krishna.ghanta.krishnamurthy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In single-link usage, the function tipc_node_timeout() still iterates
over the whole link array to handle each link. Given that the maximum
number of bearers are 3, there are 2 redundant iterations with lock
grab/release. Since this function is executing very frequently it makes
sense to optimize it.
This commit adds conditional checking to exit from the loop if the
known number of configured links has already been accessed.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Tung Nguyen <tung.q.nguyen@dektech.com.au>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The function tipc_msg_extract() is using skb_clone() to clone inner
messages from a message bundle buffer. Although this method is safe,
it has an undesired effect that each buffer clone inherits the
true-size of the bundling buffer. As a result, the buffer clone
almost always ends up with being copied anyway by the message
validation function. This makes the cloning into a sub-optimization.
In this commit we take the consequence of this realization, and copy
each inner message to a separately allocated buffer up front in the
extraction function.
As a bonus we can now eliminate the two cases where we had to copy
re-routed packets that may potentially go out on the wire again.
Signed-off-by: Tung Nguyen <tung.q.nguyen@dektech.com.au>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The poll() changes were not well thought out, and completely
unexplained. They also caused a huge performance regression, because
"->poll()" was no longer a trivial file operation that just called down
to the underlying file operations, but instead did at least two indirect
calls.
Indirect calls are sadly slow now with the Spectre mitigation, but the
performance problem could at least be largely mitigated by changing the
"->get_poll_head()" operation to just have a per-file-descriptor pointer
to the poll head instead. That gets rid of one of the new indirections.
But that doesn't fix the new complexity that is completely unwarranted
for the regular case. The (undocumented) reason for the poll() changes
was some alleged AIO poll race fixing, but we don't make the common case
slower and more complex for some uncommon special case, so this all
really needs way more explanations and most likely a fundamental
redesign.
[ This revert is a revert of about 30 different commits, not reverted
individually because that would just be unnecessarily messy - Linus ]
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
David S. Miller