Skip to content

K
Kernel

openeuler / Kernel
1 年多 前同步成功

通知 8
Star 0
Fork 0
K
Kernel
切换分支/标签
  1. 09 4月, 2019 32 次提交
  3. 08 4月, 2019 6 次提交
    • V
      net: sched: flower: insert filter to ht before offloading it to hw · 1f17f774
      由 Vlad Buslov 提交于 
      John reports:

Recent refactoring of fl_change aims to use the classifier spinlock to
avoid the need for rtnl lock. In doing so, the fl_hw_replace_filer()
function was moved to before the lock is taken. This can create problems
for drivers if duplicate filters are created (commmon in ovs tc offload
due to filters being triggered by user-space matches).

Drivers registered for such filters will now receive multiple copies of
the same rule, each with a different cookie value. This means that the
drivers would need to do a full match field lookup to determine
duplicates, repeating work that will happen in flower __fl_lookup().
Currently, drivers do not expect to receive duplicate filters.

To fix this, verify that filter with same key is not present in flower
classifier hash table and insert the new filter to the flower hash table
before offloading it to hardware. Implement helper function
fl_ht_insert_unique() to atomically verify/insert a filter.

This change makes filter visible to fast path at the beginning of
fl_change() function, which means it can no longer be freed directly in
case of error. Refactor fl_change() error handling code to deallocate the
filter with rcu timeout.

Fixes: 620da486 ("net: sched: flower: refactor fl_change")
Reported-by: NJohn Hurley <john.hurley@netronome.com>
Signed-off-by: NVlad Buslov <vladbu@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      1f17f774
    • D
      Merge branch 'rhashtable-bitlocks' · 9186c90b
      由 David S. Miller 提交于 
      NeilBrown says:

====================
Convert rhashtable to use bitlocks

This series converts rhashtable to use a per-bucket bitlock
rather than a separate array of spinlocks.
This:
  reduces memory usage
  results in slightly fewer memory accesses
  slightly improves parallelism
  makes a configuration option unnecessary

The main change from previous version is to use a distinct type for
the pointer in the bucket which has a bit-lock in it.  This
helped find two places where rht_ptr() was missed, one
in  rhashtable_free_and_destroy() in print_ht in the test code.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      9186c90b
    • N
      rhashtable: add lockdep tracking to bucket bit-spin-locks. · 149212f0
      由 NeilBrown 提交于 
      Native bit_spin_locks are not tracked by lockdep.

The bit_spin_locks used for rhashtable buckets are local
to the rhashtable implementation, so there is little opportunity
for the sort of misuse that lockdep might detect.
However locks are held while a hash function or compare
function is called, and if one of these took a lock,
a misbehaviour is possible.

As it is quite easy to add lockdep support this unlikely
possibility seems to be enough justification.

So create a lockdep class for bucket bit_spin_lock and attach
through a lockdep_map in each bucket_table.

Without the 'nested' annotation in rhashtable_rehash_one(), lockdep
correctly reports a possible problem as this lock is taken
while another bucket lock (in another table) is held.  This
confirms that the added support works.
With the correct nested annotation in place, lockdep reports
no problems.
Signed-off-by: NNeilBrown <neilb@suse.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      149212f0
    • N
      rhashtable: use bit_spin_locks to protect hash bucket. · 8f0db018
      由 NeilBrown 提交于 
      This patch changes rhashtables to use a bit_spin_lock on BIT(1) of the
bucket pointer to lock the hash chain for that bucket.

The benefits of a bit spin_lock are:
 - no need to allocate a separate array of locks.
 - no need to have a configuration option to guide the
   choice of the size of this array
 - locking cost is often a single test-and-set in a cache line
   that will have to be loaded anyway.  When inserting at, or removing
   from, the head of the chain, the unlock is free - writing the new
   address in the bucket head implicitly clears the lock bit.
   For __rhashtable_insert_fast() we ensure this always happens
   when adding a new key.
 - even when lockings costs 2 updates (lock and unlock), they are
   in a cacheline that needs to be read anyway.

The cost of using a bit spin_lock is a little bit of code complexity,
which I think is quite manageable.

Bit spin_locks are sometimes inappropriate because they are not fair -
if multiple CPUs repeatedly contend of the same lock, one CPU can
easily be starved.  This is not a credible situation with rhashtable.
Multiple CPUs may want to repeatedly add or remove objects, but they
will typically do so at different buckets, so they will attempt to
acquire different locks.

As we have more bit-locks than we previously had spinlocks (by at
least a factor of two) we can expect slightly less contention to
go with the slightly better cache behavior and reduced memory
consumption.

To enhance type checking, a new struct is introduced to represent the
  pointer plus lock-bit
that is stored in the bucket-table.  This is "struct rhash_lock_head"
and is empty.  A pointer to this needs to be cast to either an
unsigned lock, or a "struct rhash_head *" to be useful.
Variables of this type are most often called "bkt".

Previously "pprev" would sometimes point to a bucket, and sometimes a
->next pointer in an rhash_head.  As these are now different types,
pprev is NULL when it would have pointed to the bucket. In that case,
'blk' is used, together with correct locking protocol.
Signed-off-by: NNeilBrown <neilb@suse.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      8f0db018
    • N
      rhashtable: allow rht_bucket_var to return NULL. · ff302db9
      由 NeilBrown 提交于 
      Rather than returning a pointer to a static nulls, rht_bucket_var()
now returns NULL if the bucket doesn't exist.
This will make the next patch, which stores a bitlock in the
bucket pointer, somewhat cleaner.

This change involves introducing __rht_bucket_nested() which is
like rht_bucket_nested(), but doesn't provide the static nulls,
and changing rht_bucket_nested() to call this and possible
provide a static nulls - as is still needed for the non-var case.
Signed-off-by: NNeilBrown <neilb@suse.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      ff302db9
    • N
      rhashtable: use cmpxchg() in nested_table_alloc() · 7a41c294
      由 NeilBrown 提交于 
      nested_table_alloc() relies on the fact that there is
at most one spinlock allocated for every slot in the top
level nested table, so it is not possible for two threads
to try to allocate the same table at the same time.

This assumption is a little fragile (it is not explicit) and is
unnecessary as cmpxchg() can be used instead.

A future patch will replace the spinlocks by per-bucket bitlocks,
and then we won't be able to protect the slot pointer with a spinlock.

So replace rcu_assign_pointer() with cmpxchg() - which has equivalent
barrier properties.
If it the cmp fails, free the table that was just allocated.
Signed-off-by: NNeilBrown <neilb@suse.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      7a41c294
  5. 07 4月, 2019 2 次提交