Commit 7cc2e18f authored by Jason Gunthorpe's avatar Jason Gunthorpe
Browse files

RDMA/odp: Use the common interval tree library instead of generic 

ODP is working with userspace VA's in the interval tree which always fit
into an unsigned long, so we can use the common code.

This comes at a cost of a 16 byte increase in ib_umem_odp struct size due
to storing the interval tree start/last in addition to the umem
addr/length. However these values were computed and are performance
critical for the interval lookup, so this seems like a worthwhile trade
off.

Removes 2k of .text from the kernel.

Link: https://lore.kernel.org/r/20190819111710.18440-2-leon@kernel.org


Signed-off-by: default avatarLeon Romanovsky <leonro@mellanox.com>
Signed-off-by: default avatarJason Gunthorpe <jgg@mellanox.com>
parent 27b7fb1a

drivers/infiniband/Kconfig

+1 −0
Original line number Diff line number Diff line
@@ -55,6 +55,7 @@ config INFINIBAND_ON_DEMAND_PAGING 
	bool "InfiniBand on-demand paging support"

	depends on INFINIBAND_USER_MEM

	select MMU_NOTIFIER

	select INTERVAL_TREE

	default y

	---help---

	  On demand paging support for the InfiniBand subsystem.

drivers/infiniband/core/umem_odp.c

+18 −54
Original line number Diff line number Diff line
@@ -39,45 +39,13 @@ 
#include <linux/export.h>

#include <linux/vmalloc.h>

#include <linux/hugetlb.h>

#include <linux/interval_tree_generic.h>

#include <linux/interval_tree.h>

#include <linux/pagemap.h>



#include <rdma/ib_verbs.h>

#include <rdma/ib_umem.h>

#include <rdma/ib_umem_odp.h>



/*

 * The ib_umem list keeps track of memory regions for which the HW

 * device request to receive notification when the related memory

 * mapping is changed.

 *

 * ib_umem_lock protects the list.

 */



static u64 node_start(struct umem_odp_node *n)

{

	struct ib_umem_odp *umem_odp =

			container_of(n, struct ib_umem_odp, interval_tree);



	return ib_umem_start(umem_odp);

}



/* Note that the representation of the intervals in the interval tree

 * considers the ending point as contained in the interval, while the

 * function ib_umem_end returns the first address which is not contained

 * in the umem.

 */

static u64 node_last(struct umem_odp_node *n)

{

	struct ib_umem_odp *umem_odp =

			container_of(n, struct ib_umem_odp, interval_tree);



	return ib_umem_end(umem_odp) - 1;

}



INTERVAL_TREE_DEFINE(struct umem_odp_node, rb, u64, __subtree_last,

		     node_start, node_last, static, rbt_ib_umem)



static void ib_umem_notifier_start_account(struct ib_umem_odp *umem_odp)

{

	mutex_lock(&umem_odp->umem_mutex);
@@ -205,9 +173,18 @@ static void add_umem_to_per_mm(struct ib_umem_odp *umem_odp) 
	struct ib_ucontext_per_mm *per_mm = umem_odp->per_mm;



	down_write(&per_mm->umem_rwsem);

	if (likely(ib_umem_start(umem_odp) != ib_umem_end(umem_odp)))

		rbt_ib_umem_insert(&umem_odp->interval_tree,

	if (likely(ib_umem_start(umem_odp) != ib_umem_end(umem_odp))) {

		/*

		 * Note that the representation of the intervals in the

		 * interval tree considers the ending point as contained in

		 * the interval, while the function ib_umem_end returns the

		 * first address which is not contained in the umem.

		 */

		umem_odp->interval_tree.start = ib_umem_start(umem_odp);

		umem_odp->interval_tree.last = ib_umem_end(umem_odp) - 1;

		interval_tree_insert(&umem_odp->interval_tree,

				     &per_mm->umem_tree);

	}

	up_write(&per_mm->umem_rwsem);

}
@@ -217,7 +194,7 @@ static void remove_umem_from_per_mm(struct ib_umem_odp *umem_odp) 


	down_write(&per_mm->umem_rwsem);

	if (likely(ib_umem_start(umem_odp) != ib_umem_end(umem_odp)))

		rbt_ib_umem_remove(&umem_odp->interval_tree,

		interval_tree_remove(&umem_odp->interval_tree,

				     &per_mm->umem_tree);

	complete_all(&umem_odp->notifier_completion);
@@ -761,18 +738,18 @@ int rbt_ib_umem_for_each_in_range(struct rb_root_cached *root, 
				  void *cookie)

{

	int ret_val = 0;

	struct umem_odp_node *node, *next;

	struct interval_tree_node *node, *next;

	struct ib_umem_odp *umem;



	if (unlikely(start == last))

		return ret_val;



	for (node = rbt_ib_umem_iter_first(root, start, last - 1);

	for (node = interval_tree_iter_first(root, start, last - 1);

			node; node = next) {

		/* TODO move the blockable decision up to the callback */

		if (!blockable)

			return -EAGAIN;

		next = rbt_ib_umem_iter_next(node, start, last - 1);

		next = interval_tree_iter_next(node, start, last - 1);

		umem = container_of(node, struct ib_umem_odp, interval_tree);

		ret_val = cb(umem, start, last, cookie) || ret_val;

	}
@@ -780,16 +757,3 @@ int rbt_ib_umem_for_each_in_range(struct rb_root_cached *root, 
	return ret_val;

}

EXPORT_SYMBOL(rbt_ib_umem_for_each_in_range);
 


struct ib_umem_odp *rbt_ib_umem_lookup(struct rb_root_cached *root,

				       u64 addr, u64 length)

{

	struct umem_odp_node *node;



	node = rbt_ib_umem_iter_first(root, addr, addr + length - 1);

	if (node)

		return container_of(node, struct ib_umem_odp, interval_tree);

	return NULL;



}

EXPORT_SYMBOL(rbt_ib_umem_lookup);

include/rdma/ib_umem_odp.h

+12 −8
Original line number Diff line number Diff line
@@ -37,11 +37,6 @@ 
#include <rdma/ib_verbs.h>

#include <linux/interval_tree.h>



struct umem_odp_node {

	u64 __subtree_last;

	struct rb_node rb;

};



struct ib_umem_odp {

	struct ib_umem umem;

	struct ib_ucontext_per_mm *per_mm;
@@ -72,7 +67,7 @@ struct ib_umem_odp { 
	int npages;



	/* Tree tracking */

	struct umem_odp_node	interval_tree;

	struct interval_tree_node interval_tree;



	struct completion	notifier_completion;

	int			dying;
@@ -163,8 +158,17 @@ int rbt_ib_umem_for_each_in_range(struct rb_root_cached *root, 
 * Find first region intersecting with address range.

 * Return NULL if not found

 */

struct ib_umem_odp *rbt_ib_umem_lookup(struct rb_root_cached *root,

				       u64 addr, u64 length);

static inline struct ib_umem_odp *

rbt_ib_umem_lookup(struct rb_root_cached *root, u64 addr, u64 length)

{

	struct interval_tree_node *node;



	node = interval_tree_iter_first(root, addr, addr + length - 1);

	if (!node)

		return NULL;

	return container_of(node, struct ib_umem_odp, interval_tree);



}



static inline int ib_umem_mmu_notifier_retry(struct ib_umem_odp *umem_odp,

					     unsigned long mmu_seq)

CRA Git | Maintained and supported by SUSTech CRA and CCSE