DPDK patches and discussions
 help / color / mirror / Atom feed
* [dpdk-dev] [PATCH] hash: fix unlimited cuckoo path
@ 2016-10-12  0:42 Pablo de Lara
  2016-10-12  0:43 ` De Lara Guarch, Pablo
  2016-10-12  0:50 ` [dpdk-dev] [PATCH v2] " Pablo de Lara
  0 siblings, 2 replies; 4+ messages in thread
From: Pablo de Lara @ 2016-10-12  0:42 UTC (permalink / raw)
  To: dev; +Cc: bruce.richardson, Pablo de Lara

When trying to insert a new entry, if its target bucket is full,
the alternative location (bucket) of one of the entries is checked,
to try to find an empty slot, with make_space_bucket.
This function is called every time a new bucket is checked, recursively.
To avoid having a very long insert operation (and to avoid filling up
the stack), a limit in the number of pushes is introduced.

Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
---
 lib/librte_hash/rte_cuckoo_hash.c | 6 +++++-
 lib/librte_hash/rte_cuckoo_hash.h | 2 ++
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/lib/librte_hash/rte_cuckoo_hash.c b/lib/librte_hash/rte_cuckoo_hash.c
index 3324b17..51db006 100644
--- a/lib/librte_hash/rte_cuckoo_hash.c
+++ b/lib/librte_hash/rte_cuckoo_hash.c
@@ -419,6 +419,7 @@ rte_hash_reset(struct rte_hash *h)
 static inline int
 make_space_bucket(const struct rte_hash *h, struct rte_hash_bucket *bkt)
 {
+	static unsigned int nr_pushes;
 	unsigned i, j;
 	int ret;
 	uint32_t next_bucket_idx;
@@ -455,11 +456,13 @@ make_space_bucket(const struct rte_hash *h, struct rte_hash_bucket *bkt)
 			break;
 
 	/* All entries have been pushed, so entry cannot be added */
-	if (i == RTE_HASH_BUCKET_ENTRIES)
+	if (i == RTE_HASH_BUCKET_ENTRIES || nr_pushes > RTE_HASH_MAX_PUSHES)
 		return -ENOSPC;
 
 	/* Set flag to indicate that this entry is going to be pushed */
 	bkt->flag[i] = 1;
+
+	nr_pushes++;
 	/* Need room in alternative bucket to insert the pushed entry */
 	ret = make_space_bucket(h, next_bkt[i]);
 	/*
@@ -469,6 +472,7 @@ make_space_bucket(const struct rte_hash *h, struct rte_hash_bucket *bkt)
 	 * or return error
 	 */
 	bkt->flag[i] = 0;
+	nr_pushes = 0;
 	if (ret >= 0) {
 		next_bkt[i]->sig_alt[ret] = bkt->sig_current[i];
 		next_bkt[i]->sig_current[ret] = bkt->sig_alt[i];
diff --git a/lib/librte_hash/rte_cuckoo_hash.h b/lib/librte_hash/rte_cuckoo_hash.h
index c00aafa..1b8ffed 100644
--- a/lib/librte_hash/rte_cuckoo_hash.h
+++ b/lib/librte_hash/rte_cuckoo_hash.h
@@ -140,6 +140,8 @@ enum add_key_case {
 
 #define LCORE_CACHE_SIZE		64
 
+#define RTE_HASH_MAX_PUSHES             100
+
 #define RTE_HASH_BFS_QUEUE_MAX_LEN       1000
 
 #define RTE_XABORT_CUCKOO_PATH_INVALIDED 0x4
-- 
2.7.4

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [dpdk-dev] [PATCH] hash: fix unlimited cuckoo path
  2016-10-12  0:42 [dpdk-dev] [PATCH] hash: fix unlimited cuckoo path Pablo de Lara
@ 2016-10-12  0:43 ` De Lara Guarch, Pablo
  2016-10-12  0:50 ` [dpdk-dev] [PATCH v2] " Pablo de Lara
  1 sibling, 0 replies; 4+ messages in thread
From: De Lara Guarch, Pablo @ 2016-10-12  0:43 UTC (permalink / raw)
  To: dev; +Cc: Richardson, Bruce



> -----Original Message-----
> From: De Lara Guarch, Pablo
> Sent: Tuesday, October 11, 2016 5:43 PM
> To: dev@dpdk.org
> Cc: Richardson, Bruce; De Lara Guarch, Pablo
> Subject: [PATCH] hash: fix unlimited cuckoo path
> 
> When trying to insert a new entry, if its target bucket is full,
> the alternative location (bucket) of one of the entries is checked,
> to try to find an empty slot, with make_space_bucket.
> This function is called every time a new bucket is checked, recursively.
> To avoid having a very long insert operation (and to avoid filling up
> the stack), a limit in the number of pushes is introduced.
> 
> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>

Nack. Missed Fixes line.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* [dpdk-dev] [PATCH v2] hash: fix unlimited cuckoo path
  2016-10-12  0:42 [dpdk-dev] [PATCH] hash: fix unlimited cuckoo path Pablo de Lara
  2016-10-12  0:43 ` De Lara Guarch, Pablo
@ 2016-10-12  0:50 ` Pablo de Lara
  2016-10-12 15:01   ` Thomas Monjalon
  1 sibling, 1 reply; 4+ messages in thread
From: Pablo de Lara @ 2016-10-12  0:50 UTC (permalink / raw)
  To: dev; +Cc: bruce.richardson, Pablo de Lara

When trying to insert a new entry, if its target bucket is full,
the alternative location (bucket) of one of the entries is checked,
to try to find an empty slot, with make_space_bucket.
This function is called every time a new bucket is checked, recursively.
To avoid having a very long insert operation (and to avoid filling up
the stack), a limit in the number of pushes is introduced.

Fixes: 48a399119619 ("hash: replace with cuckoo hash implementation")

Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
---

Changes in v2:
- Included Fixes line

 lib/librte_hash/rte_cuckoo_hash.c | 6 +++++-
 lib/librte_hash/rte_cuckoo_hash.h | 2 ++
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/lib/librte_hash/rte_cuckoo_hash.c b/lib/librte_hash/rte_cuckoo_hash.c
index 3324b17..51db006 100644
--- a/lib/librte_hash/rte_cuckoo_hash.c
+++ b/lib/librte_hash/rte_cuckoo_hash.c
@@ -419,6 +419,7 @@ rte_hash_reset(struct rte_hash *h)
 static inline int
 make_space_bucket(const struct rte_hash *h, struct rte_hash_bucket *bkt)
 {
+	static unsigned int nr_pushes;
 	unsigned i, j;
 	int ret;
 	uint32_t next_bucket_idx;
@@ -455,11 +456,13 @@ make_space_bucket(const struct rte_hash *h, struct rte_hash_bucket *bkt)
 			break;
 
 	/* All entries have been pushed, so entry cannot be added */
-	if (i == RTE_HASH_BUCKET_ENTRIES)
+	if (i == RTE_HASH_BUCKET_ENTRIES || nr_pushes > RTE_HASH_MAX_PUSHES)
 		return -ENOSPC;
 
 	/* Set flag to indicate that this entry is going to be pushed */
 	bkt->flag[i] = 1;
+
+	nr_pushes++;
 	/* Need room in alternative bucket to insert the pushed entry */
 	ret = make_space_bucket(h, next_bkt[i]);
 	/*
@@ -469,6 +472,7 @@ make_space_bucket(const struct rte_hash *h, struct rte_hash_bucket *bkt)
 	 * or return error
 	 */
 	bkt->flag[i] = 0;
+	nr_pushes = 0;
 	if (ret >= 0) {
 		next_bkt[i]->sig_alt[ret] = bkt->sig_current[i];
 		next_bkt[i]->sig_current[ret] = bkt->sig_alt[i];
diff --git a/lib/librte_hash/rte_cuckoo_hash.h b/lib/librte_hash/rte_cuckoo_hash.h
index c00aafa..1b8ffed 100644
--- a/lib/librte_hash/rte_cuckoo_hash.h
+++ b/lib/librte_hash/rte_cuckoo_hash.h
@@ -140,6 +140,8 @@ enum add_key_case {
 
 #define LCORE_CACHE_SIZE		64
 
+#define RTE_HASH_MAX_PUSHES             100
+
 #define RTE_HASH_BFS_QUEUE_MAX_LEN       1000
 
 #define RTE_XABORT_CUCKOO_PATH_INVALIDED 0x4
-- 
2.7.4

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [dpdk-dev] [PATCH v2] hash: fix unlimited cuckoo path
  2016-10-12  0:50 ` [dpdk-dev] [PATCH v2] " Pablo de Lara
@ 2016-10-12 15:01   ` Thomas Monjalon
  0 siblings, 0 replies; 4+ messages in thread
From: Thomas Monjalon @ 2016-10-12 15:01 UTC (permalink / raw)
  To: Pablo de Lara; +Cc: dev, bruce.richardson

2016-10-12 01:50, Pablo de Lara:
> When trying to insert a new entry, if its target bucket is full,
> the alternative location (bucket) of one of the entries is checked,
> to try to find an empty slot, with make_space_bucket.
> This function is called every time a new bucket is checked, recursively.
> To avoid having a very long insert operation (and to avoid filling up
> the stack), a limit in the number of pushes is introduced.
> 
> Fixes: 48a399119619 ("hash: replace with cuckoo hash implementation")
> 
> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>

Applied, thanks

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2016-10-12 15:01 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-10-12  0:42 [dpdk-dev] [PATCH] hash: fix unlimited cuckoo path Pablo de Lara
2016-10-12  0:43 ` De Lara Guarch, Pablo
2016-10-12  0:50 ` [dpdk-dev] [PATCH v2] " Pablo de Lara
2016-10-12 15:01   ` Thomas Monjalon

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).