| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Thu, 28 Feb 2019 06:59:26 +0000
From: Ido Schimmel <idosch@...lanox.com>
To: "netdev@...r.kernel.org" <netdev@...r.kernel.org>
CC: "davem@...emloft.net" <davem@...emloft.net>,
Jiri Pirko <jiri@...lanox.com>, mlxsw <mlxsw@...lanox.com>,
Ido Schimmel <idosch@...lanox.com>
Subject: [PATCH net-next 13/14] mlxsw: spectrum_acl: Remember where to
continue rehash migration
From: Jiri Pirko <jiri@...lanox.com>
Store pointer to vchunk where the migration was interrupted, as well as
ventry pointer to start from and to stop at (during rollback). This
saved pointers need to be forgotten in case of ventries list or vchunk
list changes, which is done by couple of "changed" helpers.
Signed-off-by: Jiri Pirko <jiri@...lanox.com>
Signed-off-by: Ido Schimmel <idosch@...lanox.com>
---
.../mellanox/mlxsw/spectrum_acl_tcam.c | 91 ++++++++++++++++++-
1 file changed, 86 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.c
index cfd05af3f0f6..dc118ed714b2 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.c
@@ -184,6 +184,19 @@ struct mlxsw_sp_acl_tcam_vgroup {
struct mlxsw_sp_acl_tcam_rehash_ctx {
void *hints_priv;
bool this_is_rollback;
+ struct mlxsw_sp_acl_tcam_vchunk *current_vchunk; /* vchunk being
+ * currently migrated.
+ */
+ struct mlxsw_sp_acl_tcam_ventry *start_ventry; /* ventry to start
+ * migration from in
+ * a vchunk being
+ * currently migrated.
+ */
+ struct mlxsw_sp_acl_tcam_ventry *stop_ventry; /* ventry to stop
+ * migration at
+ * a vchunk being
+ * currently migrated.
+ */
};
struct mlxsw_sp_acl_tcam_vregion {
@@ -755,6 +768,31 @@ static void mlxsw_sp_acl_tcam_vregion_rehash_work(struct work_struct *work)
mlxsw_sp_acl_tcam_vregion_rehash_work_schedule(vregion);
}
+static void
+mlxsw_sp_acl_tcam_rehash_ctx_vchunk_changed(struct mlxsw_sp_acl_tcam_vchunk *vchunk)
+{
+ struct mlxsw_sp_acl_tcam_vregion *vregion = vchunk->vregion;
+
+ /* If a rule was added or deleted from vchunk which is currently
+ * under rehash migration, we have to reset the ventry pointers
+ * to make sure all rules are properly migrated.
+ */
+ if (vregion->rehash.ctx.current_vchunk == vchunk) {
+ vregion->rehash.ctx.start_ventry = NULL;
+ vregion->rehash.ctx.stop_ventry = NULL;
+ }
+}
+
+static void
+mlxsw_sp_acl_tcam_rehash_ctx_vregion_changed(struct mlxsw_sp_acl_tcam_vregion *vregion)
+{
+ /* If a chunk was added or deleted from vregion we have to reset
+ * the current chunk pointer to make sure all chunks
+ * are properly migrated.
+ */
+ vregion->rehash.ctx.current_vchunk = NULL;
+}
+
static struct mlxsw_sp_acl_tcam_vregion *
mlxsw_sp_acl_tcam_vregion_create(struct mlxsw_sp *mlxsw_sp,
struct mlxsw_sp_acl_tcam_vgroup *vgroup,
@@ -989,6 +1027,7 @@ mlxsw_sp_acl_tcam_vchunk_create(struct mlxsw_sp *mlxsw_sp,
goto err_chunk_create;
}
+ mlxsw_sp_acl_tcam_rehash_ctx_vregion_changed(vregion);
list_add_tail(&vchunk->list, &vregion->vchunk_list);
mutex_unlock(&vregion->lock);
@@ -1012,6 +1051,7 @@ mlxsw_sp_acl_tcam_vchunk_destroy(struct mlxsw_sp *mlxsw_sp,
struct mlxsw_sp_acl_tcam_vgroup *vgroup = vchunk->vgroup;
mutex_lock(&vregion->lock);
+ mlxsw_sp_acl_tcam_rehash_ctx_vregion_changed(vregion);
list_del(&vchunk->list);
if (vchunk->chunk2)
mlxsw_sp_acl_tcam_chunk_destroy(mlxsw_sp, vchunk->chunk2);
@@ -1141,6 +1181,7 @@ static int mlxsw_sp_acl_tcam_ventry_add(struct mlxsw_sp *mlxsw_sp,
}
list_add_tail(&ventry->list, &vchunk->ventry_list);
+ mlxsw_sp_acl_tcam_rehash_ctx_vchunk_changed(vchunk);
mutex_unlock(&vregion->lock);
return 0;
@@ -1157,6 +1198,7 @@ static void mlxsw_sp_acl_tcam_ventry_del(struct mlxsw_sp *mlxsw_sp,
struct mlxsw_sp_acl_tcam_vregion *vregion = vchunk->vregion;
mutex_lock(&vregion->lock);
+ mlxsw_sp_acl_tcam_rehash_ctx_vchunk_changed(vchunk);
list_del(&ventry->list);
mlxsw_sp_acl_tcam_entry_destroy(mlxsw_sp, ventry->entry);
mutex_unlock(&vregion->lock);
@@ -1223,15 +1265,20 @@ mlxsw_sp_acl_tcam_vchunk_migrate_start(struct mlxsw_sp *mlxsw_sp,
}
vchunk->chunk2 = vchunk->chunk;
vchunk->chunk = new_chunk;
+ ctx->current_vchunk = vchunk;
+ ctx->start_ventry = NULL;
+ ctx->stop_ventry = NULL;
return 0;
}
static void
mlxsw_sp_acl_tcam_vchunk_migrate_end(struct mlxsw_sp *mlxsw_sp,
- struct mlxsw_sp_acl_tcam_vchunk *vchunk)
+ struct mlxsw_sp_acl_tcam_vchunk *vchunk,
+ struct mlxsw_sp_acl_tcam_rehash_ctx *ctx)
{
mlxsw_sp_acl_tcam_chunk_destroy(mlxsw_sp, vchunk->chunk2);
vchunk->chunk2 = NULL;
+ ctx->current_vchunk = NULL;
}
static int
@@ -1254,7 +1301,22 @@ mlxsw_sp_acl_tcam_vchunk_migrate_one(struct mlxsw_sp *mlxsw_sp,
return 0;
}
- list_for_each_entry(ventry, &vchunk->ventry_list, list) {
+ /* If the migration got interrupted, we have the ventry to start from
+ * stored in context.
+ */
+ if (ctx->start_ventry)
+ ventry = ctx->start_ventry;
+ else
+ ventry = list_first_entry(&vchunk->ventry_list,
+ typeof(*ventry), list);
+
+ list_for_each_entry_from(ventry, &vchunk->ventry_list, list) {
+ /* During rollback, once we reach the ventry that failed
+ * to migrate, we are done.
+ */
+ if (ventry == ctx->stop_ventry)
+ break;
+
err = mlxsw_sp_acl_tcam_ventry_migrate(mlxsw_sp, ventry,
vchunk->chunk, credits);
if (err) {
@@ -1265,16 +1327,25 @@ mlxsw_sp_acl_tcam_vchunk_migrate_one(struct mlxsw_sp *mlxsw_sp,
* in vchunk->chunk.
*/
swap(vchunk->chunk, vchunk->chunk2);
+ /* The rollback has to be done from beginning of the
+ * chunk, that is why we have to null the start_ventry.
+ * However, we know where to stop the rollback,
+ * at the current ventry.
+ */
+ ctx->start_ventry = NULL;
+ ctx->stop_ventry = ventry;
return err;
} else if (*credits < 0) {
/* We are out of credits, the rest of the ventries
- * will be migrated later.
+ * will be migrated later. Save the ventry
+ * which we ended with.
*/
+ ctx->start_ventry = ventry;
return 0;
}
}
- mlxsw_sp_acl_tcam_vchunk_migrate_end(mlxsw_sp, vchunk);
+ mlxsw_sp_acl_tcam_vchunk_migrate_end(mlxsw_sp, vchunk, ctx);
return 0;
}
@@ -1287,7 +1358,16 @@ mlxsw_sp_acl_tcam_vchunk_migrate_all(struct mlxsw_sp *mlxsw_sp,
struct mlxsw_sp_acl_tcam_vchunk *vchunk;
int err;
- list_for_each_entry(vchunk, &vregion->vchunk_list, list) {
+ /* If the migration got interrupted, we have the vchunk
+ * we are working on stored in context.
+ */
+ if (ctx->current_vchunk)
+ vchunk = ctx->current_vchunk;
+ else
+ vchunk = list_first_entry(&vregion->vchunk_list,
+ typeof(*vchunk), list);
+
+ list_for_each_entry_from(vchunk, &vregion->vchunk_list, list) {
err = mlxsw_sp_acl_tcam_vchunk_migrate_one(mlxsw_sp, vchunk,
vregion->region,
ctx, credits);
@@ -1315,6 +1395,7 @@ mlxsw_sp_acl_tcam_vregion_migrate(struct mlxsw_sp *mlxsw_sp,
* to vregion->region.
*/
swap(vregion->region, vregion->region2);
+ ctx->current_vchunk = NULL;
ctx->this_is_rollback = true;
err2 = mlxsw_sp_acl_tcam_vchunk_migrate_all(mlxsw_sp, vregion,
ctx, credits);
--
2.20.1
Powered by blists - more mailing lists