aboutsummaryrefslogtreecommitdiffstats
path: root/debian/patches/ubuntu-backport-39-lpm-fix-freeing-in-compatibility-mode.patch
diff options
context:
space:
mode:
authorChristian Ehrhardt <christian.ehrhardt@canonical.com>2016-06-15 20:29:06 +0200
committerChristian Ehrhardt <christian.ehrhardt@canonical.com>2016-06-15 20:37:12 +0200
commit199e427d5dea3051eec1726eb9c857aa3d5fdde2 (patch)
tree5477b3bd4868676a210bbd3ef67a469d34814e61 /debian/patches/ubuntu-backport-39-lpm-fix-freeing-in-compatibility-mode.patch
parentb5cdd645c9fc62341d55aebbfc93a1b648415512 (diff)
Merge Ubuntu DPDK packaging as of 15th June 2016
As discussed this shall be our initial baseline. If history is needed for any sort of debugging or analysis it can be found at https://code.launchpad.net/~ubuntu-server/dpdk/+git/dpdk Change-Id: Ie95c7effbbea34d723df14f6451c1f782000cbc1 Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
Diffstat (limited to 'debian/patches/ubuntu-backport-39-lpm-fix-freeing-in-compatibility-mode.patch')
-rw-r--r--debian/patches/ubuntu-backport-39-lpm-fix-freeing-in-compatibility-mode.patch52
1 files changed, 52 insertions, 0 deletions
diff --git a/debian/patches/ubuntu-backport-39-lpm-fix-freeing-in-compatibility-mode.patch b/debian/patches/ubuntu-backport-39-lpm-fix-freeing-in-compatibility-mode.patch
new file mode 100644
index 00000000..4224cad8
--- /dev/null
+++ b/debian/patches/ubuntu-backport-39-lpm-fix-freeing-in-compatibility-mode.patch
@@ -0,0 +1,52 @@
+Description: backport of dpdk 16.07 fix 7cc3f2c2
+
+Author: Christian Ehrhardt <christian.ehrhardt@canonical.com>
+Date: Tue Apr 12 15:49:27 2016 +0200
+
+ lpm: fix freeing in compatibility mode
+
+ Back then when we fixed the missing free lpm I was to quickly to say yes
+ if it applies not only to the lpm6 but also to all of the lpm code.
+
+ It turned out to not apply to all of them. In rte_lpm_create_v20 there
+ is an unexpected fused allocation:
+ mem_size = sizeof(*lpm) + (sizeof(lpm->rules_tbl[0]) * max_rules);
+ [...]
+ lpm = (struct rte_lpm_v20 *)rte_zmalloc_socket(mem_name,mem_size,
+ RTE_CACHE_LINE_SIZE, socket_id);
+
+ That causes lpm->rules_tbl not to have an own struct malloc_elem that
+ can be derived via RTE_PTR_SUB(data, MALLOC_ELEM_HEADER_LEN) in
+ malloc_elem_from_data.
+ Due to that the rte_lpm_free_v20 accidentially misderives the elem and
+ assumes it is ELEM_FREE triggering in malloc_elem_free
+ if (!malloc_elem_cookies_ok(elem) || elem->state !=
+ return -1;
+
+ While it seems counter-intuitive the way to properly remove rules_tbl in
+ the old fused allocation style of rte_lpm_free_v20 is to not remove it.
+
+ The newer rte_lpm_free_v1604 is safe because in rte_lpm_create_v1604
+ rules_tbl is a separate allocation.
+
+ Fixes: d4c18f0a1d5d ("lpm: fix missing free")
+
+ Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
+ Acked-by: Olivier Matz <olivier.matz@6wind.com>
+
+Forwarded: yes (in DPDK 16.07)
+Author: Christian Ehrhardt <christian.ehrhardt@canonical.com>
+Last-Update: 2016-05-17
+
+Index: dpdk/lib/librte_lpm/rte_lpm.c
+===================================================================
+--- dpdk.orig/lib/librte_lpm/rte_lpm.c
++++ dpdk/lib/librte_lpm/rte_lpm.c
+@@ -373,7 +373,6 @@ rte_lpm_free_v20(struct rte_lpm_v20 *lpm
+
+ rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);
+
+- rte_free(lpm->rules_tbl);
+ rte_free(lpm);
+ rte_free(te);
+ }