net/mlx4: workaround verbs error after plug-out

Current mlx4 OFED version has bug which returns error to
ibv destroy functions when the device was plugged out, in
spite of the resources were destroyed correctly.

Hence, failsafe PMD was aborted, only in debug mode, when
it tries to remove the device in plug-out process.

The workaround added option to replace all claim_zero
assertions with debugging messages, by the way, this option
affects non ibv destroy assertions.

DPDK 18.02 release should work with Mellanox OFED-4.2 which will
include the verbs fix to this bug, then, this patch can
be removed.

Signed-off-by: Matan Azrad <matan@mellanox.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>

diff --git a/config/common_base b/config/common_base
index 78056050fec50f347184db612c9efa6efab6c9fc..5e97a08b619745a308fc139715b803910b73a917 100644 (file)
--- a/config/common_base
+++ b/config/common_base
@@ -213,6 +213,7 @@ CONFIG_RTE_LIBRTE_FM10K_INC_VECTOR=y
 #
 CONFIG_RTE_LIBRTE_MLX4_PMD=n
 CONFIG_RTE_LIBRTE_MLX4_DEBUG=n
+CONFIG_RTE_LIBRTE_MLX4_DEBUG_BROKEN_VERBS=n
 CONFIG_RTE_LIBRTE_MLX4_SGE_WR_N=4
 CONFIG_RTE_LIBRTE_MLX4_MAX_INLINE=0
 CONFIG_RTE_LIBRTE_MLX4_TX_MP_CACHE=8

diff --git a/doc/guides/nics/mlx4.rst b/doc/guides/nics/mlx4.rst
index d5bf2b3b2cf85f89483a8aa2469c946831c6d08f..f8885b23fc54c01f401177e1dd7185726e9ea0f2 100644 (file)
--- a/doc/guides/nics/mlx4.rst
+++ b/doc/guides/nics/mlx4.rst
@@ -119,6 +119,14 @@ These options can be modified in the ``.config`` file.
   adds additional run-time checks and debugging messages at the cost of
   lower performance.
 
+- ``CONFIG_RTE_LIBRTE_MLX4_DEBUG_BROKEN_VERBS`` (default **n**)
+
+  Mellanox OFED versions earlier than 4.2 may return false errors from
+  Verbs object destruction APIs after the device is plugged out.
+  Enabling this option replaces assertion checks that cause the program
+  to abort with harmless debugging messages as a workaround.
+  Relevant only when CONFIG_RTE_LIBRTE_MLX4_DEBUG is enabled.
+
 - ``CONFIG_RTE_LIBRTE_MLX4_SGE_WR_N`` (default **4**)
 
   Number of scatter/gather elements (SGEs) per work request (WR). Lowering

diff --git a/drivers/net/mlx4/Makefile b/drivers/net/mlx4/Makefile
index 755c8a4f3a4c038ea5f860935bf5a82609fc8fa0..c045bd79b604840b8af065943d2f8cd85314806c 100644 (file)
--- a/drivers/net/mlx4/Makefile
+++ b/drivers/net/mlx4/Makefile
@@ -84,6 +84,10 @@ ifdef CONFIG_RTE_LIBRTE_MLX4_SOFT_COUNTERS
 CFLAGS += -DMLX4_PMD_SOFT_COUNTERS=$(CONFIG_RTE_LIBRTE_MLX4_SOFT_COUNTERS)
 endif
 
+ifeq ($(CONFIG_RTE_LIBRTE_MLX4_DEBUG_BROKEN_VERBS),y)
+CFLAGS += -DMLX4_PMD_DEBUG_BROKEN_VERBS
+endif
+
 include $(RTE_SDK)/mk/rte.lib.mk
 
 # Generate and clean-up mlx4_autoconf.h.

diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index a2e0ae7befa4cdf7c04669101ce8678e30a72508..c0ade4f1ae3596d0ec647e223802c2b099b9c140 100644 (file)
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
@@ -182,7 +182,13 @@ enum {
                (DEBUG__(__VA_ARGS__), 0)       \
        })[0])
 #define DEBUG(...) DEBUG_(__VA_ARGS__, '\n')
+#ifndef MLX4_PMD_DEBUG_BROKEN_VERBS
 #define claim_zero(...) assert((__VA_ARGS__) == 0)
+#else /* MLX4_PMD_DEBUG_BROKEN_VERBS */
+#define claim_zero(...) \
+       (void)(((__VA_ARGS__) == 0) || \
+               DEBUG("Assertion `(" # __VA_ARGS__ ") == 0' failed (IGNORED)."))
+#endif /* MLX4_PMD_DEBUG_BROKEN_VERBS */
 #define claim_nonzero(...) assert((__VA_ARGS__) != 0)
 #define claim_positive(...) assert((__VA_ARGS__) >= 0)
 #else /* NDEBUG */