dpdk/0013-bonding-fix-crash-when-no-slave-device.patch
Nirmoy Das 3b422e301c Accepting request 399089 from home:ndas:branches:network
- Applied all the fixes recommended by upstream for 
  v2.2 stable release(bsc#981996).
	[+0008-app-testpmd-handle-SIGINT-and-SIGTERM.patch,
	+0009-bonding-copy-entire-config-structure-in-mode-4.patch,
	+0010-bonding-fix-active-slaves-with-no-primary.patch,
	+0011-bonding-do-not-ignore-multicast-in-mode-4.patch,
	+0012-bonding-do-not-activate-slave-twice.patch,
	+0013-bonding-fix-crash-when-no-slave-device.patch,
	+0014-bonding-fix-detach-of-bonded-device.patch,
	+0015-bonding-fix-detach-of-slave-devices.patch,
	+0016-eal-linux-support-built-in-kernel-modules.patch,
	+0017-examples-l3fwd-handle-SIGINT-and-SIGTERM.patch,
	+0018-fm10k-fix-VLAN-flag-in-scattered-Rx.patch,
	+0019-i40e-base-fix-driver-load-failure.patch,
	+0020-i40e-base-fix-missing-check-for-stopped-admin-queue.patch,
	+0021-i40e-fix-inverted-check-for-no-refcount.patch,
	+0022-i40e-fix-overflow.patch,
	+0023-i40e-fix-VLAN-filtering.patch,
	+0024-mempool-fix-leak-when-creation-fails.patch,
	+0025-pcap-fix-captured-frame-length.patch,
	+0026-port-fix-crash-for-ethdev-writer-nodrop.patch,
	+0027-port-fix-crash-for-ring-writer-nodrop.patch,
	+0028-tools-fix-unbinding-failure-handling.patch,
	+0029-tools-support-Python-3-in-bind-script.patch,
	+0030-tools-support-binding-to-built-in-kernel-modules.patch,
	+0031-vhost-fix-leak-of-fds-and-mmaps.patch,
	+0032-virtio-fix-crash-in-statistics-functions.patch,
	+0033-virtio-fix-descriptors-pointing-to-the-same-buffer.patch,
	+0034-virtio-fix-restart.patch]

OBS-URL: https://build.opensuse.org/request/show/399089
OBS-URL: https://build.opensuse.org/package/show/network/dpdk?expand=0&rev=3
2016-05-31 11:30:42 +00:00

37 lines
1.4 KiB
Diff

From 2186fff3675d4e774cecc8f918c05063e0367d28 Mon Sep 17 00:00:00 2001
From: Bernard Iremonger <bernard.iremonger@intel.com>
Date: Mon, 7 Mar 2016 11:40:40 +0000
Subject: [PATCH] bonding: fix crash when no slave device
If a bonded device is created when there are no slave devices
there is a loop in bond_ethdev_promiscuous_enable() which results
in a segmentation fault.
The solution is to initialise the current_primary_port to an
invalid port value when the bonded port is created.
Fixes: 2efb58cbab6e ("bond: new link bonding library")
Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
drivers/net/bonding/rte_eth_bond_api.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/bonding/rte_eth_bond_api.c b/drivers/net/bonding/rte_eth_bond_api.c
index c6c0ed8..3fca764 100644
--- a/drivers/net/bonding/rte_eth_bond_api.c
+++ b/drivers/net/bonding/rte_eth_bond_api.c
@@ -231,7 +231,7 @@ rte_eth_bond_create(const char *name, uint8_t mode, uint8_t socket_id)
internals->port_id = eth_dev->data->port_id;
internals->mode = BONDING_MODE_INVALID;
- internals->current_primary_port = 0;
+ internals->current_primary_port = RTE_MAX_ETHPORTS + 1;
internals->balance_xmit_policy = BALANCE_XMIT_POLICY_LAYER2;
internals->xmit_hash = xmit_l2_hash;
internals->user_defined_mac = 0;
--
2.6.2