
We missed the two patches during the mlnx-ofa_kernel driver upgradtion. We encouter the following error during load the module mlx_compat when IBT(Indirect Branch Tracking) feature is enabled for CPU and CONFIG_X86_KERNEL_IBT is enable by the kernel config file. "Missing ENDBR: __init_backport+0x0/0xff0 [mlx_compat]" Verification: * An ISO image can be built with this commit applied to a repo project of StarlingX's kernel 6.12.18 branch. * The ISO image can be installed to a R760 server without "Missing ENDBR: __init_backport+0x0/0xff0 [mlx_compat]" in dmesg log. The issue can be repduced in the lab without the fix. Closes-Bug: 2110611 Change-Id: I017ee18055368614d404f10ad5d4d05a6b842ad0 Signed-off-by: Jiping Ma <jiping.ma2@windriver.com>
63 lines
2.2 KiB
Diff
63 lines
2.2 KiB
Diff
From c6a43833d3b9bf97049e373d20a6eb7daff91c36 Mon Sep 17 00:00:00 2001
|
|
From: Steven Webster <steven.webster@windriver.com>
|
|
Date: Thu, 11 Aug 2022 12:32:48 -0400
|
|
Subject: [PATCH] Enable mlx5 onboard udev name
|
|
|
|
This patch prioritizes the udev renaming rule for mlx5 controlled
|
|
devices to take the onboard name over the slot/path name.
|
|
|
|
This is consistent with the 70-persistent-net.rules file written
|
|
by the StarlingX installer.
|
|
|
|
It is also consistent with the naming order in the 99-default.link
|
|
file.
|
|
|
|
Without this patch, there could be an inconsistency with the
|
|
70-persistent-net.rules first re-naming the device to its
|
|
slot/path name, and then being overridden by the Mellanox specific
|
|
82-net-setup-link.rules.
|
|
|
|
Signed-off-by: Steven Webster <steven.webster@windriver.com>
|
|
[jm: Adapted the patch for context change]
|
|
Signed-off-by: Jiping Ma <jiping.ma2@windriver.com>
|
|
---
|
|
ofed_scripts/vf-net-link-name.sh | 13 +++++++++----
|
|
1 file changed, 9 insertions(+), 4 deletions(-)
|
|
|
|
diff --git a/ofed_scripts/vf-net-link-name.sh b/ofed_scripts/vf-net-link-name.sh
|
|
index e22ae76..1134629 100755
|
|
--- a/ofed_scripts/vf-net-link-name.sh
|
|
+++ b/ofed_scripts/vf-net-link-name.sh
|
|
@@ -93,7 +93,9 @@ function test_if_pf() {
|
|
}
|
|
|
|
if [ "$skip" == "0" ]; then
|
|
- if [ -n "$ID_NET_NAME_SLOT" ]; then
|
|
+ if [ -n "$ID_NET_NAME_ONBOARD" ]; then
|
|
+ NAME="${ID_NET_NAME_ONBOARD%%np[[:digit:]]}"
|
|
+ elif [ -n "$ID_NET_NAME_SLOT" ]; then
|
|
NAME="${ID_NET_NAME_SLOT%%np[[:digit:]]}"
|
|
elif [ -n "$ID_NET_NAME_PATH" ]; then
|
|
NAME="${ID_NET_NAME_PATH%%np[[:digit:]]}"
|
|
@@ -226,11 +228,14 @@ for cnt in {1..2}; do
|
|
continue
|
|
fi
|
|
|
|
- parent_path=`get_pci_name $pci ID_NET_NAME_SLOT`
|
|
+ parent_path=`get_pci_name $pci ID_NET_NAME_ONBOARD`
|
|
if [ -z "$parent_path" ]; then
|
|
- parent_path=`get_pci_name $pci ID_NET_NAME_PATH`
|
|
+ parent_path=`get_pci_name $pci ID_NET_NAME_SLOT`
|
|
if [ -z "$parent_path" ]; then
|
|
- continue
|
|
+ parent_path=`get_pci_name $pci ID_NET_NAME_PATH`
|
|
+ if [ -z "$parent_path" ]; then
|
|
+ continue
|
|
+ fi
|
|
fi
|
|
fi
|
|
echo "NAME=${parent_path}_$PORT"
|
|
--
|
|
2.43.0
|
|
|