# mlx4\_core : Missing UAR, aborting

**URL:** <https://forums.developer.nvidia.com/t/mlx4-core-missing-uar-aborting/207322>\
**Category:** Mellanox OFED\
**Created:** [November 9, 2017, 5:35pm UTC](https://forums.developer.nvidia.com/t/mlx4-core-missing-uar-aborting/207322 "2017-11-09T17:35:22Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![semenmsu](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@semenmsu](https://forums.developer.nvidia.com/u/semenmsu)\
**Post date:** [November 9, 2017, 5:35pm UTC](https://forums.developer.nvidia.com/t/mlx4-core-missing-uar-aborting/207322/1 "2017-11-09T17:35:22Z")

</div>

OS: Ubuntu 16.04.3 LTS

$ lspci | grep Mellanox

01:00.0 Ethernet controller: Mellanox Technologies MT27520 Family [ConnectX-3 Pro]

$ mstflint -d 01:00.0 q

Image type: FS2FW Version: 2.42.5000FW Release Date: 5.9.2017Product Version: 02.42.50.00Rom Info: type=PXE version=3.4.752 devid=4103Device ID: 4103PSID: MT\_2340111023

I have installed all drivers, ran ‘/etc/init.d/mlnx-en.d retstart’ , got :

Unloading NIC driver: [OK]

Loading NIC driver: [OK]

BUT can’t see mellanox interface (ifconfig -a)

when I type now ‘dmesg | grep mlx’ , I get:

[0.991278] mlx\_compat: loading out-of-tree module taints kernel.[0.991286] mlx\_compat: module verification failed: signature and/or required key missing - tainting kernel[0.992456] mlx4\_core: Mellanox ConnectX core driver v4.1-1.0.2 (27 Jun 2017)[0.992479] mlx4\_core: Initializing 0000:01:00.0[0.992621] mlx4\_core 0000:01:00.0: Missing UAR, aborting

What is the problem here?

P.S. I had no problem with this NIC on my computer, when used Ubuntu 14.04 LTS

---

<div class="post-metadata">

**Author:** ![yuriis1](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@yuriis1](https://forums.developer.nvidia.com/u/yuriis1)\
**Post date:** [March 31, 2018, 10:04pm UTC](https://forums.developer.nvidia.com/t/mlx4-core-missing-uar-aborting/207322/2 "2018-03-31T22:04:30Z")

</div>

Is this issue still actual to you?

I had the same issue on ConnectX-3 Pro, caused by large number of VFs I put into the HCA configuration file.

**Symptoms** (dmesg output):

_Mar 31 17:36:17 macpro kernel: pci 0000:06:00.0: [15b3:1007] type 00 class 0x028000_

_Mar 31 17:36:17 macpro kernel: pci 0000:06:00.0: reg 0x10: [mem 0xf7900000-0xf79fffff 64bit]_

_Mar 31 17:36:17 macpro kernel: pci 0000:06:00.0: reg 0x18: [mem 0xf4000000-0xf47fffff 64bit pref]_

_Mar 31 17:36:17 macpro kernel: pci 0000:06:00.0: reg 0x30: [mem 0xf7800000-0xf78fffff pref]_

_Mar 31 17:36:17 macpro kernel: pci 0000:06:00.0: reg 0x134: [mem 0x00000000-0x007fffff 64bit pref]_

_Mar 31 17:36:17 macpro kernel: pci 0000:06:00.0: VF(n) BAR2 space: [mem 0x00000000-0x03ffffff 64bit pref] (contains BAR2 for 8 VFs)_

_Mar 31 17:36:17 macpro kernel: pci 0000:06:00.0: BAR 2: no space for [mem size 0x00800000 64bit pref]_

_Mar 31 17:36:17 macpro kernel: pci 0000:06:00.0: BAR 2: failed to assign [mem size 0x00800000 64bit pref]_

_…_

_Mar 31 21:09:48 macpro kernel: mlx4\_core: Mellanox ConnectX core driver v4.3-1.0.1_

_Mar 31 21:09:48 macpro kernel: mlx4\_core: Initializing 0000:06:00.0_

_Mar 31 21:09:48 macpro kernel: mlx4\_core 0000:06:00.0: enabling device (0100 → 0102)_

_Mar 31 21:09:48 macpro kernel: mlx4\_core 0000:06:00.0: Missing UAR, aborting_

**Solution:**

1. Collect FW-related information about your HCA, you need to know PSID

_# mstflint -d 06:00.0 q full_

_Image type: FS2_

_FW Version: 2.42.8016_

_FW Release Date: 21.3.2018_

_MIC Version: 2.0.0_

_Config Sectors: 2_

_PRS Name: cx3pro\_MCX354A\_fdr\_09v.prs_

_Rom Info: type=PXE version=3.4.752_

_Device ID: 4103_

_…_

_ **PSID: MT\_1090111019** _

1. Download FW for your particular HCA and unpack it.

- In my case it is _ConnectX3Pro-rel-2\_42\_6000-web.tgz_

- the archive contains set of \*.ini files for different HCAs based on the same chip.

- find the configuration file for your PSID:

_# grep **MT\_1090111019** \*.ini_

_ **MCX354A-FCC\_Ax.ini** :PSID = MT\_1090111019_

1. Dump HCA configuration file, compare it to the original configuration file:

_# mstflint -d 06:00.0 dc current.ini_

_# diff -u MCX354A-FCC\_Ax.ini current.ini_

1. Next, add a few parameters into the [HCA] section to disable sr\_iov and reduce the number of VFs to 4 (safe value in my case):

_[HCA]_

_hca\_header\_subsystem\_id = 0x0003_

_hca\_header\_device\_id = 0x1007_

_dpdp\_en = true_

_eth\_xfi\_en = true_

_mdio\_en\_port1 = 0_

_ **num\_pfs = 1** _

_ **total\_vfs = 4** _

_ **sriov\_en = false** _

1. Create the new firmware image, which includes the new configuration file

\*# mlxburn -fw ./fw-ConnectX3Pro-rel.mlx -c current-sriov.ini \*

_-wrimage ./fw-ConnectX3Pro-rel-2\_42\_8016-MCX354A-FCC\_Ax-FlexBoot-3.4.752.bin_

1. Burn it

_mlxfwmanager -d /dev/mst/mt4103\_pciconf0 -i fw-ConnectX3Pro-rel-2\_42\_8016-MCX354A-FCC\_Ax-FlexBoot-3.4.752.bin -u_

1. Reboot and check that _mlx4\_core_ driver loads now for your HCA.

---

<div class="post-metadata">

**Author:** ![samerka](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/samerka/32/367020_2.png) [@samerka](https://forums.developer.nvidia.com/u/samerka)\
**Post date:** [April 1, 2018, 6:23am UTC](https://forums.developer.nvidia.com/t/mlx4-core-missing-uar-aborting/207322/3 "2018-04-01T06:23:53Z")

</div>

Hi,

The issue is related to PCI allocation in Ubuntu 16.04.3 (Ubuntu kernel cannot allocate enough memory to bring up network interfaces during boot)

Please add the following line to grub2 :

- edit /etc/default/grub

- add GRUB\_CMDLINE\_LINUX\_DEFAULT=" pci=realloc=off"

- update-grub

- Reboot

Thanks,

Samer

---

<div class="post-metadata">

**Author:** ![charoldpctech](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@charoldpctech](https://forums.developer.nvidia.com/u/charoldpctech)\
**Post date:** [March 1, 2023, 3:11pm UTC](https://forums.developer.nvidia.com/t/mlx4-core-missing-uar-aborting/207322/4 "2023-03-01T15:11:06Z")

</div>

This didn’t work for me, but enabling SR-IOV in the motherboard BIOS did in case anyone stumbles across this same issue. Under chipset features on an AMD motherboard.
