Bug 641495

Summary: [Broadcom 5.7 bug] bnx2 v2.0.8+ has scatter gather problem on AMD Dinar system
Product: Red Hat Enterprise Linux 5 Reporter: John Feeney <jfeeney>
Component: kernelAssignee: John Feeney <jfeeney>
Status: CLOSED DUPLICATE QA Contact: Network QE <network-qe>
Severity: high Docs Contact:
Priority: high    
Version: 5.6CC: aaswath, adaora.onyia, agospoda, anderson, andrew.patterson, andriusb, benlu, bnagendr, bugproxy, bzeranski, enarvaez, gideonn, hjia, james.weinert, jburke, jjarvis, joerg.roedel, joseph.szczypek, jtorrice, kthiruma, ltroan, mchan, niran, nobody+PNT0273897, noboru.obata.ar, peterm, rpacheco, sandy.garza, sbest, seiji.aguchi.tr, tyasui
Target Milestone: rcKeywords: OtherQA, Reopened
Target Release: 5.7   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: 568601 Environment:
Last Closed: 2011-02-09 16:27:39 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On: 568601    
Bug Blocks: 531800, 537251, 566321, 566322, 566323, 580580, 600363, 617024, 618260, 621244, 655135, 668558    

Comment 2 Andrius Benokraitis 2010-11-19 19:46:37 UTC
Any updates on this? If we don't get a patch for this soon, it will be deferred to 5.7.

Comment 6 Andrius Benokraitis 2010-11-30 16:21:17 UTC
Deferring to 5.7 based on where we are in the 5.6 dev cycle and what we have in hand.

Comment 7 Beth Zeranski 2010-12-08 21:03:08 UTC

*** This bug has been marked as a duplicate of bug 660375 ***

Comment 8 John Feeney 2010-12-08 21:20:43 UTC
Even though this looks like a bnx2 upgrade issue, it is a much more complex problem and so this is not a duplicate of 660375. As this bz states above, on AMD systems with a bnx2 NIC, with scatter gather enabled, and not passthrough disabled, there is a problem transmitting. In my opinion, the problem is not solely in bnx2.

Putting back to ASSIGNED state.

Comment 9 Michael Chan 2010-12-16 00:49:29 UTC
May be related to bug #663509.

Comment 10 John Feeney 2011-01-19 16:07:45 UTC
Yes Michael, I believe you are correct about bz663509. Since bz663509 is closed errata in RHEL5.6, I will try to reproduce it but if anyone has any info on it, I would appreciate an update.

Comment 11 John Feeney 2011-02-09 16:27:39 UTC
I disabled Passthrough on a AMD Dinar system via the amd-iommu=share cmd line option and found the problem did not happen. Given that the original problem manifested when files were transmitted to another system, I copied a bunch of rpms without issue. I copied the files back from the remote system successfully also. Scatter gather was also enabled while testing. See below for details. 

Note: amd-iommu=isolate worked too.

Thus I am setting this a duplicate of bz663509.

[root@amd-dinar-01 stuff]# scp * jfeeney.18.189:stuff
jfeeney.18.189's password: 
kernel-2.6.18-233.el5.bz663509.2.x86_64.rpm   100%   19MB  19.3MB/s   00:01    
kernel-2.6.32-112.el6.src.rpm                 100%   68MB  34.0MB/s   00:02    
kernel-2.6.32-112.el6.x86_64.rpm              100%   23MB  23.0MB/s   00:01    
kernel-debug-2.6.18-233.el5.bz663509.2.x86_64 100%   20MB  19.6MB/s   00:01    
kernel-firmware-2.6.32-112.el6.noarch.rpm     100% 1959KB   1.9MB/s   00:00    
kernel-headers-2.6.18-233.el5.bz663509.2.x86_ 100% 1182KB   1.2MB/s   00:00    
kernel-xen-2.6.18-233.el5.bz663509.2.x86_64.r 100%   20MB  20.2MB/s   00:01    
[root@amd-dinar-01 stuff]# ethtool -i eth0
driver: bnx2
version: 2.0.8-rh
firmware-version: bc 4.6.2
bus-info: 0000:01:00.0
[root@amd-dinar-01 stuff]# ifconfig
eth0      Link encap:Ethernet  HWaddr 00:00:1A:1A:A0:F2  
          inet addr:10.16.42.32  Bcast:10.16.47.255  Mask:255.255.248.0
          inet6 addr: 2620:52:0:102f:200:1aff:fe1a:a0f2/64 Scope:Global
          inet6 addr: fe80::200:1aff:fe1a:a0f2/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:517317 errors:0 dropped:0 overruns:0 frame:0
          TX packets:255284 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:202699359 (193.3 MiB)  TX bytes:340227696 (324.4 MiB)
          Interrupt:58 Memory:ca000000-ca012800 

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:58 errors:0 dropped:0 overruns:0 frame:0
          TX packets:58 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:11643 (11.3 KiB)  TX bytes:11643 (11.3 KiB)

[root@amd-dinar-01 stuff]# lspci -v |grep Eth
01:00.0 Ethernet controller: Broadcom Corporation NetXtreme II BCM5709 Gigabit Ethernet (rev 20)
	Subsystem: Broadcom Corporation NetXtreme II BCM5709 Gigabit Ethernet
01:00.1 Ethernet controller: Broadcom Corporation NetXtreme II BCM5709 Gigabit Ethernet (rev 20)
	Subsystem: Broadcom Corporation NetXtreme II BCM5709 Gigabit Ethernet
[root@amd-dinar-01 stuff]# 
[root@amd-dinar-01 stuff]#  ethtool -k eth0
Offload parameters for eth0:
Cannot get device udp large send offload settings: Operation not supported
rx-checksumming: on
tx-checksumming: on
scatter-gather: on
tcp segmentation offload: on
udp fragmentation offload: off
generic segmentation offload: off
generic-receive-offload: off
[root@amd-dinar-01 stuff]#

*** This bug has been marked as a duplicate of bug 663509 ***