Menu
▾
▴
Re: [E1000-devel] i350 - NVM Checksum Is Not Valid
|
From: Jeremiah B. <jb...@ev...> - 2015-03-04 23:55:53
|
A follow up to the last fix :
The 0xa0 value isn't random, it is the beginning byte of the Mac
address of the card.
If you have questions about what value to place in the 0x0 position,
use the Value returned from doing the following :
sudo ethtool -e eth1 offset 0x0 length 1
Offset Values
------ ------
0x0000 a0
Then use that value for position 0x0
ethtool -E eth1 magic 0x15218086 offset 0x0 value 0xa0
~Jeremy
On Wed, Mar 4, 2015 at 3:34 PM, Jeremiah Brock <jb...@ev...> wrote:
> Ok so for any other lonely traveler coming across this issue - here is
> what I did to enable the nic to work out of the box even with older
> igb Kernel drivers :
>
> sudo su
> ethtool -E eth1 magic 0x15218086 offset 0x0 value 0xa0
>
> Somethings to note
>
> 1. the magic number is the reverse of the two groups of digits
> returned in the following :
>
> sudo lshw -numeric -C network | grep product | grep 8086
>
> product: I350 Gigabit Network Connection [8086:1521]
> product: I350 Gigabit Network Connection [8086:1521]
> product: I350 Gigabit Network Connection [8086:1521]
> product: I350 Gigabit Network Connection [8086:1521]
>
> 2. This only need be applied to one of the interfaces on the card
> which will cause all of them to re-calculate.
>
>
> Since I originally modified the igb_main.c to be able to initialize
> the nic, I went back and uncommented the block :
>
> cd igb/igb-5.2.15/src
> vi igb_main.c
>
> /* make sure the NVM is good */
> if (e1000_validate_nvm_checksum(hw) < 0) {
> dev_err(pci_dev_to_dev(pdev), "The NVM Checksum Is Not"
> " Valid\n");
> err = -EIO;
> goto err_eeprom;
> }
>
>
> Then re-installed the driver
>
> sudo su
> modprobe -r igb
> make clean
> make
> make install
> modprobe igb
> echo igb >> /etc/modules
> update-initramfs -u
>
> Confirmed everything is well
> dmesg | grep igb
> ifconfig -a
>
>
> I rebooted the system onto the Security Onion 1204 live cd and
> confirmed that the 4 nics showed up as well (when they had not
> originally).
>
> Voila!
>
> Thank you Todd and John for your help.
>
> On Wed, Mar 4, 2015 at 1:52 PM, nick <xer...@gm...> wrote:
>>
>>
>> On 2015-03-04 04:34 PM, Jeremiah Brock wrote:
>>> Hello John,
>>>
>>> Attached is the output of sudo ethtool -e eth1
>>>
>>> ~J
>> Jermiah,
>> There is no attachment for the ethtool output in your email. Unless it's inline at the bottom of this conversation.
>> Would you mind resending with output of ethtool on your card if not send already.
>> Thanks,
>> Nick
>>>
>>> On Wed, Mar 4, 2015 at 1:23 PM, Ronciak, John <joh...@in...> wrote:
>>>>> With that commented everything works splendid
>>>> Well not really as none of the HW configuration setting have been loaded out of the EEPROM. The NIC may still work but you could run into problems like the PHY not being configured correctly which can cause link problem as certain cable lengths for example.
>>>>
>>>>> but I would like to update
>>>>> the eeprom as that will allow the card to effectively work out of the box
>>>>> IMHO.
>>>> So you need to figure out if the EEPROM isn't configured correctly or if you are having trouble reading the EEPROM on that system. So with the checksum code commented out as you have it now see what happens when you dump the contents of the EEPROM using ethtool. It should give us a clue as to what the issue is. Your NIC may not have been programmed and you'll need to return it for one that has been programmed correctly.
>>>>
>>>> Cheers,
>>>> John
>>>>
>>>>> -----Original Message-----
>>>>> From: Jeremiah Brock [mailto:jb...@ev...]
>>>>> Sent: Wednesday, March 4, 2015 11:35 AM
>>>>> To: Ronciak, John
>>>>> Cc: e10...@li...
>>>>> Subject: Re: [E1000-devel] i350 - NVM Checksum Is Not Valid
>>>>>
>>>>> Hello John,
>>>>>
>>>>> Purchased the Nic new off of Amazon.
>>>>>
>>>>> I have run the Live instance of Security Union and the most Recently
>>>>> Updated as of last night.
>>>>>
>>>>> I am running the most recent IGB driver (5.2.15) as shown in modinfo igb .
>>>>>
>>>>> The only fix I have been able to conjure up so far is a hack to the igb_main.c
>>>>> commenting out the following and re-compiling :
>>>>>
>>>>> # Comment out the following entire piece of code in src/igb_main.c
>>>>> /* make sure the NVM is good */
>>>>> /*if (e1000_validate_nvm_checksum(hw) < 0) {
>>>>> dev_err(pci_dev_to_dev(pdev), "The NVM Checksum Is Not"
>>>>> " Valid\n");
>>>>> err = -EIO;
>>>>> goto err_eeprom;
>>>>> }
>>>>> */
>>>>>
>>>>> With that commented everything works splendid but I would like to update
>>>>> the eeprom as that will allow the card to effectively work out of the box
>>>>> IMHO.
>>>>>
>>>>> ~J
>>>>>
>>>>>
>>>>> On Wed, Mar 4, 2015 at 10:54 AM, Ronciak, John <joh...@in...>
>>>>> wrote:
>>>>>> Where did the NIC come from? Is it a new NIC? Is this only happening with
>>>>> this Security Onion 1204 (probably very old driver)? Can you boot a live
>>>>> image of Ubuntu 14.10 or even 14.04 to see if that works?
>>>>>>
>>>>>> Cheers,
>>>>>> John
>>>>>>
>>>>>>> -----Original Message-----
>>>>>>> From: Jeremiah Brock [mailto:jb...@ev...]
>>>>>>> Sent: Wednesday, March 4, 2015 9:47 AM
>>>>>>> To: e10...@li...
>>>>>>> Subject: [E1000-devel] i350 - NVM Checksum Is Not Valid
>>>>>>>
>>>>>>> Hello Everyone,
>>>>>>>
>>>>>>> I have an interesting issue with an i350-t4 v01 card running
>>>>>>> under Security Onion 1204.
>>>>>>>
>>>>>>> The PCI bus recognized the card and 4 interfaces but IGB will not
>>>>>>> initialize them with DMESG showing :
>>>>>>>
>>>>>>> [ 1.970552] igb 0000:02:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
>>>>>>> [ 1.970571] igb 0000:02:00.0: setting latency timer to 64
>>>>>>> [ 1.970909] igb 0000:02:00.0: irq 46 for MSI/MSI-X
>>>>>>> [ 1.970916] igb 0000:02:00.0: irq 47 for MSI/MSI-X
>>>>>>> [ 2.001797] igb 0000:02:00.0: The NVM Checksum Is Not Valid
>>>>>>> [ 2.021179] igb 0000:02:00.0: PCI INT A disabled
>>>>>>> [ 2.021185] igb: probe of 0000:02:00.0 failed with error -5
>>>>>>>
>>>>>>>
>>>>>>> A quick google search brought me to you guys and a thread a few
>>>>>>> months back :
>>>>>>> http://sourceforge.net/p/e1000/mailman/message/33090729/
>>>>>>>
>>>>>>> Todd mentions in the last entry that " The easiest way to fix
>>>>>>> this is to write one byte to the EEPROM with a newer driver and let
>>>>>>> it recalculate the checksums."
>>>>>>>
>>>>>>> # modinfo igb
>>>>>>> filename: /lib/modules/3.2.0-77-generic/kernel/drivers/net/igb/igb.ko
>>>>>>> version: 5.2.15
>>>>>>>
>>>>>>>
>>>>>>> Can someone please help me with how this is accomplished?
>>>>>>>
>>>>>>> Thank you!
>>>>>>>
>>>>>>> ~J
>>>>>>>
>>>>>>> ---------------------------------------------------------------------
>>>>>>> --------- Dive into the World of Parallel Programming The Go Parallel
>>>>>>> Website, sponsored by Intel and developed in partnership with
>>>>>>> Slashdot Media, is your hub for all things parallel software
>>>>>>> development, from weekly thought leadership blogs to news, videos,
>>>>>>> case studies, tutorials and more. Take a look and join the
>>>>>>> conversation now. http://goparallel.sourceforge.net/
>>>>>>> _______________________________________________
>>>>>>> E1000-devel mailing list
>>>>>>> E10...@li...
>>>>>>> https://lists.sourceforge.net/lists/listinfo/e1000-devel
>>>>>>> To learn more about Intel® Ethernet, visit
>>>>>>> http://communities.intel.com/community/wired
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
>>>>>>> by Intel and developed in partnership with Slashdot Media, is your hub for all
>>>>>>> things parallel software development, from weekly thought leadership blogs to
>>>>>>> news, videos, case studies, tutorials and more. Take a look and join the
>>>>>>> conversation now. http://goparallel.sourceforge.net/
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> E1000-devel mailing list
>>>>>>> E10...@li...
>>>>>>> https://lists.sourceforge.net/lists/listinfo/e1000-devel
>>>>>>> To learn more about Intel® Ethernet, visit http://communities.intel.com/community/wired
>
>
>
> --
|
