Bug 8320 - harddrake2 does not show quad-port ethernet card
Summary: harddrake2 does not show quad-port ethernet card
Status: RESOLVED FIXED
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: 2
Hardware: x86_64 Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Thierry Vignaud
QA Contact:
URL:
Whiteboard:
Keywords: NEEDINFO
: 4905 (view as bug list)
Depends on: 8373
Blocks:
  Show dependency treegraph
 
Reported: 2012-12-07 08:28 CET by Markus Mertens
Modified: 2013-01-15 19:29 CET (History)
3 users (show)

See Also:
Source RPM: drakxtools
CVE:
Status comment:


Attachments
lspcidrake -v (16.26 KB, text/plain)
2012-12-07 15:04 CET, Markus Mertens
Details
lspci -vvvvnn (160.57 KB, text/plain)
2012-12-10 08:25 CET, Markus Mertens
Details

Description Markus Mertens 2012-12-07 08:28:43 CET
Description of problem:

A quad-port ethernet card (Intel PRO/1000 PT Quad Port LP Server Adapter) is not being displayed in "mcc->browse and configure hardware" although all ports can be seen with lspci and they can be configured via "mcc->network center". The same quad-port card could be seen in another workstation where it was plugged into a PCIe Gen2 slot. Now it is plugged into a PCIe Gen3 slot on a Sandybridge mainboard.

# lspci | grep Copper
86:00.0 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (Copper) (rev 06)
86:00.1 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (Copper) (rev 06)
87:00.0 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (Copper) (rev 06)
87:00.1 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (Copper) (rev 06)
# harddrake2
SIOCETHTOOL: Operation not supported
SIOCETHTOOL: Operation not supported
#



Version-Release number of selected component (if applicable):


How reproducible:

It is a permanent bug.



Steps to Reproduce:

1. install the quad-port card in a PCI Gen3 slot

2. boot the system

3. call mcc

4. none of the ports is being displayed in "mcc->browse and configure hardware".

5. configuring all ports with "mcc->network center" works
Comment 1 Thierry Vignaud 2012-12-07 12:56:24 CET
Please _attach_ (not paste) output of "lspcidrake -v"
eg: run "lspcidrake -v >pci.txt" and attach the resulting 'pci.txt' file

Keywords: (none) => NEEDINFO
CC: (none) => thierry.vignaud
Source RPM: (none) => drakxtools

Thierry Vignaud 2012-12-07 12:56:32 CET

Assignee: bugsquad => thierry.vignaud

Comment 2 Markus Mertens 2012-12-07 15:04:21 CET
Created attachment 3208 [details]
lspcidrake -v
Comment 3 Thierry Vignaud 2012-12-07 16:20:02 CET
Does this really came from the same machine?
What's the output of "lspci -vvvvnn"? (to be attached too)
Comment 4 Markus Mertens 2012-12-10 08:25:23 CET
Created attachment 3221 [details]
lspci -vvvvnn
Comment 5 Markus Mertens 2012-12-10 08:26:41 CET
Yes, it is the same maschine (Fujitsu CELSIUS R920).
Comment 6 Thierry Vignaud 2012-12-11 14:50:34 CET
There's definitively sg fishy here as both uses libpci.so...

The following commands show we lspcidrake is missing 37 devices on your machines!
perl -pi -e 'undef $_ if /^hub|^usbhid/..//'<sav/pci.txt>pci.txt2
egrep '^[a-z0-9]' lspci.txt |egrep -o '\[[0-9a-f]*:[0-9a-f]*\]'|sort>lst2
perl -pi -e 's!.*vendor:([0-9a-f]+) device:([0-9a-f]+).*![\1:\2]!'<pci.txt|sort>lst1
diff -u lst?|vi -
wc -l lst?
  163 lst1
  171 lst2

100 looks very suspicious, isn't it?
and MAX_DEVICES is set to 100 in ldetect...
Ooops...
Fixed in git
I wonder how we never got more reports over the years...

Status: NEW => RESOLVED
Resolution: (none) => FIXED

Comment 7 Thierry Vignaud 2012-12-11 18:48:44 CET
*** Bug 4905 has been marked as a duplicate of this bug. ***

CC: (none) => AndrewL733

Comment 8 Markus Mertens 2012-12-12 09:15:37 CET
(In reply to comment #6)
> There's definitively sg fishy here as both uses libpci.so...
> 

At least I know that /dev/sg5 is a TapeLibrary which is connected via a LSI SAS HBA. If I manually include the kernel module "st" in the initrd I can use it. Mageia does not detect it automatically and it also does not add it with dracut to the initrd once you change/update/upgrade the kernel via mcc.

Just to clarify: inspecting the hardware with mcc works, but the required module for this hardware is not loaded/included when you modify the bootloader with mcc.


> The following commands show we lspcidrake is missing 37 devices on your
> machines!
> perl -pi -e 'undef $_ if /^hub|^usbhid/..//'<sav/pci.txt>pci.txt2
> egrep '^[a-z0-9]' lspci.txt |egrep -o '\[[0-9a-f]*:[0-9a-f]*\]'|sort>lst2
> perl -pi -e 's!.*vendor:([0-9a-f]+)
> device:([0-9a-f]+).*![\1:\2]!'<pci.txt|sort>lst1
> diff -u lst?|vi -
> wc -l lst?
>   163 lst1
>   171 lst2
> 
> 100 looks very suspicious, isn't it?
> and MAX_DEVICES is set to 100 in ldetect...
> Ooops...
> Fixed in git
> I wonder how we never got more reports over the years...

The BIOS of my former workstation could not handle too many PCIe cards before its memory got exhausted (could not initialize extended ROM). But nowadays more and more BIOSes can handle more additional hardware. I guess lspcidrake has just to catch up with this development.
Comment 9 user7 2013-01-15 00:29:45 CET
Thierry: Please correct me if I'm mistaken.

Markus: There seems to be an update candidate linked at bug 8373 addressing this issue. Could you kindly test it and report back (on bug 8373) if it fixes the problem for you?

CC: (none) => wassi

Comment 10 Markus Mertens 2013-01-15 07:31:00 CET
Several bugs seem to be related to this one:

https://bugs.mageia.org/show_bug.cgi?id=8369
https://bugs.mageia.org/show_bug.cgi?id=8373

But there is no solution, yet.
claire robinson 2013-01-15 10:21:40 CET

Blocks: (none) => 8373

claire robinson 2013-01-15 11:13:45 CET

Blocks: 8373 => (none)
Depends on: (none) => 8373

Comment 11 Thierry Vignaud 2013-01-15 19:29:14 CET
Of course there's, as shown by yourself once the update was applied.

Note You need to log in before you can comment on or make changes to this bug.