[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

FC5/FC6t1 Anaconda dd (driver disk) problem / help needed for debugging


First of all, this is not an "end user support request".

I need help debugging possible anaconda bug.. or to find out reason why my
driver disks won't work with fc5/fc6test1. I'm willing to help and debug to
resolve this.

So.. I have hardware with Qlogic qla4xxx iSCSI HBA. These are not supported
by the default kernels, so I need to create driver disk for FC installer.

I have successfully created qla4xxx driver disk (floppy) image for RHEL4 and
CentOS4. These images work well. dd images generated in the same way won't
work with fc5/fc6test.. Looking at the anaconda repo, it seems there has
been quite many changes since rhel4 to fc5/fc6test1.. that's the reason I
believe this might be an anaconda bug.. 

I used ddiskit 0.9.5 to create the images, and then added some missing files
("modules.pcimap" and "pci.ids") from the "official" Qlogic RHEL4 dd images.
FC5/FC6test1 installers will give errors without those two missing files, so
I needed to add them manually to the dd images.

But anyway.. these dd images won't work with the FC installers. Either the
installer won't give any error, or then it says "No devices of the appropriate
type were found on this driver disk". In both cases the driver module
(qla4xxx) is not loaded.

The driver modules are built for the kernels used by the installers. 
I am able to manually insmod the drivers during the installer (after
manually extracting the modules from the dd). So the drivers/modules itself
are fine - there's something else wrong with the dd images I create - or a
bug in anaconda..

It seems that during the first time I try to load the dd, installer won't
give any error.. and it seems to load the tg3 module for the NIC?! (when looking
VC3). When I try to load the dd second time, it will give the following error: 
"No devices of the appropriate type were found on this driver disk".

I have also verified that the PCI ids are correct (with lspci -n).
RHEL4/CentOS4 install just fine to the same box using same kind of selfmade dd.

Discussion about the problem:

Contents of my dd images (at the end of the msg):

Anaconda bug:

All help appreciated. Thanks!

-- Pasi Kärkkäinen
                                .     .
                              /    -    \

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]