Severe problem during download of Firmware makes PCI card useless

[Ingo Hoffmann]

Jouni,

it took some time to build everything but now I have finished your 
proposed procedure and everything seems to work again! (For now I have 
just performed a short test, I'll test more tomorrow!)

So thank you very much for this excellent support! (I really was afraid 
that I couldn't use the card anymore, so I really appreciate your 
immediate help!)

Regards Ingo

PS: The compilation of the driver works also with the latest snapshot, 
I'm talking about the thread: Problem with compiling 
hostap-driver-0.1.1.tar.gz
Thanks again!

Jouni Malinen schrieb:

>On Sun, Nov 02, 2003 at 04:37:16PM -0800, Jouni Malinen wrote:
>
>  
>
>>I will take a look at what would be needed to change the driver to allow
>>Genesis-mode operations with corrupted PRI f/w (i.e., continue minimal
>>operation mode even if initialization fails).
>>    
>>
>
>I added some preliminary code for this. A new module parameter,
>no_primary, can be used to allow initialization continue even if the
>card initialization times out. This might not be enough for all cases,
>but at least it was enough for this particular case.
>
>  
>
>>I'll try to reproduce this issue by flashing a PCI Prism2.5 card with
>>the images you used.. In the mean while, you could also consider
>>upgrading the flash with the DOS-based tool. Depending on what get
>>corrupted in flash, it might be enough to just write the PRI and STA
>>images again. Then again, it is also possible that PDA was corrupted
>>and it would also need to be fixed.
>>    
>>
>
>I was able to reproduce the issue by downloading the same images. I had
>a bit newer versions in the card before doing this, so the results may
>in theory be different, but I would assume we have the cards pretty much
>in identical state now.
>
>It looks like the PRI image was not completely destroyed, but something
>is broken.. Trying to write anything to flash at this point fails.
>However, I was able to download another PRI image to RAM using Genesis
>mode. Unfortunately, this did not fix flash download, but at least I
>could download RAM downloadable STA image afterwards. Thus, the card is
>actually working again.
>
>So, the bad news is that I don't yet know how to fix the flash contents.
>The good news is that the card can be made to work with RAM downloadable
>images. In addition, since I now have one of my cards in the same state,
>I have some extra motivation to figure out a way to fix the flash.. ;-)
>
>You can use following steps to get the card into working state using RAM
>download (this is of course assuming your card ended up in the same
>state than mine):
>
>- update to the latest CVS snapshot
>- load hostap_pci with no_primary set: insmod hostap_pci.o no_primary=1
>- download volatile primary firmware using Genesis mode; optionally,
>  make it persist hardware reset (but not driver unload):
>  prism2_srec -gp wlan0 ak010101.hex
>- download volatile STA firmware (again, with persist mode):
>  prism2_srec -rp wlan0 rf010704.hex 
>
>Card should now be in fully operational state (well, apart from flash
>download). Please let me know, if this does not work with your card
>(include dmesg and prism2_srec output).
>
>  
>