#8737 NORM Not Tri: strange SSID confuses OFW

[Zarro Boogs per Child]

#8737: strange SSID confuses OFW
-----------------------------------+----------------------------------------
   Reporter:  dsd                  |       Owner:  wmb at firmworks.com
       Type:  defect               |      Status:  new              
   Priority:  normal               |   Milestone:  Not Triaged      
  Component:  ofw - open firmware  |     Version:  not specified    
 Resolution:                       |    Keywords:                   
Next_action:  never set            |    Verified:  0                
  Blockedby:                       |    Blocking:                   
-----------------------------------+----------------------------------------

Comment(by wmb at firmworks.com):

 Replying to [comment:2 dsd]:
 > This is a malformed probe response.
 >
 > The final good tagged parameter is 00 03 a4 00 00
 > 00 = SSID tag
 > 03 = length 3
 > a4 00 00 = the SSID (!?)

 My best guess is that this tag is not actually good, but rather is junk
 data that happens to be interpretable as a valid tag.  This scan result
 already has a reasonable SSID tag, so it's unlikely that there will be two
 such tags.


 >
 > Next tag is code 0x27, OFW doesn't seem to recognise but wireshark says
 this is a measurement report.

 I suspect that we are looking at two 32-bit little-endian numbers:
 0x00a40300 followed by 0x00a42700.


 > The length is claimed to be 0xa4 (164) but then there are less than 164
 bytes following --> this is where OFW gets confused.
 >
 > Mitch says he's hardened this up a bit in the source. Which is good,
 because my wireshark session showed some even more corrupted probe
 responses at other times.

-- 
Ticket URL: <http://dev.laptop.org/ticket/8737#comment:3>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system