[TAG] Error Compiling Atmel Wireless Driver

[Lih-Chern Chiu]

Hi Thomas, Ben, and Peter,

Thanks for all the speedy and helpful response! :)

Peter wrote:
>It seems that if you edit the following
>two files to remove a couple of what are called
>"casts" in the C language, the driver compiles OK:
>
>atmelwlandriver/src/usb/rx.c
>atmelwlandriver/src/usb/vnetusba.c

That worked for me as well. I edited line 749 from rx.c:

--- original/rx.c 2005-09-20 00:01:46.000000000 -0700
+++ rx.c    2005-09-20 00:02:27.000000000 -0700
@@ -746,7 +746,7 @@
                      Adapter->RxInBuff.WirelessPacket +
                      MGMT_FRAME_BODY_OFFSET;

-                 (UCHAR) Adapter->ErrorCode =
+                 Adapter->ErrorCode =
                      (UCHAR) *
                      (PUSHORT) &
                      AssociationRspnsFrame

And line 846 from vnetusba.c:

--- original/vnetusba.c 2005-09-20 00:02:08.000000000 -0700
+++ vnetusba.c    2005-09-20 00:02:38.000000000 -0700
@@ -843,7 +843,7 @@
   u64* pTime = (u64*) tCurrentTime;
   struct timeval timev;
   do_gettimeofday(&timev);
-  (u64) (*pTime) = timev.tv_sec * 10000000;
+  (*pTime) = timev.tv_sec * 10000000;
 };

 int atmel_assoc_info_event (PVNet_ADAPTER Adapter)

Peter wrote:
>Still, on 2.6.13 I do get a warning about
>wireless_send_event being undefined, hopefully
>you won't see that :)

With my 2.6.11 kernel, I didn't get any warnings.

Thomas wrote:
> Posting us the offending line, might prove useful.
[snip]
>Either way, letting us know the version of GCC
> you're using, would help.

Ben wrote:
> In fact, given the way that GCC reports errors, it would be useful if
> you could copy-and-paste that entire function or code block into your
> next email.

Oops, I will make sure to do that next time! However, for the scope of 
this problem, is the code from above sufficient enough? As for GCC, I am 
using version 4.0.

Peter wrote:
>It could be that you have somehow got a corrupted download,
>or are having a filesystem corruption.
>
>Especially, the error message on the line just before seems
>to indicate something like that:
>
> ??~MgmtFrameProcessing??T:

My apologies with the copy & paste. The original output is actually 
'MgmtFrameProcessing', but somehow the ' character didn't convert 
properly. Just to make sure, I also ran md5sum and verified that my 
download is clean.

Ben wrote:
>Installing the latest GCC
>broke my ability to compile a kernel; I had to blow away everything 
>back
>to 3.3 and reinstall slowly and carefully (and only up to the
>penultimate version of GCC) in order to get it working again.

I didn't have any problems with kernel compiling, but I will try 
compiling the driver again with GCC 3.3 and see if that makes any 
difference.

Peter wrote:
>It compiles OK here on Debian Sarge (stable) with kernel 2.6.8-2,
>so at least that indocates that the code is probably correct.

What version of GCC are you using?

Sincerely,
Lih-Chern Chiu