Re: bcm43xx finally working, with encryption too!!!
Milan Toth wrote:
no luck :(
twoo diferent kernels 2.6.15-rc5 and 2.6.15.1 also your source
of bcm43xx and sofmac and no luck. When we can expect official
module in kernel?
thanks for your help and time Ruben :)
Hi Milan
i got curious after reading the posts by Ruben so
I experimented a bit. I built the kernel 2.6.15 without the
options regarding the ieee82011 modules,
(here is the diff between my two configurations)
< # Linux kernel version: 2.6.15
< # Tue Jan 10 01:36:09 2006
---
> # Linux kernel version: 2.6.15nowifi
> # Mon Jan 16 23:37:17 2006
610,611c610,611
< CONFIG_IEEE80211_CRYPT_CCMP=m
< CONFIG_IEEE80211_CRYPT_TKIP=m
---
> # CONFIG_IEEE80211_CRYPT_CCMP is not set
> # CONFIG_IEEE80211_CRYPT_TKIP is not set
1083,1093c1083,1086
< # CONFIG_IPW_DEBUG is not set
< CONFIG_IPW2200=m
< CONFIG_AIRO=m
< CONFIG_HERMES=m
< # CONFIG_APPLE_AIRPORT is not set
< CONFIG_PLX_HERMES=m
< CONFIG_TMD_HERMES=m
< # CONFIG_NORTEL_HERMES is not set
< CONFIG_PCI_HERMES=m
< C< CONFIG_PCI_ATMEL=m
---
> # CONFIG_IPW2200 is not set
> # CONFIG_AIRO is not set
> # CONFIG_HERMES is not set
> # CONFIG_ATMEL is not set
1098,1099d1090
< # CONFIG_PCMCIA_HERMES is not set
< # CONFIG_PCMCIA_SPECTRUM is not set
1101d1091
< # CONFIG_PCMCIA_ATMEL is not set
1108,1112c1098
< CONFIG_HOSTAP=m
< CONFIG_HOSTAP_FIRMWARE=y
< CONFIG_HOSTAP_PLX=m
< CONFIG_HOSTAP_PCI=m
< CONFIG_HOSTAP_CS=m
---
> # CONFIG_HOSTAP is not set
I emphasise that in both cases i use KERNEL 2.6.15 !!!!!!!!
(so I can't say about -rc7 or 2.6.15.1!!!)
Then I tried to built bcm43xx first against the kernel's own ieee80211 stuff
and second against the softmac-snapshot. The result was that in the
first case
the module just doesn't build which a little search shows that is due
not only that
kernel-source files for ieee* are not identical with those in
softmac-snapshot!
For instance:
diff ieee80211_crypt.h
/usr/src/softmac-snapshot/include/net/ieee80211_crypt.h
26c26,28
< #include <linux/skbuff.h>
---
> #include <linux/types.h>
> #include <linux/list.h>
> #include <asm/atomic.h>
31a34,36
> struct sk_buff;
> struct module;
>
89a95,96
> struct ieee80211_device;
and
# diff ieee80211.h /usr/src/softmac-snapshot/include/net/ieee80211.h
366,367c366,368
< * information for frames received. Not setting these will not
cause
< * any adverse affects. */
---
> * information for frames received.
> * For ieee80211_rx_mgt, you need to set at least the 'len'
parameter.
> */
1090a1092
> /* make sure to set stats->len */
1122a1125,1132
> extern int ieee80211_wx_set_auth(struct net_device *dev,
> struct iw_request_info *info,
> union iwreq_data *wrqu,
> char *extra);
> extern int ieee80211_wx_get_auth(struct net_device *dev,
> struct iw_request_info *info,
> union iwreq_data *wrqu,
> char *extra);
but of also because ieee80211softmac.h simply does not exist at all in
kernel-source.
which must be also the explanation of getting various error messages
(check with dmesg)
So it is not without reason that the README file in bcm43xx states
explicitly that you need to compile against
SOFTMAC source and not the kernel-source made respective modules!
I tried also another combination building the kernel both WITHOUT and
WITH the ieee80211 modules
and then built the bcm43xx against softmac-snapshot on top!!!. Then I
noticed that
in the first case (kernel WITHOUT ieee80211 related modules) compiling
the bcm43xx module (I repeat against softmac-snapshot!) failed (perhaps
due to some incompatibilities of the rest of the kernel stuff???)
while in the second case (kernel built WITH ieee80211 related modules)
the module did compile eventually
because the softmac stuff essentially is "overwritting" kernel's own
original ieee80211 stuff!!!!
Which means that in that case the ieee80211 modules are just the
softmac-snapshot's stuff and not the
original kernel's !!!
As a conclusion, then, I think that if someone sticks with kernel 2.6.15
and the softmac-snapshot source,
there would be no problems to encounter anymore!
(PS. sorry for the length of the reply, but I felt that I had to share
all that!)
best regards
Emmanuel Galatoulas
Reply to: