[PATCH] Fixed bug in Wireless Extension Code (CONFIG_WEXT_CORE).
mi...@mspradling.com
When CONFIG_WEXT_CORE is not enabled and CONFIG_WEXT_PRIV is enabled
there currently is a few build errors. This is because several
functions call function call_commit_handler which is currently only
defined when CONFIG_WEXT_CORE is enabled.
Signed-off-by: Michael Spradling <mi...@mspradling.com>
---
include/net/wext.h | 4 ++++
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/include/net/wext.h b/include/net/wext.h
index 4f6e742..773c8c6 100644
--- a/include/net/wext.h
+++ b/include/net/wext.h
@@ -24,6 +24,10 @@ static inline int compat_wext_handle_ioctl(struct net *net, unsigned int cmd,
{
return -EINVAL;
}
+static inline int call_commit_handler(struct net_device *dev)
+{
+ return -EINVAL;
+}
#endif
#ifdef CONFIG_WEXT_PROC
--
1.6.5.2
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majo...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Johannes Berg
Please send wireless patches to the correct list (linux-wireless) in the
future.
On Sun, 2009-11-22 at 19:36 -0500, mi...@mspradling.com wrote:
> When CONFIG_WEXT_CORE is not enabled and CONFIG_WEXT_PRIV is enabled
> there currently is a few build errors.
This should never happen unless there's as bug in Kconfig, please
provide the config that creates this problem so we can fix that issue
rather than work around it. When PRIV is enabled but not CORE, this
patch may fix compilation but it doesn't fix that the configuration
makes no sense at all.
johannes
Michael Spradling
I have included my config file just in case I may be wrong.
Thank you
I at first assumed there was a bug in Kconfig and looked into this.
However, I found this solution is not a hack. PRIV does not depend
on CORE. Some header files in "PRIV CODE" check and see if CORE is
defined. If it is not defined it appears to impliment dummy functions
that return an error code. All but one of these dummy functions are
implimented. This patch just implements one that was forgoten.
Michael Spradling
Johannes Berg
> I don't think its a Kconfig issue, see notes below. However,
> I have included my config file just in case I may be wrong.
You config file doesn't have _any_ wireless options set. Not even
CONFIG_NET.
johannes
Michael Spradling
If I come across this again I will email you directly.
--
Mike Spradling
--
Randy Dunlap
> Sorry, I lost my config file and haven't been able to reproduce it.
> If I come across this again I will email you directly.
> --
> Mike Spradling
>
> On Mon, Nov 23, 2009 at 04:53:46PM +0100, Johannes Berg wrote:
>> On Mon, 2009-11-23 at 10:42 -0500, Michael Spradling wrote:
>>> I don't think its a Kconfig issue, see notes below. However,
>>> I have included my config file just in case I may be wrong.
>> You config file doesn't have _any_ wireless options set. Not even
>> CONFIG_NET.
>>
>> johannes
All of the cases that I have seen where call_commit_handler() is undefined
have been kconfig errors -- mostly in drivers/staging/.
If you do have this problem again, please be sure to send the problem
.config file.
--
~Randy