From: Dan Carpenter <hidden> Date: 2011-01-05 08:20:39
On Wed, Jan 05, 2011 at 02:01:01AM +0100, Linus Walleij wrote:
2011/1/4 Dan Carpenter [off-list ref]:
quoted
There was a semi-colon missing and it broke the compile.
How come my setup compiles then... Oh well, probably a more
forgiving compiler.
The only explanation I can think of is that you're compiling the driver
into the kernel instead of as a module? A lot of people like to have
their keyboard compiled in instead of loaded later.
You could try a "make drivers/input/keyboard/tc3589x-keypad.i" and look
at the precompiled code.
regards,
dan carpenter
On Wed, Jan 05, 2011 at 02:01:01AM +0100, Linus Walleij wrote:
quoted
2011/1/4 Dan Carpenter [off-list ref]:
quoted
There was a semi-colon missing and it broke the compile.
How come my setup compiles then... Oh well, probably a more
forgiving compiler.
The only explanation I can think of is that you're compiling the driver
into the kernel instead of as a module?
Yes that's it, hm the macro __MODULE_INFO() resolves to
nothing if compiled-in and resolves to something needing
a semicolon otherwise.
This makes some errors pass under the radar, looks
syntactically unsound to me :-(
I'll cook up a patch.
Yours,
Linus Walleij
There was a semi-colon missing and it broke the compile.
Signed-off-by: Dan Carpenter <redacted>
Dmitry can you take this fixup into the input tree?
Russell pulled the tc3589x driver into his tree and Torvalds
pulled it into his tree, so if you are based on Torvalds' tree
for today it should apply fine.
(Sorry for the mistake, I've attempted to fix the root cause
by a patch to make this throw compile errors under all
circumstances to Rusty Russell.)
Yours,
Linus Walleij
On Fri, Jan 07, 2011 at 11:12:38AM +0100, Linus Walleij wrote:
2011/1/4 Dan Carpenter [off-list ref]:
quoted
There was a semi-colon missing and it broke the compile.
Signed-off-by: Dan Carpenter <redacted>
Dmitry can you take this fixup into the input tree?
Russell pulled the tc3589x driver into his tree and Torvalds
pulled it into his tree, so if you are based on Torvalds' tree
for today it should apply fine.
I normally merge with mainline at -rc1 or -rc2 points so I do not have
the driver yet.
I think this is one of those times when forwarding patch directly to
Linus instead of waiting for it to trickle through all the trees is the
most sensible solution.
Thanks.
--
Dmitry
I think this is one of those times when forwarding patch directly to
Linus instead of waiting for it to trickle through all the trees is the
most sensible solution.
OK said and done Linus sends to Linus...
Yours,
Linus Walleij