2005-10-31 05:38:55

by Paul Mackerras

Subject: Fix drivers/macintosh/adbhid.c stupid breakage

Commit c7f7a569d9b4ea7c53ab6fcd1377895312d8372b ("[PATCH] Input:
convert drivers/macintosh to dynamic input_dev allocation") breaks any
machine with an ADB keyboard or mouse, which includes my G4
powerbook. Was it given any testing at all?

The problem is that adbhid[]->input is NULL, so the kernel oopses with
a null pointer dereference as soon as I press a key. The following
patch fixes it.

Signed-off-by: Paul Mackerras <[email protected]>
---
diff -urN powerpc-merge/drivers/macintosh/adbhid.c merge-hack/drivers/macintosh/adbhid.c
--- powerpc-merge/drivers/macintosh/adbhid.c 2005-10-31 13:15:26.000000000 +1100
+++ merge-hack/drivers/macintosh/adbhid.c 2005-10-31 16:30:31.000000000 +1100
@@ -723,6 +723,7 @@

sprintf(hid->phys, "adb%d:%d.%02x/input", id, default_id, original_handler_id);

+ hid->input = input_dev;
hid->id = default_id;
hid->original_handler_id = original_handler_id;
hid->current_handler_id = current_handler_id;

2005-10-31 05:46:43

by Dmitry Torokhov

Subject: Re: Fix drivers/macintosh/adbhid.c stupid breakage

On Monday 31 October 2005 00:38, Paul Mackerras wrote:
> Commit c7f7a569d9b4ea7c53ab6fcd1377895312d8372b ("[PATCH] Input:
> convert drivers/macintosh to dynamic input_dev allocation") breaks any
> machine with an ADB keyboard or mouse, which includes my G4
> powerbook. Was it given any testing at all?
>

No it wasn't because I only have x86 to test with. HOwever it was in -mm
tree for a while...

Sorry about the breakage.

--
Dmitry