staging: speakup: kobjects.c: Use correct values when changing voice.
authorChristopher Brannon <chris@the-brannons.com>
Wed, 11 Sep 2013 17:05:27 +0000 (10:05 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 17 Sep 2013 14:47:46 +0000 (07:47 -0700)
When a new voice is selected, we set volume and pitch appropriate for
the voice.  We need to use the numeric index corresponding to the
voice when indexing into the volume and pitch tables, rather than
the raw user input that was used to select the voice.
Note that using the raw input can also lead to an invalid memory read
in the case of invalid or malicious user input.

Signed-off-by: Christopher Brannon <chris@the-brannons.com>
Acked-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/staging/speakup/kobjects.c

index 61a3f7a5eb6be148acc1c9ce88c4a7cb7278c277..f31afa2cffb71597df483615730c6a0e92fffdd6 100644 (file)
@@ -651,7 +651,10 @@ ssize_t spk_var_store(struct kobject *kobj, struct kobj_attribute *attr,
                * If voice was just changed, we might need to reset our default
                * pitch and volume.
                */
-               if (param->var_id == VOICE) {
+               if (param->var_id == VOICE && synth &&
+                   (ret == 0 || ret == -ERESTART)) {
+                       var_data = param->data;
+                       value = var_data->u.n.value;
                        spk_reset_default_value("pitch", synth->default_pitch,
                                value);
                        spk_reset_default_value("vol", synth->default_vol,