Skip to contentSkip to conversation

toshiba_acpi.c - fix busted keyboard backlight logic

PATCH In review Valdis Kletnieks 2 participants 2 messages Aug 20, 2014 – Aug 21, 2014 x86 platform drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 toshiba_acpi.c - fix busted keyboard backlight logic

Both the original and the first attempted patch managed to get the logic wrong. So we fix it so we only continue if sscanf returns exactly one value, and it has to be either 1 or 2. Apparently, programmers are getting out and hacking code without remembering de Morgan's Laws...

Patch is against the file that results from Ari's revert of the previous patch.

--

platform/x86/toshiba_acpi.c renamed from platform/x86/toshiba_acpi.c.orig +1 -1
@@ -1258,7 +1258,7 @@ static ssize_t toshiba_kbd_bl_mode_store
 	int mode = -1;
 	int time = -1;
 
-	if (sscanf(buf, "%i", &mode) != 1 && (mode != 2 || mode != 1))
+	if (sscanf(buf, "%i", &mode) != 1 || !(mode == 2 || mode == 1))
 		return -EINVAL;
 
 	/* Set the Keyboard Backlight Mode where:
#1 lore raw
Quoted 14 lines · Valdis Kletnieks

Both the original and the first attempted patch managed to get the logic wrong. So we fix it so we only continue if sscanf returns exactly one value, and it has to be either 1 or 2. Apparently, programmers are getting out and hacking code without remembering de Morgan's Laws...

Patch is against the file that results from Ari's revert of the previous patch.

--

platform/x86/toshiba_acpi.c static ssize_t toshiba_kbd_bl_mode_store · line 1264

Quoted 8 lines
platform/x86/toshiba_acpi.c renamed from platform/x86/toshiba_acpi.c.orig +1 -1
@@ -1258,7 +1258,7 @@ static ssize_t toshiba_kbd_bl_mode_store
 	int mode = -1;
 	int time = -1;
 
-	if (sscanf(buf, "%i", &mode) != 1 && (mode != 2 || mode != 1))
+	if (sscanf(buf, "%i", &mode) != 1 || !(mode == 2 || mode == 1))
 		return -EINVAL;
 
 	/* Set the Keyboard Backlight Mode where:

Looks good to me.

It seems Matthew already reverted the bad commit in mainline commit 8039aabb6c9f802bca04cc77ca210060a5b53916, but didn't fix the original logic yet.

Matthew, could you apply Valdis' patch so this gets fixed once and for all?

Best regards,
Ari Sundholm
[email protected]