From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from services.gcu-squad.org (zoneX.GCU-Squad.org [194.213.125.0]) (using TLSv1 with cipher ECDHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTPS id 805CB2C00C2 for ; Thu, 26 Sep 2013 20:10:16 +1000 (EST) Date: Thu, 26 Sep 2013 11:39:36 +0200 From: Jean Delvare To: Christian Kujau Subject: Re: therm_adt746x: -3 invalid for parameter limit_adjust Message-ID: <20130926113936.24c10b78@endymion.delvare> In-Reply-To: References: Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Cc: colin@colino.net, Andrew Morton , Jingoo Han , linuxppc-dev@lists.ozlabs.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Hi Christian, On Thu, 26 Sep 2013 02:16:16 -0700 (PDT), Christian Kujau wrote: > Hi, > > after upgrading from 3.11 to 3.12-rc2, the therm_adt746x module could not > be loaded any more: > > therm_adt746x: `-2' invalid for parameter `limit_adjust' > > I've alwasy passed "limit_adjust=-3" (negative 3) to the module via > modprobe.conf, to lower the max temperature. Up until 3.11, loading the > module would print: > > adt746x: Lowering max temperatures from 73, 80, 109 to 67, 47, 67 > > I can load the module without limit_adjust or with positive values, but > really wanted to lower the temperature, so that the fan would kick in > earlier. For reference, this is what happens in 3.12-rc2: > > v--- limit_adjust > | > 0: adt746x: Lowering max temperatures from 73, 80, 109 to 70, 50, 70 > 1: adt746x: Lowering max temperatures from 73, 80, 109 to 71, 51, 71 > 2: adt746x: Lowering max temperatures from 73, 80, 109 to 72, 52, 72 > 3: adt746x: Lowering max temperatures from 73, 80, 109 to 73, 53, 73 > 4: adt746x: Lowering max temperatures from 73, 80, 109 to 74, 54, 74 > 10: adt746x: Lowering max temperatures from 73, 80, 109 to 80, 60, 80 > > Was passing negative values to therm_adt746x ever supported? As far as I can see, yes. > drivers/macintosh/therm_adt746x.c hasn't been touched in a while, this > means some other change did it. > > Before I attempt a full git bisect, any hints what could have caused this? I think it is a bug in: commit 6072ddc8520b86adfac6939ca32fb6e6c4de017a Author: Jingoo Han Date: Thu Sep 12 15:14:07 2013 -0700 kernel: replace strict_strto*() with kstrto*() The change was a good idea but the code itself is not, it has kstrtoul in many places where kstrtol should be used. Please try the following patch, hopefully that should fix your problem: From: Jean Delvare Subject: kernel/params: Fix handling of signed integer types Commit 6072ddc8520b86adfac6939ca32fb6e6c4de017a broke the handling of signed integer types, fix it. Reported-by: Christian Kujau Signed-off-by: Jean Delvare Cc: Jingoo Han Cc: Andrew Morton --- kernel/params.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) --- linux-3.12-rc2.orig/kernel/params.c 2013-09-24 00:41:09.000000000 +0200 +++ linux-3.12-rc2/kernel/params.c 2013-09-26 11:32:43.434586197 +0200 @@ -254,11 +254,11 @@ int parse_args(const char *doing, STANDARD_PARAM_DEF(byte, unsigned char, "%hhu", unsigned long, kstrtoul); -STANDARD_PARAM_DEF(short, short, "%hi", long, kstrtoul); +STANDARD_PARAM_DEF(short, short, "%hi", long, kstrtol); STANDARD_PARAM_DEF(ushort, unsigned short, "%hu", unsigned long, kstrtoul); -STANDARD_PARAM_DEF(int, int, "%i", long, kstrtoul); +STANDARD_PARAM_DEF(int, int, "%i", long, kstrtol); STANDARD_PARAM_DEF(uint, unsigned int, "%u", unsigned long, kstrtoul); -STANDARD_PARAM_DEF(long, long, "%li", long, kstrtoul); +STANDARD_PARAM_DEF(long, long, "%li", long, kstrtol); STANDARD_PARAM_DEF(ulong, unsigned long, "%lu", unsigned long, kstrtoul); int param_set_charp(const char *val, const struct kernel_param *kp) -- Jean Delvare