linux-kernel - Re: [PATCH 3/3] staging: xgifb: Initialize uninitialized variables

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <4C87385B.1000101@gmail.com>
Date:	Wed, 08 Sep 2010 09:16:43 +0200
From:	Jiri Slaby <jirislaby@...il.com>
To:	Greg KH <gregkh@...e.de>
CC:	Javier Martinez Canillas <martinez.javier@...il.com>,
	Bill Pemberton <wfp5p@...ginia.edu>,
	Arnaud Patard <apatard@...driva.com>,
	Randy Dunlap <randy.dunlap@...cle.com>,
	devel@...verdev.osuosl.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 3/3] staging: xgifb: Initialize uninitialized variables

On 09/08/2010 02:38 AM, Greg KH wrote:
> On Tue, Sep 07, 2010 at 10:46:27AM +0200, Jiri Slaby wrote:
>> On 09/07/2010 07:34 AM, Javier Martinez Canillas wrote:
>>> Current patch solves compilation warnings in staging/xgifb for using possibly uninitialized variables.
>>>
>>> Signed-off-by: Javier Martinez Canillas <martinez.javier@...il.com>
>>> ---
>>>  drivers/staging/xgifb/vb_setmode.c |    4 ++--
>>>  1 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/staging/xgifb/vb_setmode.c b/drivers/staging/xgifb/vb_setmode.c
>>> index c4db2fc..720a592 100644
>>> --- a/drivers/staging/xgifb/vb_setmode.c
>>> +++ b/drivers/staging/xgifb/vb_setmode.c
>>> @@ -376,7 +376,7 @@ void InitTo330Pointer(unsigned char ChipType, struct vb_device_info *pVBInfo)
>>>  unsigned char XGISetModeNew(struct xgi_hw_device_info *HwDeviceExtension,
>>>  			unsigned short ModeNo)
>>>  {
>>> -	unsigned short ModeIdIndex;
>>> +	unsigned short ModeIdIndex = 0;
>>>  	/* unsigned char *pVBInfo->FBAddr = HwDeviceExtension->pjVideoMemoryAddress; */
>>>  	struct vb_device_info VBINF;
>>>  	struct vb_device_info *pVBInfo = &VBINF;
>>
>> No, please don't. This is a compiler bug.
>>
>> The first use in that function is:
>> XGI_SearchModeID( ModeNo , &ModeIdIndex, pVBInfo );
>>
>> and that function contains:
>> if (XYZ) {
>>   for( *ModeIdIndex = 0 ; ; ( *ModeIdIndex )++ )
>>     ...
>> } else {
>>   for( *ModeIdIndex = 0 ; ; ( *ModeIdIndex )++ )
>>     ...
>> }
>>
>> So your compiler is foolish if that emits a warning.
> 
> No, I don't see a way that the compiler could really figure that type of
> logic out.  It doesn't know that XGI_SearchModeID() always will set that
> field.  Heck, at first glance I couldn't even figure it out :)

No, the compiler could. Trust me :). Actually you've just hidden a bug
in XGI_SearchModeID. There is #ifdef LINUX and LINUX is defined in
main.h. But that file is not included in vb_setmode.c. Hence the
uninitialized warnings...

> So I don't mind applying this patch, as I'll just constantly keep
> getting people sending me the same change, when they could be off
> working on other things that this driver needs.

So instead of hiding bugs, please fix the right ones.

thanks,
-- 
js
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/