[PATCH -mm] megaraid gcc 4.1 warning fix

[PATCH -mm] megaraid gcc 4.1 warning fix

[Daniel Walker]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 2785 bytes --]

Fixes the following warning,

drivers/scsi/megaraid.c: In function ‘issue_scb’:
drivers/scsi/megaraid.c:1153: warning: passing argument 2 of ‘writel’ makes pointer from integer without a cast
drivers/scsi/megaraid.c: In function ‘issue_scb_block’:
drivers/scsi/megaraid.c:1216: warning: passing argument 2 of ‘writel’ makes pointer from integer without a cast
drivers/scsi/megaraid.c:1229: warning: passing argument 2 of ‘writel’ makes pointer from integer without a cast
drivers/scsi/megaraid.c:1231: warning: passing argument 1 of ‘readl’ makes pointer from integer without a cast
drivers/scsi/megaraid.c: In function ‘megaraid_isr_memmapped’:
drivers/scsi/megaraid.c:1361: warning: passing argument 1 of ‘readl’ makes pointer from integer without a cast
drivers/scsi/megaraid.c:1368: warning: passing argument 2 of ‘writel’ makes pointer from integer without a cast
drivers/scsi/megaraid.c:1387: warning: passing argument 2 of ‘writel’ makes pointer from integer without a cast
drivers/scsi/megaraid.c:1391: warning: passing argument 1 of ‘readl’ makes pointer from integer without a cast
drivers/scsi/megaraid.c: In function ‘megadev_ioctl’:
drivers/scsi/megaraid.c:3665: warning: ignoring return value of ‘copy_to_user’, declared with attribute warn_unused_result


Signed-Off-By: Daniel Walker <dwalker@mvista.com>

Index: linux-2.6.16/drivers/scsi/megaraid.c
===================================================================
--- linux-2.6.16.orig/drivers/scsi/megaraid.c
+++ linux-2.6.16/drivers/scsi/megaraid.c
@@ -73,10 +73,10 @@ static unsigned short int max_mbox_busy_
 module_param(max_mbox_busy_wait, ushort, 0);
 MODULE_PARM_DESC(max_mbox_busy_wait, "Maximum wait for mailbox in microseconds if busy (default=MBOX_BUSY_WAIT=10)");
 
-#define RDINDOOR(adapter)		readl((adapter)->base + 0x20)
-#define RDOUTDOOR(adapter)		readl((adapter)->base + 0x2C)
-#define WRINDOOR(adapter,value)		writel(value, (adapter)->base + 0x20)
-#define WROUTDOOR(adapter,value)	writel(value, (adapter)->base + 0x2C)
+#define RDINDOOR(adapter)		readl((void*)((adapter)->base + 0x20))
+#define RDOUTDOOR(adapter)		readl((void*)((adapter)->base + 0x2C))
+#define WRINDOOR(adapter,value)		writel(value, (void *)((adapter)->base + 0x20))
+#define WROUTDOOR(adapter,value)	writel(value, (void*)((adapter)->base + 0x2C))
 
 /*
  * Global variables
@@ -3662,8 +3662,9 @@ megadev_ioctl(struct inode *inode, struc
 			 * Send the request sense data also, irrespective of
 			 * whether the user has asked for it or not.
 			 */
-			copy_to_user(upthru->reqsensearea,
-					pthru->reqsensearea, 14);
+			if (copy_to_user(upthru->reqsensearea,
+					pthru->reqsensearea, 14))
+				return -EFAULT;
 
 freemem_and_return:
 			if( pthru->dataxferlen ) {

Re: [PATCH -mm] megaraid gcc 4.1 warning fix

[Pekka Enberg]

Hi Daniel,

On 5/10/06, Daniel Walker <dwalker@mvista.com> wrote:
> -#define RDINDOOR(adapter)              readl((adapter)->base + 0x20)
> -#define RDOUTDOOR(adapter)             readl((adapter)->base + 0x2C)
> -#define WRINDOOR(adapter,value)                writel(value, (adapter)->base + 0x20)
> -#define WROUTDOOR(adapter,value)       writel(value, (adapter)->base + 0x2C)
> +#define RDINDOOR(adapter)              readl((void*)((adapter)->base + 0x20))
> +#define RDOUTDOOR(adapter)             readl((void*)((adapter)->base + 0x2C))
> +#define WRINDOOR(adapter,value)                writel(value, (void *)((adapter)->base + 0x20))
> +#define WROUTDOOR(adapter,value)       writel(value, (void*)((adapter)->base + 0x2C))

This looks wrong. adapter->base should be void __iomem *.

                                        Pekka

Re: [PATCH -mm] megaraid gcc 4.1 warning fix

[Alan Cox]

On Maw, 2006-05-09 at 19:56 -0700, Daniel Walker wrote:
> Fixes the following warning,
> 
> drivers/scsi/megaraid.c: In function ‘issue_scb’:
> drivers/scsi/megaraid.c:1153: warning: passing argument 2 of ‘writel’ makes pointer from integer without a cast

And adds an exploitable memory leak. Please don't fix "bugs" blindly but
check that the error path still releases all the relevant resources.

In this case its probably sufficient just to set rval and fal through,
but each one needs to be reviewed properly.

Alan

Re: [PATCH -mm] megaraid gcc 4.1 warning fix

[Daniel Walker]

On Wed, 2006-05-10 at 11:39 +0100, Alan Cox wrote:
> On Maw, 2006-05-09 at 19:56 -0700, Daniel Walker wrote:
> > Fixes the following warning,
> > 
> > drivers/scsi/megaraid.c: In function ‘issue_scb’:
> > drivers/scsi/megaraid.c:1153: warning: passing argument 2 of ‘writel’ makes pointer from integer without a cast
> 
> And adds an exploitable memory leak. Please don't fix "bugs" blindly but
> check that the error path still releases all the relevant resources.

I did nothing blindly .. I have a queue of 50+ more of these. I
researched every error path , and made sure the warning wasn't an actual
problem. If it was a problem , or I wasn't sure if the fix was 100% I
sent it to LKML ..

Daniel

Re: [PATCH -mm] megaraid gcc 4.1 warning fix

[Daniel Walker]

On Wed, 2006-05-10 at 11:39 +0100, Alan Cox wrote:
> On Maw, 2006-05-09 at 19:56 -0700, Daniel Walker wrote:
> > Fixes the following warning,
> > 
> > drivers/scsi/megaraid.c: In function ‘issue_scb’:
> > drivers/scsi/megaraid.c:1153: warning: passing argument 2 of ‘writel’ makes pointer from integer without a cast
> 
> And adds an exploitable memory leak. Please don't fix "bugs" blindly but
> check that the error path still releases all the relevant resources.
> 
> In this case its probably sufficient just to set rval and fal through,
> but each one needs to be reviewed properly.

The writel on 1153 is attached to a memory leak, or I add one in this
patch ?

Daniel

Re: [PATCH -mm] megaraid gcc 4.1 warning fix

[Daniel Walker]

On Wed, 2006-05-10 at 16:08 +0100, Alan Cox wrote:
> On Mer, 2006-05-10 at 07:21 -0700, Daniel Walker wrote:
> > The writel on 1153 is attached to a memory leak, or I add one in this
> > patch ?
> 
> You exit without cleaning up.


Your talking about this warning ,

drivers/scsi/megaraid.c: In function ‘megadev_ioctl’:
drivers/scsi/megaraid.c:3665: warning: ignoring return value of ‘copy_to_user’, declared with attribute warn_unused_result


Ahh, I see .. I'll set rval as you suggested ..

Daniel

Re: [PATCH -mm] megaraid gcc 4.1 warning fix

[Alan Cox]

On Mer, 2006-05-10 at 07:21 -0700, Daniel Walker wrote:
> The writel on 1153 is attached to a memory leak, or I add one in this
> patch ?

You exit without cleaning up.