public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] I2O: mark i2o_config broken on 64-bit
@ 2006-10-01 15:56 Jeff Garzik
  2006-10-01 16:11 ` Roland Dreier
  0 siblings, 1 reply; 3+ messages in thread
From: Jeff Garzik @ 2006-10-01 15:56 UTC (permalink / raw)
  To: markus.lidel, Andrew Morton, LKML


Code comments, reading the code, and gcc warnings all indicate that this
module isn't safe at all on 64-bit.

Disable this one I2O module on 64-bit, and update the bug indicator
comments to be to more grep-able.

Signed-off-by: Jeff Garzik <jeff@garzik.org>

diff --git a/drivers/message/i2o/Kconfig b/drivers/message/i2o/Kconfig
index 6443392..0661599 100644
--- a/drivers/message/i2o/Kconfig
+++ b/drivers/message/i2o/Kconfig
@@ -56,7 +56,7 @@ config I2O_EXT_ADAPTEC_DMA64
 
 config I2O_CONFIG
 	tristate "I2O Configuration support"
-	depends on I2O
+	depends on I2O && 32BIT
 	---help---
 	  Say Y for support of the configuration interface for the I2O adapters.
 	  If you have a RAID controller from Adaptec and you want to use the
diff --git a/drivers/message/i2o/i2o_config.c b/drivers/message/i2o/i2o_config.c
index 7d23e08..69666e8 100644
--- a/drivers/message/i2o/i2o_config.c
+++ b/drivers/message/i2o/i2o_config.c
@@ -600,7 +600,7 @@ static int i2o_cfg_passthru32(struct fil
 			rcode = -EFAULT;
 			goto cleanup;
 		}
-		// TODO 64bit fix
+		// FIXME: broken on 64-bit
 		sg = (struct sg_simple_element *)((&msg->u.head[0]) +
 						  sg_offset);
 		sg_count =
@@ -640,7 +640,7 @@ static int i2o_cfg_passthru32(struct fil
 			/* Copy in the user's SG buffer if necessary */
 			if (sg[i].
 			    flag_count & 0x04000000 /*I2O_SGL_FLAGS_DIR */ ) {
-				// TODO 64bit fix
+				// FIXME: broken on 64-bit
 				if (copy_from_user
 				    (p->virt,
 				     (void __user *)(unsigned long)sg[i].
@@ -652,7 +652,7 @@ static int i2o_cfg_passthru32(struct fil
 					goto sg_list_cleanup;
 				}
 			}
-			//TODO 64bit fix
+			//FIXME: broken on 64-bit
 			sg[i].addr_bus = (u32) p->phys;
 		}
 	}
@@ -667,7 +667,7 @@ static int i2o_cfg_passthru32(struct fil
 		u32 msg[I2O_OUTBOUND_MSG_FRAME_SIZE];
 		/* Copy back the Scatter Gather buffers back to user space */
 		u32 j;
-		// TODO 64bit fix
+		// FIXME: broken on 64-bit
 		struct sg_simple_element *sg;
 		int sg_size;
 
@@ -688,7 +688,7 @@ static int i2o_cfg_passthru32(struct fil
 		sg_count =
 		    (size - sg_offset * 4) / sizeof(struct sg_simple_element);
 
-		// TODO 64bit fix
+		// FIXME: broken on 64-bit
 		sg = (struct sg_simple_element *)(msg + sg_offset);
 		for (j = 0; j < sg_count; j++) {
 			/* Copy out the SG list to user's buffer if necessary */
@@ -696,7 +696,7 @@ static int i2o_cfg_passthru32(struct fil
 			    (sg[j].
 			     flag_count & 0x4000000 /*I2O_SGL_FLAGS_DIR */ )) {
 				sg_size = sg[j].flag_count & 0xffffff;
-				// TODO 64bit fix
+				// FIXME: broken on 64-bit
 				if (copy_to_user
 				    ((void __user *)(u64) sg[j].addr_bus,
 				     sg_list[j].virt, sg_size)) {
@@ -833,7 +833,7 @@ static int i2o_cfg_passthru(unsigned lon
 			rcode = -EFAULT;
 			goto cleanup;
 		}
-		// TODO 64bit fix
+		// FIXME: broken on 64-bit
 		sg = (struct sg_simple_element *)((&msg->u.head[0]) +
 						  sg_offset);
 		sg_count =
@@ -870,7 +870,7 @@ static int i2o_cfg_passthru(unsigned lon
 			/* Copy in the user's SG buffer if necessary */
 			if (sg[i].
 			    flag_count & 0x04000000 /*I2O_SGL_FLAGS_DIR */ ) {
-				// TODO 64bit fix
+				// FIXME: broken on 64-bit
 				if (copy_from_user
 				    (p, (void __user *)sg[i].addr_bus,
 				     sg_size)) {
@@ -881,7 +881,7 @@ static int i2o_cfg_passthru(unsigned lon
 					goto sg_list_cleanup;
 				}
 			}
-			//TODO 64bit fix
+			//FIXME: broken on 64-bit
 			sg[i].addr_bus = virt_to_bus(p);
 		}
 	}
@@ -896,7 +896,7 @@ static int i2o_cfg_passthru(unsigned lon
 		u32 msg[128];
 		/* Copy back the Scatter Gather buffers back to user space */
 		u32 j;
-		// TODO 64bit fix
+		// FIXME: broken on 64-bit
 		struct sg_simple_element *sg;
 		int sg_size;
 
@@ -917,7 +917,7 @@ static int i2o_cfg_passthru(unsigned lon
 		sg_count =
 		    (size - sg_offset * 4) / sizeof(struct sg_simple_element);
 
-		// TODO 64bit fix
+		// FIXME: broken on 64-bit
 		sg = (struct sg_simple_element *)(msg + sg_offset);
 		for (j = 0; j < sg_count; j++) {
 			/* Copy out the SG list to user's buffer if necessary */
@@ -925,7 +925,7 @@ static int i2o_cfg_passthru(unsigned lon
 			    (sg[j].
 			     flag_count & 0x4000000 /*I2O_SGL_FLAGS_DIR */ )) {
 				sg_size = sg[j].flag_count & 0xffffff;
-				// TODO 64bit fix
+				// FIXME: broken on 64-bit
 				if (copy_to_user
 				    ((void __user *)sg[j].addr_bus, sg_list[j],
 				     sg_size)) {

^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH] I2O: mark i2o_config broken on 64-bit
  2006-10-01 15:56 [PATCH] I2O: mark i2o_config broken on 64-bit Jeff Garzik
@ 2006-10-01 16:11 ` Roland Dreier
  2006-10-01 16:45   ` Markus Lidel
  0 siblings, 1 reply; 3+ messages in thread
From: Roland Dreier @ 2006-10-01 16:11 UTC (permalink / raw)
  To: Jeff Garzik; +Cc: markus.lidel, Andrew Morton, LKML

 > -			//TODO 64bit fix
 > +			//FIXME: broken on 64-bit
 >  			sg[i].addr_bus = (u32) p->phys;

This looks worse than just broken on 64 bit.  I didn't even attempt to
understand what's going on here, but would this even work on 32 bit
systems that have physical addresses above 4 gigs (eg i386 with PAE)?

 - R.

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] I2O: mark i2o_config broken on 64-bit
  2006-10-01 16:11 ` Roland Dreier
@ 2006-10-01 16:45   ` Markus Lidel
  0 siblings, 0 replies; 3+ messages in thread
From: Markus Lidel @ 2006-10-01 16:45 UTC (permalink / raw)
  To: Roland Dreier; +Cc: Jeff Garzik, Andrew Morton, LKML

Hello,

Roland Dreier wrote:
>  > -			//TODO 64bit fix
>  > +			//FIXME: broken on 64-bit
>  >  			sg[i].addr_bus = (u32) p->phys;
> This looks worse than just broken on 64 bit.  I didn't even attempt to
> understand what's going on here, but would this even work on 32 bit
> systems that have physical addresses above 4 gigs (eg i386 with PAE)?

It's a kind of user-space driver :-D This part of the driver just send 
the data generated in the user-space program 1:1 to the controller. It 
works on 64-bit, but i doubt it works with 64-bit DMA addresses. This 
part of the code was copied from "drivers/scsi/dpt_i2o.c". I already 
tried to replace it with an IMHO nicer sysfs solution, but because this 
also required an "workaround" it was dropped again.



Best regards,


Markus Lidel
------------------------------------------
Markus Lidel (Senior IT Consultant)

Shadow Connect GmbH
Carl-Reisch-Weg 12
D-86381 Krumbach
Germany

Phone:  +49 82 82/99 51-0
Fax:    +49 82 82/99 51-11

E-Mail: Markus.Lidel@shadowconnect.com
URL:    http://www.shadowconnect.com

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2006-10-01 16:45 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-10-01 15:56 [PATCH] I2O: mark i2o_config broken on 64-bit Jeff Garzik
2006-10-01 16:11 ` Roland Dreier
2006-10-01 16:45   ` Markus Lidel

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox