* new 64 bit math link fail in xfs in linux-next today
@ 2014-04-16 16:09 ` Paul Gortmaker
0 siblings, 0 replies; 11+ messages in thread
From: Paul Gortmaker @ 2014-04-16 16:09 UTC (permalink / raw)
To: Dave Chinner; +Cc: linux-next@vger.kernel.org, xfs
Hi Dave,
Not sure if this has been reported yet, but this new failure showed
up in the xtensa link of today's linux-next builds.
http://kisskb.ellerman.id.au/kisskb/buildresult/10938384/
Paul.
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: new 64 bit math link fail in xfs in linux-next today
2014-04-16 16:09 ` Paul Gortmaker
(?)
@ 2014-04-16 20:59 ` Mark Tinguely
2014-04-16 21:45 ` Paul Gortmaker
-1 siblings, 1 reply; 11+ messages in thread
From: Mark Tinguely @ 2014-04-16 20:59 UTC (permalink / raw)
To: Paul Gortmaker; +Cc: linux-next@vger.kernel.org, xfs
On 04/16/14 11:09, Paul Gortmaker wrote:
> Hi Dave,
>
> Not sure if this has been reported yet, but this new failure showed
> up in the xtensa link of today's linux-next builds.
>
> http://kisskb.ellerman.id.au/kisskb/buildresult/10938384/
>
> Paul.
>
Try this works on i386. sorry if my mailer wraps the line:
The roundup in commit 68c1fb5d should be a roundup_64()
because it is desired to round a 64 bit type by an integer
and that will result in a 64 bit value. On 32 bit machines
using roundup() in this case will result in the error:
ERROR: "__divdi3" [fs/xfs/xfs.ko] undefined!
Signed-off-by: Mark Tinguely <tinguely@sgi.com>
---
fs/xfs/xfs_aops.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Index: b/fs/xfs/xfs_aops.c
===================================================================
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -1365,7 +1365,7 @@ __xfs_get_blocks(
if (offset < i_size_read(inode) &&
offset + mapping_size >= i_size_read(inode)) {
/* limit mapping to block that spans EOF */
- mapping_size = roundup(i_size_read(inode) - offset,
+ mapping_size = roundup_64(i_size_read(inode) - offset,
1 << inode->i_blkbits);
}
if (mapping_size > LONG_MAX)
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 11+ messages in thread* Re: new 64 bit math link fail in xfs in linux-next today
2014-04-16 20:59 ` Mark Tinguely
@ 2014-04-16 21:45 ` Paul Gortmaker
0 siblings, 0 replies; 11+ messages in thread
From: Paul Gortmaker @ 2014-04-16 21:45 UTC (permalink / raw)
To: Mark Tinguely; +Cc: Dave Chinner, linux-next@vger.kernel.org, xfs
On 14-04-16 04:59 PM, Mark Tinguely wrote:
> On 04/16/14 11:09, Paul Gortmaker wrote:
>> Hi Dave,
>>
>> Not sure if this has been reported yet, but this new failure showed
>> up in the xtensa link of today's linux-next builds.
>>
>> http://kisskb.ellerman.id.au/kisskb/buildresult/10938384/
>>
>> Paul.
>>
>
> Try this works on i386. sorry if my mailer wraps the line:
Yep, also works for xtensa too; feel free to add a reported-by
and/or tested-by from me if you want.
Thanks,
Paul.
--
>
> The roundup in commit 68c1fb5d should be a roundup_64()
> because it is desired to round a 64 bit type by an integer
> and that will result in a 64 bit value. On 32 bit machines
> using roundup() in this case will result in the error:
>
> ERROR: "__divdi3" [fs/xfs/xfs.ko] undefined!
>
> Signed-off-by: Mark Tinguely <tinguely@sgi.com>
> ---
> fs/xfs/xfs_aops.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> Index: b/fs/xfs/xfs_aops.c
> ===================================================================
> --- a/fs/xfs/xfs_aops.c
> +++ b/fs/xfs/xfs_aops.c
> @@ -1365,7 +1365,7 @@ __xfs_get_blocks(
> if (offset < i_size_read(inode) &&
> offset + mapping_size >= i_size_read(inode)) {
> /* limit mapping to block that spans EOF */
> - mapping_size = roundup(i_size_read(inode) - offset,
> + mapping_size = roundup_64(i_size_read(inode) - offset,
> 1 << inode->i_blkbits);
> }
> if (mapping_size > LONG_MAX)
>
^ permalink raw reply [flat|nested] 11+ messages in thread* Re: new 64 bit math link fail in xfs in linux-next today
@ 2014-04-16 21:45 ` Paul Gortmaker
0 siblings, 0 replies; 11+ messages in thread
From: Paul Gortmaker @ 2014-04-16 21:45 UTC (permalink / raw)
To: Mark Tinguely; +Cc: linux-next@vger.kernel.org, xfs
On 14-04-16 04:59 PM, Mark Tinguely wrote:
> On 04/16/14 11:09, Paul Gortmaker wrote:
>> Hi Dave,
>>
>> Not sure if this has been reported yet, but this new failure showed
>> up in the xtensa link of today's linux-next builds.
>>
>> http://kisskb.ellerman.id.au/kisskb/buildresult/10938384/
>>
>> Paul.
>>
>
> Try this works on i386. sorry if my mailer wraps the line:
Yep, also works for xtensa too; feel free to add a reported-by
and/or tested-by from me if you want.
Thanks,
Paul.
--
>
> The roundup in commit 68c1fb5d should be a roundup_64()
> because it is desired to round a 64 bit type by an integer
> and that will result in a 64 bit value. On 32 bit machines
> using roundup() in this case will result in the error:
>
> ERROR: "__divdi3" [fs/xfs/xfs.ko] undefined!
>
> Signed-off-by: Mark Tinguely <tinguely@sgi.com>
> ---
> fs/xfs/xfs_aops.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> Index: b/fs/xfs/xfs_aops.c
> ===================================================================
> --- a/fs/xfs/xfs_aops.c
> +++ b/fs/xfs/xfs_aops.c
> @@ -1365,7 +1365,7 @@ __xfs_get_blocks(
> if (offset < i_size_read(inode) &&
> offset + mapping_size >= i_size_read(inode)) {
> /* limit mapping to block that spans EOF */
> - mapping_size = roundup(i_size_read(inode) - offset,
> + mapping_size = roundup_64(i_size_read(inode) - offset,
> 1 << inode->i_blkbits);
> }
> if (mapping_size > LONG_MAX)
>
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 11+ messages in thread
* [PATCH] xfs: fix 32bit __divdi3 undefined
2014-04-16 16:09 ` Paul Gortmaker
(?)
(?)
@ 2014-04-16 21:51 ` Mark Tinguely
2014-04-16 21:57 ` Eric Sandeen
2014-04-16 22:10 ` Dave Chinner
-1 siblings, 2 replies; 11+ messages in thread
From: Mark Tinguely @ 2014-04-16 21:51 UTC (permalink / raw)
To: XFS Filesystem; +Cc: Paul Gortmaker
[-- Attachment #1: xfs-fix-32bit-__divdi3-undefined.patch --]
[-- Type: text/plain, Size: 1157 bytes --]
The roundup in commit 68c1fb5d should be a roundup_64()
because it is desired to round a 64 bit type by an integer
and that will result in a 64 bit value. On 32 bit machines
using roundup() in this case will result in the error:
ERROR: "__divdi3" [fs/xfs/xfs.ko] undefined!
Reported-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Tested-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Mark Tinguely <tinguely@sgi.com>
---
fs/xfs/xfs_aops.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Index: b/fs/xfs/xfs_aops.c
===================================================================
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -1365,7 +1365,7 @@ __xfs_get_blocks(
if (offset < i_size_read(inode) &&
offset + mapping_size >= i_size_read(inode)) {
/* limit mapping to block that spans EOF */
- mapping_size = roundup(i_size_read(inode) - offset,
+ mapping_size = roundup_64(i_size_read(inode) - offset,
1 << inode->i_blkbits);
}
if (mapping_size > LONG_MAX)
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 11+ messages in thread* Re: [PATCH] xfs: fix 32bit __divdi3 undefined
2014-04-16 21:51 ` [PATCH] xfs: fix 32bit __divdi3 undefined Mark Tinguely
@ 2014-04-16 21:57 ` Eric Sandeen
2014-04-16 22:10 ` Dave Chinner
1 sibling, 0 replies; 11+ messages in thread
From: Eric Sandeen @ 2014-04-16 21:57 UTC (permalink / raw)
To: Mark Tinguely, XFS Filesystem; +Cc: Paul Gortmaker
On 4/16/14, 4:51 PM, Mark Tinguely wrote:
> The roundup in commit 68c1fb5d should be a roundup_64()
> because it is desired to round a 64 bit type by an integer
> and that will result in a 64 bit value. On 32 bit machines
> using roundup() in this case will result in the error:
>
> ERROR: "__divdi3" [fs/xfs/xfs.ko] undefined!
>
> Reported-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> Tested-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> Signed-off-by: Mark Tinguely <tinguely@sgi.com>
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
> ---
> fs/xfs/xfs_aops.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> Index: b/fs/xfs/xfs_aops.c
> ===================================================================
> --- a/fs/xfs/xfs_aops.c
> +++ b/fs/xfs/xfs_aops.c
> @@ -1365,7 +1365,7 @@ __xfs_get_blocks(
> if (offset < i_size_read(inode) &&
> offset + mapping_size >= i_size_read(inode)) {
> /* limit mapping to block that spans EOF */
> - mapping_size = roundup(i_size_read(inode) - offset,
> + mapping_size = roundup_64(i_size_read(inode) - offset,
> 1 << inode->i_blkbits);
> }
> if (mapping_size > LONG_MAX)
>
>
> _______________________________________________
> xfs mailing list
> xfs@oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs
>
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 11+ messages in thread* Re: [PATCH] xfs: fix 32bit __divdi3 undefined
2014-04-16 21:51 ` [PATCH] xfs: fix 32bit __divdi3 undefined Mark Tinguely
2014-04-16 21:57 ` Eric Sandeen
@ 2014-04-16 22:10 ` Dave Chinner
2014-04-16 22:30 ` Mark Tinguely
1 sibling, 1 reply; 11+ messages in thread
From: Dave Chinner @ 2014-04-16 22:10 UTC (permalink / raw)
To: Mark Tinguely; +Cc: Paul Gortmaker, XFS Filesystem
On Wed, Apr 16, 2014 at 04:51:56PM -0500, Mark Tinguely wrote:
> The roundup in commit 68c1fb5d should be a roundup_64()
> because it is desired to round a 64 bit type by an integer
> and that will result in a 64 bit value. On 32 bit machines
> using roundup() in this case will result in the error:
>
> ERROR: "__divdi3" [fs/xfs/xfs.ko] undefined!
>
> Reported-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> Tested-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> Signed-off-by: Mark Tinguely <tinguely@sgi.com>
> ---
> fs/xfs/xfs_aops.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> Index: b/fs/xfs/xfs_aops.c
> ===================================================================
> --- a/fs/xfs/xfs_aops.c
> +++ b/fs/xfs/xfs_aops.c
> @@ -1365,7 +1365,7 @@ __xfs_get_blocks(
> if (offset < i_size_read(inode) &&
> offset + mapping_size >= i_size_read(inode)) {
> /* limit mapping to block that spans EOF */
> - mapping_size = roundup(i_size_read(inode) - offset,
> + mapping_size = roundup_64(i_size_read(inode) - offset,
> 1 << inode->i_blkbits);
> }
> if (mapping_size > LONG_MAX)
I'm not going to apply this as a patch - I'm going to rebase the
branch with the fix in the original patch so we don't have a bisect
breakage in the branch. Don't worry, I'll add the fact you fixed the
bug into the commit message...
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 11+ messages in thread* Re: [PATCH] xfs: fix 32bit __divdi3 undefined
2014-04-16 22:10 ` Dave Chinner
@ 2014-04-16 22:30 ` Mark Tinguely
0 siblings, 0 replies; 11+ messages in thread
From: Mark Tinguely @ 2014-04-16 22:30 UTC (permalink / raw)
To: Dave Chinner; +Cc: Paul Gortmaker, XFS Filesystem
On 04/16/14 17:10, Dave Chinner wrote:
> On Wed, Apr 16, 2014 at 04:51:56PM -0500, Mark Tinguely wrote:
>> The roundup in commit 68c1fb5d should be a roundup_64()
>> because it is desired to round a 64 bit type by an integer
>> and that will result in a 64 bit value. On 32 bit machines
>> using roundup() in this case will result in the error:
>>
>> ERROR: "__divdi3" [fs/xfs/xfs.ko] undefined!
>>
>> Reported-by: Paul Gortmaker<paul.gortmaker@windriver.com>
>> Tested-by: Paul Gortmaker<paul.gortmaker@windriver.com>
>> Signed-off-by: Mark Tinguely<tinguely@sgi.com>
>> ---
>> fs/xfs/xfs_aops.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> Index: b/fs/xfs/xfs_aops.c
>> ===================================================================
>> --- a/fs/xfs/xfs_aops.c
>> +++ b/fs/xfs/xfs_aops.c
>> @@ -1365,7 +1365,7 @@ __xfs_get_blocks(
>> if (offset< i_size_read(inode)&&
>> offset + mapping_size>= i_size_read(inode)) {
>> /* limit mapping to block that spans EOF */
>> - mapping_size = roundup(i_size_read(inode) - offset,
>> + mapping_size = roundup_64(i_size_read(inode) - offset,
>> 1<< inode->i_blkbits);
>> }
>> if (mapping_size> LONG_MAX)
>
> I'm not going to apply this as a patch - I'm going to rebase the
> branch with the fix in the original patch so we don't have a bisect
> breakage in the branch. Don't worry, I'll add the fact you fixed the
> bug into the commit message...
>
> Cheers,
>
> Dave.
Makes sense.
--Mark.
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: new 64 bit math link fail in xfs in linux-next today
2014-04-16 16:09 ` Paul Gortmaker
@ 2014-04-16 21:57 ` Dave Chinner
-1 siblings, 0 replies; 11+ messages in thread
From: Dave Chinner @ 2014-04-16 21:57 UTC (permalink / raw)
To: Paul Gortmaker; +Cc: xfs, linux-next@vger.kernel.org
On Wed, Apr 16, 2014 at 12:09:37PM -0400, Paul Gortmaker wrote:
> Hi Dave,
>
> Not sure if this has been reported yet, but this new failure showed
> up in the xtensa link of today's linux-next builds.
>
> http://kisskb.ellerman.id.au/kisskb/buildresult/10938384/
fs/built-in.o: In function `__xfs_get_blocks':
/home/dave/src/build/x86-64/xfsdev/fs/xfs/xfs_aops.c:1368: undefined reference to `__divdi3'
Fmeh. The hunk that throws the error:
@@ -1354,6 +1362,12 @@ __xfs_get_blocks(
ASSERT(mapping_size > 0);
if (mapping_size > size)
mapping_size = size;
+ if (offset < i_size_read(inode) &&
+ offset + mapping_size >= i_size_read(inode)) {
+ /* limit mapping to block that spans EOF */
+ mapping_size = roundup(i_size_read(inode) - offset,
+ 1 << inode->i_blkbits);
+ }
if (mapping_size > LONG_MAX)
mapping_size = LONG_MAX;
You'd think a generic function like roundup() would handle that sort
of 64/32 bit type problem, yes? But it doesn't, despite what the
comment implies:
/* The `const' in roundup() prevents gcc-3.3 from calling __divdi3 */
#define roundup(x, y) ( \
{ \
const typeof(y) __y = y; \
(((x) + (__y - 1)) / __y) * __y; \
} \
)
It doesn't even have type checking that might warn you of impending
problems...
<sigh>
/me looks around in the XFS code for a bit
... and finds roundup_64() in fs/xfs/xfs_linux.h. That uses do_div()
to avoid this problem. And it's a static inline function, so it has
type checking, too.
And so this hunk fixes the problem:
@@ -1365,7 +1365,7 @@ __xfs_get_blocks(
if (offset < i_size_read(inode) &&
offset + mapping_size >= i_size_read(inode)) {
/* limit mapping to block that spans EOF */
- mapping_size = roundup(i_size_read(inode) - offset,
+ mapping_size = roundup_64(i_size_read(inode) - offset,
1 << inode->i_blkbits);
}
if (mapping_size > LONG_MAX)
Can you test it, please, and I'll push out an updated branch
hopefully in time for today's linux-next build....
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
^ permalink raw reply [flat|nested] 11+ messages in thread* Re: new 64 bit math link fail in xfs in linux-next today
@ 2014-04-16 21:57 ` Dave Chinner
0 siblings, 0 replies; 11+ messages in thread
From: Dave Chinner @ 2014-04-16 21:57 UTC (permalink / raw)
To: Paul Gortmaker; +Cc: linux-next@vger.kernel.org, xfs
On Wed, Apr 16, 2014 at 12:09:37PM -0400, Paul Gortmaker wrote:
> Hi Dave,
>
> Not sure if this has been reported yet, but this new failure showed
> up in the xtensa link of today's linux-next builds.
>
> http://kisskb.ellerman.id.au/kisskb/buildresult/10938384/
fs/built-in.o: In function `__xfs_get_blocks':
/home/dave/src/build/x86-64/xfsdev/fs/xfs/xfs_aops.c:1368: undefined reference to `__divdi3'
Fmeh. The hunk that throws the error:
@@ -1354,6 +1362,12 @@ __xfs_get_blocks(
ASSERT(mapping_size > 0);
if (mapping_size > size)
mapping_size = size;
+ if (offset < i_size_read(inode) &&
+ offset + mapping_size >= i_size_read(inode)) {
+ /* limit mapping to block that spans EOF */
+ mapping_size = roundup(i_size_read(inode) - offset,
+ 1 << inode->i_blkbits);
+ }
if (mapping_size > LONG_MAX)
mapping_size = LONG_MAX;
You'd think a generic function like roundup() would handle that sort
of 64/32 bit type problem, yes? But it doesn't, despite what the
comment implies:
/* The `const' in roundup() prevents gcc-3.3 from calling __divdi3 */
#define roundup(x, y) ( \
{ \
const typeof(y) __y = y; \
(((x) + (__y - 1)) / __y) * __y; \
} \
)
It doesn't even have type checking that might warn you of impending
problems...
<sigh>
/me looks around in the XFS code for a bit
... and finds roundup_64() in fs/xfs/xfs_linux.h. That uses do_div()
to avoid this problem. And it's a static inline function, so it has
type checking, too.
And so this hunk fixes the problem:
@@ -1365,7 +1365,7 @@ __xfs_get_blocks(
if (offset < i_size_read(inode) &&
offset + mapping_size >= i_size_read(inode)) {
/* limit mapping to block that spans EOF */
- mapping_size = roundup(i_size_read(inode) - offset,
+ mapping_size = roundup_64(i_size_read(inode) - offset,
1 << inode->i_blkbits);
}
if (mapping_size > LONG_MAX)
Can you test it, please, and I'll push out an updated branch
hopefully in time for today's linux-next build....
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 11+ messages in thread