From: Jan-Simon Möller <[email protected]>
Author: PaX Team <pageexec at freemail.hu>
ML-Post: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20120507/142707.html
URL: http://llvm.linuxfoundation.org
Merge: Jan-Simon Möller <dl9pf at gmx.de>
Description:
Clang chokes on the notation "insw (%%dx)" but works for "insw %%dx" (outsw likewise);
GNU as accepts both forms.
Signed-off-by: Jan-Simon Möller <[email protected]>
CC: Jan-Simon Möller <[email protected]>
CC: PaX Team <[email protected]>
CC: Behan Webster <[email protected]>
CC: "James E.J. Bottomley" <[email protected]>
CC: [email protected]
CC: [email protected]
---
drivers/scsi/in2000.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/in2000.h b/drivers/scsi/in2000.h
index 5821e1f..cd8d852 100644
--- a/drivers/scsi/in2000.h
+++ b/drivers/scsi/in2000.h
@@ -72,7 +72,7 @@ int __dummy_1,__dummy_2; \
orl %%ecx, %%ecx \n \
jz 1f \n \
rep \n \
- insw (%%dx),%%es:(%%edi) \n \
+ insw %%dx,%%es:(%%edi) \n \
1: " \
: "=D" (sp) ,"=c" (__dummy_1) ,"=d" (__dummy_2) /* output */ \
: "2" (f), "0" (sp), "1" (i) /* input */ \
@@ -87,7 +87,7 @@ int __dummy_1,__dummy_2; \
orl %%ecx, %%ecx \n \
jz 1f \n \
rep \n \
- outsw %%ds:(%%esi),(%%dx) \n \
+ outsw %%ds:(%%esi),%%dx \n \
1: " \
: "=S" (sp) ,"=c" (__dummy_1) ,"=d" (__dummy_2)/* output */ \
: "2" (f), "0" (sp), "1" (i) /* input */ \
--
1.8.1.4
On Tue, Aug 20, 2013 at 11:19:06PM +0200, [email protected] wrote:
> From: Jan-Simon M??ller <[email protected]>
>
> Author: PaX Team <pageexec at freemail.hu>
> ML-Post: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20120507/142707.html
> URL: http://llvm.linuxfoundation.org
>
> Merge: Jan-Simon M??ller <dl9pf at gmx.de>
>
> Description:
> Clang chokes on the notation "insw (%%dx)" but works for "insw %%dx" (outsw likewise);
> GNU as accepts both forms.
>
> Signed-off-by: Jan-Simon M??ller <[email protected]>
I think the right fix is to simply kill the FAST_READ_IO/FAST_WRITE_IO
defines and their associated inline assembly entirely.
> >
> > Description:
> > Clang chokes on the notation "insw (%%dx)" but works for "insw %%dx"
> > (outsw likewise); GNU as accepts both forms.
> >
> > Signed-off-by: Jan-Simon M??ller <[email protected]>
>
> I think the right fix is to simply kill the FAST_READ_IO/FAST_WRITE_IO
> defines and their associated inline assembly entirely.
No objection ;) .
On Wednesday 21 August 2013 09:26:14 Jan-Simon M?ller wrote:
> > > Description:
> > > Clang chokes on the notation "insw (%%dx)" but works for "insw %%dx"
> > > (outsw likewise); GNU as accepts both forms.
> > >
> > > Signed-off-by: Jan-Simon M??ller <[email protected]>
> >
> > I think the right fix is to simply kill the FAST_READ_IO/FAST_WRITE_IO
> > defines and their associated inline assembly entirely.
>
> No objection ;) .
@Christoph Hellwig:
Do you take care of removing it then ?
BR,
JS
Ping.
@Christoph do you take care of removing then ?
Or can we merge this patch as-is now.
Best,
JS
On Wednesday 21 August 2013 15:11:33 Jan-Simon M?ller wrote:
> On Wednesday 21 August 2013 09:26:14 Jan-Simon M?ller wrote:
> > > > Description:
> > > > Clang chokes on the notation "insw (%%dx)" but works for "insw %%dx"
> > > > (outsw likewise); GNU as accepts both forms.
> > > >
> > > > Signed-off-by: Jan-Simon M??ller <[email protected]>
> > >
> > > I think the right fix is to simply kill the FAST_READ_IO/FAST_WRITE_IO
> > > defines and their associated inline assembly entirely.
> >
> > No objection ;) .
>
> @Christoph Hellwig:
> Do you take care of removing it then ?
>
> BR,
> JS