Re: [BUGFIX] Fix AES-NI CTR optimization compiling failure withgas 2.16.1

From: Huang Ying
Date: Wed Mar 24 2010 - 02:24:51 EST


Hi, Andrew,

On Wed, 2010-03-24 at 05:23 +0800, Andrew Morton wrote:
> On Fri, 12 Mar 2010 15:01:47 +0800
> Huang Ying <ying.huang@xxxxxxxxx> wrote:
>
> > Andrew Morton reported that AES-NI CTR optimization failed to compile
> > with gas 2.16.1, the error message is as follow:
> >
> > arch/x86/crypto/aesni-intel_asm.S: Assembler messages:
> > arch/x86/crypto/aesni-intel_asm.S:752: Error: suffix or operands invalid for `movq'
> > arch/x86/crypto/aesni-intel_asm.S:753: Error: suffix or operands invalid for `movq'
> >
> > To fix this, a gas macro is defined to assemble movq with 64bit
> > general purpose registers and XMM registers. The macro will generate
> > the raw .byte sequence for needed instructions.
> >
>
> Still no go.
>
> arch/x86/crypto/aesni-intel_asm.S: Assembler messages:
> arch/x86/crypto/aesni-intel_asm.S:752: Error: bad expression
> arch/x86/crypto/aesni-intel_asm.S:752: Error: junk at end of line, first unrecognized character is `1'
> arch/x86/crypto/aesni-intel_asm.S:753: Error: bad expression
> arch/x86/crypto/aesni-intel_asm.S:753: Error: junk at end of line, first unrecognized character is `1'
>
> it doesn't like this:
>
> REG_TYPE movq_r64_xmm_opd1_type \opd1

Try to compile the latest kernel with binutils 2.16.1 this time to debug
and verify my fix. Now, with patch followed, aesni-intel_asm.S can be
compiled. But I failed to compile whole kernel with binutils 2.16.1.
First ld report unknown options, after use latest ld instead of that of
2.16.1, "as" reports errors as follow:

AS arch/x86/kernel/acpi/realmode/wakeup.o
cc1: error: CPU you selected does not support x86-64 instruction set
cc1: error: CPU you selected does not support x86-64 instruction set
cc1: warning: -mregparm is ignored in 64-bit mode

Can you tell me how to setup the building environment with binutils
2.16.1? I just build it by hand and make the installed bin directory the
first one in $PATH.

Best Regards,
Huang Ying
------------------------------------------------------------------------>
Subject: [BUGFIX] Fix another AES-NI CTR optimization compiling failure with gas 2.16.1

The previous AES-NI CTR optimization compiling failure gas 2.16.1 fix
introduces another compiling failure by itself. This patch fixes that.

Reported-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Huang Ying <ying.huang@xxxxxxxxx>
---
arch/x86/include/asm/inst.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

--- a/arch/x86/include/asm/inst.h
+++ b/arch/x86/include/asm/inst.h
@@ -120,9 +120,9 @@
.macro REG_TYPE type reg
R64_NUM reg_type_r64 \reg
XMM_NUM reg_type_xmm \reg
- .if reg_type_r64 != REG_NUM_INVALID
+ .if reg_type_r64 <> REG_NUM_INVALID
\type = REG_TYPE_R64
- .elseif reg_type_xmm != REG_NUM_INVALID
+ .elseif reg_type_xmm <> REG_NUM_INVALID
\type = REG_TYPE_XMM
.else
\type = REG_TYPE_INVALID


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