Help!

[PATCH 1/2] x86, crypto, Generate .byte code for some new ..

 
  

Post new topic   General Reply to Topic (not reply to a specific post)    Forums Home -> Kernel RSS
Next:  CVE-2009-2584  
Author Message
Huang Ying
External


Since: May 27, 2009
Posts: 14



PostPosted: Thu Nov 05, 2009 7:10 am    Post subject: [PATCH 1/2] x86, crypto, Generate .byte code for some new instructions via gas macro
Archived from groups: linux>kernel (more info?)

It will take some time for binutils (gas) to support some newly added
instructions, such as SSE4.1 instructions or the AES-NI instructions
found in upcoming Intel CPU.

To make the source code can be compiled by old binutils, .byte code is
used instead of the assembly instruction. But the readability and
flexibility of raw .byte code is not good.

This patch solves the issue of raw .byte code via generating it via
assembly instruction like gas macro. The syntax is as close as
possible to real assembly instruction.

Some helper macros such as MODRM is not a full feature
implementation. It can be extended when necessary.

Signed-off-by: Huang Ying <ying.huang DeleteThis @intel.com>
---
arch/x86/include/asm/inst.h | 150 ++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 150 insertions(+)

--- /dev/null
+++ b/arch/x86/include/asm/inst.h
@@ -0,0 +1,150 @@
+/*
+ * Generate .byte code for some instructions not supported by old
+ * binutils.
+ */
+#ifndef X86_ASM_INST_H
+#define X86_ASM_INST_H
+
+#ifdef __ASSEMBLY__
+
+ .macro XMM_NUM opd xmm
+ .ifc \xmm,%xmm0
+ \opd = 0
+ .endif
+ .ifc \xmm,%xmm1
+ \opd = 1
+ .endif
+ .ifc \xmm,%xmm2
+ \opd = 2
+ .endif
+ .ifc \xmm,%xmm3
+ \opd = 3
+ .endif
+ .ifc \xmm,%xmm4
+ \opd = 4
+ .endif
+ .ifc \xmm,%xmm5
+ \opd = 5
+ .endif
+ .ifc \xmm,%xmm6
+ \opd = 6
+ .endif
+ .ifc \xmm,%xmm7
+ \opd = 7
+ .endif
+ .ifc \xmm,%xmm8
+ \opd = 8
+ .endif
+ .ifc \xmm,%xmm9
+ \opd = 9
+ .endif
+ .ifc \xmm,%xmm10
+ \opd = 10
+ .endif
+ .ifc \xmm,%xmm11
+ \opd = 11
+ .endif
+ .ifc \xmm,%xmm12
+ \opd = 12
+ .endif
+ .ifc \xmm,%xmm13
+ \opd = 13
+ .endif
+ .ifc \xmm,%xmm14
+ \opd = 14
+ .endif
+ .ifc \xmm,%xmm15
+ \opd = 15
+ .endif
+ .endm
+
+ .macro PFX_OPD_SIZE
+ .byte 0x66
+ .endm
+
+ .macro PFX_REX opd1 opd2
+ .if (\opd1 | \opd2) & 8
+ .byte 0x40 | ((\opd1 & Cool >> 3) | ((\opd2 & Cool >> 1)
+ .endif
+ .endm
+
+ .macro MODRM mod opd1 opd2
+ .byte \mod | (\opd1 & 7) | ((\opd2 & 7) << 3)
+ .endm
+
+ .macro PSHUFB_XMM xmm1 xmm2
+ XMM_NUM pshufb_opd1 \xmm1
+ XMM_NUM pshufb_opd2 \xmm2
+ PFX_OPD_SIZE
+ PFX_REX pshufb_opd1 pshufb_opd2
+ .byte 0x0f, 0x38, 0x00
+ MODRM 0xc0 pshufb_opd1 pshufb_opd2
+ .endm
+
+ .macro PCLMULQDQ imm8 xmm1 xmm2
+ XMM_NUM clmul_opd1 \xmm1
+ XMM_NUM clmul_opd2 \xmm2
+ PFX_OPD_SIZE
+ PFX_REX clmul_opd1 clmul_opd2
+ .byte 0x0f, 0x3a, 0x44
+ MODRM 0xc0 clmul_opd1 clmul_opd2
+ .byte \imm8
+ .endm
+
+ .macro AESKEYGENASSIST rcon xmm1 xmm2
+ XMM_NUM aeskeygen_opd1 \xmm1
+ XMM_NUM aeskeygen_opd2 \xmm2
+ PFX_OPD_SIZE
+ PFX_REX aeskeygen_opd1 aeskeygen_opd2
+ .byte 0x0f, 0x3a, 0xdf
+ MODRM 0xc0 aeskeygen_opd1 aeskeygen_opd2
+ .byte \rcon
+ .endm
+
+ .macro AESIMC xmm1 xmm2
+ XMM_NUM aesimc_opd1 \xmm1
+ XMM_NUM aesimc_opd2 \xmm2
+ PFX_OPD_SIZE
+ PFX_REX aesimc_opd1 aesimc_opd2
+ .byte 0x0f, 0x38, 0xdb
+ MODRM 0xc0 aesimc_opd1 aesimc_opd2
+ .endm
+
+ .macro AESENC xmm1 xmm2
+ XMM_NUM aesenc_opd1 \xmm1
+ XMM_NUM aesenc_opd2 \xmm2
+ PFX_OPD_SIZE
+ PFX_REX aesenc_opd1 aesenc_opd2
+ .byte 0x0f, 0x38, 0xdc
+ MODRM 0xc0 aesenc_opd1 aesenc_opd2
+ .endm
+
+ .macro AESENCLAST xmm1 xmm2
+ XMM_NUM aesenclast_opd1 \xmm1
+ XMM_NUM aesenclast_opd2 \xmm2
+ PFX_OPD_SIZE
+ PFX_REX aesenclast_opd1 aesenclast_opd2
+ .byte 0x0f, 0x38, 0xdd
+ MODRM 0xc0 aesenclast_opd1 aesenclast_opd2
+ .endm
+
+ .macro AESDEC xmm1 xmm2
+ XMM_NUM aesdec_opd1 \xmm1
+ XMM_NUM aesdec_opd2 \xmm2
+ PFX_OPD_SIZE
+ PFX_REX aesdec_opd1 aesdec_opd2
+ .byte 0x0f, 0x38, 0xde
+ MODRM 0xc0 aesdec_opd1 aesdec_opd2
+ .endm
+
+ .macro AESDECLAST xmm1 xmm2
+ XMM_NUM aesdeclast_opd1 \xmm1
+ XMM_NUM aesdeclast_opd2 \xmm2
+ PFX_OPD_SIZE
+ PFX_REX aesdeclast_opd1 aesdeclast_opd2
+ .byte 0x0f, 0x38, 0xdf
+ MODRM 0xc0 aesdeclast_opd1 aesdeclast_opd2
+ .endm
+#endif
+
+#endif


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


Since: Sep 15, 2009
Posts: 5



PostPosted: Thu Nov 05, 2009 1:10 pm    Post subject: Re: [PATCH 1/2] x86, crypto, Generate .byte code for some new instructions via gas macro [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)

On Thu, Nov 5, 2009 at 1:44 AM, Huang Ying <ying.huang.DeleteThis@intel.com> wrote:
> It will take some time for binutils (gas) to support some newly added
> instructions, such as SSE4.1 instructions or the AES-NI instructions
> found in upcoming Intel CPU.
>
> To make the source code can be compiled by old binutils, .byte code is
> used instead of the assembly instruction. But the readability and
> flexibility of raw .byte code is not good.
>
> This patch solves the issue of raw .byte code via generating it via
> assembly instruction like gas macro. The syntax is as close as
> possible to real assembly instruction.
>
> Some helper macros such as MODRM is not a full feature
> implementation. It can be extended when necessary.
>
> Signed-off-by: Huang Ying <ying.huang.DeleteThis@intel.com>
> ---
>  arch/x86/include/asm/inst.h |  150 ++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 150 insertions(+)
>
> --- /dev/null
> +++ b/arch/x86/include/asm/inst.h
> @@ -0,0 +1,150 @@
> +/*
> + * Generate .byte code for some instructions not supported by old
> + * binutils.
> + */
> +#ifndef X86_ASM_INST_H
> +#define X86_ASM_INST_H
> +
> +#ifdef __ASSEMBLY__
> +
> +       .macro XMM_NUM opd xmm
> +       .ifc \xmm,%xmm0
> +       \opd = 0
> +       .endif
> +       .ifc \xmm,%xmm1
> +       \opd = 1
> +       .endif
> +       .ifc \xmm,%xmm2
> +       \opd = 2
> +       .endif
> +       .ifc \xmm,%xmm3
> +       \opd = 3
> +       .endif
> +       .ifc \xmm,%xmm4
> +       \opd = 4
> +       .endif
> +       .ifc \xmm,%xmm5
> +       \opd = 5
> +       .endif
> +       .ifc \xmm,%xmm6
> +       \opd = 6
> +       .endif
> +       .ifc \xmm,%xmm7
> +       \opd = 7
> +       .endif
> +       .ifc \xmm,%xmm8
> +       \opd = 8
> +       .endif
> +       .ifc \xmm,%xmm9
> +       \opd = 9
> +       .endif
> +       .ifc \xmm,%xmm10
> +       \opd = 10
> +       .endif
> +       .ifc \xmm,%xmm11
> +       \opd = 11
> +       .endif
> +       .ifc \xmm,%xmm12
> +       \opd = 12
> +       .endif
> +       .ifc \xmm,%xmm13
> +       \opd = 13
> +       .endif
> +       .ifc \xmm,%xmm14
> +       \opd = 14
> +       .endif
> +       .ifc \xmm,%xmm15
> +       \opd = 15
> +       .endif
> +       .endm
> +
> +       .macro PFX_OPD_SIZE
> +       .byte 0x66
> +       .endm
> +
> +       .macro PFX_REX opd1 opd2
> +       .if (\opd1 | \opd2) & 8
> +       .byte 0x40 | ((\opd1 & Cool >> 3) | ((\opd2 & Cool >> 1)
> +       .endif
> +       .endm
> +
> +       .macro MODRM mod opd1 opd2
> +       .byte \mod | (\opd1 & 7) | ((\opd2 & 7) << 3)
> +       .endm
> +
> +       .macro PSHUFB_XMM xmm1 xmm2
> +       XMM_NUM pshufb_opd1 \xmm1
> +       XMM_NUM pshufb_opd2 \xmm2
> +       PFX_OPD_SIZE
> +       PFX_REX pshufb_opd1 pshufb_opd2
> +       .byte 0x0f, 0x38, 0x00
> +       MODRM 0xc0 pshufb_opd1 pshufb_opd2
> +       .endm
> +
> +       .macro PCLMULQDQ imm8 xmm1 xmm2
> +       XMM_NUM clmul_opd1 \xmm1
> +       XMM_NUM clmul_opd2 \xmm2
> +       PFX_OPD_SIZE
> +       PFX_REX clmul_opd1 clmul_opd2
> +       .byte 0x0f, 0x3a, 0x44
> +       MODRM 0xc0 clmul_opd1 clmul_opd2
> +       .byte \imm8
> +       .endm
> +
> +       .macro AESKEYGENASSIST rcon xmm1 xmm2
> +       XMM_NUM aeskeygen_opd1 \xmm1
> +       XMM_NUM aeskeygen_opd2 \xmm2
> +       PFX_OPD_SIZE
> +       PFX_REX aeskeygen_opd1 aeskeygen_opd2
> +       .byte 0x0f, 0x3a, 0xdf
> +       MODRM 0xc0 aeskeygen_opd1 aeskeygen_opd2
> +       .byte \rcon
> +       .endm
> +
> +       .macro AESIMC xmm1 xmm2
> +       XMM_NUM aesimc_opd1 \xmm1
> +       XMM_NUM aesimc_opd2 \xmm2
> +       PFX_OPD_SIZE
> +       PFX_REX aesimc_opd1 aesimc_opd2
> +       .byte 0x0f, 0x38, 0xdb
> +       MODRM 0xc0 aesimc_opd1 aesimc_opd2
> +       .endm
> +
> +       .macro AESENC xmm1 xmm2
> +       XMM_NUM aesenc_opd1 \xmm1
> +       XMM_NUM aesenc_opd2 \xmm2
> +       PFX_OPD_SIZE
> +       PFX_REX aesenc_opd1 aesenc_opd2
> +       .byte 0x0f, 0x38, 0xdc
> +       MODRM 0xc0 aesenc_opd1 aesenc_opd2
> +       .endm
> +
> +       .macro AESENCLAST xmm1 xmm2
> +       XMM_NUM aesenclast_opd1 \xmm1
> +       XMM_NUM aesenclast_opd2 \xmm2
> +       PFX_OPD_SIZE
> +       PFX_REX aesenclast_opd1 aesenclast_opd2
> +       .byte 0x0f, 0x38, 0xdd
> +       MODRM 0xc0 aesenclast_opd1 aesenclast_opd2
> +       .endm
> +
> +       .macro AESDEC xmm1 xmm2
> +       XMM_NUM aesdec_opd1 \xmm1
> +       XMM_NUM aesdec_opd2 \xmm2
> +       PFX_OPD_SIZE
> +       PFX_REX aesdec_opd1 aesdec_opd2
> +       .byte 0x0f, 0x38, 0xde
> +       MODRM 0xc0 aesdec_opd1 aesdec_opd2
> +       .endm
> +
> +       .macro AESDECLAST xmm1 xmm2
> +       XMM_NUM aesdeclast_opd1 \xmm1
> +       XMM_NUM aesdeclast_opd2 \xmm2
> +       PFX_OPD_SIZE
> +       PFX_REX aesdeclast_opd1 aesdeclast_opd2
> +       .byte 0x0f, 0x38, 0xdf
> +       MODRM 0xc0 aesdeclast_opd1 aesdeclast_opd2
> +       .endm
> +#endif
> +
> +#endif

It would be nice to document which version of GAS added support for
each instruction, so that if/when that version becomes the minimum
supported these macros can be removed.

--
Brian Gerst
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo.DeleteThis@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Back to top
H. Peter Anvin
External


Since: Feb 08, 2006
Posts: 1055



PostPosted: Thu Nov 05, 2009 1:10 pm    Post subject: Re: [PATCH 1/2] x86, crypto, Generate .byte code for some new instructions via gas macro [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)

On 11/05/2009 06:23 AM, Herbert Xu wrote:
> On Thu, Nov 05, 2009 at 02:44:15PM +0800, Huang Ying wrote:
>> It will take some time for binutils (gas) to support some newly added
>> instructions, such as SSE4.1 instructions or the AES-NI instructions
>> found in upcoming Intel CPU.
>>
>> To make the source code can be compiled by old binutils, .byte code is
>> used instead of the assembly instruction. But the readability and
>> flexibility of raw .byte code is not good.
>>
>> This patch solves the issue of raw .byte code via generating it via
>> assembly instruction like gas macro. The syntax is as close as
>> possible to real assembly instruction.
>>
>> Some helper macros such as MODRM is not a full feature
>> implementation. It can be extended when necessary.
>>
>> Signed-off-by: Huang Ying <ying.huang.RemoveThis@intel.com>
>
> Once Ingo acks this I'll throw it into cryptodev.
>

I'm not Ingo, but:

Acked-by: H. Peter Anvin <hpa.RemoveThis@zytor.com>

--
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel. I don't speak on their behalf.

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


Since: May 15, 2004
Posts: 212



PostPosted: Thu Nov 05, 2009 1:10 pm    Post subject: Re: [PATCH 1/2] x86, crypto, Generate .byte code for some new instructions via gas macro [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)

On Thu, Nov 05, 2009 at 02:44:15PM +0800, Huang Ying wrote:
> It will take some time for binutils (gas) to support some newly added
> instructions, such as SSE4.1 instructions or the AES-NI instructions
> found in upcoming Intel CPU.
>
> To make the source code can be compiled by old binutils, .byte code is
> used instead of the assembly instruction. But the readability and
> flexibility of raw .byte code is not good.
>
> This patch solves the issue of raw .byte code via generating it via
> assembly instruction like gas macro. The syntax is as close as
> possible to real assembly instruction.
>
> Some helper macros such as MODRM is not a full feature
> implementation. It can be extended when necessary.
>
> Signed-off-by: Huang Ying <ying.huang.DeleteThis@intel.com>

Once Ingo acks this I'll throw it into cryptodev.

Thanks!
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert.DeleteThis@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo.DeleteThis@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Back to top
Ingo Molnar
External


Since: May 15, 2006
Posts: 3112



PostPosted: Sun Nov 08, 2009 7:10 am    Post subject: Re: [PATCH 1/2] x86, crypto, Generate .byte code for some new instructions via gas macro [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)

* H. Peter Anvin <hpa.RemoveThis@zytor.com> wrote:

> On 11/05/2009 06:23 AM, Herbert Xu wrote:
> > On Thu, Nov 05, 2009 at 02:44:15PM +0800, Huang Ying wrote:
> >> It will take some time for binutils (gas) to support some newly added
> >> instructions, such as SSE4.1 instructions or the AES-NI instructions
> >> found in upcoming Intel CPU.
> >>
> >> To make the source code can be compiled by old binutils, .byte code is
> >> used instead of the assembly instruction. But the readability and
> >> flexibility of raw .byte code is not good.
> >>
> >> This patch solves the issue of raw .byte code via generating it via
> >> assembly instruction like gas macro. The syntax is as close as
> >> possible to real assembly instruction.
> >>
> >> Some helper macros such as MODRM is not a full feature
> >> implementation. It can be extended when necessary.
> >>
> >> Signed-off-by: Huang Ying <ying.huang.RemoveThis@intel.com>
> >
> > Once Ingo acks this I'll throw it into cryptodev.
> >
>
> I'm not Ingo, but:
>
> Acked-by: H. Peter Anvin <hpa.RemoveThis@zytor.com>

Fine to me too and for future variations thereof.

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


Since: May 15, 2004
Posts: 212



PostPosted: Mon Nov 09, 2009 3:10 pm    Post subject: Re: [PATCH 1/2] x86, crypto, Generate .byte code for some new instructions via gas macro [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)

On Sun, Nov 08, 2009 at 11:56:19AM +0100, Ingo Molnar wrote:
>
> > I'm not Ingo, but:
> >
> > Acked-by: H. Peter Anvin <hpa.TakeThisOut@zytor.com>
>
> Fine to me too and for future variations thereof.

Thanks guys I'll put this into cryptodev now.
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert.TakeThisOut@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo.TakeThisOut@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Back to top
Display posts from previous:   
Post new topic   General Reply to Topic (not reply to a specific post)    Forums Home -> Kernel All times are: Eastern Time (US & Canada) (change)
Page 1 of 1

 
You can post new topics in this forum
You can reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum