[PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[Haijun Zhang]

Add this file to help detect cpu type in runtime.
These macros will be more favorable for driver
to apply errata and workaround to specified cpu type.

Signed-off-by: Haijun Zhang <Haijun.Zhang@freescale.com>
Signed-off-by: Zhao Chenhui <chenhui.zhao@freescale.com>
---
changes v2:
	- Remove inline function
	- Ignore E bit of SOC
	- Add Macro to get current soc version

 arch/powerpc/include/asm/mpc85xx.h | 53 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 53 insertions(+)
 create mode 100644 arch/powerpc/include/asm/mpc85xx.h

diff --git a/arch/powerpc/include/asm/mpc85xx.h b/arch/powerpc/include/asm/mpc85xx.h
new file mode 100644
index 0000000..20af5e8
--- /dev/null
+++ b/arch/powerpc/include/asm/mpc85xx.h
@@ -0,0 +1,53 @@
+/*
+ * MPC85xx cpu type detection
+ *
+ * Copyright 2011-2012 Freescale Semiconductor, Inc.
+ *
+ * This is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __ASM_PPC_MPC85XX_H
+#define __ASM_PPC_MPC85XX_H
+
+#define SVR_REV(svr)	((svr) & 0xFF)		/* SOC design resision */
+#define SVR_MAJ(svr)	(((svr) >>  4) & 0xF)	/* Major revision field*/
+#define SVR_MIN(svr)	(((svr) >>  0) & 0xF)	/* Minor revision field*/
+
+/* Some parts define SVR[0:23] as the SOC version */
+#define SVR_SOC_VER(svr) (((svr) >> 8) & 0xFFF7FF)	/* SOC Version fields */
+
+/* Get current SOC Version */
+#define GET_SVR()	(mfspr(SPRN_SVR))
+
+#define IS_SVR_REV(svr, maj, min) \
+	((SVR_MAJ(svr) == (maj)) && (SVR_MIN(svr) == (min)))
+
+#define SVR_8533	0x803400
+#define SVR_8535	0x803701
+#define SVR_8536	0x803700
+#define SVR_8540	0x803000
+#define SVR_8541	0x807200
+#define SVR_8543	0x803200
+#define SVR_8544	0x803401
+#define SVR_8545	0x803102
+#define SVR_8547	0x803101
+#define SVR_8548	0x803100
+#define SVR_8555	0x807100
+#define SVR_8560	0x807000
+#define SVR_8567	0x807501
+#define SVR_8568	0x807500
+#define SVR_8569	0x808000
+#define SVR_8572	0x80E000
+#define SVR_P1010	0x80f900
+#define SVR_P2041	0x821001
+#define SVR_P3041	0x821103
+#define SVR_P5010	0x822100
+#define SVR_P5020	0x822000
+#define SVR_P5040	0x820400
+#define SVR_T4240	0x824800
+#define SVR_B4860	0x868800
+
+#endif
-- 
1.8.0

[PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[Haijun Zhang]

Add this file to help detect cpu type in runtime.
These macros will be more favorable for driver
to apply errata and workaround to specified cpu type.

Signed-off-by: Haijun Zhang <Haijun.Zhang@freescale.com>
Signed-off-by: Zhao Chenhui <chenhui.zhao@freescale.com>
---
changes v2:
	- Remove inline function
	- Ignore E bit of SOC
	- Add Macro to get current soc version

 arch/powerpc/include/asm/mpc85xx.h | 53 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 53 insertions(+)
 create mode 100644 arch/powerpc/include/asm/mpc85xx.h

diff --git a/arch/powerpc/include/asm/mpc85xx.h b/arch/powerpc/include/asm/mpc85xx.h
new file mode 100644
index 0000000..20af5e8
--- /dev/null
+++ b/arch/powerpc/include/asm/mpc85xx.h
@@ -0,0 +1,53 @@
+/*
+ * MPC85xx cpu type detection
+ *
+ * Copyright 2011-2012 Freescale Semiconductor, Inc.
+ *
+ * This is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __ASM_PPC_MPC85XX_H
+#define __ASM_PPC_MPC85XX_H
+
+#define SVR_REV(svr)	((svr) & 0xFF)		/* SOC design resision */
+#define SVR_MAJ(svr)	(((svr) >>  4) & 0xF)	/* Major revision field*/
+#define SVR_MIN(svr)	(((svr) >>  0) & 0xF)	/* Minor revision field*/
+
+/* Some parts define SVR[0:23] as the SOC version */
+#define SVR_SOC_VER(svr) (((svr) >> 8) & 0xFFF7FF)	/* SOC Version fields */
+
+/* Get current SOC Version */
+#define GET_SVR()	(mfspr(SPRN_SVR))
+
+#define IS_SVR_REV(svr, maj, min) \
+	((SVR_MAJ(svr) == (maj)) && (SVR_MIN(svr) == (min)))
+
+#define SVR_8533	0x803400
+#define SVR_8535	0x803701
+#define SVR_8536	0x803700
+#define SVR_8540	0x803000
+#define SVR_8541	0x807200
+#define SVR_8543	0x803200
+#define SVR_8544	0x803401
+#define SVR_8545	0x803102
+#define SVR_8547	0x803101
+#define SVR_8548	0x803100
+#define SVR_8555	0x807100
+#define SVR_8560	0x807000
+#define SVR_8567	0x807501
+#define SVR_8568	0x807500
+#define SVR_8569	0x808000
+#define SVR_8572	0x80E000
+#define SVR_P1010	0x80f900
+#define SVR_P2041	0x821001
+#define SVR_P3041	0x821103
+#define SVR_P5010	0x822100
+#define SVR_P5020	0x822000
+#define SVR_P5040	0x820400
+#define SVR_T4240	0x824800
+#define SVR_B4860	0x868800
+
+#endif
-- 
1.8.0

Re: [PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[Scott Wood]

On 07/10/2013 02:39:01 AM, Haijun Zhang wrote:
> +/* Get current SOC Version */
> +#define GET_SVR()	(mfspr(SPRN_SVR))

Unnecessary parens.  Why do we need GET_SVR(), versus opencoding it?   
Note that U-Boot (which this is patterned after) doesn't have  
GET_SVR(), so code that wants to work on both can't use it anyway.

> +#define IS_SVR_REV(svr, maj, min) \
> +	((SVR_MAJ(svr) == (maj)) && (SVR_MIN(svr) == (min)))

I don't think IS_SVR_REV is needed.  Callers can just do "if  
(SVR_REV(svr) == 0x30)" or whatever, especially since we're relying on  
them to do this for greater/less than comparisons.

-Scott

Re: [PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[Scott Wood]

On 07/10/2013 02:39:01 AM, Haijun Zhang wrote:
> +/* Get current SOC Version */
> +#define GET_SVR()	(mfspr(SPRN_SVR))

Unnecessary parens.  Why do we need GET_SVR(), versus opencoding it?  =20
Note that U-Boot (which this is patterned after) doesn't have =20
GET_SVR(), so code that wants to work on both can't use it anyway.

> +#define IS_SVR_REV(svr, maj, min) \
> +	((SVR_MAJ(svr) =3D=3D (maj)) && (SVR_MIN(svr) =3D=3D (min)))

I don't think IS_SVR_REV is needed.  Callers can just do "if =20
(SVR_REV(svr) =3D=3D 0x30)" or whatever, especially since we're relying on =
=20
them to do this for greater/less than comparisons.

-Scott=

RE: [PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[David Laight]

> > +#define IS_SVR_REV(svr, maj, min) \
> > +	((SVR_MAJ(svr) == (maj)) && (SVR_MIN(svr) == (min)))
> 
> I don't think IS_SVR_REV is needed.  Callers can just do "if
> (SVR_REV(svr) == 0x30)" or whatever, especially since we're relying on
> them to do this for greater/less than comparisons.

Not only that, I'd guess that 'maj' and 'min' are likely to be
constants - so you'd want to combine them and compare against 'svr'
rather than have two conditionals.

	David

RE: [PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[David Laight]

> > +#define IS_SVR_REV(svr, maj, min) \
> > +	((SVR_MAJ(svr) =3D=3D (maj)) && (SVR_MIN(svr) =3D=3D (min)))
>=20
> I don't think IS_SVR_REV is needed.  Callers can just do "if
> (SVR_REV(svr) =3D=3D 0x30)" or whatever, especially since we're =
relying on
> them to do this for greater/less than comparisons.

Not only that, I'd guess that 'maj' and 'min' are likely to be
constants - so you'd want to combine them and compare against 'svr'
rather than have two conditionals.

	David

RE: [PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[Zhang Haijun-B42677]

Thanks.

Regards
Haijun.


> -----Original Message-----
> From: David Laight [mailto:David.Laight@ACULAB.COM]
> Sent: Thursday, July 11, 2013 4:44 PM
> To: Wood Scott-B07421; Zhang Haijun-B42677
> Cc: Zhao Chenhui-B35336; linux-mmc@vger.kernel.org; Wrobel Heinz-R39252;
> Fleming Andy-AFLEMING; Zhang Haijun-B42677; cbouatmailru@gmail.com;
> cjb@laptop.org; linuxppc-dev@lists.ozlabs.org; Xie Xiaobo-R63061
> Subject: RE: [PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type
> detection
> 
> > > +#define IS_SVR_REV(svr, maj, min) \
> > > +	((SVR_MAJ(svr) == (maj)) && (SVR_MIN(svr) == (min)))
> >
> > I don't think IS_SVR_REV is needed.  Callers can just do "if
> > (SVR_REV(svr) == 0x30)" or whatever, especially since we're relying on
> > them to do this for greater/less than comparisons.
> 
> Not only that, I'd guess that 'maj' and 'min' are likely to be constants
> - so you'd want to combine them and compare against 'svr'
> rather than have two conditionals.
[Haijun Wrote:] yes, e.g: 1.0, 1.1 or 2.0. 'Major revision field' + '.' + 'Minor revision field '
I had resent the patch(V2). I'm not so clear about what you want to express.^_^  
> 
> 	David
> 
> 
> 

RE: [PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[Zhang Haijun-B42677]

Thanks.

Regards
Haijun.


> -----Original Message-----
> From: David Laight [mailto:David.Laight@ACULAB.COM]
> Sent: Thursday, July 11, 2013 4:44 PM
> To: Wood Scott-B07421; Zhang Haijun-B42677
> Cc: Zhao Chenhui-B35336; linux-mmc@vger.kernel.org; Wrobel Heinz-R39252;
> Fleming Andy-AFLEMING; Zhang Haijun-B42677; cbouatmailru@gmail.com;
> cjb@laptop.org; linuxppc-dev@lists.ozlabs.org; Xie Xiaobo-R63061
> Subject: RE: [PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type
> detection
>=20
> > > +#define IS_SVR_REV(svr, maj, min) \
> > > +	((SVR_MAJ(svr) =3D=3D (maj)) && (SVR_MIN(svr) =3D=3D (min)))
> >
> > I don't think IS_SVR_REV is needed.  Callers can just do "if
> > (SVR_REV(svr) =3D=3D 0x30)" or whatever, especially since we're relying=
 on
> > them to do this for greater/less than comparisons.
>=20
> Not only that, I'd guess that 'maj' and 'min' are likely to be constants
> - so you'd want to combine them and compare against 'svr'
> rather than have two conditionals.
[Haijun Wrote:] yes, e.g: 1.0, 1.1 or 2.0. 'Major revision field' + '.' + '=
Minor revision field '
I had resent the patch(V2). I'm not so clear about what you want to express=
.^_^ =20
>=20
> 	David
>=20
>=20
>=20

Re: [PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[Scott Wood]

On 07/11/2013 03:43:35 AM, David Laight wrote:
> > > +#define IS_SVR_REV(svr, maj, min) \
> > > +	((SVR_MAJ(svr) == (maj)) && (SVR_MIN(svr) == (min)))
> >
> > I don't think IS_SVR_REV is needed.  Callers can just do "if
> > (SVR_REV(svr) == 0x30)" or whatever, especially since we're relying  
> on
> > them to do this for greater/less than comparisons.
> 
> Not only that, I'd guess that 'maj' and 'min' are likely to be
> constants - so you'd want to combine them and compare against 'svr'
> rather than have two conditionals.

Yes, that's what I meant -- 0x30 would represent revision 3.0.

-Scott

Re: [PATCH 1/4 V2] powerpc/85xx: Add support for 85xx cpu type detection

[Scott Wood]

On 07/11/2013 03:43:35 AM, David Laight wrote:
> > > +#define IS_SVR_REV(svr, maj, min) \
> > > +	((SVR_MAJ(svr) =3D=3D (maj)) && (SVR_MIN(svr) =3D=3D (min)))
> >
> > I don't think IS_SVR_REV is needed.  Callers can just do "if
> > (SVR_REV(svr) =3D=3D 0x30)" or whatever, especially since we're relying=
 =20
> on
> > them to do this for greater/less than comparisons.
>=20
> Not only that, I'd guess that 'maj' and 'min' are likely to be
> constants - so you'd want to combine them and compare against 'svr'
> rather than have two conditionals.

Yes, that's what I meant -- 0x30 would represent revision 3.0.

-Scott=