Patchwork [U-Boot,v4] ARM: Avoid compiler optimization for readb, writeb and friends.

login
register
mail settings
Submitter Wolfgang Denk
Date Jan. 9, 2011, 10:19 p.m.
Message ID <1294611584-6098-1-git-send-email-wd@denx.de>
Download mbox | patch
Permalink /patch/78056/
State Accepted
Commit 3c0659b535b075be124c3d2a0714e55e65c46737
Delegated to: Albert ARIBAUD
Headers show

Comments

Wolfgang Denk - Jan. 9, 2011, 10:19 p.m.
From: Alexander Holler <holler@ahsoftware.de>

gcc 4.5.1 seems to ignore (at least some) volatile definitions,
avoid that as done in the kernel.

Reading C99 6.7.3 8 and the comment 114) there, I think it is a bug of that
gcc version to ignore the volatile type qualifier used e.g. in __arch_getl().
Anyway, using a definition as in the kernel headers avoids such optimizations when
gcc 4.5.1 is used.

Maybe the headers as used in the current linux-kernel should be used,
but to avoid large changes, I've just added a small change to the current headers.

Signed-off-by: Alexander Holler <holler@ahsoftware.de>
Signed-off-by: Dirk Behme <dirk.behme@googlemail.com>
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Alessandro Rubini <rubini-list@gnudd.com>
---
 arch/arm/include/asm/io.h |   32 ++++++++++++++++++++------------
 1 files changed, 20 insertions(+), 12 deletions(-)
Thomas Weber - Jan. 12, 2011, 3:17 p.m.
Am 09.01.2011 23:19, schrieb Wolfgang Denk:
> From: Alexander Holler <holler@ahsoftware.de>
> 
> gcc 4.5.1 seems to ignore (at least some) volatile definitions,
> avoid that as done in the kernel.
> 
> Reading C99 6.7.3 8 and the comment 114) there, I think it is a bug of that
> gcc version to ignore the volatile type qualifier used e.g. in __arch_getl().
> Anyway, using a definition as in the kernel headers avoids such optimizations when
> gcc 4.5.1 is used.
> 
> Maybe the headers as used in the current linux-kernel should be used,
> but to avoid large changes, I've just added a small change to the current headers.
> 
> Signed-off-by: Alexander Holler <holler@ahsoftware.de>
> Signed-off-by: Dirk Behme <dirk.behme@googlemail.com>
> Signed-off-by: Wolfgang Denk <wd@denx.de>
> Cc: Alessandro Rubini <rubini-list@gnudd.com>
> ---
>  arch/arm/include/asm/io.h |   32 ++++++++++++++++++++------------
>  1 files changed, 20 insertions(+), 12 deletions(-)
> 
> diff --git a/arch/arm/include/asm/io.h b/arch/arm/include/asm/io.h
> index ff1518e..3886f15 100644
> --- a/arch/arm/include/asm/io.h
> +++ b/arch/arm/include/asm/io.h
> @@ -117,21 +117,29 @@ extern inline void __raw_readsl(unsigned int addr, void *data, int longlen)
>  		*buf++ = __arch_getl(addr);
>  }
>  
> -#define __raw_writeb(v,a)		__arch_putb(v,a)
> -#define __raw_writew(v,a)		__arch_putw(v,a)
> -#define __raw_writel(v,a)		__arch_putl(v,a)
> +#define __raw_writeb(v,a)	__arch_putb(v,a)
> +#define __raw_writew(v,a)	__arch_putw(v,a)
> +#define __raw_writel(v,a)	__arch_putl(v,a)
>  
> -#define __raw_readb(a)			__arch_getb(a)
> -#define __raw_readw(a)			__arch_getw(a)
> -#define __raw_readl(a)			__arch_getl(a)
> +#define __raw_readb(a)		__arch_getb(a)
> +#define __raw_readw(a)		__arch_getw(a)
> +#define __raw_readl(a)		__arch_getl(a)
>  
> -#define writeb(v,a)			__arch_putb(v,a)
> -#define writew(v,a)			__arch_putw(v,a)
> -#define writel(v,a)			__arch_putl(v,a)
> +/*
> + * TODO: The kernel offers some more advanced versions of barriers, it might
> + * have some advantages to use them instead of the simple one here.
> + */
> +#define dmb()		__asm__ __volatile__ ("" : : : "memory")
> +#define __iormb()	dmb()
> +#define __iowmb()	dmb()
> +
> +#define writeb(v,c)	({ __iowmb(); __arch_putb(v,c); v; })
> +#define writew(v,c)	({ __iowmb(); __arch_putw(v,c); v; })
> +#define writel(v,c)	({ __iowmb(); __arch_putl(v,c); v; })
>  
> -#define readb(a)			__arch_getb(a)
> -#define readw(a)			__arch_getw(a)
> -#define readl(a)			__arch_getl(a)
> +#define readb(c)	({ u8  __v = __arch_getb(c); __iormb(); __v; })
> +#define readw(c)	({ u16 __v = __arch_getw(c); __iormb(); __v; })
> +#define readl(c)	({ u32 __v = __arch_getl(c); __iormb(); __v; })
>  
>  /*
>   * The compiler seems to be incapable of optimising constants

Tested-by: Thomas Weber <weber@corscience.de>

on Devkit8000 with codesourcery arm2010.09 (gcc4.5.1) and arm2010q1 (gcc
4.4.1)
Alexander Holler - Jan. 12, 2011, 3:39 p.m.
Am 09.01.2011 23:19, schrieb Wolfgang Denk:

> gcc 4.5.1 seems to ignore (at least some) volatile definitions,
> avoid that as done in the kernel.

Have had a look at the asm generated by gcc 4.5.1, looks good.

The wrong optimization in arch/arm/cpu/armv7/omap3/clock.c is gone and 
the writeb in drivers/mtd/nand/omap_gpmc.c doesn't have the problem as 
the v1-patch.

Reagards,

Alexander
Wolfgang Denk - Jan. 12, 2011, 4:40 p.m.
Dear Alexander Holler,

In message <4D2DCB18.20409@ahsoftware.de> you wrote:
> Am 09.01.2011 23:19, schrieb Wolfgang Denk:
> 
> > gcc 4.5.1 seems to ignore (at least some) volatile definitions,
> > avoid that as done in the kernel.
> 
> Have had a look at the asm generated by gcc 4.5.1, looks good.
> 
> The wrong optimization in arch/arm/cpu/armv7/omap3/clock.c is gone and 
> the writeb in drivers/mtd/nand/omap_gpmc.c doesn't have the problem as 
> the v1-patch.

Thanks - but please send a formal Acked-by: and/or Tested-by: .

Best regards,

Wolfgang Denk
Alexander Holler - Jan. 12, 2011, 4:49 p.m.
Am 12.01.2011 17:40, schrieb Wolfgang Denk:
> Dear Alexander Holler,
>
> In message<4D2DCB18.20409@ahsoftware.de>  you wrote:
>> Am 09.01.2011 23:19, schrieb Wolfgang Denk:
>>
>>> gcc 4.5.1 seems to ignore (at least some) volatile definitions,
>>> avoid that as done in the kernel.
>>
>> Have had a look at the asm generated by gcc 4.5.1, looks good.
>>
>> The wrong optimization in arch/arm/cpu/armv7/omap3/clock.c is gone and
>> the writeb in drivers/mtd/nand/omap_gpmc.c doesn't have the problem as
>> the v1-patch.
>
> Thanks - but please send a formal Acked-by: and/or Tested-by: .

Oh, as I'm still listed as the author, I thought that isn't necessary.

I don't know if I should paste the whole patch (this is my first ack ;) 
), but here are both:

Acked-by: Alexander Holler <holler@ahsoftware.de>
Tested-by: Alexander Holler <holler@ahsoftware.de>

Regards,

Alexander
Albert ARIBAUD - Jan. 15, 2011, 1:13 p.m.
Le 12/01/2011 17:49, Alexander Holler a écrit :

>>>> Signed-off-by: Alexander Holler <holler@ahsoftware.de>
>>>> Signed-off-by: Dirk Behme <dirk.behme@googlemail.com>
>>>> Signed-off-by: Wolfgang Denk <wd@denx.de>

>>> Tested-by: Thomas Weber <weber@corscience.de>

> Acked-by: Alexander Holler<holler@ahsoftware.de>
> Tested-by: Alexander Holler<holler@ahsoftware.de>

Applied to u-boot-arm, thanks.

Amicalement,

Patch

diff --git a/arch/arm/include/asm/io.h b/arch/arm/include/asm/io.h
index ff1518e..3886f15 100644
--- a/arch/arm/include/asm/io.h
+++ b/arch/arm/include/asm/io.h
@@ -117,21 +117,29 @@  extern inline void __raw_readsl(unsigned int addr, void *data, int longlen)
 		*buf++ = __arch_getl(addr);
 }
 
-#define __raw_writeb(v,a)		__arch_putb(v,a)
-#define __raw_writew(v,a)		__arch_putw(v,a)
-#define __raw_writel(v,a)		__arch_putl(v,a)
+#define __raw_writeb(v,a)	__arch_putb(v,a)
+#define __raw_writew(v,a)	__arch_putw(v,a)
+#define __raw_writel(v,a)	__arch_putl(v,a)
 
-#define __raw_readb(a)			__arch_getb(a)
-#define __raw_readw(a)			__arch_getw(a)
-#define __raw_readl(a)			__arch_getl(a)
+#define __raw_readb(a)		__arch_getb(a)
+#define __raw_readw(a)		__arch_getw(a)
+#define __raw_readl(a)		__arch_getl(a)
 
-#define writeb(v,a)			__arch_putb(v,a)
-#define writew(v,a)			__arch_putw(v,a)
-#define writel(v,a)			__arch_putl(v,a)
+/*
+ * TODO: The kernel offers some more advanced versions of barriers, it might
+ * have some advantages to use them instead of the simple one here.
+ */
+#define dmb()		__asm__ __volatile__ ("" : : : "memory")
+#define __iormb()	dmb()
+#define __iowmb()	dmb()
+
+#define writeb(v,c)	({ __iowmb(); __arch_putb(v,c); v; })
+#define writew(v,c)	({ __iowmb(); __arch_putw(v,c); v; })
+#define writel(v,c)	({ __iowmb(); __arch_putl(v,c); v; })
 
-#define readb(a)			__arch_getb(a)
-#define readw(a)			__arch_getw(a)
-#define readl(a)			__arch_getl(a)
+#define readb(c)	({ u8  __v = __arch_getb(c); __iormb(); __v; })
+#define readw(c)	({ u16 __v = __arch_getw(c); __iormb(); __v; })
+#define readl(c)	({ u32 __v = __arch_getl(c); __iormb(); __v; })
 
 /*
  * The compiler seems to be incapable of optimising constants