2018-11-12 15:56:44

by Changbin Du

[permalink] [raw]
Subject: [PATCH] openrisc: make function cache_loop() inline

The third operand of mtspr instruction must be a constraint. To guarantee
this condition, function cache_loop() which uses macro mtspr() must be
inlined. So let's force it as 'inline'. This is to fix compiling error with
new option CONFIG_NO_AUTO_INLINE.

In file included from arch/openrisc/mm/cache.c:17:0:
arch/openrisc/mm/cache.c: In function 'cache_loop':
arch/openrisc/include/asm/spr.h:20:27: warning: asm operand 0 probably doesn't match constraints
^
arch/openrisc/mm/cache.c:29:3: note: in expansion of macro 'mtspr'
mtspr(reg, line);
^~~~~
arch/openrisc/include/asm/spr.h:20:27: error: impossible constraint in 'asm'

Signed-off-by: Changbin Du <[email protected]>
Reported-by: kbuild test robot <[email protected]>
Cc: Stafford Horne <[email protected]>
Cc: Masahiro Yamada <[email protected]>
---
arch/openrisc/mm/cache.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/openrisc/mm/cache.c b/arch/openrisc/mm/cache.c
index b747bf1fc1b6..4a4b2b6e006b 100644
--- a/arch/openrisc/mm/cache.c
+++ b/arch/openrisc/mm/cache.c
@@ -20,7 +20,7 @@
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>

-static void cache_loop(struct page *page, const unsigned int reg)
+static inline void cache_loop(struct page *page, const unsigned int reg)
{
unsigned long paddr = page_to_pfn(page) << PAGE_SHIFT;
unsigned long line = paddr & ~(L1_CACHE_BYTES - 1);
--
2.17.1



2018-11-13 20:50:46

by Stafford Horne

[permalink] [raw]
Subject: Re: [PATCH] openrisc: make function cache_loop() inline

On Mon, Nov 12, 2018 at 11:55:44PM +0800, Changbin Du wrote:
> The third operand of mtspr instruction must be a constraint. To guarantee
> this condition, function cache_loop() which uses macro mtspr() must be
> inlined. So let's force it as 'inline'. This is to fix compiling error with
> new option CONFIG_NO_AUTO_INLINE.

Thanks for looking at this.

Do you mean, must be a 'constant'?

I think this fix is fine for now. Can you add it to your 'kernel hacking: GCC
optimization for better debug experience' series?

> In file included from arch/openrisc/mm/cache.c:17:0:
> arch/openrisc/mm/cache.c: In function 'cache_loop':
> arch/openrisc/include/asm/spr.h:20:27: warning: asm operand 0 probably doesn't match constraints
> ^
> arch/openrisc/mm/cache.c:29:3: note: in expansion of macro 'mtspr'
> mtspr(reg, line);
> ^~~~~
> arch/openrisc/include/asm/spr.h:20:27: error: impossible constraint in 'asm'
>
> Signed-off-by: Changbin Du <[email protected]>
> Reported-by: kbuild test robot <[email protected]>
> Cc: Stafford Horne <[email protected]>
> Cc: Masahiro Yamada <[email protected]>

Acked-by: Stafford Horne <[email protected]>

> ---
> arch/openrisc/mm/cache.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/openrisc/mm/cache.c b/arch/openrisc/mm/cache.c
> index b747bf1fc1b6..4a4b2b6e006b 100644
> --- a/arch/openrisc/mm/cache.c
> +++ b/arch/openrisc/mm/cache.c
> @@ -20,7 +20,7 @@
> #include <asm/cacheflush.h>
> #include <asm/tlbflush.h>
>
> -static void cache_loop(struct page *page, const unsigned int reg)
> +static inline void cache_loop(struct page *page, const unsigned int reg)
> {
> unsigned long paddr = page_to_pfn(page) << PAGE_SHIFT;
> unsigned long line = paddr & ~(L1_CACHE_BYTES - 1);
> --
> 2.17.1
>

2018-11-14 02:20:51

by Changbin Du

[permalink] [raw]
Subject: Re: [PATCH] openrisc: make function cache_loop() inline

On Wed, Nov 14, 2018 at 05:49:53AM +0900, Stafford Horne wrote:
> On Mon, Nov 12, 2018 at 11:55:44PM +0800, Changbin Du wrote:
> > The third operand of mtspr instruction must be a constraint. To guarantee
> > this condition, function cache_loop() which uses macro mtspr() must be
> > inlined. So let's force it as 'inline'. This is to fix compiling error with
> > new option CONFIG_NO_AUTO_INLINE.
>
> Thanks for looking at this.
>
> Do you mean, must be a 'constant'?
>
yes, it is 'constant'. sorry for the misstake.
> I think this fix is fine for now. Can you add it to your 'kernel hacking: GCC
> optimization for better debug experience' series?
>
hi Yamada, could you add this patch to your kbuild tree? thanks!

> > In file included from arch/openrisc/mm/cache.c:17:0:
> > arch/openrisc/mm/cache.c: In function 'cache_loop':
> > arch/openrisc/include/asm/spr.h:20:27: warning: asm operand 0 probably doesn't match constraints
> > ^
> > arch/openrisc/mm/cache.c:29:3: note: in expansion of macro 'mtspr'
> > mtspr(reg, line);
> > ^~~~~
> > arch/openrisc/include/asm/spr.h:20:27: error: impossible constraint in 'asm'
> >
> > Signed-off-by: Changbin Du <[email protected]>
> > Reported-by: kbuild test robot <[email protected]>
> > Cc: Stafford Horne <[email protected]>
> > Cc: Masahiro Yamada <[email protected]>
>
> Acked-by: Stafford Horne <[email protected]>
>
> > ---
> > arch/openrisc/mm/cache.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/arch/openrisc/mm/cache.c b/arch/openrisc/mm/cache.c
> > index b747bf1fc1b6..4a4b2b6e006b 100644
> > --- a/arch/openrisc/mm/cache.c
> > +++ b/arch/openrisc/mm/cache.c
> > @@ -20,7 +20,7 @@
> > #include <asm/cacheflush.h>
> > #include <asm/tlbflush.h>
> >
> > -static void cache_loop(struct page *page, const unsigned int reg)
> > +static inline void cache_loop(struct page *page, const unsigned int reg)
> > {
> > unsigned long paddr = page_to_pfn(page) << PAGE_SHIFT;
> > unsigned long line = paddr & ~(L1_CACHE_BYTES - 1);
> > --
> > 2.17.1
> >

--
Thanks,
Changbin Du