LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* [PATCH] m68k: Implement ndelay() as an inline function to force type checking/casting
@ 2018-05-13 14:02 Boris Brezillon
  2018-05-13 22:38 ` Andy Shevchenko
                   ` (2 more replies)
  0 siblings, 3 replies; 5+ messages in thread
From: Boris Brezillon @ 2018-05-13 14:02 UTC (permalink / raw)
  To: Geert Uytterhoeven, linux-m68k
  Cc: David Woodhouse, Brian Norris, Boris Brezillon, Marek Vasut,
	Richard Weinberger, linux-mtd, Miquel Raynal, linux-kernel

ndelay() is supposed to take an unsigned long, but if you define
ndelay() as a macro and the caller pass an unsigned long long instead
of an unsigned long, the unsigned long long to unsigned long cast is
not done and we end up with an "undefined reference to `__udivdi3'"
error at link time.

Fix that by making ndelay() an inline function and then defining dummy
ndelay() macro that redirects to the ndelay() function (it's how most
archs do to implement ndelay()).

Fixes: c8ee038bd148 ("m68k: Implement ndelay() based on the existing udelay() logic")
Signed-off-by: Boris Brezillon <boris.brezillon@bootlin.com>
---
Hello Geert,

This patch is fixing the bug reported by kbuild test robot here [1].
I could have patched the PSEC_TO_NSEC() macro to cast the result of
the division on a u32, but I thought making m68k consistent with what
other archs do would be preferable.

Let me know if don't like the solution, and I'll patch the ndelay()
caller instead.

Regards,

Boris

[1]https://lkml.org/lkml/2018/5/13/72
---
 arch/m68k/include/asm/delay.h | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/arch/m68k/include/asm/delay.h b/arch/m68k/include/asm/delay.h
index 7f474121e4ca..66ba159002a3 100644
--- a/arch/m68k/include/asm/delay.h
+++ b/arch/m68k/include/asm/delay.h
@@ -115,6 +115,13 @@ static inline void __udelay(unsigned long usecs)
  */
 #define	HZSCALE		(268435456 / (1000000 / HZ))
 
-#define ndelay(n) __delay(DIV_ROUND_UP((n) * ((((HZSCALE) >> 11) * (loops_per_jiffy >> 11)) >> 6), 1000))
+static inline void ndelay(unsigned long nsec)
+{
+	__delay(DIV_ROUND_UP(nsec *
+			     ((((HZSCALE) >> 11) *
+			       (loops_per_jiffy >> 11)) >> 6),
+			     1000));
+}
+#define ndelay(n) ndelay(n)
 
 #endif /* defined(_M68K_DELAY_H) */
-- 
2.14.1

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] m68k: Implement ndelay() as an inline function to force type checking/casting
  2018-05-13 14:02 [PATCH] m68k: Implement ndelay() as an inline function to force type checking/casting Boris Brezillon
@ 2018-05-13 22:38 ` Andy Shevchenko
  2018-05-14  7:23 ` Miquel Raynal
  2018-05-14  9:29 ` Geert Uytterhoeven
  2 siblings, 0 replies; 5+ messages in thread
From: Andy Shevchenko @ 2018-05-13 22:38 UTC (permalink / raw)
  To: Boris Brezillon
  Cc: Geert Uytterhoeven, linux-m68k, David Woodhouse, Brian Norris,
	Marek Vasut, Richard Weinberger, open list:MEMORY TECHNOLOGY...,
	Miquel Raynal, Linux Kernel Mailing List

On Sun, May 13, 2018 at 5:02 PM, Boris Brezillon
<boris.brezillon@bootlin.com> wrote:
> ndelay() is supposed to take an unsigned long, but if you define
> ndelay() as a macro and the caller pass an unsigned long long instead
> of an unsigned long, the unsigned long long to unsigned long cast is
> not done and we end up with an "undefined reference to `__udivdi3'"
> error at link time.

> -#define ndelay(n) __delay(DIV_ROUND_UP((n) * ((((HZSCALE) >> 11) * (loops_per_jiffy >> 11)) >> 6), 1000))
> +static inline void ndelay(unsigned long nsec)
> +{
> +       __delay(DIV_ROUND_UP(nsec *
> +                            ((((HZSCALE) >> 11) *

One pair of parens is redundant.

> +                              (loops_per_jiffy >> 11)) >> 6),
> +                            1000));

Can't you keep as one line as in original?

> +}
> +#define ndelay(n) ndelay(n)
>
>  #endif /* defined(_M68K_DELAY_H) */
> --
> 2.14.1
>



-- 
With Best Regards,
Andy Shevchenko

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] m68k: Implement ndelay() as an inline function to force type checking/casting
  2018-05-13 14:02 [PATCH] m68k: Implement ndelay() as an inline function to force type checking/casting Boris Brezillon
  2018-05-13 22:38 ` Andy Shevchenko
@ 2018-05-14  7:23 ` Miquel Raynal
  2018-05-14  9:29 ` Geert Uytterhoeven
  2 siblings, 0 replies; 5+ messages in thread
From: Miquel Raynal @ 2018-05-14  7:23 UTC (permalink / raw)
  To: Boris Brezillon
  Cc: Geert Uytterhoeven, linux-m68k, David Woodhouse, Brian Norris,
	Marek Vasut, Richard Weinberger, linux-mtd, linux-kernel

Hi Boris,

On Sun, 13 May 2018 16:02:12 +0200, Boris Brezillon
<boris.brezillon@bootlin.com> wrote:

> ndelay() is supposed to take an unsigned long, but if you define
> ndelay() as a macro and the caller pass an unsigned long long instead
> of an unsigned long, the unsigned long long to unsigned long cast is
> not done and we end up with an "undefined reference to `__udivdi3'"
> error at link time.
> 
> Fix that by making ndelay() an inline function and then defining dummy
> ndelay() macro that redirects to the ndelay() function (it's how most
> archs do to implement ndelay()).
> 
> Fixes: c8ee038bd148 ("m68k: Implement ndelay() based on the existing udelay() logic")
> Signed-off-by: Boris Brezillon <boris.brezillon@bootlin.com>

Thanks for fixing it.

Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>


-- 
Miquel Raynal, Bootlin (formerly Free Electrons)
Embedded Linux and Kernel engineering
https://bootlin.com

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] m68k: Implement ndelay() as an inline function to force type checking/casting
  2018-05-13 14:02 [PATCH] m68k: Implement ndelay() as an inline function to force type checking/casting Boris Brezillon
  2018-05-13 22:38 ` Andy Shevchenko
  2018-05-14  7:23 ` Miquel Raynal
@ 2018-05-14  9:29 ` Geert Uytterhoeven
  2018-05-14 10:52   ` Geert Uytterhoeven
  2 siblings, 1 reply; 5+ messages in thread
From: Geert Uytterhoeven @ 2018-05-14  9:29 UTC (permalink / raw)
  To: Boris Brezillon
  Cc: linux-m68k, David Woodhouse, Brian Norris, Marek Vasut,
	Richard Weinberger, MTD Maling List, Miquel Raynal,
	Linux Kernel Mailing List, Michael Schmitz

Hi Boris,

On Sun, May 13, 2018 at 4:02 PM, Boris Brezillon
<boris.brezillon@bootlin.com> wrote:
> ndelay() is supposed to take an unsigned long, but if you define
> ndelay() as a macro and the caller pass an unsigned long long instead
> of an unsigned long, the unsigned long long to unsigned long cast is
> not done and we end up with an "undefined reference to `__udivdi3'"
> error at link time.
>
> Fix that by making ndelay() an inline function and then defining dummy
> ndelay() macro that redirects to the ndelay() function (it's how most
> archs do to implement ndelay()).
>
> Fixes: c8ee038bd148 ("m68k: Implement ndelay() based on the existing udelay() logic")
> Signed-off-by: Boris Brezillon <boris.brezillon@bootlin.com>
> ---
> Hello Geert,
>
> This patch is fixing the bug reported by kbuild test robot here [1].
> I could have patched the PSEC_TO_NSEC() macro to cast the result of
> the division on a u32, but I thought making m68k consistent with what
> other archs do would be preferable.
>
> Let me know if don't like the solution, and I'll patch the ndelay()
> caller instead.

Thanks for your patch!

With the comment a few lines above removed:

    - * This is a macro so that the const version can factor out the first
    - * multiply and shift.

Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>

>
> Regards,
>
> Boris
>
> [1]https://lkml.org/lkml/2018/5/13/72
> ---
>  arch/m68k/include/asm/delay.h | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
>
> diff --git a/arch/m68k/include/asm/delay.h b/arch/m68k/include/asm/delay.h
> index 7f474121e4ca..66ba159002a3 100644
> --- a/arch/m68k/include/asm/delay.h
> +++ b/arch/m68k/include/asm/delay.h
> @@ -115,6 +115,13 @@ static inline void __udelay(unsigned long usecs)
>   */
>  #define        HZSCALE         (268435456 / (1000000 / HZ))
>
> -#define ndelay(n) __delay(DIV_ROUND_UP((n) * ((((HZSCALE) >> 11) * (loops_per_jiffy >> 11)) >> 6), 1000))
> +static inline void ndelay(unsigned long nsec)
> +{
> +       __delay(DIV_ROUND_UP(nsec *
> +                            ((((HZSCALE) >> 11) *
> +                              (loops_per_jiffy >> 11)) >> 6),
> +                            1000));
> +}
> +#define ndelay(n) ndelay(n)
>
>  #endif /* defined(_M68K_DELAY_H) */

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] m68k: Implement ndelay() as an inline function to force type checking/casting
  2018-05-14  9:29 ` Geert Uytterhoeven
@ 2018-05-14 10:52   ` Geert Uytterhoeven
  0 siblings, 0 replies; 5+ messages in thread
From: Geert Uytterhoeven @ 2018-05-14 10:52 UTC (permalink / raw)
  To: Boris Brezillon
  Cc: linux-m68k, David Woodhouse, Brian Norris, Marek Vasut,
	Richard Weinberger, MTD Maling List, Miquel Raynal,
	Linux Kernel Mailing List, Michael Schmitz

Hi Boris,

On Mon, May 14, 2018 at 11:29 AM, Geert Uytterhoeven
<geert@linux-m68k.org> wrote:
> On Sun, May 13, 2018 at 4:02 PM, Boris Brezillon
> <boris.brezillon@bootlin.com> wrote:
>> ndelay() is supposed to take an unsigned long, but if you define
>> ndelay() as a macro and the caller pass an unsigned long long instead
>> of an unsigned long, the unsigned long long to unsigned long cast is
>> not done and we end up with an "undefined reference to `__udivdi3'"
>> error at link time.
>>
>> Fix that by making ndelay() an inline function and then defining dummy
>> ndelay() macro that redirects to the ndelay() function (it's how most
>> archs do to implement ndelay()).
>>
>> Fixes: c8ee038bd148 ("m68k: Implement ndelay() based on the existing udelay() logic")
>> Signed-off-by: Boris Brezillon <boris.brezillon@bootlin.com>
>> ---
>> Hello Geert,
>>
>> This patch is fixing the bug reported by kbuild test robot here [1].
>> I could have patched the PSEC_TO_NSEC() macro to cast the result of
>> the division on a u32, but I thought making m68k consistent with what
>> other archs do would be preferable.
>>
>> Let me know if don't like the solution, and I'll patch the ndelay()
>> caller instead.
>
> Thanks for your patch!
>
> With the comment a few lines above removed:
>
>     - * This is a macro so that the const version can factor out the first
>     - * multiply and shift.
>
> Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>

I've just posted a fix for the root cause ("[PATCH] mtd: nand: Fix return
type of __DIVIDE() when called with 32-bit",
https://lkml.org/lkml/2018/5/14/578), so the fix for m68k's ndelay() is not
that urgent anymore.

As your patch is still a valid fix, I'll apply it, and queue it up for v4.18.

Thanks!

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2018-05-14 10:52 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-05-13 14:02 [PATCH] m68k: Implement ndelay() as an inline function to force type checking/casting Boris Brezillon
2018-05-13 22:38 ` Andy Shevchenko
2018-05-14  7:23 ` Miquel Raynal
2018-05-14  9:29 ` Geert Uytterhoeven
2018-05-14 10:52   ` Geert Uytterhoeven

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).