Re: MIPS: Outline udelay and fix a few issues.

Re: MIPS: Outline udelay and fix a few issues.

[Atsushi Nemoto]

On Mon, 08 Jun 2009 16:50:51 +0100, linux-mips@linux-mips.org wrote:
> Outlining fixes the issue were on certain CPUs such as the R10000 family
> the delay loop would need an extra cycle if it overlaps a cacheline
> boundary.
> 
> The rewrite also fixes build errors with GCC 4.4 which was changed in
> way incompatible with the kernel's inline assembly.
> 
> Relying on pure C for computation of the delay value removes the need for
> explicit.  The price we pay is a slight slowdown of the computation - to
> be fixed on another day.

Please fix this commit.

------------------------------------------------------
From: Atsushi Nemoto <nemoto@toshiba-tops.co.jp>
Subject: [PATCH] fix __ndelay build error and add 'ull' suffix for 32-bit kernel

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
---
 arch/mips/lib/delay.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/mips/lib/delay.c b/arch/mips/lib/delay.c
index f69c6b5..6b3b1de 100644
--- a/arch/mips/lib/delay.c
+++ b/arch/mips/lib/delay.c
@@ -43,7 +43,7 @@ void __udelay(unsigned long us)
 {
 	unsigned int lpj = current_cpu_data.udelay_val;
 
-	__delay((us * 0x000010c7 * HZ * lpj) >> 32);
+	__delay((us * 0x000010c7ull * HZ * lpj) >> 32);
 }
 EXPORT_SYMBOL(__udelay);
 
@@ -51,6 +51,6 @@ void __ndelay(unsigned long ns)
 {
 	unsigned int lpj = current_cpu_data.udelay_val;
 
-	__delay((us * 0x00000005 * HZ * lpj) >> 32);
+	__delay((ns * 0x00000005ull * HZ * lpj) >> 32);
 }
 EXPORT_SYMBOL(__ndelay);


Also, a block comment in delay.c should be updated...

---
Atsushi Nemoto

Re: MIPS: Outline udelay and fix a few issues.

[Florian Fainelli]

Le Tuesday 09 June 2009 04:12:48 Atsushi Nemoto, vous avez écrit :
> Subject: [PATCH] fix __ndelay build error and add 'ull' suffix for 32-bit
> kernel

Fixes the build error for me. Thanks !
-- 
Best regards, Florian Fainelli
Email : florian@openwrt.org
http://openwrt.org
-------------------------------

Re: MIPS: Outline udelay and fix a few issues.

[Ralf Baechle]

On Tue, Jun 09, 2009 at 11:12:48AM +0900, Atsushi Nemoto wrote:

> > Relying on pure C for computation of the delay value removes the need for
> > explicit.  The price we pay is a slight slowdown of the computation - to
> > be fixed on another day.
> 
> Please fix this commit.

Sigh.  I wonder how this wrong version made it into my tree.  Oh well,
applied.  No time to fuzz around before 2.6.30 even though I'd like to
avoid the 64-bit arithmetic.

Applied.  Thanks!

  Ralf

Re: MIPS: Outline udelay and fix a few issues.

[Geert Uytterhoeven]

On Tue, Jun 9, 2009 at 14:20, Ralf Baechle<ralf@linux-mips.org> wrote:
> On Tue, Jun 09, 2009 at 11:12:48AM +0900, Atsushi Nemoto wrote:
>> > Relying on pure C for computation of the delay value removes the need for
>> > explicit.  The price we pay is a slight slowdown of the computation - to
>> > be fixed on another day.
>>
>> Please fix this commit.
>
> Sigh.  I wonder how this wrong version made it into my tree.  Oh well,
> applied.  No time to fuzz around before 2.6.30 even though I'd like to
> avoid the 64-bit arithmetic.
>
> Applied.  Thanks!

I can confirm that after this patch, 2.6.30 builds and boots fine on RBTX4927.

Please submit it for 2.6.30-rc1 and 2.6.30.1 ;-)

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