Re: x86: ppc fixes for find_first_bit

Re: x86: ppc fixes for find_first_bit

[Alexander van Heukelum]

Hello Thomas,

I see Ingo has applied three fixes to the x86-tree:
    find_first_bit() ppc fix
    powerpc: fix powerpc build
    find_next_bit() fix

Could you please give some insight in what went wrong with
ppc and powerpc?

"find_first_bit() ppc fix" disables the use of find_first_bit
for every user of GENERIC_FIND_NEXT_BIT=y. It replaces it by a
macro to call find_next_bit with offset=0. It should be possible
for an arch to use GENERIC_FIND_NEXT_BIT=y and implement 
find_first_bit by itself.

"powerpc: fix powerpc build" removes the private 'implementation'
of asm-generic/bitops/find.h. It seems correct code to me. What
was the problem here? If it is duplicate declarations, then
I would suggest putting #ifndef GENERIC_FIND_NEXT_BIT around
them.

"find_next_bit() fix" changes asm-generic/bitops/find.h to
declare find_next_bit only if CONFIG_GENERIC_FIND_NEXT_BIT=n.
That is indeed a good change. It would be better if this
file disappeared completely, though.

Greetings,
    Alexander
-- 
  Alexander van Heukelum
  heukelum@fastmail.fm

-- 
http://www.fastmail.fm - One of many happy users:
  http://www.fastmail.fm/docs/quotes.html

Re: x86: ppc fixes for find_first_bit

[Ingo Molnar]

* Alexander van Heukelum <heukelum@fastmail.fm> wrote:

> Hello Thomas,
> 
> I see Ingo has applied three fixes to the x86-tree:
>     find_first_bit() ppc fix
>     powerpc: fix powerpc build
>     find_next_bit() fix
> 
> Could you please give some insight in what went wrong with
> ppc and powerpc?
> 
> "find_first_bit() ppc fix" disables the use of find_first_bit
> for every user of GENERIC_FIND_NEXT_BIT=y. It replaces it by a
> macro to call find_next_bit with offset=0. It should be possible
> for an arch to use GENERIC_FIND_NEXT_BIT=y and implement 
> find_first_bit by itself.
> 
> "powerpc: fix powerpc build" removes the private 'implementation'
> of asm-generic/bitops/find.h. It seems correct code to me. What
> was the problem here? If it is duplicate declarations, then
> I would suggest putting #ifndef GENERIC_FIND_NEXT_BIT around
> them.
> 
> "find_next_bit() fix" changes asm-generic/bitops/find.h to
> declare find_next_bit only if CONFIG_GENERIC_FIND_NEXT_BIT=n.
> That is indeed a good change. It would be better if this
> file disappeared completely, though.

we had trouble making ppc64 defconfig build fine with your bitops 
changes applied (Thomas might still have the build failure logs). The 
fixes are ad-hoc band-aids to get it to build. We used crosscompilers to 
build on ppc64.

	Ingo

[PATCH] x86: fix find_next_bit breakage on ppc and powerpc

[Alexander van Heukelum]

Powerpc (and ppc) have their have some code in their bitops.h
which used to be exacly the same as asm-generic/bitops/find.h.
Include this header instead.

This should also fix the compile problems due to the generic
find_next_bit changes. Those were fixed by Thomas Gleixner in
asm-generic/bitops/find.h earlier.

Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>

---

On Wed, Apr 16, 2008 at 02:57:24PM +0200, Ingo Molnar wrote:
> * Alexander van Heukelum <heukelum@fastmail.fm> wrote:
> > Hello Thomas,
> > 
> > I see Ingo has applied three fixes to the x86-tree:
> >     find_first_bit() ppc fix
> >     powerpc: fix powerpc build
> >     find_next_bit() fix
> > 
> > Could you please give some insight in what went wrong with
> > ppc and powerpc?
> > 
> > "find_first_bit() ppc fix" disables the use of find_first_bit
> > for every user of GENERIC_FIND_NEXT_BIT=y. It replaces it by a
> > macro to call find_next_bit with offset=0. It should be possible
> > for an arch to use GENERIC_FIND_NEXT_BIT=y and implement 
> > find_first_bit by itself.
> > 
> > "powerpc: fix powerpc build" removes the private 'implementation'
> > of asm-generic/bitops/find.h. It seems correct code to me. What
> > was the problem here? If it is duplicate declarations, then
> > I would suggest putting #ifndef GENERIC_FIND_NEXT_BIT around
> > them.
> > 
> > "find_next_bit() fix" changes asm-generic/bitops/find.h to
> > declare find_next_bit only if CONFIG_GENERIC_FIND_NEXT_BIT=n.
> > That is indeed a good change. It would be better if this
> > file disappeared completely, though.
> 
> we had trouble making ppc64 defconfig build fine with your bitops 
> changes applied (Thomas might still have the build failure logs). The 
> fixes are ad-hoc band-aids to get it to build. We used crosscompilers to 
> build on ppc64.

Hello,

Yeah, I reproduced the breakage on x86 by putting an #include
<asm-generic/bitops/find.h> in asm-x86/bitops.h. It's complaining
a lot, then. Sorry about that breakage. I should have put more
thought in the possibility of breakage due to that header file
(I did look at it, but I thought it was harmless).

Thomas' "find_next_bit() fix" is certainly needed and correct.
Acked-by: Alexander van Heukelum <heukelum@fastmail.fm>.

Could you/Thomas try the following on top of that fix? (i.e., with
"find_next_bit() fix" and "find_first_bit() ppc fix" removed?)
I think it should work then.

Architectures that use CONFIG_GENERIC_FIND_NEXT_BIT=y and include
asm-generic/bitops/find.h should be able to switch to the generic
find_first_bit implementation by setting GENERIC_FIND_FIRST_BIT=y
in asm-$ARCH/Kconfig and removing the #include <asm-generic/bitops/find.h>
from their asm-$ARCH/bitops.h.

Greetings,
	Alexander

 include/asm-powerpc/bitops.h |   17 +----------------
 1 files changed, 1 insertions(+), 16 deletions(-)

diff --git a/include/asm-powerpc/bitops.h b/include/asm-powerpc/bitops.h
index 2fc0c45..e2dbb53 100644
--- a/include/asm-powerpc/bitops.h
+++ b/include/asm-powerpc/bitops.h
@@ -318,23 +318,8 @@ static __inline__ unsigned long __fls(unsigned long x)
 	return __ilog2(x);
 }
 #include <asm-generic/bitops/fls64.h>
-
 #include <asm-generic/bitops/hweight.h>
-
-#define find_first_zero_bit(addr, size) find_next_zero_bit((addr), (size), 0)
-unsigned long find_next_zero_bit(const unsigned long *addr,
-				 unsigned long size, unsigned long offset);
-/**
- * find_first_bit - find the first set bit in a memory region
- * @addr: The address to start the search at
- * @size: The maximum size to search
- *
- * Returns the bit-number of the first set bit, not the number of the byte
- * containing a bit.
- */
-#define find_first_bit(addr, size) find_next_bit((addr), (size), 0)
-unsigned long find_next_bit(const unsigned long *addr,
-			    unsigned long size, unsigned long offset);
+#include <asm-generic/bitops/find.h>
 
 /* Little-endian versions */
 

Re: [PATCH] x86: fix find_next_bit breakage on ppc and powerpc

[Ingo Molnar]

* Alexander van Heukelum <heukelum@mailshack.com> wrote:

> Powerpc (and ppc) have their have some code in their bitops.h which 
> used to be exacly the same as asm-generic/bitops/find.h. Include this 
> header instead.
> 
> This should also fix the compile problems due to the generic 
> find_next_bit changes. Those were fixed by Thomas Gleixner in 
> asm-generic/bitops/find.h earlier.
> 
> Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>

thanks, applied. I dropped:

  Subject: powerpc: fix powerpc build
  From: Thomas Gleixner <tglx@linutronix.de>

and i'm keeping:

  Subject: generic: find_next_bit() fix
  From: Thomas Gleixner <tglx@linutronix.de>

should i also drop:

  Subject: find_first_bit() ppc fix
  From: Thomas Gleixner <tglx@linutronix.de>

?

	Ingo

Re: [PATCH] x86: fix find_next_bit breakage on ppc and powerpc

[Alexander van Heukelum]

On Wed, 16 Apr 2008 16:40:51 +0200, "Ingo Molnar" <mingo@elte.hu> said:
> 
> * Alexander van Heukelum <heukelum@mailshack.com> wrote:
> 
> > Powerpc (and ppc) have their have some code in their bitops.h which 
> > used to be exacly the same as asm-generic/bitops/find.h. Include this 
> > header instead.
> > 
> > This should also fix the compile problems due to the generic 
> > find_next_bit changes. Those were fixed by Thomas Gleixner in 
> > asm-generic/bitops/find.h earlier.
> > 
> > Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
> 
> thanks, applied. I dropped:
> 
>   Subject: powerpc: fix powerpc build
>   From: Thomas Gleixner <tglx@linutronix.de>
> 
> and i'm keeping:
> 
>   Subject: generic: find_next_bit() fix
>   From: Thomas Gleixner <tglx@linutronix.de>
> 
> should i also drop:
> 
>   Subject: find_first_bit() ppc fix
>   From: Thomas Gleixner <tglx@linutronix.de>
> 
> ?

Yes, that problem should be fixed by my patch.

Thanks,
    Alexander

> 	Ingo
-- 
  Alexander van Heukelum
  heukelum@fastmail.fm

-- 
http://www.fastmail.fm - Does exactly what it says on the tin

Re: [PATCH] x86: fix find_next_bit breakage on ppc and powerpc

[Paul Mackerras]

Ingo Molnar writes:
> 
> * Alexander van Heukelum <heukelum@mailshack.com> wrote:
> 
> > Powerpc (and ppc) have their have some code in their bitops.h which 
> > used to be exacly the same as asm-generic/bitops/find.h. Include this 
> > header instead.
> > 
> > This should also fix the compile problems due to the generic 
> > find_next_bit changes. Those were fixed by Thomas Gleixner in 
> > asm-generic/bitops/find.h earlier.
> > 
> > Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
> 
> thanks, applied. I dropped:

Why are powerpc (and ppc) patches
- not being sent to the powerpc maintainer (me)
- not being cc'd to the linuxppc-dev@ozlabs.org list
- ending up going through the x86 tree?

How come patches to unify x86_32 and x86_64 bitops need to end up
touching powerpc?

Paul.

Re: [PATCH] x86: fix find_next_bit breakage on ppc and powerpc

[Alexander van Heukelum]

On Thu, 17 Apr 2008 08:55:12 +1000, "Paul Mackerras" <paulus@samba.org>
said:
> Ingo Molnar writes:
> > 
> > * Alexander van Heukelum <heukelum@mailshack.com> wrote:
> > 
> > > Powerpc (and ppc) have their have some code in their bitops.h which 
> > > used to be exacly the same as asm-generic/bitops/find.h. Include this 
> > > header instead.
> > > 
> > > This should also fix the compile problems due to the generic 
> > > find_next_bit changes. Those were fixed by Thomas Gleixner in 
> > > asm-generic/bitops/find.h earlier.
> > > 
> > > Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
> > 
> > thanks, applied. I dropped:
> 
> Why are powerpc (and ppc) patches
> - not being sent to the powerpc maintainer (me)
> - not being cc'd to the linuxppc-dev@ozlabs.org list
> - ending up going through the x86 tree?

Hello,

My apologies for that. The patches that are now in x86#testing were
needed because of changes I introduced. Thomas Gleixner found a
problem with the patches that caused compile problems for basically
all archs with GENERIC_FIND_NEXT_BIT=y, and his fix was to change
asm-generic/bitops/find.h. However, ppc and powerpc did things
differently... (x86 too, but they have special permissions ;) )

(Better late than never) Please consider applying the patch in
        http://lkml.org/lkml/2008/4/16/107
as:

[PATCH] powerpc: use asm-generic/bitops/find.h in bitops.h

Powerpc (and ppc) have some code in their bitops.h that is
exacly the same as asm-generic/bitops/find.h. Include this
header instead of the private implementation.

> How come patches to unify x86_32 and x86_64 bitops need to end up
> touching powerpc?

This was not a pure unification. Originally I wanted to convert
both x86_64 and i386 to the existing generic bitops. Andi Kleen,
however, objected because x86_64 would than lose a certain
optimization for small bitmaps. I moved this optimization to
the generic code, and broke non-x86. Everyone except ppc/powerpc
was fixed by Thomas Gleixner (in a generic header file).
http://lkml.org/lkml/2008/4/16/107 just changes ppc/powerpc in
such a way that Thomas' fix works there too.

Greetings,
    Alexander

> Paul.
-- 
  Alexander van Heukelum
  heukelum@fastmail.fm

-- 
http://www.fastmail.fm - The professional email service