diff mbox series

[net-next,v1] mm: fix build on powerpc with GCC 14

Message ID 20240913192036.3289003-1-almasrymina@google.com (mailing list archive)
State New
Headers show
Series [net-next,v1] mm: fix build on powerpc with GCC 14 | expand

Commit Message

Mina Almasry Sept. 13, 2024, 7:20 p.m. UTC
Building net-next with powerpc with GCC 14 compiler results in this
build error:

/home/sfr/next/tmp/ccuSzwiR.s: Assembler messages:
/home/sfr/next/tmp/ccuSzwiR.s:2579: Error: operand out of domain (39 is
not a multiple of 4)
make[5]: *** [/home/sfr/next/next/scripts/Makefile.build:229:
net/core/page_pool.o] Error 1

Root caused in this thread:
https://lore.kernel.org/netdev/913e2fbd-d318-4c9b-aed2-4d333a1d5cf0@cs-soprasteria.com/

We try to access offset 40 in the pointer returned by this function:

static inline unsigned long _compound_head(const struct page *page)
{
	unsigned long head = READ_ONCE(page->compound_head);

	if (unlikely(head & 1))
		return head - 1;
	return (unsigned long)page_fixed_fake_head(page);
}

The GCC 14 (but not 11) compiler optimizes this by doing:

ld page + 39

Rather than:

ld (page - 1) + 40

Causing an unaligned read error. Fix this by bitwise operand instead of
an arthimetic operation to clear the pointer, which probably
communicates the intention of the code a bit better anyway.

Cc: Simon Horman <horms@kernel.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: David Miller <davem@davemloft.net>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Networking <netdev@vger.kernel.org>
Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Cc: Linux Next Mailing List <linux-next@vger.kernel.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: "linuxppc-dev@lists.ozlabs.org" <linuxppc-dev@lists.ozlabs.org>

Suggested-by: LEROY Christophe <christophe.leroy2@cs-soprasteria.com>
Signed-off-by: Mina Almasry <almasrymina@google.com>

---
 include/linux/page-flags.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Comments

Matthew Wilcox Sept. 13, 2024, 7:22 p.m. UTC | #1
On Fri, Sep 13, 2024 at 07:20:36PM +0000, Mina Almasry wrote:
> +++ b/include/linux/page-flags.h
> @@ -239,8 +239,8 @@ static inline unsigned long _compound_head(const struct page *page)
>  {
>  	unsigned long head = READ_ONCE(page->compound_head);
>  
> -	if (unlikely(head & 1))
> -		return head - 1;
> +	if (unlikely(head & 1UL))
> +		return head & ~1UL;
>  	return (unsigned long)page_fixed_fake_head(page);

NAK, that pessimises compound_head().
Christophe Leroy Sept. 14, 2024, 6:50 a.m. UTC | #2
Hi,

Le 13/09/2024 à 21:22, Matthew Wilcox a écrit :
> On Fri, Sep 13, 2024 at 07:20:36PM +0000, Mina Almasry wrote:
>> +++ b/include/linux/page-flags.h
>> @@ -239,8 +239,8 @@ static inline unsigned long _compound_head(const struct page *page)
>>   {
>>   	unsigned long head = READ_ONCE(page->compound_head);
>>   
>> -	if (unlikely(head & 1))
>> -		return head - 1;
>> +	if (unlikely(head & 1UL))
>> +		return head & ~1UL;
>>   	return (unsigned long)page_fixed_fake_head(page);
> 
> NAK, that pessimises compound_head().
> 

Can you please give more details on what the difference is ?

I can't see what it pessimises. In both cases, you test if the value is 
odd, when it is odd you make it even.

Christophe
Matthew Wilcox Sept. 15, 2024, 10:44 a.m. UTC | #3
On Sat, Sep 14, 2024 at 08:50:46AM +0200, Christophe Leroy wrote:
> Hi,
> 
> Le 13/09/2024 à 21:22, Matthew Wilcox a écrit :
> > On Fri, Sep 13, 2024 at 07:20:36PM +0000, Mina Almasry wrote:
> > > +++ b/include/linux/page-flags.h
> > > @@ -239,8 +239,8 @@ static inline unsigned long _compound_head(const struct page *page)
> > >   {
> > >   	unsigned long head = READ_ONCE(page->compound_head);
> > > -	if (unlikely(head & 1))
> > > -		return head - 1;
> > > +	if (unlikely(head & 1UL))
> > > +		return head & ~1UL;
> > >   	return (unsigned long)page_fixed_fake_head(page);
> > 
> > NAK, that pessimises compound_head().
> > 
> 
> Can you please give more details on what the difference is ?
> 
> I can't see what it pessimises. In both cases, you test if the value is odd,
> when it is odd you make it even.

On x86, for example, it is perfectly valid to load a 64-bit value from
an offset of 0x2f relative to a pointer.  So there's no need to make it
even.
diff mbox series

Patch

diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
index 5769fe6e4950..ea4005d2d1a9 100644
--- a/include/linux/page-flags.h
+++ b/include/linux/page-flags.h
@@ -239,8 +239,8 @@  static inline unsigned long _compound_head(const struct page *page)
 {
 	unsigned long head = READ_ONCE(page->compound_head);
 
-	if (unlikely(head & 1))
-		return head - 1;
+	if (unlikely(head & 1UL))
+		return head & ~1UL;
 	return (unsigned long)page_fixed_fake_head(page);
 }