[net-next-2.6 PATCH v2] x86: Align skb w/ start of cacheline on newer core 2/Xeon Arch

[net-next-2.6 PATCH v2] x86: Align skb w/ start of cacheline on newer core 2/Xeon Arch

[Jeff Kirsher]

From: Alexander Duyck <alexander.h.duyck@intel.com>

x86 architectures can handle unaligned accesses in hardware, and it has
been shown that unaligned DMA accesses can be expensive on Nehalem
architectures.  As such we should overwrite NET_IP_ALIGN to resolve
this issue.

Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: x86@kernel.org
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---

 arch/x86/include/asm/system.h |    9 +++++++++
 1 files changed, 9 insertions(+), 0 deletions(-)

diff --git a/arch/x86/include/asm/system.h b/arch/x86/include/asm/system.h
index b8fe48e..b4293fc 100644
--- a/arch/x86/include/asm/system.h
+++ b/arch/x86/include/asm/system.h
@@ -457,4 +457,13 @@ static inline void rdtsc_barrier(void)
 	alternative(ASM_NOP3, "lfence", X86_FEATURE_LFENCE_RDTSC);
 }
 
+#ifdef CONFIG_MCORE2
+/*
+ * We handle most unaligned accesses in hardware.  On the other hand
+ * unaligned DMA can be quite expensive on some Nehalem processors.
+ *
+ * Based on this we disable the IP header alignment in network drivers.
+ */
+#define NET_IP_ALIGN	0
+#endif
 #endif /* _ASM_X86_SYSTEM_H */

Re: [net-next-2.6 PATCH v2] x86: Align skb w/ start of cacheline on newer core 2/Xeon Arch

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Tue, 29 Jun 2010 21:38:00 -0700

> From: Alexander Duyck <alexander.h.duyck@intel.com>
> 
> x86 architectures can handle unaligned accesses in hardware, and it has
> been shown that unaligned DMA accesses can be expensive on Nehalem
> architectures.  As such we should overwrite NET_IP_ALIGN to resolve
> this issue.
> 
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Ingo Molnar <mingo@redhat.com>
> Cc: "H. Peter Anvin" <hpa@zytor.com>
> Cc: x86@kernel.org
> Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Can I get an x86'er ACK on this?  I can merge it in via net-next-2.6
which is probably most convenient for people who want to see the
networking performance effects of this change.

Re: [net-next-2.6 PATCH v2] x86: Align skb w/ start of cacheline on newer core 2/Xeon Arch

[H. Peter Anvin]

On 06/30/2010 02:28 PM, David Miller wrote:
> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> Date: Tue, 29 Jun 2010 21:38:00 -0700
> 
>> From: Alexander Duyck <alexander.h.duyck@intel.com>
>>
>> x86 architectures can handle unaligned accesses in hardware, and it has
>> been shown that unaligned DMA accesses can be expensive on Nehalem
>> architectures.  As such we should overwrite NET_IP_ALIGN to resolve
>> this issue.
>>
>> Cc: Thomas Gleixner <tglx@linutronix.de>
>> Cc: Ingo Molnar <mingo@redhat.com>
>> Cc: "H. Peter Anvin" <hpa@zytor.com>
>> Cc: x86@kernel.org
>> Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
>> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> 
> Can I get an x86'er ACK on this?  I can merge it in via net-next-2.6
> which is probably most convenient for people who want to see the
> networking performance effects of this change.

Acked-by: H. Peter Anvin <hpa@zytor.com>

	-hpa

Re: [net-next-2.6 PATCH v2] x86: Align skb w/ start of cacheline on newer core 2/Xeon Arch

[David Miller]

From: "H. Peter Anvin" <hpa@zytor.com>
Date: Wed, 30 Jun 2010 14:32:33 -0700

> On 06/30/2010 02:28 PM, David Miller wrote:
>> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
>> Date: Tue, 29 Jun 2010 21:38:00 -0700
>> 
>>> From: Alexander Duyck <alexander.h.duyck@intel.com>
>>>
>>> x86 architectures can handle unaligned accesses in hardware, and it has
>>> been shown that unaligned DMA accesses can be expensive on Nehalem
>>> architectures.  As such we should overwrite NET_IP_ALIGN to resolve
>>> this issue.
>>>
>>> Cc: Thomas Gleixner <tglx@linutronix.de>
>>> Cc: Ingo Molnar <mingo@redhat.com>
>>> Cc: "H. Peter Anvin" <hpa@zytor.com>
>>> Cc: x86@kernel.org
>>> Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
>>> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
>> 
>> Can I get an x86'er ACK on this?  I can merge it in via net-next-2.6
>> which is probably most convenient for people who want to see the
>> networking performance effects of this change.
> 
> Acked-by: H. Peter Anvin <hpa@zytor.com>

Applied, thanks everyone.

Re: [net-next-2.6 PATCH v2] x86: Align skb w/ start of cacheline on newer core 2/Xeon Arch

[Andi Kleen]

Jeff Kirsher <jeffrey.t.kirsher@intel.com> writes:


Sorry for the late comment.

>  
> +#ifdef CONFIG_MCORE2
> +/*
> + * We handle most unaligned accesses in hardware.  On the other hand
> + * unaligned DMA can be quite expensive on some Nehalem processors.
> + *
> + * Based on this we disable the IP header alignment in network drivers.
> + */
> +#define NET_IP_ALIGN	0
> +#endif
>  #endif /* _ASM_X86_SYSTEM_H */

The ifdef should be imho dropped and the option be made unconditional
for all x86. I am not aware of any x86 core where unalignment is really
slow. This would increase the chance of it actually working on many
configurations which do not necessarily optimize for Core2.

-Andi
-- 
ak@linux.intel.com -- Speaking for myself only.

Re: [net-next-2.6 PATCH v2] x86: Align skb w/ start of cacheline on newer core 2/Xeon Arch

[Alexander Duyck]

Andi Kleen wrote:
> Jeff Kirsher <jeffrey.t.kirsher@intel.com> writes:
> 
> 
> Sorry for the late comment.
> 
>>  
>> +#ifdef CONFIG_MCORE2
>> +/*
>> + * We handle most unaligned accesses in hardware.  On the other hand
>> + * unaligned DMA can be quite expensive on some Nehalem processors.
>> + *
>> + * Based on this we disable the IP header alignment in network drivers.
>> + */
>> +#define NET_IP_ALIGN	0
>> +#endif
>>  #endif /* _ASM_X86_SYSTEM_H */
> 
> The ifdef should be imho dropped and the option be made unconditional
> for all x86. I am not aware of any x86 core where unalignment is really
> slow. This would increase the chance of it actually working on many
> configurations which do not necessarily optimize for Core2.
> 
> -Andi

Seems to make sense to me.  I will see about generating a patch that 
drops the check for CONFIG_MCORE2.

Thanks,

Alex

Re: [net-next-2.6 PATCH v2] x86: Align skb w/ start of cacheline on newer core 2/Xeon Arch

[H. Peter Anvin]

On 07/01/2010 01:37 PM, Alexander Duyck wrote:
> Andi Kleen wrote:
>> Jeff Kirsher <jeffrey.t.kirsher@intel.com> writes:
>>
>>
>> Sorry for the late comment.
>>
>>>  
>>> +#ifdef CONFIG_MCORE2
>>> +/*
>>> + * We handle most unaligned accesses in hardware.  On the other hand
>>> + * unaligned DMA can be quite expensive on some Nehalem processors.
>>> + *
>>> + * Based on this we disable the IP header alignment in network drivers.
>>> + */
>>> +#define NET_IP_ALIGN	0
>>> +#endif
>>>  #endif /* _ASM_X86_SYSTEM_H */
>>
>> The ifdef should be imho dropped and the option be made unconditional
>> for all x86. I am not aware of any x86 core where unalignment is really
>> slow. This would increase the chance of it actually working on many
>> configurations which do not necessarily optimize for Core2.
>>
>> -Andi
> 
> Seems to make sense to me.  I will see about generating a patch that 
> drops the check for CONFIG_MCORE2.
> 

Just drop the #ifdef ... #endif.

	-hpa