[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [lwip-devel] TCP errors going from 1.4.1 to 2.0.3
From: |
goldsimon |
Subject: |
Re: [lwip-devel] TCP errors going from 1.4.1 to 2.0.3 |
Date: |
Thu, 09 Nov 2017 15:18:37 +0100 |
User-agent: |
K-9 Mail for Android |
Bill Auerbach wrote:
>>When flushing caches, it's important that struct pbuf is
>cache-aligned, i.e. flushing/invalidating
>>the payload region does not interfere with the rest of the struct.
>
>I didn't think this is the case flushing with write-through.
You're using write through cache? I thought the Nios had a write back cache.
But it's some years ago that I worked With that processor.
What I meant is after cache flushing, the cpu does not reference the pbuf
payload any more, so it doesn't get loaded into the cache again. The pbuf
members do get accessed and will be loaded into the cache, so you have to
ensure they don't share a cache line with the payload.
I'll see if I can benchmark that on an STM Nucleo-144, some of them have
cache...
Cheers,
Simon