-----Original Message-----
From: Thomas Monjalon [mailto:thomas.monjalon@6wind.com]
Sent: Thursday, September 8, 2016 10:36 AM
To: Gowrishankar Muthukrishnan <redacted>
Cc: dev@dpdk.org; Dumitrescu, Cristian <redacted>;
Chao Zhu [off-list ref]; Richardson, Bruce
[off-list ref]; Ananyev, Konstantin
[off-list ref]; Pradeep [off-list ref]
Subject: Re: [dpdk-dev] [PATCH v6 9/9] table: align rte table hash structs for
cache line size

2016-08-31 17:29, Dumitrescu, Cristian:

quoted

From: Gowrishankar Muthukrishnan

quoted

rte table hash structs rte_bucket_4_8, rte_bucket_4_16 and
rte_bucket_4_32 have
to be cache aligned as required by their corresponding hash create

functions

quoted

quoted

rte_table_hash_create_key8_lru etc.

Hi Gowrishankar,

My understanding is you are trying to work around the check invoked by

the hash table create functions that verifies the size of the bucket header
structure is a multiple of the cache line, right?

quoted

Given that the size of this structure is 1x, 2x or 3x times 64 bytes, the check

passes on IA CPUs (cache line of 64 bytes; explicit alignment to cache line size
is not needed in order to make the size of the structure a multiple of cache
line), but not on PPC CPUs (cache line of 128 bytes), correct?

quoted

I don't think your proposal provides the best way to fix this issue, since

your code leads to a considerable increase in the memory consumption used
per bucket in most cases:

quoted

	- 100% more memory for 8-byte key hash table
	- 0% more for 16-byte key hash table (code does not fix anything,

explicit alignment is not needed)

quoted

	- 50% more for 32-byte key hash table

I suggest you simply change the check: instead of validating this data

structure is a multiple of cache line size, validate it is a multiple of 64 bytes.

Any news please?
The whole series is blocked for this patch.
Should we expect a v7?

Yes, I think we should. Small fix for a considerable benefit.