1/ discard 'struct unx_cred'. We don't need any data that
is not already in 'struct rpc_cred'.
2/ Don't keep these creds in a hash table. When a credential
is needed, simply allocate it. When not needed, discard it.
This can easily be faster than performing a lookup on
a shared hash
On Thu, Nov 08 2018, Chuck Lever wrote:
>
> Point taken, but having a single mempool for all RPC transports
> and users is also going to be a shared resource that can
> bottleneck.
Agreed.
mempools will only access the pre-allocated memory if a regular
kmalloc(GFP_NOWAIT) fails. I asked an mm dev
> On Nov 7, 2018, at 8:41 PM, NeilBrown wrote:
>
> On Wed, Nov 07 2018, Chuck Lever wrote:
>
>> Hi Neil-
>>
>>
>>> On Nov 6, 2018, at 11:12 PM, NeilBrown wrote:
>>>
>>> 1/ discard 'struct unx_cred'. We don't need any data that
>>> is not already in 'struct rpc_cred'.
>>> 2/ Don't keep t
On Wed, Nov 07 2018, Chuck Lever wrote:
> Hi Neil-
>
>
>> On Nov 6, 2018, at 11:12 PM, NeilBrown wrote:
>>
>> 1/ discard 'struct unx_cred'. We don't need any data that
>> is not already in 'struct rpc_cred'.
>> 2/ Don't keep these creds in a hash table. When a credential
>> is needed, simp
Hi Neil-
> On Nov 6, 2018, at 11:12 PM, NeilBrown wrote:
>
> 1/ discard 'struct unx_cred'. We don't need any data that
> is not already in 'struct rpc_cred'.
> 2/ Don't keep these creds in a hash table. When a credential
> is needed, simply allocate it. When not needed, discard it.
> T
1/ discard 'struct unx_cred'. We don't need any data that
is not already in 'struct rpc_cred'.
2/ Don't keep these creds in a hash table. When a credential
is needed, simply allocate it. When not needed, discard it.
This can easily be faster than performing a lookup on
a shared hash
6 matches
Mail list logo