On Tue, 22 Sep 2009, Darren Reed wrote:
> venugopal iyer wrote:
>>
>> Thanks, Darren.
>>
>> On Mon, 21 Sep 2009, Darren Reed wrote:
>>
>>> On 18/09/09 12:29 PM, Kais Belgaied wrote:
>>>> ...
>>>>
>>>> + rxringavailcnt
>>>>
>>>> + A read-only property that specifies the number of rings
>>>> available
>>>> + on the receive side.
>>>>
>>>> + rxringcnt
>>>>
>>>> + Specifies the number of receive rings side for the MAC client.
>>>> + A value of 0 means this MAC client should not be assigned any RX
>>>> + ring. A non-0 value means reserve that many rings for this MAC
>>>> + client, if available, and fail if not. If this property is not
>>>> + specified the MAC client may get one RX ring, if available, or
>>>> + will be software based.
>>>> +
>>>> + rxhwavailclnt
>>>> +
>>>> + A read-only property that specifies the number of additional
>>>> + RX hardware-based MAC clients that can be created.
>>>> +
>>>> + txringavailcnt
>>>> +
>>>> + A read-only property that specifies the number of rings
>>>> available
>>>> + on the transmit side.
>>>> +
>>>> + txringcnt
>>>> +
>>>> + Specifies the number of transmit rings for the MAC client.
>>>> + A value of 0 means this MAC client should not be assigned need
>>>> any
>>>> + TX ring. A non-0 value means reserve that many rings for this
>>>> MAC
>>>> + client, if available, and fail if not. If this property is not
>>>> + specified the MAC client may get one TX ring, if available, or
>>>> + will be software based.
>>>> +
>>>> + txhwavailclnt
>>>> +
>>>> + A read-only property that specifics the number of additional
>>>> + TX hardware-based MAC clients that can be created.
>>>> +
>>>> +
>>>>
>>>
>>> The only comment I have is the naming, specifically the "cnt" and "clnt"
>>> on the end, seems ... I don't know... awkward/cumbersome?
>>>
>>> For example, "rxringavailclnt" and "rxringcnt" both are associated with
>>> mac client but only one mentions "clnt" in its name.
>>
>> I think you mean rxhwavailclnt and rxringcnt, right?
>> If so, then rxhwavailclnt is the number of clients that can be created and
>> we
>> wanted to have the client in the name, I can add cnt to rxhwavailclnt if
>> that
>> makes it any better.
>>
>>>
>>> Additionally, they are all a "count" of something, so "cnt" should be
>>> present on all, right?
>>
>> yes, they are counts. As mentioned above if rxhwavailclntcnt and
>> txhwavailclntcnt makes it better, I am fine with it.
>
> And that is the point, these names look awful :-(
>
> For example, if you look through any of statistics in kstat, you generally
> don't see "cnt" or "count".
>
>
>>> I'd like to suggest thinking of simpler names that do not include
>>> redundant information such as "clnt" and "cnt."
>>>
>>> For example, if "rxringcnt" became "rxrings", is any meaning really lost?
>>
>> We started off with rxrings, but wanted to make it explicit that it is
>> the number or rx rings and not, say, as ring index.
>
> If it was just "rxring", then perhaps I would agree that maybe it might be
> something else...
>
> But, for example, we have "Inbound Packets", not "Inbound Packet Count" in
> "netstat -i" output.
>
> To me, the label "rxrings" does not imply it could be about anything else but
> rx rings. If it were to be about an index, then it would be "rxringindex" or
> "rxringindexes" - the name becomes more specific.
So, do you prefer:
rxrings/txrings
rxringsavail/txringsavail
rxhwavailclnt/txhwavailclnt
thanks,
-venu
>
> Darren
>
>
