useless ath9k commit logs

useless ath9k commit logs

[Jiri Slaby]

Hi Sujith and Vasanthakumar,

could I ask you to ever explain why your ath9k patches are needed? Your
commit logs often describe the change, but not why you are doing that
(e.g. what exact issue it fixes). This is hard for us to decide whether
it should be backported to distro kernels or not.

For example
    ath9k: Fix bug in ANI channel handling

    When processing MIB interrupts, OFDM and CCK error
    handling routines for low RSSI values have to be invoked
    only when the channel mode is 11G/11B. Since HT channels
    will also fall under the bands 2Ghz/5Ghz, check appropriately.
or
    ath9k: Maintain monotonicity of PER while going across different phy

    Monotonicity of packet error rate should be kept when moving
    from one phy to another (legacy to ht, ht single stream to dual,
    etc). Current code skips updating per for other phys.

I don't know from that, what bug, under what circumstanses these fix. If
any -- e.g. a theoretical issue.

thanks,
-- 
js

useless ath9k commit logs

[Sujith]

Jiri Slaby wrote:
>     ath9k: Fix bug in ANI channel handling
> 
>     When processing MIB interrupts, OFDM and CCK error
>     handling routines for low RSSI values have to be invoked
>     only when the channel mode is 11G/11B. Since HT channels
>     will also fall under the bands 2Ghz/5Ghz, check appropriately.

This was a long standing bug. The original behaviour was changed
by the commit: 38b33707a1ec77f7b4c92ae41cfe93318014f5bf (in wireless-testing).

The bug was identified by code review.
And yep, this should have been mentioned in the commit log.
Sorry about that.

>     ath9k: Maintain monotonicity of PER while going across different phy
> 
>     Monotonicity of packet error rate should be kept when moving
>     from one phy to another (legacy to ht, ht single stream to dual,
>     etc). Current code skips updating per for other phys.

This is also a bug fix. Rate control gets bumpy when switching between
APs operating in different modes. This was identified by internal testing.

Hope this helps.

Sujith

Re: useless ath9k commit logs

[Jiri Slaby]

On 12/09/2009 12:35 PM, Sujith wrote:
> This is also a bug fix. Rate control gets bumpy when switching between
> APs operating in different modes. This was identified by internal testing.
> 
> Hope this helps.

Ah, thanks, there are many of them though, these were just examples.
Please add such a description to each changelog from now on. I think
many kernel distro packagers will appreciate that.

thanks again,
-- 
js