Is 'ath' going to be updated for 7.1-RELEASE ?

[Sam Leffler]

Jack Twilley wrote:
> Sam Leffler wrote:
>> Jack Twilley wrote:
>>> Sam Leffler wrote:
>>>> Ashish Shukla à ¤†à ¤¶à ¥€à ¤· 
>>>> à ¤¶à ¥à ¤•à ¥à ¤² wrote:
>>>>> Hi,
>>>>>
>>>>> Just came across the release schedule of 7.1-RELEASE. I'm 
>>>>> wondering if
>>>>> new version of 'ath' driver (for Athereos Wireless chipsets) is going
>>>>> to be included or not ? If yes, I'll be glad to try the betas :).
>>>>>
>>>>>   
>>>> Not sure what "new version" means.  I'm unlikely to have any time to
>>>> look at existing ath problems in RELENG_7 before the release.
>>> I am hoping against hope it means "whatever it takes to make the Asus
>>> Eee PC work without having to follow the steps on the EeeBSD page".  I
>>> rebuilt kernel and world today and forgot *again* to do the madwifi
>>> dance.  It's getting old.
>>
>> The required hal is not even committed to HEAD yet; there's no way it's
>> going to get into RELENG_7 w/o significant testing.  As to "getting old"
>> others are free to pitch in to make things happen.  I have very little
>> time for freebsd right now and what time I do have is focused on higher
>> priority issues.
>
> All I can offer is the ability to test.  I do not have the skill 
> necessary to fix this or I would have written a fix and asked for it 
> to be committed.  It is amazingly frustrating to know that there is a 
> workaround and to see that others consider this a low priority issue.
>
> If not you, then who should I ask about committing the new HAL to HEAD 
> and helping me check out whatever I need to check out to test the fix 
> so the rest of us can have wireless Ethernet working out of the box?

I have repeatedly asked for testing of the 0.10.5.6 hal sitting in 
http://www.freebsd.org/~sam on HEAD but not gotten enough results to 
commit to HEAD.  I will not commit a hal to any tree that causes 
regressions.  Atheros long ago yanked the testing resources I used to 
validate hal's for release so I am uncertain whether this hal will 
introduce regressions.  If things go bad I simply don't have the time to 
deal with it so nothing has gone in.

Past that a backport to RELENG_7 is straightforward but again nothing 
happens until code is in HEAD.

I am asking around for someone else to deal with a commit to HEAD.

    Sam