Re: [Gnewsense-dev] Re: low wifi connectivity (rtl8187)

[Zhang Le]

On 10:29 Sun 05 Apr     , yanhua wrote:
> The chip vendor have their considerations. They have no plan to port it  
> to the main kernel tree. we have tried it ourselves. But it just didn't  
> work as the vendor providing driver.

FYI: http://www.kroah.com/log/linux/free_drivers.html

And this: http://www.catb.org/~esr/jargon/html/T/top-post.html
Please try not to top post whenever bottom posting makes more sense.
Thanks.

Zhang, Le

>
> For the long run, we will do it. It may take a long time.
> zhangfx 写道:
>> The problem is, current usable driver is from hardware vendor, they  
>> maintain a modified deprecated protocol  for the driver, it will  
>> conflict with the trend of mainstream, can hardly be accepted into  
>> main stream.
>>
>> The proper way is to fix the issues of main stream driver(there is  
>> one). But we lack of documents to do this. By comparing vendor's  
>> driver and the main stream might lead to a solution, but it is not   
>> easy because some issues seems quite tricky.
>>
>> Regards
>> Robert Millan 写道:
>>> On Sat, Apr 04, 2009 at 02:29:58AM +0800, Zhang Le wrote:
>>>  
>>>> On 19:50 Fri 03 Apr     , Robert Millan wrote:
>>>>    
>>>>> On Fri, Apr 03, 2009 at 10:59:17PM +0800, yanhua wrote:
>>>>>      
>>>>>> Robert Millan 写道:
>>>>>>        
>>>>>>> On Wed, Apr 01, 2009 at 09:19:19PM +0800, fxzhang wrote:
>>>>>>>            
>>>>>>>> Have you tried the version of    
>>>>>>>> http://dev.lemote.com/drupal/sites/default/files/rtl8187B_linux_26.1051.0116.2009_driver.tar.bz2
>>>>>>>>  
>>>>>>>> 
>>>>>>>>
>>>>>>>>
>>>>>>>> with a matching network manager, it is almost faultless now.
>>>>>>>>                 
>>>>>>> I think Graziano tried it, but for a working solution I need 
>>>>>>> the source code.
>>>>>>> We don't publish sourceless binaries in gNewSense :-)
>>>>>>>
>>>>>>>            
>>>>>>>> Yanhua: you can arrange to post the source software.
>>>>>>>>                 
>>>>>>> Great, thanks.  Can this code be merged into your 
>>>>>>> linux_loongson git tree?
>>>>>>>
>>>>>>> It would be much easier to handle than a separate build system.
>>>>>>>             
>>>>>> Have you got the source code now?
>>>>>>         
>>>>> A few times :-)
>>>>>
>>>>>      
>>>>>> The 8187b driver is somewhat imcompatible with current linux  
>>>>>> kernel. It  heavily modified the ieee80211 stack, which is  
>>>>>> deprecated by the current  linux implementaion. So we think  
>>>>>> merging it in the kernel tree is not  proper.
>>>>>>         
>>>>> Is resyncing with recent ieee80211 stack out of the question?
>>>>>       
>>>> IMHO, in the long run migrating the code to the new framework is 
>>>> the right thing
>>>> to do.
>>>>
>>>> Otherwise, we would be stuck with an old linux kernel, or the code  
>>>> would just die
>>>> and we can't benefit from it any more.
>>>>     
>>>
>>> What's wrong with the new version of the driver anyway?  Perhaps it's 
>>> just a
>>> matter of adjusting a register.  I'm afraid I'm no expert in this  
>>> area...
>>>
>>>   
>>
>>
>>
>
>
> -- 
> 晏华
>
>