Re: [PATCH v2] rpc.mountd: let mountd consult /etc/services for port

[Mi Jinlong <mijinlong@cn.fujitsu.com>]

Steve Dickson:
> On 05/28/2011 12:45 PM, Jim Rees wrote:
>> Chuck Lever wrote:
>>
>>   So, we could also address this by getting rid of the legacy RPC behavior
>>   instead.  However, in cases like this, I typically choose to stick with
>>   legacy behavior, since that improves backwards compatibility.  It already
>>   works like this for legacy RPC builds, so in some sense we are stuck with
>>   it.
>>
>> Ok.  I remember the earlier discussion but forgot that the patch simply
>> restores older behavior that was changed by tirpc.  In that case I'll go
>> along with Mi's patch.
> I am all for restoring the older behavior, but unfortunately this patch
> does not do it. 
> 
> I when back and took a look at how the  nfs-utils-1.2.2 code worked.
> While its true both mountd and statd read ports from /etc/service, 
> they did not fail when those ports were already taken. They just 
> bound to random ephemeral ports, which is probably the reason none of 
> us noticed they were reading ports out of /etc/services. With
> Mi's patch, both daemons fail when the ports in /etc/service are
> already taken. 

Hi steve,

  Do you mean the daemons can't run?

  I test nfs-utils-1.2.2 at fedora15, mountd can bind to the port 
  reading form /etc/service. 

  So, can you post some error or other message here?

thanks,
Mi Jinlong