[PVFS-developers] Re: Kernel Build Problem

[Rob Ross]

We should break this into two separate feature tests.

On Fri, 23 Jan 2004, Porter Don wrote:

>  That is exactly the case.
> 
> -----Original Message-----
> From: Nathan Poznick
> To: 'pvfs-developers@www.beowulf-underground.org'
> Sent: 1/23/04 10:46 AM
> Subject: Re: [PVFS-developers] Re: Kernel Build Problem
> 
> Thus spake Robert Latham:
> > b: when you apply the patch, configure doesn't figure out that it
> >    needs to use the workaround?
> 
> I think it's that configure doesn't figure out that it needs to use the
> workaround - in combination with the fact that it won't need all the
> workarounds for this kernel.  For instance, it appears that the
> 2.4.18-blah Redhat kernels need the for_each_process() to
> for_each_task() conversion, but NOT the current->sigmask_lock to
> current->sighand->siglock conversion.  Things might need to be broken
> out into two different configure tests / defines.
> 
> Oh the joys of Redhat kernels :)
> 
> -- 
> Nathan Poznick <poznick@conwaycorp.net>
> 
> Pride is the master sin of the devil, and the devil is the father of
> lies. - Edwin Hubbel Chapin
> 
> _______________________________________________
> PVFS-developers mailing list
> PVFS-developers@www.beowulf-underground.org
> http://beowulf-underground.org/mailman/listinfo/pvfs-developers
> 
>