[OpenAFS] OpenAFS 1.2.5 Won't Build Under Kernel 2.4.18-5

[Marc Schmitt]

5. Bug IDs fixed (http://bugzilla.redhat.com/bugzilla for more info):

65110 - DRM modules were not built for i386 kernel
65149 - Highpoint proc info segfaults when using cat
65207 - Kernel panic: No init found, with DAC960 RAID controller
64921 - NFS version 3 hangs
65772 - close() hangs on file in NFS-mounted dir using
65069 - Copy 300k data crashed NetApp
64984 - Redhat 7.3: nfs writes very slow.
64838 - /proc/ksyms is truncated with cat, dd works
65176 - broken urandom (PATCH INCLUDED)
64825 - Interrupt counts are all 1000000000 too high
64959 - problem with airo.o and airo_cs.o wireless modules
66143 - System hang after 5-12 h IO stress - flushtlb problem?
65264 - iwconfig WEP key configuration fails 7.2 worked fine. Using airo
65182 - RFE: i586-up kernel missing from 2.4.18-4 errata


The NFS fixes are quite interesting for me in terms of performance as we 
still have the homes in NFS (don`t laugh...).
I`ve never seen problems with 2.4.18-4 that end in a crash, though.

Greetz
	Marc


Dirk Heinrichs wrote:
> On Thu, 20 Jun 2002, Computers in Medicine wrote:
> 
> 
>>I recently installed kernel 2.4.18-5 from Red Hat RPMs and now OpenAFS 1.2.5 won't build.  The procedure below worked fine for 2.4.18-4 and most other kernels before it.  I'm not expect at compiling programs under Linux (read I just make and pray), but since it's always worked well before we compile the kernels ourselves to be able to immediately update Linux kernels before packages are made.  Unfortunately, it didn't work this time and we would very much appreciate any hints as to why. Here's what we're doing.
> 
> What's the difference between -4 and -5, then. If you don't need the
> changes then you could also stay with -4.
> The first thing I do everytime after installing any distro is to get
> vanilla kernel sources and compile those, because you never really know
> which ugly patches they put in.
> 
> 
>>../linux/sched.h:454:15: #if with no expression
> 
> Looks like some of RH's patching went wrong, or they applied a broken
> patch. I would revert to -4.
> 
> Bye...
> 
> 	Dirk