[ENBD] Re: make test fails with 2.4.32 on linux 2.4.22

[P.T. Breuer]

In article <1078214671.1494.108.camel at pc19> you wrote:
> Sorry if I interrupt but in the 2.4.32 archive I have tested "make test"
> as always shown these errors even when It works in true usage. May be
> you should try enbd-test instead ? 

It's possible. I haven't run "make test" for a long time. In 2.4.32
there is a shell script that does the test, and I just load the modules
and run

   nbd/enbd-maketest --noreload --notest [other opts]

and then follow that with

  /tmp/enbd-test /dev/nda

to do the test. But that is surely what the "make test" itself does, as
one can see below!

> On Mon, 2004-03-01 at 21:05, Peter T. Breuer wrote:
> > "Also sprach Anders Blomdell:"
> > > export OPTASYNC="" OPTBLK="-b 1024" OPTCLISSL="" OPTJRNL="" OPTLOCK="" 
> > > OPTMD5SUM="" OPTORDER="-w 500" OPTRAID="-l" OPTRCACHE="-h 0" OPTRO="" 
> > > OPTSASYNC="" OPTSIG="-i "NBDabcdefNBD"" OPTSOCK="-n 4" OPTSRVSSL="" OPTTIME=
> > > "-t 120" OPT_REQ_TIMEO="-p 5" BUILD="/tmp" VPATH="/tmp" DEVICE="/dev/nda" 
> > > SERVER="localhost" CLIENT="localhost" EXPORT="/tmp/core0 /tmp/core1" BLKSIZ=
> > > "1024" SRVSIZ="4096" RAHEAD="20" MERGE_REQS="0" CHK_PART="0" ; 
> > > /usr/src/nbd-2.4.32/nbd/enbd-maketest

There. It ran nbd/enbd-maketest, after setting some environment
variables. Maybe the takeup of environment variables in the script
isn't good?

Peter