On Thu, Jan 04, 2001 at 03:04:31PM -0800, Timothy J. Wood wrote:
> > $ rdsinit -f basher_log basher_data 0x210000 0x50000000 0x100000 
> > 0x100000 80 32
> 
> I get:
> 
> rvm_initialize succeeded.
> Going to initialize data file to zero, could take awhile.
> done.
> ?  ERROR: rds_zap_heap RVM_EINTERNAL.
> 
>    If this rings a bell with you, any advice would be appreciated.  
> Otherwise, I'll just take a look at what is going on :)

Erhm, I probably got the arguments wrong, let's see.

Usage: rdsinit [-f] log data_seg [ datalen saddr hlen slen nl chunk ] 
  where
    -f       supplied parameter are firm, do not ask for confirm
    log      is the name of log file/device
    data_seg is the name of data segment file/device
    datalen  gives the length of data segment
    saddr    is the starting address of rvm
    hlen     is the heap length
    slen     is the static length
    nl       is the number of free list used by RDS
    chunk    is the chunksize

maybe the 0x50000000 should be something like 0x20000000 on linux, at
least that is what venus is using.

datalen must be at least one memory page larger than hlen + slen.
nl and chunk are some allocator parameters that I typically leave at 80
and 32.

Alas, RVM isn't very clear in it's error reporting and represents almost
anything with RVM_EINTERNAL.

Jan