Re[2]: [Gc] vm_portect failed

Ivan Maidanski ivmai at mail.ru
Wed Sep 30 02:16:31 PDT 2009


Hi!

"Boehm, Hans" <hans.boehm at hp.com> wrote:
> This is the first I've heard of it.
> 
> It would be good to know whether the CVS collector by itself built with
> "./configure; make; make check" fails in the same way.
> 
> In either case, I would try to debug the failing vm_protect call to see what error code it's returning and why, possibly after applying a patch like the following to at least fix the abort messages:
> 
> --- os_dep.c	2009-09-25 10:39:10.000000000 -0700
> +++ os_dep.c.revised	2009-09-29 11:36:25.000000000 -0700
> @@ -2601,12 +2601,12 @@
>  #   define PROTECT(addr,len) \
>          if(vm_protect(GC_task_self,(vm_address_t)(addr),(vm_size_t)(len), \
>                  FALSE,VM_PROT_READ) != KERN_SUCCESS) { \
> -            ABORT("vm_portect failed"); \
> +            ABORT("vm_protect (PROTECT) failed"); \
>          }
>  #   define UNPROTECT(addr,len) \
>          if(vm_protect(GC_task_self,(vm_address_t)(addr),(vm_size_t)(len), \
>                  FALSE,VM_PROT_READ|VM_PROT_WRITE) != KERN_SUCCESS) { \
> -            ABORT("vm_portect failed"); \
> +            ABORT("vm_protect (UNPROTECT) failed"); \
>          }
>  # else
> 
> (I'll check this in at some point if nobody beats me to it.)

I've checked in this simple one.

> 
> Hans
> .... 

Bye.


More information about the Gc mailing list