Re: [Gcl-devel] HEAD and GBC on Windows.

[Camm Maguire]

Greetings!

"Mike Thomas" <address@hidden> writes:

> Hi Camm.
> 
> The long standing problem I face here is that GBC is not enabled in
> o/main.c when the function malloc() is called to make space for a string
> when raw_pre_gcl.exe is first started up during the build.  (Either line
> 201 or 206 causes the problem.)
> 
> On Windows I have defined "malloc" to be "gcl_malloc" to stop GCL's
> malloc being called by the C runtime before main() starts up (causing
> the same GBC not enabled issue before main evn begins).
> 
> Apparently on Linux GBC is not called during those early calls to GCL's
> malloc so I suppose that either my configuration is returning values
> which are wrong, or not able to be dealt with by the GCL runtime as it
> stands, or Linux is using a different malloc to the one used on Windows
> or some other problem exists.
> 
> Here is some debugging output.  Any comments would be appreciated as I
> want to remove these problems ASAP (within currently unpredictable time
> constraints).
> 
> ================================================================
> 
> 
> (gdb) r
> Starting program: c:\cvs\head\gcl/unixport/raw_pre_gcl.exe
> 
> Breakpoint 2, gcl_malloc (size=5) at alloc.c:1569
> 1569            if (in_malloc)
> (gdb) n
> 1571            in_malloc=1;
> (gdb) n
> 1584            if (GBC_enable==0) {
> (gdb) n
> 1585               if ( initflag ==0)
> (gdb) n
> 1586                 gcl_init_alloc();
> (gdb) n
> 
> Breakpoint 3, gcl_init_alloc () at alloc.c:936
> 936       if (gcl_alloc_initialized) return;
> (gdb) n
> 961         malloc_list = Cnil;
> (gdb) n
> 962         enter_mark_origin(&malloc_list);
> (gdb) n
> 966       set_maxpage();
> (gdb) n
> 985       INIT_ALLOC;
> 

Here is the origin of the problem it seems.  From what I can see,
mingw is using the def from att.h, which calls sbrk(0) with some page
rounding.  Either this is not setting core_end above DBEGIN, or (more
likely) something is wrong with the arithmetic in

#define CORE_PAGES (((ufixnum)core_end-DBEGIN)>>PAGEWIDTH)

BTW, PAGESIZE is always a power of 2 on the machines I work on, and
indeed is defined (if memory serves) as (1<<PAGEWIDTH).

I'm not sure if I follow the meaning of reserve vs. commit below.

What is PAGEWIDTH?  Is DBEGIN an unsigned fixnum?  What is core_end
here? 

Take care, and thank you as always!!!!!



> 
> ;; NOTE Cam: allocate_heap() reserves heap but commits nothing. 
> ;; Apparently commits are done by sbrk() / alloc_heap().
> ;; allocate_heap() lives in "o/unexnt.c"
> 
> allocate_heap: base 3000000, end 3d000000, reserved_heap_size 3a000000,
> PAGESIZE 1000, MAXPAGE 40000, MAXCORE 42fff000, INIT_NRBDIV 9,
> INIT_HOLEDIV 3, HOLEDIV 4
> 
> 
> (gdb) n
> 988       holepage = CORE_PAGES/INIT_HOLEDIV;
> (gdb) n
> 994       new_holepage = CORE_PAGES/HOLEDIV;
> (gdb) n
> 995       nrbpage = CORE_PAGES/INIT_NRBDIV;
> (gdb) n
> 998         fprintf ( stderr,
> 
> gcl_init_alloc: PAGESIZE 1000, MAXPAGE 40000, MAXCORE 42fff000,
> CORE_PAGES 0, INIT_NRBDIV 9, INIT_HOLEDIV 3, HOLEDIV 4
> 
> 
> (gdb) n
> 1011      alloc_page(-(holepage + nrbpage));
> (gdb) p holepage
> $1 = 0
> (gdb) p new_holepage
> $2 = 0
> (gdb) p nrbpage
> No symbol "nrbpage" in current context.
> (gdb) n
> 
> Breakpoint 4, alloc_page (n=0) at alloc.c:105
> 105             e = heap_end;
> (gdb) n
> 106             if (n >= 0) {
> (gdb) p e
> $3 = 0x3000000 <Address 0x3000000 out of bounds>
> (gdb) n
> 107                     if (n >=
> (gdb) n
> 112                             holepage = new_holepage + n;
> (gdb) n
> 119                             if(in_signal_handler)
> (gdb) n
> 125                             hole_overrun=1;
> (gdb) n
> 126                             GBC(t_relocatable);
> (gdb)
> 
> =======================================================
> 
> Cheers
> 
> Mike Thomas.
> 
> 
> _______________________________________________
> Gcl-devel mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/gcl-devel
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah