Bochs/bochs/patches/patch.highmem
2002-09-25 13:29:11 +00:00

115 lines
4.4 KiB
Plaintext

----------------------------------------------------------------------
Patch name: patch.highmem
Author: Zwane Mwaikambo <zwane@linuxpower.ca>
Date: Tue, 24 Sep 2002 21:13:36 -0400 (EDT)
Detailed description:
This patch allows me to allocate upto 1G of memory, came very handy when i
was doing highmem testing in a linux simulation. Yes the 1G stack
allocation is ugly, but it isn't there for the common case (non debug).
How does this perform on Windows? On my overcommit 2.4.19-pre5-ac3 box
bochs starts much faster too.
Patch was created with:
cvs diff -u
Apply patch to what version:
cvs checked out on DATE, release version VER
Instructions:
To patch, go to main bochs directory.
Type "patch -p1 < THIS_PATCH_FILE".
----------------------------------------------------------------------
From zwane@linuxpower.ca Wed Sep 25 09:30:05 2002
Date: Tue, 24 Sep 2002 21:13:36 -0400 (EDT)
From: Zwane Mwaikambo <zwane@linuxpower.ca>
To: Bochs Devel <bochs-developers@lists.sourceforge.net>
Subject: [Bochs-developers] [PATCH] bochs-highmem
This patch allows me to allocate upto 1G of memory, came very handy when i
was doing highmem testing in a linux simulation. Yes the 1G stack
allocation is ugly, but it isn't there for the common case (non debug).
How does this perform on Windows? On my overcommit 2.4.19-pre5-ac3 box
bochs starts much faster too.
Index: bochs-cvs/config.h.in
===================================================================
RCS file: /cvsroot/bochs/bochs/config.h.in,v
retrieving revision 1.79
diff -u -r1.79 config.h.in
--- bochs-cvs/config.h.in 24 Sep 2002 20:01:56 -0000 1.79
+++ bochs-cvs/config.h.in 25 Sep 2002 00:57:02 -0000
@@ -287,7 +287,7 @@
// is not used, and this section can be ignored.
// =================================================================
-#define BX_MAX_DIRTY_PAGE_TABLE_MEGS 64
+#define BX_MAX_DIRTY_PAGE_TABLE_MEGS 1024
// Compile in support for virtual/linear/physical breakpoints.
// Set to 1, only those you need. Recommend using only linear
Index: bochs-cvs/main.cc
===================================================================
RCS file: /cvsroot/bochs/bochs/main.cc,v
retrieving revision 1.146
diff -u -r1.146 main.cc
--- bochs-cvs/main.cc 24 Sep 2002 22:49:52 -0000 1.146
+++ bochs-cvs/main.cc 25 Sep 2002 00:57:15 -0000
@@ -1609,7 +1609,7 @@
BX_CPU(0)->reset(BX_RESET_HARDWARE);
#else
// SMP initialization
- bx_mem_array[0] = new BX_MEM_C ();
+ bx_mem_array[0] = new BX_MEM_C (bx_options.memory.Osize->get() * 1024*1024);
bx_mem_array[0]->init_memory(bx_options.memory.Osize->get () * 1024*1024);
// First load the optional ROM images
Index: bochs-cvs/memory/misc_mem.cc
===================================================================
RCS file: /cvsroot/bochs/bochs/memory/misc_mem.cc,v
retrieving revision 1.30
diff -u -r1.30 misc_mem.cc
--- bochs-cvs/memory/misc_mem.cc 19 Sep 2002 19:17:20 -0000 1.30
+++ bochs-cvs/memory/misc_mem.cc 25 Sep 2002 00:57:15 -0000
@@ -123,6 +123,7 @@
BX_DEBUG(("Init $Id: patch.highmem,v 1.1 2002-09-25 13:29:11 bdenney Exp $"));
// you can pass 0 if memory has been allocated already through
// the constructor, or the desired size of memory if it hasn't
+ BX_INFO(("%.2fMB", (float)(BX_MEM_THIS megabytes) ));
if (BX_MEM_THIS vector == NULL) {
// memory not already allocated, do now...
@@ -131,21 +132,8 @@
BX_MEM_THIS megabytes = memsize / (1024*1024);
BX_INFO(("%.2fMB", (float)(BX_MEM_THIS megabytes) ));
}
- // initialize all memory to 0x00
- memset(BX_MEM_THIS vector, 0x00, BX_MEM_THIS len);
-
- // initialize ROM area (0xc0000 .. 0xfffff) to 0xff
- memset(BX_MEM_THIS vector + 0xc0000, 0xff, 0x40000);
-
-#if BX_PCI_SUPPORT
- // initialize PCI shadow RAM area (0xc0000 .. 0xfffff) to 0x00
- memset(BX_MEM_THIS shadow, 0x00, 0x40000);
-#endif
#if BX_DEBUGGER
- // initialize dirty pages table
- memset(dbg_dirty_pages, 0, sizeof(dbg_dirty_pages));
-
if (megabytes > BX_MAX_DIRTY_PAGE_TABLE_MEGS) {
BX_INFO(("Error: memory larger than dirty page table can handle"));
BX_PANIC(("Error: increase BX_MAX_DIRTY_PAGE_TABLE_MEGS"));
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
bochs-developers mailing list
bochs-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bochs-developers