[Qemu-devel] [Patch] x86 breakpoints and memory examination

[Qemu-devel] [Patch] x86 breakpoints and memory examination

[Eddie Kohler]

[-- Attachment #1: Type: text/plain, Size: 498 bytes --]

Hi all,

The following patch is useful for whole-system mode debugging and breakpoint 
setting on i386 QEMU.  It seems like both breakpoints and memory examination 
used semi-physical addresses, in that neither took account of segment 
translation.  This patch (relative to 0.9.1) adds segment translation.

I'm not sure this is the right way to do it (in fact, I'm sure that the memory 
examination patch is the WRONG way to do it) and would appreciate pointers or 
advice.

Thanks,
Eddie Kohler


[-- Attachment #2: qemu-0.9.1-i386-debug.patch --]
[-- Type: text/x-patch, Size: 1056 bytes --]

diff -ru qemu-0.9.1/target-i386/helper2.c qemu-0.9.1-p/target-i386/helper2.c
--- qemu-0.9.1/target-i386/helper2.c	2008-01-06 11:38:45.000000000 -0800
+++ qemu-0.9.1-p/target-i386/helper2.c	2008-01-12 23:56:34.000000000 -0800
@@ -1081,6 +1081,7 @@
 {
     uint32_t pde_addr, pte_addr;
     uint32_t pde, pte, paddr, page_offset, page_size;
+    addr += env->segs[R_DS].base;
 
     if (env->cr[4] & CR4_PAE_MASK) {
         uint32_t pdpe_addr, pde_addr, pte_addr;
diff -ru qemu-0.9.1/target-i386/translate.c qemu-0.9.1-p/target-i386/translate.c
--- qemu-0.9.1/target-i386/translate.c	2008-01-06 11:38:45.000000000 -0800
+++ qemu-0.9.1-p/target-i386/translate.c	2008-01-13 00:00:23.000000000 -0800
@@ -6758,7 +6758,7 @@
     for(;;) {
         if (env->nb_breakpoints > 0) {
             for(j = 0; j < env->nb_breakpoints; j++) {
-                if (env->breakpoints[j] == pc_ptr) {
+                if (env->breakpoints[j] == pc_ptr - dc->cs_base) {
                     gen_debug(dc, pc_ptr - dc->cs_base);
                     break;
                 }

Re: [Qemu-devel] [Patch] x86 breakpoints and memory examination

[Mulyadi Santosa]

Hi..

> The following patch is useful for whole-system mode debugging and breakpoint
> setting on i386 QEMU.  It seems like both breakpoints and memory examination
> used semi-physical addresses, in that neither took account of segment
> translation.  This patch (relative to 0.9.1) adds segment translation.
>
> I'm not sure this is the right way to do it (in fact, I'm sure that the memory
> examination patch is the WRONG way to do it) and would appreciate pointers or
> advice.

I believe (IIRC) jasson wessel had fixed such problem about a year
ago..but I agree that it seems qemu doesn't take segmentation into
account (zero based segment start address, like most OS does these
days I think). In this case, i think your patch does the right thing.

regards,

Mulyadi.