Date: Tue, 09 Sep 1997 21:42:44 +0200 From: "Helmut F. Wirth" <hfwirth@ping.at> To: Bruce Evans <bde@zeta.org.au> Cc: emulation@freebsd.org Subject: Re: Doscmd, debugging with gdb Message-ID: <3415A6B4.41C67EA6@ping.at> References: <199709090457.OAA28185@godzilla.zeta.org.au>
next in thread | previous in thread | raw e-mail | index | archive | help
Bruce Evans wrote:
>
> There must be more to it than that, since gdb can set breakpoints
> in shared libraries although it doesn't understand symbols in shared
> libraries. Perhaps it is relocating text addresses by 0x110000 when
> it should use 0.
>
> Bruce
Bruce,
I have investigated a little.. indeed there seems to be more.
Here is the ktrace output from gdb in this situation:
524 gdb CALL write(0x1,0xe1000,0xc)
524 gdb GIO fd 1 wrote 12 bytes
"Continuing.
"
524 gdb RET write 12/0xc
524 gdb CALL ptrace(PT_READ_I,0x208,0x127b0c,0)
524 gdb RET ptrace 4031315/0x3d8353
524 gdb CALL ptrace(PT_READ_I,0x208,0x127b0c,0)
524 gdb RET ptrace 4031315/0x3d8353
* 524 gdb CALL ptrace(PT_WRITE_D,0x208,0x127b0c,0x3dcc53)
* 524 gdb RET ptrace -1 errno 14 Bad address
* 524 gdb CALL ptrace(PT_WRITE_I,0x208,0x127b0c,0x3dcc53)
* 524 gdb RET ptrace -1 errno 14 Bad address
524 gdb CALL ptrace(PT_READ_I,0x208,0x127b0c,0)
524 gdb RET ptrace 4031315/0x3d8353
524 gdb CALL ptrace(PT_WRITE_D,0x208,0x127b0c,0x3dcc53)
524 gdb RET ptrace -1 errno 14 Bad address
524 gdb CALL ptrace(PT_WRITE_I,0x208,0x127b0c,0x3dcc53)
524 gdb RET ptrace -1 errno 14 Bad address
524 gdb CALL write(0x2,0xcb460,0x1c)
524 gdb GIO fd 2 wrote 28 bytes
"Cannot insert breakpoint 1:
"
524 gdb RET write 28/0x1c
524 gdb CALL write(0x1,0xe1000,0x36)
524 gdb GIO fd 1 wrote 54 bytes
"Error accessing memory address 0x127b0d: Bad address.
"
The lines marked with (*) show the EFAULT (14) error. The address lies
inside the emulator, but not in the VM86 area.
Here is the map from /proc/PID/map (map is from doscmd, not gdb):
0x0 0x10000 7 0 rwx vnode
0x10000 0xe0000 11 0 rwx default
0xe0000 0xf0000 0 0 rwx none
0xf0000 0x100000 1 0 rwx default
0x100000 0x110000 14 0 rwx vnode
0x110000 0x1ef000 52 0 rwx default
0x8003000 0x8013000 13 0 r-x COW vnode
0x8013000 0x8015000 0 2 rwx COW vnode
0x8015000 0x801f000 0 10 rwx default
0x8020000 0x8023000 3 0 r-x COW vnode
0x8023000 0x8024000 0 1 rwx COW vnode
0x8024000 0x8089000 48 0 r-x COW vnode
0x8089000 0x808d000 0 4 rwx COW vnode
0x808d000 0x80a4000 0 9 rwx default
0xefbde000 0xefbfe000 1 4 rwx default
As you can see, the area in question is mapped as default. The vnode
entries from 0x0-0x10000 and 0x100000-0x110000 are the emulation
of the memory wrap around (High memory area), the vnode from
0xe0000-0xf0000 is the (currently unmapped) EMS segment.
The address requested by gdb is 0x127b0d and in range of
0x110000-0x1ef000. I checked the address in the symbol table, it is
correct. It should work .. ?
Any ideas ?
Thank you
Helmut
--
Helmut F. Wirth
Email: hfwirth@ping.at
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3415A6B4.41C67EA6>