public inbox for gdb@sourceware.org
 help / color / mirror / Atom feed
From: "Amit S. Kale" <amitkale@linsyssoft.com>
To: Mark Kettenis <mark.kettenis@xs4all.nl>
Cc: gdb@sources.redhat.com
Subject: Re: kgdb support for gdb: patch 1
Date: Thu, 28 Oct 2004 10:56:00 -0000	[thread overview]
Message-ID: <200410261501.10647.amitkale@linsyssoft.com> (raw)
In-Reply-To: <200410211722.i9LHM4in000841@elgar.sibelius.xs4all.nl>

[-- Attachment #1: Type: text/plain, Size: 1911 bytes --]

Sorry about the compile error. I have done the changes you have suggested 
below. Here is a new patch.

-Amit

On Thursday 21 Oct 2004 10:52 pm, Mark Kettenis wrote:
>    From: "Amit S. Kale" <amitkale@linsyssoft.com>
>    Date: Thu, 21 Oct 2004 18:21:08 +0530
>
>    Hi,
>
>    I haven't added any support for detecting kernel osabi
>    automatically.  Detecting it with present kernel binary structure
>    is rather shaky. It's safer to let users run the command "set osabi
>    Linux-kernel".
>
> Fair enough.  Indeed it seems that vmlinux is a fairly standard static
> ELF executable.
>
>    I'll be later sending more patches.
>
>    Please review it and let me know if it's ok for inclusion in gdb.
>
> Of course it isn't, since this can't compile ;-).  You seem to be in
> limbo betwen LINUXKERNEL and LINUX_KERNEL.  Please choose the variant
> *with* the underscore.
>
> Anyway, I think for the kernel you shouldn't use the Linux register
> cache layout, bur rather the normal register layout.  The Linux
> register cache includes the "orig_eax" pseudo register that is
> implemented by the kernel.  My guess is that the kernel itself doesn't
> have it.  I'm not completely sure though, but unless you can motivate
> why you do need "orig_eax" please remove these lines:
>
>    +  set_gdbarch_num_regs (gdbarch, I386_LINUX_NUM_REGS);
>    +  set_gdbarch_register_name (gdbarch, i386_linux_register_name);
>    +  set_gdbarch_register_reggroup_p (gdbarch,
> i386_linux_register_reggroup_p);
>
> These lines will need adjustments too:
>
>    +  tdep->gregset_reg_offset = i386_linux_gregset_reg_offset;
>    +  tdep->gregset_num_regs = ARRAY_SIZE (i386_linux_gregset_reg_offset);
>    +  tdep->sizeof_gregset = 17 * 4;
>
> Setting tdep->gregset_num_regs to I386_NUM_GREGS should be OK.
>
> Apart from that, you'll also need to provide a ChangeLog.
>
> Can you post an update patch?
>
> Cheers,
>
> Mark

[-- Attachment #2: lk.patch --]
[-- Type: text/x-diff, Size: 2693 bytes --]

Index: gdb/gdb/i386-linux-tdep.c
===================================================================
--- gdb.orig/gdb/i386-linux-tdep.c	2004-08-07 02:28:28.000000000 +0530
+++ gdb/gdb/i386-linux-tdep.c	2004-10-26 14:59:29.000000000 +0530
@@ -27,6 +27,7 @@
 #include "inferior.h"
 #include "osabi.h"
 #include "reggroups.h"
+#include "elf-bfd.h"
 
 #include "gdb_string.h"
 
@@ -416,6 +417,25 @@
   set_gdbarch_skip_solib_resolver (gdbarch, glibc_skip_solib_resolver);
 }
 
+static void
+i386_linux_kernel_init_abi (struct gdbarch_info info, struct gdbarch *gdbarch)
+{
+  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
+
+  /* Linux Kernel uses ELF format for vmlinux file.  */
+  i386_elf_init_abi (info, gdbarch);
+
+  /* We can use the register offsets used for linux applications. Only
+   * first 16 elements of the array will be used for kernel. Offsets of those
+   * are identical for applications and kernel.  */
+  tdep->gregset_reg_offset = i386_linux_gregset_reg_offset;
+  tdep->gregset_num_regs = I386_NUM_GREGS;
+  tdep->sizeof_gregset = I386_NUM_GREGS * 4;
+
+  tdep->jb_pc_offset = 20;	/* From <bits/setjmp.h>.  */
+
+}
+
 /* Provide a prototype to silence -Wmissing-prototypes.  */
 extern void _initialize_i386_linux_tdep (void);
 
@@ -424,4 +444,6 @@
 {
   gdbarch_register_osabi (bfd_arch_i386, 0, GDB_OSABI_LINUX,
 			  i386_linux_init_abi);
+  gdbarch_register_osabi (bfd_arch_i386, 0, GDB_OSABI_LINUX_KERNEL,
+			  i386_linux_kernel_init_abi);
 }
Index: gdb/gdb/defs.h
===================================================================
--- gdb.orig/gdb/defs.h	2004-10-12 15:36:14.000000000 +0530
+++ gdb/gdb/defs.h	2004-10-26 14:46:27.000000000 +0530
@@ -1031,6 +1031,7 @@
   GDB_OSABI_SOLARIS,
   GDB_OSABI_OSF1,
   GDB_OSABI_LINUX,
+  GDB_OSABI_LINUX_KERNEL,
   GDB_OSABI_FREEBSD_AOUT,
   GDB_OSABI_FREEBSD_ELF,
   GDB_OSABI_NETBSD_AOUT,
Index: gdb/gdb/osabi.c
===================================================================
--- gdb.orig/gdb/osabi.c	2004-07-03 02:57:17.000000000 +0530
+++ gdb/gdb/osabi.c	2004-10-26 14:46:27.000000000 +0530
@@ -57,6 +57,7 @@
   "Solaris",
   "OSF/1",
   "GNU/Linux",
+  "Linux-kernel",
   "FreeBSD a.out",
   "FreeBSD ELF",
   "NetBSD a.out",
Index: gdb/gdb/ChangeLog
===================================================================
--- gdb.orig/gdb/ChangeLog	2004-10-26 14:45:24.000000000 +0530
+++ gdb/gdb/ChangeLog	2004-10-26 14:57:25.000000000 +0530
@@ -1,3 +1,7 @@
+2004-10-26  Amit S. Kale <amitkale@linsyssoft.com>
+	* i386-linux-tdep.c: Added a new osabi Linux-kernel for debugging
+	a remotely running linux kernel.
+
 2004-10-25  Andrew Cagney  <cagney@gnu.org>
 
 	* NEWS: Add Changes since GDB 6.3 section.

      reply	other threads:[~2004-10-26  9:32 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-10-21 17:22 Amit S. Kale
2004-10-22 12:28 ` Mark Kettenis
2004-10-28 10:56   ` Amit S. Kale [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200410261501.10647.amitkale@linsyssoft.com \
    --to=amitkale@linsyssoft.com \
    --cc=gdb@sources.redhat.com \
    --cc=mark.kettenis@xs4all.nl \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).