This is the mail archive of the gdb@sourceware.cygnus.com mailing list for the GDB project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]

Re: gdb in CVS is totally broken on Linux/i386

To: "H . J . Lu" <hjl at lucon dot org>
Subject: Re: gdb in CVS is totally broken on Linux/i386
From: Jim Blandy <jimb at cygnus dot com>
Date: 15 Nov 1999 13:52:22 -0500
Cc: GDB <gdb at sourceware dot cygnus dot com>
References: <19991112214208.CAB231B493@ocean.lucon.org><5mso2bs6lm.fsf@jtc.redbacknetworks.com><19991112143845.A11639@lucon.org>


> The register fetch/store codes in i386-linux-nat.c are mostly bogus.
> You cannot do
> 
> fpregset_t buf;
> 
> ptrace (PTRACE_GETFPREGS, inferior_pid, 0, (int) &buf);
> 
> since fpregset_t is not what the kernel expects. I got "segmentation
> fault" even for "info reg". Jim, could you please take a look at
> i386lnx-nat.c in my gdb 4.17.0.14 to see what the working one looks
> like?

Hmm.  It looks to me like `fpregset_t' has exactly the layout used by
the kernel.  Both of them exactly match the format used by the
Pentium's own FSAVE/FRSTOR functions.

This shouldn't affect the ability to call functions, though --- that
only involves the integer registers.  Is there a problem there?

References:
- Call function in gdb
  - From: H.J. Lu
- Re: Call function in gdb
  - From: J.T. Conklin
- gdb in CVS is totally broken on Linux/i386
  - From: H . J . Lu

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]