[PATCH] Use `unsigned long' for Linux/x86 debug registers

[PATCH] Use `unsigned long' for Linux/x86 debug registers

[Mark Kettenis]

It just occurred to me that Linux/x86 should really be using `unsigned
long' instead of `long' as the type for debug register values.

Checked in.

Mark


Index: ChangeLog
from  Mark Kettenis  <kettenis@gnu.org>

	* i386-linux-nat.c (i386_linux_dr_get): Change type of return
	value to `unsigned long'.  Change type of `value' to `unsigned
	long'.
	(i386_linux_dr_set): Change type of second argument to `unsigned
	long'.
	(i386_linux_dr_set_control): Change type of first argument to
	`unsigned long'.
	(i386_linux_dr_get_status): Change type of return value to
	unsigned long.
	* config/i386/nm-linux.h (i386_linux_dr_set_control,
	i386_linux_dr_get_status): Adjust prototypes accordingly. 

Index: i386-linux-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386-linux-nat.c,v
retrieving revision 1.25
diff -u -p -r1.25 i386-linux-nat.c
--- i386-linux-nat.c 2001/03/28 08:35:14 1.25
+++ i386-linux-nat.c 2001/04/01 10:47:28
@@ -701,11 +701,11 @@ store_inferior_registers (int regno)
 }
 \f
 
-static long
+static unsigned long
 i386_linux_dr_get (int regnum)
 {
   int tid;
-  long value;
+  unsigned long value;
 
   /* FIXME: kettenis/2001-01-29: It's not clear what we should do with
      multi-threaded processes here.  For now, pretend there is just
@@ -731,7 +731,7 @@ i386_linux_dr_get (int regnum)
 }
 
 static void
-i386_linux_dr_set (int regnum, long value)
+i386_linux_dr_set (int regnum, unsigned long value)
 {
   int tid;
 
@@ -748,7 +748,7 @@ i386_linux_dr_set (int regnum, long valu
 }
 
 void
-i386_linux_dr_set_control (long control)
+i386_linux_dr_set_control (unsigned long control)
 {
   i386_linux_dr_set (DR_CONTROL, control);
 }
@@ -769,7 +769,7 @@ i386_linux_dr_reset_addr (int regnum)
   i386_linux_dr_set (DR_FIRSTADDR + regnum, 0L);
 }
 
-long
+unsigned long
 i386_linux_dr_get_status (void)
 {
   return i386_linux_dr_get (DR_STATUS);
@@ -916,7 +916,7 @@ child_resume (int pid, int step, enum ta
 	      CORE_ADDR sp = read_register (SP_REGNUM);
 	      CORE_ADDR addr = sp;
 	      unsigned long int eflags;
-	      
+
 	      if (syscall == SYS_rt_sigreturn)
 		addr = read_memory_integer (sp + 8, 4) + 20;
 
Index: config/i386/nm-linux.h
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/nm-linux.h,v
retrieving revision 1.9
diff -u -p -r1.9 nm-linux.h
--- config/i386/nm-linux.h 2001/03/21 21:22:49 1.9
+++ config/i386/nm-linux.h 2001/04/01 10:47:29
@@ -43,7 +43,7 @@ extern CORE_ADDR register_u_addr (CORE_A
 
 /* Provide access to the i386 hardware debugging registers.  */
 
-extern void i386_linux_dr_set_control (long control);
+extern void i386_linux_dr_set_control (unsigned long control);
 #define I386_DR_LOW_SET_CONTROL(control) \
   i386_linux_dr_set_control (control)
 
@@ -55,7 +55,7 @@ extern void i386_linux_dr_reset_addr (in
 #define I386_DR_LOW_RESET_ADDR(regnum) \
   i386_linux_dr_reset_addr (regnum)
 
-extern long i386_linux_dr_get_status (void);
+extern unsigned long i386_linux_dr_get_status (void);
 #define I386_DR_LOW_GET_STATUS() \
   i386_linux_dr_get_status ()
 

Re: [PATCH] Use `unsigned long' for Linux/x86 debug registers

[Eli Zaretskii]

On Sun, 1 Apr 2001, Mark Kettenis wrote:

> It just occurred to me that Linux/x86 should really be using `unsigned
> long' instead of `long' as the type for debug register values.

They should indeed be unsigned, but why `unsigned long' instead of 
`unsigned int'?  i386-nat.c uses the latter, and ia32 debug registers
are 32-bit wide.

Re: [PATCH] Use `unsigned long' for Linux/x86 debug registers

[Mark Kettenis]

   Date: Sun, 1 Apr 2001 13:04:44 +0200 (IST)
   From: Eli Zaretskii <eliz@is.elta.co.il>

   On Sun, 1 Apr 2001, Mark Kettenis wrote:

   > It just occurred to me that Linux/x86 should really be using `unsigned
   > long' instead of `long' as the type for debug register values.

   They should indeed be unsigned, but why `unsigned long' instead of 
   `unsigned int'?  i386-nat.c uses the latter, and ia32 debug registers
   are 32-bit wide.

Well an `unsigned int' *might* be 16-bits, wheras ISO C guarantees
that `unsigned long' is at least 32-bits.  All that matters, is that
the variables that we use to store ia32 debug registers are at least
32-bit wide.  I was planning on changing i386-nat.c to use `unsigned
long' in the relevant places.  This all doesn't really matter for
native stuff, but if we ever move this code to i386-tdep.c it might
matter.  Better fix it now than be sorry later :-).

Mark