PATCH: S/390: don't try to read zero bytes

PATCH: S/390: don't try to read zero bytes

[Jim Blandy]

2001-12-20  Jim Blandy  <jimb@redhat.com>

	* s390-tdep.c (s390_readinstruction): Don't call
	info->read_memory_func to read zero bytes.  Some targets'
	xfer_memory functions can't cope with that.

Index: gdb/s390-tdep.c
===================================================================
RCS file: /cvs/cvsfiles/devo/gdb/s390-tdep.c,v
retrieving revision 2.33
diff -c -r2.33 s390-tdep.c
*** gdb/s390-tdep.c	2001/12/20 03:25:51	2.33
--- gdb/s390-tdep.c	2001/12/20 07:41:37
***************
*** 132,139 ****
    if ((*info->read_memory_func) (at, &instr[0], 2, info))
      return -1;
    instrlen = s390_instrlen[instr[0] >> 6];
!   if ((*info->read_memory_func) (at + 2, &instr[2], instrlen - 2, info))
!     return -1;
    return instrlen;
  }
  
--- 132,142 ----
    if ((*info->read_memory_func) (at, &instr[0], 2, info))
      return -1;
    instrlen = s390_instrlen[instr[0] >> 6];
!   if (instrlen > 2)
!     {
!       if ((*info->read_memory_func) (at + 2, &instr[2], instrlen - 2, info))
!         return -1;
!     }
    return instrlen;
  }
  

Re: PATCH: S/390: don't try to read zero bytes

[Andrew Cagney]

> 2001-12-20  Jim Blandy  <jimb@redhat.com>
> 
> * s390-tdep.c (s390_readinstruction): Don't call
> 	info->read_memory_func to read zero bytes.  Some targets'
> 	xfer_memory functions can't cope with that.
> 


Jim, which?

While slightly stupid, the target should be able to handle that.

	Andrew

Re: PATCH: S/390: don't try to read zero bytes

[Jim Blandy]

Andrew Cagney <ac131313@cygnus.com> writes:
> > 2001-12-20  Jim Blandy  <jimb@redhat.com>
> > 
> > * s390-tdep.c (s390_readinstruction): Don't call
> > 	info->read_memory_func to read zero bytes.  Some targets'
> > 	xfer_memory functions can't cope with that.
> 
> Jim, which?
> 
> While slightly stupid, the target should be able to handle that.

That was my feeling, as well.  xfer_memory, in exec.c, says:

  if (len <= 0)
    internal_error (__FILE__, __LINE__, "failed internal consistency check");

In my opinion, the test is wrong, the error message is lousy, and it
should be an assert anyway.

Re: PATCH: S/390: don't try to read zero bytes

[Jim Blandy]

Andrew Cagney <ac131313@cygnus.com> writes:
> > 2001-12-20  Jim Blandy  <jimb@redhat.com>
> > 
> > * s390-tdep.c (s390_readinstruction): Don't call
> > 	info->read_memory_func to read zero bytes.  Some targets'
> > 	xfer_memory functions can't cope with that.
> > 
> 
> 
> Jim, which?
> 
> While slightly stupid, the target should be able to handle that.

Oh, wait, I understand now.

A target's to_xfer_memory function is supposed to return the number of
bytes transferred, or zero if it can't transfer the bytes.  If you
make a zero-length request, there's no way for it to indicate success.
D'oh!

Re: PATCH: S/390: don't try to read zero bytes

[Andrew Cagney]

> Andrew Cagney <ac131313@cygnus.com> writes:
> 
>> > 2001-12-20  Jim Blandy  <jimb@redhat.com>
>> > 
>> > * s390-tdep.c (s390_readinstruction): Don't call
>> > info->read_memory_func to read zero bytes.  Some targets'
>> > xfer_memory functions can't cope with that.
>> > 
> 
>> 
>> 
>> Jim, which?
>> 
>> While slightly stupid, the target should be able to handle that.
> 
> 
> Oh, wait, I understand now.
> 
> A target's to_xfer_memory function is supposed to return the number of
> bytes transferred, or zero if it can't transfer the bytes.  If you
> make a zero-length request, there's no way for it to indicate success.
> D'oh!


Can you file a CR against this stupidity?  I'll add it to my list of 
things to untangle.

Andrew