pending/2543: RE: ARM single step problem with instruction 0xa5ffffeb

["Lincy Julien \(Neo-Soft\)" <Julien dot Lincy at thomson dot net>]

>Number:         2543
>Category:       pending
>Synopsis:       RE: ARM single step problem with instruction 0xa5ffffeb
>Confidential:   yes
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          change-request
>Submitter-Id:   unknown
>Arrival-Date:   Wed Oct 22 14:48:01 UTC 2008
>Closed-Date:
>Last-Modified:
>Originator:     
>Release:        
>Organization:
>Environment:
>Description:
 Okay thanks for answering
 We are going to look for an endianness problem in our environment=20
 
 -----Original Message-----
 From: Daniel Jacobowitz [mailto:dan@codesourcery.com]=20
 Sent: mercredi 22 octobre 2008 15:00
 To: Lincy Julien (Neo-Soft)
 Cc: gdb-gnats@sources.redhat.com
 Subject: Re: ARM single step problem with instruction 0xa5ffffeb
 
 On Wed, Oct 22, 2008 at 11:41:01AM -0000, julien.lincy@thomson.net
 wrote:
 > >Description:
 > I have a problem single stepping with a simple C program on ARM, with
 the instruction 0xa5ffffeb.
 >=20
 > 8538:       a5ffffeb        ldrgeb  pc, [pc, #4075]!        ; 952b
 <__clz_tab+0x6b>
 >=20
 > GDB stops with the error "Invalid update to pc in instruction".
 >=20
 > The C program i'm trying to test is attached.
 > The ARM I use is an ARM v1136 (a bit specific...).
 
 I think there's something wrong with your environment.  That's not a
 valid instruction; PC is always aligned, so pc+4075 is unaligned.
 The middle of __clz_tab doesn't have function pointers in it.  And you
 can't do PC-relative loads with writeback.
 
 GDB is right to complain about that instruction.
 
 --=20
 Daniel Jacobowitz
 CodeSourcery
>How-To-Repeat:
>Fix:
>Release-Note:
>Audit-Trail:
>Unformatted: