[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: gas/ia64 mov r=imm problem
From: |
Jim Wilson |
Subject: |
Re: gas/ia64 mov r=imm problem |
Date: |
29 Jun 2004 16:29:56 -0700 |
On Fri, 2004-06-18 at 15:51, Belochapka, Konstantin wrote:
> Intel Itaniaum assembler manual states that mov r=imm pseudo instruction is
> actually addl r=imm22,r0, but GAS uses adds r=imm14,r0 instruction witch
> causes
> certain problems especially when operand of the instruction is an descriptor
> offset ( for example mov address@hidden(a) ) .
> I have made a crude fix for this problem:
I don't see any advantage to using adds instead of addl. They schedule
the same, have the same performance, and the docs say we should use addl
always. So I think we should just drop the pseudo instruction entry
that maps mov to adds.
The following patch does this, and seems to give the right result for
your testcase. I now get the desired PLTOFF22 reloc instead of a IMM14
reloc for "mov address@hidden(a)".
--
Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com
patch.mov.adds
Description: Text document