[llvm] r225941 - X86: validate 'int' instruction
craig.topper at gmail.com
Wed Jan 14 00:56:46 PST 2015
I agree. I've wanted to start looking into this, but it gets into a little
trouble because the gnu assember throws a truncation warning instead of an
error on some cases. We don't have a great way to handle that in the MC
On Tue, Jan 13, 2015 at 11:58 PM, Joerg Sonnenberger <
joerg at britannica.bec.de> wrote:
> On Wed, Jan 14, 2015 at 05:10:21AM -0000, Saleem Abdulrasool wrote:
> > Author: compnerd
> > Date: Tue Jan 13 23:10:21 2015
> > New Revision: 225941
> > URL: http://llvm.org/viewvc/llvm-project?rev=225941&view=rev
> > Log:
> > X86: validate 'int' instruction
> > The int instruction takes as an operand an 8-bit immediate value.
> Validate that
> > the input is valid rather than silently truncating the value.
> I think this is the wrong approach. We don't want to handle every case
> of i8imm specially, but should check the immediates in one place. The
> problem occurs with pretty much any instruction that has immediates.
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the llvm-commits