← Back to branch summary

~ubuntu-branches/ubuntu/maverick/clamav/maverick-backports

« back to all changes in this revision

Viewing changes to libclamav/c++/llvm/test/CodeGen/X86/overlap-shift.ll

  • Committer: Bazaar Package Importer
  • Author(s): Stephen Gran, Stephen Gran, Michael Tautschnig
  • Date: 2010-04-26 21:41:18 UTC
  • mfrom: (2.1.6 squeeze)
  • Revision ID: james.westby@ubuntu.com-20100426214118-i6lo606wnh7ywfj6
Tags: 0.96+dfsg-4
http://bugs.debian.org/579271
http://bugs.debian.org/579068
[ Stephen Gran ]
* Fixed typo in clamav-milter's postinst

[ Michael Tautschnig ]
* Fixed typo in clamav-freshclam's postinst (closes: #579271)
* Debconf translation updates
  - Portuguese (closes: #579068)

Show diffs side-by-side

added added

removed removed

Lines of Context:
libclamav/c++/llvm/test/CodeGen/X86/overlap-shift.ll
 
1
;; X's live range extends beyond the shift, so the register allocator
 
2
;; cannot coalesce it with Y.  Because of this, a copy needs to be
 
3
;; emitted before the shift to save the register value before it is
 
4
;; clobbered.  However, this copy is not needed if the register
 
5
;; allocator turns the shift into an LEA.  This also occurs for ADD.
 
6
 
 
7
; Check that the shift gets turned into an LEA.
 
8
 
 
9
; RUN: llc < %s -march=x86 -x86-asm-syntax=intel | \
 
10
; RUN:   not grep {mov E.X, E.X}
 
11
 
 
12
@G = external global i32                ; <i32*> [#uses=1]
 
13
 
 
14
define i32 @test1(i32 %X) {
 
15
        %Z = shl i32 %X, 2              ; <i32> [#uses=1]
 
16
        volatile store i32 %Z, i32* @G
 
17
        ret i32 %X
 
18
}
 
19