T0b1/wasmtime
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

x64: fix CvtFloatToUintSeq: do not clobber src. (#4842)

Browse Source 
This slipped through the regalloc2 operand code update in #4811: the
CvtFloatToUintSeq pseudo-instruction actually clobbers its source. It
was marked as a "mod" operand in the original and I mistakenly
converted it to a "use" as I had not seen the actual clobber. The
instruction now takes an extra temp and makes a copy of `src` in the
appropriate place.

Fixes #4840.
This commit is contained in:
Chris Fallin
2022-09-01 15:46:57 -07:00
committed by GitHub
parent 08e7a7f1a0
commit 385bd0cbf8
5 changed files with 51 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
cranelift/codegen/src/isa/x64/inst.isle
View File

@@ -289,7 +289,8 @@
(src Xmm)
(dst WritableGpr)
(tmp_gpr WritableGpr)
(tmp_xmm WritableXmm))
(tmp_xmm WritableXmm)
(tmp_xmm2 WritableXmm))
;; A sequence to compute min/max with the proper NaN semantics for xmm
;; registers.
@@ -3213,8 +3214,9 @@
(dst WritableGpr (temp_writable_gpr))
(tmp_xmm WritableXmm (temp_writable_xmm))
(tmp_xmm2 WritableXmm (temp_writable_xmm))
(tmp_gpr WritableGpr (temp_writable_gpr))
(_ Unit (emit (MInst.CvtFloatToUintSeq out_size src_size is_saturating src dst tmp_gpr tmp_xmm))))
(_ Unit (emit (MInst.CvtFloatToUintSeq out_size src_size is_saturating src dst tmp_gpr tmp_xmm tmp_xmm2))))
dst))
(decl cvt_float_to_sint_seq (Type Value bool) Gpr)