paweld/fpc
1
0
Fork 0
mirror of https://gitlab.com/freepascal.org/fpc/source.git synced 2025-08-31 11:50:51 +02:00
Code Issues Packages Projects Releases Wiki Activity

* explicitly emit a sub instruction in tx86inlinenode.second_high, resolves #36530

Browse Source 
git-svn-id: trunk@43888 -
This commit is contained in:
florian 2020-01-07 21:46:07 +00:00
parent 7a3a4244d1
commit 11287f018f
1 changed files with 5 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
compiler/x86/nx86inl.pas
View File

@ -1288,8 +1288,11 @@ implementation
{ by subtracting 1 here, we get the -1 into the register we need if the dyn. array is nil and the carry
flag is set in this case, so we can jump depending on it
when loading the actual high value, we have to take care later of the decreased value }
hlcg.a_op_const_reg(current_asmdata.CurrAsmList,OP_SUB,left.resultdef,1,left.location.register);
when loading the actual high value, we have to take care later of the decreased value
do not use the cgs, as they might emit dec instead of a sub instruction, however with dec the trick
we are using is not working as dec does not touch the carry flag }
current_asmdata.CurrAsmList.concat(taicpu.op_const_reg(A_SUB,TCGSize2OpSize[def_cgsize(left.resultdef)],1,left.location.register));
{ volatility of the dyn. array refers to the volatility of the
string pointer, not of the string data }
cg.a_jmp_flags(current_asmdata.CurrAsmList,F_C,donelab);