TRegNameTable is defined in compiler/rgbase.pas and is an array of
strings, limited to the maximum length of the used register names.
r22792 added a long register name but did not scale the string-size
enough, resulting in the compiler built breaking for arm.
git-svn-id: trunk@22817 -
RS_INVALID superregister (instead of sometimes RS_NO and sometimes
RS_INVALID)
* check for RS_INVALID in tcg.g_save_registers() and ignore such entries
git-svn-id: trunk@21622 -
depend on that (and correct a number of cases where this was wrong)
* set the correct subreg type for xmm x86_64 parameter registers
(resolved mantis #14067)
git-svn-id: trunk@13410 -
+ added ppc740 (g3), ppc7400 (G4) and ppc970 (G5) as ppc32 cpu
targets
* initialise optimizecputype by default to ppc7400 for ppc32 and to
ppc970 for ppc64
* merged ppc32/ppc64 overflow checking code and use the ppc64 one
in case cputype or optimizecputype >= ppc970, because one of
the instructions used in the ppc32 version no longer exists on the
ppc970 (although it's emulated in the kernel on at least Mac OS X)
* moved some other support routines and constants to ppcgen which
were needed for the overflow checking (were identical for ppc32 and
ppc64)
git-svn-id: trunk@6323 -
was not being initialised -> getfinaldestination always failed, which
caused wrong optimizations in some cases)
* changed the inverse_cond into a function, because tasmcond is a record
on ppc
+ added a compare_conditions() function for the same reason
+ refaddr to treference added
+ refsymbol to treference added
* top_local stuff moved to an extra record to save memory
+ aint introduced
* tppufile.get/putint64/aint implemented
ppc, other processors should however only require maxregvars and
maxfpuregvars constants in cpubase.pas. Remember to take scratch-
registers into account when defining that value.
* fixed initialisation of fpscr register to avoid spurious SIGPFE's
(uses mtfsb0 instruction, so added extra define in options.pas to avoid
requiring to start with a cross compiler)
* fixed overflow flag definition
* fixed cg.g_overflowcheck() for signed numbers (jump over call to
FPC_OVERFLOW if *no* overflow instead of if overflow :)
