Message ID | 1310636503-8956-5-git-send-email-tsnsaito@gmail.com |
---|---|
State | New |
Headers | show |
On 14 July 2011 10:41, Tsuneo Saito <tsnsaito@gmail.com> wrote: > stfa/stdfa/stqfa instructions should raise fp_disabled exceptions > if %pstate.PEF==0 or %fprs.FEF==0. Not really related to this patch but I happened to notice it: we generate a trap if dc->fpu_enabled is clear. That flag is set with: dc->fpu_enabled = cpu_fpu_enabled(env); which looks at the CPUState passed into gen_intermediate_code_internal. I think it should be pulling the fpu-enabled state out of the tb->flags instead. Otherwise you'll have problems if you have a TB with (1) a trap-if-no-fpu insn (2) a change of the fpu-enabled state (3) a load/store that faults. -- PMM
On Thu, Jul 14, 2011 at 1:39 PM, Peter Maydell <peter.maydell@linaro.org> wrote: > On 14 July 2011 10:41, Tsuneo Saito <tsnsaito@gmail.com> wrote: >> stfa/stdfa/stqfa instructions should raise fp_disabled exceptions >> if %pstate.PEF==0 or %fprs.FEF==0. > > Not really related to this patch but I happened to notice it: > we generate a trap if dc->fpu_enabled is clear. That flag is set with: > dc->fpu_enabled = cpu_fpu_enabled(env); > which looks at the CPUState passed into gen_intermediate_code_internal. > > I think it should be pulling the fpu-enabled state out of the > tb->flags instead. Otherwise you'll have problems if you have > a TB with (1) a trap-if-no-fpu insn (2) a change of the fpu-enabled > state (3) a load/store that faults. Nice catch, this also affects Sparc32.
On Thu, Jul 14, 2011 at 1:39 PM, Peter Maydell <peter.maydell@linaro.org> wrote: > On 14 July 2011 10:41, Tsuneo Saito <tsnsaito@gmail.com> wrote: >> stfa/stdfa/stqfa instructions should raise fp_disabled exceptions >> if %pstate.PEF==0 or %fprs.FEF==0. > > Not really related to this patch but I happened to notice it: > we generate a trap if dc->fpu_enabled is clear. That flag is set with: > dc->fpu_enabled = cpu_fpu_enabled(env); > which looks at the CPUState passed into gen_intermediate_code_internal. > > I think it should be pulling the fpu-enabled state out of the > tb->flags instead. Otherwise you'll have problems if you have > a TB with (1) a trap-if-no-fpu insn (2) a change of the fpu-enabled > state (3) a load/store that faults. I applied a patch to fix this and similar bug with AM bit.
diff --git a/target-sparc/translate.c b/target-sparc/translate.c index a5a8eaf..5f92dbb 100644 --- a/target-sparc/translate.c +++ b/target-sparc/translate.c @@ -4732,6 +4732,9 @@ static void disas_sparc_insn(DisasContext * dc) switch (xop) { #ifdef TARGET_SPARC64 case 0x34: /* V9 stfa */ + if (gen_trap_ifnofpu(dc, cpu_cond)) { + goto jmp_insn; + } gen_stf_asi(cpu_addr, insn, 4, rd); break; case 0x36: /* V9 stqfa */ @@ -4739,6 +4742,9 @@ static void disas_sparc_insn(DisasContext * dc) TCGv_i32 r_const; CHECK_FPU_FEATURE(dc, FLOAT128); + if (gen_trap_ifnofpu(dc, cpu_cond)) { + goto jmp_insn; + } r_const = tcg_const_i32(7); gen_helper_check_align(cpu_addr, r_const); tcg_temp_free_i32(r_const); @@ -4746,6 +4752,9 @@ static void disas_sparc_insn(DisasContext * dc) } break; case 0x37: /* V9 stdfa */ + if (gen_trap_ifnofpu(dc, cpu_cond)) { + goto jmp_insn; + } gen_stf_asi(cpu_addr, insn, 8, DFPREG(rd)); break; case 0x3c: /* V9 casa */
stfa/stdfa/stqfa instructions should raise fp_disabled exceptions if %pstate.PEF==0 or %fprs.FEF==0. Signed-off-by: Tsuneo Saito <tsnsaito@gmail.com> --- target-sparc/translate.c | 9 +++++++++ 1 files changed, 9 insertions(+), 0 deletions(-)