| Message ID | 1422616301-38365-1-git-send-email-dmalcolm@redhat.com |
|---|---|
| State | New |
| Headers | show |
On 30/01/15 11:11, David Malcolm wrote: > libgccjit can make repeated in-process calls to the compiler code. > > On arm I see failures where e.g. the top of the generated .s file has: > on the 1st iteration: > .arch armv7-a > on the 2nd iteration: > .cpu armv7-a > without changing any options. > > What's happening is the 1st time through arm_option_override, > arm_selected_cpu is NULL, but is set to non-NULL at: > arm_selected_cpu = arm_selected_arch; > The 2nd time through, arm_selected_cpu hasn't been reset, and so the > condition: > > if (arm_selected_arch) > { > if (arm_selected_cpu) > { > /* Check for conflict between mcpu and march. */ > > fires, and this clause fires: > > /* -mcpu wins. */ > arm_selected_arch = NULL; > > so we have a NULL arm_selected_arch and a non-NULL arm_selected_cpu. > > The fix is to reset arm-specific state, setting > arm_selected_{arch|cpu|tune} > to NULL at the top of arm_option_override, to prevent state from > previous in-process runs from affecting things. > > With this, and the previous patches, most of the jit testsuite passes > on arm, with the only remaining failures being of the form: > WARNING: Timed out executing test case > on the very large test cases (test-combination.c and test-threads.c). > > OK for stage 4? > > gcc/ChangeLog: > PR jit/64810 > * config/arm/arm.c (arm_option_override): Set > arm_selected_arch/cpu/tune to NULL on entry. OK. R. > --- > gcc/config/arm/arm.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c > index 246298a..c7c0ecf 100644 > --- a/gcc/config/arm/arm.c > +++ b/gcc/config/arm/arm.c > @@ -2627,6 +2627,10 @@ arm_gimplify_va_arg_expr (tree valist, tree type, gimple_seq *pre_p, > static void > arm_option_override (void) > { > + arm_selected_arch = NULL; > + arm_selected_cpu = NULL; > + arm_selected_tune = NULL; > + > if (global_options_set.x_arm_arch_option) > arm_selected_arch = &all_architectures[arm_arch_option]; > >
diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c index 246298a..c7c0ecf 100644 --- a/gcc/config/arm/arm.c +++ b/gcc/config/arm/arm.c @@ -2627,6 +2627,10 @@ arm_gimplify_va_arg_expr (tree valist, tree type, gimple_seq *pre_p, static void arm_option_override (void) { + arm_selected_arch = NULL; + arm_selected_cpu = NULL; + arm_selected_tune = NULL; + if (global_options_set.x_arm_arch_option) arm_selected_arch = &all_architectures[arm_arch_option];
libgccjit can make repeated in-process calls to the compiler code. On arm I see failures where e.g. the top of the generated .s file has: on the 1st iteration: .arch armv7-a on the 2nd iteration: .cpu armv7-a without changing any options. What's happening is the 1st time through arm_option_override, arm_selected_cpu is NULL, but is set to non-NULL at: arm_selected_cpu = arm_selected_arch; The 2nd time through, arm_selected_cpu hasn't been reset, and so the condition: if (arm_selected_arch) { if (arm_selected_cpu) { /* Check for conflict between mcpu and march. */ fires, and this clause fires: /* -mcpu wins. */ arm_selected_arch = NULL; so we have a NULL arm_selected_arch and a non-NULL arm_selected_cpu. The fix is to reset arm-specific state, setting arm_selected_{arch|cpu|tune} to NULL at the top of arm_option_override, to prevent state from previous in-process runs from affecting things. With this, and the previous patches, most of the jit testsuite passes on arm, with the only remaining failures being of the form: WARNING: Timed out executing test case on the very large test cases (test-combination.c and test-threads.c). OK for stage 4? gcc/ChangeLog: PR jit/64810 * config/arm/arm.c (arm_option_override): Set arm_selected_arch/cpu/tune to NULL on entry. --- gcc/config/arm/arm.c | 4 ++++ 1 file changed, 4 insertions(+)