Moreover, GCC10 complains about gimple definition. For example,
doing a 'scripts/gcc-plugin.sh g++ g++ gcc' returns:
In file included from <stdin>:1:
./gcc-plugins/gcc-common.h:852:13: error: redefinition of ‘static bool is_a_helper<T>::test(U*) [with U = const gimple; T = const ggoto*]’
852 | inline bool is_a_helper<const ggoto *>::test(const_gimple gs)
| ^~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ./gcc-plugins/gcc-common.h:125,
from <stdin>:1:
/usr/lib/gcc/x86_64-redhat-linux/10/plugin/include/gimple.h:1037:1: note: ‘static bool is_a_helper<T>::test(U*) [with U = const gimple; T = const ggoto*]’ previously declared here
1037 | is_a_helper <const ggoto *>::test (const gimple *gs)
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from <stdin>:1:
./gcc-plugins/gcc-common.h:859:13: error: redefinition of ‘static bool is_a_helper<T>::test(U*) [with U = const gimple; T = const greturn*]’
859 | inline bool is_a_helper<const greturn *>::test(const_gimple gs)
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ./gcc-plugins/gcc-common.h:125,
from <stdin>:1:
/usr/lib/gcc/x86_64-redhat-linux/10/plugin/include/gimple.h:1489:1: note: ‘static bool is_a_helper<T>::test(U*) [with U = const gimple; T = const greturn*]’ previously declared here
1489 | is_a_helper <const greturn *>::test (const gimple *gs)
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
A hacky way for solving this is to ignore them for GCC10.
Signed-off-by: Frédéric Pierret (fepitre) <[email protected]>
---
scripts/gcc-plugins/gcc-common.h | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/scripts/gcc-plugins/gcc-common.h b/scripts/gcc-plugins/gcc-common.h
index 17f06079a712..9ad76b7f3f10 100644
--- a/scripts/gcc-plugins/gcc-common.h
+++ b/scripts/gcc-plugins/gcc-common.h
@@ -35,7 +35,9 @@
#include "ggc.h"
#include "timevar.h"
+#if BUILDING_GCC_VERSION < 10000
#include "params.h"
+#endif
#if BUILDING_GCC_VERSION <= 4009
#include "pointer-set.h"
@@ -847,6 +849,7 @@ static inline gimple gimple_build_assign_with_ops(enum tree_code subcode, tree l
return gimple_build_assign(lhs, subcode, op1, op2 PASS_MEM_STAT);
}
+#if BUILDING_GCC_VERSION < 10000
template <>
template <>
inline bool is_a_helper<const ggoto *>::test(const_gimple gs)
@@ -860,6 +863,7 @@ inline bool is_a_helper<const greturn *>::test(const_gimple gs)
{
return gs->code == GIMPLE_RETURN;
}
+#endif
static inline gasm *as_a_gasm(gimple stmt)
{
--
2.25.2
On Tue, Apr 07, 2020 at 01:32:59PM +0200, Frédéric Pierret (fepitre) wrote:
> Moreover, GCC10 complains about gimple definition. For example,
> doing a 'scripts/gcc-plugin.sh g++ g++ gcc' returns:
>
> In file included from <stdin>:1:
> ./gcc-plugins/gcc-common.h:852:13: error: redefinition of ‘static bool is_a_helper<T>::test(U*) [with U = const gimple; T = const ggoto*]’
> 852 | inline bool is_a_helper<const ggoto *>::test(const_gimple gs)
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~
> In file included from ./gcc-plugins/gcc-common.h:125,
> from <stdin>:1:
> /usr/lib/gcc/x86_64-redhat-linux/10/plugin/include/gimple.h:1037:1: note: ‘static bool is_a_helper<T>::test(U*) [with U = const gimple; T = const ggoto*]’ previously declared here
> 1037 | is_a_helper <const ggoto *>::test (const gimple *gs)
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~
> In file included from <stdin>:1:
> ./gcc-plugins/gcc-common.h:859:13: error: redefinition of ‘static bool is_a_helper<T>::test(U*) [with U = const gimple; T = const greturn*]’
> 859 | inline bool is_a_helper<const greturn *>::test(const_gimple gs)
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
> In file included from ./gcc-plugins/gcc-common.h:125,
> from <stdin>:1:
> /usr/lib/gcc/x86_64-redhat-linux/10/plugin/include/gimple.h:1489:1: note: ‘static bool is_a_helper<T>::test(U*) [with U = const gimple; T = const greturn*]’ previously declared here
> 1489 | is_a_helper <const greturn *>::test (const gimple *gs)
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> A hacky way for solving this is to ignore them for GCC10.
Hi! Thanks for the patch. I don't think this is a hack: it's the right
thing to do here, yes? GCC 10 includes this helper in gimple.h, so we
can ifdef it out in gcc-common.h.
-Kees
>
> Signed-off-by: Frédéric Pierret (fepitre) <[email protected]>
> ---
> scripts/gcc-plugins/gcc-common.h | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/scripts/gcc-plugins/gcc-common.h b/scripts/gcc-plugins/gcc-common.h
> index 17f06079a712..9ad76b7f3f10 100644
> --- a/scripts/gcc-plugins/gcc-common.h
> +++ b/scripts/gcc-plugins/gcc-common.h
> @@ -35,7 +35,9 @@
> #include "ggc.h"
> #include "timevar.h"
>
> +#if BUILDING_GCC_VERSION < 10000
> #include "params.h"
> +#endif
>
> #if BUILDING_GCC_VERSION <= 4009
> #include "pointer-set.h"
> @@ -847,6 +849,7 @@ static inline gimple gimple_build_assign_with_ops(enum tree_code subcode, tree l
> return gimple_build_assign(lhs, subcode, op1, op2 PASS_MEM_STAT);
> }
>
> +#if BUILDING_GCC_VERSION < 10000
> template <>
> template <>
> inline bool is_a_helper<const ggoto *>::test(const_gimple gs)
> @@ -860,6 +863,7 @@ inline bool is_a_helper<const greturn *>::test(const_gimple gs)
> {
> return gs->code == GIMPLE_RETURN;
> }
> +#endif
>
> static inline gasm *as_a_gasm(gimple stmt)
> {
> --
> 2.25.2
>
>
--
Kees Cook
On 2020-04-07 18:45, Kees Cook wrote:
>
> Hi! Thanks for the patch. I don't think this is a hack: it's the right
> thing to do here, yes? GCC 10 includes this helper in gimple.h, so we
> can ifdef it out in gcc-common.h.
>
> -Kees
Hi Kees,
Thank you very much for your comment. Would you like me to rephrase the commit including your comment too? "Hacky" mostly meaning humble modification from my point of view :)
Best regards,
Frédéric
On Tue, Apr 07, 2020 at 07:22:55PM +0200, Fr?d?ric Pierret wrote:
>
>
> On 2020-04-07 18:45, Kees Cook wrote:
> >
> > Hi! Thanks for the patch. I don't think this is a hack: it's the right
> > thing to do here, yes? GCC 10 includes this helper in gimple.h, so we
> > can ifdef it out in gcc-common.h.
> >
> > -Kees
> Hi Kees,
> Thank you very much for your comment. Would you like me to rephrase the commit including your comment too? "Hacky" mostly meaning humble modification from my point of view :)
Heh, no worries. I've just reproduced the failure you found with gcc 10,
and I've updated your commit log (and added -Wno-format-diag to the plug
builds). Here's what I've got in my tree now:
https://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git/commit/?h=for-next/gcc-plugins&id=dda632f1bc6da784baab8069e26547e3f4144dbe
Thanks for the patch!
-Kees
--
Kees Cook