diff mbox

[testsuite] Skip gcc.dg/pic-2.c and gcc.dg/pie-2.c for MIPS.

Message ID A614194ED15B4844BC4C9FB7F21FCD9270457D3D@hhmail02.hh.imgtec.org
State New
Headers show

Commit Message

Toma Tabacu March 15, 2017, 3:26 p.m. UTC
Hi,

The gcc.dg/pic-2.c and gcc.dg/pie-2.c tests are failing for MIPS targets
because __PIC__ is always set to 1 for MIPS.

This patch makes the testsuite skip those two tests for all MIPS targets.

Tested with mips-mti-elf and mips-mti-linux-gnu.

Should I have fixed this in target-supports.exp instead ?
I was worried that doing so would complicate the fpic and pie effective target
checks too much.

Regards,
Toma

gcc/testsuite/

	* gcc.dg/pic-2.c: Skip for MIPS.
	* gcc.dg/pie-2.c: Skip for MIPS.

Comments

Matthew Fortune March 15, 2017, 3:36 p.m. UTC | #1
Toma Tabacu <Toma.Tabacu@imgtec.com> writes:
> The gcc.dg/pic-2.c and gcc.dg/pie-2.c tests are failing for MIPS targets
> because __PIC__ is always set to 1 for MIPS.
> 
> This patch makes the testsuite skip those two tests for all MIPS
> targets.
> 
> Tested with mips-mti-elf and mips-mti-linux-gnu.
> 
> Should I have fixed this in target-supports.exp instead ?
> I was worried that doing so would complicate the fpic and pie effective
> target checks too much.

I think the skip is OK here. I'd like to get Catherine's opinion on
this though too. I don't think we should change the definition of __PIC__
for -fPIC on MIPS as multi-got solves 'most' issues. If we start trying to
figure out what __PIC__ should mean on MIPS then we will get into a big
mess with -mxgot as that is arguably __PIC__==2 but I expect there will be
several differing opinions.

Thanks,
Matthew

> 
> Regards,
> Toma
> 
> gcc/testsuite/
> 
> 	* gcc.dg/pic-2.c: Skip for MIPS.
> 	* gcc.dg/pie-2.c: Skip for MIPS.
> 
> diff --git a/gcc/testsuite/gcc.dg/pic-2.c b/gcc/testsuite/gcc.dg/pic-2.c
> index 59ce8e2..bccec13 100644
> --- a/gcc/testsuite/gcc.dg/pic-2.c
> +++ b/gcc/testsuite/gcc.dg/pic-2.c
> @@ -1,6 +1,7 @@
>  /* { dg-do compile } */
>  /* { dg-require-effective-target fpic } */
>  /* { dg-options "-fPIC" } */
> +/* { dg-skip-if "__PIC__ is always 1 for MIPS" { mips*-*-* } } */
> 
>  #if __PIC__ != 2
>  # error __PIC__ is not 2!
> diff --git a/gcc/testsuite/gcc.dg/pie-2.c b/gcc/testsuite/gcc.dg/pie-2.c
> index 7bdc4ac..1838745 100644
> --- a/gcc/testsuite/gcc.dg/pie-2.c
> +++ b/gcc/testsuite/gcc.dg/pie-2.c
> @@ -1,6 +1,7 @@
>  /* { dg-do compile } */
>  /* { dg-options "-fPIE" } */
>  /* { dg-require-effective-target pie } */
> +/* { dg-skip-if "__PIC__ is always 1 for MIPS" { mips*-*-* } } */
> 
>  #if __PIC__ != 2
>  # error __PIC__ is not 2!
Moore, Catherine March 20, 2017, 9:10 p.m. UTC | #2
> -----Original Message-----
> From: Matthew Fortune [mailto:Matthew.Fortune@imgtec.com]
> Sent: Wednesday, March 15, 2017 11:37 AM
> To: Moore, Catherine <Catherine_Moore@mentor.com>
> Cc: Toma Tabacu <Toma.Tabacu@imgtec.com>; gcc-
> patches@gcc.gnu.org
> Subject: RE: [PATCH,testsuite] Skip gcc.dg/pic-2.c and gcc.dg/pie-2.c for
> MIPS.
> 
> Toma Tabacu <Toma.Tabacu@imgtec.com> writes:
> > The gcc.dg/pic-2.c and gcc.dg/pie-2.c tests are failing for MIPS targets
> > because __PIC__ is always set to 1 for MIPS.
> >
> > This patch makes the testsuite skip those two tests for all MIPS
> > targets.
> >
> > Tested with mips-mti-elf and mips-mti-linux-gnu.
> >
> > Should I have fixed this in target-supports.exp instead ?
> > I was worried that doing so would complicate the fpic and pie effective
> > target checks too much.
> 
> I think the skip is OK here. I'd like to get Catherine's opinion on
> this though too. I don't think we should change the definition of __PIC__
> for -fPIC on MIPS as multi-got solves 'most' issues. If we start trying to
> figure out what __PIC__ should mean on MIPS then we will get into a big
> mess with -mxgot as that is arguably __PIC__==2 but I expect there will
> be  several differing opinions.
> 
I think the skip is the right way to go as well.   The patch is OK with me.
Catherine
Toma Tabacu March 21, 2017, 11:09 a.m. UTC | #3
> > From: Matthew Fortune
> >
> > I think the skip is OK here. I'd like to get Catherine's opinion on
> > this though too. I don't think we should change the definition of __PIC__
> > for -fPIC on MIPS as multi-got solves 'most' issues. If we start trying to
> > figure out what __PIC__ should mean on MIPS then we will get into a big
> > mess with -mxgot as that is arguably __PIC__==2 but I expect there will
> > be  several differing opinions.
> >
> From: Catherine Moore
>
> I think the skip is the right way to go as well.   The patch is OK with me.
> Catherine
> 

Committed as r246311.

Thanks,
Toma
diff mbox

Patch

diff --git a/gcc/testsuite/gcc.dg/pic-2.c b/gcc/testsuite/gcc.dg/pic-2.c
index 59ce8e2..bccec13 100644
--- a/gcc/testsuite/gcc.dg/pic-2.c
+++ b/gcc/testsuite/gcc.dg/pic-2.c
@@ -1,6 +1,7 @@ 
 /* { dg-do compile } */
 /* { dg-require-effective-target fpic } */
 /* { dg-options "-fPIC" } */
+/* { dg-skip-if "__PIC__ is always 1 for MIPS" { mips*-*-* } } */
 
 #if __PIC__ != 2
 # error __PIC__ is not 2!
diff --git a/gcc/testsuite/gcc.dg/pie-2.c b/gcc/testsuite/gcc.dg/pie-2.c
index 7bdc4ac..1838745 100644
--- a/gcc/testsuite/gcc.dg/pie-2.c
+++ b/gcc/testsuite/gcc.dg/pie-2.c
@@ -1,6 +1,7 @@ 
 /* { dg-do compile } */
 /* { dg-options "-fPIE" } */
 /* { dg-require-effective-target pie } */
+/* { dg-skip-if "__PIC__ is always 1 for MIPS" { mips*-*-* } } */
 
 #if __PIC__ != 2
 # error __PIC__ is not 2!