diff mbox series

[2/2] ltp-testsuite: drop redundant arch dependency entries

Message ID 2f61c63d53f8ba4ea0b65942b57f67a156fbf778.1516964598.git.baruch@tkos.co.il
State Accepted
Commit eb2bb6648fa09a7b59973a9f889e74d098058bd0
Headers show
Series [1/2] ltp-testsuite: disable numa tests | expand

Commit Message

Baruch Siach Jan. 26, 2018, 11:03 a.m. UTC
BR2_TOOLCHAIN_HAS_SYNC_4 is enabled for all these architectures except
arc. There is no need to list them explicitly.

Cc: Romain Naour <romain.naour@gmail.com>
Cc: Petr Vorel <petr.vorel@gmail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
---
 package/ltp-testsuite/Config.in | 5 -----
 1 file changed, 5 deletions(-)

Comments

Petr Vorel Jan. 26, 2018, 11:35 a.m. UTC | #1
> BR2_TOOLCHAIN_HAS_SYNC_4 is enabled for all these architectures except
> arc. There is no need to list them explicitly.

> Cc: Romain Naour <romain.naour@gmail.com>
> Cc: Petr Vorel <petr.vorel@gmail.com>
> Signed-off-by: Baruch Siach <baruch@tkos.co.il>
> ---
>  package/ltp-testsuite/Config.in | 5 -----
>  1 file changed, 5 deletions(-)

> diff --git a/package/ltp-testsuite/Config.in b/package/ltp-testsuite/Config.in
> index 4c00e40f330e..711c859e6caf 100644
> --- a/package/ltp-testsuite/Config.in
> +++ b/package/ltp-testsuite/Config.in
> @@ -5,11 +5,6 @@ config BR2_PACKAGE_LTP_TESTSUITE_ARCH_SUPPORTS
>  	# provided in ltp-testsuite
>  	default y if BR2_TOOLCHAIN_HAS_SYNC_4
>  	default y if BR2_arc
> -	default y if BR2_i386
> -	default y if BR2_powerpc
> -	default y if BR2_powepc64
> -	default y if BR2_powerpc64le
> -	default y if BR2_x86_64
>  	# does not build on nios2, cachectl.h issue
>  	depends on !BR2_nios2
Acked-by: Petr Vorel <petr.vorel@gmail.com>


Kind regards,
Petr
Peter Korsgaard Jan. 27, 2018, 10:17 a.m. UTC | #2
>>>>> "Baruch" == Baruch Siach <baruch@tkos.co.il> writes:

 > BR2_TOOLCHAIN_HAS_SYNC_4 is enabled for all these architectures except
 > arc. There is no need to list them explicitly.

 > Cc: Romain Naour <romain.naour@gmail.com>
 > Cc: Petr Vorel <petr.vorel@gmail.com>
 > Signed-off-by: Baruch Siach <baruch@tkos.co.il>

Committed, thanks.
Thomas Petazzoni Jan. 27, 2018, 10:07 p.m. UTC | #3
Hello,

On Fri, 26 Jan 2018 13:03:18 +0200, Baruch Siach wrote:
> BR2_TOOLCHAIN_HAS_SYNC_4 is enabled for all these architectures except
> arc. There is no need to list them explicitly.
> 
> Cc: Romain Naour <romain.naour@gmail.com>
> Cc: Petr Vorel <petr.vorel@gmail.com>
> Signed-off-by: Baruch Siach <baruch@tkos.co.il>
> ---
>  package/ltp-testsuite/Config.in | 5 -----
>  1 file changed, 5 deletions(-)
> 
> diff --git a/package/ltp-testsuite/Config.in b/package/ltp-testsuite/Config.in
> index 4c00e40f330e..711c859e6caf 100644
> --- a/package/ltp-testsuite/Config.in
> +++ b/package/ltp-testsuite/Config.in
> @@ -5,11 +5,6 @@ config BR2_PACKAGE_LTP_TESTSUITE_ARCH_SUPPORTS
>  	# provided in ltp-testsuite
>  	default y if BR2_TOOLCHAIN_HAS_SYNC_4
>  	default y if BR2_arc
> -	default y if BR2_i386
> -	default y if BR2_powerpc
> -	default y if BR2_powepc64
> -	default y if BR2_powerpc64le
> -	default y if BR2_x86_64

Is this really a correct patch ? It's not because
BR2_TOOLCHAIN_HAS_SYNC_4 already covers
i386/powerpc/powerpc64/powerpc64le/x86_64 that it makes sense to use
BR2_TOOLCHAIN_HAS_SYNC_4 instead.

What if an architecture gets added to BR2_TOOLCHAIN_HAS_SYNC_4, but
isn't supported by ltp-testsuite ?

Best regards,

Thomas
Baruch Siach Jan. 28, 2018, 3:24 a.m. UTC | #4
Hi Thomas,

On Sat, Jan 27, 2018 at 11:07:07PM +0100, Thomas Petazzoni wrote:
> On Fri, 26 Jan 2018 13:03:18 +0200, Baruch Siach wrote:
> > BR2_TOOLCHAIN_HAS_SYNC_4 is enabled for all these architectures except
> > arc. There is no need to list them explicitly.
> > 
> > Cc: Romain Naour <romain.naour@gmail.com>
> > Cc: Petr Vorel <petr.vorel@gmail.com>
> > Signed-off-by: Baruch Siach <baruch@tkos.co.il>
> > ---
> >  package/ltp-testsuite/Config.in | 5 -----
> >  1 file changed, 5 deletions(-)
> > 
> > diff --git a/package/ltp-testsuite/Config.in b/package/ltp-testsuite/Config.in
> > index 4c00e40f330e..711c859e6caf 100644
> > --- a/package/ltp-testsuite/Config.in
> > +++ b/package/ltp-testsuite/Config.in
> > @@ -5,11 +5,6 @@ config BR2_PACKAGE_LTP_TESTSUITE_ARCH_SUPPORTS
> >  	# provided in ltp-testsuite
> >  	default y if BR2_TOOLCHAIN_HAS_SYNC_4
> >  	default y if BR2_arc
> > -	default y if BR2_i386
> > -	default y if BR2_powerpc
> > -	default y if BR2_powepc64
> > -	default y if BR2_powerpc64le
> > -	default y if BR2_x86_64
> 
> Is this really a correct patch ? It's not because
> BR2_TOOLCHAIN_HAS_SYNC_4 already covers
> i386/powerpc/powerpc64/powerpc64le/x86_64 that it makes sense to use
> BR2_TOOLCHAIN_HAS_SYNC_4 instead.
> 
> What if an architecture gets added to BR2_TOOLCHAIN_HAS_SYNC_4, but
> isn't supported by ltp-testsuite ?

My understanding, based on the commit log of 583f59702bdd 
(package/ltp-testsuite: refine architecture dependency), is that when the 
toolchain implements __sync_add_and_fetch() there is no need for the 
architecture specific fallback implementation that ltp-testsuite provides. 
This is the case already for e.g. ARM. So as long as the added architecture is 
covered by BR2_TOOLCHAIN_HAS_SYNC_4, we should be fine.

baruch
diff mbox series

Patch

diff --git a/package/ltp-testsuite/Config.in b/package/ltp-testsuite/Config.in
index 4c00e40f330e..711c859e6caf 100644
--- a/package/ltp-testsuite/Config.in
+++ b/package/ltp-testsuite/Config.in
@@ -5,11 +5,6 @@  config BR2_PACKAGE_LTP_TESTSUITE_ARCH_SUPPORTS
 	# provided in ltp-testsuite
 	default y if BR2_TOOLCHAIN_HAS_SYNC_4
 	default y if BR2_arc
-	default y if BR2_i386
-	default y if BR2_powerpc
-	default y if BR2_powepc64
-	default y if BR2_powerpc64le
-	default y if BR2_x86_64
 	# does not build on nios2, cachectl.h issue
 	depends on !BR2_nios2