public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856]
@ 2021-07-15 13:04 Christophe LYON
  2021-08-04  9:11 ` Christophe Lyon
  2021-08-23  9:03 ` Kyrylo Tkachov
  0 siblings, 2 replies; 6+ messages in thread
From: Christophe LYON @ 2021-07-15 13:04 UTC (permalink / raw)
  To: gcc Patches

[-- Attachment #1: Type: text/plain, Size: 1943 bytes --]

This is a followup to Srinath's recent patch: the newly added test is
failing e.g. on arm-linux-gnueabihf without R/M profile multilibs.

It is also failing on arm-eabi with R/M profile multilibs if the
execution engine does not support v8.1-M instructions.

The patch avoids this by adding check_effective_target_FUNC_multilib
in target-supports.exp which effectively checks whether the target
supports linking and execution, like what is already done for other
ARM effective targets.  pr100856.c is updated to use it instead of
arm_v8_1m_main_cde_mve_ok (which makes the testcase a bit of a
duplicate with check_effective_target_FUNC_multilib).

In addition, I noticed that requiring MVE does not seem necessary and
this enables the test to pass even when targeting a CPU without MVE:
since the test does not involve actual CDE instructions, it can pass
on other architecture versions.  For instance, when requiring MVE, we
have to use cortex-m55 under QEMU for the test to pass because the
memset() that comes from v8.1-m.main+mve multilib uses LOB
instructions (DLS) (memset is used during startup).  Keeping
arm_v8_1m_main_cde_mve_ok would mean we would enable the test provided
we have the right multilibs, causing a runtime error if the simulator
does not support LOB instructions (e.g. when targeting cortex-m7).

I do not update sourcebuild.texi since the CDE effective targets are
already collectively documented.

Finally, the patch fixes two typos in comments.

2021-07-15  Christophe Lyon  <christophe.lyon@foss.st.com>

         PR target/100856
         gcc/
         * config/arm/arm.opt: Fix typo.
         * config/arm/t-rmprofile: Fix typo.

         gcc/testsuite/
         * gcc.target/arm/acle/pr100856.c: Use arm_v8m_main_cde_multilib
         and arm_v8m_main_cde.
         * lib/target-supports.exp: Add 
check_effective_target_FUNC_multilib for ARM CDE.



[-- Attachment #2: 0001-arm-Fix-multilib-mapping-for-CDE-extensions-PR100856.patch --]
[-- Type: text/plain, Size: 4790 bytes --]

From baa9ed42d986dd2569697ac8903b3ca70ad73bb9 Mon Sep 17 00:00:00 2001
From: Christophe Lyon <christophe.lyon@foss.st.com>
Date: Thu, 15 Jul 2021 12:57:18 +0000
Subject: [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856]

This is a followup to Srinath's recent patch: the newly added test is
failing e.g. on arm-linux-gnueabihf without R/M profile multilibs.

It is also failing on arm-eabi with R/M profile multilibs if the
execution engine does not support v8.1-M instructions.

The patch avoids this by adding check_effective_target_FUNC_multilib
in target-supports.exp which effectively checks whether the target
supports linking and execution, like what is already done for other
ARM effective targets.  pr100856.c is updated to use it instead of
arm_v8_1m_main_cde_mve_ok (which makes the testcase a bit of a
duplicate with check_effective_target_FUNC_multilib).

In addition, I noticed that requiring MVE does not seem necessary and
this enables the test to pass even when targeting a CPU without MVE:
since the test does not involve actual CDE instructions, it can pass
on other architecture versions.  For instance, when requiring MVE, we
have to use cortex-m55 under QEMU for the test to pass because the
memset() that comes from v8.1-m.main+mve multilib uses LOB
instructions (DLS) (memset is used during startup).  Keeping
arm_v8_1m_main_cde_mve_ok would mean we would enable the test provided
we have the right multilibs, causing a runtime error if the simulator
does not support LOB instructions (e.g. when targeting cortex-m7).

I do not update sourcebuild.texi since the CDE effective targets are
already collectively documented.

Finally, the patch fixes two typos in comments.

2021-07-15  Christophe Lyon  <christophe.lyon@foss.st.com>

	PR target/100856
	gcc/
	* config/arm/arm.opt: Fix typo.
	* config/arm/t-rmprofile: Fix typo.

	gcc/testsuite/
	* gcc.target/arm/acle/pr100856.c: Use arm_v8m_main_cde_multilib
	and arm_v8m_main_cde.
	* lib/target-supports.exp: Add
	check_effective_target_FUNC_multilib for ARM CDE.
---
 gcc/config/arm/arm.opt                       |  2 +-
 gcc/config/arm/t-rmprofile                   |  2 +-
 gcc/testsuite/gcc.target/arm/acle/pr100856.c |  4 ++--
 gcc/testsuite/lib/target-supports.exp        | 18 ++++++++++++++++++
 4 files changed, 22 insertions(+), 4 deletions(-)

diff --git a/gcc/config/arm/arm.opt b/gcc/config/arm/arm.opt
index af478a946b2..7417b55122a 100644
--- a/gcc/config/arm/arm.opt
+++ b/gcc/config/arm/arm.opt
@@ -82,7 +82,7 @@ EnumValue
 Enum(arm_arch) String(native) Value(-1) DriverOnly
 
 ; Set to the name of target architecture which is required for
-; multilib linking.  This option is undocumented becuase it
+; multilib linking.  This option is undocumented because it
 ; should not be used by the users.
 mlibarch=
 Target RejectNegative JoinedOrMissing NoDWARFRecord DriverOnly Undocumented
diff --git a/gcc/config/arm/t-rmprofile b/gcc/config/arm/t-rmprofile
index 3e75fcc9635..a6036bf0a51 100644
--- a/gcc/config/arm/t-rmprofile
+++ b/gcc/config/arm/t-rmprofile
@@ -54,7 +54,7 @@ MULTILIB_REQUIRED	+= mthumb/march=armv8.1-m.main+mve/mfloat-abi=hard
 MULTILIB_MATCHES	+= march?armv6s-m=march?armv6-m
 
 # For all MULITIB_MATCHES for v8-m and above add mlibarch? on the right hand side
-# of = in the variant string instead of march?.  This is needed becuase all the
+# of = in the variant string instead of march?.  This is needed because all the
 # MULITIB_MATCHES variant strings are compared with mlibarch option for multilib
 # linking.
 
diff --git a/gcc/testsuite/gcc.target/arm/acle/pr100856.c b/gcc/testsuite/gcc.target/arm/acle/pr100856.c
index 5bc030e2e46..adbe1ab08f7 100644
--- a/gcc/testsuite/gcc.target/arm/acle/pr100856.c
+++ b/gcc/testsuite/gcc.target/arm/acle/pr100856.c
@@ -1,6 +1,6 @@
 /* { dg-do run } */
-/* { dg-require-effective-target arm_v8_1m_main_cde_mve_ok } */
-/* { dg-add-options arm_v8_1m_main_cde_mve } */
+/* { dg-require-effective-target arm_v8m_main_cde_multilib } */
+/* { dg-add-options arm_v8m_main_cde } */
 
 #include "arm_cde.h"
 
diff --git a/gcc/testsuite/lib/target-supports.exp b/gcc/testsuite/lib/target-supports.exp
index 7f78c5593ac..c29247c5bcf 100644
--- a/gcc/testsuite/lib/target-supports.exp
+++ b/gcc/testsuite/lib/target-supports.exp
@@ -5518,6 +5518,24 @@ foreach { armfunc armflag armdef arminc } {
 	    global et_FUNC_flags
 	    return "$flags $et_FUNC_flags"
 	}
+
+	proc check_effective_target_FUNC_multilib { } {
+	    if { ! [check_effective_target_FUNC_ok] } {
+		return 0;
+	    }
+	    return [check_runtime FUNC_multilib {
+		#if !(DEF)
+		#error "DEF failed"
+		#endif
+		#include <arm_cde.h>
+		INC
+		int
+		main (void)
+		{
+		    return 0;
+		}
+	    } [add_options_for_FUNC ""]]
+        }
     }]
 }
 
-- 
2.25.1


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856]
  2021-07-15 13:04 [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856] Christophe LYON
@ 2021-08-04  9:11 ` Christophe Lyon
  2021-08-11 14:06   ` Christophe Lyon
  2021-08-23  9:03 ` Kyrylo Tkachov
  1 sibling, 1 reply; 6+ messages in thread
From: Christophe Lyon @ 2021-08-04  9:11 UTC (permalink / raw)
  To: Christophe LYON; +Cc: gcc Patches

ping?

On Thu, 15 Jul 2021 at 15:07, Christophe LYON via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> This is a followup to Srinath's recent patch: the newly added test is
> failing e.g. on arm-linux-gnueabihf without R/M profile multilibs.
>
> It is also failing on arm-eabi with R/M profile multilibs if the
> execution engine does not support v8.1-M instructions.
>
> The patch avoids this by adding check_effective_target_FUNC_multilib
> in target-supports.exp which effectively checks whether the target
> supports linking and execution, like what is already done for other
> ARM effective targets.  pr100856.c is updated to use it instead of
> arm_v8_1m_main_cde_mve_ok (which makes the testcase a bit of a
> duplicate with check_effective_target_FUNC_multilib).
>
> In addition, I noticed that requiring MVE does not seem necessary and
> this enables the test to pass even when targeting a CPU without MVE:
> since the test does not involve actual CDE instructions, it can pass
> on other architecture versions.  For instance, when requiring MVE, we
> have to use cortex-m55 under QEMU for the test to pass because the
> memset() that comes from v8.1-m.main+mve multilib uses LOB
> instructions (DLS) (memset is used during startup).  Keeping
> arm_v8_1m_main_cde_mve_ok would mean we would enable the test provided
> we have the right multilibs, causing a runtime error if the simulator
> does not support LOB instructions (e.g. when targeting cortex-m7).
>
> I do not update sourcebuild.texi since the CDE effective targets are
> already collectively documented.
>
> Finally, the patch fixes two typos in comments.
>
> 2021-07-15  Christophe Lyon  <christophe.lyon@foss.st.com>
>
>          PR target/100856
>          gcc/
>          * config/arm/arm.opt: Fix typo.
>          * config/arm/t-rmprofile: Fix typo.
>
>          gcc/testsuite/
>          * gcc.target/arm/acle/pr100856.c: Use arm_v8m_main_cde_multilib
>          and arm_v8m_main_cde.
>          * lib/target-supports.exp: Add
> check_effective_target_FUNC_multilib for ARM CDE.
>
>

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856]
  2021-08-04  9:11 ` Christophe Lyon
@ 2021-08-11 14:06   ` Christophe Lyon
  2021-08-16 11:51     ` Christophe LYON
  0 siblings, 1 reply; 6+ messages in thread
From: Christophe Lyon @ 2021-08-11 14:06 UTC (permalink / raw)
  To: Christophe Lyon; +Cc: Christophe LYON, gcc Patches

ping?
https://gcc.gnu.org/pipermail/gcc-patches/2021-July/575310.html


On Wed, Aug 4, 2021 at 11:13 AM Christophe Lyon via Gcc-patches <
gcc-patches@gcc.gnu.org> wrote:

> ping?
>
> On Thu, 15 Jul 2021 at 15:07, Christophe LYON via Gcc-patches
> <gcc-patches@gcc.gnu.org> wrote:
> >
> > This is a followup to Srinath's recent patch: the newly added test is
> > failing e.g. on arm-linux-gnueabihf without R/M profile multilibs.
> >
> > It is also failing on arm-eabi with R/M profile multilibs if the
> > execution engine does not support v8.1-M instructions.
> >
> > The patch avoids this by adding check_effective_target_FUNC_multilib
> > in target-supports.exp which effectively checks whether the target
> > supports linking and execution, like what is already done for other
> > ARM effective targets.  pr100856.c is updated to use it instead of
> > arm_v8_1m_main_cde_mve_ok (which makes the testcase a bit of a
> > duplicate with check_effective_target_FUNC_multilib).
> >
> > In addition, I noticed that requiring MVE does not seem necessary and
> > this enables the test to pass even when targeting a CPU without MVE:
> > since the test does not involve actual CDE instructions, it can pass
> > on other architecture versions.  For instance, when requiring MVE, we
> > have to use cortex-m55 under QEMU for the test to pass because the
> > memset() that comes from v8.1-m.main+mve multilib uses LOB
> > instructions (DLS) (memset is used during startup).  Keeping
> > arm_v8_1m_main_cde_mve_ok would mean we would enable the test provided
> > we have the right multilibs, causing a runtime error if the simulator
> > does not support LOB instructions (e.g. when targeting cortex-m7).
> >
> > I do not update sourcebuild.texi since the CDE effective targets are
> > already collectively documented.
> >
> > Finally, the patch fixes two typos in comments.
> >
> > 2021-07-15  Christophe Lyon  <christophe.lyon@foss.st.com>
> >
> >          PR target/100856
> >          gcc/
> >          * config/arm/arm.opt: Fix typo.
> >          * config/arm/t-rmprofile: Fix typo.
> >
> >          gcc/testsuite/
> >          * gcc.target/arm/acle/pr100856.c: Use arm_v8m_main_cde_multilib
> >          and arm_v8m_main_cde.
> >          * lib/target-supports.exp: Add
> > check_effective_target_FUNC_multilib for ARM CDE.
> >
> >
>

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856]
  2021-08-11 14:06   ` Christophe Lyon
@ 2021-08-16 11:51     ` Christophe LYON
  2021-08-23  8:58       ` Christophe LYON
  0 siblings, 1 reply; 6+ messages in thread
From: Christophe LYON @ 2021-08-16 11:51 UTC (permalink / raw)
  To: gcc Patches

ping?

On 11/08/2021 16:06, Christophe Lyon wrote:
> ping?
> https://gcc.gnu.org/pipermail/gcc-patches/2021-July/575310.html 
> <https://gcc.gnu.org/pipermail/gcc-patches/2021-July/575310.html>
>
>
> On Wed, Aug 4, 2021 at 11:13 AM Christophe Lyon via Gcc-patches 
> <gcc-patches@gcc.gnu.org <mailto:gcc-patches@gcc.gnu.org>> wrote:
>
>     ping?
>
>     On Thu, 15 Jul 2021 at 15:07, Christophe LYON via Gcc-patches
>     <gcc-patches@gcc.gnu.org <mailto:gcc-patches@gcc.gnu.org>> wrote:
>     >
>     > This is a followup to Srinath's recent patch: the newly added
>     test is
>     > failing e.g. on arm-linux-gnueabihf without R/M profile multilibs.
>     >
>     > It is also failing on arm-eabi with R/M profile multilibs if the
>     > execution engine does not support v8.1-M instructions.
>     >
>     > The patch avoids this by adding check_effective_target_FUNC_multilib
>     > in target-supports.exp which effectively checks whether the target
>     > supports linking and execution, like what is already done for other
>     > ARM effective targets.  pr100856.c is updated to use it instead of
>     > arm_v8_1m_main_cde_mve_ok (which makes the testcase a bit of a
>     > duplicate with check_effective_target_FUNC_multilib).
>     >
>     > In addition, I noticed that requiring MVE does not seem
>     necessary and
>     > this enables the test to pass even when targeting a CPU without MVE:
>     > since the test does not involve actual CDE instructions, it can pass
>     > on other architecture versions.  For instance, when requiring
>     MVE, we
>     > have to use cortex-m55 under QEMU for the test to pass because the
>     > memset() that comes from v8.1-m.main+mve multilib uses LOB
>     > instructions (DLS) (memset is used during startup). Keeping
>     > arm_v8_1m_main_cde_mve_ok would mean we would enable the test
>     provided
>     > we have the right multilibs, causing a runtime error if the
>     simulator
>     > does not support LOB instructions (e.g. when targeting cortex-m7).
>     >
>     > I do not update sourcebuild.texi since the CDE effective targets are
>     > already collectively documented.
>     >
>     > Finally, the patch fixes two typos in comments.
>     >
>     > 2021-07-15  Christophe Lyon  <christophe.lyon@foss.st.com
>     <mailto:christophe.lyon@foss.st.com>>
>     >
>     >          PR target/100856
>     >          gcc/
>     >          * config/arm/arm.opt: Fix typo.
>     >          * config/arm/t-rmprofile: Fix typo.
>     >
>     >          gcc/testsuite/
>     >          * gcc.target/arm/acle/pr100856.c: Use
>     arm_v8m_main_cde_multilib
>     >          and arm_v8m_main_cde.
>     >          * lib/target-supports.exp: Add
>     > check_effective_target_FUNC_multilib for ARM CDE.
>     >
>     >
>

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856]
  2021-08-16 11:51     ` Christophe LYON
@ 2021-08-23  8:58       ` Christophe LYON
  0 siblings, 0 replies; 6+ messages in thread
From: Christophe LYON @ 2021-08-23  8:58 UTC (permalink / raw)
  To: gcc-patches

ping?


On 16/08/2021 13:51, Christophe LYON via Gcc-patches wrote:
> ping?
>
> On 11/08/2021 16:06, Christophe Lyon wrote:
>> ping?
>> https://gcc.gnu.org/pipermail/gcc-patches/2021-July/575310.html 
>> <https://gcc.gnu.org/pipermail/gcc-patches/2021-July/575310.html>
>>
>>
>> On Wed, Aug 4, 2021 at 11:13 AM Christophe Lyon via Gcc-patches 
>> <gcc-patches@gcc.gnu.org <mailto:gcc-patches@gcc.gnu.org>> wrote:
>>
>>     ping?
>>
>>     On Thu, 15 Jul 2021 at 15:07, Christophe LYON via Gcc-patches
>>     <gcc-patches@gcc.gnu.org <mailto:gcc-patches@gcc.gnu.org>> wrote:
>>     >
>>     > This is a followup to Srinath's recent patch: the newly added
>>     test is
>>     > failing e.g. on arm-linux-gnueabihf without R/M profile multilibs.
>>     >
>>     > It is also failing on arm-eabi with R/M profile multilibs if the
>>     > execution engine does not support v8.1-M instructions.
>>     >
>>     > The patch avoids this by adding 
>> check_effective_target_FUNC_multilib
>>     > in target-supports.exp which effectively checks whether the target
>>     > supports linking and execution, like what is already done for 
>> other
>>     > ARM effective targets.  pr100856.c is updated to use it instead of
>>     > arm_v8_1m_main_cde_mve_ok (which makes the testcase a bit of a
>>     > duplicate with check_effective_target_FUNC_multilib).
>>     >
>>     > In addition, I noticed that requiring MVE does not seem
>>     necessary and
>>     > this enables the test to pass even when targeting a CPU without 
>> MVE:
>>     > since the test does not involve actual CDE instructions, it can 
>> pass
>>     > on other architecture versions.  For instance, when requiring
>>     MVE, we
>>     > have to use cortex-m55 under QEMU for the test to pass because the
>>     > memset() that comes from v8.1-m.main+mve multilib uses LOB
>>     > instructions (DLS) (memset is used during startup). Keeping
>>     > arm_v8_1m_main_cde_mve_ok would mean we would enable the test
>>     provided
>>     > we have the right multilibs, causing a runtime error if the
>>     simulator
>>     > does not support LOB instructions (e.g. when targeting cortex-m7).
>>     >
>>     > I do not update sourcebuild.texi since the CDE effective 
>> targets are
>>     > already collectively documented.
>>     >
>>     > Finally, the patch fixes two typos in comments.
>>     >
>>     > 2021-07-15  Christophe Lyon <christophe.lyon@foss.st.com
>>     <mailto:christophe.lyon@foss.st.com>>
>>     >
>>     >          PR target/100856
>>     >          gcc/
>>     >          * config/arm/arm.opt: Fix typo.
>>     >          * config/arm/t-rmprofile: Fix typo.
>>     >
>>     >          gcc/testsuite/
>>     >          * gcc.target/arm/acle/pr100856.c: Use
>>     arm_v8m_main_cde_multilib
>>     >          and arm_v8m_main_cde.
>>     >          * lib/target-supports.exp: Add
>>     > check_effective_target_FUNC_multilib for ARM CDE.
>>     >
>>     >
>>

^ permalink raw reply	[flat|nested] 6+ messages in thread

* RE: [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856]
  2021-07-15 13:04 [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856] Christophe LYON
  2021-08-04  9:11 ` Christophe Lyon
@ 2021-08-23  9:03 ` Kyrylo Tkachov
  1 sibling, 0 replies; 6+ messages in thread
From: Kyrylo Tkachov @ 2021-08-23  9:03 UTC (permalink / raw)
  To: Christophe LYON; +Cc: gcc-patches



> -----Original Message-----
> From: Gcc-patches <gcc-patches-
> bounces+kyrylo.tkachov=arm.com@gcc.gnu.org> On Behalf Of Christophe
> LYON via Gcc-patches
> Sent: 15 July 2021 14:05
> To: gcc Patches <gcc-patches@gcc.gnu.org>
> Subject: [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856]
> 
> This is a followup to Srinath's recent patch: the newly added test is
> failing e.g. on arm-linux-gnueabihf without R/M profile multilibs.
> 
> It is also failing on arm-eabi with R/M profile multilibs if the
> execution engine does not support v8.1-M instructions.
> 
> The patch avoids this by adding check_effective_target_FUNC_multilib
> in target-supports.exp which effectively checks whether the target
> supports linking and execution, like what is already done for other
> ARM effective targets.  pr100856.c is updated to use it instead of
> arm_v8_1m_main_cde_mve_ok (which makes the testcase a bit of a
> duplicate with check_effective_target_FUNC_multilib).
> 
> In addition, I noticed that requiring MVE does not seem necessary and
> this enables the test to pass even when targeting a CPU without MVE:
> since the test does not involve actual CDE instructions, it can pass
> on other architecture versions.  For instance, when requiring MVE, we
> have to use cortex-m55 under QEMU for the test to pass because the
> memset() that comes from v8.1-m.main+mve multilib uses LOB
> instructions (DLS) (memset is used during startup).  Keeping
> arm_v8_1m_main_cde_mve_ok would mean we would enable the test
> provided
> we have the right multilibs, causing a runtime error if the simulator
> does not support LOB instructions (e.g. when targeting cortex-m7).
> 
> I do not update sourcebuild.texi since the CDE effective targets are
> already collectively documented.

Ok.
Sorry for the delay, I was on holiday.
Thanks,
Kyrill

> 
> Finally, the patch fixes two typos in comments.
> 
> 2021-07-15  Christophe Lyon  <christophe.lyon@foss.st.com>
> 
>          PR target/100856
>          gcc/
>          * config/arm/arm.opt: Fix typo.
>          * config/arm/t-rmprofile: Fix typo.
> 
>          gcc/testsuite/
>          * gcc.target/arm/acle/pr100856.c: Use arm_v8m_main_cde_multilib
>          and arm_v8m_main_cde.
>          * lib/target-supports.exp: Add
> check_effective_target_FUNC_multilib for ARM CDE.
> 


^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2021-08-23  9:03 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-07-15 13:04 [PATCH] arm: Fix multilib mapping for CDE extensions [PR100856] Christophe LYON
2021-08-04  9:11 ` Christophe Lyon
2021-08-11 14:06   ` Christophe Lyon
2021-08-16 11:51     ` Christophe LYON
2021-08-23  8:58       ` Christophe LYON
2021-08-23  9:03 ` Kyrylo Tkachov

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).