public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* Jit testsuite updates for Darwin
@ 2024-01-13 13:57 Iain Sandoe
  2024-01-13 13:57 ` [PATCH 1/4] testsuite, jit: test-alias-attribute.c requires alias support Iain Sandoe
  0 siblings, 1 reply; 9+ messages in thread
From: Iain Sandoe @ 2024-01-13 13:57 UTC (permalink / raw)
  To: dmalcolm, gcc-patches

This is a short series of patches to support non-ELF targets in
the jit tests where these use target-specific capabilities (e.g.
alias support, or scanning for assembler directives.)

An alternative would be to skip these tests for Darwin.

Tested on x86_64 Darwin, and about to test Linux,
OK for trunk if testing is also sucessful on Linux?
thanks
Iain



^ permalink raw reply	[flat|nested] 9+ messages in thread

* [PATCH 1/4] testsuite, jit: test-alias-attribute.c requires alias support.
  2024-01-13 13:57 Jit testsuite updates for Darwin Iain Sandoe
@ 2024-01-13 13:57 ` Iain Sandoe
  2024-01-13 13:57   ` [PATCH 2/4] testsuite, jit: Handle whitespace in test-link-section-assembler.c Iain Sandoe
  2024-01-16 19:39   ` [PATCH 1/4] testsuite, jit: test-alias-attribute.c requires alias support David Malcolm
  0 siblings, 2 replies; 9+ messages in thread
From: Iain Sandoe @ 2024-01-13 13:57 UTC (permalink / raw)
  To: dmalcolm, gcc-patches

Add a dg-require-alias to cover this.

gcc/testsuite/ChangeLog:

	* jit.dg/test-alias-attribute.c: Require target alias
	support.

Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
---
 gcc/testsuite/jit.dg/test-alias-attribute.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/gcc/testsuite/jit.dg/test-alias-attribute.c b/gcc/testsuite/jit.dg/test-alias-attribute.c
index eb29003dfc9..4741aba5011 100644
--- a/gcc/testsuite/jit.dg/test-alias-attribute.c
+++ b/gcc/testsuite/jit.dg/test-alias-attribute.c
@@ -1,4 +1,5 @@
 /* { dg-do compile { target x86_64-*-* } } */
+/* { dg-require-alias "" } */
 
 #include <stdlib.h>
 #include <stdio.h>
-- 
2.39.2 (Apple Git-143)


^ permalink raw reply	[flat|nested] 9+ messages in thread

* [PATCH 2/4] testsuite, jit: Handle whitespace in test-link-section-assembler.c.
  2024-01-13 13:57 ` [PATCH 1/4] testsuite, jit: test-alias-attribute.c requires alias support Iain Sandoe
@ 2024-01-13 13:57   ` Iain Sandoe
  2024-01-13 13:57     ` [PATCH 3/4] testsuite, jit: Allow for target-specific assembler scans Iain Sandoe
  2024-01-16 19:40     ` [PATCH 2/4] testsuite, jit: Handle whitespace in test-link-section-assembler.c David Malcolm
  2024-01-16 19:39   ` [PATCH 1/4] testsuite, jit: test-alias-attribute.c requires alias support David Malcolm
  1 sibling, 2 replies; 9+ messages in thread
From: Iain Sandoe @ 2024-01-13 13:57 UTC (permalink / raw)
  To: dmalcolm, gcc-patches

Darwin has a different .section directive that has more fields and
uses different whitespace.  Amend the whitespace in the scan-asm to
be more flexible.

gcc/testsuite/ChangeLog:

	* jit.dg/test-link-section-assembler.c: Accept any whitespace
	between the .section directive and its arguments.

Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
---
 gcc/testsuite/jit.dg/test-link-section-assembler.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/gcc/testsuite/jit.dg/test-link-section-assembler.c b/gcc/testsuite/jit.dg/test-link-section-assembler.c
index a90b00e9a82..a78e9fd26ef 100644
--- a/gcc/testsuite/jit.dg/test-link-section-assembler.c
+++ b/gcc/testsuite/jit.dg/test-link-section-assembler.c
@@ -34,4 +34,4 @@ create_code (gcc_jit_context *ctxt, void *user_data)
 }
 
 /* { dg-final { jit-verify-output-file-was-created "" } } */
-/* { dg-final { jit-verify-assembler-output ".section	.my_section" } } */
+/* { dg-final { jit-verify-assembler-output ".section\\s.my_section" } } */
-- 
2.39.2 (Apple Git-143)


^ permalink raw reply	[flat|nested] 9+ messages in thread

* [PATCH 3/4] testsuite, jit: Allow for target-specific assembler scans.
  2024-01-13 13:57   ` [PATCH 2/4] testsuite, jit: Handle whitespace in test-link-section-assembler.c Iain Sandoe
@ 2024-01-13 13:57     ` Iain Sandoe
  2024-01-13 13:57       ` [PATCH 4/4] testsuite,jit: Handle Darwin/Mach-O in assembler tests Iain Sandoe
  2024-01-16 19:43       ` [PATCH 3/4] testsuite, jit: Allow for target-specific assembler scans David Malcolm
  2024-01-16 19:40     ` [PATCH 2/4] testsuite, jit: Handle whitespace in test-link-section-assembler.c David Malcolm
  1 sibling, 2 replies; 9+ messages in thread
From: Iain Sandoe @ 2024-01-13 13:57 UTC (permalink / raw)
  To: dmalcolm, gcc-patches

If we want to support multiple object formats and to allow for
scan-assembler tests, we need to make it possible to adjust the
tests on a per-target basis.

This adds similar mechamisms to jit-verify-assembler-output{,-not}
to those used for the general scan-assembler dg directives.

As an aside; it would, perhaps, be possible to integrate this more
with scanasm.exp (which would also give access to function body
scanning) but I did not attempt that for this patch.

After this, we can accept things like:
... { jit-verify-assembler-output-not "......" { target { ! *-*-darwin* } } } }
or
... { jit-verify-assembler-output "......" { target *-*-darwin* } } }

gcc/testsuite/ChangeLog:

	* jit.dg/jit.exp: Accept target clauses in jit-verify-assembler
	handling.

Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
---
 gcc/testsuite/jit.dg/jit.exp | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)

diff --git a/gcc/testsuite/jit.dg/jit.exp b/gcc/testsuite/jit.dg/jit.exp
index 56972064d30..286cfa8192a 100644
--- a/gcc/testsuite/jit.dg/jit.exp
+++ b/gcc/testsuite/jit.dg/jit.exp
@@ -875,9 +875,23 @@ proc jit-verify-assembler { args } {
 proc jit-verify-assembler-output { args } {
     verbose "jit-verify-assembler: $args"
 
+    if { [llength $args] > 3 } {
+	error "jit-verify-assembler-output: too many arguments"
+	return
+    }
+
     set dg-output-text [lindex $args 0]
     verbose "dg-output-text: ${dg-output-text}"
 
+    if { [llength $args] >= 2 } {
+	switch [dg-process-target [lindex $args 1]] {
+	    "S" { }
+	    "N" { return }
+	    "F" { setup_xfail "*-*-*" }
+	    "P" { }
+	}
+    }
+
     upvar 2 name name
     verbose "name: $name"
 
@@ -907,9 +921,23 @@ proc jit-verify-assembler-output { args } {
 proc jit-verify-assembler-output-not { args } {
     verbose "jit-verify-assembler: $args"
 
+    if { [llength $args] > 3 } {
+	error "jit-verify-assembler-output-not: too many arguments"
+	return
+    }
+
     set dg-output-text [lindex $args 0]
     verbose "dg-output-text: ${dg-output-text}"
 
+    if { [llength $args] >= 2 } {
+	switch [dg-process-target [lindex $args 1]] {
+	    "S" { }
+	    "N" { return }
+	    "F" { setup_xfail "*-*-*" }
+	    "P" { }
+	}
+    }
+
     upvar 2 name name
     verbose "name: $name"
 
-- 
2.39.2 (Apple Git-143)


^ permalink raw reply	[flat|nested] 9+ messages in thread

* [PATCH 4/4] testsuite,jit: Handle Darwin/Mach-O in assembler tests.
  2024-01-13 13:57     ` [PATCH 3/4] testsuite, jit: Allow for target-specific assembler scans Iain Sandoe
@ 2024-01-13 13:57       ` Iain Sandoe
  2024-01-16 19:52         ` David Malcolm
  2024-01-16 19:43       ` [PATCH 3/4] testsuite, jit: Allow for target-specific assembler scans David Malcolm
  1 sibling, 1 reply; 9+ messages in thread
From: Iain Sandoe @ 2024-01-13 13:57 UTC (permalink / raw)
  To: dmalcolm, gcc-patches

Several of the jit tests check for assembler-specific output
which differs on Mach-O from ELF.

This patch uses the facility to make the scans targer-dependent
and adds handling for darwin.

gcc/testsuite/ChangeLog:

	* jit.dg/test-always_inline-attribute.c: Handle Darwin in
	jit-verify-assembler-output.
	* jit.dg/test-noinline-attribute.c: Likewise.
	* jit.dg/test-setting-alignment.c: Likewise.
	* jit.dg/test-used-attribute.c: Likewise.
	* jit.dg/test-variable-attribute.c: Likewise.
	* jit.dg/test-weak-attribute.c: Likewise.

Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
---
 gcc/testsuite/jit.dg/test-always_inline-attribute.c | 10 +++++++---
 gcc/testsuite/jit.dg/test-noinline-attribute.c      | 10 +++++++---
 gcc/testsuite/jit.dg/test-setting-alignment.c       |  3 ++-
 gcc/testsuite/jit.dg/test-used-attribute.c          | 10 +++++++---
 gcc/testsuite/jit.dg/test-variable-attribute.c      | 10 ++++++----
 gcc/testsuite/jit.dg/test-weak-attribute.c          |  3 ++-
 6 files changed, 31 insertions(+), 15 deletions(-)

diff --git a/gcc/testsuite/jit.dg/test-always_inline-attribute.c b/gcc/testsuite/jit.dg/test-always_inline-attribute.c
index 5c3f386663f..f7201a17140 100644
--- a/gcc/testsuite/jit.dg/test-always_inline-attribute.c
+++ b/gcc/testsuite/jit.dg/test-always_inline-attribute.c
@@ -148,6 +148,10 @@ int foo () {
 
 /* { dg-final { jit-verify-output-file-was-created "" } } */
 /* Check that the "removed" function was inlined, but not the others */
-/* { dg-final { jit-verify-assembler-output-not ".type\\s+removed,\\s+@function" } } */
-/* { dg-final { jit-verify-assembler-output ".type\\s+not_removed,\\s+@function" } } */
-/* { dg-final { jit-verify-assembler-output ".type\\s+foo,\\s+@function" } } */
+/* { dg-final { jit-verify-assembler-output-not ".type\\s+removed,\\s+@function" { target { ! *-*-darwin* } } } } */
+/* { dg-final { jit-verify-assembler-output ".type\\s+not_removed,\\s+@function" { target { ! *-*-darwin* } } } } */
+/* { dg-final { jit-verify-assembler-output ".type\\s+foo,\\s+@function" { target { ! *-*-darwin* } } } } */
+
+/* { dg-final { jit-verify-assembler-output-not "\\n_removed:" { target *-*-darwin* } } } */
+/* { dg-final { jit-verify-assembler-output "\\n_not_removed:" { target *-*-darwin* } } } */
+/* { dg-final { jit-verify-assembler-output "\\n_foo:" { target *-*-darwin* } } } */
diff --git a/gcc/testsuite/jit.dg/test-noinline-attribute.c b/gcc/testsuite/jit.dg/test-noinline-attribute.c
index eac6cae6b6a..acfea8fc55e 100644
--- a/gcc/testsuite/jit.dg/test-noinline-attribute.c
+++ b/gcc/testsuite/jit.dg/test-noinline-attribute.c
@@ -114,6 +114,10 @@ int foo () {
 
 /* { dg-final { jit-verify-output-file-was-created "" } } */
 /* Check that the "removed" function was inlined, but not the others */
-/* { dg-final { jit-verify-assembler-output-not ".type\\s+removed.isra.0,\\s+@function" } } */
-/* { dg-final { jit-verify-assembler-output ".type\\s+not_removed.isra.0,\\s+@function" } } */
-/* { dg-final { jit-verify-assembler-output ".type\\s+foo,\\s+@function" } } */
+/* { dg-final { jit-verify-assembler-output-not ".type\\s+removed.isra.0,\\s+@function" { target { ! *-*-darwin* } } } } */
+/* { dg-final { jit-verify-assembler-output ".type\\s+not_removed.isra.0,\\s+@function" { target { ! *-*-darwin* } } } } */
+/* { dg-final { jit-verify-assembler-output ".type\\s+foo,\\s+@function" { target { ! *-*-darwin* } } } } */
+
+/* { dg-final { jit-verify-assembler-output-not "\\n_removed.isra.0:" { target *-*-darwin* } } } */
+/* { dg-final { jit-verify-assembler-output "\\n_not_removed.isra.0:" { target *-*-darwin* } } } */
+/* { dg-final { jit-verify-assembler-output "\\n_foo:" { target *-*-darwin* } } } */
diff --git a/gcc/testsuite/jit.dg/test-setting-alignment.c b/gcc/testsuite/jit.dg/test-setting-alignment.c
index 8489df9c6b9..14edc723f61 100644
--- a/gcc/testsuite/jit.dg/test-setting-alignment.c
+++ b/gcc/testsuite/jit.dg/test-setting-alignment.c
@@ -62,5 +62,6 @@ create_code (gcc_jit_context *ctxt, void *user_data)
 }
 
 /* { dg-final { jit-verify-output-file-was-created "" } } */
-/* { dg-final { jit-verify-assembler-output ".comm	foo,4,8" } } */
+/* { dg-final { jit-verify-assembler-output ".comm	foo,4,8" { target { ! *-*-darwin* } } } } */
+/* { dg-final { jit-verify-assembler-output ".comm\\s_foo,4,3" { target *-*-darwin* } } } */
 /* { dg-final { jit-verify-assembler-output "movl	-16\\\(%rbp\\\), %eax" } } */
diff --git a/gcc/testsuite/jit.dg/test-used-attribute.c b/gcc/testsuite/jit.dg/test-used-attribute.c
index cb20952c687..446c5c5c31b 100644
--- a/gcc/testsuite/jit.dg/test-used-attribute.c
+++ b/gcc/testsuite/jit.dg/test-used-attribute.c
@@ -107,6 +107,10 @@ int foo() {
 
 /* { dg-final { jit-verify-output-file-was-created "" } } */
 /* Check that the "removed" function was inlined, but not the others */
-/* { dg-final { jit-verify-assembler-output-not ".type\\s+removed,\\s+@function" } } */
-/* { dg-final { jit-verify-assembler-output ".type\\s+not_removed,\\s+@function" } } */
-/* { dg-final { jit-verify-assembler-output ".type\\s+foo,\\s+@function" } } */
+/* { dg-final { jit-verify-assembler-output-not ".type\\s+removed,\\s+@function" { target { ! *-*-darwin* } } } } */
+/* { dg-final { jit-verify-assembler-output ".type\\s+not_removed,\\s+@function" { target { ! *-*-darwin* } } } } */
+/* { dg-final { jit-verify-assembler-output ".type\\s+foo,\\s+@function" { target { ! *-*-darwin* } } } } */
+
+/* { dg-final { jit-verify-assembler-output-not "\\n_removed:" { target *-*-darwin* } } } */
+/* { dg-final { jit-verify-assembler-output "\\n_not_removed:" { target *-*-darwin* } } } */
+/* { dg-final { jit-verify-assembler-output "\\n_foo:" { target *-*-darwin* } } } */
diff --git a/gcc/testsuite/jit.dg/test-variable-attribute.c b/gcc/testsuite/jit.dg/test-variable-attribute.c
index ea854ff4a9f..219a47d8817 100644
--- a/gcc/testsuite/jit.dg/test-variable-attribute.c
+++ b/gcc/testsuite/jit.dg/test-variable-attribute.c
@@ -40,7 +40,9 @@ int PUBLIC = 12;
 
 /* { dg-final { jit-verify-output-file-was-created "" } } */
 /* Check that the attribute was applied correctly */
-/* { dg-final { jit-verify-assembler-output ".hidden\\s+PRIVATE" } } */
-/* { dg-final { jit-verify-assembler-output ".globl\\s+PRIVATE" } } */
-/* { dg-final { jit-verify-assembler-output-not ".hidden\\s+PUBLIC" } } */
-/* { dg-final { jit-verify-assembler-output ".globl\\s+PUBLIC" } } */
+/* { dg-final { jit-verify-assembler-output ".hidden\\s+PRIVATE" { target { ! *-*-darwin* } } } } */
+/* { dg-final { jit-verify-assembler-output ".private_extern\\s+_PRIVATE" { target *-*-darwin* } } } */
+/* { dg-final { jit-verify-assembler-output ".globl\\s+_?PRIVATE" } } */
+/* { dg-final { jit-verify-assembler-output-not ".hidden\\s+PUBLIC" { target { ! *-*-darwin* } } } } */
+/* { dg-final { jit-verify-assembler-output-not ".private_extern\\s+_PUBLIC" { target *-*-darwin* } } } */
+/* { dg-final { jit-verify-assembler-output ".globl\\s+_?PUBLIC" } } */
diff --git a/gcc/testsuite/jit.dg/test-weak-attribute.c b/gcc/testsuite/jit.dg/test-weak-attribute.c
index 546ade1c3c4..3ca9485bf41 100644
--- a/gcc/testsuite/jit.dg/test-weak-attribute.c
+++ b/gcc/testsuite/jit.dg/test-weak-attribute.c
@@ -38,4 +38,5 @@ void f () {}
 
 /* { dg-final { jit-verify-output-file-was-created "" } } */
 /* Check that the attribute was applied correctly */
-/* { dg-final { jit-verify-assembler-output ".weak\\s+f" } } */
+/* { dg-final { jit-verify-assembler-output ".weak\\s+f" { target { ! *-*-darwin* } } } } */
+/* { dg-final { jit-verify-assembler-output ".weak_definition\\s+_f" { target *-*-darwin* } } } */
-- 
2.39.2 (Apple Git-143)


^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH 1/4] testsuite, jit: test-alias-attribute.c requires alias support.
  2024-01-13 13:57 ` [PATCH 1/4] testsuite, jit: test-alias-attribute.c requires alias support Iain Sandoe
  2024-01-13 13:57   ` [PATCH 2/4] testsuite, jit: Handle whitespace in test-link-section-assembler.c Iain Sandoe
@ 2024-01-16 19:39   ` David Malcolm
  1 sibling, 0 replies; 9+ messages in thread
From: David Malcolm @ 2024-01-16 19:39 UTC (permalink / raw)
  To: iain, gcc-patches

On Sat, 2024-01-13 at 13:57 +0000, Iain Sandoe wrote:
> Add a dg-require-alias to cover this.

Sorry about the breakage; LGTM for trunk.

Thanks
Dave

> 
> gcc/testsuite/ChangeLog:
> 
>         * jit.dg/test-alias-attribute.c: Require target alias
>         support.
> 
> Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
> ---
>  gcc/testsuite/jit.dg/test-alias-attribute.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/gcc/testsuite/jit.dg/test-alias-attribute.c
> b/gcc/testsuite/jit.dg/test-alias-attribute.c
> index eb29003dfc9..4741aba5011 100644
> --- a/gcc/testsuite/jit.dg/test-alias-attribute.c
> +++ b/gcc/testsuite/jit.dg/test-alias-attribute.c
> @@ -1,4 +1,5 @@
>  /* { dg-do compile { target x86_64-*-* } } */
> +/* { dg-require-alias "" } */
>  
>  #include <stdlib.h>
>  #include <stdio.h>


^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH 2/4] testsuite, jit: Handle whitespace in test-link-section-assembler.c.
  2024-01-13 13:57   ` [PATCH 2/4] testsuite, jit: Handle whitespace in test-link-section-assembler.c Iain Sandoe
  2024-01-13 13:57     ` [PATCH 3/4] testsuite, jit: Allow for target-specific assembler scans Iain Sandoe
@ 2024-01-16 19:40     ` David Malcolm
  1 sibling, 0 replies; 9+ messages in thread
From: David Malcolm @ 2024-01-16 19:40 UTC (permalink / raw)
  To: iain, gcc-patches

On Sat, 2024-01-13 at 13:57 +0000, Iain Sandoe wrote:
> Darwin has a different .section directive that has more fields and
> uses different whitespace.  Amend the whitespace in the scan-asm to
> be more flexible.

LGTM for trunk, assuming your testing was successful.

Thanks
Dave

> 
> gcc/testsuite/ChangeLog:
> 
>         * jit.dg/test-link-section-assembler.c: Accept any whitespace
>         between the .section directive and its arguments.
> 
> Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
> ---
>  gcc/testsuite/jit.dg/test-link-section-assembler.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/gcc/testsuite/jit.dg/test-link-section-assembler.c
> b/gcc/testsuite/jit.dg/test-link-section-assembler.c
> index a90b00e9a82..a78e9fd26ef 100644
> --- a/gcc/testsuite/jit.dg/test-link-section-assembler.c
> +++ b/gcc/testsuite/jit.dg/test-link-section-assembler.c
> @@ -34,4 +34,4 @@ create_code (gcc_jit_context *ctxt, void
> *user_data)
>  }
>  
>  /* { dg-final { jit-verify-output-file-was-created "" } } */
> -/* { dg-final { jit-verify-assembler-output ".section  .my_section"
> } } */
> +/* { dg-final { jit-verify-assembler-output ".section\\s.my_section"
> } } */


^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH 3/4] testsuite, jit: Allow for target-specific assembler scans.
  2024-01-13 13:57     ` [PATCH 3/4] testsuite, jit: Allow for target-specific assembler scans Iain Sandoe
  2024-01-13 13:57       ` [PATCH 4/4] testsuite,jit: Handle Darwin/Mach-O in assembler tests Iain Sandoe
@ 2024-01-16 19:43       ` David Malcolm
  1 sibling, 0 replies; 9+ messages in thread
From: David Malcolm @ 2024-01-16 19:43 UTC (permalink / raw)
  To: iain, gcc-patches

On Sat, 2024-01-13 at 13:57 +0000, Iain Sandoe wrote:
> If we want to support multiple object formats and to allow for
> scan-assembler tests, we need to make it possible to adjust the
> tests on a per-target basis.
> 
> This adds similar mechamisms to jit-verify-assembler-output{,-not}
> to those used for the general scan-assembler dg directives.
> 
> As an aside; it would, perhaps, be possible to integrate this more
> with scanasm.exp (which would also give access to function body
> scanning) but I did not attempt that for this patch.
> 
> After this, we can accept things like:
> ... { jit-verify-assembler-output-not "......" { target { ! *-*-
> darwin* } } } }
> or
> ... { jit-verify-assembler-output "......" { target *-*-darwin* } } }


Thanks; LGTM for trunk, assuming your testing was successful

Dave

> 
> gcc/testsuite/ChangeLog:
> 
>         * jit.dg/jit.exp: Accept target clauses in jit-verify-
> assembler
>         handling.
> 
> Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
> ---
>  gcc/testsuite/jit.dg/jit.exp | 28 ++++++++++++++++++++++++++++
>  1 file changed, 28 insertions(+)
> 
> diff --git a/gcc/testsuite/jit.dg/jit.exp
> b/gcc/testsuite/jit.dg/jit.exp
> index 56972064d30..286cfa8192a 100644
> --- a/gcc/testsuite/jit.dg/jit.exp
> +++ b/gcc/testsuite/jit.dg/jit.exp
> @@ -875,9 +875,23 @@ proc jit-verify-assembler { args } {
>  proc jit-verify-assembler-output { args } {
>      verbose "jit-verify-assembler: $args"
>  
> +    if { [llength $args] > 3 } {
> +       error "jit-verify-assembler-output: too many arguments"
> +       return
> +    }
> +
>      set dg-output-text [lindex $args 0]
>      verbose "dg-output-text: ${dg-output-text}"
>  
> +    if { [llength $args] >= 2 } {
> +       switch [dg-process-target [lindex $args 1]] {
> +           "S" { }
> +           "N" { return }
> +           "F" { setup_xfail "*-*-*" }
> +           "P" { }
> +       }
> +    }
> +
>      upvar 2 name name
>      verbose "name: $name"
>  
> @@ -907,9 +921,23 @@ proc jit-verify-assembler-output { args } {
>  proc jit-verify-assembler-output-not { args } {
>      verbose "jit-verify-assembler: $args"
>  
> +    if { [llength $args] > 3 } {
> +       error "jit-verify-assembler-output-not: too many arguments"
> +       return
> +    }
> +
>      set dg-output-text [lindex $args 0]
>      verbose "dg-output-text: ${dg-output-text}"
>  
> +    if { [llength $args] >= 2 } {
> +       switch [dg-process-target [lindex $args 1]] {
> +           "S" { }
> +           "N" { return }
> +           "F" { setup_xfail "*-*-*" }
> +           "P" { }
> +       }
> +    }
> +
>      upvar 2 name name
>      verbose "name: $name"
>  


^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH 4/4] testsuite,jit: Handle Darwin/Mach-O in assembler tests.
  2024-01-13 13:57       ` [PATCH 4/4] testsuite,jit: Handle Darwin/Mach-O in assembler tests Iain Sandoe
@ 2024-01-16 19:52         ` David Malcolm
  0 siblings, 0 replies; 9+ messages in thread
From: David Malcolm @ 2024-01-16 19:52 UTC (permalink / raw)
  To: iain, gcc-patches

On Sat, 2024-01-13 at 13:57 +0000, Iain Sandoe wrote:
> Several of the jit tests check for assembler-specific output
> which differs on Mach-O from ELF.
> 
> This patch uses the facility to make the scans targer-dependent
> and adds handling for darwin.

Sorry about the issues with the tests.

Thanks for the patch.  I wonder if there's a better way to express ELF
vs Mach-O, and if these tests should really be rewritten to e.g. use
dumpfiles from an earlier point in the optimizer where the results
might be less target-specific.

In the meantime, LGTM for trunk, assuming your testing was successful.

Thanks again for the patch
Dave


> 
> gcc/testsuite/ChangeLog:
> 
>         * jit.dg/test-always_inline-attribute.c: Handle Darwin in
>         jit-verify-assembler-output.
>         * jit.dg/test-noinline-attribute.c: Likewise.
>         * jit.dg/test-setting-alignment.c: Likewise.
>         * jit.dg/test-used-attribute.c: Likewise.
>         * jit.dg/test-variable-attribute.c: Likewise.
>         * jit.dg/test-weak-attribute.c: Likewise.
> 
> Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
> ---
>  gcc/testsuite/jit.dg/test-always_inline-attribute.c | 10 +++++++---
>  gcc/testsuite/jit.dg/test-noinline-attribute.c      | 10 +++++++---
>  gcc/testsuite/jit.dg/test-setting-alignment.c       |  3 ++-
>  gcc/testsuite/jit.dg/test-used-attribute.c          | 10 +++++++---
>  gcc/testsuite/jit.dg/test-variable-attribute.c      | 10 ++++++----
>  gcc/testsuite/jit.dg/test-weak-attribute.c          |  3 ++-
>  6 files changed, 31 insertions(+), 15 deletions(-)
> 
> diff --git a/gcc/testsuite/jit.dg/test-always_inline-attribute.c
> b/gcc/testsuite/jit.dg/test-always_inline-attribute.c
> index 5c3f386663f..f7201a17140 100644
> --- a/gcc/testsuite/jit.dg/test-always_inline-attribute.c
> +++ b/gcc/testsuite/jit.dg/test-always_inline-attribute.c
> @@ -148,6 +148,10 @@ int foo () {
>  
>  /* { dg-final { jit-verify-output-file-was-created "" } } */
>  /* Check that the "removed" function was inlined, but not the others
> */
> -/* { dg-final { jit-verify-assembler-output-not
> ".type\\s+removed,\\s+@function" } } */
> -/* { dg-final { jit-verify-assembler-output
> ".type\\s+not_removed,\\s+@function" } } */
> -/* { dg-final { jit-verify-assembler-output
> ".type\\s+foo,\\s+@function" } } */
> +/* { dg-final { jit-verify-assembler-output-not
> ".type\\s+removed,\\s+@function" { target { ! *-*-darwin* } } } } */
> +/* { dg-final { jit-verify-assembler-output
> ".type\\s+not_removed,\\s+@function" { target { ! *-*-darwin* } } } }
> */
> +/* { dg-final { jit-verify-assembler-output
> ".type\\s+foo,\\s+@function" { target { ! *-*-darwin* } } } } */
> +
> +/* { dg-final { jit-verify-assembler-output-not "\\n_removed:" {
> target *-*-darwin* } } } */
> +/* { dg-final { jit-verify-assembler-output "\\n_not_removed:" {
> target *-*-darwin* } } } */
> +/* { dg-final { jit-verify-assembler-output "\\n_foo:" { target *-*-
> darwin* } } } */
> diff --git a/gcc/testsuite/jit.dg/test-noinline-attribute.c
> b/gcc/testsuite/jit.dg/test-noinline-attribute.c
> index eac6cae6b6a..acfea8fc55e 100644
> --- a/gcc/testsuite/jit.dg/test-noinline-attribute.c
> +++ b/gcc/testsuite/jit.dg/test-noinline-attribute.c
> @@ -114,6 +114,10 @@ int foo () {
>  
>  /* { dg-final { jit-verify-output-file-was-created "" } } */
>  /* Check that the "removed" function was inlined, but not the others
> */
> -/* { dg-final { jit-verify-assembler-output-not
> ".type\\s+removed.isra.0,\\s+@function" } } */
> -/* { dg-final { jit-verify-assembler-output
> ".type\\s+not_removed.isra.0,\\s+@function" } } */
> -/* { dg-final { jit-verify-assembler-output
> ".type\\s+foo,\\s+@function" } } */
> +/* { dg-final { jit-verify-assembler-output-not
> ".type\\s+removed.isra.0,\\s+@function" { target { ! *-*-darwin* } }
> } } */
> +/* { dg-final { jit-verify-assembler-output
> ".type\\s+not_removed.isra.0,\\s+@function" { target { ! *-*-darwin*
> } } } } */
> +/* { dg-final { jit-verify-assembler-output
> ".type\\s+foo,\\s+@function" { target { ! *-*-darwin* } } } } */
> +
> +/* { dg-final { jit-verify-assembler-output-not
> "\\n_removed.isra.0:" { target *-*-darwin* } } } */
> +/* { dg-final { jit-verify-assembler-output
> "\\n_not_removed.isra.0:" { target *-*-darwin* } } } */
> +/* { dg-final { jit-verify-assembler-output "\\n_foo:" { target *-*-
> darwin* } } } */
> diff --git a/gcc/testsuite/jit.dg/test-setting-alignment.c
> b/gcc/testsuite/jit.dg/test-setting-alignment.c
> index 8489df9c6b9..14edc723f61 100644
> --- a/gcc/testsuite/jit.dg/test-setting-alignment.c
> +++ b/gcc/testsuite/jit.dg/test-setting-alignment.c
> @@ -62,5 +62,6 @@ create_code (gcc_jit_context *ctxt, void
> *user_data)
>  }
>  
>  /* { dg-final { jit-verify-output-file-was-created "" } } */
> -/* { dg-final { jit-verify-assembler-output ".comm     foo,4,8" } }
> */
> +/* { dg-final { jit-verify-assembler-output ".comm     foo,4,8" {
> target { ! *-*-darwin* } } } } */
> +/* { dg-final { jit-verify-assembler-output ".comm\\s_foo,4,3" {
> target *-*-darwin* } } } */
>  /* { dg-final { jit-verify-assembler-output "movl      -
> 16\\\(%rbp\\\), %eax" } } */
> diff --git a/gcc/testsuite/jit.dg/test-used-attribute.c
> b/gcc/testsuite/jit.dg/test-used-attribute.c
> index cb20952c687..446c5c5c31b 100644
> --- a/gcc/testsuite/jit.dg/test-used-attribute.c
> +++ b/gcc/testsuite/jit.dg/test-used-attribute.c
> @@ -107,6 +107,10 @@ int foo() {
>  
>  /* { dg-final { jit-verify-output-file-was-created "" } } */
>  /* Check that the "removed" function was inlined, but not the others
> */
> -/* { dg-final { jit-verify-assembler-output-not
> ".type\\s+removed,\\s+@function" } } */
> -/* { dg-final { jit-verify-assembler-output
> ".type\\s+not_removed,\\s+@function" } } */
> -/* { dg-final { jit-verify-assembler-output
> ".type\\s+foo,\\s+@function" } } */
> +/* { dg-final { jit-verify-assembler-output-not
> ".type\\s+removed,\\s+@function" { target { ! *-*-darwin* } } } } */
> +/* { dg-final { jit-verify-assembler-output
> ".type\\s+not_removed,\\s+@function" { target { ! *-*-darwin* } } } }
> */
> +/* { dg-final { jit-verify-assembler-output
> ".type\\s+foo,\\s+@function" { target { ! *-*-darwin* } } } } */
> +
> +/* { dg-final { jit-verify-assembler-output-not "\\n_removed:" {
> target *-*-darwin* } } } */
> +/* { dg-final { jit-verify-assembler-output "\\n_not_removed:" {
> target *-*-darwin* } } } */
> +/* { dg-final { jit-verify-assembler-output "\\n_foo:" { target *-*-
> darwin* } } } */
> diff --git a/gcc/testsuite/jit.dg/test-variable-attribute.c
> b/gcc/testsuite/jit.dg/test-variable-attribute.c
> index ea854ff4a9f..219a47d8817 100644
> --- a/gcc/testsuite/jit.dg/test-variable-attribute.c
> +++ b/gcc/testsuite/jit.dg/test-variable-attribute.c
> @@ -40,7 +40,9 @@ int PUBLIC = 12;
>  
>  /* { dg-final { jit-verify-output-file-was-created "" } } */
>  /* Check that the attribute was applied correctly */
> -/* { dg-final { jit-verify-assembler-output ".hidden\\s+PRIVATE" } }
> */
> -/* { dg-final { jit-verify-assembler-output ".globl\\s+PRIVATE" } }
> */
> -/* { dg-final { jit-verify-assembler-output-not ".hidden\\s+PUBLIC"
> } } */
> -/* { dg-final { jit-verify-assembler-output ".globl\\s+PUBLIC" } }
> */
> +/* { dg-final { jit-verify-assembler-output ".hidden\\s+PRIVATE" {
> target { ! *-*-darwin* } } } } */
> +/* { dg-final { jit-verify-assembler-output
> ".private_extern\\s+_PRIVATE" { target *-*-darwin* } } } */
> +/* { dg-final { jit-verify-assembler-output ".globl\\s+_?PRIVATE" }
> } */
> +/* { dg-final { jit-verify-assembler-output-not ".hidden\\s+PUBLIC"
> { target { ! *-*-darwin* } } } } */
> +/* { dg-final { jit-verify-assembler-output-not
> ".private_extern\\s+_PUBLIC" { target *-*-darwin* } } } */
> +/* { dg-final { jit-verify-assembler-output ".globl\\s+_?PUBLIC" } }
> */
> diff --git a/gcc/testsuite/jit.dg/test-weak-attribute.c
> b/gcc/testsuite/jit.dg/test-weak-attribute.c
> index 546ade1c3c4..3ca9485bf41 100644
> --- a/gcc/testsuite/jit.dg/test-weak-attribute.c
> +++ b/gcc/testsuite/jit.dg/test-weak-attribute.c
> @@ -38,4 +38,5 @@ void f () {}
>  
>  /* { dg-final { jit-verify-output-file-was-created "" } } */
>  /* Check that the attribute was applied correctly */
> -/* { dg-final { jit-verify-assembler-output ".weak\\s+f" } } */
> +/* { dg-final { jit-verify-assembler-output ".weak\\s+f" { target {
> ! *-*-darwin* } } } } */
> +/* { dg-final { jit-verify-assembler-output ".weak_definition\\s+_f"
> { target *-*-darwin* } } } */


^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2024-01-16 19:52 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-01-13 13:57 Jit testsuite updates for Darwin Iain Sandoe
2024-01-13 13:57 ` [PATCH 1/4] testsuite, jit: test-alias-attribute.c requires alias support Iain Sandoe
2024-01-13 13:57   ` [PATCH 2/4] testsuite, jit: Handle whitespace in test-link-section-assembler.c Iain Sandoe
2024-01-13 13:57     ` [PATCH 3/4] testsuite, jit: Allow for target-specific assembler scans Iain Sandoe
2024-01-13 13:57       ` [PATCH 4/4] testsuite,jit: Handle Darwin/Mach-O in assembler tests Iain Sandoe
2024-01-16 19:52         ` David Malcolm
2024-01-16 19:43       ` [PATCH 3/4] testsuite, jit: Allow for target-specific assembler scans David Malcolm
2024-01-16 19:40     ` [PATCH 2/4] testsuite, jit: Handle whitespace in test-link-section-assembler.c David Malcolm
2024-01-16 19:39   ` [PATCH 1/4] testsuite, jit: test-alias-attribute.c requires alias support David Malcolm

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).