[PATCH][libgloss][AArch64] Fix flags for crt0 compilation.

[PATCH][libgloss][AArch64] Fix flags for crt0 compilation.

[Tamar Christina]

[-- Attachment #1: Type: text/plain, Size: 814 bytes --]

Hi All,

Previous patch to support nosys.specs accidentally broke
validation specs because ARM_RDI_MONITOR was never passed
to the build rule for crt0.

This fixed the compile for nosys and validation specs
but nosys won't run because of existing limitations to
aarch64's syscalls.c, it requires semihosting to get
commandline arguments and heap info without having a
fallback method as ARM does.

Tested on aarch64-none-elf with specs:
 - nosys.specs
 - aem-validation.specs
 - rdimon.specs

Previously patch was only tested with qemu, which seems
is always able to write to the std handles. So also
cross-checked with aem this time.

Ok for master?

Ps. I don't have commit right.

Thanks (sorry for the repeated fixes. All build system configurations should have caught up now.),
Tamar

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-Previous-patch-to-support-nosys.specs-accidentally.patch --]
[-- Type: text/x-patch; name="0001-Previous-patch-to-support-nosys.specs-accidentally.patch", Size: 1302 bytes --]

From e119b6c7ceca2ecef397002537fbc4208af1b93c Mon Sep 17 00:00:00 2001
From: Tamar Christina <tamar.christina@arm.com>
Date: Thu, 20 Jul 2017 13:28:40 +0100
Subject: [PATCH] Previous patch to support nosys.specs accidentally broke
 validation specs because ARM_RDI_MONITOR was never passed to the build rule
 for crt0.

This fixed the compile for nosys and validation specs
but nosys won't run because of existing limitations to
aarch64's syscalls.c, it requires semihosting to get
commandline arguments and heap info without having a
fallback method as ARM does.

Signed-off-by: Tamar Christina <tamar.christina@arm.com>
---
 libgloss/aarch64/Makefile.in | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libgloss/aarch64/Makefile.in b/libgloss/aarch64/Makefile.in
index 546557e3c..8111f262c 100644
--- a/libgloss/aarch64/Makefile.in
+++ b/libgloss/aarch64/Makefile.in
@@ -118,7 +118,7 @@ test:
 
 #
 crt0.o: crt0.S
-	$(CC) $(CFLAGS_FOR_TARGET) $($(MULTI_FLAGS_FOR_TARGET)) $(CFLAGS) $(INCLUDES) -o $@ -c $<
+	$(CC) $(CFLAGS_FOR_TARGET) $($(MULTI_FLAGS_FOR_TARGET)) $(CFLAGS) $(INCLUDES) -DARM_RDI_MONITOR -o $@ -c $<
 
 rdimon-crt0${${MULTILIBNAME}}.o: crt0.S
 	$(CC) $(CFLAGS_FOR_TARGET) $($(MULTI_FLAGS_FOR_TARGET)) $(CFLAGS) $(INCLUDES) -DARM_RDI_MONITOR -o $@ -c $<
-- 
2.13.0

Re: [PATCH][libgloss][AArch64] Fix flags for crt0 compilation.

[Corinna Vinschen]

[-- Attachment #1: Type: text/plain, Size: 1663 bytes --]

On Jul 21 08:28, Tamar Christina wrote:
> Hi All,
> 
> Previous patch to support nosys.specs accidentally broke
> validation specs because ARM_RDI_MONITOR was never passed
> to the build rule for crt0.
> 
> This fixed the compile for nosys and validation specs
> but nosys won't run because of existing limitations to
> aarch64's syscalls.c, it requires semihosting to get
> commandline arguments and heap info without having a
> fallback method as ARM does.
> 
> Tested on aarch64-none-elf with specs:
>  - nosys.specs
>  - aem-validation.specs
>  - rdimon.specs
> 
> Previously patch was only tested with qemu, which seems
> is always able to write to the std handles. So also
> cross-checked with aem this time.
> 
> Ok for master?
> 
> Ps. I don't have commit right.
> 
> Thanks (sorry for the repeated fixes. All build system configurations should have caught up now.),
> Tamar

> From e119b6c7ceca2ecef397002537fbc4208af1b93c Mon Sep 17 00:00:00 2001
> From: Tamar Christina <tamar.christina@arm.com>
> Date: Thu, 20 Jul 2017 13:28:40 +0100
> Subject: [PATCH] Previous patch to support nosys.specs accidentally broke
>  validation specs because ARM_RDI_MONITOR was never passed to the build rule
>  for crt0.
> 
> This fixed the compile for nosys and validation specs
> but nosys won't run because of existing limitations to
> aarch64's syscalls.c, it requires semihosting to get
> commandline arguments and heap info without having a
> fallback method as ARM does.
> 
> Signed-off-by: Tamar Christina <tamar.christina@arm.com>

Pushed.


Thanks,
Corinna

-- 
Corinna Vinschen
Cygwin Maintainer
Red Hat

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]