[Buildroot] [PATCH] gdb: add fix for gdb 7.12 and gdb 8.0 build on noMMU platforms

[Buildroot] [PATCH] gdb: add fix for gdb 7.12 and gdb 8.0 build on noMMU platforms

[Thomas Petazzoni]

This adds a patch to gdb 7.12 and gdb 8.x, which fixes the build on
noMMU platforms. It is not needed for older versions of gdb, since
it's related to the switch of gdb to C++ in the 7.12 release.

Fixes:

../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
       child_stack = xmalloc (STACK_SIZE * 4);

The patch has already been merged upstream, as of commit
ffce45d2243e5f52f411e314fc4e1a69f431a81f, and will therefore be part
of future gdb releases.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 ...-linux-ptrace.c-add-missing-gdb_byte-cast.patch | 41 ++++++++++++++++++++++
 ...-linux-ptrace.c-add-missing-gdb_byte-cast.patch | 41 ++++++++++++++++++++++
 2 files changed, 82 insertions(+)
 create mode 100644 package/gdb/7.12.1/0006-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
 create mode 100644 package/gdb/8.0/0004-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch

diff --git a/package/gdb/7.12.1/0006-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch b/package/gdb/7.12.1/0006-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
new file mode 100644
index 0000000..4eb72e4
--- /dev/null
+++ b/package/gdb/7.12.1/0006-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
@@ -0,0 +1,41 @@
+From 09a2c3e0164545324a1ddee70f5c9fdee71e2079 Mon Sep 17 00:00:00 2001
+From: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
+Date: Sun, 18 Jun 2017 23:09:43 +0200
+Subject: [PATCH] nat/linux-ptrace.c: add missing gdb_byte* cast
+
+On noMMU platforms, the following code gets compiled:
+
+  child_stack = xmalloc (STACK_SIZE * 4);
+
+Where child_stack is a gdb_byte*, and xmalloc() returns a void*. While
+the lack of cast is valid in C, it is not in C++, causing the
+following build failure:
+
+../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
+../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
+       child_stack = xmalloc (STACK_SIZE * 4);
+
+Therefore, this commit adds the appropriate cast.
+
+Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
+[Upstream commit: ffce45d2243e5f52f411e314fc4e1a69f431a81f]
+---
+ gdb/nat/linux-ptrace.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/gdb/nat/linux-ptrace.c b/gdb/nat/linux-ptrace.c
+index 3447e07..33833e2 100644
+--- a/gdb/nat/linux-ptrace.c
++++ b/gdb/nat/linux-ptrace.c
+@@ -270,7 +270,7 @@ linux_fork_to_function (gdb_byte *child_stack, int (*function) (void *))
+ #define STACK_SIZE 4096
+ 
+     if (child_stack == NULL)
+-      child_stack = xmalloc (STACK_SIZE * 4);
++      child_stack = (gdb_byte*) xmalloc (STACK_SIZE * 4);
+ 
+     /* Use CLONE_VM instead of fork, to support uClinux (no MMU).  */
+ #ifdef __ia64__
+-- 
+2.9.4
+
diff --git a/package/gdb/8.0/0004-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch b/package/gdb/8.0/0004-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
new file mode 100644
index 0000000..4eb72e4
--- /dev/null
+++ b/package/gdb/8.0/0004-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
@@ -0,0 +1,41 @@
+From 09a2c3e0164545324a1ddee70f5c9fdee71e2079 Mon Sep 17 00:00:00 2001
+From: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
+Date: Sun, 18 Jun 2017 23:09:43 +0200
+Subject: [PATCH] nat/linux-ptrace.c: add missing gdb_byte* cast
+
+On noMMU platforms, the following code gets compiled:
+
+  child_stack = xmalloc (STACK_SIZE * 4);
+
+Where child_stack is a gdb_byte*, and xmalloc() returns a void*. While
+the lack of cast is valid in C, it is not in C++, causing the
+following build failure:
+
+../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
+../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
+       child_stack = xmalloc (STACK_SIZE * 4);
+
+Therefore, this commit adds the appropriate cast.
+
+Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
+[Upstream commit: ffce45d2243e5f52f411e314fc4e1a69f431a81f]
+---
+ gdb/nat/linux-ptrace.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/gdb/nat/linux-ptrace.c b/gdb/nat/linux-ptrace.c
+index 3447e07..33833e2 100644
+--- a/gdb/nat/linux-ptrace.c
++++ b/gdb/nat/linux-ptrace.c
+@@ -270,7 +270,7 @@ linux_fork_to_function (gdb_byte *child_stack, int (*function) (void *))
+ #define STACK_SIZE 4096
+ 
+     if (child_stack == NULL)
+-      child_stack = xmalloc (STACK_SIZE * 4);
++      child_stack = (gdb_byte*) xmalloc (STACK_SIZE * 4);
+ 
+     /* Use CLONE_VM instead of fork, to support uClinux (no MMU).  */
+ #ifdef __ia64__
+-- 
+2.9.4
+
-- 
2.9.4

[Buildroot] [PATCH] gdb: add fix for gdb 7.12 and gdb 8.0 build on noMMU platforms

[Romain Naour]

Hi Thomas,

Le 03/08/2017 ? 09:15, Thomas Petazzoni a ?crit :
> This adds a patch to gdb 7.12 and gdb 8.x, which fixes the build on
> noMMU platforms. It is not needed for older versions of gdb, since
> it's related to the switch of gdb to C++ in the 7.12 release.
> 
> Fixes:
> 
> ../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
> ../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
>        child_stack = xmalloc (STACK_SIZE * 4);
> 
> The patch has already been merged upstream, as of commit
> ffce45d2243e5f52f411e314fc4e1a69f431a81f, and will therefore be part
> of future gdb releases.

What's the arch/config are you using?
It's seems no toolchain able to build gdb on noMMU case are present in the
autobuilder.

Best regards,
Romain

> 
> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> ---
>  ...-linux-ptrace.c-add-missing-gdb_byte-cast.patch | 41 ++++++++++++++++++++++
>  ...-linux-ptrace.c-add-missing-gdb_byte-cast.patch | 41 ++++++++++++++++++++++
>  2 files changed, 82 insertions(+)
>  create mode 100644 package/gdb/7.12.1/0006-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
>  create mode 100644 package/gdb/8.0/0004-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
> 
> diff --git a/package/gdb/7.12.1/0006-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch b/package/gdb/7.12.1/0006-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
> new file mode 100644
> index 0000000..4eb72e4
> --- /dev/null
> +++ b/package/gdb/7.12.1/0006-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
> @@ -0,0 +1,41 @@
> +From 09a2c3e0164545324a1ddee70f5c9fdee71e2079 Mon Sep 17 00:00:00 2001
> +From: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> +Date: Sun, 18 Jun 2017 23:09:43 +0200
> +Subject: [PATCH] nat/linux-ptrace.c: add missing gdb_byte* cast
> +
> +On noMMU platforms, the following code gets compiled:
> +
> +  child_stack = xmalloc (STACK_SIZE * 4);
> +
> +Where child_stack is a gdb_byte*, and xmalloc() returns a void*. While
> +the lack of cast is valid in C, it is not in C++, causing the
> +following build failure:
> +
> +../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
> +../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
> +       child_stack = xmalloc (STACK_SIZE * 4);
> +
> +Therefore, this commit adds the appropriate cast.
> +
> +Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> +[Upstream commit: ffce45d2243e5f52f411e314fc4e1a69f431a81f]
> +---
> + gdb/nat/linux-ptrace.c | 2 +-
> + 1 file changed, 1 insertion(+), 1 deletion(-)
> +
> +diff --git a/gdb/nat/linux-ptrace.c b/gdb/nat/linux-ptrace.c
> +index 3447e07..33833e2 100644
> +--- a/gdb/nat/linux-ptrace.c
> ++++ b/gdb/nat/linux-ptrace.c
> +@@ -270,7 +270,7 @@ linux_fork_to_function (gdb_byte *child_stack, int (*function) (void *))
> + #define STACK_SIZE 4096
> + 
> +     if (child_stack == NULL)
> +-      child_stack = xmalloc (STACK_SIZE * 4);
> ++      child_stack = (gdb_byte*) xmalloc (STACK_SIZE * 4);
> + 
> +     /* Use CLONE_VM instead of fork, to support uClinux (no MMU).  */
> + #ifdef __ia64__
> +-- 
> +2.9.4
> +
> diff --git a/package/gdb/8.0/0004-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch b/package/gdb/8.0/0004-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
> new file mode 100644
> index 0000000..4eb72e4
> --- /dev/null
> +++ b/package/gdb/8.0/0004-nat-linux-ptrace.c-add-missing-gdb_byte-cast.patch
> @@ -0,0 +1,41 @@
> +From 09a2c3e0164545324a1ddee70f5c9fdee71e2079 Mon Sep 17 00:00:00 2001
> +From: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> +Date: Sun, 18 Jun 2017 23:09:43 +0200
> +Subject: [PATCH] nat/linux-ptrace.c: add missing gdb_byte* cast
> +
> +On noMMU platforms, the following code gets compiled:
> +
> +  child_stack = xmalloc (STACK_SIZE * 4);
> +
> +Where child_stack is a gdb_byte*, and xmalloc() returns a void*. While
> +the lack of cast is valid in C, it is not in C++, causing the
> +following build failure:
> +
> +../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
> +../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
> +       child_stack = xmalloc (STACK_SIZE * 4);
> +
> +Therefore, this commit adds the appropriate cast.
> +
> +Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> +[Upstream commit: ffce45d2243e5f52f411e314fc4e1a69f431a81f]
> +---
> + gdb/nat/linux-ptrace.c | 2 +-
> + 1 file changed, 1 insertion(+), 1 deletion(-)
> +
> +diff --git a/gdb/nat/linux-ptrace.c b/gdb/nat/linux-ptrace.c
> +index 3447e07..33833e2 100644
> +--- a/gdb/nat/linux-ptrace.c
> ++++ b/gdb/nat/linux-ptrace.c
> +@@ -270,7 +270,7 @@ linux_fork_to_function (gdb_byte *child_stack, int (*function) (void *))
> + #define STACK_SIZE 4096
> + 
> +     if (child_stack == NULL)
> +-      child_stack = xmalloc (STACK_SIZE * 4);
> ++      child_stack = (gdb_byte*) xmalloc (STACK_SIZE * 4);
> + 
> +     /* Use CLONE_VM instead of fork, to support uClinux (no MMU).  */
> + #ifdef __ia64__
> +-- 
> +2.9.4
> +
> 

[Buildroot] [PATCH] gdb: add fix for gdb 7.12 and gdb 8.0 build on noMMU platforms

[Thomas Petazzoni]

Hello,

On Fri, 4 Aug 2017 22:43:26 +0200, Romain Naour wrote:

> > ../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
> > ../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
> >        child_stack = xmalloc (STACK_SIZE * 4);
> > 
> > The patch has already been merged upstream, as of commit
> > ffce45d2243e5f52f411e314fc4e1a69f431a81f, and will therefore be part
> > of future gdb releases.  
> 
> What's the arch/config are you using?

This was happening on m68k Coldfire, if i remember correctly.

> It's seems no toolchain able to build gdb on noMMU case are present in the
> autobuilder.

We have a m68k Coldfire configuration in the autobuilders. However, the
pre-built toolchain is built without BR2_PTHREAD_DEBUG, and therefore
the target gdb is never built because it depends on
BR2_TOOLCHAIN_HAS_THREADS_DEBUG.

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

[Buildroot] [PATCH] gdb: add fix for gdb 7.12 and gdb 8.0 build on noMMU platforms

[Romain Naour]

Hi Thomas,

Le 05/08/2017 ? 14:23, Thomas Petazzoni a ?crit :
> Hello,
> 
> On Fri, 4 Aug 2017 22:43:26 +0200, Romain Naour wrote:
> 
>>> ../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
>>> ../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
>>>        child_stack = xmalloc (STACK_SIZE * 4);
>>>
>>> The patch has already been merged upstream, as of commit
>>> ffce45d2243e5f52f411e314fc4e1a69f431a81f, and will therefore be part
>>> of future gdb releases.  
>>
>> What's the arch/config are you using?
> 
> This was happening on m68k Coldfire, if i remember correctly.
> 
>> It's seems no toolchain able to build gdb on noMMU case are present in the
>> autobuilder.
> 
> We have a m68k Coldfire configuration in the autobuilders. However, the
> pre-built toolchain is built without BR2_PTHREAD_DEBUG, and therefore
> the target gdb is never built because it depends on
> BR2_TOOLCHAIN_HAS_THREADS_DEBUG.

I got an error with gdbserver 8.0 with a m68k toolchain build today on master.

In file included from build-gnulib-gdbserver/import/stdint.h:556:0,
                 from ../common/common-defs.h:55,
                 from ../common/buffer.c:20:
/home/naourr/toolchains/nommu/host/m68k-buildroot-uclinux-uclibc/include/c++/6.4.0/cwchar:58:5:
erreur : conflicting declaration ? typedef struct rpl_mbstate_t rpl_mbstate_t ?
   } mbstate_t;
     ^
build-gnulib-gdbserver/import/wchar.h:486:13: note : previous declaration as
? typedef int rpl_mbstate_t ?
 typedef int rpl_mbstate_t;
             ^~~~~~~~~~~~~

The toolchain used the following defconfig:
BR2_m68k=y
BR2_m68k_cf5208=y
BR2_PTHREAD_DEBUG=y
BR2_TOOLCHAIN_BUILDROOT_CXX=y

gdbserver now require wchar support ?

Also I had to apply [1] to avoid the link issue when thread_db.h is included.

[1] http://patchwork.ozlabs.org/patch/798191/

Best regards,
Romain

> 
> Best regards,
> 
> Thomas
> 

[Buildroot] [PATCH] gdb: add fix for gdb 7.12 and gdb 8.0 build on noMMU platforms

[Arnout Vandecappelle]

On 03-08-17 09:15, Thomas Petazzoni wrote:
> This adds a patch to gdb 7.12 and gdb 8.x, which fixes the build on
> noMMU platforms. It is not needed for older versions of gdb, since
> it's related to the switch of gdb to C++ in the 7.12 release.
> 
> Fixes:
> 
> ../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
> ../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
>        child_stack = xmalloc (STACK_SIZE * 4);
> 
> The patch has already been merged upstream, as of commit
> ffce45d2243e5f52f411e314fc4e1a69f431a81f, and will therefore be part
> of future gdb releases.
> 
> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

 Since this is an upstream patch, and the additional issue reported by Romain is
something different, I applied to master, thanks.

 BTW, in another thread Thomas P. requested that a build fix which doesn't have
an accompanying autobuild failure should have a defconfig to reproduce :-).

 Regards,
 Arnout

-- 
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF

[Buildroot] [PATCH] gdb: add fix for gdb 7.12 and gdb 8.0 build on noMMU platforms

[Thomas Petazzoni]

Hello,

On Thu, 10 Aug 2017 11:25:27 +0200, Arnout Vandecappelle wrote:

> > Fixes:
> > 
> > ../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
> > ../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
> >        child_stack = xmalloc (STACK_SIZE * 4);
> > 
> > The patch has already been merged upstream, as of commit
> > ffce45d2243e5f52f411e314fc4e1a69f431a81f, and will therefore be part
> > of future gdb releases.
> > 
> > Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>  
> 
>  Since this is an upstream patch, and the additional issue reported by Romain is
> something different, I applied to master, thanks.
> 
>  BTW, in another thread Thomas P. requested that a build fix which doesn't have
> an accompanying autobuild failure should have a defconfig to reproduce :-).

Right, will do next time. You should have dropped my patch, I could
have done a respin with a defconfig included in the commit log.

Thanks!

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

[Buildroot] [PATCH] gdb: add fix for gdb 7.12 and gdb 8.0 build on noMMU platforms

[Peter Korsgaard]

>>>>> "Thomas" == Thomas Petazzoni <thomas.petazzoni@free-electrons.com> writes:

 > This adds a patch to gdb 7.12 and gdb 8.x, which fixes the build on
 > noMMU platforms. It is not needed for older versions of gdb, since
 > it's related to the switch of gdb to C++ in the 7.12 release.

 > Fixes:

 > ../nat/linux-ptrace.c: In function 'int linux_fork_to_function(gdb_byte*, int (*)(void*))':
 > ../nat/linux-ptrace.c:273:29: error: invalid conversion from 'void*' to 'gdb_byte* {aka unsigned char*}' [-fpermissive]
 >        child_stack = xmalloc (STACK_SIZE * 4);

 > The patch has already been merged upstream, as of commit
 > ffce45d2243e5f52f411e314fc4e1a69f431a81f, and will therefore be part
 > of future gdb releases.

 > Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Committed to 2017.02.x after dropping the gdb-8.x patch, thanks.

-- 
Bye, Peter Korsgaard