[gdbserver] compiling latest server.c (rev. 1.76) with MinGW for running on win32

[gdbserver] compiling latest server.c (rev. 1.76) with MinGW for running on win32

[Dr. Rolf Jansen]

[-- Attachment #1: Type: text/plain, Size: 2373 bytes --]

The latest server.c rev 1.76 does produce warnings when compiled with  
the MinGW tools:

server.c: In function 'start_inferior':
server.c:107: warning: implicit declaration of function 'alloca'
server.c:107: warning: incompatible implicit declaration of built-in  
function 'alloca'
server.c: In function 'handle_search_memory_1':
server.c:356: warning: implicit declaration of function 'memmem'
server.c:356: warning: assignment makes pointer from integer without a  
cast


In addition, linking of gdbserver.exe fails because of undefined  
references:

remote-utils.o:remote-utils.c:(.text+0x1057): undefined reference to  
`_disable_packet_Tthread'
server.o:server.c:(.text+0x9c6): undefined reference to  
`_disable_packet_qC'
server.o:server.c:(.text+0xa6f): undefined reference to  
`_disable_packet_qfThreadInfo'
server.o:server.c:(.text+0x20b4): undefined reference to  
`_disable_packet_vCont'
server.o:server.c:(.text+0x264a): undefined reference to  
`_disable_packet_vCont'
server.o:server.c:(.text+0x2670): undefined reference to  
`_disable_packet_Tthread'
server.o:server.c:(.text+0x2696): undefined reference to  
`_disable_packet_qC'
server.o:server.c:(.text+0x26bc): undefined reference to  
`_disable_packet_qfThreadInfo'
server.o:server.c:(.text+0x26df): undefined reference to  
`_disable_packet_vCont'
server.o:server.c:(.text+0x26e9): undefined reference to  
`_disable_packet_Tthread'
server.o:server.c:(.text+0x26f3): undefined reference to  
`_disable_packet_qC'
server.o:server.c:(.text+0x26fd): undefined reference to  
`_disable_packet_qfThreadInfo'
collect2: ld returned 1 exit status


MinGW defines alloca() in <malloc.h> and memmem() is not built-in. I  
think this should eventually be addressed in the configure script,  
however, as a quick fix, I added at the top of server.c:

#if USE_WIN32API
#include <malloc.h>
void *memmem (const void *haystack_start, size_t haystack_len, const  
void *needle_start, size_t needle_len);
#endif


In order to resolve the linking errors, the definition of the  
respective variables must be moved out of the conditional block  
(#ifdef SIGTTOU ... #endif), because they are used also within non- 
conditional code in remote-utils.c and in server.c.

I attached a .diff to this message, wich addresses both issues,  
although, there might be better ways to address 'em.

Best regards

Rolf Jansen


[-- Attachment #2: server.diff --]
[-- Type: application/octet-stream, Size: 1878 bytes --]

*** HEAD/src/gdb/gdbserver/server.c	   Fri Jul 18 14:07:19 2008
--- LOCAL/src/gdb/gdbserver/server.c	Fri Jul 18 14:07:01 2008
***************
*** 28,33 ****
--- 28,37 ----
  #if HAVE_SYS_WAIT_H
  #include <sys/wait.h>
  #endif
+ #if USE_WIN32API
+ #include <malloc.h>
+ void *memmem (const void *haystack_start, size_t haystack_len, const void *needle_start, size_t needle_len);
+ #endif
  
  unsigned long cont_thread;
  unsigned long general_thread;
*************** int terminal_fd;
*** 67,80 ****
  /* TERMINAL_FD's original foreground group.  */
  pid_t old_foreground_pgrp;
  
- /* Set if you want to disable optional thread related packets support
-    in gdbserver, for the sake of testing GDB against stubs that don't
-    support them.  */
- int disable_packet_vCont;
- int disable_packet_Tthread;
- int disable_packet_qC;
- int disable_packet_qfThreadInfo;
- 
  /* Hand back terminal ownership to the original foreground group.  */
  
  static void
--- 71,76 ----
*************** restore_old_foreground_pgrp (void)
*** 84,89 ****
--- 80,93 ----
  }
  #endif
  
+ /* Set if you want to disable optional thread related packets support
+    in gdbserver, for the sake of testing GDB against stubs that don't
+    support them.  */
+ int disable_packet_vCont = 0;
+ int disable_packet_Tthread = 0;
+ int disable_packet_qC = 0;
+ int disable_packet_qfThreadInfo = 0;
+ 
  static int
  target_running (void)
  {
*************** main (int argc, char *argv[])
*** 1299,1304 ****
--- 1303,1309 ----
  	  gdbserver_show_disableable (stdout);
  	  exit (0);
  	}
+ #ifdef SIGTTOU
        else if (strncmp (*next_arg,
  			"--disable-packet=",
  			sizeof ("--disable-packet=") - 1) == 0)
*************** main (int argc, char *argv[])
*** 1334,1339 ****
--- 1339,1345 ----
  		}
  	    }
  	}
+ #endif
        else
  	{
  	  fprintf (stderr, "Unknown argument: %s\n", *next_arg);

[-- Attachment #3: Type: text/plain, Size: 3 bytes --]

Re: [gdbserver] compiling latest server.c (rev. 1.76) with MinGW for running on win32

[Pedro Alves]

Hi Rolf,

On Friday 18 July 2008 18:32:32, Dr. Rolf Jansen wrote:
> The latest server.c rev 1.76 does produce warnings when compiled with  
> the MinGW tools:
>
> server.c: In function 'start_inferior':
> server.c:107: warning: implicit declaration of function 'alloca'
> server.c:107: warning: incompatible implicit declaration of built-in  
> function 'alloca'
> server.c: In function 'handle_search_memory_1':
> server.c:356: warning: implicit declaration of function 'memmem'
> server.c:356: warning: assignment makes pointer from integer without a  
> cast

> MinGW defines alloca() in <malloc.h> and memmem() is not built-in. I  
> think this should eventually be addressed in the configure script,  
> however, as a quick fix, I added at the top of server.c:


> #if USE_WIN32API
> #include <malloc.h>

This bit be done with...

#if HAVE_MALLOC_H
#include <malloc.h>
#endif

... instead.  We're already doing that in utils.c.


> void *memmem (const void *haystack_start, size_t haystack_len, const  
> void *needle_start, size_t needle_len);
> #endif

Hmmm, shouldn't we be picking up memmem's definition from gnulib's string.h,
since we're using memmem from gnulib?

> In addition, linking of gdbserver.exe fails because of undefined  
> references:
>
> remote-utils.o:remote-utils.c:(.text+0x1057): undefined reference to  
> `_disable_packet_Tthread'

> In order to resolve the linking errors, the definition of the  
> respective variables must be moved out of the conditional block  
> (#ifdef SIGTTOU ... #endif), because they are used also within non-
> conditional code in remote-utils.c and in server.c.
>

Ooops.  Yes, it should move out of the #if block.  It was just an
oversight.  No need to add the SIGTTOU checks around the
--disable-packet handling.

-- 
Pedro Alves

Re: [gdbserver] compiling latest server.c (rev. 1.76) with MinGW for running on win32

[Dr. Rolf Jansen]

Hi Pedro,

many thanks for the reply!

Am 18.07.2008 um 16:32 schrieb Pedro Alves:

> Hi Rolf,
>
> On Friday 18 July 2008 18:32:32, Dr. Rolf Jansen wrote:
>> The latest server.c rev 1.76 does produce warnings when compiled with
>> the MinGW tools:
>>
>> ...
>>
>> MinGW defines alloca() in <malloc.h> and memmem() is not built-in. I
>> think this should eventually be addressed in the configure script,
>> however, as a quick fix, I added at the top of server.c:
>>
>>
>> #if USE_WIN32API
>> #include <malloc.h>
>
> This bit be done with...
>
> #if HAVE_MALLOC_H
> #include <malloc.h>
> #endif
>
> ... instead.  We're already doing that in utils.c.

YES, I can confirm this, this works for server.c too.

Great!

>> void *memmem (const void *haystack_start, size_t haystack_len, const
>> void *needle_start, size_t needle_len);
>> #endif
>
> Hmmm, shouldn't we be picking up memmem's definition from gnulib's  
> string.h,
> since we're using memmem from gnulib?

Ah, now I got it.

I fear gnulib's string.h will not work in my case, because I am  
compiling gdb at a Mac OS X box using the Mac gcc-4.2-tool chain with  
host=powerpc-apple-darwin9.4.0 and target=i386-mingw32msvc. And from  
the same source tree I am cross-compiling gdbserver using the MinGW- 
gcc-4.3.1-tool chain with build=powerpc-apple-darwin9.4.0 and host/ 
target=i386-pc-mingw32msvc.

Since gnulib belongs to the gdb source tree, and Mac OS X has memmem()  
and the other stuff built-in, gnulib's string.h would never be built.

This kind of issue with gnulib when compiling gdb and cross-compiling  
gdbserver was discussed recently already, and it only came back into  
my mind now. So, I am sorry about reporting a false issue.

In my local code I simply keep the declaration of memmem(), however it  
is probably not a good idea to patch HEAD.server.c with that.



>> In order to resolve the linking errors, the definition of the
>> respective variables must be moved out of the conditional block
>> (#ifdef SIGTTOU ... #endif), because they are used also within non-
>> conditional code in remote-utils.c and in server.c.
> Ooops.  Yes, it should move out of the #if block.  It was just an
> oversight.  No need to add the SIGTTOU checks around the
> --disable-packet handling.

OK, I took it out, and I confirm that this works too.

Best regards

Rolf Jansen

Re: [gdbserver] compiling latest server.c (rev. 1.76) with MinGW for running on win32

[Pedro Alves]

[-- Attachment #1: Type: text/plain, Size: 169 bytes --]

Hi Rolf,

Sorry for dropping out on you.

I checked that gdbserver now builds OK with the attached patch
on mingw32 and linux.

Daniel, OK to check in?

-- 
Pedro Alves

[-- Attachment #2: memmem.diff --]
[-- Type: text/x-diff, Size: 2892 bytes --]

2008-07-31  Rolf Jansen  <rj@surtec.com>
	    Pedro Alves  <pedro@codesourcery.com>

	* configure.ac: Check for memmem declaration.
	* server.c [HAVE_MALLOC_H]: Include malloc.h.
	(disable_packet_vCont, disable_packet_Tthread, disable_packet_qC)
	(disable_packet_qfThreadInfo): Unconditionally compile.
	* server.h [!HAVE_DECL_MEMMEM]: Declare memmem.
	* configure, config.in: Regenerate.

---
 gdb/gdbserver/configure.ac |    2 +-
 gdb/gdbserver/server.c     |   19 +++++++++++--------
 gdb/gdbserver/server.h     |    4 ++++
 3 files changed, 16 insertions(+), 9 deletions(-)

Index: src/gdb/gdbserver/configure.ac
===================================================================
--- src.orig/gdb/gdbserver/configure.ac	2008-06-06 12:25:11.000000000 +0100
+++ src/gdb/gdbserver/configure.ac	2008-07-31 17:10:46.000000000 +0100
@@ -61,7 +61,7 @@ AC_TRY_LINK([
   [AC_MSG_RESULT(no)])
 fi
 
-AC_CHECK_DECLS([strerror, perror])
+AC_CHECK_DECLS([strerror, perror, memmem])
 
 AC_CHECK_TYPES(socklen_t, [], [],
 [#include <sys/types.h>
Index: src/gdb/gdbserver/server.c
===================================================================
--- src.orig/gdb/gdbserver/server.c	2008-07-08 00:05:33.000000000 +0100
+++ src/gdb/gdbserver/server.c	2008-07-31 17:10:46.000000000 +0100
@@ -28,6 +28,9 @@
 #if HAVE_SYS_WAIT_H
 #include <sys/wait.h>
 #endif
+#if HAVE_MALLOC_H
+#include <malloc.h>
+#endif
 
 unsigned long cont_thread;
 unsigned long general_thread;
@@ -67,14 +70,6 @@ int terminal_fd;
 /* TERMINAL_FD's original foreground group.  */
 pid_t old_foreground_pgrp;
 
-/* Set if you want to disable optional thread related packets support
-   in gdbserver, for the sake of testing GDB against stubs that don't
-   support them.  */
-int disable_packet_vCont;
-int disable_packet_Tthread;
-int disable_packet_qC;
-int disable_packet_qfThreadInfo;
-
 /* Hand back terminal ownership to the original foreground group.  */
 
 static void
@@ -84,6 +79,14 @@ restore_old_foreground_pgrp (void)
 }
 #endif
 
+/* Set if you want to disable optional thread related packets support
+   in gdbserver, for the sake of testing GDB against stubs that don't
+   support them.  */
+int disable_packet_vCont;
+int disable_packet_Tthread;
+int disable_packet_qC;
+int disable_packet_qfThreadInfo;
+
 static int
 target_running (void)
 {
Index: src/gdb/gdbserver/server.h
===================================================================
--- src.orig/gdb/gdbserver/server.h	2008-07-02 12:35:07.000000000 +0100
+++ src/gdb/gdbserver/server.h	2008-07-31 17:10:46.000000000 +0100
@@ -50,6 +50,10 @@ extern void perror (const char *);
 #endif
 #endif
 
+#if !HAVE_DECL_MEMMEM
+extern void *memmem (const void *, size_t , const void *, size_t);
+#endif
+
 #ifndef ATTR_NORETURN
 #if defined(__GNUC__) && (__GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 7))
 #define ATTR_NORETURN __attribute__ ((noreturn))

Re: [gdbserver] compiling latest server.c (rev. 1.76) with MinGW for running on win32

[Daniel Jacobowitz]

On Thu, Jul 31, 2008 at 05:18:40PM +0100, Pedro Alves wrote:
> Hi Rolf,
> 
> Sorry for dropping out on you.
> 
> I checked that gdbserver now builds OK with the attached patch
> on mingw32 and linux.
> 
> Daniel, OK to check in?

This is OK, thanks.

-- 
Daniel Jacobowitz
CodeSourcery