[Qemu-devel] [Patch] Make usermode stacksize (-s) configurable at compile-time

[Qemu-devel] [Patch] Make usermode stacksize (-s) configurable at compile-time

[Jan-Simon Möller]

[-- Attachment #1: Type: text/plain, Size: 2935 bytes --]

Hi !

We encontered problems with the low default value for the stacksize in usermode (ok, whats "low"). 
For environments like scratchbox, its hard to use "-s" as qemu is called by binfmt mechanism.
The attached patch makes this configurable at compile-time.

Signed-off-by: Jan-Simon Möller <dl9pf@gmx.de>


diff --git a/configure b/configure                                
index 43d87c5..076f45a 100755                                     
--- a/configure                                                   
+++ b/configure                                                   
@@ -220,6 +220,7 @@ uname_release=""                              
 io_thread="no"                                                   
 mixemu="no"                                                      
 kerneldir=""                                                     
+user_mode_stacksize=""                                           
 aix="no"                                                         
 blobs="yes"                                                      
 pkgversion=""                                                    
@@ -557,6 +558,8 @@ for opt do                                    
   ;;                                                             
   --kerneldir=*) kerneldir="$optarg"                             
   ;;
+  --user-mode-stacksize=*) user_mode_stacksize="$optarg"
+  ;;
   --with-pkgversion=*) pkgversion=" ($optarg)"
   ;;
   --disable-docs) docs="no"
@@ -713,6 +716,7 @@ echo "  --enable-linux-aio       enable Linux AIO support"
 echo "  --enable-io-thread       enable IO thread"
 echo "  --disable-blobs          disable installing provided firmware blobs"
 echo "  --kerneldir=PATH         look for kernel includes in PATH"
+echo "  --user-mode-stacksize=   set default stack size in bytes (as -s, only in usermode)"
 echo ""
 echo "NOTE: The object files are built at the place where configure is launched"
 exit 1
@@ -2576,3 +2580,10 @@ d=libuser
 mkdir -p $d
 rm -f $d/Makefile
 ln -s $source_path/Makefile.user $d/Makefile
+
+# change default stacksize in usermode
+#
+if test "$user_mode_stacksize" != "" ; then
+ echo "user_mode_stack   $user_mode_stacksize"
+ echo "QEMU_CFLAGS+=-DUSER_MODE_STACKSIZE=$user_mode_stacksize" >> $config_host_mak
+fi
diff --git a/linux-user/main.c b/linux-user/main.c
index 81a1ada..30c0a87 100644
--- a/linux-user/main.c
+++ b/linux-user/main.c
@@ -51,7 +51,11 @@ const char *qemu_uname_release = CONFIG_UNAME_RELEASE;
 /* XXX: on x86 MAP_GROWSDOWN only works if ESP <= address + 32, so
    we allocate a bigger stack. Need a better solution, for example
    by remapping the process stack directly at the right place */
+#if defined(USER_MODE_STACKSIZE)
+unsigned long x86_stack_size = USER_MODE_STACKSIZE;
+#else
 unsigned long x86_stack_size = 512 * 1024;
+#endif

 void gemu_log(const char *fmt, ...)
 {


[-- Attachment #2: usermode_stacksize.diff --]
[-- Type: text/x-patch, Size: 1867 bytes --]

diff --git a/configure b/configure
index 43d87c5..076f45a 100755
--- a/configure
+++ b/configure
@@ -220,6 +220,7 @@ uname_release=""
 io_thread="no"
 mixemu="no"
 kerneldir=""
+user_mode_stacksize=""
 aix="no"
 blobs="yes"
 pkgversion=""
@@ -557,6 +558,8 @@ for opt do
   ;;
   --kerneldir=*) kerneldir="$optarg"
   ;;
+  --user-mode-stacksize=*) user_mode_stacksize="$optarg"
+  ;;
   --with-pkgversion=*) pkgversion=" ($optarg)"
   ;;
   --disable-docs) docs="no"
@@ -713,6 +716,7 @@ echo "  --enable-linux-aio       enable Linux AIO support"
 echo "  --enable-io-thread       enable IO thread"
 echo "  --disable-blobs          disable installing provided firmware blobs"
 echo "  --kerneldir=PATH         look for kernel includes in PATH"
+echo "  --user-mode-stacksize=   set default stack size in bytes (as -s, only in usermode)"
 echo ""
 echo "NOTE: The object files are built at the place where configure is launched"
 exit 1
@@ -2576,3 +2580,10 @@ d=libuser
 mkdir -p $d
 rm -f $d/Makefile
 ln -s $source_path/Makefile.user $d/Makefile
+
+# change default stacksize in usermode
+#
+if test "$user_mode_stacksize" != "" ; then
+ echo "user_mode_stack   $user_mode_stacksize"
+ echo "QEMU_CFLAGS+=-DUSER_MODE_STACKSIZE=$user_mode_stacksize" >> $config_host_mak
+fi
diff --git a/linux-user/main.c b/linux-user/main.c
index 81a1ada..30c0a87 100644
--- a/linux-user/main.c
+++ b/linux-user/main.c
@@ -51,7 +51,11 @@ const char *qemu_uname_release = CONFIG_UNAME_RELEASE;
 /* XXX: on x86 MAP_GROWSDOWN only works if ESP <= address + 32, so
    we allocate a bigger stack. Need a better solution, for example
    by remapping the process stack directly at the right place */
+#if defined(USER_MODE_STACKSIZE)
+unsigned long x86_stack_size = USER_MODE_STACKSIZE;
+#else
 unsigned long x86_stack_size = 512 * 1024;
+#endif
 
 void gemu_log(const char *fmt, ...)
 {

Re: [Qemu-devel] [Patch] Make usermode stacksize (-s) configurable at compile-time

[Laurent Desnogues]

2009/10/25 Jan-Simon Möller <dl9pf@gmx.de>:
> Hi !
>
> We encontered problems with the low default value for the stacksize in usermode (ok, whats "low").
> For environments like scratchbox, its hard to use "-s" as qemu is called by binfmt mechanism.
> The attached patch makes this configurable at compile-time.
>
> Signed-off-by: Jan-Simon Möller <dl9pf@gmx.de>
>
>
> diff --git a/configure b/configure
> index 43d87c5..076f45a 100755
> --- a/configure
> +++ b/configure
> @@ -220,6 +220,7 @@ uname_release=""
>  io_thread="no"
>  mixemu="no"
>  kerneldir=""
> +user_mode_stacksize=""
>  aix="no"
>  blobs="yes"
>  pkgversion=""
> @@ -557,6 +558,8 @@ for opt do
>   ;;
>   --kerneldir=*) kerneldir="$optarg"
>   ;;
> +  --user-mode-stacksize=*) user_mode_stacksize="$optarg"
> +  ;;
>   --with-pkgversion=*) pkgversion=" ($optarg)"
>   ;;
>   --disable-docs) docs="no"
> @@ -713,6 +716,7 @@ echo "  --enable-linux-aio       enable Linux AIO support"
>  echo "  --enable-io-thread       enable IO thread"
>  echo "  --disable-blobs          disable installing provided firmware blobs"
>  echo "  --kerneldir=PATH         look for kernel includes in PATH"
> +echo "  --user-mode-stacksize=   set default stack size in bytes (as -s, only in usermode)"
>  echo ""
>  echo "NOTE: The object files are built at the place where configure is launched"
>  exit 1
> @@ -2576,3 +2580,10 @@ d=libuser
>  mkdir -p $d
>  rm -f $d/Makefile
>  ln -s $source_path/Makefile.user $d/Makefile
> +
> +# change default stacksize in usermode
> +#
> +if test "$user_mode_stacksize" != "" ; then
> + echo "user_mode_stack   $user_mode_stacksize"
> + echo "QEMU_CFLAGS+=-DUSER_MODE_STACKSIZE=$user_mode_stacksize" >> $config_host_mak
> +fi

Wouldn't it be better to set this earlier only if target_user_only
is set to yes and use cflags?

Something like this:

if test "$target_user_only" = "yes" -a "$user_mode_stacksize" != ""; then
  cflags="-DUSER_MODE_STACKSIZE $cflags"
fi

just after the test for pie (line 2478) for instance.

The informative echo should go with the others (line 1853).


> diff --git a/linux-user/main.c b/linux-user/main.c
> index 81a1ada..30c0a87 100644
> --- a/linux-user/main.c
> +++ b/linux-user/main.c
> @@ -51,7 +51,11 @@ const char *qemu_uname_release = CONFIG_UNAME_RELEASE;
>  /* XXX: on x86 MAP_GROWSDOWN only works if ESP <= address + 32, so
>    we allocate a bigger stack. Need a better solution, for example
>    by remapping the process stack directly at the right place */
> +#if defined(USER_MODE_STACKSIZE)
> +unsigned long x86_stack_size = USER_MODE_STACKSIZE;
> +#else
>  unsigned long x86_stack_size = 512 * 1024;
> +#endif

I'd prefer:

#ifndef USER_MODE_STACKSIZE
#define x86_stack_size (512 * 1024)
#endif
unsigned long x86_stack_size = USER_MODE_STACKSIZE;


Laurent

>  void gemu_log(const char *fmt, ...)
>  {
>
>

Re: [Qemu-devel] [Patch] Make usermode stacksize (-s) configurable at compile-time

[Jan-Simon Möller]

Am Sonntag 25 Oktober 2009 16:54:16 schrieb Laurent Desnogues:
> Wouldn't it be better to set this earlier only if target_user_only
> is set to yes and use cflags?
> 
> Something like this:
> 
> if test "$target_user_only" = "yes" -a "$user_mode_stacksize" != ""; then
>   cflags="-DUSER_MODE_STACKSIZE $cflags"
> fi
> 
> just after the test for pie (line 2478) for instance.
Done. 
> The informative echo should go with the others (line 1853).
Its not clear at this stage, if its "default" or the cmdline value, 
as $target_user_only isn't evaluated until line 2171 .

I'd leave therefore the echo in line 2491.
> 
>
> I'd prefer:
> 
> #ifndef USER_MODE_STACKSIZE
> #define x86_stack_size (512 * 1024)
> #endif
> unsigned long x86_stack_size = USER_MODE_STACKSIZE;
> 
You mean probably?
#ifndef USER_MODE_STACKSIZE
#define USER_MODE_STACKSIZE (512 * 1024)
#endif
unsigned long x86_stack_size = USER_MODE_STACKSIZE;

Done.

Patch will follow in next mail as reply.

Tnx and have phun !

Jan-Simon

Re: [Qemu-devel] [Patch] Make usermode stacksize (-s) configurable at compile-time

[Jan-Simon Möller]

[-- Attachment #1: Type: text/plain, Size: 3500 bytes --]

We encontered problems with the low default value for the stacksize in usermode (ok, whats "low"). 
For environments like scratchbox, its hard to use "-s" as qemu is called by binfmt mechanism.
The attached patch makes this configurable at compile-time.

Signed-off-by: Jan-Simon Möller <dl9pf@gmx.de>

diff --git a/configure b/configure                                     
index 43d87c5..dad9175 100755                                          
--- a/configure                                                        
+++ b/configure                                                        
@@ -220,6 +220,7 @@ uname_release=""                                   
 io_thread="no"                                                        
 mixemu="no"                                                           
 kerneldir=""                                                          
+user_mode_stacksize=""                                                
 aix="no"                                                              
 blobs="yes"                                                           
 pkgversion=""                                                         
@@ -557,6 +558,8 @@ for opt do                                         
   ;;                                                                  
   --kerneldir=*) kerneldir="$optarg"                                  
   ;;                                                                  
+  --user-mode-stacksize=*) user_mode_stacksize="$optarg"              
+  ;;                                                                  
   --with-pkgversion=*) pkgversion=" ($optarg)"                        
   ;;                                                                  
   --disable-docs) docs="no"                                           
@@ -713,6 +716,7 @@ echo "  --enable-linux-aio       enable Linux AIO support"
 echo "  --enable-io-thread       enable IO thread"
 echo "  --disable-blobs          disable installing provided firmware blobs"
 echo "  --kerneldir=PATH         look for kernel includes in PATH"
+echo "  --user-mode-stacksize=   set default stack size in bytes (as -s, only in usermode)"
 echo ""
 echo "NOTE: The object files are built at the place where configure is launched"
 exit 1
@@ -2481,6 +2485,16 @@ if test "$target_user_only" = "yes" -a "$static" = "no" -a \
   ldflags="-pie $ldflags"
 fi

+# change default stacksize in usermode
+#
+if test "$target_user_only" = "yes" -a "$user_mode_stacksize" != "" ; then
+ cflags="-DUSER_MODE_STACKSIZE=$user_mode_stacksize $cflags"
+ echo "user_mode_stack   $user_mode_stacksize"
+else
+ echo "user_mode_stack   default"
+fi
+
+
 if test "$target_softmmu" = "yes" -a \( \
         "$TARGET_ARCH" = "microblaze" -o \
         "$TARGET_ARCH" = "cris" \) ; then
diff --git a/linux-user/main.c b/linux-user/main.c
index 81a1ada..9ac2421 100644
--- a/linux-user/main.c
+++ b/linux-user/main.c
@@ -51,7 +51,10 @@ const char *qemu_uname_release = CONFIG_UNAME_RELEASE;
 /* XXX: on x86 MAP_GROWSDOWN only works if ESP <= address + 32, so
    we allocate a bigger stack. Need a better solution, for example
    by remapping the process stack directly at the right place */
-unsigned long x86_stack_size = 512 * 1024;
+#ifndef USER_MODE_STACKSIZE
+#define USER_MODE_STACKSIZE (512 * 1024)
+#endif
+unsigned long x86_stack_size = USER_MODE_STACKSIZE;

 void gemu_log(const char *fmt, ...)
 {


[-- Attachment #2: usermode_stacksize.diff --]
[-- Type: text/x-patch, Size: 2084 bytes --]

diff --git a/configure b/configure
index 43d87c5..dad9175 100755
--- a/configure
+++ b/configure
@@ -220,6 +220,7 @@ uname_release=""
 io_thread="no"
 mixemu="no"
 kerneldir=""
+user_mode_stacksize=""
 aix="no"
 blobs="yes"
 pkgversion=""
@@ -557,6 +558,8 @@ for opt do
   ;;
   --kerneldir=*) kerneldir="$optarg"
   ;;
+  --user-mode-stacksize=*) user_mode_stacksize="$optarg"
+  ;;
   --with-pkgversion=*) pkgversion=" ($optarg)"
   ;;
   --disable-docs) docs="no"
@@ -713,6 +716,7 @@ echo "  --enable-linux-aio       enable Linux AIO support"
 echo "  --enable-io-thread       enable IO thread"
 echo "  --disable-blobs          disable installing provided firmware blobs"
 echo "  --kerneldir=PATH         look for kernel includes in PATH"
+echo "  --user-mode-stacksize=   set default stack size in bytes (as -s, only in usermode)"
 echo ""
 echo "NOTE: The object files are built at the place where configure is launched"
 exit 1
@@ -2481,6 +2485,16 @@ if test "$target_user_only" = "yes" -a "$static" = "no" -a \
   ldflags="-pie $ldflags"
 fi
 
+# change default stacksize in usermode
+#
+if test "$target_user_only" = "yes" -a "$user_mode_stacksize" != "" ; then
+ cflags="-DUSER_MODE_STACKSIZE=$user_mode_stacksize $cflags"
+ echo "user_mode_stack   $user_mode_stacksize"
+else
+ echo "user_mode_stack   default"
+fi
+
+
 if test "$target_softmmu" = "yes" -a \( \
         "$TARGET_ARCH" = "microblaze" -o \
         "$TARGET_ARCH" = "cris" \) ; then
diff --git a/linux-user/main.c b/linux-user/main.c
index 81a1ada..9ac2421 100644
--- a/linux-user/main.c
+++ b/linux-user/main.c
@@ -51,7 +51,10 @@ const char *qemu_uname_release = CONFIG_UNAME_RELEASE;
 /* XXX: on x86 MAP_GROWSDOWN only works if ESP <= address + 32, so
    we allocate a bigger stack. Need a better solution, for example
    by remapping the process stack directly at the right place */
-unsigned long x86_stack_size = 512 * 1024;
+#ifndef USER_MODE_STACKSIZE
+#define USER_MODE_STACKSIZE (512 * 1024)
+#endif
+unsigned long x86_stack_size = USER_MODE_STACKSIZE;
 
 void gemu_log(const char *fmt, ...)
 {

[Qemu-devel] Re: [Patch] Make usermode stacksize (-s) configurable at compile-time

[Juan Quintela]

"Jan-Simon Möller" <dl9pf@gmx.de> wrote:
> @@ -2481,6 +2485,16 @@ if test "$target_user_only" = "yes" -a "$static" = "no" -a \
>    ldflags="-pie $ldflags"
>  fi
>
> +# change default stacksize in usermode
> +#
> +if test "$target_user_only" = "yes" -a "$user_mode_stacksize" != "" ; then
> + cflags="-DUSER_MODE_STACKSIZE=$user_mode_stacksize $cflags"
> + echo "user_mode_stack   $user_mode_stacksize"

Put this line together with:
  echo "PIE user targets  $user_pie"

> +else
> + echo "user_mode_stack   default"
> +fi
> +
> +
>  if test "$target_softmmu" = "yes" -a \( \
>          "$TARGET_ARCH" = "microblaze" -o \
>          "$TARGET_ARCH" = "cris" \) ; then
> diff --git a/linux-user/main.c b/linux-user/main.c
> index 81a1ada..9ac2421 100644
> --- a/linux-user/main.c
> +++ b/linux-user/main.c
> @@ -51,7 +51,10 @@ const char *qemu_uname_release = CONFIG_UNAME_RELEASE;
>  /* XXX: on x86 MAP_GROWSDOWN only works if ESP <= address + 32, so
>     we allocate a bigger stack. Need a better solution, for example
>     by remapping the process stack directly at the right place */
> -unsigned long x86_stack_size = 512 * 1024;
> +#ifndef USER_MODE_STACKSIZE
> +#define USER_MODE_STACKSIZE (512 * 1024)
> +#endif

If we are changing the code in configure anyways, why don't add the
default there?

Something like this (I tested that it compiled, but didn't tested
usermode).  I am assuming that you only want it for linux-user.

>From 0dc910f5d1b2e1dd7120abf32d25db190cd394e5 Mon Sep 17 00:00:00 2001
From: Juan Quintela <quintela@redhat.com>
Date: Sun, 25 Oct 2009 22:06:52 +0100
Subject: [PATCH] User mode stacksize implementation


Signed-off-by: Juan Quintela <quintela@redhat.com>
---
 configure         |    5 +++++
 linux-user/main.c |    2 +-
 2 files changed, 6 insertions(+), 1 deletions(-)

diff --git a/configure b/configure
index 43d87c5..82b9825 100755
--- a/configure
+++ b/configure
@@ -225,6 +225,7 @@ blobs="yes"
 pkgversion=""
 check_utests="no"
 user_pie="no"
+user_mode_stacksize="512*1024"

 # OS specific
 if check_define __linux__ ; then
@@ -517,6 +518,8 @@ for opt do
   ;;
   --disable-user-pie) user_pie="no"
   ;;
+  --user-mode-stacksize=*) user_mode_stacksize="$optarg"
+  ;;
   --enable-uname-release=*) uname_release="$optarg"
   ;;
   --sparc_cpu=*)
@@ -1841,6 +1844,7 @@ echo "uname -r          $uname_release"
 echo "NPTL support      $nptl"
 echo "GUEST_BASE        $guest_base"
 echo "PIE user targets  $user_pie"
+echo "User mode stacksize  $user_mode_stacksize"
 echo "vde support       $vde"
 echo "IO thread         $io_thread"
 echo "Linux AIO support $linux_aio"
@@ -2361,6 +2365,7 @@ if test "$target_user_only" = "yes" ; then
 fi
 if test "$target_linux_user" = "yes" ; then
   echo "CONFIG_LINUX_USER=y" >> $config_target_mak
+  echo "CONFIG_USER_MODE_STACKSIZE=$user_mode_stacksize" >> $config_target_mak
 fi
 if test "$target_darwin_user" = "yes" ; then
   echo "CONFIG_DARWIN_USER=y" >> $config_target_mak
diff --git a/linux-user/main.c b/linux-user/main.c
index 81a1ada..5d0f849 100644
--- a/linux-user/main.c
+++ b/linux-user/main.c
@@ -51,7 +51,7 @@ const char *qemu_uname_release = CONFIG_UNAME_RELEASE;
 /* XXX: on x86 MAP_GROWSDOWN only works if ESP <= address + 32, so
    we allocate a bigger stack. Need a better solution, for example
    by remapping the process stack directly at the right place */
-unsigned long x86_stack_size = 512 * 1024;
+unsigned long x86_stack_size = CONFIG_USER_MODE_STACKSIZE;

 void gemu_log(const char *fmt, ...)
 {
-- 
1.6.2.5

Re: [Qemu-devel] Re: [Patch] Make usermode stacksize (-s) configurable at compile-time

[Jan-Simon Möller]

Am Sonntag 25 Oktober 2009 22:09:10 schrieb Juan Quintela:
> If we are changing the code in configure anyways, why don't add the
> default there?
> 
> Something like this (I tested that it compiled, but didn't tested
> usermode).  I am assuming that you only want it for linux-user.

I like this version.

Signed-off-by: Jan-Simon Möller <dl9pf@gmx.de>

> 
> >From 0dc910f5d1b2e1dd7120abf32d25db190cd394e5 Mon Sep 17 00:00:00 2001
> From: Juan Quintela <quintela@redhat.com>
> Date: Sun, 25 Oct 2009 22:06:52 +0100
> Subject: [PATCH] User mode stacksize implementation
> 

Best,
Jan-Simon

Re: [Qemu-devel] [Patch] Make usermode stacksize (-s) configurable at compile-time

[Jamie Lokier]

Jan-Simon Möller wrote:
> We encontered problems with the low default value for the stacksize in usermode (ok, whats "low"). 
> For environments like scratchbox, its hard to use "-s" as qemu is called by binfmt mechanism.
> The attached patch makes this configurable at compile-time.

Note, this will only change the stack size for the main thread, at
least with LinuxThreads.  Other thread stacks are allocated by the
pthread (LinuxThreads or NPTL), and their stack size is changed by
pthread_attr_setstacksize instead.

-- Jamie

Re: [Qemu-devel] [Patch] Make usermode stacksize (-s) configurable at compile-time

[Riku Voipio]

On Sun, Oct 25, 2009 at 06:49:44PM +0100, Jan-Simon Möller wrote:
> We encontered problems with the low default value for the stacksize in usermode (ok, whats "low"). 
> For environments like scratchbox, its hard to use "-s" as qemu is called by binfmt mechanism.
> The attached patch makes this configurable at compile-time.

fwiw in scratchbox we have a wrapper in between binfmt and qemu (called misc_runner) that sets any
necessary command line parameters for qemu. You might want to consider a similar approach so we
don't need to add a configure option (and thus hardcode the setting) to qemu for every command
line option that might need changing.

> Signed-off-by: Jan-Simon Möller <dl9pf@gmx.de>
> 
> diff --git a/configure b/configure                                     
> index 43d87c5..dad9175 100755                                          
> --- a/configure                                                        
> +++ b/configure                                                        
> @@ -220,6 +220,7 @@ uname_release=""                                   
>  io_thread="no"                                                        
>  mixemu="no"                                                           
>  kerneldir=""                                                          
> +user_mode_stacksize=""                                                
>  aix="no"                                                              
>  blobs="yes"                                                           
>  pkgversion=""                                                         
> @@ -557,6 +558,8 @@ for opt do                                         
>    ;;                                                                  
>    --kerneldir=*) kerneldir="$optarg"                                  
>    ;;                                                                  
> +  --user-mode-stacksize=*) user_mode_stacksize="$optarg"              
> +  ;;                                                                  
>    --with-pkgversion=*) pkgversion=" ($optarg)"                        
>    ;;                                                                  
>    --disable-docs) docs="no"                                           
> @@ -713,6 +716,7 @@ echo "  --enable-linux-aio       enable Linux AIO support"
>  echo "  --enable-io-thread       enable IO thread"
>  echo "  --disable-blobs          disable installing provided firmware blobs"
>  echo "  --kerneldir=PATH         look for kernel includes in PATH"
> +echo "  --user-mode-stacksize=   set default stack size in bytes (as -s, only in usermode)"
>  echo ""
>  echo "NOTE: The object files are built at the place where configure is launched"
>  exit 1
> @@ -2481,6 +2485,16 @@ if test "$target_user_only" = "yes" -a "$static" = "no" -a \
>    ldflags="-pie $ldflags"
>  fi
> 
> +# change default stacksize in usermode
> +#
> +if test "$target_user_only" = "yes" -a "$user_mode_stacksize" != "" ; then
> + cflags="-DUSER_MODE_STACKSIZE=$user_mode_stacksize $cflags"
> + echo "user_mode_stack   $user_mode_stacksize"
> +else
> + echo "user_mode_stack   default"
> +fi
> +
> +
>  if test "$target_softmmu" = "yes" -a \( \
>          "$TARGET_ARCH" = "microblaze" -o \
>          "$TARGET_ARCH" = "cris" \) ; then
> diff --git a/linux-user/main.c b/linux-user/main.c
> index 81a1ada..9ac2421 100644
> --- a/linux-user/main.c
> +++ b/linux-user/main.c
> @@ -51,7 +51,10 @@ const char *qemu_uname_release = CONFIG_UNAME_RELEASE;
>  /* XXX: on x86 MAP_GROWSDOWN only works if ESP <= address + 32, so
>     we allocate a bigger stack. Need a better solution, for example
>     by remapping the process stack directly at the right place */
> -unsigned long x86_stack_size = 512 * 1024;
> +#ifndef USER_MODE_STACKSIZE
> +#define USER_MODE_STACKSIZE (512 * 1024)
> +#endif
> +unsigned long x86_stack_size = USER_MODE_STACKSIZE;
> 
>  void gemu_log(const char *fmt, ...)
>  {
> 

> diff --git a/configure b/configure
> index 43d87c5..dad9175 100755
> --- a/configure
> +++ b/configure
> @@ -220,6 +220,7 @@ uname_release=""
>  io_thread="no"
>  mixemu="no"
>  kerneldir=""
> +user_mode_stacksize=""
>  aix="no"
>  blobs="yes"
>  pkgversion=""
> @@ -557,6 +558,8 @@ for opt do
>    ;;
>    --kerneldir=*) kerneldir="$optarg"
>    ;;
> +  --user-mode-stacksize=*) user_mode_stacksize="$optarg"
> +  ;;
>    --with-pkgversion=*) pkgversion=" ($optarg)"
>    ;;
>    --disable-docs) docs="no"
> @@ -713,6 +716,7 @@ echo "  --enable-linux-aio       enable Linux AIO support"
>  echo "  --enable-io-thread       enable IO thread"
>  echo "  --disable-blobs          disable installing provided firmware blobs"
>  echo "  --kerneldir=PATH         look for kernel includes in PATH"
> +echo "  --user-mode-stacksize=   set default stack size in bytes (as -s, only in usermode)"
>  echo ""
>  echo "NOTE: The object files are built at the place where configure is launched"
>  exit 1
> @@ -2481,6 +2485,16 @@ if test "$target_user_only" = "yes" -a "$static" = "no" -a \
>    ldflags="-pie $ldflags"
>  fi
>  
> +# change default stacksize in usermode
> +#
> +if test "$target_user_only" = "yes" -a "$user_mode_stacksize" != "" ; then
> + cflags="-DUSER_MODE_STACKSIZE=$user_mode_stacksize $cflags"
> + echo "user_mode_stack   $user_mode_stacksize"
> +else
> + echo "user_mode_stack   default"
> +fi
> +
> +
>  if test "$target_softmmu" = "yes" -a \( \
>          "$TARGET_ARCH" = "microblaze" -o \
>          "$TARGET_ARCH" = "cris" \) ; then
> diff --git a/linux-user/main.c b/linux-user/main.c
> index 81a1ada..9ac2421 100644
> --- a/linux-user/main.c
> +++ b/linux-user/main.c
> @@ -51,7 +51,10 @@ const char *qemu_uname_release = CONFIG_UNAME_RELEASE;
>  /* XXX: on x86 MAP_GROWSDOWN only works if ESP <= address + 32, so
>     we allocate a bigger stack. Need a better solution, for example
>     by remapping the process stack directly at the right place */
> -unsigned long x86_stack_size = 512 * 1024;
> +#ifndef USER_MODE_STACKSIZE
> +#define USER_MODE_STACKSIZE (512 * 1024)
> +#endif
> +unsigned long x86_stack_size = USER_MODE_STACKSIZE;
>  
>  void gemu_log(const char *fmt, ...)
>  {

Re: [Qemu-devel] [Patch] Make usermode stacksize (-s) configurable at compile-time

[Jan-Simon Möller]

> > The attached patch makes this configurable at compile-time.
> 
> fwiw in scratchbox we have a wrapper in between binfmt and qemu (called misc_runner) that sets any
> necessary command line parameters for qemu. You might want to consider a similar approach so we
> don't need to add a configure option (and thus hardcode the setting) to qemu for every command
> line option that might need changing.
I got your point. But OTOH I just make the default configurable.

IMHO this leads to:
* add an (optional) config-file for the usermode to overwrite defaults
 or
* extend and generalize the wrapper and use it in binfmt instead of qemu directly

Let me sleep over it a bit - for now: other opinions ?

Best,
Jan-Simon