[Qemu-devel] [PATCH] configure: Drop AIX host support

[Qemu-devel] [PATCH] configure: Drop AIX host support

[Peter Maydell]

Nobody has mentioned AIX host support on the mailing list for years,
and we have no test systems for it so it is most likely broken.
We've advertised in configure for two releases now that we plan
to drop support for this host OS, and have had no complaints.
Drop the AIX host support code.

We can also drop the now-unused AIX version of sys_cache_info().

Note that the _CALL_AIX define used in the PPC tcg backend is
also used for Linux PPC64, and so that code should not be removed.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 configure        | 13 +------------
 util/cacheinfo.c | 11 +----------
 2 files changed, 2 insertions(+), 22 deletions(-)

diff --git a/configure b/configure
index fb7e34a..931b3ba 100755
--- a/configure
+++ b/configure
@@ -350,7 +350,6 @@ cocoa="no"
 softmmu="yes"
 linux_user="no"
 bsd_user="no"
-aix="no"
 blobs="yes"
 pkgversion=""
 pie=""
@@ -554,8 +553,6 @@ elif check_define __NetBSD__; then
   targetos='NetBSD'
 elif check_define __APPLE__; then
   targetos='Darwin'
-elif check_define _AIX; then
-  targetos='AIX'
 else
   # This is a fatal error, but don't report it yet, because we
   # might be going to just print the --help text, or it might
@@ -767,10 +764,6 @@ SunOS)
   LIBS="$solarisnetlibs $LIBS"
   libs_qga="$solarisnetlibs $libs_qga"
 ;;
-AIX)
-  aix="yes"
-  make="${MAKE-gmake}"
-;;
 Haiku)
   haiku="yes"
   QEMU_CFLAGS="-DB_USE_POSITIVE_POSIX_ERRORS $QEMU_CFLAGS"
@@ -4214,7 +4207,7 @@ elif compile_prog "" "$pthread_lib -lrt" ; then
 fi
 
 if test "$darwin" != "yes" -a "$mingw32" != "yes" -a "$solaris" != yes -a \
-        "$aix" != "yes" -a "$haiku" != "yes" ; then
+        "$haiku" != "yes" ; then
     libs_softmmu="-lutil $libs_softmmu"
 fi
 
@@ -5523,10 +5516,6 @@ if test "$darwin" = "yes" ; then
   echo "CONFIG_DARWIN=y" >> $config_host_mak
 fi
 
-if test "$aix" = "yes" ; then
-  echo "CONFIG_AIX=y" >> $config_host_mak
-fi
-
 if test "$solaris" = "yes" ; then
   echo "CONFIG_SOLARIS=y" >> $config_host_mak
 fi
diff --git a/util/cacheinfo.c b/util/cacheinfo.c
index 593940f..db5172d 100644
--- a/util/cacheinfo.c
+++ b/util/cacheinfo.c
@@ -15,16 +15,7 @@ int qemu_dcache_linesize = 0;
  * Operating system specific detection mechanisms.
  */
 
-#if defined(_AIX)
-# include <sys/systemcfg.h>
-
-static void sys_cache_info(int *isize, int *dsize)
-{
-    *isize = _system_configuration.icache_line;
-    *dsize = _system_configuration.dcache_line;
-}
-
-#elif defined(_WIN32)
+#if defined(_WIN32)
 
 static void sys_cache_info(int *isize, int *dsize)
 {
-- 
2.7.4

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Philippe Mathieu-Daudé]

On 09/04/2017 02:19 PM, Peter Maydell wrote:
> Nobody has mentioned AIX host support on the mailing list for years,
> and we have no test systems for it so it is most likely broken.
> We've advertised in configure for two releases now that we plan
> to drop support for this host OS, and have had no complaints.
> Drop the AIX host support code.
> 
> We can also drop the now-unused AIX version of sys_cache_info().
> 
> Note that the _CALL_AIX define used in the PPC tcg backend is
> also used for Linux PPC64, and so that code should not be removed.
> 
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

:~(
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>

> ---
>   configure        | 13 +------------
>   util/cacheinfo.c | 11 +----------
>   2 files changed, 2 insertions(+), 22 deletions(-)
> 
> diff --git a/configure b/configure
> index fb7e34a..931b3ba 100755
> --- a/configure
> +++ b/configure
> @@ -350,7 +350,6 @@ cocoa="no"
>   softmmu="yes"
>   linux_user="no"
>   bsd_user="no"
> -aix="no"
>   blobs="yes"
>   pkgversion=""
>   pie=""
> @@ -554,8 +553,6 @@ elif check_define __NetBSD__; then
>     targetos='NetBSD'
>   elif check_define __APPLE__; then
>     targetos='Darwin'
> -elif check_define _AIX; then
> -  targetos='AIX'
>   else
>     # This is a fatal error, but don't report it yet, because we
>     # might be going to just print the --help text, or it might
> @@ -767,10 +764,6 @@ SunOS)
>     LIBS="$solarisnetlibs $LIBS"
>     libs_qga="$solarisnetlibs $libs_qga"
>   ;;
> -AIX)
> -  aix="yes"
> -  make="${MAKE-gmake}"
> -;;
>   Haiku)
>     haiku="yes"
>     QEMU_CFLAGS="-DB_USE_POSITIVE_POSIX_ERRORS $QEMU_CFLAGS"
> @@ -4214,7 +4207,7 @@ elif compile_prog "" "$pthread_lib -lrt" ; then
>   fi
>   
>   if test "$darwin" != "yes" -a "$mingw32" != "yes" -a "$solaris" != yes -a \
> -        "$aix" != "yes" -a "$haiku" != "yes" ; then
> +        "$haiku" != "yes" ; then
>       libs_softmmu="-lutil $libs_softmmu"
>   fi
>   
> @@ -5523,10 +5516,6 @@ if test "$darwin" = "yes" ; then
>     echo "CONFIG_DARWIN=y" >> $config_host_mak
>   fi
>   
> -if test "$aix" = "yes" ; then
> -  echo "CONFIG_AIX=y" >> $config_host_mak
> -fi
> -
>   if test "$solaris" = "yes" ; then
>     echo "CONFIG_SOLARIS=y" >> $config_host_mak
>   fi
> diff --git a/util/cacheinfo.c b/util/cacheinfo.c
> index 593940f..db5172d 100644
> --- a/util/cacheinfo.c
> +++ b/util/cacheinfo.c
> @@ -15,16 +15,7 @@ int qemu_dcache_linesize = 0;
>    * Operating system specific detection mechanisms.
>    */
>   
> -#if defined(_AIX)
> -# include <sys/systemcfg.h>
> -
> -static void sys_cache_info(int *isize, int *dsize)
> -{
> -    *isize = _system_configuration.icache_line;
> -    *dsize = _system_configuration.dcache_line;
> -}
> -
> -#elif defined(_WIN32)
> +#if defined(_WIN32)
>   
>   static void sys_cache_info(int *isize, int *dsize)
>   {
> 

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Thomas Huth]

On 04.09.2017 19:19, Peter Maydell wrote:
> Nobody has mentioned AIX host support on the mailing list for years,
> and we have no test systems for it so it is most likely broken.
> We've advertised in configure for two releases now that we plan
> to drop support for this host OS, and have had no complaints.

We just added this some weeks ago to our qemu-doc.texi:

"Prior to the 2.10.0 release there was no official policy on how long
features would be deprecated prior to their removal, nor any documented
list of which features were deprecated. Thus any features deprecated
prior to 2.10.0 will be treated as if they were first deprecated in the
2.10.0 release."

Should we maybe wait two more releases (for version 3.0 ;-)) 'till we
remove it?

Anyway, the patch looks fine to me, so:

Reviewed-by: Thomas Huth <thuth@redhat.com>

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Peter Maydell]

On 4 September 2017 at 19:09, Thomas Huth <thuth@redhat.com> wrote:
> On 04.09.2017 19:19, Peter Maydell wrote:
>> Nobody has mentioned AIX host support on the mailing list for years,
>> and we have no test systems for it so it is most likely broken.
>> We've advertised in configure for two releases now that we plan
>> to drop support for this host OS, and have had no complaints.
>
> We just added this some weeks ago to our qemu-doc.texi:
>
> "Prior to the 2.10.0 release there was no official policy on how long
> features would be deprecated prior to their removal, nor any documented
> list of which features were deprecated. Thus any features deprecated
> prior to 2.10.0 will be treated as if they were first deprecated in the
> 2.10.0 release."

We pretty clearly say in the configure message
    echo "The QEMU project intends to remove support for this host OS in"
    echo "a future release if nobody volunteers to maintain it and to"
    echo "provide a build host for our continuous integration setup."

and my intention when I wrote that was absolutely to dump old
OSes in a couple of releases... The amount of code removed here
isn't great so I guess waiting a little longer doesn't hurt,
but I feel like somebody's moved the goalposts in adding that
language and not listing the deprecated host OSes and CPUs
in the list of deprecated features in 2.10 :-(

thanks
-- PMM

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Laurent Vivier]

Le 04/09/2017 à 19:19, Peter Maydell a écrit :
> Nobody has mentioned AIX host support on the mailing list for years,
> and we have no test systems for it so it is most likely broken.
> We've advertised in configure for two releases now that we plan
> to drop support for this host OS, and have had no complaints.
> Drop the AIX host support code.
> 
> We can also drop the now-unused AIX version of sys_cache_info().
> 
> Note that the _CALL_AIX define used in the PPC tcg backend is
> also used for Linux PPC64, and so that code should not be removed.
> 
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Reviewed-by: Laurent Vivier <laurent@vivier.eu>

> ---
>  configure        | 13 +------------
>  util/cacheinfo.c | 11 +----------
>  2 files changed, 2 insertions(+), 22 deletions(-)
> 
> diff --git a/configure b/configure
> index fb7e34a..931b3ba 100755
> --- a/configure
> +++ b/configure
> @@ -350,7 +350,6 @@ cocoa="no"
>  softmmu="yes"
>  linux_user="no"
>  bsd_user="no"
> -aix="no"
>  blobs="yes"
>  pkgversion=""
>  pie=""
> @@ -554,8 +553,6 @@ elif check_define __NetBSD__; then
>    targetos='NetBSD'
>  elif check_define __APPLE__; then
>    targetos='Darwin'
> -elif check_define _AIX; then
> -  targetos='AIX'
>  else
>    # This is a fatal error, but don't report it yet, because we
>    # might be going to just print the --help text, or it might
> @@ -767,10 +764,6 @@ SunOS)
>    LIBS="$solarisnetlibs $LIBS"
>    libs_qga="$solarisnetlibs $libs_qga"
>  ;;
> -AIX)
> -  aix="yes"
> -  make="${MAKE-gmake}"
> -;;
>  Haiku)
>    haiku="yes"
>    QEMU_CFLAGS="-DB_USE_POSITIVE_POSIX_ERRORS $QEMU_CFLAGS"
> @@ -4214,7 +4207,7 @@ elif compile_prog "" "$pthread_lib -lrt" ; then
>  fi
>  
>  if test "$darwin" != "yes" -a "$mingw32" != "yes" -a "$solaris" != yes -a \
> -        "$aix" != "yes" -a "$haiku" != "yes" ; then
> +        "$haiku" != "yes" ; then
>      libs_softmmu="-lutil $libs_softmmu"
>  fi
>  
> @@ -5523,10 +5516,6 @@ if test "$darwin" = "yes" ; then
>    echo "CONFIG_DARWIN=y" >> $config_host_mak
>  fi
>  
> -if test "$aix" = "yes" ; then
> -  echo "CONFIG_AIX=y" >> $config_host_mak
> -fi
> -
>  if test "$solaris" = "yes" ; then
>    echo "CONFIG_SOLARIS=y" >> $config_host_mak
>  fi
> diff --git a/util/cacheinfo.c b/util/cacheinfo.c
> index 593940f..db5172d 100644
> --- a/util/cacheinfo.c
> +++ b/util/cacheinfo.c
> @@ -15,16 +15,7 @@ int qemu_dcache_linesize = 0;
>   * Operating system specific detection mechanisms.
>   */
>  
> -#if defined(_AIX)
> -# include <sys/systemcfg.h>
> -
> -static void sys_cache_info(int *isize, int *dsize)
> -{
> -    *isize = _system_configuration.icache_line;
> -    *dsize = _system_configuration.dcache_line;
> -}
> -
> -#elif defined(_WIN32)
> +#if defined(_WIN32)
>  
>  static void sys_cache_info(int *isize, int *dsize)
>  {
> 

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Kamil Rytarowski]

[-- Attachment #1: Type: text/plain, Size: 3012 bytes --]

On 04.09.2017 19:19, Peter Maydell wrote:
> Nobody has mentioned AIX host support on the mailing list for years,
> and we have no test systems for it so it is most likely broken.
> We've advertised in configure for two releases now that we plan
> to drop support for this host OS, and have had no complaints.
> Drop the AIX host support code.
> 

There is AIX support in pkgsrc, but the activity is minimal. It
certainly does not pass perl and the amount of buildable packages is tiny.

> We can also drop the now-unused AIX version of sys_cache_info().
> 
> Note that the _CALL_AIX define used in the PPC tcg backend is
> also used for Linux PPC64, and so that code should not be removed.
> 
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---
>  configure        | 13 +------------
>  util/cacheinfo.c | 11 +----------
>  2 files changed, 2 insertions(+), 22 deletions(-)
> 
> diff --git a/configure b/configure
> index fb7e34a..931b3ba 100755
> --- a/configure
> +++ b/configure
> @@ -350,7 +350,6 @@ cocoa="no"
>  softmmu="yes"
>  linux_user="no"
>  bsd_user="no"
> -aix="no"
>  blobs="yes"
>  pkgversion=""
>  pie=""
> @@ -554,8 +553,6 @@ elif check_define __NetBSD__; then
>    targetos='NetBSD'
>  elif check_define __APPLE__; then
>    targetos='Darwin'
> -elif check_define _AIX; then
> -  targetos='AIX'
>  else
>    # This is a fatal error, but don't report it yet, because we
>    # might be going to just print the --help text, or it might
> @@ -767,10 +764,6 @@ SunOS)
>    LIBS="$solarisnetlibs $LIBS"
>    libs_qga="$solarisnetlibs $libs_qga"
>  ;;
> -AIX)
> -  aix="yes"
> -  make="${MAKE-gmake}"
> -;;
>  Haiku)
>    haiku="yes"
>    QEMU_CFLAGS="-DB_USE_POSITIVE_POSIX_ERRORS $QEMU_CFLAGS"
> @@ -4214,7 +4207,7 @@ elif compile_prog "" "$pthread_lib -lrt" ; then
>  fi
>  
>  if test "$darwin" != "yes" -a "$mingw32" != "yes" -a "$solaris" != yes -a \
> -        "$aix" != "yes" -a "$haiku" != "yes" ; then
> +        "$haiku" != "yes" ; then
>      libs_softmmu="-lutil $libs_softmmu"
>  fi
>  
> @@ -5523,10 +5516,6 @@ if test "$darwin" = "yes" ; then
>    echo "CONFIG_DARWIN=y" >> $config_host_mak
>  fi
>  
> -if test "$aix" = "yes" ; then
> -  echo "CONFIG_AIX=y" >> $config_host_mak
> -fi
> -
>  if test "$solaris" = "yes" ; then
>    echo "CONFIG_SOLARIS=y" >> $config_host_mak
>  fi
> diff --git a/util/cacheinfo.c b/util/cacheinfo.c
> index 593940f..db5172d 100644
> --- a/util/cacheinfo.c
> +++ b/util/cacheinfo.c
> @@ -15,16 +15,7 @@ int qemu_dcache_linesize = 0;
>   * Operating system specific detection mechanisms.
>   */
>  
> -#if defined(_AIX)
> -# include <sys/systemcfg.h>
> -
> -static void sys_cache_info(int *isize, int *dsize)
> -{
> -    *isize = _system_configuration.icache_line;
> -    *dsize = _system_configuration.dcache_line;
> -}
> -
> -#elif defined(_WIN32)
> +#if defined(_WIN32)
>  
>  static void sys_cache_info(int *isize, int *dsize)
>  {
> 



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

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Markus Armbruster]

Peter Maydell <peter.maydell@linaro.org> writes:

> On 4 September 2017 at 19:09, Thomas Huth <thuth@redhat.com> wrote:
>> On 04.09.2017 19:19, Peter Maydell wrote:
>>> Nobody has mentioned AIX host support on the mailing list for years,
>>> and we have no test systems for it so it is most likely broken.
>>> We've advertised in configure for two releases now that we plan
>>> to drop support for this host OS, and have had no complaints.
>>
>> We just added this some weeks ago to our qemu-doc.texi:
>>
>> "Prior to the 2.10.0 release there was no official policy on how long
>> features would be deprecated prior to their removal, nor any documented
>> list of which features were deprecated. Thus any features deprecated
>> prior to 2.10.0 will be treated as if they were first deprecated in the
>> 2.10.0 release."

Daniel's commit eb22aeca65f.

> We pretty clearly say in the configure message
>     echo "The QEMU project intends to remove support for this host OS in"
>     echo "a future release if nobody volunteers to maintain it and to"
>     echo "provide a build host for our continuous integration setup."
>
> and my intention when I wrote that was absolutely to dump old
> OSes in a couple of releases... The amount of code removed here
> isn't great so I guess waiting a little longer doesn't hurt,
> but I feel like somebody's moved the goalposts in adding that
> language and not listing the deprecated host OSes and CPUs
> in the list of deprecated features in 2.10 :-(

AIX host has been *clearly* deprecated since 2.9.0 (commit 898be3e).
Failure to include that in a summary added later and elsewhere does not
undo that.  It's simply a bug in the summary.  We can discuss whether a
grace period of two releases is enough.  For what it's worth, I think it
is in this case.

Let's fix the bugs in the summary for 2.10.1.

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Thomas Huth]

On 05.09.2017 09:16, Markus Armbruster wrote:
> Peter Maydell <peter.maydell@linaro.org> writes:
> 
>> On 4 September 2017 at 19:09, Thomas Huth <thuth@redhat.com> wrote:
>>> On 04.09.2017 19:19, Peter Maydell wrote:
>>>> Nobody has mentioned AIX host support on the mailing list for years,
>>>> and we have no test systems for it so it is most likely broken.
>>>> We've advertised in configure for two releases now that we plan
>>>> to drop support for this host OS, and have had no complaints.
>>>
>>> We just added this some weeks ago to our qemu-doc.texi:
>>>
>>> "Prior to the 2.10.0 release there was no official policy on how long
>>> features would be deprecated prior to their removal, nor any documented
>>> list of which features were deprecated. Thus any features deprecated
>>> prior to 2.10.0 will be treated as if they were first deprecated in the
>>> 2.10.0 release."
> 
> Daniel's commit eb22aeca65f.
> 
>> We pretty clearly say in the configure message
>>     echo "The QEMU project intends to remove support for this host OS in"
>>     echo "a future release if nobody volunteers to maintain it and to"
>>     echo "provide a build host for our continuous integration setup."
>>
>> and my intention when I wrote that was absolutely to dump old
>> OSes in a couple of releases... The amount of code removed here
>> isn't great so I guess waiting a little longer doesn't hurt,
>> but I feel like somebody's moved the goalposts in adding that
>> language and not listing the deprecated host OSes and CPUs
>> in the list of deprecated features in 2.10 :-(
> 
> AIX host has been *clearly* deprecated since 2.9.0 (commit 898be3e).
> Failure to include that in a summary added later and elsewhere does not
> undo that.
Yes, sure, I fully agree that it is marked as deprecated since 2.9. I
certainly did *not* want to say that grace period for the AIX removal
has now to start with QEMU 2.11 - sorry if my mail was not clear in that
way. I just wanted to discuss whether we should apply the "any features
deprecated prior to 2.10.0 will be treated as if they were first
deprecated in the 2.10.0 release" idea here, too.

OTOH, AIX support is really very, very like broken since years and
nobody complained (since there haven't been any AIX-related bug fix
commits in years), so I guess nobody will complain if we remove it right
now already.

 Thomas

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Greg Kurz]

[-- Attachment #1: Type: text/plain, Size: 3000 bytes --]

On Mon,  4 Sep 2017 18:19:00 +0100
Peter Maydell <peter.maydell@linaro.org> wrote:

> Nobody has mentioned AIX host support on the mailing list for years,
> and we have no test systems for it so it is most likely broken.
> We've advertised in configure for two releases now that we plan
> to drop support for this host OS, and have had no complaints.
> Drop the AIX host support code.
> 
> We can also drop the now-unused AIX version of sys_cache_info().
> 
> Note that the _CALL_AIX define used in the PPC tcg backend is
> also used for Linux PPC64, and so that code should not be removed.
> 

Yeah this is used by the old big-endian ABI.

> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---

Reviewed-by: Greg Kurz <groug@kaod.org>

>  configure        | 13 +------------
>  util/cacheinfo.c | 11 +----------
>  2 files changed, 2 insertions(+), 22 deletions(-)
> 
> diff --git a/configure b/configure
> index fb7e34a..931b3ba 100755
> --- a/configure
> +++ b/configure
> @@ -350,7 +350,6 @@ cocoa="no"
>  softmmu="yes"
>  linux_user="no"
>  bsd_user="no"
> -aix="no"
>  blobs="yes"
>  pkgversion=""
>  pie=""
> @@ -554,8 +553,6 @@ elif check_define __NetBSD__; then
>    targetos='NetBSD'
>  elif check_define __APPLE__; then
>    targetos='Darwin'
> -elif check_define _AIX; then
> -  targetos='AIX'
>  else
>    # This is a fatal error, but don't report it yet, because we
>    # might be going to just print the --help text, or it might
> @@ -767,10 +764,6 @@ SunOS)
>    LIBS="$solarisnetlibs $LIBS"
>    libs_qga="$solarisnetlibs $libs_qga"
>  ;;
> -AIX)
> -  aix="yes"
> -  make="${MAKE-gmake}"
> -;;
>  Haiku)
>    haiku="yes"
>    QEMU_CFLAGS="-DB_USE_POSITIVE_POSIX_ERRORS $QEMU_CFLAGS"
> @@ -4214,7 +4207,7 @@ elif compile_prog "" "$pthread_lib -lrt" ; then
>  fi
>  
>  if test "$darwin" != "yes" -a "$mingw32" != "yes" -a "$solaris" != yes -a \
> -        "$aix" != "yes" -a "$haiku" != "yes" ; then
> +        "$haiku" != "yes" ; then
>      libs_softmmu="-lutil $libs_softmmu"
>  fi
>  
> @@ -5523,10 +5516,6 @@ if test "$darwin" = "yes" ; then
>    echo "CONFIG_DARWIN=y" >> $config_host_mak
>  fi
>  
> -if test "$aix" = "yes" ; then
> -  echo "CONFIG_AIX=y" >> $config_host_mak
> -fi
> -
>  if test "$solaris" = "yes" ; then
>    echo "CONFIG_SOLARIS=y" >> $config_host_mak
>  fi
> diff --git a/util/cacheinfo.c b/util/cacheinfo.c
> index 593940f..db5172d 100644
> --- a/util/cacheinfo.c
> +++ b/util/cacheinfo.c
> @@ -15,16 +15,7 @@ int qemu_dcache_linesize = 0;
>   * Operating system specific detection mechanisms.
>   */
>  
> -#if defined(_AIX)
> -# include <sys/systemcfg.h>
> -
> -static void sys_cache_info(int *isize, int *dsize)
> -{
> -    *isize = _system_configuration.icache_line;
> -    *dsize = _system_configuration.dcache_line;
> -}
> -
> -#elif defined(_WIN32)
> +#if defined(_WIN32)
>  
>  static void sys_cache_info(int *isize, int *dsize)
>  {


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Daniel P. Berrange]

On Mon, Sep 04, 2017 at 08:09:17PM +0200, Thomas Huth wrote:
> On 04.09.2017 19:19, Peter Maydell wrote:
> > Nobody has mentioned AIX host support on the mailing list for years,
> > and we have no test systems for it so it is most likely broken.
> > We've advertised in configure for two releases now that we plan
> > to drop support for this host OS, and have had no complaints.
> 
> We just added this some weeks ago to our qemu-doc.texi:
> 
> "Prior to the 2.10.0 release there was no official policy on how long
> features would be deprecated prior to their removal, nor any documented
> list of which features were deprecated. Thus any features deprecated
> prior to 2.10.0 will be treated as if they were first deprecated in the
> 2.10.0 release."
> 
> Should we maybe wait two more releases (for version 3.0 ;-)) 'till we
> remove it?

The deprecation policy is for features which currently work and which we
expect to have actual users. Neither is the case with AIX support, so I
don't see any reason to keep it longer.

Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Thomas Huth]

On 05.09.2017 10:04, Daniel P. Berrange wrote:
> On Mon, Sep 04, 2017 at 08:09:17PM +0200, Thomas Huth wrote:
>> On 04.09.2017 19:19, Peter Maydell wrote:
>>> Nobody has mentioned AIX host support on the mailing list for years,
>>> and we have no test systems for it so it is most likely broken.
>>> We've advertised in configure for two releases now that we plan
>>> to drop support for this host OS, and have had no complaints.
>>
>> We just added this some weeks ago to our qemu-doc.texi:
>>
>> "Prior to the 2.10.0 release there was no official policy on how long
>> features would be deprecated prior to their removal, nor any documented
>> list of which features were deprecated. Thus any features deprecated
>> prior to 2.10.0 will be treated as if they were first deprecated in the
>> 2.10.0 release."
>>
>> Should we maybe wait two more releases (for version 3.0 ;-)) 'till we
>> remove it?
> 
> The deprecation policy is for features which currently work and which we
> expect to have actual users. Neither is the case with AIX support, so I
> don't see any reason to keep it longer.

OK, then let's remove the dead AIX code in 2.11 already!

 Thomas

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Stefan Hajnoczi]

On Mon, Sep 04, 2017 at 06:19:00PM +0100, Peter Maydell wrote:
> Nobody has mentioned AIX host support on the mailing list for years,
> and we have no test systems for it so it is most likely broken.
> We've advertised in configure for two releases now that we plan
> to drop support for this host OS, and have had no complaints.
> Drop the AIX host support code.
> 
> We can also drop the now-unused AIX version of sys_cache_info().
> 
> Note that the _CALL_AIX define used in the PPC tcg backend is
> also used for Linux PPC64, and so that code should not be removed.
> 
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---
>  configure        | 13 +------------
>  util/cacheinfo.c | 11 +----------
>  2 files changed, 2 insertions(+), 22 deletions(-)

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Peter Maydell]

On 5 September 2017 at 08:36, Thomas Huth <thuth@redhat.com> wrote:
> OTOH, AIX support is really very, very like broken since years

It turns out there's an AIX box in the gcc compile farm, so just
out of curiosity I had a look.

(1) we don't recognize the cpu so you have to pass --cpu=ppc64
(2) we unconditionally pass -m64 for ppc64, which the AIX gcc
doesn't recognize, so you have to hack configure to avoid that
(3) after all that, AIX gcc doesn't have a working __thread for TLS

So yes, definitively broken and likely has been for years.

thanks
-- PMM

Re: [Qemu-devel] [PATCH] configure: Drop AIX host support

[Peter Maydell]

On 4 September 2017 at 18:19, Peter Maydell <peter.maydell@linaro.org> wrote:
> Nobody has mentioned AIX host support on the mailing list for years,
> and we have no test systems for it so it is most likely broken.
> We've advertised in configure for two releases now that we plan
> to drop support for this host OS, and have had no complaints.
> Drop the AIX host support code.
>
> We can also drop the now-unused AIX version of sys_cache_info().
>
> Note that the _CALL_AIX define used in the PPC tcg backend is
> also used for Linux PPC64, and so that code should not be removed.
>
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---
>  configure        | 13 +------------
>  util/cacheinfo.c | 11 +----------
>  2 files changed, 2 insertions(+), 22 deletions(-)

Applied to master, thanks.

-- PMM