[Xenomai-core] Patch selection by prepare-kernel.sh (again)

[Xenomai-core] Patch selection by prepare-kernel.sh (again)

[Jan Kiszka]

Philippe,

what was the idea behind commit #3256 [1]? It breaks the original
behavior of selection adeos-ipipe-2.6.23-... in the presence of 2.6.23.x
- and leaves a bad first impression for beginners (I received such a
notice here...). Was this for the special case
adeos-ipipe-2.6.20.21-i386-1.12-02.patch? But I think the common case
should take precedence over this corner case here. Unless there are more
issues with the original version, I would vote for restoring it.

Alternatively, we could add a safety belt by trying to apply the patch
with --dry-run first and only continue if this works out without
failures. Can hack this up if it's acceptable.

Jan

[1]http://svn.gna.org/viewcvs/xenomai/trunk/scripts/prepare-kernel.sh?rev=3257&view=diff&r1=3257&r2=3256&p1=trunk/scripts/prepare-kernel.sh&p2=/trunk/scripts/prepare-kernel.sh

-- 
Siemens AG, Corporate Technology, CT SE 2
Corporate Competence Center Embedded Linux

Re: [Xenomai-core] Patch selection by prepare-kernel.sh (again)

[Jan Kiszka]

[-- Attachment #1: Type: text/plain, Size: 993 bytes --]

Jan Kiszka wrote:
> Philippe,
> 
> what was the idea behind commit #3256 [1]? It breaks the original
> behavior of selection adeos-ipipe-2.6.23-... in the presence of 2.6.23.x
> - and leaves a bad first impression for beginners (I received such a
> notice here...). Was this for the special case
> adeos-ipipe-2.6.20.21-i386-1.12-02.patch? But I think the common case
> should take precedence over this corner case here. Unless there are more
> issues with the original version, I would vote for restoring it.
> 
> Alternatively, we could add a safety belt by trying to apply the patch
> with --dry-run first and only continue if this works out without
> failures. Can hack this up if it's acceptable.

Oh, this check already exists!

So let's try this version, which additionally cleans up the related
documentation. Catches both 2.6.23* and 2.6.24* for me here, but I may
miss some cases (2.4?).

Jan

-- 
Siemens AG, Corporate Technology, CT SE 2
Corporate Competence Center Embedded Linux

[-- Attachment #2: improve-patch-detection.patch --]
[-- Type: text/x-patch, Size: 2692 bytes --]

---
 README.INSTALL            |    5 ++++-
 scripts/prepare-kernel.sh |    8 ++++----
 2 files changed, 8 insertions(+), 5 deletions(-)

Index: xenomai/README.INSTALL
===================================================================
--- xenomai.orig/README.INSTALL
+++ xenomai/README.INSTALL
@@ -66,7 +66,6 @@ support:
 $ cd xenomai-2.4
 $ scripts/prepare-kernel.sh
   --linux=/usr/src/linux-2.6.23-ipipe
-  --adeos=/tmp/adeos-ipipe-2.6.23-i386-1.10-12.patch
   --arch=i386
 
 Note: The script will infer the location of the Xenomai kernel code
@@ -75,6 +74,10 @@ if /usr/src/xenomai-2.4/script/prepare-k
 Xenomai's kernel support available from /usr/src/xenomai-2.4/ksrc will
 be bound to the target kernel.
 
+Second note: If the script fails to figure out a matching Adeos patch,
+it will ask for the correct patch file. Suitable patches are available
+under ksrc/arch/<target-arch>/patches.
+
 1.2 Configuring and building the target kernel
 ----------------------------------------------
 
Index: xenomai/scripts/prepare-kernel.sh
===================================================================
--- xenomai.orig/scripts/prepare-kernel.sh
+++ xenomai/scripts/prepare-kernel.sh
@@ -353,10 +353,10 @@ eval linux_`grep '^PATCHLEVEL =' $linux_
 eval linux_`grep '^SUBLEVEL =' $linux_tree/Makefile | sed -e 's, ,,g'`
 eval linux_`grep '^VERSION =' $linux_tree/Makefile | sed -e 's, ,,g'`
 
-linux_version="$linux_VERSION.$linux_PATCHLEVEL.$linux_SUBLEVEL$linux_EXTRAVERSION"
+linux_version="$linux_VERSION.$linux_PATCHLEVEL.$linux_SUBLEVEL"
 
 if test x$verbose = x1; then
-echo "Preparing kernel $linux_version in $linux_tree..."
+echo "Preparing kernel $linux_version$linux_EXTRAVERSION in $linux_tree..."
 fi
 
 if test -r $linux_tree/include/linux/ipipe.h; then
@@ -369,7 +369,7 @@ elif test -r $linux_tree/include/linux/a
    exit 2
 else
    if test x$adeos_patch = x; then
-      default_adeos_patch=`( ls $xenomai_root/ksrc/arch/$xenomai_arch/patches/adeos-ipipe-$linux_version-$linux_arch-*|sort -r ) 2>/dev/null | head -n1`
+      default_adeos_patch="`( ls $xenomai_root/ksrc/arch/$xenomai_arch/patches/adeos-ipipe-$linux_version*-{$linux_arch,$xenomai_arch}-*|sort -r ) 2>/dev/null | head -n1`"
    fi
    if test x$default_adeos_patch = x; then
       default_adeos_patch=/dev/null
@@ -393,7 +393,7 @@ else
    curdir=$PWD
    cd $linux_tree && patch --dry-run -p1 -f < $adeos_patch || { 
         cd $curdir;
-        echo "$me: Unable to patch kernel $linux_version with `basename $adeos_patch`." >&2
+        echo "$me: Unable to patch kernel $linux_version$linux_EXTRAVERSION with `basename $adeos_patch`." >&2
         exit 2;
    }
    patch -p1 -f -s < $adeos_patch

Re: [Xenomai-core] Patch selection by prepare-kernel.sh (again)

[Jan Kiszka]

[-- Attachment #1: Type: text/plain, Size: 1110 bytes --]

Jan Kiszka wrote:
> Jan Kiszka wrote:
>> Philippe,
>>
>> what was the idea behind commit #3256 [1]? It breaks the original
>> behavior of selection adeos-ipipe-2.6.23-... in the presence of 2.6.23.x
>> - and leaves a bad first impression for beginners (I received such a
>> notice here...). Was this for the special case
>> adeos-ipipe-2.6.20.21-i386-1.12-02.patch? But I think the common case
>> should take precedence over this corner case here. Unless there are more
>> issues with the original version, I would vote for restoring it.
>>
>> Alternatively, we could add a safety belt by trying to apply the patch
>> with --dry-run first and only continue if this works out without
>> failures. Can hack this up if it's acceptable.
> 
> Oh, this check already exists!
> 
> So let's try this version, which additionally cleans up the related
> documentation. Catches both 2.6.23* and 2.6.24* for me here, but I may
> miss some cases (2.4?).

The documentation hunk was nonsense, here is a hopefully better version.

Jan

-- 
Siemens AG, Corporate Technology, CT SE 2
Corporate Competence Center Embedded Linux

[-- Attachment #2: improve-patch-detection.patch --]
[-- Type: text/x-patch, Size: 3120 bytes --]

---
 README.INSTALL            |   11 +++++------
 scripts/prepare-kernel.sh |    8 ++++----
 2 files changed, 9 insertions(+), 10 deletions(-)

Index: xenomai/README.INSTALL
===================================================================
--- xenomai.orig/README.INSTALL
+++ xenomai/README.INSTALL
@@ -51,8 +51,10 @@ $ scripts/prepare-kernel.sh --linux=<lin
     valid either way.
 
     --adeos specifies the path of the Adeos patch to apply against the
-    kernel tree. This parameter can be omitted if Adeos has already
-    been patched in. In any case, the script will not try to apply it
+    kernel tree. Suitable patches are available with Xenomai under
+    ksrc/arch/<target-arch>/patches. This parameter can be omitted if
+    Adeos has already been patched in or the script shall suggest an
+    appropriate one. In any case, the script will not try to apply it
     again whenever a former patch is detected.
 
     --arch tells the script about the target architecture. If
@@ -64,10 +66,7 @@ located at /usr/src/linux-2.6.23-ipipe i
 support:
 
 $ cd xenomai-2.4
-$ scripts/prepare-kernel.sh
-  --linux=/usr/src/linux-2.6.23-ipipe
-  --adeos=/tmp/adeos-ipipe-2.6.23-i386-1.10-12.patch
-  --arch=i386
+$ scripts/prepare-kernel.sh --linux=/usr/src/linux-2.6.23-ipipe
 
 Note: The script will infer the location of the Xenomai kernel code
 from its own location within the Xenomai source tree. In other words,
Index: xenomai/scripts/prepare-kernel.sh
===================================================================
--- xenomai.orig/scripts/prepare-kernel.sh
+++ xenomai/scripts/prepare-kernel.sh
@@ -353,10 +353,10 @@ eval linux_`grep '^PATCHLEVEL =' $linux_
 eval linux_`grep '^SUBLEVEL =' $linux_tree/Makefile | sed -e 's, ,,g'`
 eval linux_`grep '^VERSION =' $linux_tree/Makefile | sed -e 's, ,,g'`
 
-linux_version="$linux_VERSION.$linux_PATCHLEVEL.$linux_SUBLEVEL$linux_EXTRAVERSION"
+linux_version="$linux_VERSION.$linux_PATCHLEVEL.$linux_SUBLEVEL"
 
 if test x$verbose = x1; then
-echo "Preparing kernel $linux_version in $linux_tree..."
+echo "Preparing kernel $linux_version$linux_EXTRAVERSION in $linux_tree..."
 fi
 
 if test -r $linux_tree/include/linux/ipipe.h; then
@@ -369,7 +369,7 @@ elif test -r $linux_tree/include/linux/a
    exit 2
 else
    if test x$adeos_patch = x; then
-      default_adeos_patch=`( ls $xenomai_root/ksrc/arch/$xenomai_arch/patches/adeos-ipipe-$linux_version-$linux_arch-*|sort -r ) 2>/dev/null | head -n1`
+      default_adeos_patch="`( ls $xenomai_root/ksrc/arch/$xenomai_arch/patches/adeos-ipipe-$linux_version*-{$linux_arch,$xenomai_arch}-*|sort -r ) 2>/dev/null | head -n1`"
    fi
    if test x$default_adeos_patch = x; then
       default_adeos_patch=/dev/null
@@ -393,7 +393,7 @@ else
    curdir=$PWD
    cd $linux_tree && patch --dry-run -p1 -f < $adeos_patch || { 
         cd $curdir;
-        echo "$me: Unable to patch kernel $linux_version with `basename $adeos_patch`." >&2
+        echo "$me: Unable to patch kernel $linux_version$linux_EXTRAVERSION with `basename $adeos_patch`." >&2
         exit 2;
    }
    patch -p1 -f -s < $adeos_patch

Re: [Xenomai-core] Patch selection by prepare-kernel.sh (again)

[Philippe Gerum]

Jan Kiszka wrote:
> Jan Kiszka wrote:
>> Jan Kiszka wrote:
>>> Philippe,
>>>
>>> what was the idea behind commit #3256 [1]? It breaks the original
>>> behavior of selection adeos-ipipe-2.6.23-... in the presence of 2.6.23.x
>>> - and leaves a bad first impression for beginners (I received such a
>>> notice here...). Was this for the special case
>>> adeos-ipipe-2.6.20.21-i386-1.12-02.patch? But I think the common case
>>> should take precedence over this corner case here. Unless there are more
>>> issues with the original version, I would vote for restoring it.
>>>
>>> Alternatively, we could add a safety belt by trying to apply the patch
>>> with --dry-run first and only continue if this works out without
>>> failures. Can hack this up if it's acceptable.
>> Oh, this check already exists!
>>
>> So let's try this version, which additionally cleans up the related
>> documentation. Catches both 2.6.23* and 2.6.24* for me here, but I may
>> miss some cases (2.4?).
> 
> The documentation hunk was nonsense, here is a hopefully better version.
> 

Makes sense and works fine for the kernel/arch combos I've tested so
far. Merged, thanks.

-- 
Philippe.

Re: [Xenomai-core] Patch selection by prepare-kernel.sh (again)

[Gilles Chanteperdrix]

[-- Attachment #1: Type: text/plain, Size: 1449 bytes --]

On Jan 22, 2008 5:23 PM, Philippe Gerum <rpm@xenomai.org> wrote:
>
> Jan Kiszka wrote:
> > Jan Kiszka wrote:
> >> Jan Kiszka wrote:
> >>> Philippe,
> >>>
> >>> what was the idea behind commit #3256 [1]? It breaks the original
> >>> behavior of selection adeos-ipipe-2.6.23-... in the presence of 2.6.23.x
> >>> - and leaves a bad first impression for beginners (I received such a
> >>> notice here...). Was this for the special case
> >>> adeos-ipipe-2.6.20.21-i386-1.12-02.patch? But I think the common case
> >>> should take precedence over this corner case here. Unless there are more
> >>> issues with the original version, I would vote for restoring it.
> >>>
> >>> Alternatively, we could add a safety belt by trying to apply the patch
> >>> with --dry-run first and only continue if this works out without
> >>> failures. Can hack this up if it's acceptable.
> >> Oh, this check already exists!
> >>
> >> So let's try this version, which additionally cleans up the related
> >> documentation. Catches both 2.6.23* and 2.6.24* for me here, but I may
> >> miss some cases (2.4?).
> >
> > The documentation hunk was nonsense, here is a hopefully better version.
> >
>
> Makes sense and works fine for the kernel/arch combos I've tested so
> far. Merged, thanks.

configure.in probably needs a similar fix for the --enable-linux-build
option to work with the -x86 patches.

-- 
                                               Gilles Chanteperdrix

[-- Attachment #2: xeno-x86-configure.diff --]
[-- Type: application/octet-stream, Size: 319 bytes --]

Index: configure.in
===================================================================
--- configure.in	(révision 3431)
+++ configure.in	(copie de travail)
@@ -337,6 +337,7 @@
   fi
   case "$ADEOS_PATCH" in
 	*-powerpc-*) XENO_LINUX_ARCH=powerpc;;
+	*-x86-*) XENO_LINUX_ARCH=x86;;
   esac
   AC_SUBST(ADEOS_PATCH)