[PATCH] [RFC] implement size_append in perl

[PATCH] [RFC] implement size_append in perl

[Alexander Shishkin]

The current implementation has been several times fixed to work around
different shells and build environments and it remains a somewhat tricky
shell script. This patch replaces all the dependencies by a single line
of perl. Build dependency on perl is not new, so that shouldn't be a
problem.

The new implementation does exactly what it's supposed to: it outputs
a little-endian 32-bit unsigned integer of a file's length.

This might simplify matters with different build environments and
different shells and is also considerably shorter and hopefully more
readable.

Signed-off-by: Alexander Shishkin <virtuoso@slind.org>
Cc: Sebastian Dalfuß <sd@sedf.de>
Cc: Oliver Hartkopp <oliver@hartkopp.net>
Cc: Michael Guntsche <mike@it-loops.com>
Cc: Jonathan Nieder <jrnieder@gmail.com>
Cc: Michael Tokarev <mjt@tls.msk.ru>
Cc: Alek Du <alek.du@intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Marek <mmarek@suse.cz>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Albin Tonnerre <albin.tonnerre@free-electrons.com>
Cc: linux-kbuild@vger.kernel.org
Cc: linux-kernel@vger.kernel.org

---
 scripts/Makefile.lib |   16 +---------------
 1 files changed, 1 insertions(+), 15 deletions(-)

diff --git a/scripts/Makefile.lib b/scripts/Makefile.lib
index 54fd1b7..5bfd872 100644
--- a/scripts/Makefile.lib
+++ b/scripts/Makefile.lib
@@ -213,21 +213,7 @@ cmd_gzip = (cat $(filter-out FORCE,$^) | gzip -f -9 > $@) || \
 
 # Bzip2 and LZMA do not include size in file... so we have to fake that;
 # append the size as a 32-bit littleendian number as gzip does.
-size_append = printf $(shell						\
-dec_size=0;								\
-for F in $1; do								\
-	fsize=$$(stat -c "%s" $$F);					\
-	dec_size=$$(expr $$dec_size + $$fsize);				\
-done;									\
-printf "%08x\n" $$dec_size |						\
-	sed 's/\(..\)/\1 /g' | {					\
-		read ch0 ch1 ch2 ch3;					\
-		for ch in $$ch3 $$ch2 $$ch1 $$ch0; do			\
-			printf '%s%03o' '\\' $$((0x$$ch)); 		\
-		done;							\
-	}								\
-)
-
+size_append = $(PERL) -e 'print pack("V", (stat "$(lastword $1)")[7])'
 quiet_cmd_bzip2 = BZIP2   $@
 cmd_bzip2 = (cat $(filter-out FORCE,$^) | \
 	bzip2 -9 && $(call size_append, $(filter-out FORCE,$^))) > $@ || \
-- 
1.7.1

Re: [PATCH] [RFC] implement size_append in perl

[H. Peter Anvin]

On 07/14/2010 07:01 AM, Alexander Shishkin wrote:
> +size_append = $(PERL) -e 'print pack("V", (stat "$(lastword $1)")[7])'

Why the "lastword"?

	-hpa

Re: [PATCH] [RFC] implement size_append in perl

[Alexander Shishkin]

On Wed, Jul 14, 2010 at 07:54:43 -0700, H. Peter Anvin wrote:
> On 07/14/2010 07:01 AM, Alexander Shishkin wrote:
> > +size_append = $(PERL) -e 'print pack("V", (stat "$(lastword $1)")[7])'
> 
> Why the "lastword"?

There's a whitespace right after it due to it being part of the target
spec "$(obj)/piggy.$(suffix_y) FORCE", where "FORCE" is being filtered
out.

Regards,
--
Alex

Re: [PATCH] [RFC] implement size_append in perl

[Michal Marek]

On 07/14/2010 06:12 PM, Alexander Shishkin wrote:
> On Wed, Jul 14, 2010 at 07:54:43 -0700, H. Peter Anvin wrote:
>> On 07/14/2010 07:01 AM, Alexander Shishkin wrote:
>>> +size_append = $(PERL) -e 'print pack("V", (stat "$(lastword $1)")[7])'
>>
>> Why the "lastword"?
> 
> There's a whitespace right after it due to it being part of the target
> spec "$(obj)/piggy.$(suffix_y) FORCE", where "FORCE" is being filtered
> out.

Then you should use $(strip $1), which is more obvious IMO.

Michal

Re: [PATCH] [RFC] implement size_append in perl

[Sam Ravnborg]

On Wed, Jul 14, 2010 at 05:01:03PM +0300, Alexander Shishkin wrote:
> The current implementation has been several times fixed to work around
> different shells and build environments and it remains a somewhat tricky
> shell script. This patch replaces all the dependencies by a single line
> of perl. Build dependency on perl is not new, so that shouldn't be a
> problem.
> 
> The new implementation does exactly what it's supposed to: it outputs
> a little-endian 32-bit unsigned integer of a file's length.
> 
> This might simplify matters with different build environments and
> different shells and is also considerably shorter and hopefully more
> readable.

Reminds me that I had a patch that moved this to mkpiggy.c.
Goggle gave this hit:

    http://lists.openwall.net/linux-kernel/2009/12/20/92

I will try to update this to cover arm also and check
if other archs needs this.

In other word - move thes size stuff to mkpiggy so we
skip this extra step.

	Sam

Re: [PATCH] [RFC] implement size_append in perl

[Alexander Shishkin]

On Wed, Jul 14, 2010 at 05:23:37 +0200, Sam Ravnborg wrote:
> On Wed, Jul 14, 2010 at 05:01:03PM +0300, Alexander Shishkin wrote:
> > The current implementation has been several times fixed to work around
> > different shells and build environments and it remains a somewhat tricky
> > shell script. This patch replaces all the dependencies by a single line
> > of perl. Build dependency on perl is not new, so that shouldn't be a
> > problem.
> > 
> > The new implementation does exactly what it's supposed to: it outputs
> > a little-endian 32-bit unsigned integer of a file's length.
> > 
> > This might simplify matters with different build environments and
> > different shells and is also considerably shorter and hopefully more
> > readable.
> 
> Reminds me that I had a patch that moved this to mkpiggy.c.
> Goggle gave this hit:
> 
>     http://lists.openwall.net/linux-kernel/2009/12/20/92

Ok, thanks for the reference.

> I will try to update this to cover arm also and check
> if other archs needs this.
> 
> In other word - move thes size stuff to mkpiggy so we
> skip this extra step.

Hm, maybe this mkpiggy can be made arch-independent and moved to scripts/
as well?

Regards,
--
Alex

Re: [PATCH] [RFC] implement size_append in perl

[H. Peter Anvin]

On 07/14/2010 09:17 AM, Alexander Shishkin wrote:
> 
> Hm, maybe this mkpiggy can be made arch-independent and moved to scripts/
> as well?
> 

It depends on which architectures want this wrapping format.

	-hpa

Re: [PATCH] [RFC] implement size_append in perl

[Alexander Shishkin]

On Wed, Jul 14, 2010 at 09:57:52 -0700, H. Peter Anvin wrote:
> On 07/14/2010 09:17 AM, Alexander Shishkin wrote:
> > 
> > Hm, maybe this mkpiggy can be made arch-independent and moved to scripts/
> > as well?
> > 
> 
> It depends on which architectures want this wrapping format.

Yes, well, when it's more than one, it already kind of makes sense.

Regards,
--
Alex