All of lore.kernel.org
 help / color / mirror / Atom feed
From: Philippe Gerum <rpm@xenomai.org>
To: Wolfgang Grandegger <wolfgang.grandegger@domain.hid>
Cc: adeos-main <adeos-main@gna.org>
Subject: Re: [Adeos-main] Re: Patch/Code maintanance
Date: Thu, 03 Jun 2004 11:09:02 +0200	[thread overview]
Message-ID: <1086253742.680.8.camel@domain.hid> (raw)
In-Reply-To: <40C03997.5040805@domain.hid>

On Fri, 2004-06-04 at 10:57, Wolfgang Grandegger wrote:
> On 06/03/2004 10:31 AM Philippe Gerum wrote:
> > On Thu, 2004-06-03 at 21:08, Wolfgang Grandegger wrote:
> >> Hi Philippe,
> >> 
> >> I started porting ADEOS to PowerPC with "2.4r11c1" and now I want to
> >> catch up with "24r13c2" before I release it. I realized that there are
> >> quite some differences and I wonder how you maintain the ADEOS code,
> >> especially how do you get or handle the files in "linux" and "patches".
> >> 
> > 
> > There are two major trees, one for 2.4, the other for 2.6, respectively
> > available under linux/v2.4 and linux/v2.6.
> > 
> > For each supported kernel version, I simply create some "shadow tree"
> > under linux/v2.x/<kernel-ver>-<arch>, only containing the kernel files
> > modified by Adeos.
> > 
> > The generic Adeos support which is the same for any version of the host
> > kernel in a given major branch is stored under linux/v2.x/adeos-core.
> > 
> > Basically, merging a kernel-specific tree with the Adeos generic support
> > gives you all the files applicable to a genuine kernel tree in order to
> > Adeos-enable it. This way, for working on Adeos, you just need to untar
> > the vanilla kernel tree of interest, and replace the set of affected
> > files by symlinks to the corresponding files in the CVS tree.
> > 
> > When a development milestone has been reached for a given kernel release
> > / Adeos revision, I simply build a patch against the corresponding
> > vanilla tree, and feed patches/ with it. The same patch is then copied
> > to the download area on GNA, and a news is published there.
> 
> OK, and when you modify files on one shadow tree you adapt the files on
> the other trees manually!? Hm, unfortunately, this is the most tedious part.
> 

Indeed yes, especially for x86 with an increasing number of supported
versions (I tend to discontinue the oldest one from times to times,
though). But fortunately, not that many changes occur in the
kernel-dependent portions these days, but most of them concern the
generic Adeos parts.

> I have now updated to 2.4r13c2 and it still works :-). I took the files
> from the "adeos-core" sub-directory as they are and adapted the rest
> manually. Finally, beginning of next week I want to check the PowerPC
> port in. Most parts do not affect other code. Just the examples need
> some attention as I have not tested them on X86. They now supports cross
> compilation and add some PowerPC related timing code for the nice
> latency test modules from "Der Herr Hofrat". I have attached a patch in
> case you (or somebody else) want to have a look.
> 

I will. Many thanks for your contribution.

> Thanks.
> 
> Wolfgang.
> 
> 
> 
> ______________________________________________________________________
> + diff -u linux/examples/measurement/irq_jitter.c.ORIG linux/examples/measurement/irq_jitter.c
> --- linux/examples/measurement/irq_jitter.c.ORIG	Sun Feb 29 19:31:34 2004
> +++ linux/examples/measurement/irq_jitter.c	Fri Jun  4 00:38:28 2004
> @@ -46,8 +46,22 @@
>  #if defined(CONFIG_X86) && defined(CONFIG_X86_TSC)
>  #include <asm/msr.h> /* rdtscll - ia32 specific */
>  #define hwtimer(x) rdtscll(x)
> +#elif defined(CONFIG_PPC)
> +#include <asm/time.h>
> +static inline void ppc_timebase(long long *x)
> +{
> +    unsigned long tbl, tbu1, tbu2;
> +    do 
> +	{
> +	tbu1 = get_tbu();
> +	tbl  = get_tbl();
> +	tbu2 = get_tbu();	
> +	} while (tbu1 != tbu2);
> +    *x = (long long)tbl + ((long long)tbu1 << 32);
> +}
> +#define hwtimer(x) ppc_timebase(&(x))
>  #else
> -	#error "Only X86 supported for now"
> +	#error "Only X86 and PPC supported for now"
>  #endif
>  
> 
> @@ -88,7 +102,8 @@
>  			     IPIPE_DYNAMIC_MASK);
>  	}
>  
> -    scaller = cpu_khz/1000;
> +    /* might be incorrect due to rounding effects */
> +    scaller = (unsigned long)sys_info.cpufreq / 1000000;
>      period = (1000/HZ)*1000;
>      printk("scaller set to %ld\n",(unsigned long) scaller);
>      printk("period set to %ld\n",(long) period);
> + diff -u linux/examples/measurement/Makefile.ORIG linux/examples/measurement/Makefile
> --- linux/examples/measurement/Makefile.ORIG	Sun Feb 29 19:31:34 2004
> +++ linux/examples/measurement/Makefile	Fri Jun  4 00:00:56 2004
> @@ -3,21 +3,35 @@
>  # Environment" feature in the general setup when configuring this
>  # kernel.
>  
> +ifndef LINUXSRC
>  LINUXSRC = /usr/src/linux
> -ARCH = i686
> +endif
> +
> +DEBUG	= -g -ggdb
> +DEFINES = -D__KERNEL__ -DMODULE $(DEBUG)
>  
> -DEFINES = -D__KERNEL__  -DMODULE -DEXPORT_SYMTAB
> +include $(LINUXSRC)/.config
> +
> +ifdef CONFIG_PPC
> +CFLAGS = $(DEFINES) -O2 -Wall -Wstrict-prototypes -Wno-trigraphs \
> +         -fno-strict-aliasing -fno-common -fomit-frame-pointer \
> +	 -fsigned-char -msoft-float -pipe -ffixed-r2 \
> +	 -Wno-uninitialized -mmultiple -mstring \
> +	-I$(LINUXSRC)/include -I$(LINUXSRC)/arch/ppc
> +else
> +ARCH = i686
>  CFLAGS = $(DEFINES) -I$(LINUXSRC)/include -Wall -Wstrict-prototypes \
>  	-Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing \
>  	-fno-common -mpreferred-stack-boundary=2 -march=${ARCH}
> +endif
>  
>  all: irq_jitter.o virtual_irq_jitter.o
>  
>  virtual_irq_jitter.o: virtual_irq_jitter.c
> -	gcc $(CFLAGS) -o $@ -c $<
> +	$(CROSS_COMPILE)gcc $(CFLAGS) -o $@ -c $<
>  
>  irq_jitter.o: irq_jitter.c
> -	gcc $(CFLAGS) -o $@ -c $<
> +	$(CROSS_COMPILE)gcc $(CFLAGS) -o $@ -c $<
>  
>  clean:
>  	rm -f *.o
> + diff -u linux/examples/measurement/virtual_irq_jitter.c.ORIG linux/examples/measurement/virtual_irq_jitter.c
> --- linux/examples/measurement/virtual_irq_jitter.c.ORIG	Fri Jun  4 00:24:42 2004
> +++ linux/examples/measurement/virtual_irq_jitter.c	Fri Jun  4 01:04:49 2004
> @@ -134,9 +134,13 @@
>          unsigned int a1, a2;
>          unsigned long long res;
>  
> +#ifdef __LITTLE_ENDIAN
>          a1 = ((unsigned int*)&a)[0];
>          a2 = ((unsigned int*)&a)[1];
> -
> +#else
> +        a1 = ((unsigned int*)&a)[1];
> +        a2 = ((unsigned int*)&a)[0];
> +#endif
>          res = a1/b0 +
>                  (unsigned long long)a2 * (unsigned long long)(0xffffffff/b0) +
>                  a2 / b0 +
> + diff -u linux/examples/simple/Makefile.ORIG linux/examples/simple/Makefile
> --- linux/examples/simple/Makefile.ORIG	Sun Feb 29 19:31:33 2004
> +++ linux/examples/simple/Makefile	Fri Jun  4 00:01:24 2004
> @@ -3,23 +3,38 @@
>  # Environment" feature in the general setup when configuring this
>  # kernel.
>  
> +ifndef LINUXSRC
>  LINUXSRC = /usr/src/linux
> +endif
>  
> -DEFINES = -D__KERNEL__  -DMODULE -DEXPORT_SYMTAB
> -CFLAGS = $(DEFINES) -I$(LINUXSRC)/include -Wall -Wstrict-prototypes -Wno-trigraphs \
> -        -O2 -fomit-frame-pointer -fno-strict-aliasing -fno-common \
> -        -mpreferred-stack-boundary=2 -march=i686
> +DEBUG	= -g -ggdb
> +DEFINES = -D__KERNEL__ -DMODULE $(DEBUG)
> +
> +include $(LINUXSRC)/.config
> +
> +ifdef CONFIG_PPC
> +CFLAGS = $(DEFINES) -O2 -Wall -Wstrict-prototypes -Wno-trigraphs \
> +         -fno-strict-aliasing -fno-common -fomit-frame-pointer \
> +	 -fsigned-char -msoft-float -pipe -ffixed-r2 \
> +	 -Wno-uninitialized -mmultiple -mstring \
> +	-I$(LINUXSRC)/include -I$(LINUXSRC)/arch/ppc
> +else
> +ARCH = i686
> +CFLAGS = $(DEFINES) -I$(LINUXSRC)/include -Wall -Wstrict-prototypes \
> +	-Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing \
> +	-fno-common -mpreferred-stack-boundary=2 -march=${ARCH}
> +endif
>  
>  all: adtest.o interdomain_irq.o interdomain_mutex.o
>  
>  adtest.o: adtest.c
> -	gcc $(CFLAGS) -o $@ -c $<
> +	$(CROSS_COMPILE)gcc $(CFLAGS) -o $@ -c $<
>  
>  interdomain_irq.o: interdomain_irq.c
> -	gcc $(CFLAGS) -o $@ -c $<
> +	$(CROSS_COMPILE)gcc $(CFLAGS) -o $@ -c $<
>  
>  interdomain_mutex.o: interdomain_mutex.c
> -	gcc $(CFLAGS) -o $@ -c $<
> +	$(CROSS_COMPILE)gcc $(CFLAGS) -o $@ -c $<
>  
>  clean:
>  	rm -f *.o
-- 

Philippe.



  reply	other threads:[~2004-06-03  9:09 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-06-03 19:08 [Adeos-main] Patch/Code maintanance Wolfgang Grandegger
2004-06-03  8:31 ` [Adeos-main] " Philippe Gerum
2004-06-04  8:57   ` Wolfgang Grandegger
2004-06-03  9:09     ` Philippe Gerum [this message]
2004-06-04 15:54     ` Wolfgang Denk
2004-06-07  8:40       ` Wolfgang Grandegger

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1086253742.680.8.camel@domain.hid \
    --to=rpm@xenomai.org \
    --cc=adeos-main@gna.org \
    --cc=wolfgang.grandegger@domain.hid \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.