2.5 questions

2.5 questions

[Kyle Harris]

How do I get the latest code base for 2.5 kernels? I pulled the latest cvs 
but it appears this must be for 2.4 since it doesn't compile. Is there a 
separate branch for 2.5?

How stable is mtd for 2.5? I'm seeing some very strange behavior where if I 
make slight mods to even script files, the files become unusable and cause 
the system to hang. I'm running 2.5.30-rmk1-pxa1 with jffs2 root fs.

Thanks, Kyle.

Re: 2.5 questions

[David Woodhouse]

kharris@nexus-tech.net said:
> How do I get the latest code base for 2.5 kernels? I pulled the latest
> cvs  but it appears this must be for 2.4 since it doesn't compile. Is
> there a  separate branch for 2.5?

Should do. What breaks? It'll only be cosmetic, surely?

> How stable is mtd for 2.5? I'm seeing some very strange behavior where
> if I  make slight mods to even script files, the files become unusable
> and cause  the system to hang. I'm running 2.5.30-rmk1-pxa1 with jffs2
> root fs. 

Not expected to break. Not highly tested either. More details please. 
SysRq-T is your friend.

--
dwmw2

Re: 2.5 questions

[Kyle Harris]

On Friday 13 September 2002 03:24 am, David Woodhouse wrote:

> kharris@nexus-tech.net said:
> > How do I get the latest code base for 2.5 kernels? I pulled the latest
> > cvs  but it appears this must be for 2.4 since it doesn't compile. Is
> > there a  separate branch for 2.5?
>
> Should do. What breaks? It'll only be cosmetic, surely?

I ran the pathcin.sh script to set this up. Below is first error, which is 
2.4 - 2.5 related (MINOR() and kdev_t stuff).

  arm-linux-gcc -Wp,-MD,./.mtdchar.o.d -D__KERNEL__ 
-I/home/kyle/handheld/homebase/kernel/linux-2.5.30.newmtd/include -Wall 
-Wstrict-prototypes -Wno-trigraphs -Os -fno-strict-aliasing -fno-common -g 
-mapcs -mapcs-32 -D__LINUX_ARM_ARCH__=5 -march=armv4 -Wa,-mxscale  
-mtune=strongarm  -mshort-load-bytes -msoft-float -Wa,-mno-fpu -nostdinc 
-iwithprefix include    -DKBUILD_BASENAME=mtdchar   -c -o mtdchar.o mtdchar.c
mtdchar.c: In function `mtd_open':
mtdchar.c:63: invalid operands to binary &
make[2]: *** [mtdchar.o] Error 1
make[2]: Leaving directory 
`/home/kyle/handheld/homebase/kernel/linux-2.5.30.newmtd/drivers/mtd'
make[1]: *** [mtd] Error 2
make[1]: Leaving directory 
`/home/kyle/handheld/homebase/kernel/linux-2.5.30.newmtd/drivers'
make: *** [drivers] Error 2

>
> > How stable is mtd for 2.5? I'm seeing some very strange behavior where
> > if I  make slight mods to even script files, the files become unusable
> > and cause  the system to hang. I'm running 2.5.30-rmk1-pxa1 with jffs2
> > root fs.
>
> Not expected to break. Not highly tested either. More details please.
> SysRq-T is your friend.

I'm still not entirely sure what's happening. It may be a race condition at 
boot that is breaking something and completely unrelated to the minor changes 
I made to the root fs. I'm not familiar sith SysRq-T. Where do I find more 
about this?

Thanks. Kyle.

Re: 2.5 questions

[David Woodhouse]

kharris@nexus-tech.net said:
>  I ran the pathcin.sh script to set this up. Below is first error,
> which is  2.4 - 2.5 related (MINOR() and kdev_t stuff). 

OK, trivial to fix if anyone actually cares then. I just haven't received or
applied a patch to do so yet.

--
dwmw2

Re: 2.5 questions

[Kyle Harris]

On Friday 13 September 2002 08:45 am, David Woodhouse wrote:

> kharris@nexus-tech.net said:
> >  I ran the pathcin.sh script to set this up. Below is first error,
> > which is  2.4 - 2.5 related (MINOR() and kdev_t stuff).
>
> OK, trivial to fix if anyone actually cares then. I just haven't received
> or applied a patch to do so yet.

Yes. I fixed this one and then the compile failed someplace else. So it 
appeared that the code was not compatible with 2.5. So that's why I asked.

Kyle.

Re: 2.5 questions

[David Woodhouse]

kharris@nexus-tech.net said:
>  Yes. I fixed this one and then the compile failed someplace else. So
> it  appeared that the code was not compatible with 2.5. So that's why
> I asked.

Well send me the first patch anyway, making sure it builds on both 2.4 and 
2.5, and then report the second one. If it's the fact that nftl doesn't 
compile, frown at Thomas :)

--
dwmw2

Re: 2.5 questions

[David Woodhouse]

kharris@nexus-tech.net said:
>  Yes. I fixed this one and then the compile failed someplace else. So
> it  appeared that the code was not compatible with 2.5. So that's why
> I asked. 

Most of it now compiles with 2.5. Only the block devices remain -- they 
need more attention and probably we'll end up splitting them into separate 
files for <=2.4 and >=2.5.

--
dwmw2