All of lore.kernel.org
 help / color / mirror / Atom feed
* cross-compiling flags?
@ 2008-11-04 15:56 Hollis Blanchard
  2008-11-06  2:51 ` Liu Yu
                   ` (3 more replies)
  0 siblings, 4 replies; 5+ messages in thread
From: Hollis Blanchard @ 2008-11-04 15:56 UTC (permalink / raw)
  To: kvm-ppc

Hi Yu, Christian tells me you have some weird cross-build setup that
brings you to want to use gcc's -idirafter flag for zlib headers. I
don't like using something that obscure. Can you elaborate on the errors
you get when using -I ?

-- 
Hollis Blanchard
IBM Linux Technology Center


^ permalink raw reply	[flat|nested] 5+ messages in thread

* RE: cross-compiling flags?
  2008-11-04 15:56 cross-compiling flags? Hollis Blanchard
@ 2008-11-06  2:51 ` Liu Yu
  2008-11-06  4:39 ` Hollis Blanchard
                   ` (2 subsequent siblings)
  3 siblings, 0 replies; 5+ messages in thread
From: Liu Yu @ 2008-11-06  2:51 UTC (permalink / raw)
  To: kvm-ppc




> -----Original Message-----
> From: Hollis Blanchard [mailto:hollisb@us.ibm.com] 
> Sent: Tuesday, November 04, 2008 11:57 PM
> To: Liu Yu-B13201
> Cc: kvm-ppc
> Subject: cross-compiling flags?
> 
> Hi Yu, Christian tells me you have some weird cross-build setup that
> brings you to want to use gcc's -idirafter flag for zlib headers. I
> don't like using something that obscure. Can you elaborate on 
> the errors
> you get when using -I ?
> 


Hi Hollis,

Sorry, I attended FTF yesterday.

I don't know what libc headers are used in your case. The host libc
header files?
In my case, the libc headers are already appointed to a
toolchain-combined version in LTIB(assume you know it).
But zlib.h is not existing so I had to use qemu-cflag to appoint another
dir

The problem I met is that my zlib.h is mixed with a lot of other libc
header files.
After I '-I' this dir, the build adopted libc header files in this dir,
then I met a lot of errors such as "redifination of int8_t".

I think it's because that the dir appointed by '-I' will be considerred
prior,
but the dir appointed by '-idirafter' will be considerred at last.
That's why I can pass the building by using '-idirafter'.

I think now that the qemu-cflags is additional options based on the
configuration default options,
the options in qemu-cflags should not influence the default opions in
configuration.

What do you think of it? Is this approach the best way?



^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: cross-compiling flags?
  2008-11-04 15:56 cross-compiling flags? Hollis Blanchard
  2008-11-06  2:51 ` Liu Yu
@ 2008-11-06  4:39 ` Hollis Blanchard
  2008-11-06  7:44 ` Liu Yu
  2008-11-06 15:39 ` Hollis Blanchard
  3 siblings, 0 replies; 5+ messages in thread
From: Hollis Blanchard @ 2008-11-06  4:39 UTC (permalink / raw)
  To: kvm-ppc

On Wednesday 05 November 2008 20:51:18 Liu Yu wrote:
> 
> > -----Original Message-----
> > From: Hollis Blanchard [mailto:hollisb@us.ibm.com] 
> > Sent: Tuesday, November 04, 2008 11:57 PM
> > To: Liu Yu-B13201
> > Cc: kvm-ppc
> > Subject: cross-compiling flags?
> > 
> > Hi Yu, Christian tells me you have some weird cross-build setup that
> > brings you to want to use gcc's -idirafter flag for zlib headers. I
> > don't like using something that obscure. Can you elaborate on 
> > the errors
> > you get when using -I ?
> > 
> 
> I don't know what libc headers are used in your case. The host libc
> header files?

When I cross-compile qemu, it's using the cross-toolchain glibc headers. Since 
my toolchain is from crosstool, those headers are 
at /opt/crosstool/gcc-3.4.5-glibc-2.3.6/powerpc-440-linux-gnu/powerpc-440-linux-gnu

> In my case, the libc headers are already appointed to a
> toolchain-combined version in LTIB(assume you know it).
> But zlib.h is not existing so I had to use qemu-cflag to appoint another
> dir

If your zlib headers are mixed in with the target libc headers, then why do 
you need any -I at all?

I have heard of LTIB, but I don't know how its toolchain is installed.

-- 
Hollis Blanchard
IBM Linux Technology Center

^ permalink raw reply	[flat|nested] 5+ messages in thread

* RE: cross-compiling flags?
  2008-11-04 15:56 cross-compiling flags? Hollis Blanchard
  2008-11-06  2:51 ` Liu Yu
  2008-11-06  4:39 ` Hollis Blanchard
@ 2008-11-06  7:44 ` Liu Yu
  2008-11-06 15:39 ` Hollis Blanchard
  3 siblings, 0 replies; 5+ messages in thread
From: Liu Yu @ 2008-11-06  7:44 UTC (permalink / raw)
  To: kvm-ppc



> -----Original Message-----
> From: Hollis Blanchard [mailto:hollisb@us.ibm.com] 
> Sent: Thursday, November 06, 2008 12:40 PM
> To: Liu Yu-B13201
> Cc: kvm-ppc
> Subject: Re: cross-compiling flags?
> 
> On Wednesday 05 November 2008 20:51:18 Liu Yu wrote:
> > 
> > > -----Original Message-----
> > > From: Hollis Blanchard [mailto:hollisb@us.ibm.com] 
> > > Sent: Tuesday, November 04, 2008 11:57 PM
> > > To: Liu Yu-B13201
> > > Cc: kvm-ppc
> > > Subject: cross-compiling flags?
> > > 
> > > Hi Yu, Christian tells me you have some weird cross-build 
> setup that
> > > brings you to want to use gcc's -idirafter flag for zlib 
> headers. I
> > > don't like using something that obscure. Can you elaborate on 
> > > the errors
> > > you get when using -I ?
> > > 
> > 
> > I don't know what libc headers are used in your case. The host libc
> > header files?
> 
> When I cross-compile qemu, it's using the cross-toolchain 
> glibc headers. Since 
> my toolchain is from crosstool, those headers are 
> at 
> /opt/crosstool/gcc-3.4.5-glibc-2.3.6/powerpc-440-linux-gnu/pow
> erpc-440-linux-gnu

Yes, me too.

> 
> > In my case, the libc headers are already appointed to a
> > toolchain-combined version in LTIB(assume you know it).
> > But zlib.h is not existing so I had to use qemu-cflag to 
> appoint another
> > dir
> 
> If your zlib headers are mixed in with the target libc 
> headers, then why do 
> you need any -I at all?

The zlib headers are mixed in powerpc local libc headers not the toolchain cross libc headers.
And it seems the local libc headers is unfit for cross building.


^ permalink raw reply	[flat|nested] 5+ messages in thread

* RE: cross-compiling flags?
  2008-11-04 15:56 cross-compiling flags? Hollis Blanchard
                   ` (2 preceding siblings ...)
  2008-11-06  7:44 ` Liu Yu
@ 2008-11-06 15:39 ` Hollis Blanchard
  3 siblings, 0 replies; 5+ messages in thread
From: Hollis Blanchard @ 2008-11-06 15:39 UTC (permalink / raw)
  To: kvm-ppc

On Thu, 2008-11-06 at 15:44 +0800, Liu Yu wrote:
> > 
> > When I cross-compile qemu, it's using the cross-toolchain 
> > glibc headers. Since 
> > my toolchain is from crosstool, those headers are 
> > at 
> > /opt/crosstool/gcc-3.4.5-glibc-2.3.6/powerpc-440-linux-gnu/pow
> > erpc-440-linux-gnu
> 
> Yes, me too.
> 
> > 
> > > In my case, the libc headers are already appointed to a
> > > toolchain-combined version in LTIB(assume you know it).
> > > But zlib.h is not existing so I had to use qemu-cflag to 
> > appoint another
> > > dir
> > 
> > If your zlib headers are mixed in with the target libc 
> > headers, then why do 
> > you need any -I at all?
> 
> The zlib headers are mixed in powerpc local libc headers not the toolchain cross libc headers.
> And it seems the local libc headers is unfit for cross building.

Since I'm not familiar with LTIB, I chatted with Kumar Gala about this
briefly, and he suggested that you post your issue to the internal
Freescale LTIB list.

My understanding is that you have a special directory for target
headers, which is used by the cross toolchain. Without understanding the
details, my suggestion would be to build a second copy of zlib from
source, and configure it with --prefix=/target/specific/directory.

-- 
Hollis Blanchard
IBM Linux Technology Center


^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2008-11-06 15:39 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-11-04 15:56 cross-compiling flags? Hollis Blanchard
2008-11-06  2:51 ` Liu Yu
2008-11-06  4:39 ` Hollis Blanchard
2008-11-06  7:44 ` Liu Yu
2008-11-06 15:39 ` Hollis Blanchard

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.