[RFC] scsi headers, in glibc-dev or linux-libc-headers-dev?

[RFC] scsi headers, in glibc-dev or linux-libc-headers-dev?

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

Currently conflicting headers are both in glibc-dev and
linux-libc-headers-dev, which isn't a good thing. According to the
interwebs we should use the linux-libc-headers version.

Anyone opposed to removing the scsi headers from glibc? It would
basically involve reverting
http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=b8bb2a62b3916470c752bd79d31322b4358fc676
and
http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=83f41716ab6a2a9d83d4ff044dcef00595ecfeb2

regards,

Koen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFL1rkKMkyGM64RGpERAuL4AJ9SNDjpJyCb2NMXbkHqTO61yUnFOQCdGud2
3sWfhdGpfGVfB7pQmiyekCk=
=ooCD
-----END PGP SIGNATURE-----

Re: [RFC] scsi headers, in glibc-dev or linux-libc-headers-dev?

[Tom Rini]

On Tue, 2010-04-27 at 12:14 +0200, Koen Kooi wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hi,
> 
> Currently conflicting headers are both in glibc-dev and
> linux-libc-headers-dev, which isn't a good thing. According to the
> interwebs we should use the linux-libc-headers version.
> 
> Anyone opposed to removing the scsi headers from glibc? It would
> basically involve reverting
> http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=b8bb2a62b3916470c752bd79d31322b4358fc676
> and
> http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=83f41716ab6a2a9d83d4ff044dcef00595ecfeb2

So, I did both of those after checking the interwebs and poking some
kernel folks (see comments in
http://bugs.openembedded.org/show_bug.cgi?id=4262 ).  The question is,
did the kernel folks change their mind again and we need to either drop
older linux-libc-headers (which I'll just assume is a no-go) or go back
to removing them in new enough linux-libc-headers.

-- 
Tom Rini <tom_rini@mentor.com>
Mentor Graphics Corporation

Re: [RFC] scsi headers, in glibc-dev or linux-libc-headers-dev?

[Tom Rini]

On Tue, 2010-04-27 at 11:21 -0700, Tom Rini wrote:
> On Tue, 2010-04-27 at 12:14 +0200, Koen Kooi wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> > 
> > Hi,
> > 
> > Currently conflicting headers are both in glibc-dev and
> > linux-libc-headers-dev, which isn't a good thing. According to the
> > interwebs we should use the linux-libc-headers version.
> > 
> > Anyone opposed to removing the scsi headers from glibc? It would
> > basically involve reverting
> > http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=b8bb2a62b3916470c752bd79d31322b4358fc676
> > and
> > http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=83f41716ab6a2a9d83d4ff044dcef00595ecfeb2
> 
> So, I did both of those after checking the interwebs and poking some
> kernel folks (see comments in
> http://bugs.openembedded.org/show_bug.cgi?id=4262 ).  The question is,
> did the kernel folks change their mind again and we need to either drop
> older linux-libc-headers (which I'll just assume is a no-go) or go back
> to removing them in new enough linux-libc-headers.

s/in new enough/with new enough/


-- 
Tom Rini <tom_rini@mentor.com>
Mentor Graphics Corporation

Re: [RFC] scsi headers, in glibc-dev or linux-libc-headers-dev?

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 27-04-10 20:21, Tom Rini wrote:
> On Tue, 2010-04-27 at 12:14 +0200, Koen Kooi wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> Hi,
>>
>> Currently conflicting headers are both in glibc-dev and
>> linux-libc-headers-dev, which isn't a good thing. According to the
>> interwebs we should use the linux-libc-headers version.
>>
>> Anyone opposed to removing the scsi headers from glibc? It would
>> basically involve reverting
>> http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=b8bb2a62b3916470c752bd79d31322b4358fc676
>> and
>> http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=83f41716ab6a2a9d83d4ff044dcef00595ecfeb2
> 
> So, I did both of those after checking the interwebs and poking some
> kernel folks (see comments in
> http://bugs.openembedded.org/show_bug.cgi?id=4262 ).  The question is,
> did the kernel folks change their mind again

I would guess so, since l-l-h 2.6.32 still has the headers....

> and we need to either drop
> older linux-libc-headers (which I'll just assume is a no-go) or go back
> to removing them in new enough linux-libc-headers.

I have no real preference, but removing them from l-l-h would be the
easiest option. My immediate problem is that I need to be able to
install both glibc-dev and l-l-h-dev on the target to show off native
qt4 compilation.

regards,

Koen

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFL1zSaMkyGM64RGpERAnFdAJ40UWtXamdRU8oTLd9ILE1BdeApuwCeJo3M
4Cg9iYa3MeFktzS1H4hP+A0=
=x1El
-----END PGP SIGNATURE-----

Re: [RFC] scsi headers, in glibc-dev or linux-libc-headers-dev?

[Tom Rini]

On Tue, 2010-04-27 at 21:01 +0200, Koen Kooi wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 27-04-10 20:21, Tom Rini wrote:
> > On Tue, 2010-04-27 at 12:14 +0200, Koen Kooi wrote:
> >> -----BEGIN PGP SIGNED MESSAGE-----
> >> Hash: SHA1
> >>
> >> Hi,
> >>
> >> Currently conflicting headers are both in glibc-dev and
> >> linux-libc-headers-dev, which isn't a good thing. According to the
> >> interwebs we should use the linux-libc-headers version.
> >>
> >> Anyone opposed to removing the scsi headers from glibc? It would
> >> basically involve reverting
> >> http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=b8bb2a62b3916470c752bd79d31322b4358fc676
> >> and
> >> http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=83f41716ab6a2a9d83d4ff044dcef00595ecfeb2
> > 
> > So, I did both of those after checking the interwebs and poking some
> > kernel folks (see comments in
> > http://bugs.openembedded.org/show_bug.cgi?id=4262 ).  The question is,
> > did the kernel folks change their mind again
> 
> I would guess so, since l-l-h 2.6.32 still has the headers....

So,
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=history;f=include/scsi/Kbuild;h=b3a0ee6b2f1c8bcefec4a219454b824538cc32cf;hb=bc113f151a73cb2195c2fb40d7d70acf8e2f9208 shows that they came back in with 2.6.31.

> > and we need to either drop
> > older linux-libc-headers (which I'll just assume is a no-go) or go back
> > to removing them in new enough linux-libc-headers.
> 
> I have no real preference, but removing them from l-l-h would be the
> easiest option. My immediate problem is that I need to be able to
> install both glibc-dev and l-l-h-dev on the target to show off native
> qt4 compilation.

Checking with the kernel folks I know again and found that yeah, ugg.
Debian currently removes the files from the kernel (which is what we
should do) and I'm going to submit a patch for the kernel and see what
happens.

-- 
Tom Rini <tom_rini@mentor.com>
Mentor Graphics Corporation